Document e1x5Q5J6DeRGg93Vy0p6B06M4

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 241, No. 1, August 15, pp. 88-94, 1985 Inhibition of Microsomal Lipid Peroxidation by Naphthoquinones: Structure-Activity Relationships and Possible Mechanisms of Action' RONALD E. TALCOTT,* MARTYN T. SMITH,?' AND DONALD D. GIANNINIS *Brain Turn& Research of Califwnia, San Berkeley, Center, Departments of Laboratory Medicine and Neurological Surgery, University Francisco, California 9U&Y; tSchoo1 of Public Health, University of Califoria, California 94720; and $Department of Pharmaceutical Chemistry, University of California, San Francisco, California 9414.3 Received November 27, 1984, and in revised form April 30, 1985 Menadione (2-methyl-1,4-naphthoquinone) is a remarkably potent inhibitor of microsomal lipid peroxidation, effective at submicromolar concentrations. Its possible mechanism of action and the relationship between naphthoquinone structure and antioxidant activity were the topics of this investigation. In the microsomal lipid- peroxidizing system dependent on NADPH and ferric pyrophosphate, menadione, at concentrations of 50 PM or higher virtually eliminated the accumulation of malondi- aldehyde and lipid hydroperoxides. In the NADPH-independent, cumene hydroperoxide- dependent system, menadione was also an effective antioxidant, but only in the presence of reducing equivalents. These and other observations indicate that a reduced form of menadione, either the hydroquinone or semiquinone, is the active antioxidant, and suggest that it may trap hydroperoxy radicals, alkoxy radicals, or other free radicals involved in propagating lipid peroxidation. Moreover, these results show that electron diversion per se cannot account for the antioxidant effects of menadione. A comparison of the antioxidant activities of eight 1,4-naphthoquinones indicated that methyl substitution of C-2, lack of steric hindrance at C-3 or C-5, and (in the case of weak acids) a relatively high pK, are favorable structural features associated with strong antioxidant activity. o iSa5 Academic PEW k. Microsomal lipid peroxidation, a process that may contribute to the toxicity of many xenobiotics (l), has been studied extensively since its discovery two decades ago (2). Much attention has been focused on the role of the microsomal electron transport system (3-5), chelated iron (6, 7), reduced oxygen (g-lo), and organic hydroperoxides (7, 11) in the initiation and propagation of the peroxidative re- action. Present understanding of the events in the reaction sequence derives principally from studies performed in two `This paper is dedicated to the memory of Dr. R. E. Talcott who died tragically in an automobile accident in 1984. `To whom correspondence should be addressed. model systems. The first of these requires NADPH, NADPH cytochrome P-450 reductase, and chelated iron, and is activated by the reduction of the iron to an oxygenbinding species (6,7). The ferrous-oxygen complex rearranges to perferryl ion, which is believed to initiate lipid peroxidation by abstracting a methylene hydrogen from an unsaturated fatty acid, creating an alkyl radical that combines with oxygen to form a lipid hydroperoxide (6). The lipid hydroperoxide then serves as a source of other free radicals as the chain reaction is propagated by iron species such as ferric cytochrome P-450 (7). In the second model system, the requirements for reducing equivalents and non-heme iron are bypassed by initiating the reaction with 0003-9861/85 $3.00 Copyright 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved. 88 A.N.'~I,'..I~U-~.rX.. ILJAN'I' ..