1716
Chem. Res. Toxicol. 2004, 17, 1716-1724
Free Radical Oxidation of (E)-Retinoic Acid by the Fenton Reagent: Competing Epoxidation and Oxidative Breakdown Pathways and Novel Products of 5,6-Epoxyretinoic Acid Transformation Lucia Panzella, Paola Manini, Alessandra Napolitano, and Marco d’Ischia* Department of Organic Chemistry and Biochemistry, University of Naples “Federico II”, Via Cinthia 4, I-80126 Naples, Italy Received July 29, 2004
The nature of the products formed by reaction of all-trans retinoic acid (1) and its major metabolite, 5,6-epoxyretinoic acid (2), with the Fenton reagent was investigated. Oxidation of 1 in a vigorously stirred biphasic medium (0.1 M phosphate buffer, pH 7.4/ethyl acetate 5:1 v/v) with Fe2+/EDTA complex (2 mol equiv) and a 10-fold excess of H2O2 proceeded smoothly to give a very complex mixture of products. Repeated TLC fractionation of the reaction mixture after methylation allowed isolation of the main products which were identified as 2 methyl ester, (7E)-7,8-epoxyretinoic acid methyl ester (6), all-(E)-2,6-dimethyl-8-(2,6,6-trimethyl-2cyclohexen-1-ylidene)-2,4,6-octatrienal (11), the novel (9E)-5,6,9,10-diepoxyretinoic acid methyl ester (7), and (9Z)-5,6,9,10-diepoxyretinoic acid methyl ester (8) (1:1 mixture of syn/anti isomers each), 5,6-epoxy-β-ionone (9), 5,6-epoxy-β-ionylideneacetaldehyde (10), and trace amounts of β-ionone (12), β-ionylideneacetaldehyde (13), and 4-oxoretinoic acid (3) methyl ester. When the oxidation was carried out with the substrate and the Fenton reagent at concentrations as low as 10 µM, the main detectable products were 2 methyl ester, 11, and 7/8. Under similar conditions, the epoxide 2 gave mainly products 7-10. A less efficient conversion of 1 and 2 but similar product patterns were observed with other oxidizing systems such as peroxidase/ H2O2 and 13-hydroperoxyoctadecadienoic acid in the presence of Fe(II). Besides providing the first detailed insight into the products formed by reaction of a retinoid with the Fenton reagent, the results of this study disclose a novel nonenzymatic route from 1 to the epoxide 2 and offer an improved chemical basis to inquire into the mechanism of the antiinflammatory, antimutagenic, and cancer chemopreventive action of these retinoids.
Introduction The vitamin A metabolite all-trans retinoic acid (1)1 and its congeners, the retinoids, are potent inducers of cell differentiation (1) and have achieved status over the past decades as some of the best recognized agents for the treatment of certain types of cancer (2-4) as well as of photoaged skin and acne (5, 6). The potent antiinflammatory and antiproliferative properties of retinoids have been related at least in part to their susceptibility to structural modifications or conversion to bioactive metabolites following enzymatic oxidation and/or interaction with reactive free radicals and toxic forms of oxygen. Enzymatic metabolic routes of 1 may be mediated by prostaglandin H (PGH)2 synthase (7, 8) as well as by several cytochrome P450s (9) and lead mainly to the 5,6epoxy (2) and the 4-oxo (3), 4-OH (4), and 18-OH (5) derivatives. * To whom correspondence should be addressed. Tel: +39-081674132. Fax: +39-081674393. E-mail:
[email protected]. 1 All-trans retinoic acid is (2E,4E,6E,8E)-3,7-dimethyl-9-(2,6,6-trimethyl-1-cyclohexen-1-yl)-2,4,6,8-nonatetraenoic acid. For the sake of simplicity, throughout this paper the term retinoic acid (1) was preferably used. 2 Abbreviations: PGH, prostaglandin H; CPBA, chloroperoxybenzoic acid; 13-HPODE, 13-(S)-hydroperoxy-9Z,11E-octadecadienoic acid; HMQC, heteronuclear multiple quantum coherence; HMBC, heteronuclear multiple bond correlation; COSY, correlation spectroscopy; ROESY, rotating Overhauser effect spectroscopy.
Nonenzymatic oxidation pathways bear particular relevance to the role of 1 and its congeners in lipid peroxidation and other harmful cell damaging reactions caused by free radicals and reactive oxygen species. The majority of data indicate that the retinoids behave as antioxidants, can inhibit iron-induced peroxidation of polyunsaturated fatty acids in lipid bilayers (10), and can reduce cellular radiation sensitivity in vitro, due purportedly to their hydroxyl radical scavenging properties (11, 12). However, under particular conditions, they may exert the opposite effect and act as prooxidants, for example,
10.1021/tx049794b CCC: $27.50 © 2004 American Chemical Society Published on Web 11/19/2004
Oxidation of Retinoic Acid by the Fenton Reagent
in homogeneous liquid-phase systems (13) and in SDS micelles (14). Insights into the chemical bases of the paradoxical antioxidant/prooxidant roles of retinoids underscored the importance of conversion to, or interaction with, free radicals, chiefly peroxyl radicals. Peroxyl radicals can add to the C-5 terminus of the conjugated polyene system of 1 to give a highly delocalized carbon-centered radical which evolves chiefly via loss of an alkoxyl radical to form the 5,6-epoxide 2 (15, 16). This conversion warranted proposal of 1 as a convenient probe for peroxyl radical (16). In the PGH synthaseinduced reaction (7, 8), on the other hand, 1 itself is oxidized through H-atom abstraction at C-4 to generate a carbon-centered radical intermediate. This couples with oxygen to give a peroxyl radical at C-4, which in turn would add to another molecule of 1 to give 2. Detection of both 3 and 4 in the oxidation mixture of the PGH synthase-promoted oxidation lent support to the H-atom abstraction mechanism, suggesting different mechanistic routes of the peroxyl radical, for example, via Russell rearrangement of tetroxide intermediates, H-atom abstraction from the environment, or reduction (7). The main oxidized metabolites of 1, 2-4, can bind the three isotypes of retinoic acid receptors, can induce the transactivation of gene transcription, and can activate several signaling pathways, probably playing an important role in cell physiology and cancer therapy (17). As noticeable examples, 3 is highly active in modulating embryogenesis (18); 2, 3, and 4 can inhibit the growth of cancer cell lines (19-21); and 2, 4, and 5 can induce maturation of NB4 promyelocytic leukemia cells (22). The epoxide 2, moreover, is as active as 1 on 12-O-tetradecanoyl phorbol-13-acetate-dependent tumor promotion in the two-stage mouse skin carcinogenesis assay (23), but is more potent than 1 in the inhibition of phorbol esterstimulated chemiluminescence in human polymorphonuclear leukocytes (24), to the point that metabolic conversion to 2 has been implicated in the biological and pharmacological responses of 1. Apart from other scattered observations (25), knowledge of the oxidative transformations of 1 is currently limited to the mechanisms of formation of 2-4, and surprisingly little is known as to whether 2 and other major metabolites may undergo further enzymatic or nonenzymatic conversion, their oxidation chemistry being totally uncharted. Another noticeable gap concerns the behavior of retinoids to the Fenton reagent, a widely used system mimicking iron-mediated oxidations of relevance to oxidative stress states. The generation of hydroxyl radicals and/or hypervalent iron-oxo intermediates by Fenton-like reactions is an established biochemical correlate of most pathological conditions with severe inflammation and may contribute to the cell damaging mechanisms in skin photoaging and acne. Normal human skin contains approximately 200 µM iron, predominantly complexed to ferritin (26, 27). The release of free ferrous ions by UV irradiation and the subsequent interaction with cellular hydrogen peroxide, via Fenton-type chemistry, or the direct UV-induced homolysis of hydrogen peroxide, are therefore likely events in inflamed skin and may concur to the generation of relatively high fluxes of hydroxyl radicals (28). These may target 1 and related retinoids, which may be present in the skin at levels of about 300 ng/g wet tissue following topical administration (29), possibly affecting their biological activity. Only a few reports have recently appeared showing the ability of
Chem. Res. Toxicol., Vol. 17, No. 12, 2004 1717
various carotenoids (30) as well as of 1 (31) to serve as scavengers of reactive free radical species produced by the Fenton reagent in vitro and in vivo. An insight into the reaction of 1 with the Fenton reagent appears therefore desirable if the pharmacological and toxicological relevance of oxidative routes of retinoids in the skin and other organs has to be assessed to a closer detail. In this paper, we report the first investigation into the reaction of 1 as well as of its 5,6-epoxide 2 with the Fenton reagent and other model oxidizing systems. Specific aims of the study were to provide a careful inventory of the oxidation products of 1, to assess whether generation of 2 is a route of general relevance to nonenzymatic pathways of oxidation of 1, and to gain an insight into the oxidation chemistry of 2.
