Different behaviors of benznidazole as free radical generator with mammalian and Trypanosoma cruzi microsomal preparations

ARCHIVES

Vol.

OF BIOCHEMISTRY

218, No. 2, October

AND

BIOPHYSICS

15, pp. 585-591,

1982

Different Behaviors of Benznidazole as Free Radical Generator with Mammalian and Trypanosoma cruzi Microsomal Preparations’ SILVIA

N. J. MOREN0,2 W. LEON,? AND

R. DOCAMPO; R. P. MASON,* A. 0. M. STOPPANI3’4

Centro de Znvestigaciones Bioenerg&icas, Consejo National de Znvestigaciones Cientificas y T&nicas, Znstituto de Quimica Biol&ica, Faeultad de Medicine, Universidad de Buenos Aires, Paraguay 2155, 1121 Buenos Aires, Argentina; *National Znstitute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709; and tlnstituto de Microbiologia, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil Received

May

28, 1982

Benznidazole (a nitroimidazole derivative used for the treatment of Chagas’ disease) is reduced by rat liver microsomes to the nitro anion radical, as indicated by ESR spectroscopy. Addition of benznidazole to rat liver microsomes produced an increase of electron flow from NADPH to molecular oxygen, and generation of both superoxide anion and hydrogen peroxide. The benznidazole-stimulated 0, consumption and 0; formation was greatly inhibited by NADP+ and p-chloromercuribenzoate but not by SKF-525-A and metyrapone. The former inhibitions indicated the involvement of NADPH-cytochrome P-450 (c) reductase, while the lack of inhibition by SKF-525-A and metyrapone ruled out any major role for cytochrome P-450 in benznidazole reduction. In contrast to nifurtimox, a nitrofuran derivative (R. Docampo and A. 0. M. Stoppani, 1979, Arch. Biochem. Biophys. 197,317-321), benznidazole was not reduced to the nitro anion radical, nor did it stimulate oxygen consumption, 0; production, and H202 generation by Trypanosoma cruzi cells or microsomal fractions. A different mechanism of benznidazole toxicity in T. cruzi and the mammalian host is postulated.

Nifurtimox,5 a nitrofuran derivative used extensively in the treatment of Chagas’

disease, is reduced to the nitro anion radical by intact Trypanosoma crud cells and fractions obtained therefrom (1). The oxidation of the nitro anion radical induces the generation of 01, H202, and possibly OH., with the oxygen radical apparently being involved in the trypanocidal action of nifurtimox (2, 3). Similarly, nifurtimox is reduced to the nitro anion radical (4) and increases the production of 01, H202, and lipid peroxides by rat liver microsomes (5). These reactions may explain the toxic effect of nifurtimox on the mammalian host (5). Nitroimidazole drugs are used in the treatment of protozoa1 infections (6), and benznidazole, a nitroimidazole derivative, is effective in the treatment of Chagas’ disease (7). Taking into account observations made by Docampo et al. (l-5) with

‘This work received financial support from the UNDP/World Bank/World Health Organization Special Programme for Research and Training in Tropical Diseases, from the Secretaria de Estado de Ciencia y Tecnologia, Argentina, from UNDP/ UNESCO, from FINEP (Grant 527/CT), Brazil, and from the Scientific Office of the American States Organization. ’ Research Fellow, Consejo National de Investigaciones Cientificas y Tkcnicas. 3 Career Investigator, Consejo National de Investigaciones Cientificas y T&micas. ’ To whom all correspondence should be addressed. 5 Abbreviations used: benznidazole (Ro 7-1051), N-benzyl-2-nitro-1-imidazoleacetamide; nifurtimox, 3-methyl-4(5’-nitrofurfurylideneamino)tetrahydro4H-1,4-thiazone-l,l’-dioxide; SKF-525-A, @-dimethylaminoethyldiphenylpropylacetate; metyrapone, 2methyl-1,2-di-3-pyridyl-l-propanone. 585

0003-9861/82/120585-07$02.00/O Copyright All rights

8 1982 by Academic Press, Inc. of reproduction in any form resewed.

