An electron paramagnetic resonance study of the interactions between the adriamycin semiquinone, hydrogen peroxide, iron-chelators, and radical scavengers

ARCHIVES

OF BIOCHEMISTRY

AND BIOPHYSICS

Vol. 286, No. 1, April, pp. 164-170, 1991

An Electron Paramagnetic Resonance Study of the Interactions between the Adriamycin Semiquinone, Hydrogen Peroxide, Iron-Chelators, and Radical Scavengers B. Kalyanaraman,

*J K. M. Morehouse,ty’

and Ronald P. Mason?

*Biophysics Section, Department of Radiology, Medical College of Wisconsin, Milwaukee, Wisconsin 53226; and TLaboratory of Molecular Biophysics, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709

Received July 25, 1990, and in revised form November

20, 1990

Anaerobic reduction of hydrogen peroxide in a xanthinelxanthine oxidase system by adriamycin semiquinone in the presence of chelators and radical scavengers was investigated by direct electron paramagnetic resonance and spin trapping techniques. Under these conditions, adriamycin semiquinone appears to react with hydrogen peroxide forming the hydroxyl radical in the presence of chelators such as ethylenediaminetetraacetic acid and diethylenetriaminepentaacetic acid. In the absence of chelators, a related, but unknown oxidant is formed. In the presence of desferrioxamine, adriamycin semiquinone does not disappear in the presence of hydrogen peroxide at a detectable rate. The presence of adventitious iron is therefore implicated during adriamycin semiquinone-catalyzed reduction of hydrogen peroxide. Formation of cY-hydroxyethyl radical and carbon dioxide radical anion from ethanol and formate, respectively, was detected by spin trapping. Both the hydroxyl radical and the related oxidant react with these scavengers, forming the corresponding radical. In the presence of scavengers from which reducing radicals are formed, the rate of consumption of hydrogen peroxide in this system is increased. This result can be explained by a radicaldriven Fenton reaction. o 1991 Academic PWSS, IN.

The anthracycline quinone antibiotics adriamycin and daunorubicin are widely used in the chemotherapeutic treatment of a variety of cancers (l-3). Unfortunately, 1 To whom correspondence should be addressed at Biophysics Section, Medical College of Wisconsin, 8701 Watertown Plank Rd., Milwaukee, WI 53226. FAX: (414) 266-4007. ’ Present address: Division of Food Chemistry and Technology, Food and Drug Administration, 200 “C” Street, S. W., Washington, DC 20204. 164

the clinical use of these very effective antitumor agents is severely limited by a dose-dependent cardiomyopathy leading to congestive heart failure (1). The cardiotoxicity of anthracyclines has been linked to their ability to undergo reduction to the semiquinone, followed by redox cycling in the presence of 02, forming 0; and HzOz (4-14). The adriamycin semiquinone, under limiting oxygen concentration, has also been shown to reduce ferric iron complexes (15) and to mobilize iron from ferritin in vitro (16). The net result is a radicaldriven Fenton reaction, which produces reactive oxidants such as the hydroxyl radical and/or ferry1 iron (Scheme 1). Two important pieces of evidence support the free radical hypothesis of cardiotoxicity (17-19). First, the in uiuo cardiotoxicity of adriamycin is partially decreased by free radical scavengers and iron chelators, and second, anthracycline analogs such as 5-iminodaunorubicin, which shows little or no cardiotoxicity, do not undergo redox cycling to any significant extent. Cardiac injury induced by adriamycin has been attributed to increased peroxidation of mitochondrial lipids (20). Recently, it was shown that adriamycin-dependent lipid peroxidation is greatly enhanced at low oxygen concentration in the presence of ferritin (21). Despite the fact that the adriamycin semiquinone reacts very rapidly with molecular oxygen (K = 10’ M-l s-l), its ability to initiate a radical-driven Fenton reaction in biological systems has also been demonstrated (22, 23). Bates and Winterbourn have shown earlier that the adriamycin semiquinone is also capable of reducing hydrogen peroxide, under very low oxygen tension, forming hydroxyl radicals (24). This reaction was presumed to occur without the participation of iron (24). Using electron paramagnetic resonance spectroscopy, Kalyanaraman et 0003~9861/91 $3.00 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

REACTIONS

OF ADRIAMYCIN

rdriemycin

165

SEMIQUINONE

scavenger-derived radicals modify the overall reaction of H202 in the Fenton reaction.

