THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1986 hy The American Society of Biological Chemists, Inc.

Vol. 261, No. 3. Issue of January 25, pp. 1193-1200,1986 Printed in U. S. A .

Mechanism of the Inhibition of Catalase by Ascorbate ROLES OF ACTIVE OXYGEN SPECIES, COPPER AND SEMIDEHYDROASCORBATE*

(Received for publication, August 15, 1985)

Allan J. DavisonS, Anthony J. Kettle, and Dolores J. Fatur From the Bioenergetics Research Laboratory. School of Kinesiology, Simon Fraser University, Burnaby, British Columbia, Canada V5A IS6

Ascorbate reversibly inhibits catalase, and this in- hibition is enhanced and rendered irreversible by the prior addition of copper(I1)-bishistidine. In the absence of copper, the inhibition was prevented and reversed by ethanol, but not by superoxide dismutase, benzoate, mannitol, thiourea, desferrioxamine, or DETAPAC. In the presence of the copper complex mannitol, benzoate, and superoxide dismutase still had no effect, but thi- ourea, desferrioxamine, DETAPAC, or additional his- tidine decreased the extent of inactivation to that seen in the absence of copper. In the presence of copper, ethanol protected at [ascorbate] < 1 mM, but was inef- fective at [ascorbate] > 2 mM, even in the absence of oxygen. Although in the absence of copper, complete removal of oxygen provided full protection against inactivation by ascorbate, this protection was not seen if the catalase was briefly preincubated with H202 prior to flushing with nitrogen, or if copper was pres- ent. In fact, if copper was present, inactivation was enhanced by the removal of oxygen. Increasing the concentration of oxygen from ambient to 100% slowed the inactivation, whether or not copper was present. It is concluded that the initial reversible inactivation involves reaction with H202 to form compound I, fol- lowed by one electron reduction of compound I to com- pound 11. In the presence of added copper, the initial (reversible) inactivation allows H202 to accumulate sufficiently to permit irreversible inactivation. Since in the presence of copper oxygen is not required, and neither the reversible nor the irreversible inactivation was prevented by conventional scavengers of active forms of oxygen, the inactivation is likely mediated by semidehydroascorbate, and/or it may involve site-spe- cific generation of the damaging intermediates.

The rapid inhibition of catalase by the ascorbate/oxygen couple and its potentiation by copper ions have attracted the attention of several workers. However, controversy rather than clarification has resulted from their efforts (1-5). Ra- tional discussion of the conflicting mechanisms of inhibition requires an understanding of those aspects of the reaction mechanism of catalase outlined in Fig. 1. The resting Fe(II1)- enzyme binds HzOz at i ts active site and undergoes a two- electron oxidation liberating HzO. At low concentrations of

*This work was supported by Grant A7568 from the Natural Sciences and Engineering Research Council of Canada and Simon Fraser University Programmes of Distinction. The costs of publica- tion of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertise- ment” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed.

H202, compound I can act peroxidatically, catalyzing two- electron oxidations of compounds such as ethanol and for- mate. Alternatively, compound I can regenerate ferricatalase by two successive one-electron reductions. The first of these univalent reductions forms inactive compound I1 which is incapable of decomposing H202. Compounds such as ethanol, methanol, NADH, and NADPH, however can transfer a single electron to it and thereby restore active catalase. Another inactive form of catalase is the enzyme-peroxide complex, compound 111, formed through the divalent oxidation of com- pound I1 by Hz02 (6).

Two alternate mechanisms have been proposed for the inactivation of catalase by ascorbate: (i) the formation of compound I1 by saturating levels of Hz02 and (ii) free radical attack on the enzyme. Originally, Chance (1) demonstrated the spectrum of compound I1 in ascorbate-treated catalase. He attributed the formation of compound I1 to the presence of H202 from ascorbate autoxidation, because the notatin system, which produces a continuous supply of HzOz, also caused the formation of compound 11. However, the amount of compound I1 formed in the presence of ascorbate consid- erably exceeded that formed in the presence of the notatin system at pH 7.0. This led Chance to speculate that the equilibrium between compounds I and I1 may be affected by the ascorbate molecule itself. In a similar investigation, Keilin and Hartree ( 2 ) showed that formation of compound 11, when catalase was incubated with ascorbate, required oxygen and was greatly accelerated by copper salts.

On the other hand, Orr (3-5) showed that, in the presence or absence of added copper, ascorbate caused degradative changes in catalase, actually cleaving the molecule to smaller peptide fragments. He could not show formation of compound I1 when catalase was incubated with ascorbate (3), perhaps because he used pH 7.0 and formation of compounds I and I1 is slow at pH 7 in comparison to pH 5.0 ( 7 ) . Orr concluded that ascorbate inactivated catalase through peroxy or hydroxl radial attack on the protein. Against Orr’s hypothesis is the observation that inactivation of catalase in hemosylate was reversed by ethanol or NADPH, a finding which suggests that the inactivated form of catalase was compound I1 (8). Ascor- bate itself has been excluded as the inhibiting species because inactivation requires oxygen (2) and is greatly accelerated by metal ions ( 2 , 4). Intermediates in the oxidation of ascorbate to dehydroascorbate and hydrogen peroxide (9) include sem- idehydroascorbate (10) and superoxide (11). The H202 can in turn be reduced by the other intermediates in Fenton type reactions to hydroxyl radicals (12, 13). Any one or a combi- nation of these radical species could inactivate catalase.

Although semidehydroascorbate is relatively unreactive and decays mainly by disproportionation (14), it reacts faster with cytochrome c than does ascorbate (15). Furthermore, the

1193

1194 Inhibition of Catalase by Ascorbate

FIG. 1. Possible redox transformations involved in reac- tions of catalase. Formal oxidation states of iron are shown in parentheses. * denotes an inactive form of catalase. In the normal action of catalase, the enzyme-peroxide complex (compound I) binds and oxidizes a further molecule of H202 regenerating ferricatalase and liberating Hz0 and 0 2 .

semiquinone radicals of 1,4-naphthoquinone-2-sulfonic acid and 6-hydroxydopamine have been implicated in the inacti- vation of catalase (16).

