Formation of Compound I in the Reaction of Native Myoglobins with Hydrogen Peroxide*

Reaction of ferric native myoglobin (Mb) with hydrogen peroxide (H2O2) was studied by the aid of stopped-flow rapid-scan spectrophotometry. In contrast to the results in previous studies where compound I was reported to be undetectable, both sperm whale and horse heart metmyoglobins (metMbs) formed a significant quantity of compound I, an oxoferryl porphyrin π-cation radical (Por+-FeIV(O)), during their reactions with H2O2. With both kinds of Mbs, formation of compound I was more clearly observed in D2O than in H2O. The compound thus formed was capable of performing monooxygenation of thioanisole to methyl phenyl sulfoxide and a 2-electron oxidation of H2O2 giving O2 and H2O as products. It was also converted into ferryl myoglobin (Por-FeIV(O)-globin+) spontaneously. Rate constants for these reactions and that for a direct conversion of metMb to ferryl Mb through the homolysis of H2O2 were determined. These results established unambiguously that native metMb can form both compound I and ferryl Mb upon reaction with H2O2 and that these high valent iron compounds serve as essential intermediates in Mb-assisted peroxidative reactions. The observed deuterium effect on the apparent stability of compound I was attributable to that effect on the hydrogen abstraction step in the 2-electron oxidation of H2O2 by compound I.


REACTION 1
Por ϩ -Fe IV (O)-globin 3 Por-Fe IV (O)-globin ϩ REACTION 2 where metMb is first oxidized by H 2 O 2 to compound I (Por ϩ -Fe IV (O)-globin) (Reaction 1), which is known to occur in the reactions of various other heme proteins such as peroxidases and possibly in the catalytic cycle of cytochromes P450. The oxidizing equivalent on the porphyrin -cation radical is then transferred to the globin moiety of Mb to produce ferryl Mb (Reaction 2), which oxidizes organic compounds such as lipids and styrenes. On the other hand, oxidation of certain other organic compounds such as sulfides by metMb has been postulated to proceed via a direct oxygen transfer from compound I in a similar way to those found in peroxidase-catalyzed monooxygenation reactions and probably in the catalytic cycle of cytochromes P450 (Reaction 3) (6 -9).
Por ϩ -Fe IV (O)-globin ϩ R 3 Por-Fe III -globin ϩ RO REACTION 3 Despite such formulations, however, the mechanisms for the peroxidative reactions catalyzed by metMb remained ambiguous in that the key intermediate in the above formula, compound I (Por ϩ -Fe IV (O)-globin), has never been isolated nor observed as a discernible entity. In this connection, Watanabe and co-workers (7,8) recently prepared a series of site-directed mutants of sperm whale metMb, in which the distal histidine (His-64) was replaced by a variety of different amino acid residues such as alanine, leucine, and serine, and they observed the formation of compound I of these mutant metMbs during the reaction with m-chloroperbenzoic acid. With native metMb, however, it was not possible to observe the formation of compound I under the same experimental conditions; only the formation of ferryl Mb was detected. They interpreted these findings to indicate that compound I of native metMbs was formed but was immediately converted to a ferryl state by accepting a reducing equivalent from the globin moiety through the distal histidine, His-64, and therefore did not accumulate as a discernible entity. On the other hand, the transfer of a reducing equivalent from the globin moiety to the heme could be slower in mutants due to the absence of His-64.
By employing stopped-flow rapid-scan spectrophotometry, we examined here the possible involvement of compound I and other high valent iron compounds in the H 2 O 2 -assisted oxidative reactions of organic compounds catalyzed by native metMb. The results have shown that compound I of native Mb is formed during its reaction with H 2 O 2 and serves as an obligatory intermediate of peroxidative reactions. Further kinetic analyses have indicated that 1) compound I abstracts a hydrogen from H 2 O 2 during the 2-electron oxidation of H 2 O 2 , 2) apparent stability of compound I is therefore mainly determined by its reaction with H 2 O 2 , and 3) native metMb carries out both homolysis and heterolysis of H 2 O 2 .

