The vanadate-stimulated oxidation of NAD(P)H by biomembranes is a superoxide-initiated free radical chain reaction

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 250, No. 1, October, pp. 139-145,1986

The Vanadate-Stimulated Oxidation of NAD(P)H by Biomembranes Superoxide-Initiated Free Radical Chain Reaction’ STEFAN LIOCHEV2 Department

of Biochemistry,

IRWIN FRIDOVICH3

AND

Duke University

Is a

Medical Center, Durham,

North Carolina

27710

Received December 27, 1985, and in revised form May 29, 1986

Rat liver microsomes catalyze a vanadate-stimulated oxidation of NAD(P)H, which is augmented by paraquat and suppressed by superoxide dismutase, but not by catalase. NADPH oxidation was a linear function of the concentration of microsomes in the absence of vanadate, but was a saturating function in the presence of vanadate. Microsomes did not catalyze a vanadate-stimulated oxidation of reduced nicotinamide mononucleotide (NMNH), but gained this ability when NADPH was also present. When the concentration of NMNH was much greater than that of NADPH a minimal average chain length could be calculated from ‘/z the ratio of NMNH oxidized per NADPH added. The term chain length, as used here, signifies the number of molecules of NMNH oxidized per initiating event. Chain length could be increased by increasing [vanadate] and [NMNH] and by decreasing pH. Chain lengths in excess of 30 could easily be achieved. The K, for NADPH, arrived at from saturation of its ability to trigger NMNH oxidation by microsomes in the presence of vanadate, was 1.5 PM. Microsomes or the outer mitochondrial membrane was able to catalyze the vanadate-stimulated oxidation of NADH or NADPH but only the oxidation of NADPH was accelerated by paraquat. The inner mitochondrial membrane was able to cause the vanadate-stimulated oxidation of NAD(P)H and in this case paraquat stimulated the oxidation of both pyridine coenzymes. Our results indicate that vanadate stimulation of NAD(P)H oxidation by biomembranes is a consequence of vanadate stimulation of NAD(P)H or NMNH oxidation by 02, rather than being due to the existence of vanadate-stimulated NAD(P)H oxidases or dehydrogenases. 0 1986 Academic Press, Inc.

Vanadate stimulates the oxidation of NAD(P)H by biomembranes (l-6) and this has been attributed to a vanadate-dependent NAD(P)H oxidase or to an NAD(P)H dehydrogenase, which is converted to an oxidase by vanadate (2, 4, 6). Since ’ This work was supported by research grants from the Council for Tobacco Research, USA, Inc., the National Science Foundation and the U.S. Army Research Office. a S.L. was a fellow of the International Atomic Energy Agency. Institutional affiliation: Institute of Physiology, Bulgarian Academy of Sciences, Acad. G. Bonchev St., bl., 23,1113, Sofia, Bulgaria. 3 To whom correspondence should be addressed. 4 Abbreviations used: SOD, superoxide dismutase;

SOD4 inhibited the vanadate-stimulated NAD(P)H oxidation (l-3,5), we considered the possibility that vanadate was able to catalyze the oxidation of NAD(P)H by 02, by a free radical chain mechanism, and we demonstrated that it could do so with enzymatic, photochemical, or chemical sources of 02 (7-9). This, in turn, suggested that vanadate might have stimulated the oxidation of NAD(P)H by biomembranes merely because the membranes could produce 0;. In this case NAD(P)H plus van-

NMNH, reduced nicotinamide paraquat. 139

mononucleotide;

P&a’,

0003-9861186 $3.00 Copyright All rights

D 1986 by Academic Press, Inc. of reproduction in any form reserved.

140

LIOCHEV

AND

adate could be viewed as an amplifying detector of 0; production; and the NAD(P)H could be considered to play two distinct roles, i.e., as a source of electrons for some OS-producing NAD(P)H oxidase and as a part of the amplifying detector. We have explored this possibility in two ways. On the one hand compounds such as menadione and paraquat, which increase 0; production, should augment the SOD-inhibitable, vanadate-stimulated NAD(P)H oxidation by biomembranes. On the other hand a dihydropyridine compound such as NMNH, which is not oxidized by bio-membranes, should show no vanadate-stimulated oxidation unless NADPH was present to allow 0; production. Moreover, in this situation with excess NMNH one molecule of NADPH could serve to facilitate the oxidation of several molecules of NMNH, reflecting the chain length; and SOD should inhibit. We now report the affirmation of these expectations. MATERIALS