586

MORENO

nifurtimox, the reduction of the nitro group can be postulated as an essential step for both the trypanocidal and toxic properties of benznidazole. One-electron reduction results in the formation of the nitro anion radical, the reoxidation of which results in the formation of oxygen radicals (8-10). In addition, reduction of the nitro group may determine the formation of unstable nitroso and hydroxylamine derivatives which cause mutagenesis (ll), single-strand breaks in DNA (12), and alkylation of proteins (13). The present investigation was undertaken to establish the validity of the postulated mechanisms of benznidazole toxicity to T. cruzi and the mammalian host. MATERIALS

AND

METHODS

Treotmcnt of animals. Wistar male rats (150-200 g) or CD male rats (150-200 g, Charles River Inc.) were used in the experiments They were fed standard rat chow and water ad libitum and were not fasted prior to use. The animals were sacrificed by decapitation and their livers rapidly removed and processed. Culture. T. cruzi (Tulahuen strain) was grown at 28°C in the liquid medium described by Docampo et al. (3). Six days after inoculation, cells were collected by centrifugation and washed with 0.15 M NaCl. The final concentration of epimastigotes was estimated as described before (3). Microsomal preparations. Rat liver was suspended in 3 vol of ice-cold 150 mM KCl, 50 mM Tris-HCl buffer (pH 7.4). Homogenization was performed in a Potter tissue grinder with a Teflon pestle. The homogenate was centrifuged at 875Og for 15 min at 5°C. The supernatant was then centrifuged at 165,000g and 4°C for 38 min to separate the microsomal fraction. The pellet was washed twice with the KCl-Tris buffer solution in the centrifuge at 165,000g for 38 min. Cell fro&m&on. To the washed epimastigote pellet, glass powder (5 g/g cell wet wt) was added, and the pellet was ground in a mortar for 5 min at 4’C. This procedure resulted in complete breakage of the cells, as revealed by phase-contrast microscopy. The homogenate was suspended in 250 mM sucrose-5 mM KC1 (10 ml/g cell wet wt). Most of the glass powder was separated by decantation and the suspension was subjected to differential centrifugation at 4°C. The fractions obtained were: (a) the nuclear-flagellar fraction (sedimented at 68Og for 10 min); (b) the mitochondrial fraction (sedimented at 30,006g for 30 min); (c) the microsomal fraction (sedimented at 105,OOOg for 60 min); (d) the supematant.

ET AL. Reagents. These were obtained from the following sources: L-epinephrine, erythrocyte superoxide dismutase (Type I), horseradish peroxidase (Type VI), p-chloromercuribenzoic acid, glucose B-phosphate, glucose-6-phosphate dehydrogenase, NADH, NAD+, and NADPH from Sigma Chemical Company, St. Louis, Missouri, and benznidaxole from HoffmanLaRoche & Company, Basel, Switzerland. SKF-525A (from Smith, Klein & French Laboratories) and metyrapone (from Ciba Pharmaceutics Co.) were kindly supplied by Dr. J. A. Castro (Citefa, Argentina). Electron spin resonance spectroscopy. ESR observations were made at room temperature, 22’C, with a Varian E-104 spectrometer equipped with a TM,,, cavity or a Varian-E-9 spectrometer as previously described (14-16). The reaction mixture (3 ml, final volume) contained 150 mM KCl, 5 mM MgClx, 50 mM Tris-HCI (pH 7.4), and a NADPH-generating system made up of NADP+ (0.38 mM), glucose 6-phosphate (5.5 mM) and glucose-6-phosphate dehydrogenase (0.67 unit/ml). Benznidaxole was dissolved in ethylene glycol monomethyl ether (5%, w/v) and added as stated under Results; control samples received the corresponding volume of solvent. The reaction mixture was gassed with nitrogen for 5 min prior to initiating the reaction with NADP+. Protein concentration was determined as previously described (15). Fresh microsomes or T. cruzi subcellular fractions less than 9 h old stored on ice were used throughout. Determination of superoxide anion and hydrogen peroxide. Production of 0; was determined by the adrenochrome assay (17) and also by the superoxide dismutase-sensitive reduction of acetylated cytochrome c (18). The formation of adrenochrome was measured at 480-575 nm, using an absorption coefficient (c) of 2.96 mM-’ cm-‘. The reaction mixture contained 150 mM KCl, 50 mM Tris buffer (pH 7.4), 1 mM epinephrine, and NADPH as indicated under Results. The reaction exhibits sigmoidal kinetics, and under these conditions it is apparently autocatalytic (19), the maximal rate occurring after a lag of as long as 5 min. The initial rates reported here were determined within 30 s of the reaction and the corresponding values were not different from zero-time extrapolation of the plots of adenochrome formation versus time (20). Reduction of acetylated cytochrome c was measured at 550-540 nm using an absorption coafficient (t) of 19 rnM-i cm-’ (18). The reaction mixture contained 150 mM KCl, 50 mM Tris buffer (pH 7.4), 10 PM acetylated cytochrome c, and NADPH as indicated under Results Hydrogen peroxide generation was measured by the horseradish peroxidase assay (16) at 417-402 nm using an absorption coefficient (c) of 50 mM-’ cm-‘. The reaction mixture contained 150 mM KCl, 50 mM Tris buffer (pH 7.4), horseradish peroxidase, and NADPH as indicated under Results

