Manganese Lipoxygenase

Linoleic acid was incubated with manganese lipoxygenase (Mn-LO) from the fungus Gäumannomyces graminis. The product consisted of (13R)-hydroperoxy-(9Z,11E)-octadecadienoic acid ((13R)-HPOD) and a new hydroperoxide, (11S)-hydroperoxy-(9Z,12Z)-octadecadienoic acid ((11S)-HPOD). Incubation of (11R)-[2H]- and (11S)-[2H]linoleic acids with Mn-LO led to the formation of hydroperoxides that largely retained and lost, respectively, the deuterium label. Conversion of the (11S)-deuteriolinoleic acid was accompanied by a primary isotope effect, which manifested itself in a strongly reduced rate of formation of hydroperoxides and in a time-dependent accumulation of deuterium in the unconverted substrate. These experiments indicated that the initial step catalyzed by Mn-LO consisted of abstraction of the pro-S hydrogen of linoleic acid to produce a linoleoyl radical. (11S)-HPOD was converted into (13R)-HPOD upon incubation with Mn-LO. The mechanism of this enzyme-catalyzed hydroperoxide rearrangement was studied in experiments carried out with18O2 gas or18O2-labeled hydroperoxides. Incubation of [11-18O2](11S)-HPOD with Mn-LO led to the formation of (13R)-HPOD, which retained 39–44% of the 18O label, whereas (11S)-HPOD incubated with Mn-LO under 18O2 produced (13R)-HPOD, which had incorporated 57% of 18O. Furthermore, analysis of the isotope content of (11S)-HPOD remaining unconverted in such incubations demonstrated that [11-18O2](11S)-HPOD suffered a time-dependent loss of 18O when exposed to Mn-LO, whereas (11S)-HPOD incorporated 18O when incubated with Mn-LO under 18O2. On the basis of these experiments, it was proposed that the conversion of (11S)-HPOD into (13R)-HPOD occurred in a non-concerted way by deoxygenation into a linoleoyl radical. Subsequent reoxygenation of this intermediate by dioxygen attack at C-13 produced (13R)-HPOD, whereas attack at C-11 regenerated (11S)-HPOD. The hydroperoxide rearrangement occurred by oxygen rebound, although, as demonstrated by the 18O experiments, the oxygen molecule released from (11S)-HPOD exchanged with surrounding molecular oxygen prior to its reincorporation.

O 2 . On the basis of these experiments, it was proposed that the conversion of (11S)-HPOD into (13R)-HPOD occurred in a non-concerted way by deoxygenation into a linoleoyl radical. Subsequent reoxygenation of this intermediate by dioxygen attack at C-13 produced (13R)-HPOD, whereas attack at C-11 regenerated (11S)-HPOD. The hydroperoxide rearrangement occurred by oxygen rebound, although, as demonstrated by the 18 O experiments, the oxygen molecule released from (11S)-HPOD exchanged with surrounding molecular oxygen prior to its reincorporation.
Lipoxygenase-catalyzed dioxygenation of polyunsaturated fatty acids leads to the formation of reactive fatty acid hy-droperoxides. Mammalian lipoxygenases can catalyze oxygenation at carbons 5,8,12, and 15 of their predominant substrate, i.e. arachidonic acid (1). Many plant lipoxygenases can also utilize arachidonic acid, although their most important substrates are the C-18 fatty acids linoleic acid and ␣-linolenic acid (2). Interest in lipoxygenases stems partly from the fact that fatty acid hydroperoxides can be further metabolized into biologically active oxylipins such as leukotrienes and jasmonates.