Superoxide, the other transient intermediate of ascorbate autoxidation, also inhibits catalase, both by a rapid reaction that is prevented and reversed by superoxide dismutase and by a slow reaction that is prevented by superoxide dismutase or ethanol and reversed by ethanol only (17, 18). The rapid inhibition was reportedly due to the formation of unstable compound 111 while the slow inhibition was attributed to production of compound I1 through the reduction of com- pound I by superoxide.

Several alternative explanations have been given for the role of copper in accelerating the cytotoxicity of ascorbate. In the mechanism proposed by Chevion and co-workers (19, 20), ascorbate plays the dual roles of reducing Cu2+ to CUI+ and of providing H202 through its autoxidation. In turn, CU'+ bound to macromolecules reacts with H202 in a Fenton reaction to produce hydroxyl radicals. However, it is unlikely that Fen- ton-derived hydroxyl radicals could attack active catalase because, in the presence of catalase, Hz02 levels should be so low as to preclude formation of substantial quantities of hydroxyl radicals via the Fenton reaction (21). In an alter- native mechanism, Golan-Goldhirsh and Whitaker (22) have proposed that Cu2+ binds to the histidine residues of proteins and then chelates O2 and semidehydroascorbate. This com- plex is then thought to promote specific degradation of histi- dine, thereby damaging the protein and inactivating the en- zyme.

Keilin and Hartree (2) were the first to investigate the inactivation of catalase by ascorbate in the presence of Cu2+ and their absorption spectra revealed formation of compounds I1 and 111. In contrast to the above-mentioned conclusions (3, 22), they proposed that Cu-induced inactivation of catalase results from the reaction between active catalase and the increased levels of H202 resulting from the presence of copper.

Because of the above-mentioned uncertainties and contra- dictions, the current study was undertaken to investigate the mechanism of inactivation of catalase by ascorbate and/or ascorbate plus Cu2+(His)21 more fully. By applying recent knowledge of the actions of scavengers of activated oxygen and transition metals, we proposed to clarify the discrepancies between previous investigations and to determine the species responsible for inactivation.

EXPERIMENTAL PROCEDURES

Materials-L-Ascorbic acid, Analar grade, was from BDH Chemi- cals. Catalase (65,000 units/mg) was obtained from Boehringer

The abbreviations used are: Cu2+(His),, copper(I1)-bishistidine; DETAPAC, diethylenetriaminepentaacetic acid; DTPA, diethylene- triamine pentaacetic acid.

Mannheim because this commercial source has been shown to be free of superoxide dismutase (23). Crystalline, bovine blood superoxide dismutase (2,800 units/mg), and ascorbate oxidase (350 units/mg) were obtained from Sigma. Desferrioxamine was a gift from the CIBA Pharmaceutical Co. All other reagents were of analytical grade.

Procedures-Deionized, distilled water was used for preparation of stock buffer solutions. Ion contamination of the water was negligible (>lo megaohm ~ m - ~ ) . Ascorbic acid stock was always freshly prepared in phosphate buffer (50 mM, pH 7.0) and neutralized with 0.1 M NaOH. A stock solution of Cu2'(His)2 was prepared by dissolving CuSOl and L(+)-histidine monohydrochloride, in the molar ratio 12, in phosphate buffer. In all cases, reagents, except catalase and Cu2+(His)2, were added to the phosphate buffer to give the desired concentrations in a final volume of 2.025 ml and were preincubated at 37 "C in sealed VirTis vials for 10 min. When used, Cu2+(His)2 was added immediately before the reaction was initiated by the addition of 0.1 ml of catalase. The incubate was thoroughly mixed, and aliquots were withdrawn and assayed for catalase activity at predetermined intervals for 1 h. Atmospheres of nitrogen (Linde high purity grade < 5 ppm of oxygen) or oxygen (Linde high purity grade) were applied in rubber-capped vials, using a VirTis gas manifold, by five cycles of repeated evacuation to boiling and reintroduction of the desired gas to a slight positive pressure. Rubber caps were carefully degassed prior to use. All reagents added to incubates under nitrogen atmos- pheres were prepared under atmospheres of nitrogen.

Catahe Assay-Catalase activity was measured using an LKB spectrophotometer coupled to a microcomputer. The assay procedure was as outlined by Aebi (24) with the following modifications. An aliquot was withdrawn from the reaction mixture and rapidly mixed into a cuvette containing enough Hz02 in phosphate buffer (50 mM, pH 7.0) to give an initial absorbance between 0.550 and 0.520 at 240 nm. The rate of change in absorbance between 0.450 and 0.400 was taken as a measure of catalase activity. Aliquot volumes between 8 and 25 p1 were chosen to give a rate of change in absorbance of between 0.02 and 0.10 absorbance units/l5 s.

Ascorbate Analysis-The method used for ascorbate analysis is a modification of that given by Chevion and Navok (25). A heating step was introduced because the reaction was not instantaneous, taking several hours to reach maximum absorbance. This finding was similar to that of Schilt and Di Tusa (26) who used PPTS as a ferroin type chromagen to assay ascorbate. This analysis involves the reduction by ascorbate of ferriphenanthroline to ferrophenanthroline and then measuring the change in absorbance at 515 nm due to the production of the ferrous complex. A stock solution of ferriphenanthroline (1.8 mM) was adjusted to pH 5 with HCl. Before assaying ascorbate, the stock was diluted 1:lO in imidazole buffer (0.1 M, pH 8.0) and a 2-ml aliquot was transfered to an air-tight screw-cap cuvette. Nitrogen was first scrubbed with 5% sodium sulfite solution to remove any traces of oxygen then bubbled through the buffered ferroin solution for 10 min. A 25-pl aliquot was injected into the air-tight cuvette which was then heated to 90 "C for exactly 10 min. Absorbance was recorded immediately. A reagent blank prepared in the same manner, except for the addition of ascorbate, gave negligibly small absorbance values. All assays were performed in duplicate. Ascorbate concentrations were calculated from the regression line of a graph of absorbance as a function of ascorbate concentration. The correlation coefficient of the regression line was 0.994 and the standard error of the estimated variance was 0.0275, giving a maximum error of +4.8%.

