Mechanism of linoleic acid hydroperoxide reaction with ... - Springer Link

ARTICLE

Mechanism of Linoleic Acid Hydroperoxide Reaction with Alkali Harold W. Gardner a'*, Thomas D. Simpson a, and Mats Hamberg b aNational Center for Agricultural Utilization Research,ARS, USDA, Peoria, Illinois 61604 and bDepartment of Medical Biochemistry and Biophysics, Karolinska Institutet, S-171 77 Stockholm, Sweden

ABSTRACT: Treatment of (13S,9Z,11E)-I 3-hydroperoxy-9,11octadecadienoic acid (13S-HPODE) with strong alkali resulted in the formation of about 75% of the corresponding hydroxy acid, (13S,9Z,11E)-13-hydroxy-9,11-octadecadienoic acid (13S-HODE), and the remaining 25% of products was a mixture of several oxidized fatty acids, the majority of which was formed from (9Z,11 R,S,12S, R)-13-oxo-11,12-epoxy-9-octadecenoic acid by Favorskii rearrangement (Gardner, H.W., et al. (1993) Lipids 28, 487-495). In the present work, isotope experiments were completed in order to get further information about the initial steps of the alkali-promoted decomposition of 13S-HPODE. 1. Reaction of [hydroperoxy-1802113S-HPODE with 5 M KOH resulted in the formation of [hydroxy-180]13S HODE and [epoxy-180](9Z,11 R,S,12 S,R)-I 3-oxo-11,12-epoxy9-octadecenoic acid; 2. treatment of a mixture of [UJ4C] 13SHODE and [hydroperoxy-1802113S-HPODE with KOH and analysis of the reaction product by radio-TLC showed that 13SHODE was stable under the reaction conditions and did not serve as precursor of other products; 3. reaction of a mixture of [U-14C]13-oxo-9,11-octadecadienoic acid (13-OODE) and [hydroperoxy-1802113S-HPODE with KOH resulted in the formation of [U-~4C-epoxy-180](9Z,11 R,S,12S, R)-13-oxo-11,12epoxy-9-octadecenoic acid; 4. treatment of a mixture of [hydroperoxy -1802] 13S-HPODE and [carboxy1-1801 ] 13 S-H PODE with KOH afforded (9Z,11R,S, 12S, R)-13-oxo-11,12-epoxy-9octadecenoic acid having an 180-labeling pattern which was in agreement with its formation by intermolecular epoxidation. It was concluded that (9Z,11R,S,12 S,R)-I 3-oxo-11,12-epoxy-9octadecenoic acid is formed from 13S-HPODE by a sequence involving initial dehydration into the 0~,[3-unsaturated ketone, 13-OODE, followed by epoxidation of the A 11 double bond of this compound by the peroxyl anion of a second molecule of 13S-HPODE. Rapid conversion of hydroperoxides by alkali appeared to require the presence of an ~,~-unsaturated ketone intermediate as an oxygen acceptor. This was supported by experiments with a saturated hydroperoxide, methyl 12-hydroper-

*To whom correspondence should be addressed at the National Center for Agricultural UtilizationResearch, ARS, USDA, Peoria, IL 61604. Abbreviations: ANS, 8-anilino-l-naphthalenesulfonic acid; GC-MS, gas chromatography-mass spectroscopy; HPLC, high-performance liquid chromatography; 13S-HODE, (13S,9Z,11E)-13-hydroxy-9,1l-octadecadienoic acid; 12-HPO,methyl 12-hydroperoxyoctadecanoate;13S-HPODE, (13S,9Z,11E)-13-hydroperoxy-9,11-octadecadienoic acid; 13-OODE, (9Z,11E)13-oxo-9,11-octadecadienoicacid; OTMS, trimethylsilyloxy;TLC, thin-layer chromatography;TMS, trimethylsilyl;UV, ultraviolet. Copyright 9 1996 by AOCS Press

oxyoctadecanoate, which was found to be much more resistant to alkali-promoted conversion than 13S-HPODE. Lipids 31, 1023-1028 (1996).

