Kinetic Studies on Spin Trapping of Superoxide and Hydroxyl Radicals Generated in NADPH-Cytochrome P-450 Reductase-Paraquat Systems EFFECT OF IRON

Electron spin resonance (ESR) studies on spin trap- ping of superoxide and hydroxyl radicals by 5,5-di-methyl- 1-pyrroline-l-oxide (DMPO) were performed in NADPH-cytochrome P-450 reductase-paraquat systems at pH 7.4. Spin adduct concentrations were de- termined by comparing ESR spectra of the adducts with the ESR spectrum of a stable radical solution. Kinetic analysis in the of desferriox- amine showed that:

Since the amount of such oxygen-radical adducts detected by ESR is only a part of the overall biochemical metabolites produced, it is necessary to elucidate kinetic meanings of the spin detection in the overall metabolism.
To clarify the quantitative relationships between enzymatic reactions and spin-trapping data of oxygen radicals, we have used the NADPH-cytochrome P-450 reductase-paraquat system as a standard system, since the primary reduction product of oxygen is thought to be only superoxide.
For instance, xanthine oxidase, a widely used superoxide-generating enzyme produces both superoxide ions and hydrogen peroxide as primary products (5,35) and the kinetic analysis is considerably more complex.
Several papers have reported that the Haber-Weiss reaction is too slow to explain the formation of hydroxyl radicals in '  biological systems (36) and it is now well accepted that iron plays an essential role in the formation of hydroxyl radicals from superoxide-generating systems. Since the role of iron is greatly modified by its chelators, we have attempted to examine the effects of three typical iron chelators, desferrioxamine B (deferoxamine), diethylenetriamine pentaacetic acid (DETAPAC), and EDTA. EXPERIMENTAL PROCEDURES AND RESULTS'