Lipoxygenases contain ferrous iron, which is oxidized into the ferric state by, e.g., hydroperoxides. The ferric form of lipoxygenases is catalytically active (3) and catalyzes the stereospecific abstraction of one hydrogen from the bis-allylic methylene group of the (1Z,4Z)-pentadiene structure of the substrate as the initial step (4). Attack by dioxygen at one of the terminal positions of the resulting pentadienyl radical results in the formation of a hydroperoxide having one pair of E/Z-conjugated double bonds. Studies of the regio-and stereochemistry of the steps occurring in the oxygenation of 8,11,14eicosatrienoic acid by soybean lipoxygenase-1 revealed that the pro-S hydrogen was stereospecifically removed from C-13 and that dioxygen was regio-and stereospecifically inserted at C-15 to produce a (15S)-hydroperoxide (4). This finding, and results of similar studies carried out with linoleic acid (9S)-lipoxygenase from corn (5), arachidonic acid (12S)-lipoxygenases from human platelets (6) and a red alga (7), arachidonic acid (5S)lipoxygenases from rat basophil leukemia cells and potato (8), and arachidonic acid (8S)-lipoxygenase from mouse epidermis (9) indicated the existence of an antarafacial relationship between hydrogen abstraction and oxygen insertion as a common feature of dioxygenations catalyzed by lipoxygenases. Interestingly, such a steric relationship has also been found for the dioxygenation catalyzed by an "R" lipoxygenase, i.e. (12R)lipoxygenase from sea urchin (10), as well as for dioxygenations catalyzed by prostaglandin endoperoxide synthases I (11) and II (12), and by ferrylmyoglobin (13). Mammalian and plant lipoxygenases so far studied catalyze production of hydroperoxides that have the "S" absolute configuration. In contrast, a number of marine invertebrates, such as starfish, sea urchin, and the coral Plexaura homomalla, express lipoxygenases, which catalyze formation of "R" hydroperoxides, as demonstrated by the configuration of hydroperoxides formed by oxygenation of arachidonic acid at the C-5, C-8, C-11, and C-12 positions. One of these enzymes, arachidonic acid (8R)-lipoxygenase, was recently cloned and sequenced (14).
Chromatographic and Instrumental Methods-RP-HPLC was performed with a column of Nucleosil 100 -5 C 18 (250 ϫ 4.6 mm) purchased from Macherey-Nagel (Dü ren, Germany). The solvent system used consisted of acetonitrile/water/2 M hydrochloric acid (60:40:0.02, v/v/v). Straight phase high performance liquid chromatography was carried out with a column of Nucleosil 50 -5 (200 ϫ 4.6 mm) and a solvent system of 2-propanol/hexane (1:99, v/v). The absorbance (217 nm) and radioactivity of high performance liquid chromatography effluents were determined on-line using a Spectromonitor III ultraviolet detector (Laboratory Data Control, Riviera Beach, FL) and a liquid scintillation counter (IN/US Systems, Tampa, FL), respectively. GLC was performed with a Hewlett-Packard (Avondale, PA) model 5890 gas chromatograph equipped with a methyl silicone capillary column (length, 25 m; film thickness, 0.33 m). Helium at a flow rate of 25 cm/s was used as the carrier gas. Retention times were converted into C-values using standards of saturated fatty acid methyl esters (24). GC-MS was carried out with a Hewlett-Packard model 5970B mass selective detector connected to a Hewlett-Packard model 5890 gas chromatograph. LC-MS was performed as described in the accompanying paper (15). Ultraviolet absorption as a function of wavelength or time was recorded with a Hitachi (Tokyo, Japan) model U-2000 UV-visible spectrophotometer. Infrared spectrometry was carried out using a Perkin-Elmer model 1650 FT-IR spectrophotometer. Radioactivity was determined with a Packard Tri-Carb model 4450 liquid scintillation counter (Packard Instruments, Downer's Grove, IL).