BENZNIDAZOLE

‘; 02=0.220mM-

4

INDUCTION

OF FREE

NADPH 1 5JS I

II-Y

lmin -

02:OmM”

FIG. 1. Effect of bensnidazole on oxygen consumption by rat liver microsomes (M) and T. cruzi epimastigotes (E). The reaction mixtures contained microsomes (4.0 mg protein/ml) or epimastigotes (1.2 mg protein/ml), 150 mM KCl, and 50 mM Tris-HCl buffer (pH 7.4). NADPH, 0.5 mM, and 1.0 mM benznidaxole (B) were added where indicated by the arrows. Other experimental conditions were as described under Materials and Methods. The values near the tracings indicate nmol O,/min/mg of protein. An Aminco-Chance spectrophotometer (American Instrument Company, Silver Spring, Md.) was utilized. Unless stated otherwise, all measurements were made at 30°C. Oxygen uptake. Oxygen uptake was measured in the Gilson polarograph using a Clark electrode. Assays of oxygen consumption were made at 30°C in a medium containing 150 mM KCl, 50 mM Tris buffer (pH 7.4), and additions as stated under Results.

RADICAL

587

GENERATION

Enzymatic reduction of 5-nitroimidazoles produces an anion radical capable of reacting with molecular oxygen to regenerate the parent nitroimidazole (15). Figure 2A shows that the incubation of benznidazole with rat liver microsomes, in the presence of a NADPH-generating system, produced a characteristic multiline spectrum that could be identified as the nitro anion radical. The value of aEo2 was 13.75 G and compared favorably with the value of 14.10 G for misonidazole (21), another 2-nitroimidazole. The radical formation depended on all three components of the system, namely, microsomes, NADPH, and nitroimidazole. Omission of NADPH or heating of microsomes in a steam bath for 10 min led to total loss of reducing activity, while the NADPH-generating system alone did not produce any characteristic signal. The spectrum of the anion radical appeared after a lag when the oxygen dissolved in the incubation medium had been exhausted. By increasing the microsomal protein concentration in the reaction mixture, the rate of oxygen consumption increased as well and, as a consequence of this, the lag was shorter (data omitted). Attempts to demonstrate benznidazole

RESULTS

Figure 1, M, shows that oxygen consumption was increased severalfold when benznidazole was introduced into a reaction mixture containing rat liver microsomes and NADPH. The benznidazole-induced oxygen uptake depended on all the components of the system, the reaction rate being negligible in the absence of either NADPH or microsomes, or after the thermal denaturation of the microsomes (experimental data omitted). In contrast to these results benznidazole did not stimulate oxygen consumption in T. cruzi intact cells (Fig. 1, E), even at a concentration as high as 3 mM (not shown).