Optical Absorption Spectra-Optical absorption spectra were re- corded with the above-mentioned spectrophotometric equipment. Ab- sorbance measurements were recorded at 0.9-nm intervals at a scan-

to be 240,000 (21). ning speed of 1 nm/s. The molecular weight of catalase was assumed

Numerical Analysis-The initial activity of catalase was taken to be the mean value for catalase incubated alone in phosphate buffer for 1 h. Each treatment effect was compared with the control for that particular set of experiments. Positive results (treatments which showed significant deviation from the control) were replicated. Esti- mates of the experimental error were +6.9% among days (based on 8 experiments performed on different days) and +2.2% within a day (based on control experiments performed in triplicate on each day).

The data were fitted to the exponential decay equation, y = ae-" + c, where setting x = 0 yields Q (the initial value) and c is the final value, while b is the time decay constant. The unknown coefficients were calculated by an iterative nonlinear regression procedure (27, 28). This provided the initial rates of enzyme activity (-ab) and final values (c) with their associated standard errors. That the initial rate

Inhibition of Catalase by Ascorbate 1195

= -ab can be shown differentiating y with respect to x and setting x equal to zero. The values of the coefficients were tested with a one- tailed t test (27) to determine the significance (p < 0.05) of differences between treatment effects.

RESULTS

Inhibition of Catalase by Ascorbate-When catalase was incubated with ascorbate there was a progressive loss of catalase activity (Fig. 2). In the initial 25 min, catalase lost 45% of its activity. In the 125 min that followed, a further 25% loss in activity was observed, during which time an oscillating steady state was reached. Subsequently catalase activity was followed for only 1 h, time enough for only a single oscillation phase (Figs. 3 and 8-10). In contrast, the variation in catalase activity, when catalase was incubated alone, was random, thus excluding external factors such as fluctuations in the temperature of the water (C0.02 “C), as a cause of the oscillations.

Trend analysis (29) using the method of orthogonal poly- nomials was applied to the data of Fig. 3 as a test of the statistical validity of these oscillations and showed significant 0, < 0.001) contributions of the cubic and quartic components, thus confirming the nonrandom nature of the oscillations. In the presence of added copper, concentrations of ascorbate <1 mM led to oscillations in the steady state levels of active catalase similar to that observed with ascorbate alone (30). At ascorbate concentrations of 2 mM or greater, oscillations were not seen.

Effects of C U ~ + ( H ~ ) ~ and Metal Chelating Agents-Whereas 2 mM ascorbate alone decreased catalase activity to approxi- mately 60% of its original activity within 1 h (Fig. 3), the further addition of Cu2+(His)2 (to 40 PM) accelerated inacti- vation, decreasing activity to about 20% of the original value in that time. Neither bovine serum albumin, desferrioxamine,

h s l oo \ 90

2oi i

0 : I I 1 I

0 20 40 60 80 100 120 140

TIME (minutes) FIG. 2. Inhibition of catalase by ascorbate. The final incubate

contained 50 mM phosphate buffer, pH 7.0, 0.20 PM catalase and 2 mM ascorbate at 37 “C. Aliquots were withdrawn and residual catalase activity was measured in terms of the rate of decrease in absorbance between 0.450 and 0.400 at 240 nm due to Hz02 decomposition. The standard deviation was 22.2% (see “Experimental Procedures”).

0 0 - 100 n n - W

lo 1 o l I I I I I

0 10 20 30 40 50 60 TIME (minutes)

FIG. 3. Effects of Cu*+(His)% and metal chelators on inhibi- tion of catalase by ascorbate. Incubation and assay conditions were as described in Fig. 2. When present, Cu2+(His)2, DTPA, and desferrioxamine were added to final concentrations of 40 PM, 0.1 mM, and 0.1 mM, respectively. The lines plotted represent: open circles, catalase only; solid circles, catalase plus ascorbate; inverted triangles, catalase plus ascorbate and DTPA; solid squares, catalase plus ascor- bate and desferrioxamine; open triangles, catalase plus ascorbate and Cu”(His)Z; and open squares, catalase plus ascorbate, Cu2+(His)Z, and DTPA. The estimated S.D. was *2.2% (see “Experimental Proce- dures”).

nor diethylenetriamine pentaacetic acid (DTPA) modified the inactivation induced by ascorbate but either desferrioxamine or DTPA protected in the presence of ascorbate plus Cu2+(His)Z, decreasing the inhibition to that of ascorbate alone. Increasing the concentration of histidine between 0.5 and 2 mM resulted in a progressive decrease in the extent of catalase inactivation until at a histidine concentration of 2 mM the rate and extent of inactivation were equivalent to those observed with ascorbate alone (Fig. 4). In contrast, the addition of bovine serum albumin to a final concentration of 100 pg/ml did not retard the inactivation of catalase by ascorbate in the presence of copper.

Effects of Superoxide Dkmutase-Whether in the presence or absence of added copper ions, superoxide dismutase had no significant effect on the inactivation of catalase induced by ascorbate even when added to a final concentration of 0.15 PM. The values for control, control plus superoxide dismutase, and control plus bovine serum albumin were, respectively, 13.0 & 1.9, 12.9 f 1.9, and 13.4 f 2.4 in the presence of copper, and 5.27 f 1.2, 3.57 f 0.9, and 5.62 f 1.8 in the absence of copper. Such a high concentration of superoxide dismutase was used because lower concentrations reportedly failed to affect ascorbate oxidation (31, 32). In these studies, due to the difficulty of permanently denaturing superoxide dismu- tase by boiling, DTPA was included as a control for the possible metal chelating effects of superoxide dismutase. The effects of bovine serum albumin as described above were used as a control to reveal any nonspecific protein effects but, as stated, none were seen.