Oxidation of Linoleic Acid by Manganese Lipoxygenase
Isolation of Reaction Products of Linoleic Acid-[1-14 C]Linoleic acid (350 M) was stirred for 30 min with Mn-LO (1.5 g) in buffer A (2.1 ml) at 23°C. Material isolated by extraction with diethyl ether was subjected to RP-HPLC radiochromatography. As seen in Fig. 1, three peaks of radioactivity appeared. The least polar material (29%; 81.1 ml of effluent) was identical to . This result indicated that compound A was a hydroperoxyoctadecadienoic acid (molecular weight, 312). The UV spectrum of compound A was featureless (Fig. 2) demonstrating the absence of, e.g., conjugated double bonds. Treatment of compound A with a small amount of perchloric acid resulted in a rapid appearance of UV absorption bands at 259, 268, and 279 nm (Fig. 2). Analysis of the esterified reaction product by GLC and GC-MS demonstrated the presence of geometrical isomers of methyl 8,10,12-and 9,11,13-octadecatrienoates as judged by their molecular weight (292) and retention times (C-19.00 (21%), C-19.13 (21%), C-19.42 (28%), and C-19.46 (30%)). The fact that a virtually identical mixture of geometrical isomers of methyl 8,10,12-and 9,11,13-octadecatrienoates was formed upon acid treatment of the methyl ester of (11R)-HOD (21) suggested that compound A was a derivative of linoleic acid substituted at the bis-allylic position (C-11). Treatment of compound A with sodium borohydride followed by esterification afforded the methyl ester of 11-HOD as judged by GLC and GC-MS. The retention time corresponded to C-19.41, a value identical to that observed for the Me 3 Si derivative of the methyl ester of authentic 11-HOD but different from the values recorded for . Reduction of compound A with sodium borodeuteride afforded 11-HOD without significant incorporation of deuterium (less than 0.5%), thus excluding the possibility of a 11keto group in compound A. Catalytic hydrogenation of compound A using platinum as catalyst resulted in the formation of a 1:1 mixture of 11-hydroxyoctadecanoate and 11-ketooctadecanoate as shown by GC-MS analysis using the authentic compounds as references. On the basis of these data, compound A was identified as a 11-hydroperoxy derivative of linoleic acid.
Partial hydrogenation of compound A followed by oxidative ozonolysis performed on the MC derivative was used to determine the double bond positions and the configuration of C-11. Analysis of the esterified ozonolysis product by GLC and GC-MS demonstrated the presence of the MC derivatives of methyl (2S)-hydroxynonanoate (less than 2% of the (2R)-hydroxynonanoate; fragment originating in the MC derivative of methyl 11-hydroxy-9-octadecenoate) and of dimethyl (2R)-hydroxydodecane-1,12-dioate (less than 2% of the (2S)-hydroxydodecanedioate; fragment originating in the MC derivative of methyl 11-hydroxy-12-octadecenoate). This experiment thus established that the two double bonds of compound A were localized in the ⌬ 9 and ⌬ 12 positions, and that the absolute configuration of the alcohol group at C-11 was "S." The data presented demonstrated that compound A was identical to (11S)-hydroperoxy-(9Z,12Z)-octadecadienoic acid ((11S)-HPOD). bated with Mn-LO and the product compositions at different times of incubation and varying substrate concentrations were determined by RP-HPLC (cf. Fig. 1). Because of the different extinction coefficients of (13R)-HPOD, (11S)-HPOD, and linoleic acid at the wavelength used for detection (217 nm), the peak areas had to be divided with relative response factors in order to obtain the molar composition of reactions products. These relative response factors, i.e. 0.0732 (linoleic acid), 0.578 ((11S)-HPOD), and 1 ((13R)-HPOD, were calculated by performing repeated injections of mixtures containing the three 14 C-labeled compounds onto the high performance liquid chromatography column and separately determining the peak areas of aborbance at 217 nm and the radioactivity present in the corresponding effluents.
As seen in Fig. 3, when a relatively low concentration of linoleic acid (175 M) was incubated with Mn-LO, (11S)-HPOD was transient in appearance. The maximum level was observed at 10 -20 min of incubation, and at 30 min only a trace amount of the hydroperoxide was detectable. In contrast, when the linoleic acid concentration was increased to 526 M, the amounts of (11S)-HPOD continued to increase during the entire incubation period of 30 min. Interestingly, analyses by RP-HPLC of products from experiments, where (11S)-HPOD had been generated from linoleic acid and then allowed to disappear, did not show any new peak(s) of UV absorption ascribable to the further conversion of (11S)-HPOD. It was conceivable that (11S)-HPOD had been converted into non-UVabsorbing compound(s); however, analysis by RP-HPLC radiochromatography of reaction products generated from [1-14 C]linoleic acid showed the presence of (13R)-HPOD as the single labeled product. These results suggested that (11S)-HPOD was convertible to (13R)-HPOD in the presence of Mn-LO, and that this conversion was slowed down in the presence of high concentrations of linoleic acid.