AND

RESULTS

oxidation of NADPH Tracing 1 in Fig. 1 demon-

Vanudab-stimulated

2

----

-----

--------------__---____

3

4

5 +MiCrOlOrne*

METHODS

Ammonium metavanadate, NADH, NADPH, NMNH, methyl viologen (paraquat), and catalase were from Sigma; sucrose and EDTA were from Mallinkrodt and menadione from Aldrich. SOD was isolated and assayed as previously described (10). Livers, removed from 250 g male albino rats immediately after decapitation, were chilled and homogenized in 10 vol of cold 0.25 M sucrose, 10 mM sodium phosphate, 0.5 mM EDTA, at pH 7.2. Inner and outer mitochondrial membranes were isolated as previously described (11). The 600 g supernatant solution from the liver homogenate was twice centrifuged for 20 min at 15,OOOg and the resultant supernatant solution was centrifuged for 1 h at 105,0009, yielding a pellet of microsomes, which were resuspended to the original volume in the homogenization buffer but lacking the EDTA. These microsomes were either used fresh or were stored at -7O’C until needed. Protein concentration was estimated calorimetrically (12). NAD(P)H or NMNH oxidations were monitored at 340 nm in 3.0 ml reaction volumes containing 100 mM sodium phosphate at 37°C. NAD(P)H or NMNH oxidation in the presence of biomembranes was corrected for the rate seen in the absence of these membranes. When inner mitochondrial membranes were used, 7.5 ~1 of a 1.0 mM solution of antimycin A in ethanol was also added (final concentration of antimycin A = 2.5 PM).

and NMNH.

FRIDOVICH

MINUTES FIG. 1. Oxidation of NADPH and NMNH by microsomes. Reaction mixtures contained 100 mM sodium phosphate at pH 7.6 and 37’C and where indicated 320 pM NADPH, 400 pM vanadate, 50 pg/ml of microsomal protein, 5 fig/ml SOD, 500 pM paraquat, and 320 FM NMNH. When NADPH was added after NMNH (tracings 3-5) it was added to 3.2 pM.

strates that vanadate markedly stimulated the oxidation of NADPH by rat liver microsomes, while subsequent addition of paraquat caused further augmentation and that SOD inhibited. Paraquat, per se, moderately increased NADPH oxidation (tracing 2). SOD had no effect on this oxidation in the absence of vanadate. Subsequent addition of vanadate stimulated this oxidation and SOD again inhibited. These effects of paraquat could be duplicated with 0.1 mM menadione (data not shown). NMNH was not detectably oxidized by microsomes, even in the presence of vanadate; but addition of only 3.2 PM NADPH allowed a rapid transient oxidation of NMNH. Repeated addition of a catalytic amount of NADPH caused another burst of NMNH oxidation (tracing 3). Complete oxidation of 3.2 PM NADPH would cause a change in

0, PRODUCTION

141

BY BIOMEMBRANES

AsdOnrnof only 0.02; whereas the AAs4,,,,,,, NAD(P)H dehydrogenases or oxidases (l-

seen when this amount of NADPH was added in the presence of 320 PM NMNH was 0.47. It is apparent that under these conditions, oxidation of one molecule of NADPH caused the oxidation of 23 molecules of NMNH. The molar extinction change at 340 nm, accompanying the oxidation of NAD(P)H or NMNH, was taken to be 6.22 X lo3 M’ cm-’ (13). Changing the order of addition of NMNH, microsomes, and vanadate had no effect on the catalytic effect of added NADPH (tracing 4); which was eliminated by SOD (tracing 5). It appears that NADPH oxidation, by an enzyme associated with the microsomal membranes, produced 0; which could then initiate the vanadate-mediated free radical chain oxidation of NMNH. The ability of inner and outer mitochondrial membranes to catalyze the vanadate-stimulated oxidation of NADH and NADPH was also examined, in the absence and in the presence of PQ’+. All membrane fractions examined (microsomes, outer and inner mitochondrial membranes) catalyzed a vanadate-stimulated oxidation of either NADH or NADPH. PQ’+ augmented this oxidation of NADPH in all membrane fractions, but increased the oxidation of NADH only with the inner mitochondrial membranes. Vanadate enhancement of the effect of paraquat on the oxidation of dihydropyridines was seen only when paraquat, in the absence of vanadate, caused some increase in this oxidation. This is as expected since only then could paraquat be mediating increased 0% production. Submitochondrial particles from rat liver have previously been reported to catalyze a vanadate-dependent oxidation of NADH in the presence of inhibitors of electron transport such as antimycin A or rotenone (14). Our results are in accord with this earlier report (14) but contradict the subsequent statement that inner mitochondrial membranes do not catalyze a vanadate-stimulated oxidation of NADH (1). Efect of varying