EXPERIMENTAL PROCEDURES
Materials-Sperm whale and horse heart metMbs were purchased from Sigma. The metMb dissolved in 10 mM potassium phosphate buffer, pH 7.4, was treated with a small amount of potassium ferricyanide to ensure the oxidation of ferrous Mb remaining in the commercial preparations. MetMb was extensively dialyzed against the same buffer and was purified by CM-cellulose column chromatography. D 2 O (99.9 atom %) was purchased from Aldrich and used without further purification. The pD value of the D 2 O buffer was determined according to the relation, pD ϭ pH obs ϩ 0.39, where pH obs is a reading of an ordinary pH meter which employs a glass electrode (20). All other reagents used in this study were of highest grade commercially available.
Stopped-flow Rapid-scan Spectrophotometry-Changes in optical absorption were measured by using an RSP-601 stopped-flow rapid-scan system (Unisoku Co., Ltd., Osaka, Japan) (21). The dead time of the measurements with the system was determined to be 5 ms by employing a pseudo first-order reaction of 2,6-dichlorophenol-indophenol with ascorbic acid (21).
Determination of Rate Constants-As described under "Results," absorption changes observed during the reaction of metMb with H 2 O 2 in the presence and absence of thioanisole were found to proceed in a biphasic manner in both H 2 O and D 2 O buffer systems. We therefore performed least squares fits of the absorption changes for each experiment according to Equation 1 which is given for describing such biphasic absorption changes with two apparent rate constants, k 1 obs and k 2 obs (22,23); where A(t) is absorption at a fixed wavelength and a time point t, and c 1 , c 2 , and c 3 are time-independent (and -dependent) fitting parameters.
The rapid-scan spectrophotometric system described above allowed us to record up to 512 time profiles of absorption at each run of the stopped-flow experiments, from which 126 time profiles at every 1.62 nm from 322 to 525 nm were selected, and they were respectively subjected to the least squares fits so as to optimize a single set of k 1 obs and k 2 obs values. Numerical calculations for the least squares fits were carried out by the aid of the techniques of singular value decomposition as described previously (21)(22)(23). A computer software used for the calculations was described elsewhere (21).
Rate constants for individual reaction steps, such as the decay of compound I to ferryl Mb, are not equivalent to k 1 obs or k 2 obs , but their values are obtainable from those of k 1 obs and k 2 obs according to kinetic equations shown below. Mathematical procedures for developing these kinetic equations are described under "Appendix." Monooxygenation of Thioanisole-Ten M metMb (solution A) and a mixture of 100 mM H 2 O 2 , 400 M thioanisole, and 0.8% methanol (solution B) were prepared in either the H 2 O or D 2 O buffer (200 mM potassium phosphate), pH (pD) 6.0, respectively. Then solutions A and B were mixed together with a 1:1 ratio and allowed to react for 1 min at 25°C. Within 1 min, the reaction ceased because of a complete conversion of metMb to ferryl Mb. Then, cumyl alcohol was added to the reaction mixture as an internal standard for the analyses on high pressure liquid chromatography. An aliquot of the reaction mixture was directly loaded on a high pressure liquid chromatographic system (Jasco, model TRI) equipped with a reverse-phase column (Waters Bondasphere 5 m CN-100 Å) and eluted with a 10% (v/v) methanol/ water solution at a flow rate of 0.55 ml/min. The eluent was monitored at 210 nm. Other details were described under appropriate figure and table legends. Fig. 1 shows changes in absorption spectra of sperm whale metMb in both Soret and visible regions during the reaction of metMb with 50 mM H 2 O 2 at pH 6.0, 25°C. The spectrum of metMb ( max ; 409, 505, and 634 nm) changed to that of ferryl Mb ( max ; 421, 551, and 586 nm) (15) in 200 ms without giving any spectral intermediate distinct from the two species (Fig. 1, A and B). However, we noticed that the isosbestic