Incubations in the Presence of Glutathione Peroxidase-Glutathione peroxidase and reduced glutathione catalyze reduction of fatty acid hydroperoxides into hydroxy acids. Trapping of hydroperoxides by this enzyme has been used in previous studies to confirm the existence of fatty acid hydroperoxides as intermediates in the biosynthesis of epoxy-hydroxy acids (25) and dihydroxy acids (26,7). If the Mn-LO-catalyzed formation of (13R)-HPOD from linoleic acid took place by the sequence linoleic acid 3 (11S)-HPOD 3 (13R)-HPOD, inclusion of glu-  1 g) at 23°C. The absorbance at 235 nm was measured using a cuvette having 1 mm path length. Aliquots (0.65 ml) were removed at 6.5, 15, and 30 min of incubation and subjected to RP-HPLC for determination of the molar ratio of (11S)-HPOD/(13R)-HPOD (Q). The amounts of (13R)-HPOD (E) were calculated from the absorbance values using ⑀ ϭ 26,000, and the amounts of (11S)-HPOD (F) were obtained from the amounts of (13R)-HPOD multiplied with Q. A second addition of enzyme (0.5 g) was made, as indicated by arrow. Panel B, same as A, but conducted with 526 M linoleic acid. Aliquots for RP-HPLC analysis were removed at 7, 14, and 30 min. A second addition of enzyme (0.8 g) was made, as indicated by arrow.
tathione peroxidase and glutathione would be expected to result in high yields of 11-HOD, the reduction product of (11S)-HPOD, and in low yields of 13-HOD, the reduction product of (13R)-HPOD. In a typical experiment, a mixture of linoleic acid (224 M), glutathione peroxidase (3 units) and reduced glutathione (3 mM) in 0.8 ml of buffer C was treated with Mn-LO (0.86 g) at 23°C. Spectrophotometric assay of the reaction mixture showed that the rate of formation of 13-H(P)OD was 2.1 nmol min Ϫ1 g Ϫ1 , a rate considerably lower than that observed in a control incubation carried out in the absence of glutathione peroxidase and reduced glutathione, i.e. 9.8 nmol min Ϫ1 g Ϫ1 . This reduced rate of formation of (13R)-HPOD in the presence of glutathione peroxidase was not due to a selective interference with formation of this hydroperoxide but to partial inhibition of linoleic acid oxygenation. Thus, analysis of the reaction product (20 min of incubation) by RP-HPLC showed that unconverted linoleic acid was the main component (75%). The remaining part of the product was due to 13-HOD (21%; formed by reduction of (13R)-HPOD) and a small percentage of 11-HOD (4%; formed by reduction of (11S)-HPOD).
Incubations of (11S)-HPOD with Soybean Lipoxygenase-In order to test the possibility that (11S)-HPOD served as a substrate for soybean lipoxygenase, the hydroperoxide (48 M) in 1 ml of buffer A was treated with soybean lipoxygenase (750 units) at 23°C. No increase in the absorbance at 235 nm could be detected (Fig. 5). In another experiment, soybean lipoxygenase (600 units) in 1 ml of buffer B was treated with (11S)-HPOD (3 or 5 M) for 2 min. Subsequently, linoleic acid (73 M) was added. Measurement of the absorbance at 235 nm showed that the rate of conversion of linoleic acid by soybean lipoxygenase treated with (11S)-HPOD was virtually identical to that of untreated lipoxygenase (Fig. 5).  Table I). In another experiment, (11S)-HPOD was incubated with Mn-LO under 18 O 2 . As seen (incubation 5, Table I), this led to the formation of (13R)-HPOD that had incorporated a significant amount of 18 Table I). Such oxygen exchange was not observed when (13R)-HPOD was treated with Mn-LO (incubation 6, Table I).  