the concentration

of mi-

cro~omes. Previous papers describing the catalysis, by biomembranes, of the vanadate-stimulated oxidation of NAD(P)H; attributed it to vanadate-dependent

6). In that case the activity should be a linear function of the concentration of microsomes. In contrast, our demonstration that vanadate catalyzes an OF-initiated free radical chain oxidation of NAD(P)H (7-9) suggests that microsomes plus NADPH generate 0,; whose ability to initiate NAD(P)H or NMNH oxidation is merely amplified by vanadate. Since radical-radical termination reactions become more rapid the greater the number of parallel chain reactions in progress; average chain length decreases with increases in the number of such parallel chain reactions (15). In that case the vanadate-stimulated oxidation of NAD(P)H by microsomes should be a saturating function of microsome concentration. Figure 2 presents the results pertinent to these predictions. Line 1 demonstrates that the rate of oxidation of NADPH is a linear function of the concentration of microsomes, as would be expected of an enzymatic, nonchain process. In contrast, line 2 shows that NADPH oxidation, in the presence of vanadate, is a saturating function of the concentration of microsomes, as expected for a free radical chain reaction. When 3.2 PM NADPH was used to

z 2 4o

- 16

4

E : z 30

-E

Iz

3;

b 20 I 2 z '0

2bN

zc

0

\E

-12 zz(3

Y

-8

2

0

100

200

MICROSOMES

300

400

I=

-4

0

-0

z 3 5

(~g/ml)

FIG. 2. Rate and chain length as a function of [microsomes]. The oxidations of NADPH or of NMNH were examined as a function of the concentration of microsome protein, under the conditions described in the legend of Fig. 1. Line 1, NADPH; line 2, NADPH plus vanadate; line 3, 320 pM NMNH plus vanadate, initiated by 3.2 pM NADPH; line 4, rate of 0; production calculated as rate of NMNH oxidation divided by chain length; line 5, chain length calculated as ‘/z [NMNH] oxidized per NADPH added.

142

LIOCHEV

AND

trigger the oxidation of NMNH whose initial concentration was 320 PM, we were able to estimate minimal chain length as 1 the ratio of NMNH oxidized. Line 3 presents the rate of NMNH oxidation in the presence of NADPH as a function of microsome concentration; while line 5 presents the calculated minimal chain lengths, which decreased from 13 at 25 pg/ml of microsomal protein to 4.5 at 350 pg/ml. Chain lengths were calculated as molecules of NMNH oxidized per $ molecule of NADPH added. The factor f was used because each equivalent of NADPH could give rise to a maximum of two equivalents of superoxide radical. The maximal rate of 0, production could be calculated by dividing the rate of NMNH oxidation by the estimated chain length and this is presented by line 4. These results indicate that the vanadate-stimulated microsomal oxidation of NAD(P)H is a free radical chain reaction initiated by 02, which is produced by a microsomal NAD(P)H oxidase. NADPH triggered NMNH oxidation. Rat liver microsomes catalyzed the rapid reduction of cytochrome c in the presence of NADH or NADPH, but not in the presence of NMNH. This confirms that microsomal enzymes can abstract electrons from NAD(P)H, but not from NMNH, and supports the view that NADPH stimulated NMNH oxidation by microsomes, in the presence of vanadate, precisely because NAD(P)H allowed some 0, production by the microsomes. The observation that 5 pg/ ml of SOD caused 93% inhibition of the NADPH-stimulated NMNH oxidation is in full accord with this view. In order to gain further confidence in this interpretation of the data, we examined the NADPH-triggered, vanadate-stimulated oxidation of NMNH by microsomes as a function of the concentration of the several reactants and of the pH. In each study we also calculated minimal chain lengths and maximal rates of 02 production, as was done for the data in Fig. 2. Line 1 in Fig. 3 presents the rate of vanadate-stimulated oxidation of NMNH by microsomes, as triggered by NADPH, as a function of [NADPH]. Line 2 presents the minimal average chain length and line 3

