Chem. Res. Toxicol. 1995,8, 807-815
807
Free Radical Oxidation of (E)-RetinoicAcid by Prostaglandin H Synthase Victor M. Samokyszyn,*>tTao Chen,? Krishna Rao Maddipati,* Thomas J. Franz,s Paul A. Lehman,§ and Roger V. Lloyd1' Department of Pharmacology & Toxicology and Department of Dermatology, University of Arkansas for Medical Sciences, Little Rock, Arkansas 72205, Cayman Chemical Corporation, Ann Arbor, Michigan 48108, and Department of Chemistry, University of Memphis, Memphis, Tennessee 38152 Received January 13, 1995@
Cooxidative metabolism of all-trans (E)-retinoic acid (RA) by prostaglandin H synthase was investigated employing ram seminal vesicle microsomes (RSVM) or purified, RSVM-derived enzyme. RA was shown to undergo hydroperoxide [HzOz or 5-phenyl-4-penten-1-yl hydroperoxide (PPHP)]- or arachidonic acid-dependent cooxidation by microsomal prostaglandin H (PGH) synthase as evidenced by W spectroscopic analysis of reaction mixtures. Cooxidation of RA by microsomal or purified PGH synthase, using PPHP as substrate, was characterized by uptake of dioxygen which was first order with respect to enzyme concentration. Dioxygen uptake was inhibited by the peroxidase reducing substrate 2-methoxyphenol. In addition, 0 2 uptake was inhibited by the spin trap nitrosobenzene. ESR spin trapping studies, using a-phenyl-N-tert-butylnitrone (PBN) as the spin trap, demonstrated the formation of RA-PBN adducts, characterized by hyperfine coupling constants of a~ = 3.2 G and U N = 15.8 G. Reverse phase HPLC analysis of reaction mixtures demonstrated the formation of 4-hydroxy-RA, 5,6epoxy-RA, 4-0xo-RA, (13Z)-retinoic acid, and other geometric isomers which were identified on the basis of cochromatography with synthetic standards, W spectroscopy, and/or mass spectrometry. Mechanisms are proposed for the hydroperoxide-dependent, PGH synthasecatalyzed oxidation of RA that are consistent with these results.
Introduction all-trans-Retinoic acid (RAY is used topically in the treatment of acne vulgaris ( 1 ) and photodamaged skin (2)and is currently under investigation for the treatment of acute promyelocytic leukemia (reviewed in ref 3). RA is also a major in vivo and in vitro metabolite of retinol and retinyl acetate, and metabolism of retinol to RA is necessary for biological activity in mammalian epidermis as well as retinol-dependent teratogenicity (reviewed in ref 4). RA undergoes extensive (-50%) photochemical isomerization to 13-cis-retinoic acid ((13Z)-RA) when topically applied to human skin, and both isomers undergo percutaneous absorption (5). In addition, RA, (13Z)-RA, and other retinoids inhibit phorbol esterdependent tumor promotion in the mouse skin two-stage (initiation-promotion) carcinogenenesis assay (reviewed in ref 6). However, under certain conditions, retinoids exhibit the opposite effect and enhance tumorigenesis ( 7 ,
* Address correspondence to this author at Division of Toxicology (Mail Slot 6381, University of Arkansas for Medical Sciences, 4301 W. Markham, Little Rock, AR 72205. Tel: 501-686-5766; FAX: 501-6868970; E-Mail:
[email protected]. +Departmentof Pharmacology & Toxicology, UAMS. f Cayman Chemical Corporation. Department of Dermatology, UAMS. I1University of Memphis. @Abstractpublished in Advance ACS Abstracts, June 15, 1995. Abbreviations: RA, all-trans-(all-E)-retinoic acid; (13Z)-RA, 13cis-retinoic acid; PGH, prostaglandin H; PGGz, prostaglandin Gz; PGH2, prostaglandin Hz; RSVM, ram seminal vesicle microsomes; PPHP, 5-phenyl-4-penten-1-yl hydroperoxide; ESR, electron spin resonance spectroscopy; HPLC, high performance liquid chromatography; t ~retention , time; PBN, a-phenyl-N-tert-butylnitrone; PCI, positive chemical ionization (mass spectrometry);NCI, negative chemical ionization (mass spectrometry); TPA, 12-0-tetradecanoylphorbol 13-acetate. 0893-228~!95/2708-0807$09.00l0
8). Phorbol esters elicit a variety of biological responses
including the stimulation of prostaglandin biosynthesis due to enhanced arachidonic acid mobilization from phospholipid stores (9). RA has itself been shown to stimulate arachidonic acid release and prostaglandin biosynthesis in cultured MDCK cells (IO),guinea pig skin (111,and fetal mouse bone culture, and in vivo in the rat as evidenced by urinary prostanoid levels (12). Prostaglandin H (PGH) synthase is found in virtually all mammalian tissues including skin, kidney medulla, lung, brain, urinary bladder, and other tissues (reviewed in refs 13 and 14). Thus, oxidation of RA and (13Z)-RA by PGH synthase may represent a pharmacologically- and/or toxicologically-relevant pathway of retinoid metabolism in skin and in other organ systems. PGH synthase catalyzes the first two reactions in the prostaglandin pathway. The cyclooxygenase activity of PGH synthase oxygenates arachidonic acid to the hydroperoxy endoperoxide prostaglandin G2 (PGGd, and its peroxidase activity reduces PGG2 to the hydroxy endoperoxide PGHz (reviewed in refs 14-16). PGHz is converted to prostaglandins, thromboxanes, or prostacyclin by tissue-specificisomerases and/or reductases. Both the cyclooxygenase and peroxidase activities of PGH synthase require heme (ferric protoporphyrin E), and various nonsteroidal antiinflammatory agents (e.g., aspirin, indomethacin) inhibit PGH synthase cyclooxygenase without inhibiting the peroxidase. The peroxidase exhibits broad substrate specificity reducing (in addition to PGGd various fatty acid hydroperoxides, alkyl hydroperoxides, and H202. Analogous to other heme peroxidases, PGH synthase's peroxidase catalytic cycle involves the generation of enzyme-derived higher oxidation states: compounds I and 11. Reaction of the enzyme with
0 1995 American Chemical Society
808 Chem. Res. Toxicol., Vol. 8, No. 5, 1995
Samokyszyn et al.