1 g) at 23°C. The absorbance at 235 nm was measured using a cuvette having 10 mm path length. The maximum rate of increase of the absorbance at 235 nm was 0.50 absorbance unit/min corresponding to a rate of formation of (13R)-HPOD equal to 18.3 nmol min Ϫ1 g Ϫ1 . B, linoleic acid (53 M) in buffer C (1.05 ml) was treated with Mn-LO (1.1 g) at 23°C and the absorbance was measured as described in A. The maximum rate of increase of the absorbance at 235 nm was 0.33 absorbance unit/min corresponding to a rate of formation of (13R)-HPOD equal to 12.1 nmol min Ϫ1 g Ϫ1 . seen in Table II, hydroperoxides produced from (11R)-[ 2 H]linoleic acid retained most of the deuterium label, whereas hydroperoxides generated from (11S)-[ 2 H]linoleic acid lost most of the label. These results demonstrated that the hydrogen abstracted from the C-11 methylene group by Mn-LO had the pro-S configuration. The isotope contents of linoleic acid remaining unconverted in incubations of (11R)-[ 2 H]-and (11S)-[ 2 H]linoleic acids were also determined. As seen (Table II), incubation of (11S)-[ 2 H]linoleic acid was accompanied by a time-dependent enrichment of deuterium in unconverted linoleic acid. The presence of a kinetic isotope effect in the enzymecatalyzed abstraction of the (11S) deuterium indicated by this experiment also manifested itself in the time course of formation of (13R)-HPOD from (11S)-[ 2 H]linoleic acid measured spectrophotometrically. As seen in Fig. 6, production of (13R)-HPOD from (11S)-[ 2 H]linoleic acid (32.6% deuterated and 67.4% undeuterated molecules) occurred by a biphasic time course. This was explainable if it is assumed that undeuterated linoleic acid present in the mixture incubated was mainly oxygenated in the early phase of the incubation (segment A in Fig. 6), whereas the gradually accumulating deuterated substrate was oxygenated at a much slower rate during the later phase (segments B and C in Fig. 6). Estimates of the rates of conversion of undeuterated and deuterated molecules could be made from the slopes of segments A-C, i.e. 7.9 nmol min Ϫ1 g Ϫ1 for segment A (roughly corresponding to oxygenation of undeuterated linoleic acid), and 0.4 -0.5 nmol min Ϫ1 g Ϫ1 for segments B and C (roughly corresponding to oxygenation of deuterated linoleic acid). The magnitude of the kinetic isotope effect estimated from these rates was k H /k D ϭ 15-22. The reported value of the isotope effect in the soybean lipoxygenasecatalyzed oxygenation of linoleic acid dideuterated at C-11 (94 -95% dideuterated molecules) is k H /k D ϭ 8 -9 (27,28); however, much larger values have been reported recently (29 -31).

DISCUSSION
Mn-LO purified from the fungus G. graminis as described in the accompanying paper (15) catalyzes conversion of linoleic acid into (13R)-HPOD as the major product. This transformation consisted of dioxygenation of a fatty acid possessing a (1Z,4Z)-pentadiene moiety into a fatty acid hydroperoxide having a 1-hydroperoxy-(2E,4Z)-pentadiene structure, thus satisfying the requirements for classifying Mn-LO as a lipoxygenase enzyme. The aim of the present study was originally to deter-mine the stereochemistry of the biosynthesis of (13R)-HPOD using stereospecifically deuterated linoleic acids. In the course of this work, it became apparent that Mn-LO not only catalyzes transformation of linoleic acid into (13R)-HPOD but also other reactions, which are not characteristic of traditional lipoxygenases.
Experiments with linoleic acids labeled with 2 H in the (11R) and (11S) positions demonstrated that Mn-LO, like soybean lipoxygenase-1, catalyzed abstraction of the pro-S hydrogen from C-11 of linoleic acid (Table II). As was found previously for soybean lipoxygenase (4, 27-31), a pronounced primary isotope effect resulting in accumulation of 2 H in the unconverted substrate was noted in these experiments (Table II, Fig. 6). This result indicated that the first step of the Mn-LO-catalyzed oxygenation, like that of the soybean lipoxygenase-catalyzed oxygenation, consisted of hydrogen abstraction from the bisallylic methylene group to produce a pentadienyl moiety. The overall steric course of formation of (13R)-HPOD from linoleic acid in the presence of Mn-LO consisted of hydrogen abstraction and oxygen insertion occurring in a suprafacial way (Fig.  7). This was in contrast to the antarafacial stereochemistry repeatedly observed for oxygenations catalyzed by soybean lipoxygenase and other lipoxygenases (4 -10).