-- A.L'l_'-I.r I1IT".. V `II OF NAPHTHOQUINONES 89 an organic hydroperoxide such as cumene hydroperoxide (7, 11). In this system, ini- tiation is accomplished when the unsatu- rated fatty acid donates a hydrogen to the microsomal peroxidase, ferric cyto- chrome P-450 (11, 12). In the presence of sufficient butylated hydroxytoluene (13, 14), glutathione (14, 15), and other antioxidants (1, 15), micro- somal lipid peroxidation does not occur and the function of the microsomal elec- tron transport system is preserved. Men- adione (2-methyl-1,4-naphthoquinone) is an especially potent antioxidant, as it confers complete protection at submi- cromolar concentrations (16), but the mechanism of its action is not well un- derstood; this was addressed in a series of experiments described in this paper. MATERIALS AND METHODS Chemicals. NADPH, TRIZMA base, hydrogen per- oxide, and thialbarbituric acid were purchased from Sigma Chemic.al Company (St. Louis, MO.); cumene hydroperoxide, menadione, and silver oxide were purchased from Pfaltz and Bauer, Inc. (Stamford, Conn.); methyl1 iodide, sodium peroxide, palmitoyl chloride, and 5-hydroxy-2-methyl-1,4-naphthoquinone were from Aldrich Chemical Company (Milwaukee, Wise.); chromium trioxide and potassium iodide were from Mallinckrodt, Inc. (Paris, KY.); 2,3-dimethyl- naphthalene arnd 1,4-naphthoquinone were from Fluka Chemical Corporation (Hauppauge, N. Y.); and ferric pyrophosphate was obtained from K and K Laboratories, Inc. (Plainview, N. Y.) All organic solvents were reagent grade, from Mallinckrodt. Eight derivatives of 1,4-naphthoquinone were tested as inhibitors of microsomal lipid peroxidation and as stimulators of microsomal NADPH oxidation. As noted above, menadione, 1,4-naphthoquinone, and 5-hydroxymenadione were from commercial sources; the other five derivatives were synthesized as follows: 2,3-dimethyl-l,&naphthoquinone was prepared by chromic acid oxidation of 2,3-dimethylnaphthalene (17), 3-hydroxymenadione was prepared from men- adione by H202/H2S04 treatment as described by Fieser (18), 3-methoxy- and 5-methoxymenadione were prepared by silver oxide-catalyzed methylation of the corresponding hydroxy compounds (19), and pentadecylmenadione was prepared from menadione by dipalmitoyl peroxide alkylation (20). The dipal- mitoyl peroxide was prepared from palmitoyl chloride and sodium peroxide as described previously (20). The synthetic yields ranged from 30 to 100% and were similar to those previously reported. The re- action products were confirmed mass spectroscopy. by low-resolution Assays of microsomal lipid peroxidation Micro- somes were prepared from the livers of eight male Sprague-Dawley rats, 200-250 g. The rats were killed by exsanguination and the livers were excised im- mediately, perfused with 1.15% KCl, and homogenized with a Brinkman Polytron homogenizer. The micro- somes were isolated by differential centrifugation (16), resuspended in 1.15% KC1 to a protein concen- tration of 15 mg/ml, and stored in aliquots at -70C prior to use. Under these storage conditions, the activity of the microsomal lipid peroxidation system remained unchanged for at least 3 months. The incubation mixture for the assays of NADPH- dependent lipid peroxidation consisted of liver mi- crosomes (0.5 mg protein/ml) suspended in 0.1 M Tris-HCl buffer, pH `7.4, containing 12.5 FM ferric pyrophosphate. The reaction was initiated by the addition of 25 ~1 of 50 mM NADPH to the 2.5 ml of incubation mixture preincubated for 1 min at 3'7C. Aliquots (1.0 ml) were withdrawn at 0 and 20 min after the NADPH was added; the malondialdehyde content of these samples was measured by the thiobarbituric acid method (16, 21). The extent of lipid peroxidation was expressed as nanomoles malondialdehyde produced per 20 minutes per milligram microsomal protein. The measurements of cumene hydroperoxide-de- pendent lipid peroxidation were performed similarly except that the ferric pyrophosphate was omitted and the reactions were initiated by the addition of 25 ~1 of 30 nIM eumene hydroperoxide to the 2.5-ml incubation mixtures. Antioxidant activity measurements. Acetone stock solutions of each of the eight naphthoquinones were prepared and were diluted serially. A graded series of at least six quinone concentrations were generated by adding 25 ~1 of each dilution to a series of 2.5-ml lipid peroxidation incubation mixtures, pH `7.4, unless otherwise indicated. The reactions were initiated by NADPH additions and the malondialdehyde accu- mulating in 20 min was measured and plotted as a