Experimental Procedures Materials. Retinoic acid (1), β-ionone (12) (96%), activated manganese oxide (85%), m-chloroperoxybenzoic acid (m-CPBA) (77%), butylated hydroxytoluene, sodium borohydride, and xylenol orange sodium salt were from Aldrich. Ammonium iron(II) sulfate hexahydrate, ammonium iron(III) sulfate dodecahydrate, and hydrogen peroxide (ca 30% solution) were purchased from Carlo Erba. Disodium EDTA dihydrate was from Fluka. Aqueous buffers were treated with Chelex-100 (BioRad, 200-400 mesh). Compound 1 methyl ester was prepared from 1 by treatment with diazomethane. Diazomethane was obtained from N-methyl-N-nitroso-p-toluenesulfonamide by treatment with KOH/ethanol (32). Caution: Diazomethane is explosive and must be collected in peroxide-free ether in a dry ice/acetone bath and kept at -20 °C. Note that diazomethane is also toxic and carcinogenic and should be handled in a fume hood. 5,6-Epoxyretinoic acid (2) methyl ester and 5,6-epoxy-β-ionone (9) were prepared from 1 methyl ester (33) and 12 (34), respectively, by treatment with m-CPBA in dry and peroxidefree diethyl ether. Compound 2 was obtained by hydrolysis of its methyl ester in KOH/ethanol under an argon atmosphere (33). 4-Oxoretinoic acid (3) methyl ester was obtained from 1 methyl ester by treatment with manganese oxide in dichloromethane (35). 4-Hydroxyretinoic acid (4) methyl ester was obtained by reduction of 3 methyl ester with sodium borohydride in water/methanol 1:1 v/v (35). 13-(S)-Hydroperoxy-(9Z,11E)octadecadienoic acid (13-HPODE) was prepared as reported (36). Methods. UV and IR spectra were performed using a Beckman DU 640 spectrophotometer and a Jasco 430 FT-IR spectrophotometer, respectively. 1H and 13C NMR spectra were recorded with a Bruker WM 400 at 400.1 and 100.6 MHz, respectively. The instrument was equipped with a 5 mm 1H/ broadband gradient probe with inverse geometry. Gradient selected versions of inverse (1H detected) heteronuclear multiple quantum coherence (HMQC) and heteronuclear multiple bond correlation (HMBC) experiments were used. The HMBC experiments used a 100 ms long-range coupling delay. 1H,1H correlation spectroscopy (COSY) and rotating Overhauser effect spectroscopy (ROESY) experiments were run using standard pulse programs from the Bruker library. Chemical shifts are reported in δ values (ppm) downfield from TMS. (CD3)2CO was used as solvent. EI- and high-resolution EI-MS spectra were obtained at 70 eV and 230 °C using a Kratos MS 50 spectrometer. Main fragmentation peaks are reported with their relative intensities (percent values are in parentheses). Analytical TLC analyses were performed on F254 0.25 mm silica gel. Preparative TLC was performed on F254 0.5 mm silica gel plates. Cyclohexanes-ethyl acetate 90:10 (eluant A) or 75: 25 containing 1% acetic acid (eluant B) were used as the eluant. Column chromatography was performed using 0.063-0.200 mm basic aluminum oxide. HPLC analyses were performed with a Gilson instrument equipped with a UV detector set at 280 nm. A dodecylsilane-
1718
Chem. Res. Toxicol., Vol. 17, No. 12, 2004
Panzella et al.
Table 1. NMR Spectral Data for 7 and 8a 7b 1H
(J, Hz)
1c 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 OCH3 a
1.10 and 1.43 (m) 1.39 (m) 1.77 and 1.80 (m) 6.06 (d, 15.6)/6.07 (d, 15.6) 5.58 (d, 15.6)/5.59 (d, 15.6) 3.41 (d, 6.8)/3.42 (d, 6.8) 6.13 (dd, 16.0, 6.8)/6.14 (dd, 16.0, 6.8) 6.58 (d, 16.0)/6.59 (d, 16.0) 5.87 (s) 0.88 (s)/0.89 (s)d 1.08 (s)/1.09 (s)d 1.11 (s)/1.12 (s) 1.41 (s)/1.42 (s) 2.28 (s)/2.29 (s) 3.67 (s)
8b 13C
34.2 36.7 17.9 30.8 65.7 71.0 128.7 136.3 62.5 65.8 132.0 138.8 152.0 120.6 167.6 26.3 26.3 21.3 16.4 14.0 51.4
1H
1.10 and 1.42 (m) 1.40 (m) 1.77 and 1.79 (m) 6.05 (d, 15.6)/6.06 (d, 15.6) 5.66 (d, 15.6)/5.67 (d, 15.6) 3.52 (d, 7.6) 5.94 (dd, 16.0, 7.6)/5.96 (dd, 16.0, 7.6) 6.58 (d, 16.0)/6.59 (d, 16.0) 5.84 (m) 0.89 (s)/0.90 (s)e 1.07 (s)/1.11 (s)e 1.12 (s)/1.13 (s) 1.46 (s)/1.47 (s) 2.20 (d, 1.2)/2.22 (d, 1.6) 3.66 (s)
Recorded in (CD3)2CO. b Each as a couple of diastereomers. c Numbering as shown in structural formula 1.
