A one-step method of 10,17-dihydro(pero)xydocosahexa-4Z,7Z,11E,13Z,15E,19Z-enoic acid synthesis by soybean lipoxygenase.

A product of lipoxygenase (LOX) oxidation of docosahexaenoic acid (DHA), 10,17-dihydro(pero)xydocosahexa-4Z,7Z,11E,13Z,15E,19Z-enoic acid [10,17(S)-diH(P)DHA] was obtained through various reaction pathways that involved DHA, 17(S)-hydro(pero)xydocosahexa-4Z,7Z,11Z,13Z,15E,19Z-enoic acid [17(S)-H(P)DHA], soybean lipoxygenase (sLOX), and potato tuber lipoxygenase (ptLOX) in various combinations. The structure of the product was confirmed by HPLC, ultraviolet (UV) light spectrometry, GC-MS, tandem MS, and NMR spectroscopy. It has been found that 10,17(S)-diH(P)DHA formed by sLOX through direct oxidation of either DHA or 17(S)-H(P)DHA was apparently identical to the product of ptLOX oxidation of the latter. The sLOX- and ptLOX-derived samples of 10,17-diHDHAs coeluted under the conditions of normal, reverse, and chiral phase HPLC analyses, displayed identical UV absorption spectra with maxima at 260, 270, and 280 nm, and had similar one-dimensional and two-dimensional proton NMR spectra. Analysis of their NMR spectra led to the conclusion that 10,17-diHDHA formed by sLOX had solely 11E,13Z,15E configuration of the conjugated triene fragment, which was identical to the previously published structure of its ptLOX-derived counterpart. Based on the cis,trans geometry of the reaction products, the conclusion is made that in the tested conditions sLOX catalyzed direct double dioxygenation of DHA. Compared with the previously described two-enzyme method that involved sLOX and ptLOX, the current simplified one-enzyme procedure uses only sLOX as the catalyst of both dioxygenation steps.

epoxide and its rearrangement in a conjugated triene with 11E,13E,15Z geometry of the double bonds. Thus, the mechanism can be verified by determining the geometry of the double bonds of the final product and/or by checking the role of the hydroperoxy group at C (17) , as its substitution with the hydroxyl group should completely prevent the formation of 16,17-epoxy-DHA, the key intermediate of the rearrangement.
At the same time, direct evaluation of the cis,trans geometry of the sLOX-derived DHA oxidation products has never been performed before. Therefore, the goals of this study were as follows: 1) to design and compare various (chemo)enzymatic methods to propose a simple and scalable biosynthetic procedure of making 10,17(S)-diH(P) DHA; and 2) to validate the stereochemistry of the target product of DHA oxidation.

Materials
The following equipment, reagents, and supplies were used in this study: DHA (Nu-Chek Prep, Inc., Elysian, MN); CD 3 OD (99.8%), PtO 2 , NaBH 4 , and bis(trimethylsilyl)trifluoroacetamide (BSTFA) (Aldrich, Milwaukee, WI); sLOX preparations type I-B and V and monododecyl ether of decaoxyethylene glycol (C 12 E 10 ) (Sigma Chemical Co., St. Louis, MO); octadecyl (C 18 ) solid phase extraction cartridges ( J. T. Baker, Philipsburg, NJ); and type 528-PP NMR tubes (Wilmad Glass Co., Inc., Buena, NJ). A 5980 series II gas chromatograph equipped with a 5971 series electron-impact (EI) mass selective detector was manufactured by Hewlett-Packard. Ultraviolet (UV) light spectra of the reaction mixtures and purified products were recorded on a Beckman DU800 spectrophotometer with a temperature-controlled unit. An Agilent DB-17HT column (30 m 3 0.25 mm column with 0.15 mm polymer layer) was used for GC-MS analysis of the reaction products. Mass spectra of the compounds were obtained with an LCQ Deca XP Max MS n spectrometer (Thermo Electron Corp., San Jose, CA) equipped with an electrospray ionization (ESI) ion source. A Waters Alliance 2695 HPLC separations module equipped with a Waters 2996 diode-array detector was used to analyze and purify the DHA products. Proton NMR spectra were taken on a 400 MHz Varian spectrometer in CD 3 OD at room temperature. 17(S)-HPDHA and 17(S)-HDHA were synthesized and analyzed as described previously (10,12).