1196 Inhibition of Catalase by Ascorbate

--- I

‘ O 0 1

EXPERIMENTAL CONDITIONS

FIG. 4. Effect of histidine concentration on the inactivation of catalase in the presence of ascorbate and Cu2+(His)s. Incu- bation mixtures contained 50 mM phosphate buffer, pH 7.0, and 0.20 @I catalase at 37 “C. Ascorbate and CuZ+(His)z were added to final concentrations of 2 mM and 40 pM, respectively. Aliquots were with- drawn and residual catalase activity was measured in terms of the rate of decrease in absorbance between 0.450 and 0.400 at 240 nm due to HzOz decomposition. Measurements were carried out as de- scribed in Fig. 4 .0 , rate of destruction; @, amount of destruction.

Effects of Hydroxyl Radical Scavengers-In the absence of added copper, the addition of benzoate, thiourea, or mannitol (both to 1 mM) led to slight decreases in the reaction rate which did not reach statistical significance, the respective rates being 5.33 f 1.5, 3.39 k 1.8, and 4.76 & 1.3 ( W loss of enzyme activity/min), as compared with the control rate of 5.29 f 1.2. The lack of significant effects of hydroxyl radical scavengers in the absence of copper contrasts with Orr’s (4) report that dimethyl sulfoxide decreased the extent of catalase inhibition induced by ascorbate. This discrepancy can be explained by traces of metal ion contaminants in Orr’s prep- arations. In the copper-stimulated inactivation, thiourea but not mannitol protected (Fig. 5), even when compared with the rate in the presence of urea which was used as a control because it is an analog of thiourea but is a poor hydroxyl radical scavenger (33). When mannitol failed to show an effect at concentrations of 1-10 wM, concentrations of up to 1 M were used, in an unsuccessful attempt to inhibit inactivation possibly mediated by “crypto-hydroxyl radicals” (34, 35).

Effects of Nitrogen and Oxygen Atmospheres-In agreement with the results of Keilin and Hartree (2), nitrogen prevented inactivation of catalase induced by ascorbate alone (Fig. 6). In a control reaction the protection offered by nitrogen, however, was shown to be reversed by the presence of 2 mM H202. In this reaction, to exclude the formation of 0 2 from the decomposition of H202 by catalase and reaction of Hz02 with ascorbate, HzOz was initially incubated with catalase for 5 min. The reaction vial was then evacuated and placed under nitrogen before the reaction was started by the addition of ascorbate. Under these circumstances the rate and final extent of inactivation of catalase by ascorbate was at least as great as the inactivation by ascorbate in the presence of air, being 5.6%/h and 42% total. When, as a control, 2 mM H202 was incubated with catalase alone for 1 h, there was no significant difference between the initial (100%) and final (105%) values of catalase activity.

Surprisingly, changing from an air-bubbled to an oxygen- bubbled medium also offered substantial protection, decreas-

eo

20

0

C E EXPERIMENTAL CONDITIONS

FIG. 5. Effects of active oxygen scavengers on the rate of destruction and the amount of destruction in a system con- taining ascorbate, Cu2+(His)z, and catalase. Incubation and as- say conditions were as described in Fig. 1. The initial rates of reactions and final values with their associated standard errors were calculated by nonlinear regression (see “Experimental Procedures”). Each of the two sets of results were performed on different days. 0, rate of destruction; @, amount of destruction.

EXPERIMENTAL CONDITIONS

FIG. 6. Effects of nitrogen, air, and oxygen atmospheres on the inactivation of catalase by ascorbate in the absence and the presence of Cu2+(His)2. Incubation and assay conditions were as described in Fig. 1. Atmospheres of the desired gas were induced into rubber-capped vials as outlined under “Experimental Proce- dures.” The rates of inactivation were measured as the decrease in the initial rate of action of catalase upon addition of the enzyme to the assay mixture. The final steady state level of catalase activity was measured 60 min after the reaction was initiated. 0, Nz; ., air; H, 0,.

ing the initial rate of inhibition by 45%. Similarly, in the copper-catalyzed inactivation, going from ambient air to pure oxygen protected, slowing the initial rate by 82% and decreas- ing the steady state destruction of catalase by 50%. In contrast to effects in the absence of added copper, however, a nitrogen atmosphere failed to protect against copper-induced inacti- vation, and instead significantly increased the final extent of damage. This aggravation by anaerobic conditions in the presence of copper was more pronounced at 1 mM than at 2 mM ascorbate.

Inhibition of Catalase by Ascorbate 1197

Effects of Ascorbate Concentration on the InitMl Rate of Inhibition-The relationship between the concentration of ascorbate and the initial rate of inhibition was investigated, primarily to determine whether or not the protection offered by 100% oxygen was due solely to a decrease in the initial ascorbate concentration. Inhibition of catalase increased with increasing concentrations of ascorbate, reaching a maximum at an ascorbate concentration of 1.5 mM (Fig. 7). In the presence of added copper, the plateau also occurred at ascor- bate concentrations of 1.5 mM, but all of the rates were of course much higher.

Effects of Ethanol-Because ethanol is a substrate of com- pounds I and 11, it was added to see if it could modulate the inhibitory action of ascorbate on catalase, and indeed ethanol (20 mM) completely prevented inactivation of catalase by ascorbate. Moreover, when added during the reaction to con- centrations of 20 or 40 mM, it reversed inactivation (Fig. 8).

In parallel e,xperiments in the presence of CuZ+(His)2 the effects of ethanol depended on the concentration of ascorbate. Thus, with 2 mM ascorbate, ethanol neither prevented nor reversed inactivation of catalase, whether under air or nitro- gen (Fig. 9). However, when ascorbate concentration was 0.2 mM, ethanol did slow the initial rate of inactivation and also reversed inactivation (by approximately 8%) when added at 30 min (Fig. 10).