coated column, 4.6 × 150 mm, 4 µm particle size (Synergi MAX RP 80A), was used with a flow rate of 1.0 mL/min. Binary gradient elution conditions were used as follows: 0.01 M ammonium acetate (pH 6.5)-methanol (50:50), solvent A; 0.01 M ammonium acetate (pH 6.5)-methanol (10:90) solvent B; from 40% to 90% B, 0-55 min, 90% B, 55-75 min; flow rate 1 mL/ min. All operations with 1 and its derivatives, including workup of reaction mixtures, were carried out in the dark or under fluorescent light. Reaction of 1 and 2 with the Fenton Reagent. To a solution of 1 or 2 (0.01 mmol) in ethyl acetate (10 mL) was added 0.1 M phosphate buffer (pH 7.4) (990 mL), followed by 1-2 mol equiv of a complex Fe(NH4)2(SO4)2‚6H2O/EDTA (1:1 molar ratio) and 1-10 mol equiv of H2O2 while the mixture was vigorously stirred at 37 °C. The reaction course was monitored periodically by HPLC analysis of the ethyl acetate phase. After 3 h, the organic layer was separated from the aqueous layer, which was extracted twice with ethyl acetate; the combined organic layers were dried over sodium sulfate and taken to dryness to give a yellow residue which was treated with diazomethane and analyzed by TLC (eluant A). In other experiments, the reaction of 1 or 2 was run as above but (i) with the substrate at 5.0 mM concentration in the ethyl acetate phase (1:5 v/v ratio with the aqueous phase) using 1.4 mM Fe(NH4)2(SO4)2‚6H2O/EDTA (1:1 molar ratio) and 10 mol equiv of H2O2. After 16 h, the mixture was worked up and analyzed as above. Other reaction conditions were as (i) but (ii) without addition of H2O2; or (iii) in the absence of EDTA; or (iv) with Fe(NH4)(SO4)2‚12H2O in place of Fe(NH4)2(SO4)2‚6H2O; or (v) under an argon atmosphere; or (vi) in the reaction medium but without the Fenton reagent. Product analysis was carried out by HPLC. Eluates corresponding to the main chromatographic peaks were collected, the eluates were extracted with ethyl acetate, and the organic layers were dried and treated with diazomethane. Product identification followed from comparison of the TLC mobility (eluant A) with that of the products isolated from the oxidation mixture of 1 or 2 as described below. Isolation of 2 Methyl Ester, 6, 7, 8, 9, and 10. For preparative purposes, the reaction of 1 with the Fenton reagent was carried out using 100 mg of the starting material. The reaction was run in ethyl acetate/0.1 M phosphate buffer (pH 7.4) 1:5 v/v (375 mL total volume), with 2 mol equiv of the complex Fe(NH4)2(SO4)2‚6H2O/EDTA (1:1 molar ratio) and 10 mol equiv of H2O2 added in 10 portions at 15 min intervals, while the mixture was taken under vigorous stirring at room temperature. After 16 h, the reaction was worked up as above, and the yellow residue obtained was treated with diazomethane
13C
(J, Hz)
34.0 36.5 17.7 30.5 65.4 70.9 129.9 132.0/132.1 63.4 65.8/65.9 132.5 138.6 151.5/151.6 120.4 167.4 26.1f 26.3f 21.0/21.5 22.4/22.5 13.7 51.2
d -fInterchangeable.
and purified on preparative TLC (eluant A) to give 6 (Rf 0.55, 10 mg, 9% yield) (37), 2 methyl ester (Rf 0.50, 10 mg, 9% yield), 7 (Rf 0.40, 2 mg, 2% yield), 8 (Rf 0.37, 2 mg, 2% yield), 9 (Rf 0.32, 1 mg, 1% yield) (38), and 10 (Rf 0.31, 1 mg, 1% yield) (39). Compounds 7, 8, 9, and 10 (7 mg, 6% yield; 6 mg, 5% yield; 1 mg, 1% yield; 1 mg, 1% yield, in that order) were also isolated from the reaction mixture of 2 (115 mg) with the Fenton reagent following fractionation on preparative TLC plates (eluant A). Compound 7 (4 mg, 9% yield) was also isolated from the reaction of 2 methyl ester (45 mg) with m-CPBA under the same conditions used for 1 methyl ester, following fractionation on preparative TLC plates (eluant A). In a separate experiment, the reaction of 1 or 2 was run as in the general procedure, using 50 mg of the starting material. After 16 h, the reaction mixture was worked up as above, treated with diazomethane, and directly analyzed by NMR. The presence of hydroperoxides in the reaction mixture was evaluated by the xylenol orange assay (40). Briefly, the mixture was incubated with 25 mM sulfuric acid in methanol/water 9:1 v/v containing 4 mM butylated hydroxytoluene, 0.1 mM xylenol orange, and 0.25 mM Fe(NH4)2(SO4)2‚6H2O at room temperature, and after 30 min the absorbance at 560 nm was measured. Another aliquot of the mixture was treated with sodium borohydride in water; the organic layer was separated from the aqueous layer, which was extracted with ethyl acetate, and the combined organic layers were dried over sodium sulfate and directly analyzed by NMR. (9E)-5,6,9,10-Diepoxyretinoic Acid Methyl Ester (7). UV: λmax (ethyl acetate) 271 nm. FT-IR (dichloromethane): νmax 1711, 1613 cm-1. 1H and 13C NMR: see Table 1. EI-MS m/z: 346 (M+, 5), 207 (100). High-resolution mass spectrometry for C21H30O4: calcd 346.2144, found 346.2114. (9Z)-5,6,9,10-Diepoxyretinoic Acid Methyl Ester (8). UV: λmax (ethyl acetate) 268 nm. FT-IR (dichloromethane): νmax 1712, 1616 cm-1. 1H and 13C NMR: see Table 1. EI/MS m/z: 346 (M+, 5), 207 (100). High-resolution mass spectrometry for C21H30O4: calcd 346.2144, found 346.2130. Isolation of 11, 12, and 13. The reaction of 1 with the Fenton reagent was carried out as above using 100 mg of the starting material. After 16 h, the reaction mixture was worked up as above except for the methylation treatment. The yellow residue was directly fractionated by preparative TLC (eluant B) to give 11 (Rf 0.70, 6 mg, 6% yield) and a fraction (Rf 0.66, 4 mg) consisting of 12 and 13 (6% overall yield) in a 2:1 ratio (NMR evidence). Identification of compounds 11 (41) and 13 (42) followed from comparison of the proton and carbon spectra with those reported. Compound 12 was identified by comparison of
Oxidation of Retinoic Acid by the Fenton Reagent the chromatographic and spectral properties with those of a commercial sample. Compound 11 was also obtained from the reaction mixture of 1 with MnO2. Briefly, 1 (50 mg) was dissolved in dichloromethane (8.6 mL) and treated with MnO2 (1.29 g). The reaction mixture was taken under stirring for 48 h, filtered to remove the oxidant, and taken to dryness. The residue was fractionated by preparative TLC (eluant B) to give 11 (29 mg, 64% yield). Oxidation of 1 or 2 with Peroxidase/H2O2. To a solution of 1 or 2 (0.01 mmol) in ethyl acetate (2 mL) was added 0.1 M phosphate buffer (pH 7.4) (10 mL), followed by horseradish peroxidase (4 U/mL final concentration) and H2O2 (0.10 mmol), while the mixture was vigorously stirred. After 16 h, the organic layer was separated from the aqueous layer, which was extracted twice with ethyl acetate. The combined organic layers were dried over sodium sulfate and taken to dryness to give a yellow residue. The residue was taken up in ethyl acetate and analyzed by HPLC or treated with diazomethane and analyzed by TLC (eluant A). Oxidation of 1 or 2 with 13-HPODE/Fe (II). To a solution of 1 or 2 (0.01 mmol) in ethyl acetate (2 mL) was added 0.1 M phosphate buffer (pH 7.4) (10 mL), followed by 13-HPODE (0.01 mmol) and the complex Fe(NH4)2(SO4)2‚6H2O/EDTA (1:1 molar ratio) (7 µmol), while the mixture was taken under vigorous stirring. After 16 h, the mixture was worked up and analyzed as described above.
