Direct detection of Fe(IV)=O intermediates in the cytochrome aa3 oxidase from Paracoccus denitrificans/H2O2 reaction

We report the first evidence for the formation of the “607- and 580-nm forms” in the cytochrome oxidase aa3/H2O2 reaction without the involvement of tyrosine 280. The pKa of the 607–580-nm transition is 7.5. The 607-nm form is also formed in the mixed valence cytochrome oxidase/O2 reaction in the absence of tyrosine 280. Steady-state resonance Raman characterization of the reaction products of both the wild-type and Y280H cytochrome aa3 from Paracoccus denitrificans indicate the formation of six-coordinate low spin species, and do not support, in contrast to previous reports, the formation of a porphyrin π-cation radical. We observe three oxygen isotope-sensitive Raman bands in the oxidized wild-type aa3/H2O2 reaction at 804, 790, and 358 cm–1. The former two are assigned to the Fe(IV)=O stretching mode of the 607- and 580-nm forms, respectively. The 14 cm–1 frequency difference between the oxoferryl species is attributed to variations in the basicity of the proximal to heme a3 His-411, induced by the oxoferryl conformations of the heme a3-CuB pocket during the 607–580-nm transition. We suggest that the 804–790 cm–1 oxoferryl transition triggers distal conformational changes that are subsequently communicated to the proximal His-411 heme a3 site. The 358 cm–1 mode has been found for the first time to accumulate with the 804 cm–1 mode in the peroxide reaction. These results indicate that the mechanism of oxygen reduction must be reexamined.

Establishing the structural and functional properties of the binuclear center in cytochrome c oxidase (CcO) 1 is essential in understanding the linkage of the structural features with the structure of transient O 2 intermediates that are important for the protein to function as a redox-linked proton pump (1)(2)(3)(4)(5)(6). CcO contains a homo-dinuclear Cu A center, one low spin heme a, and a high spin heme a 3 -Cu B binuclear center where the reduction of dioxygen to H 2 O takes place (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15). One of the unique properties of the binuclear heme a 3 -Cu B center that were determined by the crystal structures of the bovine, the Paracoccus denitrificans, and the Thermus thermophilus heme copper oxidases is the Tyr-280 -His-276 (P. denitrificans numbering) cross-linked structure (7)(8)(9)(10)(11). Tyr-280 is located at the end of the proton K-channel and is highly conserved among the heme copper oxidases. The suggestion that the O-O bond cleavage proceeds by concerted hydrogen atom transfer from the cross-linked His-Tyr species to produce the Fe(IV)ϭO/Cu B 2ϩ -His-Y ⅐ species raises important issues as to the means by which the electron transfer to these transient species is regulated for conformational transitions that are required to occur in any pumping mechanism (13,16).
The CcO/H 2 O 2 reaction has been a conundrum and subject of debate for a number of years (17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28). The reaction serves as an alternative view of the peroxy (2e Ϫ ) oxidation level. The CcO/ H 2 O 2 reaction, monitored extensively by optical absorption, forms consecutively the 607-and 580-nm forms, and their formation is dependent on the concentration of substrate and pH. Importantly, both forms were first observed in mitochondria during the reversal of electron transfer from water to cytochrome c (29). Recent studies have suggested that at high pH, addition of stoichiometric amounts of H 2 O 2 to oxidized enzyme leads to the formation of the 607-nm form having the Fe(IV)ϭO⅐⅐⅐⅐HO-Cu(II) B -Y ⅐ structure, whereas the 580-nm form having the Fe(IV)ϭO Cu(II) B -Y ⅐ structure is the result of a single protonation of the former species (14). At low pH, the 580-nm form is also produced by addition of stoichiometric amounts of H 2 O 2 to the enzyme (14). Addition of a second equivalent of H 2 O 2 converts each of the above mentioned species to Fe(IV)ϭO Cu B 2ϩ -His-Y. On the other hand, it has been reported that the formation of the 580-nm form, but not that of the 607-nm form, results in a radical-like EPR signal (24). The involvement of Cu B as the alternative electron source in the O-O cleavage process through its Cu B (III) oxidation state has been suggested (25). Recently, the broad component of the EPR spectrum of H 2 O 2 -treated oxidized bovine CcO was attributed to a tryptophan cation radical exchange coupled to the oxoferryl Fe(IV)ϭO state of heme a 3 , whereas the second observed signal to a porphyrin cation radical (28). In the later study, it was concluded that these EPR signals were not related to those demonstrating the presence of a tyrosine radical in the binuclear center (27) but were interpreted to arise from a tryptophan residue ϳ 9 Å away from heme a 3 .
Characterization of the Fe(IV)ϭO intermediates generated in the oxidized CcO/H 2 O 2 reaction and their pH and H 2 O 2 concentration dependence has been reported (21)(22)(23). It was demonstrated that under steady-state conditions, the peroxide reaction generates the 607-and 580-nm forms with 1 and 5 mM H 2 O 2 , respectively, and produces the same oxygen isotopesensitive RR modes as those found for the transient intermediates in the O 2 reaction with CcO (30 -40). In the pH 7.4 -10 range, the population of the 607-nm form with a characteristic (Fe(IV)ϭO) mode at 804 cm Ϫ1 decreased at higher pH and at high concentrations of H 2 O 2 (5 mM). Formation of the 580-nm form at non-saturating levels of H 2 O 2 yields two more oxygen isotope-sensitive Raman modes at 785 and 355 cm Ϫ1 . The first disappears above pH 8.5 and has been assigned to a second Fe(IV)ϭO species; the second disappears at pH above 9.0 without significant changes in the absorption spectrum and has been attributed to a third oxoferryl species that is not related to either of the Fe(IV)ϭO species observed at 785 and 804 cm Ϫ1 .
Recently, we demonstrated that the stability of the binuclear center is not compromised in the Y280H mutant oxidase lacking the Tyr-280 -His-276 cross-link since heme a 3 retains the same proximal environment, spin, and coordination state as in the wild-type enzyme in both oxidized and reduced states (41). As a result of the different experimental approaches in the oxidized CcO/H 2 O 2 reaction, there is no consensus on the structure of either the 607-or the 580-nm forms. The relationship between the 804 cm Ϫ1 and/or the 790 cm Ϫ1 RR modes with that at 355 cm Ϫ1 is also essential for the clarification of the structure of each intermediate. In the work presented here, we report the formation of the 607-and 580-nm forms and their associated RR bands at 804, 790, and 358 cm Ϫ1 in the reaction of oxidized wild-type aa 3 from P. denitrificans with H 2 O 2 under steady-state conditions at ambient temperatures. Our data indicate that in the pH 6.6 -7.4 range, the ϳ580-nm form is produced with a characteristic oxygen-sensitive mode at 790 cm Ϫ1 . In the pH 7.7-8.1 range, however, we obtain a mixture of both forms and two oxygen-sensitive RR bands at 804 and 358 cm Ϫ1 . Following the H 2 O 2 reaction for the Y280H mutant, we report the production of the 607-and 580-nm forms and demonstrate that Tyr-280 is obviously not involved in their formation. The mixed valence Y280H/O 2 reaction reveals similar characteristics of the mixed valence wild-type/O 2 reaction, which further confirms that the production of the 607-nm form occurs without the involvement of the Tyr-280 residue.