Measurement of enzymatic activity
The activity of ptLOX was measured spectrophotometrically by monitoring the product formation at 236 nm (maximum of absorption of conjugated dienes) and 270 nm (conjugated trienes) in a 1 ml reaction mixture composed of 100 mM DHA, 100 mM SDS, and 0.02% C 12 E 10 in 0.05 M sodium phosphate buffer, pH 6.5, at 48C (10,11,(13)(14)(15). Preparative scale synthesis of DHA metabolites with ptLOX and their normal phase HPLC purification were performed according to previously published protocols (10,11). To convert hydroperoxides of DHA to the corresponding hydroxides, the final product mixture was brought to pH z10 with concentrated NaOH and treated with z10 M excess of NaBH 4 (30 min on ice).
To monitor the course of sLOX-catalyzed reactions, the following method was used. A 1 ml aliquot of the reaction mixture (100 mM in 50 mM sodium borate buffer, pH 9) was placed in a temperature-controlled spectrophotometric quartz cuvette, the reaction was initiated by adding the enzyme, and the progress of the reaction was observed by recording sequential absorption spectra (Dt 1-5 min; range, 200-400 nm) at 48C. The conjugated diene products produced the spectra with l max 234-238 nm (e m 23,000 M 21 3 cm 21 ), whereas the conjugated triene(s) gave a characteristic triplet at 260, 270, and 280 nm (e m 40,000 M 21 3 cm 21 at 270 nm, estimated).

Synthesis of the oxygenated derivatives of DHA
For preparative scale synthesis of the sLOX DHA oxidation products, a 50 ml reaction mixture that contained 100 mM solutions of DHA, 17(S)-HPDHA, or 17(S)-HDHA dissolved in 20 mM sodium borate buffer, pH 9.0, was used. The reaction was conducted on ice to minimize chances of the formation of nonspecific oxidation and isomerization products (14,15). Commercially available sLOX preparations [type I-B from Sigma Chemical Co. or sLOX from Fluka (product 62340)] were used throughout the experiments, although in some cases an affinitypurified type V enzyme from Sigma was tested. All of the preparations were shown to be effective as catalysts of 10,17(S)-diH(P)DHA formation. The reactions were initiated by adding a sLOX stock solution (2 mg of sLOX type V, 12.5 mg of sLOX type I-B, or the same quantity of LOX from Fluka dissolved in z1 ml of the same buffer) and were allowed to proceed for z30 min. The indicated amounts of sLOX were determined in preliminary experiments to achieve maximal conversion of DHA in the target product(s). When the target products were hydroxides of DHA, z10 M excess of freshly prepared NaBH 4 solution in the same buffer was added drop-wise to the mixture of the hydroperoxides, and the reaction vessel was placed on ice for 15 min. The unreacted NaBH 4 was then decomposed with an excess of glacial acetic acid (z0.25 ml) added drop-wise (foaming), and the mixture was left on ice until the bubbling stopped. Then, the lipid mixture was slowly loaded on a 500 mg C 18 solid phase extraction cartridge and washed with 25 ml of deionized water, the excess of water from the cartridge was removed by vacuuming, and the product was eluted with 2 ml of absolute ethanol. The solvent was then evaporated to dryness under a stream of N 2 at room temperature; the products were redissolved in nitrogen-flushed ethanol (1 ml) and stored at 2808C in a glass vial with a TeflonR cup. No decomposition of 10,17(S)-diHDHA occurred within at least 2 months of storage. When the hydroperoxides of DHA were to be made, the NaBH 4 treatment step was omitted and the initial product acidified with the same amount of acetic acid was loaded directly onto the extraction cartridge. The hydroperoxides of DHA stored as described above were stable for at least 2 weeks. Slow accumulation of decomposition products was observed upon prolonged storage of the compounds.

HPLC analysis and purification of the products
The DHA oxidation products were separated by normal phase high performance liquid chromatography (NP HPLC) on a Waters m-Bondapak silica gel column (4.6 3 300 mm, 5 mm silica) at 308C in a heptane-2-propanol-acetic acid (949:50:1, v/v/v) mobile phase at a flow rate of 2 ml/min essentially as described previously (10). Analytical separations were conducted on a 5 mm Waters Spherisorb silica gel column (3.2 3 250 mm) either isocratically as described above for the preparative HPLC (the flow rate was reduced to 1 ml/min) or in a hexane-2-propanol-acetic acid gradient mixture as follows. Two solvents were prepared: 989 ml of n-hexane, 10 ml of 2-propanol, and 1 ml of glacial acetic acid (1 liter total; solvent A) and 949 ml of n-hexane, 50 ml of 2-propanol, and 1 ml of acetic acid (1 liter total; solvent B). The flow rate was maintained at 2 ml/min throughout the experiment. The elution profile was monitored spectrophotometrically with the help of the diode-array detector operating in scan mode (210-400 nm). The column was equilibrated at 308C with solvent A until UV light absorbance of the eluent at 236 and 270 nm stabilized, and a sample of the product(s) dissolved in 2-propanol was injected. Then, solvent A was pumped through the column for 3 min, after which a linear gradient from 100% solvent A to a solvent A/solvent B mixture of 50:50 (v/v) was run over the next 10 min. Then, a 5 min linear gradient to 100% solvent B was started, followed by a 5 min isocratic elution with the same solvent. In the next 1 min, the eluent was changed to 100% solvent A, and the column was reequilibrated with 100% solvent A for 6 min. The overall duration of the experiment was 30 min. In both experiments, the fractions that contained target compounds were collected, the solvent was evaporated under a stream of nitrogen, and the individual diH(P)DHAs were stored in nitrogen-saturated ethanol at 2808C.