Results and Discussion Oxidation of 1. All operations with 1 and its derivatives, including workup of reaction mixtures, were carried out in the dark or under carefully controlled conditions to avoid artifactual isomerizations or other unwanted light-induced reactions. In typical experiments, 1 (10 µM) was exposed to H2O2 (10-100 µM) in the presence of Fe2+/ EDTA complex (10-20 µM) in 0.1 M phosphate buffer (pH 7.4) containing 1% ethyl acetate at 37 °C under vigorous stirring. After 3 h, HPLC and TLC (eluant B) analysis revealed 20-40% consumption of 1 and the formation of a complex and chromatographically unresolved product profile. However, methylation of the mixture with diazomethane improved the resolution and allowed isolation by TLC fractionation. The major oxidation product was identified as 5,6-epoxide 2 methyl ester by straightforward 1H NMR analysis and comparison of its chromatographic properties with those of a synthetic standard (33). To isolate and spectrally characterize the other oxidation products of 1, the concentrations of reagents in reaction mixtures were scaled-up ([1] ) 5.0 mM in the organic phase, [Fe(II)/EDTA] ) 1.4 mM, [H2O2] ) 8.3 mM) and the mixture was vigorously stirred for 16 h in a biphasic medium consisting of 0.1 M phosphate buffer (pH 7.4)/ethyl acetate (5:1 v/v). HPLC analysis indicated ca. 40% consumption of 1 and subsequent methylation, and TLC analysis revealed the formation of a very complex mixture of products comprising, besides 2 methyl ester, five main species characterized by distinct chromatographic bands: Rf 0.55, 0.40, 0.37, 0.32, and 0.31 (eluant A). Lack of artifactual isomerization and other unwanted light- or workup-induced reactions was confirmed by control experiments in which pure 1 was subjected to the same sequence of reaction and workup steps, with the sole exception of exposure to the oxidant, and was recovered virtually unchanged with no detectable product or isomer formation. Fractionation of the methylated mixture on preparative TLC afforded eventually the five bands in a chromato-
Chem. Res. Toxicol., Vol. 17, No. 12, 2004 1719
graphically homogeneous form under a variety of conditions, and each was subjected to extensive 2D NMR. The band at Rf 0.55 was due to an epoxide product isomeric to 2 methyl ester featuring two diagnostic signals in the 1H NMR spectrum at δ 3.21 and 3.49, showing one-bond correlation with two carbon resonances at δ 58.4 and 62.7, respectively, denoting the oxirane ring. HMQC and HMBC data eventually allowed the unambiguous identification of the compound as the (7E)-7,8epoxyretinoic acid methyl ester (6). This conclusion was secured by comparison of the proton and carbon spectrum with that reported in the literature (37) for 6. The bands migrating at Rf 0.40 and 0.37 shared a molecular ion peak at m/z 346. 1H NMR analysis revealed, moreover, that each band consisted of an intimate mixture of two closely related species (∆δ < 0.017 and 0.034 for the Rf 0.40 and 0.37 compounds, respectively) in comparable amounts, thus suggesting four products in a diastereoisomeric relationship. All attempts to separate each pair of coeluting products met with failure, and eventually they were characterized as mixtures. The species at Rf 0.40 exhibited a pattern of resonances in the proton spectrum similar to that of 2 methyl ester, with the noticeable exception of the lack of the H-10 doublet around δ 6.2, and the presence of a doublet at δ 3.41/3.42 (J ) 6.8 Hz) correlating with carbon resonances at δ 65.8 and 62.5 in the HMQC and HMBC spectra, respectively. On the basis of these and other lines of evidence, the products were identified as the syn and anti isomers of the novel (9E)-5,6,9,10-diepoxyretinoic acid methyl ester (7). The epoxy group on the 9,10-positions affects the chemical shift of the H-8 and CH3-19 protons, which appeared relatively highfield (δ 5.58/5.59 and 1.41/ 1.42, respectively) with respect to 2 methyl ester (δ 6.13 and 2.00, respectively). Consistent with this structural formulation was also the ipsochromic shift of the absorption maximum (271 nm) with respect to 1, denoting interruption of the conjugated polyene system. By similar arguments, the more retained products were assigned the structures of the syn and anti isomers of (9Z)-5,6,9,10-diepoxyretinoic acid methyl ester (8). The compounds displayed closely related 1H NMR spectra and featured a doublet at δ 3.52 (J ) 7.6 Hz) giving crosspeaks in the HMQC and HMBC spectra with carbon resonances at δ 65.8/65.9 and δ 63.4, respectively. The Z configuration at the 9,10-epoxy function was inferred on the basis of ROESY experiments, showing a telltale crosspeak between the CH3-19 protons at δ 1.46 and 1.47 (syn and anti isomers) and the H-10 protons at δ 3.52. No similar correlation was apparent in the ROESY spectrum of 7. NMR resonances of compounds 7 and 8 and assignments based on 2D correlation experiments are listed in Table 1. Unfortunately, the virtually overlapped signals and the lack of accessible nuclear Ovehauser effect contacts precluded reliable assignment of each set of signals to the components of the syn/anti pairs. The 1H NMR spectrum of the compound at Rf 0.32 was suggestive of a breakdown product of 1 in which the COOCH3 ester function was lost. On the basis of HMQC and HMBC data, the product was readily assigned the structure of 5,6-epoxy-β-ionone (9). This assignment was secured by comparison with a synthetic sample (34). Product 9 was shown to be generated during the autoxidation of β-carotene (43), but was never obtained by direct oxidation of 1 or related metabolites.
1720
Chem. Res. Toxicol., Vol. 17, No. 12, 2004
The species migrating at Rf 0.31 proved likewise devoid of the carbomethoxy functionality and displayed in the 1 H NMR spectrum as a most noticeable feature a doublet integrating for one proton at δ 10.15 (J ) 8.0 Hz) giving correlation peaks in the 1H,1H COSY spectrum with a proton resonance at δ 5.91, and in the HMQC spectrum with a carbon resonance at δ 192.0. The presence of only two additional signals in the low-field region (doublets at δ 6.37 and 6.66, J ) 15.6 Hz) suggests that this product represents the 5,6-epoxy-β-ionylideneacetaldehyde (10). Spectral data of 10 were in fairly good agreement with those reported previously (39). Formation of 10 was also described by oxidation of β-carotene with lipoxygenase in the presence of linoleic acid (44). Another component of the reaction mixture comigrated with 2 methyl ester (Rf 0.50) under the above elutographic conditions (NMR evidence) and was obtained in pure form by TLC fractionation (eluant B) of the reaction mixture without the methylation treatment. The product lacked the carboxyl moiety and displayed resonances for an aldehyde group (proton resonance at δ 9.44 correlating with a deshielded carbon signal at δ 194.5). It also featured 10 signals for sp2 carbons but only two signals for CH2 groups. On this basis, the product was assigned the structure of all-(E)-2,6-dimethyl-8-(2,6,6-trimethyl2-cyclohexen-1-ylidene)-2,4,6-octatrienal (11). Notably, 11 was also obtained by chemical oxidation of 1 with MnO2. Formation of 11 by oxidation of retinal under the latter conditions was previously described (41). Other minor components of the reaction mixture could be identified using the elutographic conditions B. These included β-ionone (12) and β-ionylideneacetaldehyde (13) (42). Formation of 12 was observed in the autoxidation of 1 (45).