FRIDOVICH 25

\ IN - 1.6 o

OL OK----J 0

2

4 NADPH

6 OOJ)

8

IO”

FIG. 3. Effect of varying the concentration of NADPH. Reaction mixtures contained 100 mM sodium phosphate, 320 pM NMNH, 400 pM vanadate, 50 pg/ ml microsomes (as protein) and the indicated concentrations of NADPH at pH 7.6 and at 37°C. Line 1, nmol NMNH oxidized/min/ml; line 2 average chain length calculated as i/z the molecules of NMNH oxidized per NADPH added; line 3, nmol 0; produced/ min/ml calculated as the rate of NMNH oxidation divided by the chain length.

the rate of 0; production. These measurements were done at a relatively low concentration of microsomes (50 pg/ml, as protein). Saturation of the rate of NMNH oxidation, as a function of NADPH, would, under these circumstances, reflect primarily the saturation of the microsomal NADPH oxidase. This view is supported by the relative constancy of chain length (line 2). In that case we should expect a linear Lineweaver-Burk plot and could use such a plot to arrive at Km for NADPH and for V,. Figure 4 presents the affirmation of these expectations. Km for NADPH was calculated to be 1.5 PM, while V, is 27.5 nmol NMNH oxidized/min/ml. This V,, expressed in terms of the rate of NADPHtriggered oxidation of NMNH, compares well with the rate of vanadate-stimulated oxidation of NADPH by 50 pg/ml of microsomes (as protein), which was seen at 320 PM NADPH (Fig. 2, line 2). This clearly indicates that O;, produced by the saturable NADPH oxidase of microsomes, was then equally effective in causing the vanadate-stimulated oxidation of NADPH or of NMNH. When the data for 0, production by microsomes, as a function of [NADPH] (line

0; PRODUCTION

BY BIOMEMBRANES

143

rate of NMNH oxidation divided by the average chain length, was constant as 5 0.16 t [NMNH] was increased (line 3). This is precisely the result to be expected from the proposal that OF, produced by an enzymatic NADPH oxidase, served to cause the vanadate-stimulated oxidation of NMNH. Eflects of pH. Decreasing pH, in the I I I I I I 0' range 8.0-5.0, has been reported to in2.0 2.5 0.5 1.0 1.5 0 crease the vanadate-stimulated oxidation (NADPH , HAM)-’ of NAD(P)H by biological membranes (1); FIG. 4. Saturation of the triggering of NMNH oxileading to the conclusion that this was an dation by NADPH. The data from line 1 of Fig. 3 are effect of pH on a membrane-associated here plotted as a function of [NADPH] on reciprocal NAD(P)H oxidase. Our results suggested coordinates. K, NADPH, calculated as slope/interthe possibility that decreasing pH incept, = 1.5 /LM and V,, calculated as l/intercept = 27.6 creased the average length of the O;-ininmol NMNH/min/ml. tiated, vanadate-stimulated, free radical chain oxidation of NAD(P)H or NMNH; without significantly changing the enzy3, Fig. 3), were graphed on reciprocal co- matic rate of production of 0; by memordinates a straight line was obtained from brane-associated NAD(P)H oxidases. This which K, NADPH = 2.9 pM and V, = 2.6 was tested by observing the effect of pH on nmol O;/min/ml. This Km is in reasonable the NADPH-triggered, vanadate-stimuagreement with the value obtained from lated oxidation of NMNH by microsomes. the effect of NADPH in triggering NMNH Line 2 in Fig. 6 demonstrates that decreasoxidation by microsomes in the presence ing the pH from 8.0 to 7.1 did increase the of vanadate (Fig. 4). The V, for 0, pro- oxidation of NMNH. However, line 1 shows duction is approximately $, that for that average chain length also increased NADPH or NMNH oxidation, reflecting the sharply with decreasing pH. As a result it average length of the free radical chain re- could be calculated that the rate of 0; proactions initiated by 0; in the presence of duction, by the microsomal NADPH oxivanadate. It is interesting to note that Vijaya et al. (2) observed a biphasic response 0 15to [NADH] in the vanadate-stimulated oxidation of NADH by erythrocyte mem-6 \ branes and reported Km NADH = 3 and c E 133 PM. It now seems clear that the lower \ Km reflects saturation of an Oi-producing '2 NADH oxidase; whereas the larger Km re-2E sulted from increasing chain length with I increasing [NADH], to a limit due to more 0 frequent chain termination at very high 0 0.2 0.3 0.1 [NADH]. NMNH (mM) Studies of the NADPH-triggered oxiFIG. 5. Effect of varying the concentration of NMNH. dation of NMNH by microsomes in the Reaction mixtures contained 0.4 mM vanadate, 50 pg presence of vanadate were performed as a microsomes (as protein), 3.2 pM NADPH and the infunction of [NMNH]. The results are shown dicated concentrations of NMNH in 100 mM sodium in Fig. 5. NMNH oxidation increased with phosphate at pH 7.6 and at 37°C. Line 1, nmoles increasing [NMNH] (line 1). The average NMNH oxidized/min/ml; line 2, average chain length chain length, calculated as f NMNH oxi- (calculated as i/z NMNH oxidized/NADPH added); line dized per NADPH added, also increased 3, rate of 0; production in nmol/min/ml (calculated with increasing [NMNH] (line 2). However, from the rate of oxidation of NMNH divided by the the rate of 0; production, calculated as the chain length). i ZZ