Effects of Ascorbate Oxidase-Ascorbate oxidase oxidizes ascorbate without generating HzOz, while liberating semide- hydroascorbate as a reaction product (36). High concentra- tions of ascorbate oxidase (10 units/ml) decreased the inhi- bition of catalase by ascorbate but simultaneous addition of 2 mM H202 partially removed this protection (data not shown). Similarly, addition of H202 to catalase incubated with ascor- bate only had no effect on the extent of inhibition. When low concentrations of ascorbate oxidase (1.7 units/ml) were added 10 min after initiation, the final extent of catalase inhibition was slightly increased (by 18%) compared to ascorbate alone. This may reflect the increased production of semidehydroas- corbate by the enzyme (36).

Optical Absorption Spectra-The extinction coefficient at 404 nm decreased when catalase (2 p ~ ) was incubated with

+

ho 10

0.26 0.60 0.76 I 1.26 1.50 1.76 2 2.26 ASCORBATE CONCENTRATION (mM)

FIG. 7. Effects of ascorbate concentration on the initial rate of inhibition of catalase in the presence and absence of Cua+(His)2. Incubation and assay conditions were as described in Fig. 1. The lines plotted represent: crosses, catalase plus ascorbate plus Cu2+(His)2; solid circles, catalase plus ascorbate. The maximum initial rate of inhibition was based on the rate of inhibition induced by 2 mM ascorbate.

I 0 0

I \ ETHANOL ADDED

80 -

70 -

"

OU 1 0

0 10 20 30 40 50 60 TIME (minutes)

FIG. 8. Effects of ethanol on the inhibition of catalase by ascorbate. Incubation and assay conditions were as described in Fig. 2. The lines plotted represent: open circles, catalase alone; solid circles, catalase plus ascorbate; open squares, catalase plus ascorbate and 20 mM ethanol; solid squares, catalase plus ascorbate and 20 D M ethanol added at 20 min; open triangles, catalase plus ascorbate and 40 mM ethanol added at 20 min. The estimated S.D. was &2.2% (see "Exper- imental Procedures").

ascorbate (2 mM) or ascorbate plus copper in phosphate buffer (50 mM, pH 7.0) at 20 "C (data not shown), indicating that compound I was formed (1). Also, there was a slight shift of 2 nm toward higher wavelengths in the Soret region, which is in agreement with the earlier results (1, 2) that compound I1 is formed in this reaction. Orr's (3) inability to detect these small changes is presumably because he used much lower concentrations of catalase, and a lower pH. The presence of Cu2+(His)2 (40 pM), however, caused a progressive precipita- tion of catalase, precluding unambiguous absorption spectra.

DISCUSSION

The failure of either ascorbate or oxygen to inhibit when the other is absent and the dramatic acceleration when copper is added argue that the immediately damaging agent is a species produced in the reduction of oxygen by ascorbate. Transition metal ions do not appear to be implicated in the reversible inactivation which occurs in the absence of added copper. Moreover, we will argue that most of the other plau- sible candidates for such a role are excluded by the current scavenger studies, leaving semidehydroascorbate as the most important candidate for a role as the main damaging species.

Activated Species of Oxygen Do Not Inhibit-Although both oxygen and ascorbate are both required for the reversible inactivation of catalase, active forms of oxygen which mediate ascorbate oxidation are not the primary inhibitory species. First, the inhibition of catalase is not directly dependent on the rate of ascorbate oxidation. Thus, although metal chelat- ing agents virtually blocked ascorbate oxidation in the pres- ence of catalase, they did not slow the reversible inhibition of catalase (Fig. 3). Second, scavengers of superoxide or hydroxyl radicals had no significant effects on the initial rate of catalase

1198 Inhibition of Catalase by Ascorbate

h !s l o o o 90

80

70

60

50

40

30

20

10

I I 1 I

0 10 20 30 40 50 60 TIME (minutes)

FIG. 9. Effects of ethanol on inactivation of catalase by ascorbate plus Cu2+(His)Z under atmosdheres of air or nitro- gen. Incubation and assay conditions were as described in Figs. 1 and 2. The lines plotted represent: open circles, catalase alone; trian- gles, catalase plus ascorbate and Cu2+(His)2 under air; inverted tri- angles, catalase plus ascorbate, Cu2+(His)2, and 20 mM ethanol under air; diamonds, catalase plus ascorbate, Cu2+(His)z, and 20 mM ethanol added at 20 min under air; solid circles, catalase Flus ascorbate and Cu2+(His)2 under nitrogen; open squares, catalase plus ascorbate, Cu2+(His)2, and 20 mM ethanol under nitrogen; solid squares, catalase plus ascorbate and CuZ+(His)2 under nitrogen and 20 mM ethanol added at 20 min. The estimated S.D. was k2.2% (see “Experimental Procedures”).

inhibition (Fig. 5). Evidently, in contrast to previous proposals (2,3), the inactivation of catalase by ascorbate in the presence of copper does not directly involve free superoxide, hydroxyl radicals, or hydrogen peroxide. Conceivably site-specific pro- duction of active species of oxygen might be involved, but if so, these were not detectible by the presence of even very high concentrations of mannitol, catalase, or superoxide dismutase. Moreover, since the medium was homogeneous, the arguments regarding compartmentation often made in cellular systems cannot apply. Because other hydroxyl scavengers were inef- fective in the metal-catalyzed inactivation, the ability of thi- ourea (and presumably dimethyl sulfoxide (4)) to decrease the rate of inactivation of catalase by ascorbate plus Cu2+ must be attributed to properties other than their ability to scavenge hydroxyl radicals (e.g. their metal chelating properties (37, 38)). The arguments against participation of oxygen radical intermediates do not, however, exclude semidehydroascor- bate, which can be produced by comproportionation between ascorbate and dehydroascorbate (39) formed during preincu- bation. This possibility will later be considered in more detail.

