Kinetics of interconversion of ferrous enzymes, compound II and compound III, of wild-type synechocystis catalase-peroxidase and Y249F: proposal for the catalatic mechanism.

With the exception of catalase-peroxidases, heme peroxidases show no significant ability to oxidize hydrogen peroxide and are trapped and inactivated in the compound III form by H2O2 in the absence of one-electron donors. Interestingly, some KatG variants, which lost the catalatic activity, form compound III easily. Here, we compared the kinetics of interconversion of ferrous enzymes, compound II and compound III of wild-type Synechocystis KatG, the variant Y249F, and horseradish peroxidase (HRP). It is shown that dioxygen binding to ferrous KatG and Y249F is reversible and monophasic with apparent bimolecular rate constants of (1.2 +/- 0.3) x 10(5) M(-1) s(-1) and (1.6 +/- 0.2) x 10(5) M(-1) s(-1) (pH 7, 25 degrees C), similar to HRP. The dissociation constants (KD) of the ferrous-dioxygen were calculated to be 84 microm (wild-type KatG) and 129 microm (Y249F), higher than that in HRP (1.9 microm). Ferrous Y249F and HRP can also heterolytically cleave hydrogen peroxide, forming water and an oxoferryl-type compound II at similar rates ((2.4 +/- 0.3) x 10(5) M(-1) s(-1) and (1.1 +/- 0.2) x 10(5) M(-1) s(-1) (pH 7, 25 degrees C)). Significant differences were observed in the H2O2-mediated conversion of compound II to compound III as well as in the spectral features of compound II. When compared with HRP and other heme peroxidases, in Y249F, this reaction is significantly faster ((1.2 +/- 0.2) x 10(4) M(-1) s(-1))). Ferrous wild-type KatG was also rapidly converted by hydrogen peroxide in a two-phasic reaction via compound II to compound III (approximately 2.0 x 10(5) M(-1) s(-1)), the latter being also efficiently transformed to ferric KatG. These findings are discussed with respect to a proposed mechanism for the catalatic activity.

Catalase-peroxidases (KatGs, 1 EC 1.11.1.7) are bifunctional heme b-containing enzymes exhibiting an overwhelming catalase activity and a substantial peroxidase activity with various one-electron donors. The catalatic activity of KatGs, i.e. the dismutation of hydrogen peroxide to oxygen and water, is unique for both heme peroxidase superfamilies (i.e. the superfamily of peroxidases from archaea, bacteria, fungi, and plants and the superfamily of animal peroxidases). Despite the availability of crystal structures (1)(2)(3)(4), no structural bases for the high catalase activity of KatGs has been assigned so far. Whether in KatGs compound I plays a similar role in the catalase cycle as in monofunctional catalases (EC 1.11.1.6; see Fig. 1, Reactions 1 and 2) (5) is not clear at the moment. From a thermodynamic point of view, there is no barrier for the two-electron oxidation process, which oxidizes hydrogen peroxide to dioxygen. The standard reduction potential of the redox couple EЈ°(O 2 /H 2 O 2 ) is 280 mV (6), and normally, the oxidizing intermediates of heme peroxidases are strong oxidants (7). Similar to monofunctional catalases, compound I does not accumulate during H 2 O 2 degradation but can be trapped by using organic peroxides (8). However, this intermediate exhibits only a low reactivity toward H 2 O 2 (8).
Another peculiarity of KatGs regards the spectral features of compound II formed by one-electron reduction of compound I (see Fig. 1, Reaction 3), which in most heme peroxidases is a ferryl species with a red-shifted Soret band and two typical maxima in the visible region (7) and accumulates in the peroxidase cycle (see Fig. 1, Reactions 1, 3, and 4). Both in the absence and in the presence of one-electron donors, an accumulation of a ferryl-like compound II was never observed in wild-type KatGs, suggesting that the spectral features of KatG compound II could be similar to that of the ferric protein (8,9).