+a1

MATERIALS

,ssmiguinone territin/ iron complexes

4

/ 0;

\

gFss+ inhibitor redfcsl

SCHEME

<

inhibitor

1. Adriamycin

* OH higher

semiquinone-driven

oxidant

radical reactions.

al. also concluded that adriamycin semiquinone reacts with hydrogen peroxide and organic peroxides in the presence and absence of chelators (25). More recently, Winterbourn et al. (26, 27) have modified their original conclusions with regard to the role of iron and chelators in the reaction between adriamycin semiquinone and H,Oz as follows: (i) the presence of a catalytic amount of iron is necessary for the reduction of Hz02 by the adriamycin semiquinone; (ii) the oxidant formed in this reaction is dependent on the type of chelators present; for example, hydroxyl radicals are formed in the presence of chelators such as EDTA and DETAPAC,3 whereas a related oxidant, possibly a perferryl species (crypto-hydroxyl radical), is formed in the presence of ADP and in the absence of added chelators; and (iii) the reactivity of the related oxidant toward radical scavengers and inhibitors is very much different from that of *OH (28, 29). Other investigators have also proposed similar mechanisms to account for formation of hydroxyl radicals during reductive metabolism of adriamycin (30, 31). We have reinvestigated the anaerobic reduction of H,O, by adriamycin semiquinone in the presence of several chelators and radical scavengers using EPR and the spin trapping technique. In this paper we report that the adriamycin semiquinone reacts with Hz02 in the presence of nonchelated adventitious iron and EDTA-, DETAPACchelated adventitious iron but fails to react with Hz02 in the presence of the desferrioxamine-chelate of adventitious iron, and that hydroxyl radicals are detected by spin trapping only when EDTA and DETAPAC are included in the reaction mixture. However, in the presence of commonly used hydroxyl radical scavengers, spin trapping data show evidence for the production of scavenger-derived radicals with the generation of both . OH and the related oxidant. The EPR data also demonstrate that 3 Abbreviations used: DETAPAC, diethylenetriaminepentaacetic DMPO, 5,5-dimethy-1-pyrroline-N-oxide.

acid,

AND

METHODS

Adriamycin HCI (NSC123-127) was obtained as a gift from Dr. Birandra K. Sinha of the National Cancer Institute. Hydrogen peroxide (30%) was obtained from Fisher Scientific Co. Xanthine (Merck Chemical Co.), mannitol (Sigma Chemical Co.), P-deoxyribose (Sigma Chemical Co.), EDTA (Aldrich Chemical Co.), and DETAPAC (Aldrich Chemical Co.) were used as received. Xanthine oxidase (grade III and grade IV) was obtained from Sigma Chemical Co. 5,5-Dimethyl-l-pyrreline-N-oxide (Aldrich Chemical Co.) was purified by vacuum distillation before use. Desferrioxamine was a gift from CIBA Geigy (Suffern, NY). A Varian E-109 spectrometer operating at X-band (9.5 GHz) and employing 100 kHz modulation was used for direct ESR measurements. A typical incubation mixture for ESR consisted of adriamycin (100 FM), xanthine (400 fiM), and xanthine oxidase (0.2 units) in a 2-ml phosphate buffer (100 mM, pH 7.5). Where necessary, radical scavengers and spin trap were added to the incubation mixtures. Incubations were bubbled with nitrogen or argon gas for 15 min before addition of enzyme. Incubations were also made using Chelex-treated phosphate buffer, prepared according to procedures described by Buettner (32). The reaction mixture was then rapidly introduced into an aqueous flat cell by aspiration (33), and ESR observations were begun.