Table 1. Initial Rates of Dioxygen Uptake Associated with PPHP-Dependept, PGH Synthase-Catalyzed Oxidation of RA and (lSZ)-RAa RA (13Z)-RA RSVM (nmol of 02.min-l. 160 f 11 406 f 15 (mg of purified enzyme (pM Oymin-l) 20 f 1 51 f 7
-i
a The values represent the mean f SD of triplicate measurements. b RSVM (0.5 mg of proteidml) or purified PGH synthase (1pM) was preincubated with retinoid (150 pM) for 3 min at 37 "C followed by the addition of PPHP (200 pM). Dioxygen uptake was measured with a Clarke electrode.
an equivalent of hydroperoxide yields compound, I which is two oxidizing equivalents above the resting ferric state and consists of a heme-derived ferryl intermediate (FeW=O)and porphyrin ~d cation radical. One-electron reduction of compound I by a peroxidase reducing substrate yields the ferryl intermediate compound 11, and a second one-electron reduction regenerates the resting ferric state. These catalytic intermediates are powerful oxidants, exhibiting redox potentials of -1 V, that have been shown to oxidize a large variety of reductants by sequential one-electron transfer mechanisms including various phenols, catechols, hydroquinones, aromatic amines, 1,3-dicarbonyls, and others (16). FL4 and (13Z)-RA contain four allylic carbons, three primary allylic methyls ( C I ~C19, , and CZO) and a secondary endocylic allylic carbon at C4 of the /3-ionone ring. We hypothesized that these C-H bonds may undergo oxidation by compounds I and I1 of PGH synthase peroxidase by an H-atom abstraction mechanism because of significantly lower bond dissociation energies compared with "nonactivated" primary or secondary C-H bonds (17). The C-H bond a t C4 is expected to be significantly more reactive because bond homolysis at this position would yield a secondary (versus primary) allylic carbon radical and the position is endocyclic and should therefore yield maximum p-orbital overlap. We have previously reported that (13Z)-RA functions as a peroxidase reducing substrate and undergoes hydroperoxide- or arachidonic acid-dependent cooxidation by PGH synthase (28, 19). In the present study we report that RA also functions as a peroxidase cosubstrate for PGH synthase. We report evidence that hydroperoxide-dependent, PGH synthase-catalyzed oxidation of RA involves H-atom abstraction at C4, yielding RA-derived carbon-centered radicals which react with dioxygen, yielding peroxyl radicals. These peroxyl radicals react bimolecularly with RA by addition reactions, resulting in epoxidation across the 5,6-double bond, or react by reversible addition reactions, yielding (1321-W and other geometric isomers.
Materials and Methods Materials. RA and (13Z)-RA were purchased from Eastman Kodak (Rochester, NY)and stored at -20 "C under argon. Retinoid stock solutions were prepared fresh in MezSO (Aldrich, Milwaukee, WI), and all procedures involving retinoids were carried out in the dark or under yellow light. Synthetic 4-oxoRA, 4-hydroxy-RA, and 5,6-epoxy-RA were generous gifts from Dr. William E. Scott and Dr. Peter F. Sorter, Hoffmann LaRoche Inc. (Nutley, NJ). 5-Phenyl-4-penten-1-yl hydroperoxide (PPHP) was synthesized as described by Weller et al. (20). Arachidonic acid was purchased from Nu Chek Prep (Elysian, MN). a-PhenylN-tert-butylnitrone (PBN), nitrosobenzene, 2-methoxyphenol and (guaiacol), Tween 20, l-methyl-3-nitro-l-nitrosoguanidine, hydrogen peroxide (30%nominal concentration) were purchased from Aldrich, and the latter was titrated iodometrically using
O 0.0
Z
0.5
2.0
1.5
1.0
[PGH Synthase] (pM)
E
h
zoc -
,
,
25
50
. n o 0
\ 75
100
125
150
[Z-Methoxyphenol] (pM) Figure 1. (A) Dioxygen uptake as a function of PGH synthase concentration. RA (150 pM) was preincubated for 3 min at 37 "C in 0.1 M phosphate (pH 7 . Q containing 200 pM Tween 20, with increasing concentrations of purified PGH synthase followed by the addition of PPHP (200 pM). Dioxygen uptake was measured with a Clark electrode, and values represent the mean f S D of triplicate measurements. (B) Inhibition of PGH synthase-catalyzed oxidation of RA by 2-methoxyphenol. RA (150 pM) was preincubated with RSVM (0.5 mg of p r o t e i d m l ) and increasing concentrations of 2-methoxyphenol at 37 "C for 3 min in 0.1 M phosphate (pH 7.8) followed by the addition of PPHP (200 pM). Dioxygen uptake was measured with a Clark electrode, and the values represent the mean *SD of triplicate measurements. KIOs-standardized S20s2-. Ram seminal vesicle glands were purchased from Oxford Biomedical Research Inc. (Oxford, MI) and ram seminal vesicle microsomes (RSVM)prepared as described Marnett and Wilcox (21).Purified PGH synthase was prepared from RSVM as described by Marnett et al. (22). The enzyme was desalted over Sephadex G-25 (Sigma, St. Loius, MO) using 0.1 M phosphate (pH 7.8) containing 200 pM Tween 20 as the mobile phase, to remove diethyldithiocarbamate which was present in the enzyme preparation as a stabilizer. PGH synthase concentration was determined by measuring absorbance at 410 nm (E410= 0.125 pM-l cm-I (23)),and the enzyme was shown to be 48% holoenzyme. Heat-denatured RSVM or purified enzyme was prepared by subjecting the samples to a boiling water bath for 3 min. Reactions involving RSVM were carried out in Chelex 100 (Bio-Rad, Hercules, CAI-treated 0.1 M phosphate (pH 7.8). Reactions involving purified PGH synthase were carried out in Chelex 100-treated 0.1 M phosphate (pH 7.8) containing 200 pM Tween 20. The Tween 20 was present to stabilize the enzyme and solubilize the retinoids. All other reagents were obtained through commercial sources.