In contrast to other heme peroxidases, KatG compound III does not accumulate even under high concentrations of H 2 O 2 . In Mycobacterium KatG, 400 mM H 2 O 2 had to be added to ferric KatG to monitor the formation of compound III spectroscopically, but the observed intermediate was unstable and rapidly decayed to the ferric enzyme (10). Normally, peroxidase compound III is a ferrous-dioxy/ferric-superoxide complex (similar to oxyhemoglobin or oxymyoglobin) formed by three different routes (7): (a) the oxygenation of ferrous peroxidase (see Fig. 1, Reaction 6); (b) the reaction of ferric enzyme with superoxide anion (see Fig. 1, Reaction 8); and (c) the reaction of (ferryl)compound II with excess hydrogen peroxide (see Fig. 1, Reaction 11). Recently, the rate of reaction between ferric Mycobacterium KatG and superoxide was determined by pulse radiolysis to be 5.5 ϫ 10 5 M Ϫ1 s Ϫ1 (11). However, the reactivity of ferrous KatG is completely unknown. Neither its reactivity toward dioxygen (see Fig. 1, Reaction 6) nor its two-electron oxidation reaction to compound II mediated by H 2 O 2 (see Fig. 1, Reaction 10) (12) nor the transition of compound II to compound III has been investigated so far, although the knowledge of these reactions could help to understand the KatG-specific high catalase activity. This is most obvious when looking at the variants Y249F of Synechocystis KatG (13) or Y229F of Mycobacterium KatG (14). These tyrosines are part of a KatG-specific covalent link at the distal heme side, formed among the side chains of conserved tryptophan, tyrosine and methionine. Most interestingly, disruption of this covalent link by exchange of these tyrosines converted the bifunctional KatGs to typical monofunctional heme peroxidases. The catalase activity was dramatically decreased, the formation of (conventional) compound I could be followed even with equimolar H 2 O 2 , and a typical (ferryl-)compound II was formed, which was easily transformed to compound III (13,14).
These findings motivated us to perform the following comparative kinetic investigation of the reaction of ferrous wildtype KatG from Synechocystis, the variant Y249F, and horseradish peroxidase, respectively, with both dioxygen and hydrogen peroxide as well as of the interconversion of compound II to compound III. By using the stopped-flow technique, we were able to calculate the apparent bimolecular rate constants of the transitions ferrous protein 3 compound III, ferrous protein 3 compound II, and compound II 3 compound III and detect significant differences between the three investigated protein species.

EXPERIMENTAL PROCEDURES
Materials-Hydrogen peroxide, obtained as a 30% solution from Sigma, was diluted, and the concentration was determined using ⑀ 240 ϭ 39.4 M Ϫ1 cm Ϫ1 (15). Sodium dithionite was from Aldrich and was used either directly or from a freshly prepared anaerobic stock solution. All experiments were performed at 25°C and 50 mM phosphate buffer, pH 7.0. All solutions were made anaerobic by flushing with nitrogen gas (oxygen Ͻ 3 ppm) and stored in a glove box (Meca-Plex, Neugebauer) with a positive pressure of nitrogen (25 millibars). Recombinant wildtype KatG and the Y249F variant were expressed and purified as described previously (8,13). Horseradish peroxidase was purchased from Sigma (Typ VIA). Ferrous peroxidases were prepared in the glove box by a minimal excess of solid sodium dithionite. The presence of excess dithionite was estimated by its absorbance at 315 nm.
Methods-Oxygen concentrations were measured polarographically (YSI 5300 biological oxygen monitoring system) by utilizing a Clarktype electrode (YSI 5331 oxygen probe) inserted into a stirred water bath (YSI 5301B) at 25°C. To achieve a defined oxygen concentration, oxygen-saturated 50 mM phosphate buffer (100% oxygen saturation at 25°C corresponds to 1262 M dioxygen (16), giving a final concentration of 631 M after mixing in the stopped-flow machine) was bubbled with nitrogen for different periods, and the oxygen concentration was measured polarographically. Oxygenated buffer solutions were transported to the stopped-flow device in gas-tight syringes with no gas head space to avoid loss of oxygen or contact with oxygen.