Characterization of the products
Molecular masses of the DHA oxidation products were determined on an LCQ Deca XP Max MS n mass spectrometer with an ESI ion source operating in either negative (M 2 H 1 , free fatty acids, and/or M 1 Cl 2 adducts) or positive (M 1 H 1 and/or M 1 Na 1 adducts) mode. The following parameters were used in the direct infusion experiments with the samples dissolved in meth-anol: infusion rate of 2-10 ml/min; nitrogen as sheath gas (10 arbitrary units); capillary temperature of 3258C; data collection for at least 1 min at 5 3 5 ms microscans; spray voltage of 5 kV; capillary voltage of 214 V; tube lens offset of 25 V; and scan range of 50-1,000 mass units.
The individual products and/or product mixtures were subjected to EI GC-MS analysis after catalytic hydrogenation with H 2 /PtO 2 and trimethylsilylation with BSTFA (10). Briefly, helium was used to elute the compounds from the DB-17HT capillary column. The following elution program was used. The column was preequilibrated at 1508C. Then, the sample (z2 ml of solution in BSTFA) was injected. The column was washed for 3.5 min at the initial temperature, then the oven temperature was increased at 1.58C/min until it reached 2108C. Total ion chromatograms were recorded with a sampling rate of two per second. Later, single ion monitoring chromatograms were extracted, plotted, and integrated using a Hewlett-Packard Chemstation's built-in routine.