0.20

144

LIOCHEV

AND

FRIDOVICH

production by the membrane-associated NADPH oxidase. DISCUSSION

Having demonstrated that enzymatic, chemical, or photochemical sources of 0; cause a vanadate-stimulated oxidation of NAD(P)H by a free radical chain mechanism (‘7-g), we wished to determine whether 0; production, by membrane-associated NAD(P)H oxidases, could fully PH account for the frequently reported (l-6, FIG. 6. Effect of pH on NMNH oxidation. Reaction 14) vanadate stimulation of NAD(P)H oxmixtures contained 0.4 InM vanadate, 50 pg microidation by biomembranes. Several lines of somes (as protein), 0.32 mM NMNH and 1.6 pM evidence support an affirmative answer to NADPH in 100 sodium phosphate at 3’7°C and at the this question. (a) Paraquat, which augindicated pH. Line 2, nmol NMNH oxidized/min/ml; ments 0; production, increased the vanaline 1, average chain length; line 3, nmol O;/min/ml. date-stimulated oxidation of NADPH by Chain length and rate of 0; production were calcumicrosomes, and by both inner and outer lated as specified in the legend to Fig. 5. mitochondrial membranes and of NADH only by inner mitochondrial membrane. (b) dase, was essentially invariant with pH in SOD, which intercepts O;, inhibited this the range 8.0-7.1 (line 3). vanadate-stimulated NAD(P)H oxidation. Eflect of varying the concentration of (c) Microsomal oxidation of NADPH, in the vanadate. The strategy of using a small absence of vanadate, was a linear function amount of NADPH to trigger the oxidation of microsome concentration, as expected of a larger amount of NMNH, by micro- for an enzymatic process; while in the somes, could also be used to examine the presence of vanadate it was a saturating effects of vanadate on chain length and on function of microsome concentration, as the rate of 0; production. Reaction mix- expected for a free radical chain process, tures contained 320 PM NMNH, 6.4 PM where chain termination becomes more NADPH, 50 pg microsomes (as protein), prominent with increasing concentration 100 mM sodium phosphate, and either 0.1 of free radical intermediates. (d) NMNH, or 0.4 mM vanadate, at pH 7.6, and at 37°C. which was not directly oxidized by memWe noted that increasing vanadate from brane-associated oxidases, was not subject 0.1 to 0.4 KiM increased the rate of NMNH to vanadate-stimulated oxidation by mioxidation and the average chain length by crosomes. However, small amounts of 3.5-4.0 fold with a much smaller effect on NADPH, which can provide O;, triggered the rate of 0; production (data not shown). the vanadate-stimulated oxidation of much Since vanadate must be competing with the larger amounts of NMNH by microsomes. spontaneous dismutation reaction for the It follows that there is presently no reason 02 produced, some increase in apparent to consider the reality of vanadate-stimu02 production with increasing vanadate lated NAD(P)H oxidases. Vanadate &mconcentration should be expected. The 15% ulation of NAD(P)H oxidation by biomemincrease which we calculated upon in- branes is rather to be seen as a vanadatecreasing vanadate from 0.1 to 0.5 mM can stimulated NAD(P)H oxidation by O,, easily be accounted for in terms of smaller which is produced by vanadate-indepenlosses to the spontaneous dismutation re- dent NAD(P)H oxidases. action at the higher concentration of vanWe have previously proposed a nonadate. It thus appears that vanadate in- branching free radical chain mechanism creases average chain length rather than for the vanadate-stimulated oxidation of having any effect on the rate of 0; NAD(P)H or NMNH by 0; (7-9). Assuming