function of quinone concentration. These plots were interpolated to obtain the ICsO values; i.e., the con- centrations of quinone estimated to reduce the malondialdehyde accumulation by 50%. Menadione was also tested as an antioxidant in the cumene hydroperoxide-stimulated system. Two series of incubation mixtures were prepared; in one series NADPH (0.5 mM) was included and in the other series it was omitted. Graded concentrations of menadione (O-5 pM final concentration) were added to the flasks in each series and, after preincubating at 37C for 1 min, the flasks received cumene hydro- peroxide to initiate peroxidation. Assays of naphthoquinone-dependent NADPH ox- idation. Each of the eight naphthoquinones was tested as a substrate for the microsomal menadione 90 TALCOTT, SMITH, AND GIANNINI reductase by measuring the rate of NADPH oxidation occurring in the presence of microsomes and the quinone (22). In each case, graded concentrations were chosen so that the Michaelis constants could be estimated. A Kontron Model 800 uv-vis recording spectrophotometer equipped with an automatic sam- ple changer was used to obtain these measurements. The reference cuvette and the six sample cuvettes contained 2.5 ml of 0.1 M Tris-HCl buffer, pH 7.4 (unless otherwise indicated), 0.25 mM NADPH, and microsomes (0.5 mg protein/ml). With the mono- chromator set at 340 nm and the cuvettes warmed to 37C NADPH oxidation in the sample cuvettes was initiated by adding 25 pl of a series of naphtho- quinone stock solutions (in acetone) to the six sample cuvettes. The rates of NADPH oxidation were re- corded for 5 min and the recordings were used to calculate the initial velocities, in nanomoles NADPH oxidized per minute per milligram microsomal pro- tein. These data were used to estimate the apparent Michaelis constants (K,, V,,,,,) by the method of Wilkinson (23). Other methods. To estimate the amount of lipid hydroperoxide that had accumulated in the micro- somal membranes after 20 min of incubation with NADPH, ferric pyrophosphate, and a selected con- centration of menadione, 4 ml of the incubation mixture described above was extracted with an equal volume of chloroform:methanol, 2:l (24). The organic extract containing microsomal lipids was concentrated to dryness with a stream of Nz and the residue was redissolved in 1.0 ml of methanol. An aliquot (100 pl) of the methanolic solution was added to a sample cuvette containing 2.5 ml of freshly prepared 1% KI in glacial acetic acid. The reference cuvette contained only the KI-acetic acid reagent. The increase in absorbance at 360 nm due to the reaction of the hydroperoxide with I- (25) was recorded for 2 min; the hydroperoxide concentration TABLE I EFFECT OF MENADIONE CATALYZED MICROSOMAL ON NADPH + Fe*+- LIPID PEROXIDATION Menadione concentration mf) MDA accumulation (nmol 20 min-' mg protein-`) Hydroperoxide accumulation (nmol 20 mini' mg protein') 0 84.6 25 65.4 50 0.76 100 0.62 150 0.65 200 0.70 Note. Results of a typical experiment 15.0 16.5 1.87 1.00 1.50 0.50 are shown. TABLE II INHIBITION OF CUMENE HYDROPEROXIDE-DEPENDENT MICROSOMAL LIPID PEROXIDATION BY MENADIONE Menadione concentration (PM) Malondialdehyde accumulation (nmol 20 min-' mg protein-`) -NADPH +NADPH 0 16.2 0.2 15.8 0.5 15.4 1.0 15.7 2.0 15.0 5.0 15.6 Note. Results of a typical experiment 9.2 5.8 2.5 1.6 1.0 1.5 are shown. was calculated from a standard curve prepared with graded concentrations of cumene hydroperoxide. The apparent pK,`s of the hydroxylated menadione derivatives were estimated by recording the absor- bance of the ionized forms as a function of pH. The anions of 3-hydroxymenadione and 5-hydroxymena- dione were found to absorb maximally at 500 and 510 nm, respectively; the unionized forms showed negligible absorption at these wavelengths. RESULTS In a preliminary experiment, menadione was found to inhibit NADPH-dependent microsomal lipid peroxidation in a con- centration-dependent manner. The con- centration dependence was independent of incubation time for at least 20 min, after which no further malondialdehyde accumulation