DISCUSSION
Paraquat-mediated superoxide generation has been reported in NADPH-cytochrome P-450 reductase systems (58), glutathione reductase systems (41), animal systems (59, 60), plant systems (28, 61), and microorganisms (44,62,63). It is believed that reducing equivalents accepted by paraquat are all used to reduce oxygen to superoxide ions, but without any quantitative evidence. The kinetic data in Fig. 5 show that within experimental errors the following stoichiometry is established in the NADPH-cytochrome P-450 reductase-paraquat system, NADPH + 2 02 + NADP+ + H+ + 2 O;, and also that all the resultant superoxide ions can be trapped by DMPO when extrapolated to infinite DMPO concentration. The slopes in Fig. 5 give a value of 2.2 X lo5 M . s for kJ k$, which is much higher than the value of 3 X lo3 M.s obtained using data of kd = 3 X lo5 M-' s-' (64) and kz = 10 M-l s-l (31) at pH 7.4. If one assumes that the kd value is correct, the kZ value will be about 1.2 M-' s-l, which is significantly less than that reported by Finkelstein et al. (31). This difference is probably, in part, ascribable to the difference in the experimental conditions (ionic strength, iron chelator, etc.) The DMPO concentration that traps one-half of the superoxide ions accumulated increases with the increase in the rate of superoxide formation as can be seen in Fig. 5 and Table II. For the DMPO-OOH decay we conclude as follows: 1) DMPO-OOH decays with a half-life of 66 s (kl = 0.011 s-l) at pH 7.4 according to first order kinetics. The rate is nearly the same as that obtained in illuminated pea chloroplasts at pH 7 (28). 2) the DMPO-OOH decay is accompanied by the production of a small amount of DMPO-OH. The conversion ratio is measured to be 2.8%, that is, 0.8 pM DMPO-OH is formed during the decay of 29 ,uM DMPO-OOH, which is equal to the approximate integrated value, Jk,[DMPO-OOH] dt, obtained from the kinetic trace for formation and decay of DMPO-OOH as shown in Fig. 4A. Here, a value of 0.011 s-' is used for kl. The total amount of DMPO-OOH formed during the reaction will be greater than 29 pM since kl appears to be greater than 0.011 s-' under these experimental conditions (Table I). DMPO-OH is assumed to be stable under these conditions. Since Fig. 6C shows that DMPO-OH is formed through the reaction of DMPO with hydroxyl radicals, we conclude that 2.8% or less of DMPO-OOH decay occurs through a reaction producing hydroxyl radicals. Finkelstein et al. (34) have reported that the ratio is about 3%. The formation of DMPO-OH from DMPO-OOH has been discussed also in neutrophil systems (65). 3) DMPO-OOH reaches a steady level a few minutes after initiation of the reactions in the presence of 5-10 FM paraquat (Fig. 2). Equa-'Portions of this paper (including "Experimental Procedures, " "Results," Figs. l-13, and Tables I-III) are presented in miniprint at the end of this paper.
Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.
tion 1 is valid under these experimental conditions. The above conclusions are obtained from reactions in the presence of 100 PM deferoxamine. According to recent reports (66, 67), deferoxamine reacts with both superoxide ions and hydroxyl radicals, with rate constants of 9 x 10' and 10" M-' s-l, respectively. These reactions may be negligible in our reaction systems since the concentration of DMPO is 1000 times higher than that of deferoxamine and we could not observe any nitroxide-free radical which is a product of the reaction of deferoxamine with superoxide (67). When deferoxamine is present, the mechanism of oxygen metabolism in the NADPH-cytochrome P-450 reductase-paraquat system is relatively simple as shown in Fig. 14A. The formation of hydroxyl radicals through reactions of hydrogen peroxide with superoxide and paraquat radicals is not detectable under our experimental conditions.
2) When the superoxide generation is slow, 5 pM Fe(III)-EDTA completely inhibits the DMPO-OOH accumulation (Fig. 8C). As seen in Fig. 14B, this inhibition may occur through two mechanisms, one is reductive decomposition of DMPO-OOH by Fe(II)-EDTA (Reaction 7) and the other is a weak superoxide dismutase activity of iron-EDTA (Reactions 5, 6, and 6'). The superoxide dismutase activity of Fe(III)-EDTA is still controversial mostly because of the apparent slow dissociation of an iron-EDTA-peroxide complex (Reaction 6'). At pH 7.4, the value for ke' is estimated to be at least 1 s-i or possibly larger (48,50). The rate constant of 1 s-l would be very slow for iron-EDTA to catalyze the dismutation of superoxide ions at any significant rate under the usual assay conditions for 0; as described by Diguiseppi and Fridovich (54). Under the experimental conditions in Fig.  8C where the rate of superoxide formation is about 0.2 pM s-i (Table I) and the rate of reaction of DMPO with superoxide ions is very slow, calculations using the known rate constants shown in the legend of Fig. 14 and using a value of 1 s-i for kg' clearly show that the maximal DMPO-OOH accumulation is decreased from 6.2 (Table I) to 0.23 pM through Reactions 5, 6, and 6' in the presence of 5 pM iron-EDTA. Under the same conditions the DMPO-OOH concentration would decrease to about one-sixth the original concentration following Reaction 7 alone. Since these two mechanisms operate additively, the DMPO-OOH accumulation following both decay mechanisms would be expected to decrease to a level below the ESR sensitivity (Table III). As the superoxide generation becomes faster, superoxide ions disappear mostly through dismutation and more Fe(III)-EDTA is needed for complete suppression of DMPO-OOH formation (Fig. 1lA).
3) Reaction of the Fe(II)-EDTA is switched from Reaction 6 to Reaction 8 by the presence of 100 pM hydrogen peroxide (Figs. 8C and SB). Only a part of Reaction 8 may result in the formation of hydroxyl radicals (Reaction 8'). The other may be followed by hydrogen peroxide-consuming reactions (Reactions 9 and lo), the detailed mechanism remaining to be clarified. Rush and Koppenol (52) have suggested the possibility that hydroxyl radicals are formed via Reaction 10. The efficiency of hydroxyl radical formation in the Fenton's reaction is also a problem to be solved (72, 73). The increase of DMPO-OOH formation in the presence of hydrogen peroxide (Table III) can be explained in terms of the decrease in the Fe(B)-EDTA concentration. 4) In the presence of a certain amount of hydrogen peroxide, as shown in Table III, the increase in the rate of superoxide generation brings about a depression in hydroxyl radical production. This depression can be partially removed by increasing the concentration of hydrogen peroxide. Fig. 14B shows such competition between Reactions 6 and 8. However, the superoxide-induced destruction of DMPO-OH, recently reported by Samuni et al. (74) should also be considered.
Although the results shown in Table III are somewhat complicated, those in the presence of EDTA can be explained by the known kinetic data (Fig. 14B). The results with DE-TAPAC and endogenous chelator, however, cannot be completely explained because of lack of detailed kinetic data. A slight increase of DMPO-OOH formation by hydrogen peroxide in the presence of deferoxamine also remains unex-plained. In Table III we consider that [DMPO-OH],., as measured by ESR is nearly equal to the total amount of DMPO-OH accumulated during the course of the reaction because of the inherent stability of the DMPO-OH adduct while the total amount of DMPO-OOH actually formed is much greater than that measured by ESR as [DMPO-OOHlm, because of the inherent instability of the DMPO-OOH adduct.
It is now clear that spin trapping by DMPO can be used effectively for kinetic analysis of oxygen radicals generated in enzyme reactions even though the reaction of DMPO with superoxide is slow and its product is unstable. The most important criticism to be raised might be that the generation of hydroxyl radicals following superoxide formation is modified by the trapping of superoxide by DMPO as discussed by Britigan et al. (65) and Kleinhans and Barefoot (33) in neutrophil systems. New kinetic approaches are necessary to solve this problem, which is now under investigation in this laboratory.