Chem. Res. Toxicol., Vol. 8, No. 5, 1995 809
Peroxidase-Catalyzed Oxidation of Retinoic Acid
C
D
20 Gauss Figure 2. ESR spectrum of PBN-derived nitroxides generated during PPHP-dependent, P G k synthase-catalyzed oxidation of RA. (A) Complete system (2 mg/mL RSVM,10 mM PBN, 600 pM RA, and 800 pM PPHP at ambient temperature in 0.1 M phosphate (pH 7.8)); (B) complete system minus P P H P (C) complete system minus RA, (D) complete system containing heat-denatured RSVM instead of viable microsomes. ESR spectra were obtained using a gain of 1.6 x lo5, modulation amplitude of 2.0 G, microwave power of 20 mW, scan time of 8 min, and time constant of 1.0 s.
Analytical Methods. Electronic absorption spectra were obtained using a n Aminco DW-2a UV-visible spectrophotometer or Hewlett-Packard 8452A diode array spectrophotometer. Analytical reverse phase HPLC analysis was carried out using a Waters Novapak C18 125 4 p m 3.9 x 300 mm HPLC column and a Waters 600E HPLC with a 994 programmable photodiode array detector. The latter allowed the acquisition of on-line electronic absorption spectra of eluents. Dioxygen uptake associated with hydroperoxide-dependent, POH synthasecatalyzed oxidation of RA or (13Z)-RA was measured polarographically with a YSI 5300 oxygen monitor equipped with a YSI Clark electrode (YSI Inc., Yellow Springs, OH) and connected to an r - y chart recorder. Initial rates of dioxygen incorporation were determined from initial slopes of the LO21 versus time curves. Ambient 0 2 concentration was 210 FM in the aqueous phase, as determined by calibration of the oxygen polarograph (241, and assay conditions are described in the figure legends. Mass spectra were obtained using a Hewlett-Packard system consisting of a 5988A mass spectrometer interfaced to a 59980A particle beam interface. Samples were introduced into the interface via a Hewlett-Packard 1090 Series I1 HPLC. The HPLC was configured without a column (injector output connected directly to the particle beam input), and purified samples were introduced in neat acetonitrile a t a flow rate of 0.5 mIJ min. Ten microliters of each sample was injected. The particle beam was maintained a t 60 "C with 35 psi helium at the nebulizer. The mass spectrometer was configured in chemical ionization mode using methane as the reagent gas at a source pressure of 1 x low4Torr. The source temperature was maintained a t 250 "C. Separate injections were evaluated in both positive (PCI) and negative (NCI) chemical ionization modes with programmed scanning from mlz 50 to m l z 350. ESR spectra were obtained using a Varian E-104 spectrometer custom-interfaced wih an IBM-compatible computer for data acquisition and analysis. All spectra were stored on the computer for later analysis, and signal intensities were measured from the stored spectra using software written by Duling (NIEHS) (25). All reactions were carried out at room temperature. Immediately after preparation, solutions were aspirated into a 10.5-mm Wilmad flat cell centered in the TMllo microwave
0
0
20
40
60
80
100
[Nitrosobenzene] (yM) Figure 3. Inhibiton of PGH synthase-catalyzed oxidation of RA by nitrosobenzene. RA (150 yM), RSVM (0.5 mg of proteid mL), and increasing concentrations of nitrosobenzene were preincubated for 3 min a t 37 "C followed by the addition of PPHP (200 pM).Dioxygen uptake was measured with a Clark electrode, and the values represent the mean f S D of triplicate measurements. Inset: Plot of reciprocal velocities of 0 2 uptake versus concentration of nitrosobenzene. cavity. To avoid the possibility of transition metal contamination, polyethylene tubing rather than stainless steel was used for the aspiration. Because sample aspiration permits the flat cell to be cleaned and refilled without being disturbed, reproducibility between scans was greatly improved. The composition of reaction mixtures and instrumental parameters employed in
810 Chem. Res. Toxicol., Vol. 8, No. 5, 1995 the ESR experiments are described in the figure legends. Electronic Absorption Spectroscopy of Microsomal Reaction Mixtures. Electronic absorption spectra (220-400 nm) of RA and its metabolites in various incubation mixtures involving microsomal PGH synthase-dependent metabolism were recorded with the Aminco DW-2a spectrophotometer at ambient temperature. Incubations consisted of RSVM or heatdenatured RSVM (0.1 mg of p r o t e i d m l ) and RA (10 pM) followed by the addition of HzOz, PPHP, or arachidonic acid (100 pM). Analysis of RA Metabolites. RA (150 pM) was preincubated with purified PGH synthase (1 pM) for 3 min at 37 "C followed by the addition of PPHP or HzOz (200 pM) in a total volume of 1.3 mL. After 5 min, reaction mixtures were acidified to pH 3.0 with HC1, saturated with NaC1, and extracted with HPLC-grade ethyl acetate. The ethyl acetate was removed i n vacuo, and the residue was dissolved in HPLC-grade methanol or acetonitrile and filtered through 0.45 pm nylon filters (Scientific Resources, Inc., Eatontown, NJ). Analytical reverse phase HPLC was carried out using the chromatographic system described above and a solvent system consisting of 20% solvent B in solvent A at t = 0 followed by a linear gradient to 90% solvent B over 30 min followed by isocratic conditions for 25 min. Solvent A consisted of 5 0 5 0 methanoliwater containing 0.01 M ammonium acetate (pH 6.651, and solvent B consisted of 9O:lO methanoliwater containing 0.01 M ammonium acetate (pH 6.65). The flow rate was 1.0 m u m i n , and eluents were detected by absorbance a t 343 nm. The major metabolites were identified on the basis of cochromatography with synthetic standards, UV spectroscopy, and/or mass spectrometry. Samples for mass spectrometry were obtained by collecting the HPLCeluted material obtained from scaled-up reaction mixtures of purified enzyme, RA, and HzOz using the concentrations described above (total volume = 100 mL). The HPLC solvent was removed i n vacuo, and the corresponding methyl esters were prepared by reaction with ethereal diazomethane generated by reaction of 1-methyl-3-nitro-1-nitrosoguanidine with KOH at 4 "C (26). Caution: It should be noted that diazomethane is a potent electrophile and is potentially explosive. Thus, diazomethane generation and handling should be carried out in a fume hood, behind protective shielding, and using gloves and protective clothing.
