Biochemical Characterization of the Oxygenation of Unsaturated Fatty Acids by the Dioxygenase and Hydroperoxide Isomerase of Pseudomonas aeruginosa 42A2

We have studied oxygenation of fatty acids by cell extract of Pseudomonas aeruginosa 42A2. Oleic acid ((9Z)-18:1) was transformed to (10S)-hydroperoxy-(8E)-octadecenoic acid ((10S)-HPOME) and to (7S,10S)-dihydroxy-(8E)-octadecenoic acid (7,10-DiHOME). Experiments under oxygen-18 showed that 7,10-DiHOME contained oxygen from air and was formed sequentially from (10S)-HPOME by isomerization. (10R)-HPOME was not isomerized. The (10S)-dioxygenase and hydroperoxide isomerase activities co-eluted on ion exchange chromatography and on gel filtration with an apparent molecular size of ∼50 kDa. 16:1n-7, 18:2n-6, and 20:1n-11 were also oxygenated to 7,10-dihydroxy fatty acids, and (8Z)-18:1 was oxygenated to 6,9-dihydroxy-(7E)-octadecenoic acid. A series of fatty acids with the double bond positioned closer to ((6Z)-18:1, (5Z,9Z)-18:2) or more distant from the carboxyl group ((11Z)-, (13Z)-, and (15Z)-18:1) were poor substrates. The oxygenation mechanism was studied with [7S-2H]18:1n-9, [7R-2H]18:2n-6, and [8R-2H]18:2n-6 as substrates. The pro-R hydrogen at C-8 was lost in the biosynthesis of (10S)-HPODE, whereas the pro-S hydrogen was lost and the pro-R hydrogen was retained at C-7 during biosynthesis of the 7,10-dihydroxy metabolites. Analysis of the fatty acid composition of P. aeruginosa revealed relatively large amounts of (9E/Z)-16:1 and (11E/Z)-18:1 and only traces of 18:1n-9. We found that (11Z)-18:1 (vaccenic acid) was transformed to (11S,14S)-dihydroxy-(12E)-octadecenoic acid and to a mixture of 11- and 12-HPOME, possibly due to reverse orientation of (11Z)-18:1 at the active site compared with oleic acid. The reaction mechanism of the hydroperoxide isomerase suggests catalytic similarities to cytochrome P450.