Panzella et al.
Given the incomplete mass recovery, further information about the reaction products of 1 was sought through extensive 2D NMR analysis (COSY, HMQC, HMBC) of a crude oxidation mixture obtained as above and methylated with diazomethane. The 1H NMR spectrum displayed, besides significant amounts of unreacted 1 methyl ester, the resonances for 2 methyl ester, and those due to compounds 6, 7, and 8. Careful inspection of 1H,1H COSY and HMBC spectra provided evidence for the presence of small amounts of 3 methyl ester, the aldehydes 10, 11, and 13, as well as compounds 9 and 12. Overall, spectral analysis confirmed that the isolated products represented the major components of the reaction mixture. With products 6-13 available, a closer inspection of the mixture formed by oxidation of 10 µM 1 with 10 µM H2O2 and 10 µM Fe2+/EDTA became feasible. Besides 2 methyl ester, compounds 7, 8, 10, and 11 could be identified by careful co-chromatography and NMR analysis. In another set of experiments, the reaction of 1 with other oxidizing systems of physiological relevance was investigated. In the biphasic phosphate buffer (pH 7.4)/ ethyl acetate (5:1 v/v) medium, 1 (5 × 10-3 M in the organic phase) reacted slowly with 10 equiv of H2O2 and 4 U/mL horseradish peroxidase to give, after methylation, 2 methyl ester as the main product with low substrate consumption (ca. 20% after 16 h). A similar product profile was observed using 13-(S)-hydroperoxy-9(Z),11(E)octadecadienoic acid (13-HPODE) in the presence of Fe2+ as the oxidizing system in the same biphasic medium. Oxidation of 2. The prevalent formation of 2 by oxidation of 1 with the Fenton reagent prompted a separate set of experiments aimed at dissecting the main oxidation channels of this product under the same experimental conditions adopted for 1. Exposure of racemic 2 (10 µM) to H2O2 (100 µM) and Fe2+/EDTA (20 µM) resulted in a smooth reaction to give, after 16 h, a complex mixture of products, which, after workup, yielded two main species (actually pairs of products) identical in all respects to 7 and 8. Small amounts of 9 and 10 could also be detected (TLC evidence). Performing the reaction under preparative scale conditions allowed isolation and full spectral characterization of the products, thus providing definitive confirmation of their identity. To gain a deeper insight into the products formed by oxidation of 2 with the Fenton reagent, and to rule out artifacts due to chromatographic separation, a separate experiment was carried out where the whole crude mixture ([2] ) 5 × 10-3 M in the organic phase, 16 h reaction time) was subjected to extensive NMR analysis following methylation treatment but without chromatographic fractionation (as reported above in the case of 1). The 1H NMR spectrum displayed, besides some 2 methyl ester (ca. 33% of the mixture), the resonances for 7 (10%), 8 (10%), 9 (4%), and 10 (4%). Product yields were determined by comparing integrated areas of characteristic proton resonances for each compound, the signal at δ 3.41 (1H) in the case of 7, at δ 3.52 (2H) for 8, at δ 2.24 (3H) for 9, and at δ 10.13 (1H) for 10, with the total area contributed by the C-16 or C-17 methyl groups (δ 0.89). Especially worthy of note was the presence of signals in the region between δ 4.68-4.87, correlating in the HMQC and HMBC spectra with carbon signals at δ 89-
Oxidation of Retinoic Acid by the Fenton Reagent
94, suggestive of hydroperoxide derivatives. This conclusion was supported by the marked decrease in the signal area following reductive treatment of the mixture with aqueous NaBH4. Moreover, the reaction mixture proved positive to the xylenol orange assay specific for hydroperoxides (40). In addition, a complex group of resonances in the range between δ 4.05-4.56, displaying cross-peaks in the HMQC spectrum with carbon signals at δ 72-77, were indicative of hydroxylated derivatives, supposedly derived from hydrolytic cleavage of epoxide functionalities or CH2OH-containing species derived from hydroxylation of allylic methyl groups such as C-19 (46). Two minor proton signals at δ 3.78 and 3.83, correlating with carbon resonances at δ 62.6 and 62.4, respectively, denoted probably additional diepoxide derivatives. A group of signals in the low-field region of the proton spectrum at δ 9.37-10.60 revealed the presence in the mixture of other aldehyde species, possibly oxidative breakdown products associated with the formation of 9 and 10, products derived from oxidation of allylic methyl groups, and/or products derived from β-scission of alkoxyl radical intermediates. To gain further insight into the oxidation chemistry of 2, additional experiments were carried out involving a peroxidase/H2O2 system. In the biphasic phosphate buffer (pH 7.4)/ethyl acetate (5:1 v/v) medium, 2 (5 × 10-3 M in the organic phase) reacted slowly with 10 equiv of H2O2 and 4 U/mL horseradish peroxidase to give, after methylation, 7 and 8 along with some 9 and 10. No appreciable substrate conversion (TLC and HPLC evidence) was observed in the reaction of 2 with 13-HPODE in the presence of Fe2+ even after prolonged reaction time. Finally, the reaction of 2 with an established epoxidizing agent, m-chloroperbenzoic acid (m-CPBA), was investigated to substantiate the regioselective pattern of oxygenation at C-9 and the underlying stereochemical facets. As expected, TLC analysis revealed the formation of a single chromatographic band which proved to consist of pure 7. Mechanistic Issues. The structures of the products isolated by oxidation of 1 and 2 raised several issues concerning the mechanism of the Fenton-induced epoxidation reactions leading to 2, 6, and 7/8 as well as of the decarboxylation or cleavage processes at the various sites of the polyene chain. To address these issues, the effects of varying certain experimental parameters on the extent of degradation of 1 and 2 and on the product distributions were investigated. Omission of EDTA resulted in a lower substrate consumption (15% after 16 h) without appreciable changes in the product pattern and distribution. This is consistent with the reported effect of EDTA on the reduction potential of the Fe3+/Fe2+ couple (lowering from + 0.77 to + 0.12 V, measured in water) (47), although phosphate itself may act as complexing ligand favoring Fe(II) autoxidation (48) even if it is less efficient in the generation of hydroxyl radical in the Fenton reaction (49). A similar decrease in substrate decay was observed using Fe3+ in the place of Fe2+, whereas no reaction occurred in the absence of H2O2. Noticeably, when the Fenton-induced oxidation was carried out under an argon atmosphere, product formation was virtually inhibited, although the substrate was consumed to a comparable extent (Figure 1).
Chem. Res. Toxicol., Vol. 17, No. 12, 2004 1721
Figure 1. HPLC elution profile of the ethyl acetate extractable fraction of the reaction mixture of 2 (5.0 mM in the organic phase) with 1.4 mM Fe2+/EDTA complex and 10 mol equiv of H2O2 in 0.1 M phosphate buffer (pH 7.4)/ethyl acetate (5:1 v/v) after 16 h. The reaction was run in air (plot A) or under an argon atmosphere (plot B). Compounds eluted under peaks I and II were identified as 8 and 7 free acid, respectively, as described in the Experimental Section.