Frankel et al. (l) reported conversion by alkali of methyl linoleate hydroperoxides to their corresponding hydroxy fatty acids without explanation of possible mechanisms of transformation. Recently, this alkali transformation was investigated in more detail showing that (13S,9Z,1 IE)-13-hydroperoxy-9,11-octadecadienoic acid (13S-HPODE) was converted to ( 13S,9Z, 11E)- 13-hydroxy-9,1 l-octadecadienoic acid ( 13SHODE) in about 75% yield without loss of regio- and stereoconfiguration (2,3). The majority of remaining 25% of products were proposed (3) to arise from Favorskii Rearrangement of intermediate (9Z, 11R,S, 12S,R)- 13-oxo- 11,12-epoxy-9-octadecenoic acid. O'Brien (4), who reacted a number of "nucleophiles" with linoleic acid hydroperoxide obtained a similar conversion into hydroxyoctadecadienoic acid. Although O'Brien did not test alkali, he suggested that nucleophiles displaced the distal OH from the hydroperoxide group, and this seemed to be a plausible mechanism with the alkali transformation. However, it is difficult to reconcile how the nucleophilic hydroxyl anion could displace the distal hydroperoxide oxygen if the hydroperoxide group exists as a peroxyl anion in strong alkali. In this communication we investigate the mechanism of alkali transformation of 13S-HPODE and compare its reaction with some other hydroperoxides. We show by labeling experiments that hydroperoxide oxygen is consumed by intermolecular epoxidation of a dehydration by-product, (9Z, l 1E) ! 3-oxo-9,11-octadecadienoic acid (13-OODE). Evidence is presented to question the possibility of a nucleophilic displacement of the distal hydroperoxide oxygen.

MATERIALS AND METHODS Materials. Bovine liver catalase 27,000 units/rag was from

Sigma (St. Louis, MO). [U-14C] linoleic acid was purchased from Amersham (Buckinghamshire, United Kingdom). 1802 of 99.4 atom % Igo was from Isotec, Inc. (Miamisburg, OH). [Carboxyl- 18O~] 13S-HPODE (5) and [hydroperoxy-1802]-

1023

Lipids, Vol. 31, no. 10 (1996)

1024

H.W. GARDNER ETAL.