Results Preliminary evidence that RA undergoes PGH synthase-catalyzed metabolism was obtained by ultraviolet spectroscopic analysis of microsomal incubation mixtures. Addition of PPHP, HzOz, or arachidonic acid (100 pM) to a suspension of RA (10 pM) and RSVM (0.1 mg of proteid mL) resulted in a slight blue shift of the retinoid absorption maximum from 347 to 343 nm as well as a 40% decrease in molar absorptivity. In addition, there was an increase in absorbance at lower wavelengths (250-306 nm) with an isosbestic point at 306 nm. All spectral changes were eliminated when incubations were carried out with heat-denatured microsomes, indicating that retinoid metabolism is enzymatic. These results suggest that RA is metabolized by PGH synthase, resulting in the formation of products with altered extents of conjugation. Hydroperoxide-dependent, PGH synthase-catalyzed metabolism of RA was also characterized by uptake of dioxygen. PPHP was used as the hydroperoxide substrate in 0 2 uptake experiments because RSVM cosedimented with significant catalase activity, which resulted in dioxygen evolution when HzOz was used as substrate (data not shown). Hydroperoxide-dependent uptake of 0 2 associated with RA metabolism was observed using microsomal as well as purified PGH synthase (Table 1). We have previously reported that PGH synthase-cata-
Samokyszyn et al.
lyzed oxidation of (132)-RA is also characterized by uptake of dioxygen (18, 19). As shown in Table 1, the 132 isomer was considerably more reactive compared with RA and was characterized by -2.5 times greater initial rates of 0 2 uptake in microsomal as well as purified enzymatic systems. Dioxygen uptake in these systems was not observed in the absence of hydroperoxide or retinoid, or when viable microsomes or enzyme were substituted with heat-denatured preparations. Dioxygen uptake was first order with respect to enzyme concentration (Figure 1A) and was first order with respect to RA at the concentrations tested (0-80 pM) in microsomal systems (data not shown). 2-Methoxyphenol (guaiacol) effectively inhibited 0 2 uptake (Figure 1B). 2-Methoxyphenol is an effective peroxidase reducing substrate for PGH synthase (16,271, suggesting competition of 2-methoxyphenol with RA for oxidation by compounds I and I1 of PGH synthase peroxidase. Collectively, these results (as well as other data presented below) suggest that the higher oxidation states of PGH synthase peroxidase (compound I and 11)oxidize RA by an H-atom abstraction mechanism, generating retinoid-derived carbon-centered radicals. These radicals couple with dioxygen, yielding RA-derived peroxyl radicals, resulting in the observed uptake of dioxygen:
RH
-R' -He
Oz
ROO'
The rate of dioxygen coupling with carbon-centered radicals is diffusion limited, exhibiting rate constants on the order of -lo9 M-l s-l (28). Thus, the rate-limiting step in eq 1involves the formation of RA-derived carbon radicals. PGH synthase's peroxidase catalytic cycle, like other peroxidases, is characterized by ping pong kinetics (16). Therefore, the initial rates of the above H-atom abstraction reaction, resulting in the generation of retinoid-derived carbon radicals (R),may be expressed by: d[R'lldt = [RHl(k,[compound I1
+ k,[compound 111) (2)
where [RH] represents the concentration of RA, and kl and kz represent the rate constants for H-atom abstraction by compounds I and 11, respectively. Because reduction of compound I1 to the resting ferric state represents the rate-limiting step in the peroxidase catalytic cycle (16),eq 2 simplifies to: d[R'lldt = k,[RHl[compound I11
(3)
The steady-state concentration of compound I1 is neccessarily proportional to total enzyme concentration, thus accounting for the first order dependence on enzyme concentration (Figure 1A). The first order dependence on RA concentration is also consistent with eq 3 at less than saturating concentrations. Further evidence that RA oxidation involves a free radical mechanism (eq 1)was obtained from ESR spin trapping experiments using a-phenyl-N-tert-butylnitrone (PBN) as the spin trap. As shown in Figure 2, reaction mixtures containing RSVM, PBN, RA, and PPHP yielded ESR-detectable nitroxides characterized by hyperfine coupling constants of aN = 15.8 G and a~ = 3.2 G. No signal was observed in the absence of RA or PPHP, and a significantly diminished signal was observed when viable microsomes were substituted with heat-denatured preparations. Similar PBN-derived nitroxides, charac-
Chem. Res. Toxicol., Vol.8, No. 5,1995 811
Peroxidase-Catalyzed Oxidation of Retinoic Acid
A343
N
5
0
10
20
30 Time (min)
40
50
0
10
20
30
4Q
50
A343
Time (min)
Figure 4. HPLC elution profiles of RA metabolites generated during PPHP-dependent, purified PGH synthase-catalyzed oxidation of RA. The incubation, metabolite isolation, and HPLC conditions are described in the Materials and Methods section. Upper panel: complete system; lower panel: complete system minus enzyme.