occurred. Table I shows that menadione, at concentrations of 50 nM or higher, effectively blocked NADPH-de- pendent microsomal lipid peroxidation, as measured either by the accumulation of malondialdehyde or of lipid hydroperox- ides. It was also of interest to assess the antioxidant activity of menadione in a system where lipid peroxidation does not require NADPH. For this experiment, cu- mene hydroperoxide was chosen as the initiator of lipid peroxidation. The results (Table II) indicate that menadione (0.2- 5.0 PM) is not an effective inhibitor of cumene hydroperoxide-dependent lipid per- oxidation unless NADPH is included in the incubation mixture. These results also ANTIOXIDANT ACTIVITY OF NAPHTHOQUINONES 91 STRUCTURE-ACTIVITY TABLE RELATIONSHIPS III AMONG 1,4-NAPHTHOQUINONES Activity Structure RI R3 Inhibition of lipid peroxidation R3 [%I (PM)] Stimulation K (PM) HH CH, H CHs CH:s CHz OH CHa OCH3 CHs G&l CH, H CH, H H H H H H H OH OCHB 27 0.037 0.130 180 0.51 800 0.25 0.47 27.9 + 6.0 34.1 + 5.6 20.5 rt 5.8 57.0 + 8.7 24.2 f 9.5 7.3 +- 2.8 10.4 + 0.3 11.8 + 1.0 Note. R1 is at the 2 position; R2 at the 3 position; and R3 at the 5 position. of NADPH oxidation V max (nmol min-' mg-`) 9.1 f 1.3 13.6 f 1.3 10.8 f 2.1 0.8 + 0.1 10.5 31 1.8 0.9 f 0.01 12.9 f 1.4 8.4 f 0.3 show that, in the absence of menadione, NADPH produces a partial inhibition of the cumene h,ydroperoxide-dependent lipid peroxidation, in agreement with the results of previous workers (26). The observations noted above (Table II) suggested that the antioxidant activity of a given naphthoquinone may bear a relation to its activity as a substrate for microsomal q.uinone reductase. This possibility was examined by comparing eight 1,4-naphthoquinones as antioxidants and as stimulators of microsomal NADPH oxidation. As antioxidants, the quinones varied over a. potency range of approximately 2 X 104. Although the two weakest antioxidants, 3-hydroxymenadione and 3pentadecylmenadione, were also the two weakest quintone reductase substrates, a close quantitative relationship between inhibition of lipid peroxidation and stimulation of NA.DPH oxidation was not apparent: 1,4-naphthoquinone, menadione, 3methylmenadione, 3-methoxymenadione, Shydroxymenadione, and 5-methoxymenadione were similar in activity as NADPH oxidase stimulators, but antioxidant activity among these compounds varied widely (Table III). Particularly striking was the difference in antioxidant potency between menadione (I& = 0.037 PM) and its desmethyl analog, 1,4-naphthoquinone (ICM = 27 PM), despite being approximately equivalent in their ability to stimulate NADPH oxidation. The structure-activity results revealed a striking difference in potency between the two hydroxylated compounds, 3-hydroxymenadione (I& = 180 PM) and 5hydroxymenadione (I&,, = 0.25 PM), which was not apparent between the two corresponding methoxy compounds (Table III). These observations suggested that the difference in potency between the hydroxylated compounds may be largely explained by differences in their acid-base properties. An experimental determination of the dissociation constants yielded pK, values of 9.1 and 5.1 for 5-hydroxy- and 3-hydroxymenadione, respectively, indicating that the latter compound is approximately 99.5% ionized at pH 7.4 (Fig. 1). This finding clearly suggested that the relative inactivity of 3-hydroxymenadione may have reflected its state of ionization at pH 7.4. The antioxidant activity of 3-hydroxymenadione measured at pH 7.0 and 6.0 increased in proportion to the percentage increase in the undissociated form (Table IV). Lowering the pH also improved the ability of this compound to stimulate NADPH oxidation. These results suggest 92 TALCOTT, SMITH, AND GIANNINI FIG. 1. pH titration (A) 5-hydroxymenadione; PH curves for the anionic forms: (B) 3-hydroxymenadione. that the active antioxidant form of 3hydroxymenadione is the reducible, undissociated form. At pH 7.4, 7.0, and 6.0 the I&, value for this form appears to be nearly constant and approximately 0.9 PM, i.e., I&,,, = (apparent I&) X (fraction undissociated