These observations were consistent with the generally held mechanism for Fenton reactions (50) involving oneelectron reduction of H2O2 by the Fe2+/EDTA complex to give OH radicals or related species, such as hypervalent iron-oxo intermediates (48, 49), which trigger a free radical oxidation process mediated by molecular oxygen. As a rule, OH radicals from the Fenton reagent react with unsaturated systems (i) by H-atom abstraction from an allylic or benzylic position (51), (ii) by electron abstraction from π systems (52), or (iii) by addition. The latter, however, becomes significant only with electron-rich substrates, for example, phenolic systems (53), or in the absence of allylic hydrogens, because of the relatively electrophilic character of OH radicals. Accordingly, a plausible mechanistic framework for the oxidative conversion of 1 mediated by the Fenton system was proposed, which is outlined in Scheme 1. The initial event would be the abstraction of an allylic hydrogen at C-4 of 1 by the OH radical or a related species generated by the Fenton system, or by peroxyl/ alkoxyl radicals formed during the reaction. Alternatively, oxidation of 1 via electron abstraction would give
1722
Chem. Res. Toxicol., Vol. 17, No. 12, 2004
Panzella et al. Scheme 1
a radical cation intermediate which then depronates (7). This latter option is not usually favored from a thermodynamic point of view (54) but cannot be ruled out definitely considering the high redox potential of hydroxyl radical. In any event, the resulting carbon-centered radical would couple with oxygen to give peroxyl radicals (ROO•), which in turn could add to the C-5 position of another molecule of 1. Homolytic β-scission of the resulting peroxide bond followed by ring closure with elimination of an alkoxyl radical would eventually yield 2 (route a). Previous studies have provided solid mechanistic support for the peroxyl radical-dependent epoxidation of 1 to 2 (7, 8, 15, 16) and have shown that the regiospecificity of the reaction may be due to the 5,6-double bond behaving like a partially isolated olefinic function because of impaired p-orbital overlap imposed by geometric constraints (16). In this context, the present results not only demonstrate the generality of this regiospecific epoxidation pathway of 1 but disclose also a novel, minor epoxidation route involving conceivably addition of peroxyl radical(s) across the 7,8-double bond. This latter pathway becomes appreciable only at higher substrate and oxidant concentrations, consistent with the less favorable energetics predicted for peroxyl radical attack at sites other than the 5,6-double bond (16). Alternatively, reduction of hydroperoxide intermediates at C-8 by the Fe2+/EDTA complex and cyclization of the resulting alkoxyl radical would provide a possible contributory pathway to the epoxide 6 (route c). Such a mechanistic option, however, is conceivably precluded in the case of the 5,6-epoxide 2, as cyclization of the alkoxy radical would lead to a nondelocalized C-4 centered carbon radical (14). The detection of small amounts of 3 may be taken as a “footprint” of the proposed allylic H-atom abstraction at C-4, as argued in the case of the PGH synthase-
catalyzed reaction (7, 8, 15). Oxygen coupling of the resulting C-4 centered radical followed by Fenton-mediated decomposition of the hydroperoxide would give rise to 4 and hence to 3 by further oxidation (route b). Yet, the complexity of the product pattern and the lack of substantial formation of 4 or related 4-hydroxylated products, as inferred from 2D NMR analysis (COSY, HMQC, HMBC) of the whole oxidation mixtures of 1, suggest that the first formed carbon-centered radical of 1 can partition among several possible coupling routes with oxygen involving various sites of the polyene chain with unpaired electron density. Thus, formation of 11 would involve breakdown of hydroperoxide intermediates concomitant with decarboxylation (route d). In principle, diepoxides 7/8 as well as the carbonyl compounds 9 and 10 may arise from two different mechanistic sequences, mutually nonexclusive, in which the regiospecific 5,6-epoxidation step may precede or follow epoxidation at the 9,10-double bond or oxidative breakdown steps. Epoxide formation prior to functionalization is supported by the studies on the oxidation of 2 with the Fenton reagent, which demonstrated substantial conversion of this epoxide into products 7-10. The mechanism here could involve H-atom abstraction from C-19 to yield a delocalized carbon-centered radical which would couple with oxygen at C-10 or at C-12. The potential mechanistic fates of the resulting peroxyl radicals are shown in Scheme 2. Formation of 7/8 might involve an intermolecular, peroxyl radical-mediated mechanism akin to that reported in the epoxidation of 1. Free rotation around the C9-C10 bond would give the cis and trans isomers in comparable amounts. The formation of both cis and trans isomers clearly denotes lack of stereospecificity, at variance with the m-CPBA-induced epoxidation of 2, which proceeded with high regio- and stereospecificity yielding only the trans isomer 7. The remote location of the 5,6-
Oxidation of Retinoic Acid by the Fenton Reagent Scheme 2
Chem. Res. Toxicol., Vol. 17, No. 12, 2004 1723
the discovery of hitherto unrecognized transformation pathways of this retinoid in vivo. Other chemical highlights derive from the elucidation of the main oxidation products of 2, which revealed an unexpected facility of this epoxide to further epoxidation and oxidative breakdown of the polyene chain. Besides the possible bearing on the activity of endogenously produced 1, these results may be of particular relevance to the mechanism of the antioxidant action of 1 when topically delivered at locally high concentrations on inflamed skin with sustained lipid peroxidation. An assessment of the generation and biological activity of the novel products described in the present study is likely to open new perspectives in the current understanding of the mechanism of action of 1 and related retinoids.
Acknowledgment. This work was carried out in the frame of research programs on antitumor agents, sponsored by MIUR (PRIN 2003), and on the mechanisms of aging and photoaging (IFO 2003), sponsored by the Ministry of Health. We thank the “Centro Interdipartimentale di Metodologie Chimico-Fisiche of Naples University” for NMR facilities and mass spectra and Mrs. Silvana Corsani for technical assistance. Supporting Information Available: 1H NMR, 1H,13C HMBC, and ROESY spectra of compounds 7 and 8. This material is available free of charge via the Internet at http:// pubs.acs.org.
References epoxy functionality with respect to the 9,10-double bond explains the virtual lack of facial selectivity in the second epoxidation step, as is apparent from the syn/anti ratios of 1:1 invariably observed. Finally, a Hock-type cleavage (55) of the 10- and 12hydroperoxides would account for the observed generation of 9 and 10 from 2. Similar mechanisms involving breakdown of peroxyl radicals or Hock cleavage of hydroperoxide intermediates at C-12 and C-10 would account for the formation of carbonyl products (e.g., 12 and 13) from 1. It is clear, however, that the proposed mechanism for product formation in the oxidation of 1 and 2 relies on the available evidence and that it could be improved and detailed following additional experiments including oxygen labeling coupled with mass spectrometric analysis, measurements of oxygen uptake during the reaction course, as well as identification of minor reaction products or intermediates.