The sequential mixing stopped-flow apparatus (Model SX-18MV) and the associated computer system and software were from Applied Photophysics. All reactions were followed by either using a monochromator or using the diode-array detector (Applied Photophysics PD.1) attached to the stopped-flow machine. All tubes of the stopped-flow were flushed several times with a nitrogen-bubbled dithionite solution to remove free oxygen from the system. Final enzyme concentration varied typically between 1 and 3 M. For measuring the reaction of ferrous KatG and HRP with O 2 or H 2 O 2 , the conventional stopped-flow mode was used. Compound III formation was followed as absorbance increase at 414 nm (wild-type and Y249F), whereas compound II formation was followed as absorbance decrease at 440 nm (wild-type KatG and Y249F). HRP compound II formation was followed at 418 nm. In a typical experiment, 2-6 M ferrous peroxidase was mixed with various concentration of H 2 O 2 or O 2 in 50 mM phosphate buffer, pH 7.0. Oxygenated buffer, nitrogen-bubbled hydrogen peroxide solutions (1-200 M), and preparations of ferrous wild-type KatG, Y249F, or HRP were transported to the stopped-flow device in gas-tight syringes.
The H 2 O 2 -mediated transition of compound II to compound III was either monitored in the H 2 O 2 -mediated reaction sequence ferrous peroxidase 3 compound II 3 compound III (in the case of wild-type KatG) by using the conventional mode and the diode array detector or monitored in the sequential stopped-flow mode by preforming compound II (in the case of Y249F and HRP) and following its reaction with H 2 O 2 as absorbance increases at 414 nm. In a typical experiment, Y249F was mixed with equimolar H 2 O 2 , and finally, after a delay time of 7 s, with hydrogen peroxide. The final concentrations and conditions are as follows: 1 M Y249F, 10 -150 M H 2 O 2 , 50 mM phosphate buffer, pH 7.0, and 25°C.
All reactions were also analyzed by using the Pro-K simulation program from Applied Photophysics. Time traces were fitted using the single-exponential equation of the Applied Photophysics software, and from the slopes of the linear plots of the k obs values versus substrate concentration, the apparent second-order rate constants were obtained.

Reaction of Ferrous Wild-type KatG and the Y249F Variant
with Dioxygen-Even in the presence of a high excess of hydrogen peroxide, it is not possible to monitor spectroscopically the formation of compound III (oxyferrous peroxidase) of Synechocystis wild-type KatG (Fig. 1, Reaction 6). By contrast, only a minor excess (10-fold) of H 2 O 2 led to the formation of compound III in the variant Y249F, as indicated by the appearance of absorption bands at 414, 542, and 576 nm (13). To understand these significant differences and eventually to relate them to the dramatic differences in enzymatic properties (Y249F lost the catalase activity), it is necessary to investigate pathways of compound III formation and transition. Thus, the ferrous forms of both wild-type KatG and Y249F were formed and probed for their reactivity with dioxygen.
A good spectrum of ferrous Synechocystis wild-type KatG has its peak maxima at 438 nm and in the visible region at 558 nm with a shoulder around 580 nm ( Fig. 2A, first spectrum). Ferrous Y249F exhibited similar spectral features (Fig. 3A, first spectrum). The addition of dioxygen to ferrous wild-type KatG resulted in the rapid formation of compound III with absorbance maxima at 414, 548, and 578 nm. Fig. 2A demonstrates the spectral transition with clear isosbestic points at 391, 427, 473, 552, 573, and 590 nm. The reaction was monophasic, and the observed pseudo-first-order rate constants (k obs ) were determined by fitting the absorbance (414 nm) versus time curves to a single exponential function (Fig. 2B). The observed pseudofirst-order rate constants were directly proportional to the initial oxygen concentrations and allowed the determination of the apparent second-order rate constant for oxygen binding (k on ) from the slope of the plot (Fig. 2C). With (1.2 Ϯ 0.3) ϫ 10 5 M Ϫ1 s Ϫ1 (pH 7.0 and 25°C), the rate constant is about twice as much as that determined for horseradish peroxidase ((5.3 Ϯ 0.4) ϫ 10 4 M Ϫ1 s Ϫ1 at pH 7.0 and 25°C) (17). However, the k off values are different. In HRP, k off was reported to be Ͻ0.1 s Ϫ1 (17), resulting in K D ϭ k off /k on Ͻ 1.9 M (17), whereas k off in the case of wild-type KatG is 10 s Ϫ1 (Fig. 2C, intercept), resulting in a K D value of 84 M. As a consequence of this high dissociation rate constant, the ferrous-dioxy complex is unstable.