Analysis of the hydroperoxide product isolated following incubation of linoleic acid with Mn-LO demonstrated the presence of a second, less abundant component in addition to (13R) -HPOD (compound A, Fig. 1). The structure of this compound was determined by ultraviolet spectroscopy and mass spectrometry and by chemical methods. An important clue to the structure was provided by the finding that the hydroperoxy group of compound A was rapidly eliminated upon acid treatment to provide a mixture of 8,10,12-and 9,11,13-octadecatrienoates (Fig. 2). This type of conversion had earlier been observed with 11-HOD (21,22), thus indicating that compound A was a bis-allylic hydroperoxide. Reduction of compound A by treatment with sodium borohydride afforded 11-HOD. The geometry of the two double bonds of this product was rigorously established as Z,Z by comparison with authentic 11-HOD and with chemically prepared (9E,12Z)-, (9Z,12E)-, and (9E,12E)-11-HOD. Partial hydrogenation of the methyl ester of 11-HOD derived from compound A followed by treatment with (Ϫ)menthoxycarbonyl chloride and oxidative ozonolysis resulted in the formation of chiral fragments whose structures localized the double bonds of compound A to the ⌬ 9 and ⌬ 12 positions and demonstrated that the absolute configuration of C-11 was "S." The experiments thus allowed compound A to be formulated as (11S)-hydroperoxy-(9Z,12Z)-octadecadienoic acid ((11S)-HPOD). The estimated rates of formation of (13R)-HPOD in time segments A, B, and C were 7.9, 0.5, and 0.4 nmol min Ϫ1 g Ϫ1 , respectively. a Incubations were performed at 23°C using buffer A. b Products were isolated by RP-HPLC. Recovered linoleic acid was treated with diazomethane, whereas hydroperoxides were reduced, esterified, and converted into their Me 3 Si derivatives. The isotope content was determined by GC-MS operated in the selected ion monitoring mode.
c Numbers within parentheses indicate the percentage enrichment of 2 H relative to the deuterated linoleic acid incubated. In the case of (11S)-[ 2 H]linoleic acid, the upper limit of relative enrichment is 100 ϫ (100/32.6) ϭ 307%.
The proportion between (11S)-HPOD and (13R)-HPOD depended on the time of incubation and the substrate concentration (Fig. 3). Short times of incubation of high substrate concentrations gave the highest yields of (11S)-HPOD and a ratio (11S)-HPOD/(13R)-HPOD equal to 0.31. As shown by RP-HPLC radiochromatographic analysis, the only product present in incubations where (11S)-HPOD had been allowed to disappear was (13R)-HPOD, thus suggesting that (11S)-HPOD was converted into (13R)-HPOD in the presence of Mn-LO. This could be directly demonstrated in experiments where (11S)-HPOD was incubated with Mn-LO and the ultraviolet absorbance due to (13R)-HPOD was monitored versus time. The time course of formation of (13R)-HPOD showed a characteristic sigmoidal shape because of the presence of a lag phase (Fig. 4). The maximum rate of formation of (13R)-HPOD from (11S)-HPOD was ϳ50% greater than that of formation of (13R)-HPOD from linoleic acid. Conversion of (11S)-HPOD into (13R)-HPOD was suppressed in the presence of high concentrations of linoleic acid (Fig. 3), indicating that (11S)-HPOD and linoleic acid competed for the same catalytic site of Mn-LO. The fact that (11S)-HPOD was converted into (13R)-HPOD in the presence of Mn-LO raised the question whether (11S)-HPOD served as an obligatory intermediate in the formation of (13R)-HPOD. The enzyme glutathione peroxidase, which traps fatty acid hydroperoxides as the corresponding hydroxy compounds, has been successfully utilized to prove the existence of hydroperoxides as intermediates in the biosynthesis of various oxylipins (7,25,26). Inclusion of glutathione peroxidase and reduced glutathione in incubations of linoleic acid with Mn-LO resulted in a decreased rate of formation of (13R)-H(P)OD; however, this was not caused by trapping of (11S)-HPOD but by partial inhibition of oxygenation of linoleic acid. Importantly, the ratio of the reduction products, (11S)-HOD/(13R)-HOD, did not increase in the presence of glutathione peroxidase, thus disfavoring the hypothetical sequence linoleic acid 3 (11S)-HPOD 3 (13R)-HPOD.