Conclusions The present study has examined for the first time the products formed by oxidation of a retinoid with the Fenton reagent and has yielded, to the best of our knowledge, the first insight into the oxidation chemistry of the 5,6-epoxide 2 under biomimetic conditions. The results have confirmed the importance of 2 as a major conversion product of 1 and have thrown light on novel pathways of nonenzymatic oxidation of 1 toxicologically and pharmacologically relevant. The generation of a decarboxylation product (11) and a diepoxide (7/8) by mild oxidation of 1 at micromolar levels is especially worthy of note and may provide an important basis for
(1) Castro, D. S., Hermanson, E., Joseph, B., Walle´n, A., Aarnisalo, P., Heller, A., and Perlmann, T. (2001) Induction of cell cycle arrest and morphological differentiation by Nurr1 and retinoids in dopamine MN9D cells. J. Biol. Chem. 276, 43277-43284. (2) Hansen, L. A., Sigman, C. C., Andreola, F., Ross, S. A., Kelloff, G. J., and De Luca, L. M. (2000) Retinoids in chemoprevention and differentiation therapy. Carcinogenesis 21, 1271-1279. (3) Fontana, J. A., and Rishi, A. K. (2002) Classical and novel retinoids: their targets in cancer therapy. Leukemia 16, 463472. (4) Marnett, L. J., and Ji, C. (1994) Modulation of oxidant formation in mouse skin in vivo by tumor-promoting phorbol esters. Cancer Res. 54, 1886s-1889s. (5) Nagpal, S., and Chandraratna, R. A. S. (2000) Recent developments in receptor-selective retinoids. Curr. Pharm. Des. 6, 919931. (6) Bershad, S. (2001) Developments in topical retinoid therapy for acne. Semin. Cutan. Med. Surg. 20, 154-161. (7) Samokyszyn, V. M., Chen, T., Maddipati, K. R., Franz, T. J., Lehman, P. A., and Lloyd, R. V. (1995) Free radical oxidation of (E)-retinoic acid by prostaglandin H synthase. Chem. Res. Toxicol. 8, 807-815. (8) Freyaldenhoven, M. A., Lloyd, R. V., and Samokyszyn, V. M. (1996) Prostaglandin H synthase-catalyzed oxidation of all-trans and 13-cis-retinoic acid to carbon centered and peroxyl radical intermediates. Chem. Res. Toxicol. 9, 677-681. (9) Marrill, J., Cresteil, T., Lanotte, M., and Chabot, G. G. (2000) Identification of human cytochrome P450s involved in the formation of all trans retinoic acid principal metabolites. Mol. Pharmacol. 58, 1341-1348. (10) Samokyszyn, V. M., and Marnett, L. J. (1990) Inhibition of liver microsomal lipid peroxidation by 13-cis-retinoic acid. Free Radical Biol. Med. 8, 491-496. (11) Ruifrok, A. C., Mason, K. A., and Thames, H. D. (1996) Changes in clonogen number and radiation sensitivity in mouse jejunal crypts after treatment with dimethyl sulfoxide and retinoic acid. Radiat. Res. 145, 740-745. (12) Takahashi, N. (2000) Antioxidant properties of N-(4-hydroxyphenyl)retinamide (fenretinide). Biol. Pharm. Bull. 23, 222-225. (13) Finkelshtein, E. I.; Rubashinskaya, Y. F., and Kozlov, Y. I. (1981) Kinetics and mechanism of the autoxidation of retinoic acid and its methyl ester. Zh. Org. Khim. 17, 936-943. (14) Freyaldenhoven, M. A., Lehman, P. A., Franz, T. J., Lloyd, R. V., and Samokyszyn, V. M. (1998) Retinoic acid-dependent stimula-
1724
(15) (16)
(17) (18)
(19)
(20)
(21)
(22)
(23)
(24)
(25) (26) (27) (28) (29)
(30)
(31)
(32) (33) (34)
(35)
Chem. Res. Toxicol., Vol. 17, No. 12, 2004
tion of 2,2′-azobis(2-amidinopropane)-initiated autoxidation of linoleic acid in sodium dodecyl sulfate micelles: a novel prooxidant effect of retinoic acid. Chem. Res. Toxicol. 11, 102-110. Samokyszyn, V. M., and Marnett, L. J. (1987) Hydroperoxidedependent cooxidation of 13-cis-retinoic acid by prostaglandin H syntase. J. Biol. Chem. 262, 14119-14133. Samokyszyn, V. M., Freyaldenhoven, M. A., Chang, H. C., Freeman, J. P., and Compadre, R. L. (1997) Regiospecificity of peroxyl radical addition to (E)-retinoic acid. Chem. Res. Toxicol. 10, 795-801. Idres, N., Marill, J., and Chabot, G. G. (2002) Activation of retinoic acid receptor-dependent transcription by all-trans-retinoic acid metabolites and isomers. J. Biol. Chem. 277, 31491-31498. Pijnappel, W. W., Hendriks, H. F., Folkers, G. E., van der Brink, C. E., Dekker, E. J., Edelenbosch, C., van der Saag, P. T., and Durston, A. J. (1993) The retinoid ligand 4-oxo-retinoic acid is a highly active modulator of positional specification. Nature 366, 340-344. van der Leede, B. M., van den Brink, C. E., Pijnappel, W. W., Sonneveld, E., van der Saag, P. T., and van der Burg, B. (1997) Autoinduction of retinoic acid metabolism to polar derivatives with decreased biological activity in retinoic acid-sensitive, but not in retinoic acid-resistant human breast cancer cells. J. Biol. Chem. 272, 17921-17928. Van heusden, J., Wouters, W., Ramaekers, F. C., Krekels, M. D., Dillen, L., Borgers, M., and Smets, G. (1998) All-trans-retinoic acid metabolites significantly inhibit the proliferation of MCF-7 human breast cancer cells in vitro. Br. J. Cancer 77, 26-32. Ramp, U., Gerharz, C. D., Eifler, E., Biesalski, H. K., and Gabbert, H. E. (1994) Effects of retinoic acid metabolites on proliferation and differentiation of the clonal rhabdomyosarcoma cell line BAHAN-1C. Biol. Cell. 81, 31-37. Idres, N., Benoit, G., Flexor, M. A., Lanotte, M., and Chabot, G. G. (2001) Granulocytic differentiation of human NB4 promyelocytic leukemia cells induced by all-trans retinoic acid metabolites. Cancer Res. 61, 700-705. Verma, A. K., Slaga, T. J., Werz, P. W., Mueller, G. C., and Boutwell, R. K. (1980) Inhibition of skin tumor promotion by retinoic acid and its metabolite 5,6-epoxyretinoic acid. Cancer Res. 40, 2367-2371. Kensler, T. W., and Trush, M. A (1981) Inhibition of phorbolesterstimulated chemiluminescence in human polymorphonuclear leukocytes by retinoic acid and 5,6-epoxyretinoic acid. Cancer Res. 41, 216-222. Rockley, N. L., Halley, B. A., and Nelson, E. C. (1980) Identification of a decarboxylation product of retinoic acid. Biochim. Biophys. Acta 627, 270-275. Gorodetsky, R., Sheskin, J., and Weinreb, A. (1986) Iron, copper and zinc concentrations in normal skin and in various nonmalignant and malignant lesions. Int. J. Dermatol. 25, 440-445. Goldblum, W. R., Derby, S., and Lerner, A. B. (1953) The metal content of skin, nails and hair. J. Invest. Dermatol. 20, 13-18. Darr, D., and Fridovich, I. (1994) Free radicals in cutaneous biology. J. Invest. Dermatol. 102, 671-675. Didierjean, L., Carraux, P., Grand, D., Sass, J. O., Nau, H., and Saurat, J. H. (1996) Topical retinaldehyde increases skin content of retinoic acid and exerts biologic activity in mouse skin. J. Invest. Dermatol. 107, 714-719. Polyakov, N. E., Leshina, T. V., Konovalova, T. A., and Kispert, L. D. (2001) Carotenoids as scavengers of free radicals in a Fenton reaction: antioxidants or prooxidants? Free Radical Biol. Med. 31, 398-404. Sultana, S.; Alam, A.; Sharma, S.; and Khan, N. (2004) 13-cis Retinoic acid ameliorates benzoyl peroxide-induced oxidative stress and hyperproliferative response in murine skin: A chemopreventive study. Cancer Detect. Prev. 28, 200-207. Vogel, A. I. (1951) A Text Book of Practical Organic Chemistry Including Qualitative Organic Analysis, II ed., Vol. I, pp 842844, Longmans, Green & Co, London. John, K. V., Lakshmanan, M. R., and Cama, H. R. (1967) Preparation, properties and metabolism of 5,6-monoepoxyretinoic acid. Biochem. J. 103, 539-543. Jungalwala, F. B., and Cama, H. R. (1965) Preparation and properties of 5,6-monoepoxyvitamin A acetate, 5,6-monoepoxyvitam A alcohol, 5,6-monoepoxyvitamin A aldehyde and their corresponding 5,8-monoepoxy (furanoid) compounds. Biochem. J. 95, 17-26. Samokyszyn, V. M., Gall, W. E., Zawada, G., Freyaldenhoven, M. A., Chen, G., Mackenzie, P. I., Tephly, T. R., and Radominska-
Panzella et al.