13S-HPODE (6) were prepared as described previously. Methyl 12-hydroperoxyoctadecanoate (12-HPO) was synthesized by the method of Bascetta and Gunstone (7) from methyl 12-hydroxyoctadecanoate from Nu-Chek-Prep (Elysian, MN). 13-OODE was prepared by oxidation of 13SHODE with a CrO3-pyridine complex (8). Reaction conditions. 13S-HPOD (2.3 mM) was treated with 5 M KOH at 35~ and products were recovered as described previously (3). In certain KOH (always 5 M) treatments, the concentrations of some of the fatty acids or other reactants varied. Reaction of KOH was tested with (i) 0.75 mM uniformly 14C-labeled [U-HC]I3S-HODE and 5.0 mM [hydroperoxy-tSOz ]I3S-HPODE; (ii) 1.0 mM [carboxyiJ8OI]I3S-HPODE and !.3 mM [hydroperoxy-lSOz]13S HPODE; (iii) 0.6 mM [U-14C]13-OODE and 2.3 mM [hydroperoxy-18Oz]13S-HPODE; (iv) 3.3 mM 12-HPO; and (v) 2.3 mM 12-HPO plus 2.3 mM 13S-HPODE. Other methods. Other methods were essentially as described previously with few exceptions (3). Gas chromatography-mass speCtrometry (GC-MS) was accomplished with the same equipment, except selected ion monitoring was utilized to determine isotopic compositions. Developing solvents for thin-layer chromatography (TLC) of methyl esterifled products were hexane/ethyl acetate (8:2, vol/vol) and hexane/acetone (92:8, vol/vol). Products of 12-HPO were analyzed as methyl esters by high-performance liquid chromatography (HPLC) with a Microsorb silica column (250 x 4.6 mm, 5 la spherical particle size; Rainin, Woburn, MA) using a Spectra-Physics (San Jose, CA) SP8800 pump. The isocratic solvent used for 12-HPO products was hexane/isopropanol (191:1, voi/vol) at a flow rate of 0.75 mL/min, and eluants were detected with a Varex (Rockville, MD) ELSD IIA detector (evaporative light-scattering detector) using a detector temperature of 73~ and a gas flow meter setting of 40 mm. Methyl 13S-HPODE and methyl 13S-HODE were separated with hexane/isopropanol (397:3, vol/vol) solvent at a flow rate of 0.6 mL/min. Detection was absorbance at 235 nm with a Spectra-Physics SP8490 detector. 1420 z measurements. 13S-HPODE (385 mM) was incubated with 5 M KOH at 35~ and 1 mL aliquots were taken at various times. The solution was adjusted to pH 4 with 2 M citric acid, and the aqueous layer was extracted with an equal volume of diethyl ether followed by hexane extraction to remove fatty acids. The extracted fatty acids were saved for analyses of the 13S-HODE/I 3S-HPODE ratios. The aqueous layer, containing HzO 2, was adjusted to pH 6 (3.6 mL total), and a 2.4 mL portion was transferred to an Oz-electrode cell (Gils0n Oxygraph, model 5/6H; Middleton, W1). O z evolution was measured after addition of 10 laL catalase solution [ 1 mg bovine liver catalase (27,000 units/mg from Sigma) per mL HzO]. For determining the peroxide value of the aqueous layer, the same procedure as above was used, except the acidified reaction solution instead was extracted three times with a threefold volume of CHCl3/methanol (2:1, vol/vol) to ensure complete removal of fatty acids in the CHCI 3 layer. The per-

Lipids, Vol. 31, no. 10 (1996)

oxide value of the HzO-methanol layer was determined as described previously (9), except the thiosulfate used in titration was 1 • -4N. RESULTS A N D

DISCUSSION

Putative nucleophilic mechanism. One possible mechanism of fatty acid hydroperoxide conversion to fatty acid hydroxides by alkali would be nucleophilic attack at the distal OH of the hydroperoxide group affording fatty acid hydroxyl anion and HzO 2. This mechanism was proposed by O'Brien (4) for the conversion of hydroperoxide fatty acid into its corresponding hydroxy fatty acid by a number of nucleophiles other than alkali. Previously, our laboratory could not detect a significant amount of HzO z after 3 h of reaction (2). As shown in Table 1, these experiments were extended by (i) using higher concentrations of 13S-HPODE (385 mM), (ii) measuring HzO z at a number of shorter times of reaction, and (iii) using two methods of HzO 2 determination. Under these conditions, HzO 2 could be detected, but the values were always fairly insignificant. During the first 30 min of reaction, a range of only 0.03~0.33% yield of HzO 2 based on total 13S-HPODE was determined. Either HzO 2 was a very transient intermediate or it was not formed in significant amounts. Considering alkali as a nucleophilic reagent, a severe constraint would appear to be presented by ionization of the hydroperoxide group to peroxyl anion under these very alkaline conditions, thus preventing attack by hydroxyl anion to form HzO z. The acid dissociation constant of the perhydroxyl (hydroperoxide) radical has a pK of 4.88 (10), which should not be much different for organic hydroperoxides. In addition, it was previously shown (2) that O z evolution could not be detected during the course of reaction as measured by a War-

TABLE 1 Hydrogen Peroxide Produced During Treatment of (13S,9Z, 11 E)-I 3-hydroperoxy-9,11 -octadecadienoic Acid (13S-HPODE) with 5 N KOH a Reaction time (min) 1 15 30 45 60

H2O 2 generated (mmol/mmol 13S-HPODE) O2-Electrode b 0.0012 0.0016 0.0009 n.d. 0

Titration c

13S-HODE/13S-HPODE ratio d

0.0033 0.0015 0.0003 trace n.d.