terized by hyperfine coupling constants of a N = 15.6 G and a H = 3.6 G, have previously been detected during PPHP-dependent, PGH synthase-catalyzed oxidation of (13Z)-RA (19).We have not yet characterized the structures of the PBN-RA or PBN-(13Z)-RA nitroxide adducts. However, they may represent carbon-centered radical adducts because similar hyperfine coupling constants have been reported for various PBN-(carboncentered radical) adducts generated in other in vitro and in vivo systems (29).For example, the nitroxide adduct obtained by reaction of PBN with isopropyl radicals, generated during horseradish peroxidase-catalyzed oxidation of iproniazid, is characterized by values of aN and aH of 15.7 and 3.1 G, respectively (30). Equation 1 is further supported by the ability of the spin trap nitrosobenzene to inhibit dioxygen uptake (Figure 3). Nitroso spin traps react with carbon-centered radicals, yielding nitroxides (311. Furthermore, the bimolecular rate constant for the reaction of carboncentered radicals with nitrosobenzene, under homogeneous liquid phase conditions, is estimated to be -lo8 M-l s-l (32). Thus, the ability of nitrosobenzene to inhibit 02 uptake associated with PPHP-dependent,PGH synthase-catalyzed oxidation of RA probably reflects competition of nitrosobenzene with 02 for reaction with RA-derived carbon-centered radicals. Interestingly, the nitrosobenzene-dependent inhibition of dioxygen uptake was characterized by reciprocal initial velocities which were proportional with nitrosobenzene concentration as shown in the inset in Figure 3. This result may be interpreted by assuming that the rate of 02 uptake is
expressed by the equation: -d[02Ydt = k3[R1[O21
(4)
where K B represents the bimolecular rate constant for the coupling of R with dioxygen, yielding RA-derived peroxyl radicals. If a steady-state approximation is assumed where the rate of R formation equals the s u m of the rates of R scavenging by dioxygen and nitrosobenzene, we obtain the following equation, where K q is the bimolecular rate constant for the reaction of R with the spin trap: d[R'Ydt = 0 = k,[RHl[compound I11 - k,[R'I[O,I k,[R'l[nitrosobenzenel Solving for
(5)
[a], inserting the expression for [RI into eq
4, and taking the reciprocal, we obtain the following
equation: (-d[O,Ydt)-l
= (k,[O,]
+
k,[nitrosobenzenel)/k2k3[RHl[compound IIl[O,l ( 6 ) This rate law accounts for the linearity of the reciprocal velocity of dioxygen consumption with nitrosobenzene concentration (Figure 3, inset). To gain further mechanistic insights, and to determine the regioselectivity or regiospecificity of H-atom abstraction, we have attempted to identify several of the major RA-derived reaction products generated during PGH synthase-catalyzed oxidation of RA. RA was inclubated
Samokyszyn et al.
812 Chem. Res. Toxicol., Vol. 8, No. 5, 1995 Table 2. PCI and NCI Mass Spectrometric Patterns of Metabolites 2 (Upper) and 3 (Lower), and Proposed Fragmentation Assignments
B
C
D
Table 3. Spectral Characteristics of RA-Derived Geometric Isomers tR
metabolite 4
E
5
A
-
FOH a
PCI NCI
331 329
B+ B+D B CH3 (CH& C D + CH3 B+
B+
330 313 299 192 176 328 313 193
162
ioni- molecular zation iona -H
A
E
146 130
116 116
Cz~HloOj: 330 ioni- molecular B+ B+ B+ B+D zation ion= -H -CH3 E* B CH3 (CHd2 C D +CHs PCI 331 330 315 299 192 176 162 116 NCI 329 193 146 130 116
M
+ H+ for PCI, M - H for NCI.
Or loss of epoxide OCH3.
with PPHP in the presence or absence of purified PGH synthase, and the reaction products were isolated and analyzed by reverse phase HPLC as described in the Materials and Methods section. Complete reaction mixtures resulted in the formation of several major products (Figure 4, upper) which were not detected in the absence of enzyme (Figure 4, lower). The major peak eluting at 44.6 min represents unmetabolized RA. Metabolite 1( t ~ = 22.7 min) cochromatographed with synthetic 4-oxo-RA and exhibited identical electronic absorption spectra characterized by absorption maxima at 364 and 286 nm (A36dAZS6 = 2.7). Metabolite 2 ( t = ~ 25.3 min) cochromatographed with synthetic 4-hydroxy-RA and exhibited an identical UV spectrum characterized by a ,A of 340 nm. The PCI and NCI mass spectra of its methyl ester, generated by reaction with ethereal diazomethane, were characteristic of a hydroxylated product (Table 2, upper). Molecular ions were observed for the methyl ester as the M H+ by PCI and M - H- by NCI. Predominate loss of an additional hydrogen was observed in all spectra, which seems to be characteristic of particle beam introduced samples by chemical ionization (33). Metabolite 2 methyl ester demonstrated a significant fragment (mlz 313, PCI) showing the loss of water, confirming the presence of the hydroxyl group. Fragmentation was also characterized by loss of methoxyl and carboxyl (D and E sites in Table 2) fragments and cleavage at susceptible bonds (B and C sites in Table 2). The detection of 4-hydroxy-RA and 4-0xo-RA as metabolites are consistent with PGH synthase-catalyzed H-atom abstraction at Cq of the retinoid j3-ionone ring. Metabolite 3 ( t =~31.8 min), cochromatographed with synthetic 5,6-epoxy-M, exhibited identical U V spectra characterized by a A, of 330 nm, and its methyl ester yielded PCI and NCI mass spectra characteristic of the 5,6-epoxide (Table 2, lower). As with 4-hydroxy-methyl-
+
Amax
( m i d (nm) PCI mass spectra (mlz) 40.9 352 315 (M H)+ 314 M+ (37.71, 313 (8.3),299 (5.0), 283 (17.6), 219 (9.6), 201 (6.21, 175 (5.6), 137 (8.2), 123 (17.5) 41.8 350 315 (M + HI+ (loo), 314 M+ (501, 313 (111,299 (5.71, 283 (25.3), 219 (13.41, 201 (8.81, 175 (7.71, 137 (7.31, 123 (23.0)
+
Numbers in parentheses represent relative abundance.