(36)
(37) (38)
(39)
(40)
(41)
(42)
(43) (44)
(45)
(46)
(47)
(48)
(49)
(50) (51)
(52)
(53)
(54)
(55)
Pandya, A. (2000) 4-Hydroxyretinoic acid, a novel substrate for human liver microsomal UDP-glucuronosyltransferase(s) and recombinant UGT2B7. J. Biol. Chem. 275, 6908-6914. Napolitano, A., Camera, E., Picardo, M., and d’Ischia, M. (2002) Reactions of hydro(pero)xy derivatives of polyunsaturated fatty acid/esters with nitrite ions under acidic conditions. Unusual nitrosative breakdown of methyl 13-hydro(pero)xyoctadeca-9,11dienoate to a novel 4-nitro-2-oximinoalk-3-enal product. J. Org. Chem. 67, 1125-1132. Davalian, D., and Heathcock, C. H. (1979) Synthesis of 7,8-epoxy7,8-dihydroretinoids. J. Org. Chem. 44, 4458-4461. Acemoglu, M., Eschenmoser, W., and Eugster, C. H. (1981) Synthese von (-)-(5R,6S)-5,6-epoxy-5,6-dihydro-β-ionon. Helv. Chim. Acta 64, 2691-2695. Acemoglu, M., and Eugster, C. H. (1984) (5R,6S,5′R,6′S)-5,6,5′,6′diepoxy-β,β-carotin: synthese, spektroskopische, chiroptische und chromatographische eigenschaften. Helv. Chim. Acta 67, 184190. Jiang, Z.-Y., Hunt, J. V., and Wolff, S. P. (1992) Ferrous ion oxidation in the presence of xylenol orange for detection of lipid hydroperoxide in low-density lipoprotein. Anal. Biochem. 202, 384-389. Colmenares, L. U., and Liu, R. S. H. (1991) An unexpected C-19 product from reaction of retinal with MnO2. Structure characterization and bacteriorhodopsin analogues. Tetrahedron Lett. 32, 1933-1936. Valla, A., Andriamialisoa, Z., Prat, V., Laurent, A., Giraud, M., Labia, R., and Potier, P. (2000) Syntheses stereoselectives des acides 13E et 13Z retinoiques via un nouvel intermediaire β-methylenealdehyde en C-15. Tetrahedron 56, 7211-7215. Mordi, R. C., and Walton, J. C. (1993) Oxidative degradation of β-carotene and β-apo-8′-carotenal. Tetrahedron 49, 911-928. Wu, Z., Robinson, D. S., Hughes, R. K., Casey, R., Hardy, D., and West, S. I. (1999) Co-oxidation of β-carotene catalyzed by soybean and recombinant pea lipoxygenases. J. Agric. Food Chem. 47, 4899-4906. Oyler, A. R., Motto, M. G., Naldi, R. E., Facchine, K. L., Hamburg, P. F., Burinsky, D. J., Dunphy, R., and Cotter, M. L. (1989) Characterization of autoxidation products of retinoic acid. Tetrahedron 45, 7679-7694. Matsushima, Y., Kawachi, E., Tanaka, H., Kagechika, H., Hashimoto, Y., and Shudo, K. (1992) Differentiation-inducing activity of retinoic acid isomers and their oxidized analogues on human promyelocytic leukemia HL-60 cells. Biochem. Biophys. Res. Commun. 189, 1136-1142. Nofre, C., Cier, A., and Lefier, A. (1960) Activation de l’oxygene moleculaire en solution aqueuse par les ions metalliques at leurs chelates. Memoires Presentes a la Societe` Chimique, 530-535. Welch, K. D., Zane Davis, T., and Aust, S. D. (2002) Iron autoxidation and free radical generation: effects of buffers, ligands, and chelators. Arch. Biochem. Biophys. 397, 360-369. Yamazaki, I., and Piette, L. H. (1990) ESR spin trapping studies on the reaction of Fe2+ ions with H2O2-reactive species in oxygen toxicity in biology. J. Biol. Chem. 265, 13589-13594. Goldstein, S., Meyerstein, D., and Czapski, G. (1993) The Fenton reagents. Free Radical Biol. Med. 15, 435-445. Hayashi, Y., Ueda, Y., Nakajima, A., Yokoyama, H., Mitsuyama, Y., Ohya-Nishiguchi, H., and Kamada, H. (2002) Nitric oxide and hydroxyl radicals initiate lipid peroxidation by NMDA receptor activation. Brain Res. 941, 107-112. Polyakov, N. E., Kruppa, A. I., Leshina, T. V., Konovalova, T. A., and Kispert, L. D. (2001) Carotenoids as antioxidants: spin trapping EPR and optical study. Free Radical Biol. Med. 31, 4352. Walling, C., and Johnson, R. A. (1975) Fenton’s reagent. V. Hydroxylation and side-chain cleavage of aromatics. J. Am. Chem. Soc. 97, 363-367. Samokyszyn, V. M., Chang, H. C., and Lilia Compadre, R. (1997) A theoretical investigation of endocyclic allylic carbon-centered radical formation in retinoic acid. Bioorg. Med. Chem. Lett. 7, 2343-2348. Schneider, C., Tallman, K. A., Porter, N. A., and Brash, A. R. (2001) Two distinct pathways of formation of 4-hydroxynonenal. Mechanisms of nonenzymatic transformation of the 9- and 13hydroperoxides of linoleic acid to 4-hydroxyalkenals. J. Biol. Chem. 276, 20831-20838.
TX049794B