RESULTS

Generation of adriamycin semiquinone radical. Adriamycin semiquinone was generated by adding xanthine oxidase to a nitrogen-purged solution containing xanthine and adriamycin. After a short time lag (corresponding to the removal of residual oxygen), a one-line ESR spectrum (AHpp = 3.25 G, g = 2.0035) of the adriamycin semiquinone was observed (Fig. 1). Addition of

t

4G

’

FIG. 1. ESR spectrum of the adriamycin semiquinone. The incubation conditions are described under Materials and Methods. Spectrometer conditions: modulation amplitude, 1.0 G; microwave power, 1 mW; scan range, 40 G.

166

KALYANARAMAN,

MOREHOUSE,

FIG. 2. Time-course ESR spectra of semiquinone formed during anaerobic reduction of adriamycin by xanthine/xanthine oxidase in phosphate buffer, pH 7.4 (A) Control (no H202 addition); (B) containing 50 pM H,Os and 100 pM EDTA; (C) containing 50 pM H202 and 100 pM DETAPAC; (D) containing 50 FM HzOz and no added chelator; and (E) The short time lag containing 50 pM HzOz and 100 pM desferrioxamine. observed in (A) and (E) is attributed to the presence of residual oxygen (9,31). The enzyme was added at the point marked t. The spectra were obtained by scanning as a function of time while sitting on top of the maximum absorption of adriamycin semiquinone signal.

EDTA (100 PM) or DETAPAC (100 PM) to the incubation mixture had no effect on the formation of the semiquinone. All of the components of the incubation mixture (cf. Fig. 1) were necessary for the generation of the ESR signal. Time course studies showed a short time lag corresponding to the removal of residual oxygen (Fig. 2). Time lag in semiquinone formation due to reduction of HzOz. In the presence of added HzOz, anaerobic reduction of adriamycin in the xanthine/xanthine oxidase system exhibited considerable time lag in the formation of adriamycin semiquinone with and without the chelators EDTA and DETAPAC (Fig. 2) and o-phenanthroline (data not shown). After HzOz was depleted, the semiquinone was detected (Fig. 2). The time lag was directly proportional to [H,O,], and inversely proportional to [xanthine oxidase] under otherwise identical conditions (cf. Fig. 2) (data not shown) (25). Experiments using Chelextreated phosphate buffer also showed the same time lag under incubation conditions shown in Figs. 2B-D. However, the anaerobic reduction of adriamycin by xanthine/ xanthine oxidase in the presence of H202 and desferrioxamine did not show an extended time lag for the semiquinone formation (Fig. 2). Increasing the [H,O,] in Fig. 2E also did not affect the time lag (data not shown). These results indicate clearly that adriamycin semiquinone (i) reacts with HzOz in a reaction involving EDTA- and DE-

AND

MASON

TAPAC-chelates of adventitious transition metal(s), presumably iron and/or copper, (ii) reacts with H202 in the presence of nonchelated adventitious transition metal(s), and (iii) fails to react with HzOz in the presence of desferrioxamine chelate of the adventitious transition metal(s). Characterization of the identity of the oxidant formed during the reduction of H202 In order to identify the radical intermediates formed under these conditions (cf. Figs. 2B-E), the spin trap DMPO was included in the incubation mixtures. During the time lag, intense ESR spectra characteristic of DMPO/ . OH were observed only in the presence of EDTA (Fig. 3, left lane) and DETAPAC (Fig. 3, middle lane). However, in the absence of added chelators the ESR spectrum due to DMPO/ . OH was only barely detectable (Fig. 3, right lane) under otherwise identical experimental conditions. Note the absence of the ESR signal due to the adriamycin semiquinone (Fig. 3). On the basis of experiments using radical scavengers (ethanol, formate, etc.), we verified that the DMPO/ . OH adduct in Fig. 3 was produced from the trapping of hydroxyl radical by detecting intense spectra of DMPO/CH,‘CHOH and DMPO/ * CO; (Figs. 3B and C, left and middle lanes). In the absence of chelators, the signal intensities due to DMPO/CH,‘CHOH and DMPO/ * CO; were considerably more intense than that of DMPO/ . OH (Fig. 3, right lane), clearly implicating formation of an oxidizing intermediate other than the hydroxyl radical.