1:12 1:3.8 1:1.6 n.d. 1:0.4

al 3S-HPODE (385 mM) was reacted with 5 N KOH terminating the reaction at the times indicated by acidifying to pH 4, extracting the fatty acids, and then analyzing H202in the aqueous layer as described in the Materials and Methods section; n.d., not determined. bH202 was measured by determination of the 02 released by O2-electrode after treatment with catalase. r was measured by determination of peroxide value of the extracted aqueous layer by thiosulfate titration of 12 released after treatment with acidic KI. dThe 13S-HODE/13S-HPODE ratio of extracted fatty acids was determined by high-performance liquid chromatography separation using ultraviolet detection at 232 nm. 13S-HODE, (13S,9Z,11E)-I 3-hydroxy-9,11 -octadecenoic acid.

REACTION OF HYDROPEROXIDES WITH ALKALI

burg respirometer (a small amount of O 2 was consumed, 11% on a per mole basis). The lack of O 2 evolution was confirmed in this study by the inability to detect 1802 from the headspace of KOH treated [hydroperoxy-1802]13S-HPODE by mass spectral analysis (data not shown). These data together suggest that nucleophilic displacement or 02 elimination are probably not significant mechanisms. Disposition of peroxide oxygen. A reaction mixture of [U-laC] 13S-HODE (0.75 mM) and [hydroperoxy-1802] 13SHPODE (5 mM) in 5 M KOH at 35~ was sampled at 30, 60, and 90 rain. The methyl esterified products were separated by TLC with hexane/ethyl acetate (4:1, vol/vol) and radioscanned. Samples obtained at each time showed a single radiolabeled peak due to the methyl ester of [U-laC]I 3SHODE. Spraying the plate with 8-anitino-l-naphthalenesulfonic acid (ANS) showed methyl esterified (9Z, I 1R,S, 12S,R)13-oxo- 11,12-epoxy-9-octadecenoic acid and other products unlabeled with 14C. This experiment showed that, once formed from 13S-HPODE, 13S-HODE is remarkably stable in alkali and that (9Z, ! I R,S, 12S,R)- ! 3-oxo- 11,12-epoxy-9octadecenoic acid and other by-products do not originate from 13S-HODE. Reaction of [hydroperoxy- 1802] 13S-HPODE (98+% isotopic purity) with 5 M KOH showed that product 13S-HODE retained one 180 as expected as determined by the mass spectrum of its methyl ester/! 3-trimethylsilyioxy (OTMS) derivative: [m/z (percentage relative intensity, ion structure)] 384 (22, M+); 313 (50, [M - CH3(CH2)4]+); 227 (23, [M (CH2)TCOOCH3]+); 73 [ 100, trimethylsilyl (TMS+)]. Of significance to the mechanism, there was 180 transfer to the other products. The important product (9Z, 11R,S, 12S,R)- 13oxo-1 I, 12-epoxy-9-octadecenoic acid, intermediate to other compounds, was J80-labeled at the 11,12-epoxide. The 13ketone was completely devoid of ]so-label, presumably due to exchange with water. Compared with the same unlabeled fatty acid methyl ester (3), the important features of its mass spectrum were: 326 (28, M+); 306 (28, [M - H2180]+); 295 (22, [M - CH30]+); 266 (12); 223 (21, [M - CH3(CH2) 4 CH3OH]+); 178 (61); 169 (43, [M - (CH2)7COOCH3]+); 140 (53); 99 (86, CH3(CH2)4CO+). The 180-labeling of Favorskii Rearrangement products derived from (9Z, 1 ] R,S, 12S,R)- 13oxo- I 1,12-epoxy-9-octadecenoic acid, that is, 2-butyl-4-hydroxy-5-tetradecendioic acid and 2-butyl-3,5-tetradecadiendioic acid, were also examined by GC-MS. The mass spectrum of 2-butyl-3,5-tetradecadiendioic acid, dimethyl ester, was indistinguishable from the unlabeled compound previously reported (3) in agreement with its origin from the oxoepoxyene fatty acid with loss of the 1so-labeled I 1,12-epoxide. There are two isomers of product 2-butyl-4-hydroxy-5tetradecendioic acid. One isomer (determined as its dimethyl ester/OTMS derivative) contained one 1so-label as indicated by its m/z ion 398 (6, [M - CH3OH]+). Compared with its previously determined spectrum (3), the label was certainly not located on the most highly substituted carboxylate group as indicated by its m/z 371 (9, [M - COOCH3]+), but it was probably located at the 4-hydroxyl as indicated by its m/z 287

