Copper, Zinc Superoxide Dismutase and H2O2

Copper,zinc superoxide dismutase (Cu,Zn-SOD) catalyzes the HCO 3 − -dependent oxidation of diverse substrates. The mechanism of these oxidations involves the generation of a strong oxidant, derived from H2O2, at the active site copper. This bound oxidant then oxidizes HCO 3 − to a strong and diffusible oxidant, presumably the carbonate anion radical that leaves the active site and then oxidizes the diverse substrates. Cu,Zn-SOD is also subject to inactivation by H2O2. It is now demonstrated that the rates of HCO 3 − -dependent oxidations of NADPH and urate exceed the rate of inactivation of the enzyme by ∼100-fold. Cu,Zn-SOD is also seen to catalyze a HCO 3 − -dependent consumption of the H2O2 and that HCO 3 − does not protect Cu,Zn-SOD against inactivation by H2O2. A scheme of reactions is offered in explanation of these observations.


dependent consumption of the H 2 O 2 and that HCO 3 ؊ does not protect Cu,Zn-SOD against inactivation by H 2 O 2 . A scheme of reactions is offered in explanation of these observations.
All superoxide dismutases (SODs) 1 catalyze the conversion of O 2 . into H 2 O 2 plus O 2 and do so with great efficiency. The Cu,Zn-SOD is also able to catalyze other processes, albeit at lower rates, including: the HCO 3 Ϫ -dependent oxidation of diverse substrates by H 2 O 2 (1)(2)(3)(4)(5); the reduction of O 2 . by substrates other than O 2 . (6), referred to as superoxide reductase activity; and the elimination of NO Ϫ both aerobically and anaerobically (7,8). The inactivation of Cu,Zn-SOD by H 2 O 2 and its ability to act as a nonspecific peroxidase were described (1,9) long before the dependence of the latter activity on HCO 3 Ϫ was appreciated (2)(3)(4)(5).
The relevant literature has been marred by controversy. Thus, the production of HO ⅐ as the oxidant responsible for both inactivation by H 2 O 2 and for the peroxidase activity has been claimed (5) and denied (3,4,10). The bicarbonate effect has been attributed to facilitation of the binding of H 2 O 2 (2), and that explanation has been called into question (3,4). Bicarbonate has been reported to protect (3,11) and not to protect (4) against the inactivation of the enzyme by H 2 O 2 . Finally, the rate constant for reaction leading to the formation of the copper-bound oxidant was estimated to be only 0.1 M Ϫ1 s Ϫ1 at pH 7.4 (12), which is not fast enough to account for the rapid HCO 3 Ϫ -dependent oxidations (2,4). Clearly the mechanism of the interaction of H 2 O 2 with the Cu,Zn-SOD and the effects of HCO 3 Ϫ require further clarifica-tion. We now present data and calculations derived therefrom, which shed some light on these reactions. The following scheme of reactions reflect current knowledge. Elaborations, to be presented under "Discussion" will lead to a scheme more in tune with observed reactions and their rates.
In this scheme SH 2 denotes a substrate such as NADPH, and His denotes a histidine residue in the ligand field of the copper. This scheme is very similar to the one that we arrived at previously (4). In Reaction a the enzyme is univalently reduced by H 2 O 2 , while in Reaction b the bound oxidant is generated. We consider the oxidant to be bound rather than diffusible because HO ⅐ scavengers, such as ethanol, do not protect (1) and because the oxidation of large substrates depends upon the mediation of bicarbonate (2)(3)(4)(5). The bound oxidant can inactivate the enzyme by oxidation of a histidine residue (Reaction d) or, it can oxidize HCO 3 Ϫ to the powerful oxidant CO 3 . , which then leaves the active site and oxidizes substrates in free solution (Reaction f) or alternately stays in the active site to oxidize a histidine residue (Reaction g). Reaction g was proposed because HCO 3 Ϫ competes with histidine (Reactions d and e) and should thus protect, which it does not (4). Urate does protect the enzyme, hence Reaction h. * This work was supported by grants from the Amyotrophic Lateral Sclerosis Association and Grant R01 DK 59868 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