EDTA

DETAPAC 0

NO CHELATOR

0

FIG. 3. ESR spectra of DMPO radical adducts observed during the lag period in Figs. 2B-2D. The incubations corresponding to left, middle, and right lanes are identical to those shown in Figs. 2B-2D, except that they contained EDTA (100 PM), DETAPAC (100 pM), and no added chelator. Conditions: A, DMPO (100 mM); B, DMPO and ethanol (100 mM); and C, DMPO and formate (100 mM). (0) DMPO/*OH; (A) DMPO/CHsCHOH; (0) DMPO/. CO; adducts.

REACTIONS

OF ADRIAMYCIN

l B

0

on the ESR spectra of adriaFIG. 4. The effect of desferrioxamine mycin semiquinone and DMPO radical adducts during the reduction of adriamycin by xanthine, xanthine oxidase in phosphate buffer under nitrogen (pH 7.4). (A) Containing 50 pM HxOx, DMPO (100 mM) and desferrioxamine (100 PM); (B) same as (A) but containing ethanol (100 mM) and desferrioxamine (100 PM); (C) same as (A) but containing formate (100 mM). (0) Adriamycin semiquinone; (0) DMPO/ * CO, adduct. Note that there is no extended time lag in the formation of adriamycin semiquinone in the presence of desferrioxamine as in Fig. 3.

In the presence of desferrioxamine, the ESR spectrum of DMPO/ * OH was totally absent (Fig. 4A), as was the case in the presence of ethanol (Fig. 4B). In the presence of formate, weak signals from DMPO/ * CO; could be detected (Fig. 4C). Again, it is clear that an oxidant other than the hydroxyl radical is produced at low concentrations even in the presence of desferrioxamine. However, the adriamycin semiquinone signal was formed immediately in these systems (Fig. 4). This implies little or no turnover of hydrogen peroxide in the adriamycin semiquinone-driven Fenton reaction in the presence of desferrioxamine. Effect of radical scavengers on the rate of reduction of As can be seen in Figs. 2B-D, the adriamycin f&Q+ semiquinone concentration is not detectable until the hydrogen peroxide is depleted. The time lag can therefore be attributed to removal of hydrogen peroxide in this radical-driven Fenton system. When radical scavengers such as ethanol, formate, mannitol, etc. are present, scavengerderived radicals, namely the a-hydroxyethyl radical and the carbon dioxide radical anion, are also formed (Fig. 3), which can affect the rate of depletion of hydrogen peroxide. Time lag experiments show that, under otherwise identical conditions, addition of ethanol decreases the lag time in a concentration-dependent fashion (Fig. 5), implying that the rate of consumption of hydrogen peroxide (i.e., [H,O#t) is increased in the presence of ethanol. In the absence of hydrogen peroxide, ethanol (at these concentrations) had no effect on the rate of formation of the adriamycin semiquinone radical (data not shown). Addition of mannitol, deoxyglucose, isopropanol, and formate to this system (cf. Fig. 5) increased the rate of con-

167

SEMIQUINONE

~2?.J---t FIG. 5. Time-course ESR spectra of semiquinone formed during anaerobic reduction of adriamycin by xanthine/xanthine oxidase. (A) In the presence of 75 pM H202 and 100 pM EDTA and (B) same as (A) but containing ethanol. Note a threefolddecrease in time lag in the presence of ethanol. The enzymatic activity as measured by uric acid formation was not altered in the presence of ethanol (100 mM). The enzyme was added at the point marked t.