sLOX-catalyzed reactions
In agreement with earlier publications (10, 12), low concentrations of sLOX (5 3 10 28 M to 1 3 10 27 M) caused rapid and virtually quantitative conversion of 10 24 M DHA into 17(S)-HPDHA. There were two peaks detected under the conditions of NP HPLC. The first very hydrophobic compound (product I, Scheme 2) with a short retention time was positively identified as unreacted DHA. The major reaction product II a had the UV light absorbance spectrum of a typical conjugated diene, with l max of 238 nm (Fig. 1A) and molecular weight of 360.3 (molecular formula, C 22 H 32 O 4 ; theoretical isotopic mass, 360.2). The molecular weight of its NaBH 4 -reduced derivative was 344.3 (C 22 H 32 O 3 ; product II b ; theoretical isotopic mass, 344.2) (Fig. 1B). The structure of product II b was elucidated by EI GC-MS and by 1 H-NMR. After catalytic hydrogenation of product II b over PtO 2 in methanol followed by full trimethylsilylation of its hydroxyl and carboxyl groups, the compound produced EI MS fragments with m/z of 73, 173, 429, and 485 that positively identified it as di-trimethylsilyl-17-hydroxydocosanoic acid. A full 400 MHz 1D 1 H-NMR spectrum of product II b is shown in Fig. 1C. The spectrum displayed resonances that were similar to those published earlier for the cis,trans isomers of 9/13-hydroxy linoleyl alcohol and 9/13-hydroxy monolinoleoyl glycerol (14,15). No traces of trans,trans isomers were detected [the latter would have been seen as a quartet with d CH z6.1-6.2 ppm that belongs to the C (3) hydrogen atom of 1-hydroxy-2E,4Epentadiene fragment (11, and references therein)]. A 2D double quantum correlation spectroscopy ( 1 H, 1 H-DQCOSY) scan of product II b (Fig. 1D) along with the 1D 1 H-NMR spectrum allowed us to unambiguously deduce the structure of the compound and assign the observed resonances to particular protons of the compound, except for those that had close or equivalent values of d ( Table 1). Importantly, the protons with d z4. 15-4.17 ppm [believed to be =C (17) H-OH] produced cross-peaks with protons with d z2.3 ppm [=C (18) H 2 ] and 5.75 [=C (16) H-], but not with the protons at C (13) and C (14) . Therefore, product II a was identified as 17-hydroperoxydocosahexa-4Z,7Z,10Z, 13Z,15E,19Z-enoic acid, with the hydroxyl group at C (17) being in the S configuration (10,12).
A dramatically different result was obtained when the concentration of sLOX in the reaction mixture was increased to 0.4 3 10 26 M and greater (Fig. 2). DHA was rapidly converted to a mixture of two major products, III a and IV a , one of which had a UV light absorption spectrum of a typical conjugated triene, with l max of 260.5, 270.0, and 280.3 nm (product IV a ; Fig. 2A), whereas the other (product III a ) showed a split spectrum, with two maxima at 225.6 and 243.4 nm, similar to the spectra of 7,17-diHDHA and 10,20-diHDHA (10). The molecular masses of both products III a and IV a were estimated to be 392.1 Da, as negative mode ESI MS analysis of the products produced strong parent ions with m/z 391.1 and 427.2 (M 2 H 1 and M 1 Cl 2 , correspondingly) (Fig. 2B). These masses are indicative of isobaric compounds with the molecular formula C 22 H 32 O 6 and isotopic mass of 392.2.
After treatment of product IV a with NaBH 4 , its reduction product IV b was isolated by NP HPLC and its structure was determined by ESI MS, GC-MS, and 1D and 2D 1 H-NMR. The molecular mass of product IV b was found to be 360.1 Da (Fig. 3A), consistent with the molecular formula C 22 H 32 O 4 (isotopic mass, 360.2). The fragmenta- tion pattern of its catalytically hydrogenated and fully trimethylsilylated derivative (m/z of 73, 173, 331, 359, 393, 427, 517, and 574) confirmed that it was tri-trimethylsilyl-10,17-dihydroxydocosanoic acid.
The 1D 1 H-NMR and 2D 1 H, 1 H-DQCOSY spectra of product IV b (Fig. 3B, C) revealed that the geometry of the conjugated triene fragment was of the 11E,13Z,15E type (for detailed information on the spectra, see Table 1). The rest of the double bonds of the molecule remained unchanged during the enzyme-catalyzed oxygenation, which was confirmed by the presence of two methylene-interrupted double bonds in its structure, with d 2.84 ppm [=C (6) H 2 , 2H, triplet] and d 5.33-5.39 ppm (-CH=CH-, 6H, multiplet). The spectrum lacked features that would have been present if the product had a trans,trans fragment, being, for example, an 11E,13E,15Z or 11Z,13E,15E isomer. For instance, resonances with d z6.2 ppm, characteristic of a trans,trans conjugated double bond, were not detected. Instead, in a 1 H, 1 H-COSY experiment, it was revealed that protons with d 4.17 ppm (2H, quintet), believed to be C (10) H-OH and C (17) H-OH, produced crosspeaks with proton d 5.73 and 5.75 ppm (triplets). No such cross-peaks with cis protons of the C (13) /C (14) vinyl group (d 5.95-5.99 ppm) were discovered. A full 400 MHz 2D 1 H, 1 H-DQCOSY spectrum of product IV b is presented in Fig. 4. The spectrum was found to be essentially iden- tical to a spectrum of the ptLOX-derived 10,17(S)-diHDHA published previously (11). Interestingly, product IV b gave a single symmetrical peak on a Chiralcel OD-H column with a retention time of 58 min (Fig. 5). A similar product was formed by sLOX from 17(S)-HDHA or 17(S)-HPDHA (data not shown; see Scheme 1). It had the same retention time as product III b in all tested conditions, including chiral, reverse phase, and NP HPLC. Thus, product IV a was identified as 10,17(S)-dihydroperoxydocosahexa-4Z,7Z,11E,13Z,15E,19Z-enoic acid, whereas product IV b was identified as the corresponding hydroxide (Scheme 2). Using the analytical procedures described above, product III a was found to be a DHA derivative with hydroperoxy groups at C (7) and C (17) , similar to the compound described previously (10) as one of the two major products of ptLOX-catalyzed oxygenation of 17(S)-HDHA and 17(S)-HPDHA [a detailed report on 7,17-dihydro(pero) xydocosahexaenoic acid synthesis and characterization will be published elsewhere]. Minor amounts of various H(P)DHAs were detected among the reaction products by NP HPLC (Fig. 2) and in a single ion monitoring GC-MS experiment (Fig. 6). Among those, 17-HDHA was positively identified by its coeluting fragments with m/z 173 and 429, whereas 10-HDHA produced fragments with m/z 271 and 331. The single ion monitoring elution peaks were integrated, and the apparent ratio of 10-HDHA to 17-HDHA of 1:2 was calculated.
The elution peaks with UV light absorbance maxima at 236 6 2 nm that were observed during NP HPLC had retention times similar to those of authentic 10(S)-and 17(S)-HDHAs, although no attempts were made to isolate and further characterize individual monohydroxylated products because of their relatively low abundance.
Interestingly, at high enzyme concentrations, the sLOXcatalyzed reaction of DHA oxidation showed clear signs of autoinactivation/product inhibition. Under the implemented conditions, 1 molecule of sLOX was able to make z97 molecules of 10,17-diH(P)DHA before the reaction stopped as a result of the apparent inactivation/product inhibition of the enzyme.