MATERIALS AND METHODS
NADPH, horseradish peroxidase, and dianisidine were from Sigma. The dianisidine was twice recrystallized from ethanol. Uric acid and NaHCO 3 were from Fisher, H 2 O 2 was from Mallinkrodt, and Cu,Zn-SOD was from Chemie Grunenthal. All experiments were performed in 100 mM sodium phosphate at pH 7.4 and at 22-23°C. When NaHCO 3 Ϫ was added the buffer was readjusted to pH 7.4 with H 2 SO 4 as needed. Initial rates of oxidation of NADPH were followed at 340 nm and of urate at 295 nm. The usual concentrations of SOD and H 2 O 2 were 0.2 mg/ml and 10 mM, respectively. When [NADPH] or [urate] were varied the HCO 3 Ϫ was held constant at 50 mM in the case of NADH and at 150 mM in the case of urate. When [HCO 3 Ϫ ] was varied, urate and NADPH were held constant at 60 and 100 M, respectively. When urate oxidation was examined in the absence of HCO 3 Ϫ , the [SOD] was raised to 0.8 mg/ml. The results were then corrected to the usual condition by dividing by four.
Inactivation of SOD by H 2 O 2 was followed by removing samples at intervals and assaying remaining SOD activity by the xanthine oxidase/ cytochrome c method (13). Consumption of H 2 O 2 by Cu,Zn-SOD Ϯ HCO 3 Ϫ was followed by removing samples at intervals, assaying for residual [H 2 O 2 ] by 25-fold dilution into buffered (pH 7.4) solutions of 0.3 mM dianisidine plus 40 g/ml horseradish peroxidase, and measuring absorbance at 460 nm. The reduction of 5.0 mg/ml Cu(II),Zn SOD by 2 mM H 2 O 2 was followed at 680 nm.

HCO 3
Ϫ -dependent Oxidation of NADPH and Urate-Were the processes presented by Reactions a-h correct we would expect the bound oxidant produced by Reaction b to inactivate the enzyme by Reaction d or, in the presence of HCO 3 Ϫ , to oxidize extraneous substrates through the mediation of CO 3 . .
The rate of inactivation should in any case equal the maximal rate of oxidation of substrates such as NADPH because both depend on the rate of production of the bound oxidant. We will now show that the rate of inactivation is ϳ100-fold slower than the maximal rate of oxidation of either NADPH or urate. Figs. 1 and 2 present the rates of oxidation of urate and of NADPH as a function of HCO 3 Ϫ . Because the effect of HCO 3 Ϫ approaches a limit we can use reciprocal coordinates to obtain the maximal rates of oxidation at infinite HCO 3 Ϫ , which will be referred to as V max , and the concentration of HCO 3 Ϫ that caused half that maximal rate, which will be referred to as K m . Fig. 3 shows the data for NADPH oxidation plotted on reciprocal coordinates, and Table I, lines 1 and 2 present K m and V max for the HCO 3 Ϫ effect on both urate and NADPH oxidation obtained at the indicated concentrations of H 2 O 2 and Cu,Zn-SOD. Lines 3 and 4 present K m and V max values for NADPH and urate as the variable substrates at fixed [HCO 3 Ϫ ].
The K m for HCO 3 Ϫ was approximately equal for both NADPH and urate oxidation. This is expected because both oxidations depend upon the CO 3 . leaving the active site. Were NADPH and urate comparable processes we would expect that V max should also be the same for both oxidations. However, V max for NADPH oxidation was 5ϫ greater than V max for urate oxidation. This difference may be partially artifactual Thus oxidation of NADPH to NADP ⅐ by CO 3 . would be followed by the auto-oxidation of NADP ⅐ to NADP ϩ ϩ O 2 . with loss of 340 nm of absorbance. In contrast the oxidation of urate can involve several intermediates, such as allantoin and allantoic acid (14), and some of those may absorb at 295 nm, as does urate, and may also consume CO 3 . . It is also possible that urate, being an anion, may be less reactive with CO 3 . than is the uncharged dihydropyridine moiety of NADPH. When [HCO 3 Ϫ ] was constant at 50 mM and NADPH was varied the K m for NADPH was less than 17 M. Hence 100 M NADPH will scavenge essentially all the CO 3 . leaving the active site, and the V max for NADPH oxidation, 20 M/min, must approximate the rate of production of CO 3 . . K m for urate was also low (ϳ7.5 M). The V max obtained by extrapolating to infinite HCO 3 Ϫ and at 100 M NADPH was 50 M/min, which must then approximate the rate of production of the bound oxidant (Reaction b).
Urate oxidation in the absence of HCO 3 Ϫ was so slow that it was advisable to raise [SOD] to achieve a measurable rate. This