species). Attempts to measure the antioxidant activity of 5-hydroxymenadione at pH's approaching its pK, failed because at these pH's the microsomal lipid peroxidation system was inactivated. DISCUSSION that would otherwise catalyze lipid peroxidation. On the other hand, Wills (27) noted that menadione inhibited lipid peroxidation even when excess NADPH was present, and suggested that an antioxidant effect independent of electron diversion may contribute importantly to the inhibitory effect. Several observations now indicate that electron diversion per se cannot account for the antioxidant activity of menadione. The menadione concentration required to inhibit microsomal NADPH-dependent lipid peroxidation is approximately three orders of magnitude below the concentration required to stimulate half-maximally the rate of NADPH oxidation. Moreover, 1,4-naphthoquinone and menadione are about equally effective in stimulating NADPH oxidation but menadione is about three orders of magnitude more potent as an antioxidant. The weakest antioxidant, 3-pentadecylmenadione, was ineffective at concentrations where a certain amount of stimulation of NADPH oxidation did occur. Finally, menadione, in the presence of NADPH, inhibited cumene hydroperoxide-dependent microsomal lipid peroxidation; in this model system NADPH inhibits rather than stimulates the reaction. Thus, it seems clear that the antioxidant action of menadione is unrelated to electron diversion, but is a result of a reaction between either menadione itself or one of its reduced forms with chemical species that are critically involved in the catalysis of microsomal lipid peroxidation. It seems likely that the active antioxidant form of menadione is either the The results presented in this paper pro- vide new information on the mechanism and structure-activity relationships un- derlying the inhibition of microsomal lipid peroxidation by 1,4-naphthoquinones. In an earlier paper, Wills (27) suggested that menadione and related naphthoquinones may inhibit microsomal lipid peroxidation by more than one mechanism. Since re- ducing equivalents are consumed by mi- crosomal quinone reduction and by micro- somal lipid peroxidation, Wills (27) sug- gested that menadione may divert electrons TABLE IV EFFECT OF pH ON 3-OH MENADIONE REDUCTION AND ON ITS ANTIOXIDANT ACTIVITY Percentage undissociated PH 3-OH-MD Apparent antioxidant potency [Icso (PM)] NADPH oxidase stimulation ( vmx~m 7.4 0.50 7.0 1.05 6.0 10.50 180 0.014 75 0.028 9 0.078 ANTIOXIDANT ACTIVITY OF NAPHTHOQUINONES 93 semiquinone or hydroquinone, because of the demonstrated requirement for reduc- ing equivalents (Table II). The antioxidant effect could therefore result from reactions between the hydroquinone and free radical species involved in catalyzing the propa- gation of lipid peroxidation. Studies with quinones in chemical systems (28) indicate that hydroperoxy radicals could be among those trapped. In the cumene hydroper- oxide stimulated reaction, there is evi- dence from studies with reconstituted systems (7, 11) and with microsomes (29) that ferric cytochrome P-450 reacts with cumene hydroperoxide to form cumoxy radicals. Once again the hydroquinone could trap these radicals thereby prevent- ing the initiation of lipid peroxidation. It is also possible that the semiquinone of menadione may contribute to the antiox- idant activity by pairing with hydroperoxy or alkoxy raldicals. The free electron in menadione semiquinone is delocalized so the semiquinone may be viewed as a com- posite of eighlt separate resonance forms. The structure-activity study presented here demonstrates the importance of a single methyl group on the quinone ring (cf. antioxidant activity of menadione vs that of 1,4-naphthoquinone), suggesting that the inductive properties of the methyl substituent may serve to stabilize forms that are important for optimal antioxidant activity (28). Also, the structure-activity study demonstrates the importance of an unhindered C-3 or C-5 site, suggesting that these sites may be importantly in- volved in the antioxidant reaction. Further