points between the spectra of metMb and ferryl Mb around 417, 616, and 667 nm were dubious, suggesting that one or more other spectral species could be involved in the reaction. Such an interpretation was further supported by the biphasic curves obtained for the time-dependent changes in absorption at 417, 616, and 667 nm in the insets of Fig. 1, A and B. Hence, we carried out the next series of experiments in a D 2 O medium, hoping that the use of D 2 O would make the other spectral species prominent since D 2 O frequently affects the rates of chemical reactions through solvent isotope effects (24) and sometimes helps to detect unstable reaction intermediates. Fig. 2 shows Soret and visible absorption changes in the reaction of sperm whale metMb with 50 mM H 2 O 2 in a D 2 O buffer containing 200 mM potassium phosphate, pD 6.0, at 25°C. Spectra at 0 ms (broken lines in Fig. 2, A and B) are those of metMb, and those at 0.5 s (thin lines) were attributable to the ferryl form. At 50 ms, the Soret absorption at 409 nm had decreased to almost a half (solid thick line in Fig. 2A) and accompanying increases in absorption in the visible region with prominent peaks at ϳ550, ϳ590, and 648 nm (solid thick line in Fig. 2B). Among the peaks in the visible region, one at 648 nm (marked by an asterisk) was unique to this intermediate spectrum not being found in the spectrum of metMb nor of ferryl Mb, whereas the other two were attributable to those of ferryl Mb at 551 and 586 nm. The Soret absorption maximum with a reduced intensity and a visible band around 650 nm are characteristic to compound I of other protoheme proteins such as horseradish peroxidase and catalase (25)(26)(27) and mutant Mbs (8). The spectral changes observed in the H 2 O buffer system ( Fig. 1) can also be explained by assuming the formation of the same species; the profiles of the time-dependent changes in absorption near the isosbestic points between metMb and ferryl Mb were all biphasic being similar to each other in H 2 O and D 2 O systems. Formation of a new spectral species with similar spectral characteristics was also observed with horse heart metMb both in H 2 O and D 2 O systems (data not shown).

Detection of a New Spectral Species-
Reaction of the New Spectral Species with Thioanisole-In an attempt to characterize further the properties of the new spectral species, we examined its reaction with thioanisole (CH 3 -S-C 6 H 5 ); it has been shown that compound I of heme proteins (5)(6)(7)(8)(9) transfers an oxygen atom to thioanisole yielding methyl phenyl sulfoxide as the product. When the reaction of metMb with H 2 O 2 was carried out in the presence of thioanisole, methyl phenyl sulfoxide (CH 3 -SO-C 6 H 5 ) was formed both in H 2 O and D 2 O buffer systems in significant quantities (Table I). As seen, the amount of the product formed in D 2 O (44 nmol) was about twice as that in H 2 O (19 nmol). Formation of other products was negligible, if present. A similar result was also obtained with horse heart metMb (see also Table I). These results, together with the spectral evidence described above, indicate that the new spectral species is compound I of native myoglobin.
We then carried out stopped-flow rapid-scan experiments under varying concentrations of thioanisole in the D 2 O buffer system. When formation and decomposition of the spectral species were followed at 418 nm, the absorption first decreased and then increased, indicating that the spectral species had accumulated transiently and then decreased (inset of Fig. 3, broken line). As seen, the decrease in absorption became more shallow as thioanisole concentration was raised, suggesting that the new spectral species decomposed more rapidly as thioanisole concentration increased. This finding has further supported the above interpretation that the new spectral species is compound I; it decomposes by donating an oxygen atom to thioanisole giving the sulfoxide as product. In the following experiments, therefore, we determined the rate constants for the reaction between compound I and thioanisole.