0, PRODUCTION

that each NAD(P)H oxidized by a membrane-associated oxidase yields two 0,) we can estimate average chain length, from the NADPH-triggered oxidation of NMNH, as $ NMNH oxidized per NAD(P)H added. Under the conditions used we have obtained chain lengths in excess of 30. HOWever, the yield of 0, per NAD(P)H oxidized may be less than 2.0; some of the 0, produced by membrane associated NAD(P)H oxidases may be inaccessible to vanadate; and some of the 02 released from the membrane, or produced during the propagating steps of the chain reaction, will be lost to the spontaneous dismutation reaction. It follows that our estimated chain lengths represent minimum values. These considerations, together with the sharp increase in chain length with decreasing pH, indicate that chain lengths far in excess of 30 can easily be achieved and that the vanadate-stimulated oxidation of NAD(P)H or NMNH can therefore serve as an amplifying detector of 0; production by membranes or by soluble enzymes. It could also provide an amplifying detector for any oxidant which converted NAD(P)H to NAD(P)‘, because the NAD(P)’ would, by reacting with dioxygen, yield 0;. In addition it suggests that the toxicity of vanadate may be due, at least in part, to its ability to stimulate this Oi-dependent chain oxidation of NAD(P)H. Exploration of this possibility is underway.

BY BIOMEMBRANES

145 REFERENCES

1. RAMASARMA, T., MACKELLAR, W. C., AND CRANE, F. L. (1981) Biochim Biophys. Ada 646,88-98. 2. VIJAYA, S., CRANE, F. L., AND RAMASARMA, T. (1984) Mol. Cell. Biochenz. 62, 175-185. 3. BRISKIN, D. P., THORNLEY, W. R., AND POOLE, R. J. (1985) Arch. Bicchem Biophys. 236,228237. 4. ERDMAN, E., KRAWIETZ, W., PHILIPP, G., HACKBARZTH, I., SCHMITZ, W., SCHOLZ, H., AND CRANE, F. L. (1979) Nature (London) 282, 335336. 5. CRANE, F. L., MACKELLAR, W. C., MOORE, 0. J., RAMASARMA, T., GOLDENBERG, H., GREBING, C., AND Low, H. (1980) Biodmn. Biophys. Res Commun. 93,746-754. 6. MENON, A. S., RAU, U., RAMASARMA, T., AND CRANE, F. L. (1980) FEBSLett. 114,139-141. 7. DARR, D., AND FRIDOVICH, I. (1954) Arch Biochm. Biophys. 232,562-565. 8. DARR, D., AND FRIDOVICH, I. (1985) Arch. Biochem Biophys. 243,220-227. 9. LIOCHEV, S., AND FRIDOVICH, I. (1986) J. Free Rad. Biol. Med 1,287-292. 10. MCCORD, J. M., AND FRIDOVICH, I. (1969) J. Biol. Chem 244,6049-6055. 11. LJKJTAKOVA,S. G., RUSSANOV, E. M., AND LIOCHEV, S. (1984) Arch. Biochem Biophys. 235,636-643. 12. LOWRY, 0. H., ROSEBROUGH, N. J., FARR, H. L., AND RANDALL, R. J. (1951) J. Biol Chem. 193, 265-275. 13. HORECKER, B. L., AND KORNBERG, A. (1948) J. Biol Chem. 175,385-390. 14. BYCZKOWSKI, J. Z., ZYCKLINSKI, L., TLUCZKIEWICZ, J. (1979) Int. J. Bimhem 10,1007-1011. 15. WILKINSON, F. (1980) Chemical Kinetics and Reaction Mechanisms, Chap. 7, Van NostrandReinhold, New York.