sumption of HzOz (Table I). We verified that this effect is not due to nonenzymatic depletion of Hz02 in these systems. DISCUSSION

Radical-driven Fenton reaction is an alternative pathway for production of hydroxyl radical and other higher oxidant species in biological systems (22,23). In an ironcatalyzed Haber-Weiss reaction, the superoxide anion acts as the reductant of iron (23). Chelated Iron-Catalyzed

Haber- Weiss Mechanism

Fe3+ (chelate) + 0; + Fe2+ (chelate) + O2 Fe2+ (chelate) + H202 --f Fe3+ (chelate) + . OH $ OH0; + H202 --f .OH+OH-+02 In radical-driven

Fenton

TABLE

Effect of Inhibitors

Incubation Controlb Control + + t t

the radical

mixture

(200mM)

R.

I

on the Rate of Depletion

mannitol (60 mM) ethanol (100 mM) isopropanol (80 mM) glucose

reactions,

of H,02

(nmol/min) 10.6 22.6 25 20 15.2

n Rate of consumption of hydrogen peroxide is measured indirectly using the time lag (t) detected prior to the appearance of the adriamycin semiquinone radical. * Incubation mixture consisted of adriamycin (100 FM), xanthine (400 pM), xanthine oxidase (0.2 units), DETAPAC (100 PM), and I&O, (280 pM)in 2 ml phosphate buffer (100 mM, pH 7.5) under N1.

168

KALYANARAMAN,

MOREHOUSE,

(usually, but not always, a radical anion) acts as a reductant.

Fe3+ (chelate) + R * + Fe’+ (chelate) + R+ Fe’+ (chelate) + HzOz + . OH + OH- + Fe3+ (chelate) .OH+OH-+R+ R. +HzOz+ Fenton Reaction in the Absence of EDTA

Fe3+ + R. + Fe2+ + R+ Fe2+ + H202 + [ox] R- +H202+ [ox] +R+ ([ox] refers to Fe(IV), Fe02+, or a peroxo complex.) At high oxygen tension R * preferentially lecular oxygen, as shown below:

reacts with mo-

R. +O,+O$+R+ In summary, R * can react with trace transition metals (presumably iron) which then react with H202, forming two types of oxidants, the hydroxyl radical or a higher oxidant, depending upon the chelator. The reducing radical R. generally reacts much faster with iron-chelates than 0;. Only submicromolar amounts of adventitious transition metal or transition metal-chelates are presumably needed to maintain the radical-driven Fenton reaction as long as there is a sufficient amount of H202 present

Net reaction:

*OH + C2H50H CH,‘CHOH + Fe3+ (chelates) Fe2+ (chelate) + H202 C2H50H + H202

Similar reactions can be postulated for other scavengerderived radicals which are shown in Table II. Based on the reduction potentials, it is evident that inhibitor-derived radicals are powerful reducing species in radicaldriven Fenton reactions, even in the presence of desferrioxamine (Table II). Under experimental conditions when methyl radicals are formed (i.e., in the presence of DMSO as in Fig. 5) we did not observe any decrease in lag time due to consumption of H202. In contrast, the adriamycin semiquinone radical was not detected at all. It is possible that the methyl radicals react with the adriamycin semiquinone to form a diamagnetic product, or the rate of re-