A

10 Gauss

FIG. 2. The ESR spectrum of benznidazole anion radical after anaerobic incubation of 4 mM benznidazole with an NADPH-generating system and 4 mg/ ml of rat liver microsomal protein (A) or 5 mg/ml of T. crud epimastigotes (B). The spectra were obtained with a nominal microwave power of 5 mW and a modulation amplitude of 0.8 G. Other experimental conditions were as indicated under Materials and Methods.

588

MORENO 480-575nm SOD

SOD

nA=O.OOl

550 - 540 “In SOD I

SOD

FIG. 3. Effect of benznidazole on supsroxide-dependent adrenochrome production (upper tracings) and acetylated cytochrome c reduction (lower tracings) by rat liver microsomes. The reaction mixture contained microsomes (0.3 mg protein/ml), 150 mM KCl, 50 mM Tris buffer (pH 7.4), and 1 mM epinephrine or 10 FM acetylated cytochrome c. NADPH, 0.16 mM, 1 mM benznidazole (B), and 5 ag/ml superoxide dismutase (SOD) were added where indicated by the arrows. Other experimental conditions were as indicated under Materials and Methods and Table I. The values near the tracings indicate nmol Oi/min/mg of protein.

reduction to a nitro anion radical in the presence of T. cruzi intact cells (Fig. 2B), mitochondrial, or microsomal fractions under conditions similar to those used with rat liver microsomes were unsuccessful.

ET AL.

Figure 3, B, shows that benznidazole significantly increased the rate of adrenochrome formation by NADPH-supplemented liver microsomes and decreased the induction period of epinephrine oxidation. Omission of microsomes or NADPH from, or addition of superoxide dismutase to, the reaction mixture prevented adrenochrome formation (Fig. 3, upper tracings), thus confirming the formation of 0; via electron transfer from the nitroaromatic anion radical to oxygen. Similar results were obtained by measuring 0; generation by the rate of the superoxide dismutase-sensitive reduction of acetylated cytochrome c (Fig. 3, lower tracings). Benznidazole stimulated 0; production linearly up to 2 mM, and at this concentration the basal rate was increased about sixfold (Fig. 4). Under identical conditions, incubations in the presence of T. cruzi microsomes did not show any stimulation of 0; generation (Fig. 4). Production of H202 by liver microsomes was also stimulated by benznidazole (Fig. 5), but not by T. cruzi microsomes (Fig. 6). With the liver microsomes, benznidazole stimulated H202 production linearly up to 2 mM, and at this concentration the basal rate was increased ninefold (Fig. 6). The lack of 2:l stoichiometry between 0; generation (Fig. 3) and HzOz generation (Fig. 5) may be explained by the different con417 - A02 “In

BENZNIDAZOLE

(mM)

FIG. 4. Effect of benznidazole concentration on superoxide-dependent acetylated cytochrome c reduction by liver microsomes (LM) or T. cruzi microsomes (0.4 m g protein/ml). Experimental conditions were as described in Fig. 3, except for benznidazole concentration, which is indicated on the abscissa.

-

FIG. 5. Effect of benznidazole on hydrogen peroxide production by rat liver microsomes. The reaction mixture contained microsomes (0.3 mg protein/ml), 150 mM KCl, 50 mM Tris buffer (pH 7.4), and 0.6 pM horseradish peroxidase. NADPH, 0.16 mM, and 3 mM banznidazole (B) were added where indicated by the arrows. Other experimental conditions were as indicated under Materials and Methods. The values near the tracings indicate nmol HzOz/min/mg of protein.

BENZNIDAZOLE

0

1 BENZNIDAZOLE

INDUCTION

OF FREE

2 (mM1

FIG. 6. Effect of benznidazole concentration on hydrogen peroxide production by liver microsomes (LM) or 7’. cruzi microsomes. Experimental conditions were as described in Fig. 5 except for benznidazole concentration, which is indicated on the abscissa. T. cruzi microsomes, 0.4 mg protein/ml.

tent of superoxide dismutase (from the cytosol) in the microsomal preparations used in each experiment, which even after washing the microsomes twice, as done in the present study, may give day-to-day variation in the actual numbers. Table I shows the effect of specific inhibitors of the microsomal electron transport system on O2 consumption and 0; formation by liver microsomes in the presTABLE EFFECT