HzO, Alone Is Not Inhibitory-That hydrogen peroxide is not the primary immediate inhibitory agent follows, not only because the presence of catalase in the incubation medium precludes accumulation of HzOz, but also because Hz02 alone failed to inhibit when added to incubates of catalase, or even to catalase plus ascorbate. Moreover, an atmosphere of nitro- gen failed to protect against copper-catalyzed inactivation

lo 1 “ I 0 Ib i0 i0 10 50 60

I ~~ ~~ I

TIME (minutes) FIG. 10. Effects of low concentrations of ascorbate on the

inactivation of catalase in the presence of CuZ+(His)z. Incuba- tion and assay conditions were as described in Fig. 1 except ascorbate was present at 0.2 mM. Lines plotted represent: open circles, catalase only; solid circles, catalase and ascorbate plus CuZ+(His)z; ope? squares, catalase and ascorbate plus Cu2+(His)2 under nitrogen; sold squures, catalase and ascorbate plus Cu2+(His)z and 20 mM ethanol; triangles, catalase and ascorbate plus Cu2+(His)2 and 20 mM ethanol added at 30 min. The estimated S.D. was -+2.2% (see ”Experimental Procedures”).

although it completely blocked oxidation of ascorbate, and thus H202 formation. Furthermore, an atmosphere of 100% oxygen (which increased the initial rate of ascorbate oxida- tion, presumably increasing HzOz production) actually pro- tected rather than aggravated the inhibition of catalase (Fig. 5). On the other hand, HzOz is clearly implicated in the mechanism of inhibition, since it relieved the protection of- fered by nitrogen or by concentrations of ascorbate oxidase high enough to remove oxygen.

Hypothesis: Semidehydroascorbate Is the Damaging Spe- cies-The following rationale supports the conclusion that semidehydroascorbate is the main species that inhibits cata- lase. In addition to the rather weak argument by exclusion of the other intermediates discussed above, there are a number of other indirect lines of evidence which, taken together, are quite persuasive. Although ascorbate alone is not very dam- aging, any of the following three reagents can release ascor- bate’s inhibitory action: 1) oxygen, 2) hydrogen peroxide, or 3) cupric ions. Any of the these is capable of oxidizing ascor- bate to semidehydroascorbate. Also the enhancement of the inhibitory action of ascorbatefoxygen by the presence of either copper or ascorbate oxidase (both of which increase the yield of semidehydroascorbate) supports Chis view. Finally, the protection offered by 0, is difficult to account for on the basis of a mechanism involving the sequential actions of ascorbate and H202, unless oxygen is removing an interme- diate of ascorbate oxidation which inhibits catalase, most plausibly semidehydroascorbate.

One can calculate that the reaction between oxygen and ascorbate is kinetically (14) and thermodynamically (40) un-

Inhibition of Catalase by Ascorbate 1199

favorable (the respective reduction potentials for the couples 0 2 / 0 2 ~ and A/AHT being -0.33 and -0.21) but the rate of the reaction is nevertheless non-zero. A simple calculation (14) shows that the rate of reaction of semidehydroascorbate with Op, at saturating levels of 0 2

(A’ + 02 + A + 02’) (1)

is approximately 50% of the rate of disproportionation of this radical.

Inactivation in the Absence of Added Copper Is Reversible- Inhibition of catalase by ascorbate was never complete but reached a steady state at between 35 and 90% inhibition. This suggests that the inhibition is reversible by species present in the reaction medium. That ethanol can reverse inhibition indicates that compound I1 is the inactive form of catalase produced in this reaction. The requirement for HzOz for anaerobic inactivation implies that compound I is the likely initial target of damage.

Mechanism of the Reversible Inactivation-On the basis of the above results, the reaction pathway involved in inhibition of catalase by ascorbate can be written as follows:

AH, + 0% + A + Hz02 (2)

(3) Ferricatalase + H202 Compound I

Fe(II1) FeW)

Compound I + AH- + Compound I1 + A’ FeW) Fe(IV)

FeW) Fe(IV)

(4)

(5)

where AH% ascorbate, AT is semidehydroascorbate, and A is dehydroascorbate.

This mechanism is analogous to that proposed by Yamazaki and Piette (41) for the peroxidatic action of peroxidase. The relative extents to which ascorbate and semidehydroascorbate reduce compound I cannot be determined from the current investigation. It must be concluded, however (on the basis that semidehydroascorbate contributes at much lower concen- trations than ascorbate), that the rate constant for its reaction with compound I is considerably greater than that of ascor- bate.

Ethanol prevents inhibition by preferentially reacting with compound I to regenerate ferricatalase, i.e.

Compound I + CH3CH20H -+ ferricatalase + CH3CH0 + HzO

Compound I + A’ -+ Compound I1 + A

F e W Fe(II1) (6)

Moreover, ethanol reverses inhibition by rereducing com- pound TI to active catalase, i.e. 2 Compound I1 + CH3CH20H + 2 ferricatalase + CH3CH0

Fe(1V) (FeIII) (7)

That a steady state level of catalase activity is finally achieved, rather than complete inhibition, implies that compound I1 is reactivated by one electron transfer from single-electron do- nors in the redox system.

Compound I1 + AH- or A’ --$ ferricatalase + A’ or AH- (8)

The steady state level of activity then reflects the ratio of the relative rates of reactivation and inactivation, and the shift toward a higher level of steady state activity (as caused by the admission of oxygen) reflects an increase in the ratio of reactivation to inactivation (Fig. 6).

Mechanism of the Irreversible Inactivation in the Presence of Added Cu-Interaction between ascorbate and Cu2+ is an integral part of the irreversible inactivation, because inacti- vation in the simultaneous presence of these two reactants

was greater than the sum of their individual effects. Also, the metal chelating agent DTPA (which prevents the reaction between Cu2+ and ascorbate) decreased inactivation to that seen with ascorbate alone.

On the basis of the above discussion, it is proposed that, after initial reversible inactivation of the catalase by a metal- independent, semidehydroascorbate-dependent mechanism, H20z accumulates to the point where Fenton type reactions can occur and irreversible inhibition ensues. Such a mecha- nism reconciles the previously conflicting explanations, im- plying that the mechanism of Chance (1) precedes and is prerequisite to the mechanism of Orr (3), which presumably would not have been seen at all had he taken the precautions regarding freedom from interference from transition metal ions made mandatory by present knowledge of their partica- tion.