Binding of dioxygen to ferrous Y249F (absorption bands at 438 and 558 nm) followed similar kinetics. The transition to compound III (absorption bands at 414, 545, and 578 nm) was also monophasic with clear isosbestic points at 391, 427, 472, 550, and 586 nm, respectively (Fig. 3A). The observed firstorder rate constants were determined at 414 nm (compound III formation) and at 440 nm (oxidation of the ferrous enzyme) and plotted against the concentration of oxygen, yielding an apparent bimolecular rate constant of (1.6 Ϯ 0.2) ϫ 10 5 M Ϫ1 s Ϫ1 at 414 nm at pH 7.0 and 25°C (Fig. 3C). Similar to wild-type KatG, the finite intercept of 20 s Ϫ1 indicates reversible dioxygen binding. The calculated dissociation constant, K D , is 129 M at pH 7.0. HRP compound III reverts over a period of minutes to the native ferric state (k decay ϭ 8.2 ϫ 10 Ϫ3 s Ϫ1 ) according to Reaction 9 in Fig. 1, and the transition is independent of oxygen concentration (17). Most interestingly, wild-type KatG compound III is much more unstable. At pH 7.0 the decay rate was determined to be 2.5-3.6 s Ϫ1 , which is about 350 times higher than that of HRP compound III. By contrast, conversion of Y249F compound III to the ferric state was about 40 times slower (0.07 s Ϫ1 ) than that of wild-type KatG compound III.
From these findings, the following conclusion can be drawn. The kinetics of compound III formation by binding of dioxygen to the ferrous proteins as well as the spectral features of compound III are similar in wild-type KatG, Y249F, and HRP. However, there is a significant difference in the stability of the oxyperoxidases. In HRP, compound III is very stable; it neither decays to ferric HRP-releasing superoxide (k decay ϭ 8.2 ϫ 10 Ϫ3 s Ϫ1 ) (Fig. 1, Reaction 9) nor dissociates to ferrous HRP-releasing dioxygen (indicated by the low k off value of 0.1 s Ϫ1 (17) (Fig.  1, Reaction 7)) at reasonable rates. By contrast, KatG compound III is very unstable, and both pathways are more than 2 orders of magnitude faster (k off of 10 s Ϫ1 and k decay of 2.5-3.6 FIG. 1. Generalized reaction scheme for heme peroxidases and catalases. In the first step, hydrogen peroxide is used for compound I formation (Reaction 1). Compound I (ferryl porphyrin -cation radical or ferryl protein radical) either reacts with a second hydrogen peroxide back to the ferric enzyme (Reaction 2, i.e. the catalatic reaction) or reacts with one-electron donors via compound II (ferryl species or protein radical) to the ferric peroxidase (Reactions 3 and 4). With an excess of H 2 O 2 , compound II reacts to compound III (Reaction 11). Alternatively, compound III (ferrous-dioxy/ferric-superoxide complex) is formed upon dioxygen binding to the ferrous heme protein (Reaction 6), which is reversible (Reaction 7), or by the H 2 O 2 -mediated oxidation of the ferrous protein to compound II (Reaction 10). s Ϫ1 ). In Y249F, the kinetics of Reactions 6 and 7 are similar to wild-type KatG, but the rate of Reaction 9 is significant slower (k decay ϭ 0.07 s Ϫ1 ). The exchange of Y249 prevents the formation of the KatG-typical covalent adduct (18) but only slightly affects the binding and dissociation of fluoride (19) and dioxygen to the ferric protein. However, it decelerates compound III decay via Reaction 9 (Fig. 1).