MASON

(26,34). Consistent with this, treatment of only the buffer with Chelex 100 (which apparently does not remove all the adventitious trace transition metals) is not enough to abolish the radical-driven Fenton-type reaction (34). Based on product analysis, production of hydroxyl radical or a related oxidant has been postulated in paraquat radical- and adriamycin semiquinone radical-driven Fenton reactions (35). Although the structural identity of the oxidant is not known, it has been ascribed to a perferryl species (Fe(IV) or Fe(III)-0). To a large extent, the present spin trapping data are in agreement with these findings. Hydroxyl radicals were detected only in the presence of EDTA and DETAPAC. In the absence of added chelators, there is no evidence for the hydroxyl radical. The spin trapping data obtained in the absence of chelator, however, differ from product analysis data with respect to oxidation of scavengers. Whereas results from product studies based on the measurements of carbon dioxide demonstrate little or no oxidation of formate in the absence of chelator, the spin trapping results have clearly provided evidence for formation of CO; (Fig. 3), although the corresponding spin adduct is detected in much smaller amounts. Addition of scavengers or inhibitors to radical-driven Fenton reactions has an added caveat. Inhibitor-derived radicals can be reducing or oxidizing species (36, 37). Radicals derived from ethanol, isopropanol, mannitol, deoxyribose, and formate are of the reducing type, whereas those derived from benzoic acid and dimethylsulfoxide are of the oxidizing type. Reducing radicals can enhance the rate of turnover of hydrogen peroxide by participating in chain-radical reactions (36). For example, increased consumption of H202 in the presence of ethanol (Fig. 5) can be explained on the basis of the following reactions:

Radical-Driven Fenton Reaction in the Presence of EDTA or DETAPAC

Radical-Driven or DETAPAC

AND

+ + + +

CH;CHOH + Hz0 Fe2+ (chelate) + CH3CH0 Fe3+ (chelate) + . OH + OHCH3CH0 + OH- + Hz0

duction of H202 is extremely slow as a result of rapid oxidation of Fe’+ to Fe3+ by methyl radicals. Further investigation is needed to fully explain this finding, but it is clear that scavenger-derived radicals can affect the overall mechanism of reaction of hydrogen peroxide in the Fenton reaction. Since hydroxyl radical scavengers are frequently used in in vitro and in vivo experiments, the results obtained in the present study are clearly significant. Minotti and Aust (38) have reported that the addition of radical scavengers such as mannitol and benzoate had opposing effects on the rate of oxidation of Fe2+ in the Fenton reaction. Using the o-phenanthroline method, they

REACTIONS

OF ADRIAMYCIN TABLE

Reactions

of Reducing

169

SEMIQUINONE

II

(A) and Oxidizing

Radicals”~b~c [Refs.

(44-47)]

(A) R. +M”+‘+R++M”+

Reduction

CO; + Fe3+(EDTA)

E’(COJCO;)

(CH,);COH

+ CO2 + Fe’+(EDTA)

+ Fe3+(EDTA)

-2 (CH,),CO

03)

of radicals (V)

= -1.8

E’(CH&OCH&H&COH

+ Fe’+(EDTA)

CHsCHOH + Fe3+(EDTA) -z CH,CHO + Fe’+(EDTA) Mannitol radical t Fe3+(EDTA) + Product + Fe’+(EDTA) Adriamycin semiquinone (Adr-) + Fes+ (EDTA) --* Adriamycin

potentials

(Adr) + Fe*+ (EDTA)

= -1.6

EO(CH,CHO/CH;CHOH) E”(Adr/Adr-) = -0.33

= -1.4

R. +M”+-+R-+M”+’

. CH, + Fe’+(EDTA)

‘z

E”(CH4/.

CH, + Fe3+(EDTA)

’ Reducing radicals enhance the rate of consumption of H,O, in radical-driven Fenton reactions. * M”+’ refers to Fe3+, CU*+, C?‘, etc. ’ E”(Fe3+(EDTA)/Fe2+(EDTA)) = +0.177 V. E”(Fe3+(DETAPAC)/Fe2+(DETAPAC)) = +0.03 ferrioxamine)) = - 0.45 V.