Employing the rapid-scan spectrophotometric system and methods described under "Experimental Procedures," we performed kinetic least squares fits of the time profiles of the reaction to a double-exponential function, and we obtained k 1 obs and k 2 obs , where k 1 obs and k 2 obs denote apparent rate constants for describing the absorption changes. Then theoretical time profiles of absorption at 418 nm which were drawn by using k 1 obs and k 2 obs values thus obtained were overlaid on the experimental curves in the inset of Fig. 3 (solid line). As seen, theoretical curves agreed well with experimental data. Experimental time profiles at every other wavelengths examined (322-525 nm) were all well reproduced when the same k 1 obs and k 2 obs values as above were applied (data not shown), ensuring the validity of our kinetic analyses. When we plotted the parameter k 1 obs ϩ k 2 obs against thioanisole concentrations, the plot showed a linear increase with increasing concentration of thioanisole ( Fig.  3), where the slope corresponds to the bimolecular rate constant (k S ) between compound I and thioanisole (see "The General Formula for Reaction Systems Involving Three Spectral Species" under "Appendix"). Values of k S thus obtained were listed in Table I. Based on all these results so far described, we conclude that the new spectral species found in this study is compound I of native myoglobin.
Overall Reaction Scheme for the Reaction of MetMb with H 2 O 2 -Since it is now clear that the new spectral species is compound I, the scheme for the present reaction system can be given by Scheme I, where k 12 and k 13 are both pseudo firstorder rate constant with respect to H 2 O 2 for the formation of compound I and ferryl Mb from metMb under the experimental conditions. The latter conversion of metMb to ferryl Mb was suggested to proceed through a homolytic cleavage of H 2 O 2 by metMb (Reactions 4 and 5) (5, 28).

TABLE I Monooxygenation of thioanisole by compound I of myoglobin
The amount of methyl phenyl sulfoxide, which is the product of the monooxygenation of thioanisole by compound I, was determined for 1-min incubation of 1 ml of reaction mixture, at 25°C, pH (pD) 6 Then k 23 is the first-order rate constant for the conversion from compound I to ferryl Mb, and k 21 is the rate constant for a backward reaction from compound I to metMb. The backward reaction proceeds when an oxygen acceptor molecule such as thioanisole is present in the reaction system (Reaction 3) (6 -9). It should be noted that H 2 O 2 that converts metMb to compound I also causes a decomposition of compound I according to the following reaction.
Por ϩ -Fe IV (O)-globin ϩ HOOH 3 Por-Fe III -globin ϩ HOH ϩ O 2 REACTION 6 The above reaction, which resembles those found in catalytic cycles of catalase and certain peroxidases (25)(26)(27), was already observed for compound I of the His-64 mutant Mbs (8,9). It is a catalase-like reaction that involves a 2-electron oxidation of H 2 O 2 to dioxygen (O 2 ) and water (H 2 O). By using an oxygen electrode, we were able to find that a detectable amount of O 2 was evolved during the reaction of metMbs with H 2 O 2 , indicating that the reaction of compound I with H 2 O 2 actually proceeded in the present reaction system (data not shown).
As described previously, the least squares fits of absorption time profiles observed for the present reaction system have given two apparent rate constants, k 1 obs and k 2 obs . When specific oxygen acceptors (sulfides etc.) are absent and H 2 O 2 is excess as compared with Mb, these rate constants can be related to rate constants involved in Scheme I according to Equations 2 and 3 below, where kЈ 12 (Fig. 4, A and B). These linear dependences were reproducible with horse heart metMb (Fig. 4, C and D). These findings indicate that kЈ 21 is not negligible (see relationships I-III described above). In other words, the reaction between compound I and H 2 O 2 (Reaction 6) proceeded with a significant rate in the reaction system under the present conditions. From the plots of k 1 obs ϩ k 2 obs versus [H 2 O 2 ] (Fig. 4), we also estimated values for k 23 ; the intercept of the plots on the k 1 obs ϩ k 2 obs axis corresponds to k 23 . The rate constants thus obtained are listed in Table II.