Reaction of Ferrous HRP, Wild-type KatG, and the Y249F Variant with Hydrogen Peroxide-The high dissociation rate constant of the ferrous-dioxy complex of KatG implies the presence of ferrous KatG in steady state. Besides binding dioxygen, a ferrous heme peroxidase can also mediate the heterolytic cleavage of hydrogen peroxide, forming water and compound II (Fig. 1, Reaction 10) (20).
The spectral transitions observed upon reaction of ferrous HRP or the KatG variant Y249F with hydrogen peroxide were very similar. Fig. 4 shows the formation of a typical oxyferryl compound II after mixing 2 M ferrous Y249F with 100 M hydrogen peroxide. In the Soret region, a clear isosbestic point was at 425 nm. Compound II was built within 90 ms (Fig. 4B), but it was very unstable and transformed to compound III. Y249F compound II exhibited absorbance maxima at 418, 530, and 558 nm, exactly the same that were observed when Y249F was incubated with a very small excess of H 2 O 2 (Ͻ10 M) (13). Due to the instability of Y249F, the reaction was biphasic with about 90% of the changes in absorbance corresponding to the first and faster phase, i.e. the conversion of ferrous Y249F to compound II. By using a single-exponential equation (Fig. 4B) and plotting the pseudo-first-order rate constants versus the concentration of hydrogen peroxide (Fig. 4C), an apparent second-order rate constant of (2.4 Ϯ 0.3) ϫ 10 5 M Ϫ1 s Ϫ1 (pH 7.0, 25°C) was calculated. The finite intercept of 2.5 s Ϫ1 underlines that compound II is unstable and is subjected to a further redox transition mediated by H 2 O 2 .
Both the spectral transition and the kinetics of the twoelectron oxidation of HRP with H 2 O 2 are very similar (not shown). The calculated apparent bimolecular rate constant was determined to be (1.1 Ϯ 0.2) ϫ 10 5 M Ϫ1 s Ϫ1 (pH 7.0, 25°C) with an intercept of 0.3 s Ϫ1 . These findings underline the high similarity between the KatG variant Y249F and HRP. Not only are the kinetics of compound I and compound II formation and the UV-Vis spectra nearly identical, but in addition, reduction of both redox intermediates with ascorbate and tyrosine follow the same kinetics (13), apart from the most important fact that the catalatic activity of both enzymes is negligible.
Completely different is the H 2 O 2 -mediated oxidation of wildtype ferrous KatG. In contrast to the Y249F, no ferryl-type compound II is formed. This fits well with previous observations that wild-type KatG compound II exhibits spectral features similar to ferric KatG (8,9,13). Fig. 5A shows the spectral changes obtained in the reaction of 3 M ferrous wild-type KatG with 50 M hydrogen peroxide. The reaction again exhibited a clear isosbestic point at 421 nm. The bold spectrum depicts a spectrum similar (but not identical) to the ferric enzyme in the Soret region (408 nm) but not at higher wavelength (bands at 540 and 580 nm), indicating that some portion of the protein was already in the compound III state. The bold spectrum was selected 880 ms after mixing ferrous wild-type KatG with H 2 O 2 . It finally converted to the spectrum of the ferric state (data not shown). The reaction of ferrous wild-type KatG with H 2 O 2 was biphasic with the first and faster phase responsible for about 80 -90% of the decrease in absorption at 440 nm (Fig. 5B). By fitting the first phase using a single-exponential equation (Fig.  5B) and plotting the k obs against hydrogen peroxide concentration (Fig. 5C), an apparent bimolecular rate constant of (2.0 Ϯ 0.4) ϫ 10 5 M Ϫ1 s Ϫ1 was calculated. Although this rate corresponds to that of both the Y249F variant and HRP, the spectral transition is completely different. Neither a ferryl-type compound II spectrum nor a pure protein radical compound II spectrum, which should exhibit spectral features similar to ferric KatG (Fig. 5A), could be monitored. It is evident that it transforms to the compound III spectrum very fast. This can be inferred by inspection of Fig. 5D, which shows the reaction of ferrous wild-type KatG with a higher excess of hydrogen peroxide (3 M enzyme and 300 M H 2 O 2 ). Within 40 ms, an intermediate appeared with spectral features representative of a mixture of at least two redox intermediates with compound III dominating (Fig. 5D, bold line), which, finally, reacted back to the ferric enzyme (broken line) within a few seconds.