1025

(100, TMS 18OCH-CH=CH(CH2)TCOOCH3+). This result is in agreement with its origin from (9Z, 11R,S, 12S,R)-13-oxo11,12-epoxy-9-octadecenoic acid with retention of the epoxide oxygen as a 4-hydroxyl. In contrast, the other isomer of 2-butyi-4-hydroxy-5-tetradecendioic acid had lost much of its 180-label as indicated by its m/z 285/287 ratio of 1:0.65. Evidently, this latter isomer may be more susceptible to forming the lactone observed previously (3), thus permitting exchange of oxygen. Also, the product, 11-hydroxy-9,12-heptadecadienoic acid (3), was labeled with one ]80. Prominent features of its mass spectrum (methyl ester/OTMS derivative) were: 370 (23, M+); 313 (25, [M -CH3(CH2)3]§ 213 (46, [M - (CH2)TCOOCH3]+); 73 (100, TMS+).

Precursor to formation of (9Z,11R,S,12S,R)-13-oxo11,12-epoxy-9-octadecenoic acid. 13-OODE was a consistent, but low-yield, product in reactions of 13S-HPODE with alkali (3). It was postulated that this ketodiene fatty acid could be generated from 13S-HPODE by dehydration and subsequently act as an acceptor of peroxide oxygen to form (9Z, 11R,S, 12S,R)- 13-oxo- 11,12-epoxy-9-octadecenoic acid. This hypothesis was tested by reacting [U-14C] 13-OODE (0.6 mM) and [hydroperoxy-1802]13S-HPODE (2.3 mM) with 5 M KOH at 35~ taking aliquot samples at 1 and 40 min. The esterified products were separated by TLC with hexane/acetone (92:8, vol/vol) solvent. Radioscanning of the 1-min sample showed one radiolabeled band due to unconverted [U-14C] 13-OODE. After spraying with ANS spray no fluorescent band due to the oxoepoxyene fatty ester was visible under ultraviolet (UV) light, but unconverted 13-OODE could be seen as a UV absorbing band. The 40-min products separated into a strongly radiolabeled band due to methyl (9Z, 11R,S, 12S,R)- 13-oxo- 11,12-epoxy-9-octadecenoate (48,600 c.p.m.) and a faintly labeled band due to [U-14C] 13OODE (2,944 c.p.m.). ANS spray and UV viewing revealed the two bands as the oxoepoxyene fatty ester (fluorescent) and 13-OODE (UV absorbing). GC-MS analysis showed the same prominent ions as described above for (9Z, 11R,S, 12S,R)- 13-oxo- l 1,12-epoxy-9-octadecenoic acid derived from [hydroperoxy-180~] 13S-HPODE alone showing that it was labeled with one 18O- located at the ll,12-epoxide. This oxoepoxyene fatty acid was also generated from 13OODE simply by incubation with H202 and 5 M KOH (data not shown). Therefore, (9Z,11R,S,.12S,R)-13-oxo-ll,12epoxy-9-octadecenoic acid (OEOE in equation below) can be formed by epoxidation of 13-OODE in the presence of either 13S-HPODE or H202 and 5 M KOH as indicated by Equation 1: i 3-OODE + 13S-HPODE (or H202) + KOH --o OEOE + 13S-HODE (or H20 ) + KOH

[ 1]

Mechanism of oxygen transfer to ( 9Z, I IR,S, 12S,R)-13oxo- 11,12-epoxy-9-octadecenoic acid. Intermolecular transfer of peroxide oxygen to the 11,12-double bond of i 3-OODE in alkali is a reasonable assumption. This possibility was tested by reaction of [carboxyl- 18O 1]13S-HPODE (1.0 raM) Lipids, Vol. 31, no. 10 (1996)

H.W. GARDNER ETAL.