OF INHIBITORY ADRENOCHROME

RADICAL

ence of benznidazole. These results deserve the following comments: (a) the inhibition of activity by NADP+ indicates the participation of NADPH-cytochrome P-450 (c) reductase, since NADP+ is a competitive inhibitor of this enzyme; (b) the same conclusion is supported by the effect of p-chloromercuribenzoate, an inhibitor of electron transport in liver microsomes during NADPH oxidation (2224); (c) the lack of effect of SKF-525-A and metyrapone, two specific inhibitors of the microsomal electron transport system at the level of cytochrome P-450 (25) rules out any major role for this cytochrome in benznidazole reduction. Benznidazole concentrations lower than those increasing 0; and H202 generation by liver microsomes, but ineffective for the same purpose in T. cruzi, were effective in inhibiting the parasite growth (Fig. 7). DISCUSSION

The observations here described indicate that benznidazole is reduced by rat liver microsomes to a nitro anion radical, the reoxidation of which is accompanied by the formation of active oxygen radicals. I

OF NADPH-SUPPORTED OXYGEN CONSUMPTION, SUPEROXIDE-DEPENDENT FORMATION, AND ACETYLATED CYT~CHROME c REDUCTION BY LIVER MICROSOMES IN THE PRESENCE OF BENZNIDAZOLE’ nmol/min/mg

Inhibitor None NADP+ (1.0) p-Chloromercuribenzoate Metyrapone (0.1) SKF-525-A (0.1)

589

GENERATION

02 consumption

(mM)

(0.1)

4.4 + 0.2’ 2.9 * 0.2 (66)d 0 (100) 4.4 f 0.3 5.0 -t 0.3

Adrenochrome formation 6.0 f 0.5 1.5 * 0.2 (75) 0 (100) 5.2 + 0.4 5.1 + 0.3

of proteinb Acetylated cytochrome c reduction 5.7 f 0.7 1.4 k 0.3 (72) 0 (100) 5.8 + 0.8 5.3 k 0.7

a The incubation mixture contained microsomal protein (4 mg protein/ml), 1 mM benznidazole, and inhibitors as stated above. Other experimental conditions were as described under Materials and Methods. b Rates of Ox consumption, adrenochrome formation, and acetylated cytochrome c reduction in the absence of benznidazole were 1.5 + 0.1, 1.4 + 0.2, and 1.9 k 0.3 nmol/min/mg protein, respectively. Adrenochrome formation was completely inhibited by 5 pg/ml of superoxide dismutase. Values of acetylated cytochrome c reduction are expressed as the difference between cytochrome c reduction in the absence and the presence of 5 pg/ml of superoxide dismutase (superoxide dismutase-dependent cytochrome c reduction). ’ Mean f SD (three independent determinations). d Percentage inhibition shown in parentheses.

590

MORENO

TIME

(days)

FIG. 7. Effect of benznidazole on growth of T. cruzi. Time of culture was as indicated on the abscissa. The values near the lines indicate benznidazole concentration (pM). Other conditions were as described under Materials and Methods.

In fact, benznidazole (a) induced additional microsomal oxygen utilization (Fig. 1 and Table I); (b) increased the rate of 0; formation (Figs. 3 and 4, Table I); and (c) increased the rate of Hz02 production by microsomes (Figs. 5 and 6). In addition to these effects, evidence for the nitro radical formation was obtained by electron spin resonance spectroscopy (Fig. 2). The effects of inhibitors in Table I indicate that the initial reduction of the drug was initiated by NADPH-cytochrome P-450 (c) reductase, 2 ArNOz

+ NADPH 2 ArNO;

+ H+ + NADP+

+ 2 H’

[l]

The anion radical of the drug reacts with molecular oxygen to generate the superoxide anion, whose dismutation yields HzOz, ArNO;

+ O2 -

2 0; + 2 H+ -

ArN02

+ 0; ,

HzOz + Oz .