As described earlier, the least squares fits of the absorption time profiles employed in this study were performed according to Equation 1, where parameters c 1 , c 2 , and c 3 , respectively, vary with . We then developed kinetic equations as under "Derivations of Equations 4 and 5" under "Appendix," and we 2 T. Matsui, S. Ozaki, and Y. Watanabe, personal communication. have expressed relationships between these fitting parameters and the rate constants by using another set of parameters, l 1 and l 2 , as follows in Equations 4 and 5, where [metMb] 0 is the initial concentration of metMb used for the stopped-flow experiments; d is the length of the light pass adopted for the optical measurements, and ⑀ cpdI is the molar extinction coefficient of compound I at wavelength. These equations indicate that the value of kЈ 21 is obtainable according to Equation 5, whereas those of l 1 and l 2 are given by resolving Equation 4. Among the values of other parameters in Equations 4 and 5, however, those of ⑀ cpdI at any were not determined exactly in the present study because either metMb or ferryl Mb contributed to the observed absorption changes in the spectrophotometric detection of compound I (see Fig. 2B, for example). In the case of the H64A mutant of sperm whale Mb, on the other hand, the values of ⑀ cpdI were determined in a wavelength region from 360 to 710 nm. 2 We therefore applied the ⑀ cpdI values of H64A mutant to the present kinetic analyses for native sperm whale Mb under the assumption that the molar extinction coefficients of compound I of native Mb were very similar to those of the H64A mutant, if not entirely the same. Substituting the ⑀ cpdI values at every 10 nm from 360 to 460 nm into Equation 4, we obtained a set of simultaneous equations for l 1 and l 2 , and then a numerical solution of l 1 and l 2 was derived. The value of kЈ 21 was then obtained by substituting the l 1 and l 2 values, together with that of k 23 , into Equation 5 (Table II).
As already described, the parameter k 1 obs ϫ k 2 obs becomes roughly equivalent to kЈ 21  As we noted under "Relationships between k 1 obs and k 2 obs and kЈ 12 , kЈ 13 , and kЈ 21 at a Higher Concentration of H 2 O 2 " under "Appendix", it was a good approximation to assume the above equation when the k 1 obs and k 2 obs values at 100 mM H 2 O 2 (Fig. 4,  A and B)  All these results are depicted in Table II. Corresponding data for horse heart Mb were also derived by the same procedure by adopting the ⑀ cpdI values of the H64A mutant except for kЈ 12 and kЈ 13 values in D 2 O since it could not be a good approximation to assume Equation 6 in this case (see "Relationships between k 1 obs and k 2 obs and kЈ 12 , kЈ 13 , and kЈ 21 at a Higher Concentration of H 2 O 2 " under "Appendix").
Accordingly, we were able to describe all the rate constants in this reaction system, i.e. the reaction of native metMb with H 2 O 2 in the presence and absence of an oxygen acceptor such as thioanisole, indicating that Scheme I, which has long been suggested for H 2 O 2 -assisted reactions by Mb, is indeed correct. Based on the rate constant values given for the Scheme I, chemical mechanisms of the individual reaction steps are discussed under "Discussion." DISCUSSION It has long been postulated that oxidative reactions catalyzed by metMb using H 2 O 2 as the oxidant proceed through the formation of compound I as an essential reaction intermediate (4 -9). Up to now, however, formation of compound I has never been experimentally proven in the reaction system of native Mbs with H 2 O 2 . Evidence for the formation of compound I during the reaction of metMb with H 2 O 2 reported here comprises the following: 1) detection of the UV-visible absorption spectra typical to compound I of a protoheme protein; 2) decomposition of the spectral species upon reaction with thioanisole, yielding methylphenyl sulfoxide; and 3) all the experimental findings observed here satisfactorily fit to Scheme I which assumes the formation and decomposition of compound I in the reaction. Rate constants for all the component reactions in the scheme have been successfully determined.