The goal of the present study was to determine the mechanism of biosynthesis of 7,10-DiHOME. Our first objective was to partially purify the diol synthase activity and to determine the apparent molecular weight. Second, the mechanism of biosynthesis of 7,10-dihydroxy metabolites was investigated with (i) stereospecifically deuterated fatty acids or (ii) (10S)-and (10R)-HPOME as substrates and with (iii) experiments under oxygen-18 gas. We also performed a systematic investigation of the substrate specificity and products with a series of regioisomeric 18:1 fatty acids. Finally, we found that P. aeruginosa was rich in vaccenic acid, and we therefore investigated its oxygenation by P. aeruginosa 42A2 by semipurified cell extracts.
Preparation and Purification of Diol Synthase Cell Fraction-P. aeruginosa 42A2 was grown overnight on agar plates (30 g/liter TSA: 15 g of tryptone (digest of casein), 5 g of NaCl, 5 g of soytone, and 15 g of agar per liter) and collected. The cells were centrifuged and suspended in Ringer's solution (OD ϳ2), and 1 ml was transferred to 50 ml of TSB: 17 g of peptone, 3 g of peptone, 2.5 g of glucose, 5 g of NaCl, and 2.5 g of KH 2 PO 4 . Enzyme activity was induced by growing with 1% oleic acid (18 h, 30°C; 250 rpm). The cells were washed twice with Ringer, resuspended in 0.05 M Tris-HCl (pH 7.0; ϩ4°C), and sonicated (three times for 1 min each) on ice (Branson Sonifier B-15). After centrifugation (8000 ϫ g, 5 min, ϩ4°C), the supernatant was filtered and then concentrated and washed by diafiltration (Amicon Ultra-15, Millipore; final volume, 1 ml). An aliquot of this preparation (10 -20 l) was used for most studies of the enzyme activity and is referred to as the semipurified cell extract. The protein was purified by gel filtration (Tricorn Superdex 200 10/300 GL (GE Healthcare)), eluted with 0.05 M Tris-HCl (pH 7.0) at 1 ml/min (0.5 megapascals) in a chromatography system (Ä KTA FPLC, GE Healthcare). Molecular mass standards were BSA (66 kDa), ovalbumin (45 kDa), trypsinogen (24 kDa), and lysozyme (14.3 kDa) from Sigma. The fractions with enzyme activity were combined and purified by chromatography on Mono-Q 5/50 with a gradient in 0.05 M Tris-HCl (pH 7.0) from 0 to 1 M NaCl in 20 min (flow 2 ml/min). An aliquot (10 l) of each fraction (0.5-0.3 ml) was assayed for diol synthase activity using oleic acid as substrate, and the released products were analyzed by TLC (hexane/diethyl ether/acetic acid; 75:15:10). The plates were sprayed with molybdatophosphoric acid and heated to visualize the products.
Bioconversion Assays-Fatty acids (1-5 mM) in 0.05 M Tris-HCl (pH 7.0) were prepared from stock solutions in ethanol, and incubated with 0.02 volume of the semipurified enzyme preparation described above (usually 0.5 mg of protein/ml, Bradford assay (28)) for up to 1 h at 37°C with occasional vortexing. The reaction was terminated by acidification to pH 2-3 with 0.5 M HCl and then extracted with ethyl acetate or on a C 18 cartridge (SepPak/C 18 ). Products were usually analyzed by preparative TLC, followed by LC-MS/MS analysis. The K m of the semipurified fraction was determined in triplicate by measuring the amount of 7,10-diol (LC-MS/MS analysis) formed from 0.7, 1.1, 1.5, and 1.9 mM oleic acid (10 min, ϩ37°C). The effect of CO on the oxygenation of oleic acid was assessed in buffer with dissolved CO and under partial CO atmosphere. The effects of P450 inhibitors (piperonyl butoxide and 1-aminobenzotriazole) at 1 and 5 mM concentrations with oleic acid as a substrate were analyzed by TLC.
LC-and GC-MS Analysis-Reversed phase-HPLC with MS/MS analysis was performed with a Surveyor MS pump (ThermoFisher) and an analytical octadecyl silica column (5 m; 2.0 ϫ 150 mm; Phenomenex), which was usually eluted with methanol/water/acetic acid, 800:200:0.05, at 0.3 ml/min. The effluent was subject to electrospray ionization in an ion trap mass spectrometer (LTQ, ThermoFisher). The heated transfer capillary was set at 315°C, the ion isolation width was set at 1.5 atomic mass units, and the collision energy was set at 25-35 (arbitrary scale). Prostaglandin F1␣ (100 ng/min) was infused for tuning. Products formed from stereospecifically deuterated 18:2n-6 were analyzed as described (5). 6 The deuterium content of the [ 2 H]18:2n-6 was determined by LC-MS/MS analysis (m/z 276 -282 3 full scan) or by GC-MS analysis.
GC-MS analysis and synthesis of fatty acid methyl ester and trimethylsilyl ether derivatives were performed as described (31). Carbon values were estimated from the retention times of saturated fatty acids methyl esters (31). Steric analysis of 11,14-DiHOME was performed after ozonolysis of (Ϫ)-menthoxycarbonyl derivatives as described (32).