The failure to respond to the addition of scavengers of active oxygen not only excludes participation of the relevant radicals as participants in any Fenton system but may also imply a “site-specific” mechanism. Such a site-specific mech- anism is consistent with the known ability of Cu2+ to bind to peptide linkages in general, and to histidine residues in par- ticular (42) (such as the essential histidine residue at the active site of catalase). The progressive protection of catalase by increasing concentrations of histidine is in complete agree- ment with the site-specific mechanisms and also with results of Shinar et al. (19) for the inactivation of acetylcholine esterase by ascorbate plus Cu2+. Their interpretation of the results is that histidine protects by preventing Cup+ from binding to the enzyme. Moreover, ascorbate plus Cu2+ induces specific degradation of histidine residues (22), and semide- hydroascorbate is known to react selectively with amino acids including histidine (43). Either site-specific (“crypto-”) hy- droxyl radicals or semidehydroascorbate might be the ultimate irreversible inhibitor. The failure of even the highest concen- trations of hydroxyl radical scavengers to protect, together with the protective actions of oxygen, tend to support a role for semidehydroascorbate, but without conclusively excluding the possibility of site-specific hydroxyl radical production with a possible role for semidehydroascorbate as a Fenton electron donor.

Oscillatory Behavior-Oscillations commonly occur in ex- tensively coupled multivariate systems which involve conver- sions far from equilibrium and which contain unstable states (44). Yamasaki and co-workers (45, 46) have demonstrated that the analogous oxidation of NADH by horseradish per- oxidase leads to coupled oscillations involving molecular ox- ygen and compound 111. On the assumption that compound I1 is rereduced to active catalase by ascorbate or semidehy- droascorbate (Reaction 8), the series of oscillations (Fig. 2) can be explained by fluctuations in the relative concentrations of compounds I and 11 as the reaction progresses to equilib- rium. These oscillations deserve to be investigated further so as to establish their exact mechanism.

Mechanism of Protection by Oxygen-That an atmosphere of nitrogen fails to protect in the presence of copper histidine indicates that the cupric ions themselves allow the anaerobic formation of the inactivating intermediates by direct oxida- tion of ascorbate. Involvement of semidehydroascorbate in the system ascorbate plus Cu2+(His)z is well-documented (47). In the presence of metals, this reaction reportedly proceeds through single inner sphere electron transfer steps within a metal-chelate-ascorbate complex. The first electron transfer is followed by dissociation of the reduced metal ion and recombination of semidehydroascorbate with another metal ion in the higher valence state. However, in the presence of

1200 Inhibition of Catalase by Ascorbate

excess molecular oxygen, rapid reoxidation of the metal in the ternary complex occurs before this complex can dissociate. This allows a second reduction of the metal, this time by semidehydroascorbate. Thus, elevated concentrations of molecular oxygen prevent semidehydroascorbate from be- coming available for side reactions (such as in the current study the inactivation of catalase). This analysis suggests several further approaches for research into the specific mechanism of irreversible inactivation of catalase by ascor- bate plus Cu*+(His)2. Among these, the effects of scavengers on the anaerobic inactivation of compound I deserve a high priority.

Implications-Whether the ability of catalase to be reduced by one-electron donors is a harmful side reaction or if it serves metabolically useful ends deserves further investigation. The extent to which two successive reductions by semidehydroas- corbate generate active compound I reflects the ability of catalase to act peroxidatically toward these and related radi- cals, such as semiquinones, thus limiting radical chain prop- agation. This conclusion extends the hypothesis of Sullivan et al. (48) that 02, superoxide dismutase, and catalase form a metabolic pathway that protects against oxidative damage. Whether the rate of removal of radicals of this type by catalase is rapid enough to provide a useful contribution to their elimination remains to be determined. Against this is the current finding that, in the presence of free copper, the reversible reaction of the enzyme predisposes toward its irre- versible inactivation.

The cytotoxicity of semiquinones has been emphasized recently and the current findings demonstrate that the tox- icity of semidehydroascorbate should not be neglected. As discussed by Swartz and Dodd (49), the comparatively low reactivity of semidehydroascorbate allows it to migrate rela- tively long distances and eventually react in a site-specific manner with crucial cellular targets distant from their site of formation. Among these targets, catalase must be considered a significant one since, in situations where catalase is playing a major protective role, even its reversible inactivation may allow local accumulation of hydrogen peroxide. This in turn will be compounded by the ensuing Fenton reactions including the irreversible inactivation of catalase. Such mechanisms reportedly play a part in the known cytotoxicity of reducing agents (including ascorbate) in the presence of metal ions (50, 51). These processes would be particularly important if per- oxidases such as glutathione peroxidase were subject to simi- lar kinds of inhibition, and indeed, production of inactive compound I1 has also been demonstrated in an analogous reaction of horseradish peroxidase with a reducing agent (52).

REFERENCES 1. Chance, B. (1950) Biochem. J. 46,387-402 2. Keilin, D., and Hartree, E. F. (1951) Biochem. J . 49,8&104 3. Orr, C. W. M. (1967) Biochemistry 6,3000-3006 4. Orr, C. W. M. (1967) Biochemistry 6,2995-3000 5. Orr, C. W. M. (1966) Biochem. Biophys. Res. Commun. 23 ,854860 6. Chance, B., Sies, H., and Boveris, A. (1979) Physiol. Reu. 59,527-605

7. Freeman, B. A., and C r c p J . D. (1982) Lab. Inuest. 47,412-426 8. Eaton, J. W., Boraas, , and Etkm, N. L. (1972) in Hemoglobrn and Red

Press, New York Cell Strllcture and Function (Brewer, G. J., ed) pp. 121-131, Plenum

9. Taqui Khan, M. M., and Martell, A. E. (1967) J . Am. Chem. Soc. 89,4176- 4185

10. Bielski, B. H. (1982) in Ascorbic Acid: Chemistry, Metabolism, and Uses (Seib, P. A., and Tolbert, B. M., eds) pp. 81-100, American Chemical Society, Washington D.C.