retinoate (metabolite 21, the mass spectra of the 5,6epoxy-RA methyl ester were characterized by loss of methoxyl and carboxyl (D and E sites in Table 2) fragments and cleavage at susceptible bonds (B and C sites in Table 2). We have not identified the metabolite eluting at 30.1 min and are attempting to obtain sufficient mass of this product for unambiguous identification by NMR, FT-IR, and mass spectrometric analysis. Peaks 4 and 5 represent geometric isomers of RA, as evidenced by mass spectrometric and UV spectroscopic data (Table 3). Peak 5 ( t = ~ 41.8 min) cochromatographed with authentic (13Z)-RA and exhibited identical UV spectra characterized by a Amax of 350 nm. In addition, the methyl ester of metabolite 5, generated by reaction with ethereal diazomethane, exhibited a molecular mass (M+ = m / z 314) and fragmentation pattern characteristic of a geometric isomer (19). Metabolite 4 ( t =~ 40.9 min) also represents a geometric isomer as evidenced by the UV spectrum which was characterized by a A, of 352 nm and the mass spectrum of its methyl ester which was nearly identical to that of the 132 isomer (Table 3). However, we have not determined its absolute configuration. Peaks 6 ( t =~ 42.4 min, Amax = 338 nm) and 7 ( t =~44.0 min, Am= = 344 nm), which were poorly resolved from (132)-RA and RA,respectively, may also represent geometric isomers because of the similar retention times compared with RA,(132)-RA, and metabolite 5 as well as their electronic absorption spectra. We have not attempted to further characterize these metabolites. A summary of RA metabolites generated via PPHPdependent, PGH synthase-catalyzed oxidation is shown in Figure 5. Similar product profiles were obtained using H202, instead of PPHP, as the hydroperoxide substrate or using RSVM instead of purified enzyme (data not shown).
Discussion Our results indicate that RA undergoes hydroperoxide (HzOz or PPHP)- or arachidonic acid-dependent, PGH synthase-catalyzed metabolism as evidenced by ultraviolent spectroscopic analysis of reaction mixtures. PGH synthase-catalyzed metabolism of RA was characterized by uptake of dioxygen (Table 1, Figure 1) which was inhibited by 2-methoxyphenol (guaiacol), an effective peroxidase reducing substrate for PGH synthase (16,26). In addition, dioxygen uptake was inhibited by the spin trap nitrosobenzene (Figure 3). ESR spin trapping studies, using PBN as the spin trap, demonstrated the formation of retinoid-derived free radical intermediates, and the ESR spectrum is consistent with the formation of a carbon-centered radical adduct. We have also identified several major and minor products generated during hydroperoxide-dependent, PGH synthase-cata-
Chem. Res. Toxicol., Vol. 8, No. 5, 1995 813
Peroxidase-Catalyzed Oxidation of Retinoic Acid
0
t
t
Other Geometric Isomers and Unidentified Oxidation Products Figure 5. Summary of metabolites generated during hydroperoxide-dependent, PGH synthase-catalyzed cooxidation of RA. Numbers correspond t o HPLC peaks in Figure 4.
(pq
Scheme 1
H
H
H
0.
0.
lyzed oxidation of RA, including 4-hydroxy-RA, 5,6-epoxyRA, 4-0xo-RA, (13Z)-RA, and other geometric isomers (Figure 5). Collectively, these results suggest that the higher oxidation states of PGH synthase peroxidase (compounds I and 11) oxidize RA at the C4 position by H-atom (or electron) abstraction, yielding RA-derived carbon-centered radical intermediates. These RA carbon radicals couple with dioxygen, yielding RA-derived peroxy1 radicals (Scheme 1). This mechanism is consistent with the observed 2-methoxyphenol- and nitrosobenzeneinhibitable uptake of 0 2 , the ESR spin trapping experiment, and hydroxylation at the C4 position. Alternatively, as shown in Scheme 1, oxidation of RA could occur via an electron abstraction mechanism, yielding retinoid radical cation intermediates which subsequently deprotonate.
Our results also demonstrated that dioxygen uptake, associated with PGH synthase-catalyzed oxidation of RA, was first order with respect to enzyme concentration (Figure 1A). These results are consistent with eq 3. Futhermore, these results indicate that oxidation of RA in this system is not characterized by free radical propagation reactions. The rate law characterizing the autoxidation of hydrocarbons is represented by the following equation (28):
where k, is the propagation rate constant, [RHI the concentration of oxidizable substrate (retinoid in our system), Ri the rate of initiation, and kt the rate constant for termination. Thus, if hydroperoxide-dependent,PGH synthase-catalyzed oxidation of RA was characterized by free radical propagation reactions, the rate of dioxygen uptake would exhibit a square root dependence on enzyme concentration because enzyme-bound oxidants (compounds I and 11)function as the H-atom (or electron) abstracting species. The first order dependence of the enzyme concentration is consistent with ping pong kinetics, which characterizes the catalytic mechanism of PGH synthase peroxidase (16). Several mechanisms may account for the lack of propagation reactions. PGH synthase is very susceptible to hydroperoxide-dependent inactivation (14,15). This autoinactivation is probably a consequence of oxidation of critical amino acid residues in the active site by the peroxidase higher oxidation states andlor hydroperoxide-dependent heme degradation. Consequently, the rate of PGH synthase-catalyzed H-atom abstraction rapidly decreases, whereas hydrocarbon autoxidation requires a steady-state condition where the rate of initiation equals the rate of termination (28). In addition, conjugated polyenes containing allylic C-H bonds react preferentially with peroxyl radicals by
Samokyszyn et al.
814 Chem. Res. Toxicol., Vol. 8, No. 5, 1995
Scheme 4
Scheme 2
R
Scheme 3
c 0
addition rather than H-atom abstraction mechanisms (34). We have recently demonstrated that RA reacts with chemically-generated peroxyl radicals by addition reactions, yielding the 5,6-ep0xide.~Thus, the identification of 5,6-epoxy-RA as a major product generated during hydroperoxide-dependent, PGH synthase-catalyzed oxidation of RA is consistent with the bimolecular addition of RA-derived peroxyl radicals across the 5,g-doublebond, yielding the corresponding oxirane (Scheme 2). The detection of (13Z)-RA and other geometric isomers is consistent with reversible peroxyl radical addition reactions. Addition of peroxyl radicals across double bonds of the conjugated polyene system would impart considerable delocalized single bond character capable of undergoing C-C bond rotation. Subsequent @-scissionof the peroxide intermediate could result in geometric isomerization (Scheme 3). The detection of 4-oxo-RA suggests V. M. Samokyszyn, manuscript in preparation.