Oleate Diol Synthase Activity
Crude or semipurified cell extracts of P. aeruginosa rapidly transformed 18:1n-9 to (7S,10S)-DiHOME and (10S)-HPOME (Fig. 1). (10S)-HOME was also detected along with small amounts of 8-HOME by LC-MS/MS analysis. Preparation and purification of 1-mg amounts of the main enzymatic products in cultures of P. aeruginosa are described in the supplemental material. (10S)-HPOME was a poor substrate in comparison with oleic acid, whereas transformation of (10S)-HOME to the 7,10-diol could not be detected. A comparison of the oxidation of oleic acid and equal amounts of oleic acid plus (10S)-HPOME showed that the latter rather inhibited than augmented the biosynthesis of the 7,10-diol (TLC analysis). This diol synthase activity showed a pH optimum around pH 7.0, and the reaction rate was optimal at 37°C (supplemental Fig. S1) and was linear for at least 10 min (data not shown). The K m was estimated to be 1.7 mM (supplemental material). The formation of DiHOME appeared to be reduced ϳ30 and ϳ50% by 1 and 5 mM 1-aminobenzotriazole (a nonspecific and irreversible P450 inhibitor) (33), respectively, whereas neither CO nor the P450 inhibitor piperonyl butoxide (5 mM) (34) appeared to inhibit the diol synthase activities.

Partial Enzyme Purification
The diol synthase activity was subject to sequential purification by gel filtration and by ion exchange chromatography (Mono-Q), as shown in Fig. 2. SDS-PAGE analysis showed multiple protein bands at this stage; TLC analysis showed that the dioxygenase activity described below was not separated from the hydroperoxide isomerase activity. The elution of the enzyme activity on gel filtration in relation to protein standards (ovalbumin, 45 kDa; bovine serum albumin, 66 kDa) suggested an apparent molecular mass of ϳ50 kDa.

Substrate Specificity
We next examined the substrate specificity of the diol synthase of the semipurified cell extracts. We first studied a series of regioisomeric 18:1 fatty acids. The results are summarized in   Table 1 along with previous data on oleic, palmitoleic, and ricinoleic acids. With the exception of vaccenic acid, the common denominator of substrates was an unsaturated carbon chain from the carboxyl group to the cis double bond at position 9,10, whereas a double bond at 8,9 resulted in slow but significant conversion. Additional double bonds between the carboxyl group and the 9Z double bond inhibited transformation. 7,10-DiHOME showed an unexpected fragmentation during MS/ MS analysis due to fragments with charge migration to the -end; a systematic study of its MS/MS spectrum is reported elsewhere. 7 We also investigated 18:2n-6, 20:1n-9, 20:1n-11, and 20:2n-6 ( Table 1) and found that 18:2n-6 (supplemental Fig. S4) and 20:1n-11 were oxidized, whereas the methyl ester of oleic acid was not a substrate. The results suggested that substrate might bind the active site 'head' first because the distance from the carboxyl group to the first double bond appeared to be critical and not the distance from the -end to the double bond. We found one exception. The oxygenation of vaccenic acid might be due to reverse orientation at the active site (see below).

Oxygenation Mechanism
The Hydroperoxide Isomerase Activity-Experiments performed under oxygen-18 atmosphere with oleic acid as a substrate revealed that both hydroxyl groups of 7,10-DiHOME contained either oxygen-18 or oxygen-16, but we could not detect significant formation of the diol with both isotopes present (Fig. 3). In a separate experiment under oxygen-18, we were able to isolate [ 18 O 2 ](10S)-HPOME (Ͼ95% 18 O 2 ). The latter was transformed by the enzyme to [ 18 O 2 ]7,10-DiHOME (supplemental material). We conclude that 7,10-DiHOME is formed from (10S)-HPODE in a hydroperoxide isomerase reaction.
The (10S)-Dioxygenase Activity-Linoleic acid was oxidized slowly compared with oleic acid. We found that the deuterium label of [(8R)-2 H]18:2n-6 was lost in 10-HODE and also in 7,10dihydroxy-(8E,12Z)-octadecadienoic acid, but we confirmed that the label was retained in the lipoxygenase product   13-HODE of the very same preparation (supplemental Fig. S5). Finally, small amounts of 8-HODE were also detected. Steric analysis by CP-HPLC suggested that 8-HODE was a mixture of the S and R stereoisomers in a ϳ3:7 ratio and could in part be produced enzymatically.
It is known that (9Z)-16:1 is oxidized by the diol synthase of P. aeruginosa (14), but cis vaccenic acid has not been investigated. We found that this fatty acid was oxygenated but to an unexpected metabolite, (11S,14S)-DiHOME. The latter was identified by GC-MS analysis (Fig. 6A) after purification by TLC. The C value was 20.8. The position of the double bond was determined by ozonolysis, and the configuration of the alcohols by separation of (Ϫ)-menthoxycarbonyl derivatives of the biological product and authentic standards on GC-MS (Fig. 6B).