Among the rate constants for the reactions in Scheme I, those for kЈ 21 showed the largest D 2 O isotope effect as depicted in Table II, where the values in D 2 O were 6 and 11 times smaller than those in H 2 O for sperm whale and horse heart Mbs, respectively. Since kЈ 21 is the rate constant for the degradation of compound I by H 2 O 2 , such a decrease in kЈ 21  ing why the accumulation of compound I was more easily seen in D 2 O rather than in H 2 O. No such a remarkable isotope effect was observed for the rate constant of other reactions in the scheme (Table II). Then, from a mechanistic point of view, the observed isotope effect on kЈ 21 suggests that compound I abstracts either a proton, hydrogen atom, or hydride ion from H 2 O 2 when it performs the 2-electron oxidation of H 2 O 2 according to Reaction 6.
In peroxidase-catalyzed reactions, the O-O bond of H 2 O 2 is known to be cleaved in a heterolytic fashion (29). For the cleavage, a positive charge on an arginine residue present near the heme in the distal side (Arg-38 in horseradish peroxidase and Arg-48 in cytochrome c peroxidase) has been suggested to assist the heterolysis (29 -31). On the other hand, hydrophobic amino acid residues such as Phe-43 and Val-68 occupy the distal space in Mb (32) resulting in a lack of amino acid residue(s) analogous to the arginine residue of peroxidases. Such a deficit in a charged residue(s) in metMb is in accordance with the view that the homolytic cleavage of H 2 O 2 occurs at least in part in Mb-assisted peroxidative reactions (3)(4)(5)(6)(7)(8)(9)28). However, the ratio of the homolysis to heterolysis (homo/hetero) has not been hitherto reported, because there was no way to measure the ratio. In this regard, we were able to obtain the values for kЈ 13 and kЈ 12 which correspond to the rate constants for the homolytic and heterolytic cleavages of H 2 O 2 , respectively. Hence the homo/hetero ratio corresponds to kЈ 13 /kЈ 12 . When the kЈ 13 and kЈ 12 values in Table II were applied, the ratio was found to be 2.7 and 3.5 in H 2 O for sperm whale and horse heart Mbs, respectively. We were not able to exclude, however, the possibility that the homolytic cleavage also contributed to the kЈ 12 values in part; when the hydroxyl radical formed via the homolysis (Reaction 4) oxidizes not only the globin moiety but also the heme, it may also generate compound I. If such an oxidation reaction was partly involved, the homo/hetero ratio determined here has been underestimated since kЈ 13 and kЈ 12 were determined in this study as the rate constants for the formations of ferryl Mb and compound I, respectively. It should be also noted that the data of kЈ 12 in Table II had somewhat larger uncertainties. Nevertheless, the present results in Table  II (kЈ 13 , 3.9 -4.9 Ϯ 0.5-0.7 ϫ 10 2 M Ϫ1 s Ϫ1 ; kЈ 12 , 1.1-1.8 Ϯ 1.5 ϫ 10 2 M Ϫ1 s Ϫ1 ) clearly indicate that the homolytic cleavages indeed proceeded at significant rates which were in a comparable order of magnitude as compared with that of the heterolysis.
The results of previous studies (4,7,8,10) have suggested that His-64 plays a crucial role in the conversion of compound I to ferryl Mb (Reaction 2). As mentioned earlier, Watanabe and co-workers (7,8) prepared several mutants of sperm whale Mb in which His-64 was replaced by an aliphatic amino acid such as alanine, serine, and leucine. When the reaction of such mutant Mbs with mCPBA was studied, they observed an accumulation of compound I that was not detected with native metMb. They interpreted the findings to indicate that the role of His-64 in the reaction is to assist a quick conversion of compound I to ferryl Mb, and hence its replacement with other kinds of amino acid made k 23 smaller resulting in a slower conversion. No evidence was available, however, on whether or not the amino acid substitution indeed affected the rate of conversion; the rate constants (k 23 ) for the mutant Mbs were reported to be all about 1 s Ϫ1 at 5°C (8), whereas no data were available for the native metMb until this study. As seen in Table II, we obtained k 23 values of 4 Ϯ 6 and 3 Ϯ 7 s Ϫ1 at 25°C for sperm whale and horse heart metMbs, respectively. Although the S.D. values were large, we have assumed from these data the upper limit of the value to be 10 s Ϫ1 , which corresponds to about 2-3 s Ϫ1 at 5°C if one takes the effects of temperature into account in an ordinary way. Thus it is difficult to judge whether such a difference in k 23 values, 1:2-3, is significant enough to explain the accumulation of compound I. As will be discussed below, the substitution of His-64 may affect other factors such as the bimolecular rate constant for the reaction between mCPBA and mutant Mbs.