Effects of HCO 3
Ϫ on H 2 O 2 /Cu,Zn-SOD Interactions direct oxidation of urate should reflect the rate of Reaction h. The V max urate of 0.7 M/min, which would be 0.18 M/min at 0.2 mg/ml SOD, is more than 50ϫ smaller than the rate in the presence of saturating HCO 3 Ϫ . Urate is thus a poorer substrate for the bound oxidant than is HCO 3 Ϫ , yet urate protects while HCO 3 Ϫ appears not to do so (4). How a molecule as large as urate accesses the active site and protects against inactivation by the bound oxidant (Reaction d) remains a puzzle.
Does HCO 3 Ϫ Protect Cu,Zn-SOD against Inactivation by H 2 O 2 ?- Fig. 4 presents a semilog plot of residual SOD activity during exposure to excess H 2 O 2 . Rather than protect, as has been reported by others (3,11), HCO 3 Ϫ modestly increased the rate of inactivation. We previously reported (4) that 10 mM HCO 3 Ϫ did not protect in accordance with the results in Fig. 4 where higher concentrations were needed to show a measurable acceleration. The rate of inactivation from Fig. 4 was ϳ100fold slower than the V max for the HCO 3 Ϫ -dependent oxidation of NADPH. Thus the V max for NADPH oxidation from Table I was  Ϫ greatly increases the consumption of H 2 O 2 by Cu,Zn-SOD. The effect of HCO 3 Ϫ was saturable and the V max for this H 2 O 2 consumption was similar to that for NADPH oxidation. If the H 2 O 2 reacted rapidly with the bound oxidant its rate of consumption would be high in the absence of HCO 3 Ϫ , which is not the case. But we know that H 2 O 2 rapidly reduces the Cu(II) enzyme (Reaction a). Hence if HCO 3 Ϫ has the effect of converting the bound oxidant back to the Cu(II) enzyme by Reaction e that alone would explain the augmentation of H 2 O 2 consumption by HCO 3 Ϫ ; this will be made clear later in the text. Is it also plausible that CO 3 . Ϫ as shown in Fig. 6 the rate of reduction, followed in terms of the bleaching of absorbance at 680 nm, was very rapid. Thus half (78 M) of the SOD was bleached within 10 -15 s. Because the rate was declining rapidly within this time inter-val we estimated that the initial rate of bleaching was 78 M/5 s ϭ 16 M/s. To connect this rate to our standard condition of 6.2 M SOD and 10 mM H 2 O 2 we assume first order dependence and get (16 M/sec) (6.2/156) (10/2) ϭ 3.2 M/s. Thus the rate of reduction of the Cu(II) was faster by a factor of 4-fold than the ϳ0.8 M/sec rates of HCO 3 Ϫ -dependent oxidation of NADPH or of H 2 O 2 consumption. The maximal rates of the reactions stud-   Table II. It should be pointed out that the reactive form of H 2 O 2 in these reactions is HO 2 Ϫ (12) and the pK a of H 2 O 2 is ϳ11.6 (15). That being the case the rate constants that involve H 2 O 2 reactions in Table II are apparent rate constants and they are ϳ10,000-fold lower than the rate constants that would be calculated on the basis of [HO 2 Ϫ ] rather than upon [H 2 O 2 ].