DISCUSSION
We report that P. aeruginosa expresses a fatty acid diol synthase responsible for the sequential transformation of oleic acid to (10S)-HPOME and (7S,10S)-DiHOME. This enzyme also oxygenates, albeit slowly, linoleic acid and some other fatty acids. Transformation of stereospecifically deuterated oleic and linoleic acids suggested that 10S-hydroperoxides were formed by abstraction of the pro-R hydrogen at C-8 and antarafacial oxygenation, whereas the 7,10-diols were formed by abstraction of the pro-S hydrogen at C-7 and suprafacial oxygenation. The reaction mechanism is outlined in Fig. 7.
A systematic investigation of the substrate specificity suggested that the enzyme required a saturated carbon chain from the carboxyl group to the first cis double bond between C-8 and C-9 or between C-9 and C-10, whereas fatty acids with a longer carbon chain ((13Z)-18:1 and (15Z)-18:1) were not oxidized (Table 1). These observations and studies of (9Z)-16:1, (9Z)-20:1, and (11Z)-20:1 suggested that the oxygenated fatty acids bound to the enzyme with their carboxyl groups at a fixed position relative to the catalytic site. We found one exception; (11Z)-18:1 was transformed to (11S,14S)-DiHOME and 11-HPOME, possibly by reverse orientation at the active site ( Fig. 8  and supplemental Fig. S6).
acid and also reduced the rate of oxygenation of oleic acid. As far as is known, hydroperoxide isomerases belong to the family of P450, and suprafacial hydrogen abstraction and oxygenation are typical for this enzyme family (23). The genome of P. aeruginosa PAO1, which is closely related to 42A2 and often used for comparison, contains only two P450 homologues. Whether these two P450s possess hydroperoxide isomerase activity clearly merits further investigation.
The (10S)-DOX activities are of interest in their own right. The genome of P. aeruginosa contains no obvious homologues to heme-containing fatty acid dioxygenases (cyclooxygenases, linoleate diol synthases, 10-DOX, and ␣-DOX), and monounsaturated fatty acids are poor substrates for lipoxygenases (22). Furthermore, partial purification of the (10S)-DOX and hydroperoxide isomerase activities suggested that they were linked. The molecular mass was estimated by gel filtration to be ϳ50 kDa. The observation that (10S)-HPODE was transformed to the 7,10-diol less efficiently than its precursor, oleic acid, suggested that (10S)-DOX and hydroperoxide isomerase activities are functionally coupled. The (10S)-DOX activity probably involves hydrogen abstraction at C-8 by a catalytic metal so that the double bond migrates (9Z 3 8E), and the carbon radical centered at C-10 reacts with molecular oxygen and forms a hydroperoxide in analogy with (10R)-DOX of aspergilli (26). Determination of whether the 7,10-diol synthase consists of one enzyme with both dioxygenase and hydroperoxide isomerase activities or two separate enzymes awaits further studies.
What is the biological function of the diol synthase of P. aeruginosa? Oleic acid is the preferred substrate, and enzyme expression is induced by growing P. aeruginosa in the presence of oleic acid. This bacterium contains little oleic acid compared with palmitoleic and vaccenic acids. This rare prokaryote enzyme may be associated with the modification of the membrane lipids and/or associated with the lipid modification. Alternatively, the combined (10S)-DOX and diol synthase activities could contribute to detoxification of fatty acids in the bacterial environment, promoting colonization and growth. This bioconversion might also contribute to production of oxylipins during growth on plant tissues rich in oleic acid. In this context, the antifungal activity of (7S,10S)-DiHOME is of interest (see Ref. 15 for a review). We conclude that P. aeruginosa expresses a rare diol synthase activity never described before among bacteria.