The H 2 O 2 -mediated Conversion of Compound II to Compound III-Both Figs. 4 and 5 already indicate that KatG compound II readily converts to compound III in the presence of hydrogen peroxide. The measurements suggested that this transition is much faster in KatGs than in any other plant-type peroxidase. To determine the actual bimolecular rate constant, the sequential stopped-flow mode was used. Fig. 6A shows the direct conversion of Y249F compound II to compound III. The fast reaction was monophasic (Fig. 6B), and the bimolecular rate constant was (1.2 Ϯ 0.2) ϫ 10 4 M Ϫ1 s Ϫ1 at pH 7.0, which is much higher than that reported for other heme peroxidases like HRP (20 M Ϫ1 s Ϫ1 (21)), lignin peroxidase (70 M Ϫ1 s Ϫ1 (22)), or myeloperoxidase (78 M Ϫ1 s Ϫ1 (20)), respectively.
The elucidation of the kinetics of the reaction of ferrous wild-type KatG with H 2 O 2 faced several problems. Only with low concentrations of H 2 O 2 (Ͻ50 M) could a shift in the Soret region from 438 to 408 nm be observed, whereas at higher concentrations, the spectral features of compound III began to dominate within milliseconds (compare Fig. 5, A and D), suggesting that the reaction sequence is also ferrous wild-type KatG 3 compound II 3 compound III. The slowest rate in this transition was estimated to be Ϸ2.0 ϫ 10 5 M Ϫ1 s Ϫ1 . With respect to the spectral features of compound II, this investigation underlined that significant differences exist between wildtype KatG and Y249F, the latter corresponding to a typical ferryl-type (low spin) compound II like in HRP, whereas the spectral features of wild-type compound II suggests the presence of a high spin compound II (ferric protein plus protein radical). It has to be noted that the unexpected high rate of compound II to compound III conversion was very similar to the rate reported for the H 2 O 2 -mediated transition of compound II to compound III of the monofunctional bovine liver catalase (6.1 ϫ 10 4 M Ϫ1 s Ϫ1 ) (23).
Relevance for Catalase Activity and Proposal of the Catalatic Mechanism-In monofunctional catalases, the catalatic pathway is thought to include Reactions 1 and 2 (Fig. 1). Depending on experimental conditions, inactivation of catalase activity of monofunctional catalases is assumed to occur due either to accumulation of a (ferryl-like) compound II ("slow" inhibition) or to formation of compound III ("rapid" inhibition) at higher peroxide concentrations. The catalatic mechanism of KatG is still under discussion. Assuming a mechanism similar to monofunctional catalases is one working hypothesis; however, many findings hold against it as follows. (i) Wild-type compound I produced with peroxoacetic acid reacts extremely slow with H 2 O 2 as demonstrated by the sequential stopped-flow technique (8,9). (ii) In wild-type Synechocystis catalase-peroxidase, the chemical nature of the intermediate referred to as conventional compound I was shown to be the superposition of the oxoferryl porphyrin -cation radical, the tryptophanyl radical, and the tyrosyl radical as demonstrated by our recent multifrequency EPR spectroscopy study (24). (iii) The observation of redox intermediates with unique features in their room temperature electronic absorption spectrum suggests the existence of alternative electronic structures of compound I (13,24).