1026

TABLE 2 |ntermolecular Transfer of Hydroperoxide Oxygen to the 11,12-Epoxide of (9Z,11 R,S,12S,R)-13-Oxo-11,1 2-epoxy-9-octadecenoic acid (OEOE) Compound analyzed (as derivatives)

Isotopic composition 9% 1800; 42% carboxy11801; 49% hydroperoxide 1802 10% 1800; 40% carboxy11801; 50% hydroperoxide 1802 10% 1800; 42% carboxy11801; 48% hydroxy 1801 23% 1800; 43% carboxy11801; 34% keto 1801 29% 1800; 47% 1801; 24% 1802 29% 1800; 50% 1801; 21% 1802 9% 1800; 91% 1801; 0% 1802

13S-HPODE reactanta 13S-HPODE recovered 13S-HODE 13-Oxo-9,11 -octadecadienoate OEOE b, found OEOE, intermolecular reaction, calculated OEOE, intramolecular reaction, calculated

al mM of [carboxyl-1801113S-HPODE (96% 1801; 4% 1800) and 1.3 mM [hydroperoxy-1802113S-HPODE (87% 1802; 13% 1800) was reacted with 5 M KOH at 35~ for 40 min. See Table 1 for other abbreviations. bUnder the conditions used, any 180-label at the C-13 keto oxygen of OEOE would have been lost because of exchange with water. This was proved in a separate experiment in which OEOE produced from [hydroperoxy-1802 ] 13S-HPODE was found to contain 180-label at the epoxide oxygen, but not at the keto oxygen.

and [hydroperoxy-1802]13S-HPODE (1.3 mM) with 5 M KOH at 35~ for 40 min. The methyl esters of 13S-HODE, 13-OODE, and (9Z, l 1R,S, 12S,R)- 13-oxo- 11,12-epoxy-9-octadecenoic acid were isolated by TLC after development with hexane/acetone (92:8, vol/vol). The isolates were analyzed by GC-MS using selected ion monitoring furnishing results consistent with intermolecular epoxidation in the formation of the oxoepoxyene fatty acid (Table 2). Stability of saturated hydroperoxides. The availability of an activated double bond of 13-OODE seemed necessary in order to "reduce" 13S-HPODE into 13S-HODE. Thus, the fully saturated methyl 12-HPO was reacted with 5 M KOH. As expected, this hydroperoxide was resistant to alkali (Table 3). With extended times of incubation, 12-HPO was converted mainly into 12-oxooctadecanoic acid with only small amounts of 12-hydroxyoctadecanoic acid. 12-Oxooctadecanoic acid (methyl ester) was identified by its mass spectrum as follows: [m/z (percentage relative intensity, ion structure)] 312 (1.4, M§ 281 (17, M § CH30); 242 (22, +CH2=COH(CH2) IoCOOCH3); 227 (17, +CO(CH2)IoCOOCH3); 185 (20); 167 (15); 153 (40); 128 (61, CH3(CH2)sCOH=CH2+); 113 (73, CH3(CH2)sCO+); 98 (80); 85 (57); 69 (70); 55 (100). 12-Hydroxyoctadecanoic acid was identified by its mass spectrum as its methyl ester tri-