OH
that RA-derived peroxyl radicals may undergo coupling to tetroxide intermediates. These tetroxides may subsequently undergo a Russell rearrangement (351,yielding 4-hydroxy-RA, 4-0xo-u, and (or triplet oxygen and an excited carbonyl) (Scheme 4). This mechanism may partially account for the ultimate fate of RA-derived peroxyl radicals. Alternatively, the peroxyl radicals may abstract hydrogens from amino acid residues of the protein or may undergo reduction by microsomal low molecular weight reductants. RA, (13Z)-RA, and other retinoids have been shown to inhibit tumor promotion in the two-stage (initiationpromotion) mouse skin assay (6). The assay involves exposure of epidermis to a mutagen such as 7,12dimethylbenz[alanthracene followed by the application of a promotor such as croton oil-derived phorbol esters (e.g., 12-0-tetradecanoylphorbol13-acetate (TPA)). However, under certain conditions, RA exhibits the opposite effect and actually promotes tumorigenesis. For example, RA was shown to function as a potent promoter when applied to ultraviolet-irradiated skin ( 7 ) and as a weak promoter when applied to dimethylbenzanthracene-initiated skin in place of TPA (8).The molecular mechanisms underlying the paradoxical activities of RA in tumor promotion have not been elucidated. However, peroxyl radicals may play a role in tumor promotion. For example, Marnett et al. have unequivocally demonstrated that TPA stimulates peroxyl radical production in mouse skin (36). In addition, the free radical initiator benzoyl peroxide as well as other peroxides, hydroperoxides, and peracids are themselves effective tumor promotors in the mouse skin assay (37-39). Conversely, butylated hydroxyanisole and other phenolic antioxidants, which effectively scavenge peroxyl radicals, are potent inhibitors of TPA-dependent promotion (40). Thus, PGH synthasecatalyzed oxidation of RA, yielding RA-derived peroxyl radicals, may be relevant under these conditions and may play a role in the promotional effects of RA.
Peroxidase-Catalyzed Oxidation of Retinoic Acid
Acknowledgment. "his work was supported by Grant 1-R29ES06765-01 from the NIH.
References (1) Kelly, A. P. (1990) Acne and related disorders. In Principles and
Practice ofDermatology (Sams, W. M., Jr., and Lynch, R. J., Eds.) pp 781-796, Churchill Livingstone, New York. (2) Weiss, J. S., Ellis, C. N., Headington, J. T., Tincoff, T., Hamilton, T. A., and Voorhees, J. J. (1988) Topical tretinoin improves photoaged skin. A double-blind vehicle-controlled study. J . Am. Med. Assoc. 269, 527-532. (3) Hong, W. K., and Itri, L. M. (1994) Retinoids and human cancer. In The Retinoids. Biology, Chemistry, and Medicine. Second Edition (Sporn, M. B., Roberts, A. B., and Goodman, D. S., Eds.) pp 597-630, Raven Press, New York. Hill, D. L., and Sani, B. P. (1991) Metabolic disposition and development of new chemopreventive retinoids. Drug Metab. Rev. 23,413-438. Lehman, P. A., and Malany, A. M. (1989) Evidence of percutaneous absorption of isotretinoin from the photo-isomerization of topical tretinoin. J. Invest. Dermatol. 93, 595-599. Moon, R. C., Mehta, R. G., and Rao, K. V. N. (1994) Retinoids and cancer in experimental animals. In The Retinoids. Biology, Chemistry, and Medicine. Second Edition (Sporn,M. B., Roberts, A. B., and Goodman, D. S., Eds.) pp 573-595, Raven Press, New York. Forbes, P. D., Urbach, F., and Davies, R. E. (1979) Enhancement of experimental photocarcinogenisis by topical retinoic acid. Cancer Lett. 7, 85-90. Hennings, H., Wenk, M. L., and Donahoe, R., (1982)Retinoic acid promotion of papilloma formation in mouse skin. Cancer Lett. 16, 1-5.