reported that the rate of Fe’+ to Fe3+ oxidation was decreased by mannitol and increased by benzoate (38). As discussed earlier, the mannitol-derived radical is reducing (i.e., it facilitates reduction of Fe3+ to Fe’+) and the radical derived from benzoate is oxidizing (i.e., it facilitates oxidation of Fe’+ to Fe3+) in nature. Consequently, the rate of turnover of HzOz in this system is either enhanced or decreased. Many years ago, Beauchamp and Fridovich (39) suggested that the greater effectiveness of benzoate as compared to ethanol as an inhibitor of ethylene production could be due to its higher rate constant for reaction with ‘OH. Alternatively, this effect can also be explained based on the occurence of chain-radical reactions in the presence of ethanol. Gutteridge (40) and Halliwell et al. (41) have also postulated the back-reaction between a deoxyribose-derived radical and the iron-EDTA complex. A radical chain-reaction mechanism has been proposed during microsomal oxidation of ethanol (42). We feel that the present methodology may be useful in studying other reactions of the adriamycin semiquinone. Using ESR, evidence for the mobilization of iron from ferritin by adriamycin semiquinone and other radicals has been obtained (43). Based on the observed decrease in the ESR signal intensity of the adriamycin semiquinone in the presence of ferritin, it was concluded that a reaction occurs between the adriamycin semiquinone and ferritin. However, the present data indicate that, were this reaction to occur, one would expect to see a time lag in the appearance of the adriamycin semiquinone. In conclusion, we have obtained evidence for the formation of the hydroxyl radical and an as yet unidentified related oxidant during adriamycin semiquinone-catalyzed reduction of H202 in the presence of adventitious transition metal(s). We have also shown that scavenger-derived radicals either decrease or increase the rate of re-

CHs) = +1.6

V. E”(Fe3+(Desferrioxamine)/FeZ+(Des-

duction of HzOz through radical-driven in this system.

Fenton reactions

ACKNOWLEDGMENT This work was supported by the National GM-29035 and RR-01008.

Institutes

of Health Grants

REFERENCES 1. Lernaz, L., and Page, J. A. (1976) Cancer Treat. Rev. 3, 111-120. 2. Powis, G. (1987) in Metabolism and Action of Anti-Cancer Drugs (Powis, G., and Prough, R. A., Eds.), pp. 211-260, Taylor and Francis, New York. 3. Ferrans, V. J. (1978) Cancer Treat. Rep. 62, 955-961. 4. Handa, K., and Sato, S. (1975) Gann 66, 43-47. 5. Bachur, N. R., Gordon, S. L., and Gee, M. V. (1977) Mol. Pharmacol. 13,901-910. 6. Bather, N. R., Gordon, M. V., Gee, M. V., and Kon, H. (1979) Proc. Natl. Acad. Sci. USA 76,954-957. 7. Goodman, J., and Hochstein, Commun. 77, 797-803.

P. (1977) Biochem.

Biophys.

Res.

8. Gutierrez, P. L., Gee, M. V., and Bachur, N. R. (1983) Arch. Biochen. Biophys. 223, 668-675. 9. Kalyanaraman, B., Perez-Reyes, E., and Mason, Biochim. Biophys. Acta 630, 119-130.

R. P. (1980)

10. Powis, G., and Appel, P. A. (1980) Biochem. Pharmacol. 2572. 11. Thayer,

W. W. (1977) Chem. Biol. Interact.

12. Doroshow,

29, 2567-

19, 2655278.

J. H. (1986) Proc. Natl. Acad. Sci. USA 83,4514-4518.

13. Lown, J. W. (1985) Adu. Free Rad. Biol. Med. 1, 225-264. 14. Doroshow, J. H. (1983) Cancer Res. 43,4543-4551. 15. Butler, J., Hoey, B. M., and Swallow, A. J. (1985) FEBS L&t. 182, 95-98. 16. Thomas, C. E., and Aust, S. D. (1986) Arch. Biochem. Biophys. 248, 684-689.