methylsilyloxy derivative: [m/z (percentage relative intensity, ion structure)] 355 (2.4, M +- CH30); 339 (7, M § CH3OH - CH3); 301 (59, M+ - CH3(CH2)5); 187 (100, CH3(CHz)5CHOTMS+); 73 (49, TMS). When 13S-HPODE was also included in the reaction with 12-HPO (Table 3), the conversion of 12-HPO was more rapid implying that the unsaturation provided by 13S-HPODE was responsible. A proposed mechanism and other considerations. As shown in Figure 1, a mechanism is proposed that is consistent with the results. We propose that the first step is dehydration of 13S-HPODE to afford 13-OODE. Although 13-OODE is an intermediate that does not accumulate, 12-HPO in the presence of alkali was converted mainly to 12-oxooctadecanoic acid, inferring that conversion to the ketone is an important reaction with 13S-HPODE also. This ketodiene fatty acid provides the labile double bond that permits transfer of one distal oxygen from the peroxyl anion of 13S-HPODE to the I l, 12-epoxide of (9Z, l 1R,S, 12S,R)- 13-oxo- 11,12-epoxy9-octadecenoic acid. As discussed previously (3), this oxoepoxyene fatty acid evidently undergoes Favorskii Rearrangement through an unobserved cyclopropanone intermediate to furnish 2-butyl-4-hydroxy-5-tetradecendioic acid and 2-butyl-3,5-tetradecadiendioic acid. It seems plausible that

TABLE 3 Conversion of Methyl 12-Hydroperoxyoctadecanoate (12-HPO) by Alkali with and without 13S-HPODE a 12-HPO alone Time (h) 0 1 4 6 12

12-HPO + 13S-HPODE

Unreacted (mg)

12-HO b (mg)

12-OO b (mg)

Unreacted 12-HPO (mg)

Unreacted 13S-HPODE (mg)

12-HO + 13-HODE ~ (mg)

1.07 0.68 0.43

0.01 0.01 0.03

0 0.39 0.30

0.77 0.28

0.72 0.12

0.01 0.12

0.10

0.02

0.31

0.39

0.04

0.45

aThe reaction was always 5 M KOH at 35~ Incubations with 12-HPO alone was with 3.3 mM 12-HPO (1.09 mg 12HPO/mL); with 12-HPO plus 13S-HPODE the concentration of both was 2.3 mM (0.76 mg 12-HPO/mL and 0.72 mg 135HPODE/mL). Material from 1 mL aliquots was analyzed as methyl esters. See Table 1 for other abbreviations. bl 2-HO, methyl 12-hydroxyoctadecanoate; 12-00, methyl 12-oxooctadecanoate. CMethy112-hydroxyoctadecanoate and 13-HODE could not be separated by high-performance liquid chromatography, and thus, are tabulated together as a sum.

Lipids, Vol. 31, no. 10 (1996)

REACTION OF HYDROPEROXlDESWITH ALKALI

18Q1O. 8

1

18OH-

1027

~ I

13S-HPODE anion

~

13-OODE

/

13-OODE

/t 13-oxo-ll,12-epoxy-9-octadecenoate

H2 ~

a

'

proposedF a v ~ T i ~ : ~ 1 8 0

2-butyl-4-hydroxy-5-tetradecendioate

~"

R'

/k..~ HCOOH 2-butyl-3,5-tetradecadiendioate

11-hydroxy-9,12-heptadecadienoate FIG. 1. Proposed reaction mechanism of (13S,9Z,11 E)-I 3-hydroperoxy-9,11 -octadecadienoate (13S-HPODE) by alkali into (13S,9Z,11 E)-I 3-hydroxy-9,11-octadecadienoate (13S-HODE) and other fatty acids (as their K salts). 13-OODE, (9Z,11 E)-I 3-oxo-9,11-octadecadienoate; R = CH3(CH2)3-; R' = -(CH2)7COO- K+, the bold 180 indicates one oxygen transfer from hydroperoxy group to the epoxy group.

l l-hydroxy-9,12-heptadecadienoic acid could arise from 2-butyl-4-hydroxy-5-tetradecendioic acid by elimination of formic acid, but the only evidence for this eventuality was a consistent 180-iabeling pattern of the hydroxyl group. Although the mechanism would seem to be completely assessed by our experiments, the stoichiometry of product formation is not balanced. If one considers that "dehydration" to 13-OODE is the origin of 25% of products other than 13S-