studies with naphthoquinones that are substituted at other positions are needed to assess the importance of other sites on the naphthoquinone nucleus. An interesting topic for future research is the question of whether the pro-oxidant and antioxidiant properties of naphtho- quinones can be uncoupled. Central to this question would be a comparison of the structure-activity relationships governing superoxide generation with those govern- ing antioxidant activity. If different res- onance forms are more important to one pathway than to the other, different structure-activity relationships may be expected. If this is the case, it should be possible to maximize or minimize one activity (pro-oxidant or antioxidant) in relation to the other by varying the positions of substitution or the nature of the substituents on the naphthoquinone nucleus. This capability could be important in the design of new pharmacologically active naphthoquinones. ACKNOWLEDGMENTS This study was supported in part by PHS Grant CA13525 from the National Cancer Institute. M.T.S. thanks the University of California Cancer Research Coordinating Committee for support. The mass spec- tra were obtained by the Bio-organic, Biomedical Mass Spectrometry Resource (A. L. Burlingame, Director) supported by the NIH Division of Research Resources Grant RR01614. We thank Albert Ketter- man for expert technical assistance. REFERENCES 1. PLAA, G. L., AND WITSCHI, H. P. (1976) Annu. Rev. Pharwzacol. TwxicoL 16, 125-142. 2. HOCHSTEIN, P., AND ERNSTER, L. (1963) B&hem. Biophys. Res. Commun 12, 388394. 3. PEDERSON, T. C., AND AUST, S. D. (1972) B&hem. Biophys. Res. Commun 48, 789-794. 4. PEDERSON, T. C., BUEGE, J. A., AND AUST, S. D. (1973) J. BioL Chem. 248, 7134-7141. 5. SUGIOKA, K., AND NAKANO, M. (1976) B&him Biophys. Acta 423, 203-216. 6. SVINCEN, B. A., O'NEAL, F. O., AND AUST, S. D. (1978) Photochem. PhotobioL 28, 803-810. 7. SVINGEN, B. A., BUEGE, J. A., O'NEAL, F. O., AND AUST, S. D. (1979) .I BioL Chem. 254, 58925899. 8. FUNG, K.-L., MCCAY, P. B., POYER, J. L., KEELE, B. B., AND MISRA, H. (1973) J. BioL Chem. 248, 7792-7797. 9. KING, M. M., LAI, E. K., AND MCCAY, P. B. (1975) J. BioL Chem. 250, 6496-6502. 10. LAI, C. S., AND PIEITE, L. H. (1978) Adz. Biochem. Biophys. 190,27-38. 11. O'BRIEN, P. J., AND RAHIMTULA, A. (1975) J. Agtic. Food Chem 23, 154-158. 12. HRYCAY, E. G., AND O'BRIEN, P. J. (1971) Arch. Biochem. Biophys. 147,14-17. 13. WELTON, A. F., AND AUST, S. D. (1972) B&hem. Biophys. Res. Commun. 49, 661-666. 14. TALCOTT, R. E., DENK, H. ECKERSTORFER, R., AND SCHENKMAN, J. B. (1976) Chem BioL Interact. 12. 355-361. 15. KAPPUS, H., KIECZKA, H., SCHEULEN, M., AND 94 TALCOTT, SMITH, AND GIANNINI REMMER, H. (1977) Naunyn-Schmiedeberg's Arch. PhamnacoL 300, 179-187. 16. TALCOTT, R. E., SHU, H., AND WEI, E. T. (1979) Biochem. PhuwnacoL 28, 665-671. 17. LIN, A. J., PARDINI, R. S., COSBY, L. A., LILLIS, B. J., CHANSKY, C. W., AND SARTORELLI, A. C. (1973) J. Med. Chem. 16, 1268-1271. 18. FIESER, L. F. (1940) J. BioL Chem. 133, 391-396. 19. MACK, D. O., WOLFENSBERGER, M., GIRARDOT, J.-M., MILLER, J. A., AND JOHNSON, B. C. (1979) J. BioL Chem 254, 2656-2664. 20. LIN, A. J., PARDINI, R. S., LILLIS, B. J., AND SARTORELLI, A. C. (1974) J. Med. Chem. 17, 668-672. 21. OTTOLENGHI, A. (1959) Arch. B&hem. Biophys. 79, 355-363. 22. NISHIBAYASHI, H., OMURA, T., AND SATO, R. (1966) J B&hem. 60,172-183. 23. WILKINSON, G. N. (1961) B&hem. J. 80,324-332. 24. FOLCH, J., LEES, M., AND SLOANE-STANLEY, G. H. (1957) .I BioL Chem 226, 497-509. 25. BUEGE, J. A., AND AUST, S. D. (1977) in Methods in Enzymology (Fleischer, S., and Packer, L., eds.), Vol. 52, pp. 302-310, Academic Press, New York. 26. CAVALLINI, L., VALENTE, M., AND BINDOLI, A. (1983) Biochim. Biophys. Acta 752, 339-345. 27. WILLS, E. D. (1972) B&hem PharmacoL 21, 1879-1883. 28. SCOTT, G. (1965) Atmospheric Oxidation and Antioxidants, Elsevier, Amsterdam. 29. GRIFFIN, B. W. (1979) in Microsomes, Drug Oxi- dations, and Chemical Carcinogenesis (Coon, M. J., Conney, A. H., and Estabrook, R. W., eds.), 4th ed., pp. 319-322, Academic Press, New York.