ptLOX-catalyzed conversion of 17(S)-HDHA and 17(S)-HPDHA
In agreement with our earlier findings, ptLOX converted 17(S)-HDHA and 17(S)-HPDHA to the conjugated triene product V a , whose structure was studied by HPLC, UV light spectrometry, ESI MS, EI GC-MS, and proton NMR as described above for products II-IV. The molecular mass of product V a was found to be identical to that of product IV a (392.3 Da; molecular formula, C 22 H 32 O 6 ). The UV light spectrum of product V a taken in methanol was indistinguishable from the spectrum of product IV a . Based on the EI GC-MS fragmentation of its fully hydrogenated and trimethylsilylated derivative, product V a was identified as 10,17(S)-diHPDHA, with the hydroperoxy group at C (10) being, most likely, in the S configuration as well.
Product V a was then reduced with NaBH 4 to yield product V b . The chromatographic properties of products IV b and V b were compared in various conditions, and the compound appeared to be coeluting in the following HPLC systems: 1) on a silica gel column in several isocratic and gradient hexane-2-propanol-acetic acid mixtures (between 989:10:1 and 949:50:1, v/v/v), as described in Materials and Methods; 2) on a cyanopropyl-silica gel column in a hexane-2-propanol-acetic acid mixture (949:50:1, v/v/v); 3) on a C 18 silica gel column in isocratic and gradient methanol-water-phosphoric acid and acetonitrilewater-phosphoric acid mixtures; and 4) on the Chiralcel OD-H column in a hexane-2-propanol-acetic acid mixture (949:50: A full 1D 1 H-NMR spectrum of the isolated product V b was identical to that of product IV a , confirming our earlier prediction that 10,17(S)-diHDHA formed from 17(S)-H(P)DHA by ptLOX had the 11E,13Z,15E fragment in it. A full 2D 1 H, 1 H-COSY spectrum of product V b also showed no differences from product III b .
Only three types of protons that belonged to the 11E,13Z, 15E fragment were observed in the NMR experiments (Table 1, Figs. 1, 3, 4). This was not surprising considering the highly symmetrical nature of 10,17(S)-HDHA (Scheme 2), with four pairs of equivalent protons at C (10) / C (17) , C (11) /C (16) , C (12) /C (15) , and C (13) /C (14) . The clear absence of trans,trans vinyl protons was also indicative of a specific (i.e., purely enzymatic) mechanism of formation of the products, as any involvement of a free radical chain reaction similar to those described previously (14,15) would have produced measurable quantities of thermodynamically favorable trans,trans or all-trans isomers of the products, which was not the case in the current experiments.
Additional evidence that supports the 11E,13Z,15E arrangement of the conjugated triene came from the fact that the methine protons at C (10) and C (17) (d z4.15-4.17 ppm) produced cross-peaks with trans vinyl protons at C (11) and C (16) (d z5.75 ppm). No such cross-peaks with protons of a cis vinyl group (d z5.97 ppm) were observed. These data unambiguously identify the sLOX product of DHA oxidation as 10,17(S)-diH(P)DHA. Interestingly, this product was invariably formed no matter which synthetic procedure was implemented (Scheme 1). Because there were no visible differences in the NMR spectra of the ptLOXderived 10,17(S)-diHDHA and its sLOX-derived counterpart, the compounds were considered to be identical in all respects except for the possible (S)/(R) stereoisomerism of the hydroxyl group at C (10) [for a discussion of its putative (S)-stereochemistry, see (10,11)]. The exact alignment of DHA and 17(S)-H(P)DHA in sLOX and ptLOX catalytic centers during the formation of 10,17-diH(P) DHA remains to be investigated.  (10) , whereas ions with m/z of 173 reveal compounds with hydroxyls at C (17) .
The finding that 10,17-diH(P)DHA formed through two convenient and easily reproducible chemoenzymatic pathways consistently had the 11E,13Z,15E arrangement of the conjugated triene fragment needs to be taken into account when making this compound and studying its biological properties.