An alternative mechanism could be very similar to that proposed for the human peroxidases myeloperoxidase (20) and lactoperoxidase (25), which includes Reactions 6, 10, and 11 in Fig. 1. When compared with KatG, the catalatic activity of myeloperoxidase and lactoperoxidase is low, but the present study revealed important differences in the rates of the H 2 O 2mediated oxidation of compound II to compound III. In KatGs, the rate constant for this reaction is 1.2 ϫ 10 4 M Ϫ1 s Ϫ1 and therefore at least 2 orders of magnitude faster when compared with myeloperoxidase (78 M Ϫ1 s Ϫ1 ) (20)  KatG and could demonstrate that the spectral features of wildtype KatG, Y249F, and HRP compound III are very similar but that there are substantial differences in the stability of the corresponding dioxygen complexes. Furthermore, the two-electron oxidation of ferrous KatG by H 2 O 2 unequivocally demonstrated significant differences between wild-type KatG and Y249F (or HRP). In wild-type KatG (and in contrast to Y249F and HRP), no oxoferryl-type compound II accumulates in detectable concentrations, which confirmed earlier observations in the reaction between compound I preformed with peroxoacetic acid and one-electron donors (9,10,13), which suggested that KatG compound II could be a protein radical species with spectral features similar to ferric KatG.
The present data also suggest that in wild-type KatG, oxoferryl (compounds I and II) and dioxygen adducts (compound III) are destabilized when compared with Y249F and HRP. One significant structural difference between wild-type KatG and Y249F is that in the variant, the KatG-specific covalent adduct (which includes Tyr-249, Met-275, and Trp-122) at the distal heme side is absent (18). We now know three striking consequences; the catalase activity in Y249F is lost (whereas the peroxidase activity is unaffected), and in the presence of H 2 O 2 , an oxoferryl-type compound II is formed and easily converts to compound III (13,24). By contrast, in the presence of the covalent link, the catalase activity is high, and neither an oxoferryl-type compound I nor II accumulates, nor does compound III. Based on these data, we propose the following mechanism that has the stoichiometry of a classical catalatic activity ( In the first reaction, the ferric enzyme is oxidized to compound I, thereby reducing H 2 O 2 to water. Recent EPR experiments have shown the contribution of three different oxoferryl species obtained by reaction of wild-type KatG with peroxoacetic acid, namely an exchange-coupled porphyrin radical, a tryptophanyl, and a tyrosyl radical (13,24). Upon using H 2 O 2 in enzyme oxidation, neither the accumulation of the classical compound I (porphyrin radical-type, which typically shows a hypochromicity in the Soret peak) nor the accumulation of a protein radical species (which should exhibit an oxoferryl-type compound II spectrum like in cytochrome c peroxidase (27)) could be ever observed spectroscopically (in contrast to Y249F). This suggests a rapid conversion of compound I by the second hydrogen peroxide molecule. Based on previous findings that wild-type compound I, which accumulates with peroxoacetic acid and exhibits a 40 -50% hypochromicity in the Soret region, reacts only very slowly with H 2 O 2 , we proposed that the protein radical form has a higher reactivity toward H 2 O 2 (24). Mechanistically, the reaction of this species with H 2 O 2 would be similar to that of an oxyferryl compound II with H 2 O 2 , namely displacement of the ferryl oxygen by both oxygen atoms deriving from hydrogen peroxide as has been demonstrated for the compound II to compound III conversion of lactoperoxidase by isotopic labeling (28). The present study has demonstrated that this reaction is at least 2 orders of magnitude faster in KatGs (k app of 1.2 ϫ 10 4 M Ϫ1 s Ϫ1 ) when compared with other heme peroxidases (e.g. HRP 20 M Ϫ1 s Ϫ1 (21) or myeloperoxidase 78