The values of k S , the bimolecular rate constant for the monooxygenation of thioanisole by compound I, were 2.7 ϫ 10 4 and 3.0 ϫ 10 4 M Ϫ1 s Ϫ1 for sperm whale and horse heart Mbs, respectively (Table I). These values were smaller by 2 orders of magnitude than 1.5 ϫ 10 6 M Ϫ1 s Ϫ1 for H64A and H64S mutant Mbs (8). One of the possible causes for this difference is the steric hindrance offered by the side chain of His-64 which is greater in volume than the small side chains of Ala and Ser. If this is true, a similar difference in the bimolecular reaction rate of mCPBA with native and mutant Mbs can be expected to occur. The structure and the size of mCPBA (Cl-C 6 H 4 C(O)OOH) are not grossly different from those of thioanisole (C 6 H 5 -S-CH 3 ), and hence it is expected that His-64 can also hinder the access of mCPBA to the heme. Therefore, the replacement of His-64 by Ala or Ser possibly accelerates the reaction rate of mCPBA with metMb and in turn the formation of compound I of the mutant Mbs.
Finally a few words may be necessary for an additional reaction intermediate that had been expected to occur but was not detected in this study. Prior to the formation of compound I of heme proteins such as peroxidases, it has been considered that H 2 O 2 coordinates to a ferric heme to form a hydroperoxy (Fe III -O-O-H) intermediate termed compound 0 which undergoes a subsequent cleavage of the O-O bond in a heterolytic manner to give compound I (29,33). Furthermore, several lines of evidence have indicated that the distal histidine of peroxidases (His-42 of horseradish peroxidase and His-52 of cytochrome c peroxidase) acts as a base catalyst to abstract one proton from H 2 O 2 , accelerating the formation of compound 0 (29, 34, 35). Also for Mb, the formation of compound I has been indicated to require a precursor compound corresponding to compound 0 (Scheme II) (3-9), although such a compound has not been detected yet for the reaction of native metMb with H 2 O 2 .
Under the present experimental conditions, metMb showed a direct conversion to either compound I or ferryl Mb with rate constants k 12 or k 13 , and no accumulation of compound 0 was detected. This finding indicates that the formation of compound 0 is the rate-limiting step in the pathway in Scheme II. In other words, k 0 in the above pathway is much smaller than kЈ and kЉ. In such a case, k 12 and k 13 respectively correspond to k 0 kЈ/(kЈ ϩ kЉ) and k 0 kЉ/(kЈ ϩ kЉ), and therefore kЈ 12 ϩ kЈ 13 (Table II). The corresponding data for horse heart Mb were 3.6 ϫ 10 2 M Ϫ1 s Ϫ1 in D 2 O and 5.0 ϫ 10 2 M Ϫ1 s Ϫ1 in H 2 O, which were only slightly different from each other. These results indicate that, if the coordination reaction of H 2 O 2 to the heme involves a chemical step where a proton is abstracted from H 2 O 2 , it is not the rate-limiting one. It is likely that neither His-64 nor the other amino acid residues in Mb act as a base catalyst in the coordination of H 2 O 2 to the heme, at least under the present experimental conditions. Such a different feature of the distal histidine of Mb from that of peroxidases had already been suggested on the basis of the findings that the overall rate of the conversion from metMb to ferryl Mb upon the reaction with H 2 O 2 is 5 orders of magnitude smaller than the bimolecular reaction rate between peroxidases and H 2 O 2 (34). Further efforts to detect compound 0, if present, in the reaction of native metMb with H 2 O 2 are underway in this laboratory.