PI [31

One might expect the rate limiting reaction to be the reduction by NADPH of the heterocyclic compound. The lack of effect of specific inhibitors of cytochrome P-450 (metyrapone and SKF-525) on Oz consumption and 0~ formation (Table I) indicates that cytochrome P-450 is not involved in the reduction of benznidazole. In spite of the many similarities in the mode of action of nitrofurans and nitroimidazoles in several biological systems (26),

ET AL.

benznidazole increased 0; and HzOz production by rat liver microsomes at a concentration one order of magnitude higher than that of nifurtimox (5). This difference fits in well with the lower electronegativity of the 2-nitroimidazoles, as compared with nitrofurans, which should decrease the rate of enzymatic radical formation (27). Furthermore, the results in Figs. 3-6 allow one to assume that the effect of benznidazole on superoxide anion and hydroxide peroxide formation in mammalian tissues must necessarily be small, hardly exceeding the basal levels. The difference noted between nifurtimox’s and benznidazole’s capability to generate oxygen free radicals in mammalian cells may be significant in the treatment of Chagas’ disease when these drugs are compared for their toxicity to the host. The oxidation of benznidazole nitro anion radical by oxygen blocks any further reduction of this group, thus preventing the formation of nitroso and hydroxylamine derivatives, which are presumably responsible for mutagenesis (1 l), singlestrand breaks in DNA (12), and alkylation of proteins (13). Worth noting in this connection is the absence of carcinogenic activity in benznidazole as compared with nitro derivatives, such as 4-nitroquinoline N-oxide, whose reduction by biological electron donors is in part oxygen insensitive (28). Since no nitro anion radical could be demonstrated after incubation of benznidazole with NADPH and T. cruzi microsomal preparations, and 0; and HzOz generation was not stimulated by benznidazole under the same experimental conditions, the possibility that, in T. cruzi, benznidazole might be involved in aerobic redox cycling leading to one electron reduction of oxygen is not supported by the experimental evidence. Moreover, the results in Fig. 7 indicate that benznidazole inhibits growth of T. cruzi at concentrations which do not stimulate superoxide anion and hydrogen peroxide generation, thus indicating that the trypanocidal effect does not depend on the effect of oxygen radicals. The difference between the mechanisms of benznidazole and nifurtimox (2, 3) toxicity on T. cruzi is, therefore, signif-

BENZNIDAZOLE

INDUCTION

OF FREE

icant. If reduction of the nitro group were required for the lethal action of benznidazole, the reaction should depend on a nitroreductase different from the mammalian microsomal enzyme. In this connection it should be recalled that Peterson et al. (1979) (14) described an oxygen-insensitive nitroreductase in Escherichiu coli which does not appear to transfer one electron to nitrocompounds. Moreover, Josephy et al. (29) have shown the reduction of misonidazole [1-(2-nitro-1-imidazolyl) 3-methoxy-2-propanol] and its azo and azoxy derivatives by xanthine oxidase, under hypoxic conditions, the metabolic reduction products being responsible for the cytotoxicity of the drug. The information available on T. cruzi microsomal cytochromes is limited (30,31) and perhaps an oxygen-insensitive, non-free radical nitroreductase exists in T. cruzi as it does in E.

11. 12. 13. 14.

15.

16.

17. 18.

19.

COli. REFERENCES

20. 21.

1. DOCAMPO, R. (1980) in The Host-Invader Interplay (Van Den Bossche, H., ed.), pp. 677-681, Elsevier/North-Holland, Amsterdam. 2. DOCAMPO, R., AND STOPPANI, A. 0. M. (1979) Arch. Biochem. Biaphys. 197, 317-321. 3. DOCAMPO, R., MORENO, S. N. J., STOPPANI, A. 0. M., LEON, W., CRUZ, F. S., VILLALTA, F., AND MUNIZ, R. P. A. (1981) Biachem. Pharnlncol. 30, 1947-1951. 4. DOCAMPO, R., MASON, R. P., MORLEY, C., AND MUNIZ, R. P. A. (1981) J. Bial. Chem. 266, 10,930-10,933. 5. DOCAMPO, R., MORENO, S. N. J., AND STOPPANI, A. 0. M. (1981) Arch. Biochem. Biophys. 207, 316-324. 6. MULLER, M. (1980) in The Eukaryotic Microbial Cell (Goodey, G. W., Lloyd, D., and Trinci, A. P. J., ede.), pp. 127-142, Cambridge Univ. Press, Cambridge. 7. POLAK, A., AND RICHLE, R. (1978) Ann. Trap. Med. Parasitol. 72, 45-54. 8. MASON, R. P., AND HOLTZMAN, J. L. (1975) Bio&em. Biophys. Res. Commun. 67, 1267-1274. 9. MASON, R. P., AND HOLTZMAN, J. L. (1975) Biochemistry 14, 1626-1632. 10. ADAMS, G. E., CLARKE, E. D., JACOBS, R. S.,