170

KALYANARAMAN,

MOREHOUSE,

17. Myers, C. E., McGuire, W. P., and Young, R. (1976) Cancer Treat. Rep. 60,961-962. 18. Lown, J. W., Chen, H. H., Plambeck, J. A., and Acton, E. M. (1979) Biochem. Pharmacol. 28,2563-2568. 19. Myers, E. C., Muindi, J. R., Zweier, J., and Sinha, B. K. (1987) J. Biol. Chem. 262, 11,571-11,577. 20. Myers, C. E., McGuire, W. P., Liss, R. H., Ifrim, I., Grot Zinger, K., and Young, R. C. (1977) Science 197, 165-167. C. C. (1988) Biochem. Phurmacol. 37, 21. Vile, G. F., and Winterbourn, 2893-2897. C. C. (1987) Free Rad. Biol. Med. 3, 33-39. 22. Winterbourn, 23. Sutton, H. C., Vile, G. F., and Winterbourn, C. C. (1987) Arch. Biochem. Biophys. 256,462-471. C. C. (1982) FEBS Lett. 145, 13724. Bates, P. A., and Winterbourn, 142. 25. Kalyanaraman, B., Sealy, R. C., and Sinha, B. K. (1984) Biochim. Biophys. A& 799, 270-275. C. C., Gutteridge, J. M. C., and Halliwell, B. (1985) 26. Winterbourn, J. Free Rad. Biol. Med. 1, 43-49. 27. Winterbourn, C. C. (1983) FEBS L&t. 136,89-84. C. C., and Sutton, H. C. (1987) Arch. 28. Vile, G. F., Winterbourn, B&hem. Biophys. 259,616-626. C. C., (1988) Arch. Biochem. Biophys. 29. Vile, G. F., and Winterbourn, 267,606-613. P. J., Bannister, W. H., and Bannister, J. V. (1986) 30. Thornalley, Free Rad. Res. Commun. 2,163-171. 31. Rumyantseva, G. V., and Weiner, L. M. (1988) FEBS L&t. 234, 459-463.

AND

32. Buettner

MASON

(1988) J. Biochem. Biophys. Methods 16, 27-40.

33. Mason, R. P. (1984) in Methods in Enzymology (Packer, L., Ed.), Vol. 105, pp. 416-422, Academic Press, San Diego. 34. Morehouse, 1204-1211.

K. M., and Mason, R. P. (1988) J. Biol. Chem. 263,

35. Winterbourn,

C. C. (1987) Free Rad. Biol. Med. 3.33-39.

36. Behar, D., Samuni, A., and Fessenden, R. W. (1973) J. Phys. Chem. 77,2055-2059. 37. Koppenol, 4430. 38. Minotti,

W. H., and Rush, J. D. (1987) J. Phys. Chem. 91,4429G., and Aust, S. D. (1987) J. Biol. Chem. 262,1098-1104.

39. Beauchamp, 4646. 40. Gutteridge,

C., and Fridovich,

I. (1970) J. Biol. Chem. 18, 4641-

J. M. C. (1987) Biochem. J. 243, 709-714.

41. Halhwell, B., Gutteridge, Biochem. 165, 215-219.

J. M. C., and Aruoma, 0. I. (1987) Anal.

42. Feierman, D. E., Winston, G. W., and Cederbaum, A. I. (1985) Alcohol. Clin. Exp. Res. 9, 95-102. 43. Samokyszyn, V. M., Thomas, C. E., Reif, D. W., Saito, M., and Aust, S. D. (1988) Drug Metub. Reu. 19, 283-303. 44. Koppenol, 1,91-131. 45. Swallow,

W. H., and Butler,

A. J. (1978) Prog. React. &net.

46. Miller, D. M., Beuttner, BioL. Med. 8, 95-108. 47. Wardman,

J. (1985) Adu. Free Rad. Biol. Med. 9, 195-366.

G. R., and Aust, S. D. (1990) Free Rad.

P. (1989) J. Phys. Chem. Ref. Data 18, 1637-1755.