The mechanism of the conversion of (11S)-HPOD into (13R)-HPOD was studied by 18 O experiments (Table I). In one set of incubations, linoleic acid was treated with Mn-LO under 18 18 O 2 , showed that the oxygen molecule migrating from C-11 to C-13 during the hydroperoxide rearrangement was subject to exchange with surrounding molecular oxygen. This fact, in turn, necessitated that the conversion of (11S)-HPOD into (13R)-HPOD took place in a stepwise way involving a deoxygenated intermediate. Interestingly, the oxygen exchange also manifested itself in the isotope content of (11S)-HPOD remaining not converted. As seen in Table I, there was a time-dependent loss of 18  In the initial step, Mn-LO abstracts the pro-S hydrogen from C-11 of linoleic acid. The resulting linoleoyl radical is reversibly oxygenated at C-11 to produce an (11S)-peroxy radical that can be further converted into (11S)-HPOD. Alternatively, oxygen attack at the C-13 position of the linoleoyl radical results in irreversible formation of (13R)-HPOD. Free radical conversions are shown, although it is conceivable that heterolytic reactions may partly be involved (cf. Ref. 37). R 1 , (CH 2 ) 7 -COOH; R 2 , (CH 2 ) 4 -CH 3 . dicated that this intermediate could be reversibly oxygenated into (11S)-HPOD. Fig. 7 shows the mechanism proposed for the Mn-LO-catalyzed oxygenation of linoleic acid. The initial step consisted of abstraction of the pro-S hydrogen from C-11 to produce a linoleoyl radical. This intermediate was reversibly converted into (11S)-HPOD via the corresponding (11S)-peroxy radical, or irreversibly converted into (13R)-HPOD via the corresponding (13R)-peroxy radical. Chemical studies on fatty acid autoxidation have demonstrated that conversion of carbon-centered fatty acid radicals into peroxy radicals occurs by reversible binding of dioxygen (35). Furthermore, non-enzymatic free radical rearrangements of fatty acid hydroperoxides have been described (33,36). In a study of rearrangement of hydroperoxides derived from linoleic acid (36), the methyl ester of (9S)-HPOD (incorrectly referred to as (9R)-HPOD in Ref. 36) was found to yield a mixture of the E,Z-and E,E-isomers of 9and 13-HPOD methyl esters when kept in hexane solution under O 2 . When the reaction was performed under 18 O 2 , partial incorporation of 18 O into the hydroperoxides was observed. Although the 18 O experiments of the present study demonstrated the existence of a non-concerted pathway involving a deoxygenated intermediate for the transformation of (11S)-HPOD into (13R)-HPOD, they did not exclude the possibility of an additional mechanism contributing to the formation of (13R)-HPOD, i.e. a direct conversion of the (11S)-peroxy radical into the (13R)-peroxy radical by concerted transfer of the peroxy radical oxygen from C-11 to C-13. Studies of the viability of such a contributing pathway are under way.
It is uncertain whether the mechanism proposed for Mn-LO involving a bis-allylic peroxy radical and a bis-allylic hydroperoxide has any relevance for other lipoxygenases such as soybean lipoxygenase. 11-Hydroperoxyoctadec-12-en-9-ynoic acid has been isolated as one of several products formed from an acetylenic inhibitor, octadec-(12Z)-en-9-ynoic acid, upon incubation with soybean lipoxygenase (34). However, despite extensive studies of soybean lipoxygenase-catalyzed oxygenations, formation of bis-allylic hydroperoxides from polyunsaturated fatty acids has never been reported. In the present study, (11S)-HPOD was tested as a substrate for soybean lipoxygenase with negative result. The possible roles of the corresponding (11S)-peroxy radical, and of (11R)-HPOD or its corresponding peroxy radical, as intermediates in soybean lipoxygenase catalysis remains to be examined.