HODE, then only 25% of the total serves as a sink for peroxide oxygen. Taken to completion, this would result in a 1:1 ratio of 13S-HODE to other products; whereas, the actual ratio was 3:1 (3). Obviously, there appear to be oxygen consumptive reactions unknown to us. For example, a number of minor products were short-chain dienedioc and monoenedioic acids, such as 3-hydroxy-4-tridecenedioic acid (3), and these compounds could arise from secondary oxidations not shown Lipids, Vol. 31, no. 10 (1996)

1028

H.W. GARDNER ETAL

in Figure 1. However, the minor quantity of short-chain compounds found gives reason to doubt that these potential oxidations would entirely account for the stoichiometric imbalance. One possibility is an oxygen consumption in the "dehydration" of 13S-HPODE to 13-OODE. A reaction incorporating such a feature is known as the "Russell Mechanism" which combines two peroxyl radicals of 13S-HPODE, yielding a molecule each of 13S-HODE and 13-OODE, as well as singlet oxygen (11). This would lead to a more realistic 2:1 ratio of 13S-HODE to other products. Evidentally, this reaction is not occurring because neither oxygen nor products of reaction between 13S-HODE and singlet oxygen were observed when [U-14C]13S-HODE was incubated with 13SHPODE.

REFERENCES 1. Frankel, E.N., Evans, C.D., McConnell, D.G., and Jones, E.P. (1961) Analysis of Lipids and Oxidation Products by Partition Chromatography. Fatty Acid Hydroperoxides, J. Am. Oil Chem. Soc. 38, 134-137. 2. Simpson, T.D., and Gardner, H.W. (1993) Conversion of 13(S)Hydroperoxy-9(Z),l l(E)-octadecadienoic Acid to the Corresponding Hydroxy Fatty Acid by KOH: A Kinetic Study, Lipids 28, 325-330. 3. Gardner, H.W., Simpson, T.D., and Hamberg, M. (1993) Transformation of Fatty Acid Hydroperoxides by Alkali and Characterization of Products, Lipids 28, 487-495.

Lipids, Vol. 31, no. 10 (1996)

4. O'Brien, P.J. (1969) Intracellular Mechanisms for the Decomposition of a Lipid Peroxide. I. Decomposition of a Lipid Peroxide by Metal Ions, Heme Compounds, and Nucleophiles, Can. J. Biochem. 47, 485-492. 5. Hamberg, M. (1988) Fatty Acid Allene Oxides II. Formation of Two Macrolactones from 12,13(S)-Epoxy-9(Z),l l-octadecadienoic Acid, Chem. Phys. Lipids 46, 235-243. 6. Hamberg, M., and Gotthammar, B. (1973) New Reaction of Unsaturated Fatty Acid Hydroperoxides: Formation of 1l-Hydroxy- 12,13-epoxy-9-octadecenoic Acid from ! 3-Hydroperoxy9, i 1-octadecadienoic Acid, Lipids 8, 737-744. 7. Bascetta, E., and Gunstone, F.D. (1984) Attempted Synthesis of (9Z)-Methyl 12-Hydroperoxyoctadec-9-enoate, J. Chem. Soc. Perkin L 2217-2220. 8. Ratcliffe, R., and Rodehorst, R. (1970) Improved Procedure for Oxidations with the Chromium Trioxide-pyridine Complex, J. Org. Chem. 35, 4000---4002. 9. Anonymous (1971 ) Official Methods and Recommended Practices o f the American Oil Chemists' Society (Link, W.E., ed.) 3rd edn., Method Cd 8-53, American Oil Chemists' Society, Champaign. 10. Behar, D., Czapski, G., Rabani, J., Dorfman, L.M., and Schwarz, H.A. (1970)J. Phys. Chem. 74, 3209-3213. 11. Nakano, M., Takayama, K., Shimizu, Y., Tsuji, Y., Inaba, H., and Migita, T. (1976) Spectroscopic Evidence for the Generation of Singlet Oxygen in Self-reaction of sec-Peroxy Radicals, J. Am. Chem. Soc. 98, 1974-1975. [Received May 16, 1996; and in final revised form August 8, 1996; Revision accepted August 9, 1996]