Fischer, S. M. (1984) The role of prostaglandins in tumor production. In Mechanisms of Tumor Promotion. Volume II. Tumor Promotion and Skin Carcinogenesis (Slaga, T. J., Ed.) pp 113-126, CRC Press, Boca Raton. Levine, L., and Ohuchi, K. (1978) Retinoids as well as tumour promoters enhance deacylation of cellular lipids and prostaglandin production in MDCK cells. Nature 276, 274-275. Ziboh, V. A,, Price, B., and Fulton, J. (1975) Effects of retinoic acid on prostaglandin biosynthesis in guinea-pig skin. J . Invest. Dermatol. 66, 370-374. Viau, C. (1985)The role of prostaglandins in retinoid toxicity. Dks. Abstr. Int., B 46, 126-127. Eling, T. E., Thompson, D. C., Foureman, G. L., Curtis, J. F., and Hughs, M. F. (1990) Prostaglandin H synthase and xenobiotic oxidation. Annu. Rev. Pharmacol. Toxicol. 30, 1-45. Smith, W. L., Marnett, L. J., and DeWitt, D. L. (1991) Prostaglandin and thromboxane biosynthesis. Pharmacol. Ther. 49,153179. Smith, W. L., and Marnett, L. J. (1991) Prostaglandin endoperoxide synthase: structure and catalysis. Biochim. Biophys. Acta 1083, 1-17. Marnett, L. J., and Madipatti, K. R. (1991) Prostaglandin H synthase. In Peroxidases in Chemistry and Biology, Volume 1 (Everse, J., Everse, K. E., and Grisham, M. B., Eds.) pp 293334, CRC Press, Boca Raton. McMillen, D. F., and Golden, D. M. (1982) Hydrocarbon bond dissociation enereies. Annu. Rev. Phvs. Chem. 33. 493-532. (18) Samokyszyn, V. Sloane, B. F., Hinn, K V., and Marnett, L. J. (1984) Cooxidation of 13-cis-retinoic acid by prostaglandin H synthase. Biochem. Biophys. Res. Commun. 124, 430-436. (19) Samokyszyn, V. M., and Marnett, L. J. (1987) Hydroperoxidedependent cooxidation of 13-cis-retinoic acid by prostaglandin H synthase. J. Biol. Chem. 262, 14119-14133. (20) Weller, P. E., Markey, C. M., and Marnett, L. J. (1985) Enzymatic reduction of 5-phenyl-4-pentenyl-hydroperoxide:Detection of peroxidases and identification of peroxidase reducing substrates. Arch. Biochem. Biophys. 243, 633-643. (21) Marnett, L. J., and Wilcox, C. L. (1977) Stimulation of prosta-
k,
Chem. Res. Toxicol., Vol. 8, No. 5, 1995 816 landin biosynthesis by lipoic acid. Biochim. Biophys. Acta 487, 222-230. Marnett, L. J., Siedlik, P. H., Ochs, R. C., Pagels, W. R., Das, M., Honn, K. V., Warnock, R. H., Tainer, B. E., and Eling, T. E. (1984) Mechanism of the stimulation of prostaglandin H synthase and prostacyclin synthase by the antithrombotic and antimetastatic agent, nafazatrom. Mol. Pharmacol. 26, 328-335. Mahy, J . P., and Mansuy, D. (1991) Formation of prostaglandin synthase-iron-nitrosoalkane inhibitory complexes upon in situ oxidation of N-substituted hydroxylamines. Biochemistry 30, 4165-4172. Holtzman, J. L. (1976) Calibration of the oxygen polarograph by the dedetion of oxmen with hvpoxanthine-xanthine oxidasecatalase. Anal. Ch& 48, 229-530. (25) Duling, D. (1994) Simulation of multiple spin-trap EPR spectra. J . Magn. Reson., Ser. B 104, 105-110. (26) Black, T. H. (1983) The preparation and reactions of diazomethane. Aldrichimica Acta 16, 3-10. (27) Markey, C. M., Award, A., Weller, P. E., and Marnett, L. J. (1987) Quantitative studies of hydroperoxide reduction by prostaglandin H synthase. Reducing substrate specificity and the relationship of peroxidase to cyclooxygenase activities. J . Biol. Chem. 262, .6266-6279. Pryor, W. A. (1984) Free radicals in autoxidation and in aging, Part I, kinetics of the autoxidation of linoleic acid in SDS micelles: Calculations of radical concentrations, kinetic chain lengths, and the effects of viatimin E, Part 11,the role of radicals in chronic human diseases and in aging. In Free Radicals in Molecular Biology, Aging and Disease: Aging (Armstrong, D., Sohal, R. S., Cutler, R. G., and Slater, R. F.,Eds.) pp 1-26, Raven Press, New York. Li, A. S. W., Cummings, K. B., Roethling, H. P., Buettner, G. R., and Chignell, C. F. (1988)A spin-trappingdatabase implemented on the IBM PC/AT. J. Magn. Reson. 79, 140-142. Sinha, B. K. (1987) Activation of hydrazine derivatives to free radicals in the perfused rat liver: A spin-trapping study. Biochim. Biophys. Acta 924, 261-269. Evans, C. A. (1979) Sin trapping. Aldrichimica Acta 12, 23-29. Schmid, P., and Ingold, K. U. (1978) Kinetic applications of electron paramagnetic resonance spectroscopy. 31. Rate constants for spin trapping. 1. Primary alkyl radicals. J . Am. Chem. SOC. 100,2493-2500. Bloom, J., Lehman, P., Israel, M., Rosario, O., and Korfmacher, W. (1992) Comparison of thermospray mass spectrometry to particle beam mass spectrometry for the characterization of three anthracyclines. J . Anal. Toxicol. 16, 223-227. Ingold; K U. (1969) Peroxy radicals. ACC.Chem. Res. 2, 1-9. Russell, G. A. (1956)Deuterium-isotope effects in the autoxidation of aralkyl hydrocarbons. Mechanism of the interaction of peroxy radicals. J. Am. Chem. SOC.79, 3871-3877. Marnett, L. J., Reddy, A., Pruess-Schwartz, D., and Ji, C. (1991) Oxidation of polycyclic hydrocarbons by oxygen radicals. In Xenobiotics and Cancer (Erster et al., Eds.) pp 63-73, Japan Scientific Societies Press, Tokyo/Taylor & Francis Ltd., London. Slaga, T. J., Klein-Szanto, A. J. P., Triplett, L. L., Yotti, L. P., and Trosko, J. E. (1981)Skin tumor promoting activity of benzoyl Deroxide. a widelv used free radical generating Science - compound. 213, 1023-1025: Slaaa. T. J.. Solanki, V.. and Logani, M. (1983) Studies on the meihanism of action of antitumor-promoting agents: Suggestive evidence for the involvement of free radicals in promotion. In Radioprotectors and Anticarcinogens (Nygaard, 0. F., and Simic, M. G., Eds.) pp 471-465, Academic Press, New York. Gimenez-Conti. I.. Viaie. A.. Chesner. J., Conti, C., and Slaga, T. J. (1991) Induction of short-term markers of tumor promotion by organic peroxides. Carcinogenesis 12, 563-569. (40) Kozumbo, W. J., Seed, J. L., and Kensler, T. W. (1983) Inhibition by 2(3)-tert-butyl-4-hydroxyanisoleand other antioxidants of epidermal ornithine decarboxylase activity induced by 12-0tetradecanoyl-13-acetate. Cancer Res. 43, 2555-2559. L
TX9500078