Acknowledgments-We thank Drs. Yoshihito Watanabe, Shin-ichi Ozaki, and Toshitaka Matsui of the Institute for Molecular Science (Okazaki, Japan) for helpful discussions.

APPENDIX
The General Formula for Reaction Systems Involving Three Spectral Species-For the kinetic analyses, we first calculated singular value decomposition (36) of the absorption changes observed for the present reaction system. Upon analyses of the calculated results of singular value decomposition (22), we made a conclusion that only one spectral species other than metMb and ferryl Mb contributed to the absorption changes under the experimental conditions (data not shown). Thus, we have started from Scheme III which is a general formula for reaction systems involving metMb, ferryl Mb, and one unknown species (X) as well as all the possible reaction steps among the three species.
The application of Scheme III into the present reaction system does not necessarily mean that the system contains only three chemical species. Each step represented by one rate constant in the scheme may consist of a few chemical steps involving further intermediates each of which converts to others with an intrinsic rate constant. However, such intermediates were invisible in the present reaction system, indicating that the reaction steps to form them were very slow as compared with those of metMb, ferryl Mb, and X in Scheme III. In such a case, the intrinsic rate constants representing the invisible chemical species degenerate to one rate constant, as we exemplified in "Discussion". Therefore, Scheme III, which involves only three species and six rate constants, is sufficient to describe the kinetics of the present reaction system, regardless of possible involvement of more chemical species in the system.
Reaction systems represented by Scheme III were categorized into the competitive-consecutive reactions with reversible reaction steps by Szabó (37). When all rate constants in Scheme III are first-or pseudo first-order, the reaction system of Scheme III approaches from the initial condition ([metMb] ϭ [metMb] 0 , [ferryl Mb] ϭ [ferryl Mb] 0 , and [X] ϭ [X] 0 ) to an equilibrium state by a first-order process with two apparent rate constants, k 1 obs and k 2 obs , which relate with the rate constants for the elementary steps as in the following equations (Equations 8 and 9), where S 1 and S 2 are further given by following Equations 10 and 11 (37). If X in Scheme III can decompose to metMb upon reacting with a sulfide, k 21 is given as a sum of two rate constants, k 21 s and k 21 i ; the former is a pseudo first-order constant for the reaction between X and the sulfide, and the latter is independent of the sulfide. In the presence of an excess amount of the sulfide, then k 1 obs ϩ k 2 obs is given by Equation Since k 1 obs Ͼ k 2 obs and ␣ 1 ϩ ␣ 2 ϭ k 23 , the difference in question is k 1 obs k 23 at most. Thus the uncertainty for the calculated kЈ 21 kЈ 13 values caused by assuming Equation 6 is at most k 1 obs k 23 /k 1 obs k 2 obs (ϭ k 23 /k 2 obs ). Based on the above considerations, we estimated the uncertainties that were brought about by the application of Equation 6 to the experimental data in Fig. 4. The values obtained were less than 11% for the calculated kЈ 21 kЈ 13 values (not shown) except for that derived from data in Fig. 4D which yielded a somewhat larger uncertainty (ϳ50%). The uncertainties of less than 11% were comparable to those caused by the present experimental errors in the k 1 obs k 2 obs values (ϳ10%).  On the other hand, absorption changes at wavelength for the present reaction system are given by Equation 28. where ⑀ met , ⑀ cpdI , and ⑀ ferryl are, respectively, the molar extinction coefficients of metMb, compound I, and ferryl Mb at wavelength, and d is the length of the light pass adopted for the optical measurements. Substituting the equations for [metMb], [compound I], and [ferryl Mb] into the above equation and rearranging terms, we can express A(t) as shown in Equation 29.