22. 23.

24. 25.

26.

27.

28. 29. 30. 31.

RADICAL

GENERATION

591

STRAFORD, I. J., WALLACE, R. G., WARDMAN, P., AND WATTS, M. E. (1976) Biachem. Biaphys. Res. Commun. 72.824-829. HECHT, S. S., EL-BAYOUMY, K., TULLEY, L., AND LAVOIE, E. (1979) J. Med. Chem. 22,981-987. OLIVE, P. L., AND MCCALLA, D. R. (1977) Chem. Biol. Interact. 16, 223-233. MCCALLA, D. R., REUVERS, A., AND KAISER, C. (1971) Biachem. Pharmacol. 20, 3532-3538. PETERSON, F. J., MASON, R. P., HOVSEPIAN, J., AND HOLTZMAN, J. L. (1979) J. Biol. Chem. 254,4009-4014. PEREZ-REYES, E., KALYANARAMAN, B., AND MASON, R. P. (1980) Mol. Pharmacol. 17, 239244. DOCAMPO, R., CRUZ, F. S., BOVERIS, A., MUNIZ, R. P. A., AND ESQUIVEL, D. M. S. (1978) Arch. Biachem. Biaphys. 186, 292-297. MISRA, H. P., AND FRIDOVICH, I. (1972) J. Biol. Chem. 247, 188-192. Azz~, A., MONTECUCCO, C., AND RICHTER, C. (1975) B&hem. Biaphys. Res. Commun. 66, 597-603. MASON, R. P., PETERSON, F. J., AND HOLTZMAN, J. L. (1978) Mol. Pharmacol. 14, 665-671. TURRENS, J. F., AND BOVERIS, A. (1980) Biachem. J. 191, 421-427. AYSCOUGH, P. B., ELLIOT, A. J., AND SALMON, G. A. (1978) J. Chem. Sot. Faraday Trans. 1, 74, 511-518. ERNSTER, L., AND ORRENIUS, S. (1965) Fed. Proc. 24, 1190-1199. MASTERS, B. B. S., KAMIN, H., GIBSON, Q. H., AND WILLIAMS, C. H., JR. (1965) J. Bial. Chem. 240, 921-931. ORRENIUS, S. (1965) J. Cell. Bial. 26, 713-723. ERNSTER, L., AND NORDENBRAND, K. (1967) in Methods in Enzymology (Estabrook, R. W., and Pullman, M. E., ede.), Vol. 10, pp. 574-580, Academic Press, New York. MASON, R. P. (1982) in Free Radicals in Biology (Pryor, W. A., ed.), pp. 196-212, Academic Press, New York. CLARKE, E. D., WARDMAN, P., AND GOULDING, K. H. (1980) B&hem. Pharmacol. 29, 26842687. KATO, R., TAKAHASHI, A., AND OSHIMA, T. (1970) B&hem. Pharmacol. 19, 45-55. JOSEPHY, P. D., PALCIC, B., AND SKARSGAFZD, L. D. (1981) Biachem. Pharmacol. 30.849-853. DE BOISO, J. F., AND STOPPANI, A. 0. M. (1971) Proc. Sot. Exp. Biol. Med. 136, 215-221. AGOSIN, M., NAQUIRA, C., CAPDEVILA, J., AND PAULIN, J. (1976) Znt. J. Biochem. 7.585-593.