11. Scarpa, M., Stevanato, R., Viglino, P., and Rigo, A. (1983) J. Biol. Chem.

12. White, G. A., and Krupka, R. M. (1965)Arch. Biochem. Biophys. 110,448- 268,6695-6697

13. Winterbourn, C. C. (1981) Biochem. J. 198.125-131 461

14. Bielski, B. H., Richter, H. W., and Chan, P. C. (1975) Ann. N. Y. Acad. Sci. 268.231-237

15. Yamazaki, I. (1962) J . Biol. Chem. 237.224-229

17. Kono. Y.. and Fridovich. I. (1982) J. Blol. Chem. 257.5751-5754 16. Sullivan, S. G., and Stern, A. (1980) Bipchem. Pharmol . 29,2351-2359

18. Shimim,.N., Kobayashi; K.; and' Hayashi, K. (1984) 3. Biol. Chem. 269,

19. Shinar, E., Navok, T., and Chevion, M. (1983) J. Biol. Chem. 268,14778- 4414-4418

14783 20. Samuni, A., Aronovitch, J., Godinger, D., Chevion, M., and Czapski, G.

(1983) Eur. J . Biochem. 137,119-124 21. Nicholls, P., and Schonbaum, G. R. (1963) in The Enzymes (Boyer, P. D.,

Press, New York Lardy, H., and Myrback, K., ed) Vol. 8,2nd Ed., pp. 147-226, Academic

22. Golan-Goldhirsh, A., and Whitaker, J. R. (1984) Abstracts of the Proceed- ings of the Pacijic Slope Biochemistry Conference July 1984, Santa CNZ,

23. Halliwet B. (1973) Biochem. J. 135,379-381 Italy 18

24. Aebi, H. (1974) in Methods of Enzymatic Analysis (Bergmeyer, H. U., ed)

25. Chevion, k , andNavok, T. (1983) Anal. Biochem. 128,152-158 26. Schilt, A. A., and Di Tusa, M. R. (1982) Talunta 29,129-132 27. Cleland, W. W. (1967) Adu. Enzymol. 2 9 , 1-32 28. Gallant, A. R. (1975) Am. Stat. 29,73-81 29. Keppel, G. (1982) in Design and Analysis: A Researcher's Handbook pp.

30. Obukhova, L. K., and Emanuel, N. M. (1983) Russ. Chem. Reu. 5 2 , 353- 127-143, Prentice-Hall, Inc., Englewood Cliffs, New Jersey

31. Rigo, A., Scarpa, M., Argese, E., Ugo, P., and Vlglino, P. (1984) in Oxygen 372

Radicals in Chemistry and Biology (Bors,. W., Saran, M., and Tait, D., eds) pp. 171-176, Walter de Gruyter, Berlln

32. Halliwell, B., and Foyer, C. H. (1976) Biochem. J . 156,697-700 33. Cederbaum, A. I., Dicker, E., Rubin, E., and Cohen, G. (1977) Biochem.

34. Youngmann, E. (1984) Trends Biochem. Sci. 9,280-283 35. Borg, D. (1984) Isr. J . Chem. 24,38-53 36. Yamazaki, I., Mason, H. S., and Piette, L. (1960) J. Biol. Chem. 236,2444-

37. Gould, D. C., and Mason, H. S. (1966) in The Biochemistry o Copper 2449

(Peisach, J., Aisen, P., and Blumberg, W. E., eds) pp. 35-47, dcademic

38. Cotton, F. A., and Wilkinson, G. (1980) Aduanced Inorganic Chemistry 4th Press, New York

Ed., p 177; 188, John Wiley & Sons, New York 39. Foster, 8: V., Weisa, W., and Staudinger, H. (1965) Justus Liebigs Ann.

Chem. 6 9 0 , 1 6 40. Nishikimi, M. (1975) Biochem. Biophys. Res. Commun. 63,463-468 41. Yamazaki, I., and Piette, L. H. (1961) Biochem. Biophys. Acta 50,62-69 42. Pickart, L., Freedman, J. H., Loker, W. J., Peisach, J., Perkins, M.,

Stenkamp, R. E., and Weinstein, B. (1980) Nature 288,715-716 43. Adams, G. E., Aldrkh, J. E., Bisby, R. H., Cundall, R. B. Redpath, J. L.,

and Willson, R. L. (1972) Radiat. Res. 49, 278-289 44. Frank, U. F. (1978) Angew. Chem. Int. Ed. Engl. 17 , l -15 45. Yamazaki I., Yokota, K., and Nakajima, R. (1965) Biochem. Biophys. Res.

46. Yamazaki, I., and Yokota, K. (1967) Biochim. Biophys. Acta 132,310-320 Commin. 2 1,582-586

47. Martell, A. E. (1982) in Ascorbic Acid: Chemistry, Metabolism, and Uses (Seib, P. A,, and Tolbert, B. M., eds) pp. 153-178, American Chemical Society, Washington D. C.

48. Sullivan S. G., Winterbourn, C. C., and Stem, A. (1983) in Oxy Radicals and h e i r Scauenger Systems. (Cohen, G., and Greenwald, R. A., eds) Vol. 1 pp. 364-367, Elsevier Science Publishing Co. Inc., New York

49. Swartz, H. M., and,Dcdd, N. J. F. (1981) in Oxygen and Oxy-Radicals in Chemlstry and Bwlow (Rodgers, M. A. J., and Powers, E. L., eds) pp. 161-168. Academlc Press. New York

Vol. 2, p 673 684, Academic Press, New York

Biophys. Res. Commun. 78,1254-1262

~" ~.., - .~ ~ _ . ~ . ~ .~~~ 50. Shamber er, R J. (1984) Mutat. Res. 133, 135:159 51. Rowle 8. A. L d Halliwell, B. (1983) Clan. Scc. 64 ,649453 52. Hayasti, Y., ind Yamazaki, I. (1979) J . BioL Chem. 264,9101-9106