Because HCO 3
Ϫ does not much change the rate of inactivation of Cu,Zn-SOD by H 2 O 2 and did not affect the rate of reduction of this enzyme by H 2 O 2 (4), it cannot be the case that HCO 3 Ϫ increases the binding of H 2 O 2 as has been suggested (2). It follows that Reaction b as written in the forward direction is not affected by HCO 3 Ϫ . We have seen that the rate of inactivation of the enzyme by H 2 O 2 (Reaction d) is only ϳ1% as great as the maximal rates of HCO 3 Ϫ -dependent oxidations and therefore of Reactions b and e. This discrepancy can be explained by introducing the reverse of Reaction b (called Reaction b 1 ) and further assuming that the equilibrium lies to the left such that at equilibrium [E- In that case adding HCO 3 Ϫ would, by decreasing E-Cu(II)OH below its equilibrium concentration, pull Reaction b 1 to the right and allow the great increase in H 2 O 2 -dependent oxidations. This can be restated and hopefully clarified. Thus at the steady state, the production of Cu(II)OH, i.e. V b ϩ , is equal to its rates of consumption, i.e. Ϫ ] approaches ϱ V bϪ and V d will approach zero and the rate of production of CO 3 . will approach the rate of Reaction bϩ. Because we have seen that V bϩ is 100-fold greater than V d then V bϪ must be ϳ0.99 V bϩ . Hence in the absence of HCO 3 Ϫ 99% of the bound oxidant disappears by Reaction bϪ, leaving only 1% to inactivate the enzyme. By lowering [E-Cu (II)OH], HCO 3 Ϫ might also be expected to pre-vent the inactivation of the enzyme by Reaction d. But rather than protecting, HCO 3 Ϫ slightly (ϳ2-fold) increases the inactivation of the enzyme by H 2 O 2 . This is explained by assuming that a small fraction of the CO 3 . produced does not leave the active site before oxidizing an essential histidine (Reaction g Ϫ appeared to protect because it caused depletion of the H 2 O 2 long before the residual activity was measured. The large difference between the rates of inactivation of the enzyme and the V max of HCO 3 Ϫ -dependent oxidations of NADPH and of urate can also be explained in another way. Thus suppose that ϳ1% of the Cu(II)OH produced by Reaction b 1 goes on to dissociate OH Ϫ and thus forms Cu(III) as in Reaction j, which should be ϳ100-fold slower than V bϩ .
We could further suppose that Cu(III) was able to attack a liganding histidine while Cu(II)OH could react with HCO 3 Ϫ but not with the histidine. The Cu(III) could also be responsible for the slow HCO 3 Ϫ -independent oxidations of urate, formate, and H 2 O 2 . This could explain the slowness of those reactions as well as the protective effects of urate and formate (9,12).
Reaction bϪ is really the reduction of the Cu(II)OH by OH Ϫ . The postulation of that reaction can be avoided in still another way. Thus suppose the following reactions, in which V l Ͻ Ͻ V k .
In this case the rapid HCO 3 Ϫ -dependent oxidations follow the oxidation of HCO 3 Ϫ by E-Cu(I)H 2 O 2 , while the slow HCO 3 Ϫindependent inactivation and oxidations depend on E-Cu(II)-OH produced by the slow Reaction l. In this case the urate protects by reacting with E-Cu(II)OH in competition with the histidine residue, but is not in competition with HCO 3 Ϫ for E-Cu(I)H 2

Effects of HCO 3
Ϫ on H 2 O 2 /Cu,Zn-SOD Interactions tial rates by correcting for the concurrent inactivation of the SOD during the period of observation and for the consumption of H 2 O 2 during that period. None of these corrections introduced more than a factor of 2.0 -2.5 into the calculated rate constant. The rate of production of the bound oxidant Cu(II)OH (V bϩ ) must, in the steady state, equal its rate of consumption by the reverse reaction (V bϪ ) and by the oxidation of an active site histidine (V d ). Hence, V bϩ ϭ V bϪ ϩ V d and then V bϪ ϭ V bϩ Ϫ V d . If Reaction k actually occurs and not Reaction b 1 , V l ϭ V d and then the above equation should be replaced by V kϪ ϭ V kϩ Ϫ V l . As previously stated the rate of production of the bound oxidant can be equated to the maximal rate of HCO 3 Ϫ -dependent oxidation of NADPH or consumption of H 2 O 2 . In calculating V bϩ and k bϩ we take the [Cu(I),Zn-SOD] from the total concentration of SOD even though some fraction of the SOD must remain in the Cu(II) form even in the presence of H 2 O 2 . Hence our estimate of k bϩ is a minimal figure. Nevertheless we find k bϩ to be ϳ100fold higher than the value reported by Cabelli et al. (12). In their system formate was oxidized by the bound oxidant and the resultant CO 2 . reduced nitroblue tetrazolium. This is quite analogous to our use of HCO 3 Ϫ followed by oxidation of NADPH by CO 3 . . A possible explanation for the very low rate of Reaction bϩ found by Cabelli et al. (12) is that CO 2 . did not predominantly escape the active site but rather reduced the E-Cu(II) to E-Cu(I). This reaction is known to occur (16). Our high estimate for the rate of Reaction bϩ accounts for the rapid HCO 3 Ϫdependent oxidations reported by Hodgson and Fridovich (1), Sankarapandi and Zweier (2), and in the present work. It has become obvious from the results and discussion presented above that in the absence of HCO 3 Ϫ no significant oxidant leaves the active site of Cu,Zn-SOD in the presence of H 2 O 2 . Hence the assertion that this enzyme catalyzes the production of HO ⅐ from H 2 O 2 can be laid to rest.