EXPERIMENTAL PROCEDURES
Wild-type and Y28H mutant CcO were purified from P. denitrificans according to published procedures (42,43). The activity of isolated wild-type and mutant CcO has been reported (43). The samples were concentrated to 100 -150 M in a desired buffer (pH 6.6, MES; pH 7.4 -8.1, HEPES; pH 9.5, CHES) containing 0.1% dodecyl ␤-D-maltoside and stored in liquid nitrogen until use. In the resonance Raman experiments, the enzyme solution and an H 2 O 2 buffer solution were placed in separate syringes and driven through two eight-jet tangential Gibson type mixers in series with a syringe pump (44 -48, 70). The exit port of the mixer was designed to allow the oxidase solution (50 M after mixing) to flow through a rectangular 2-cm-long flow cell with an internal 0.2 ϫ 0.1-mm cross-section. The time resolution is determined by the flow rate, the cell cross-sectional area and the spacing between the sample, and the probing 427-nm laser beam. With the flow rate and geometry used in the present investigation, we are able to vary the time from the mixer to the laser beam from 2 to 5 min. The resonance Raman spectra were acquired as described elsewhere (41). The power incident on the CcO sample was typically 5-6 milliwatt to avoid photoeffects of the sample by the incident laser light due to the low flow rate. H 2 O 2 was purchased from Aldrich, and H 2 18 O 2 was purchased from Icon. The optical absorbance spectra were recorded with a PerkinElmer Life Sciences Lamda 20 UV-visible spectrometer. Fig. 1A shows the optical difference spectra upon reacting 6 M oxidized CcO with 3 mM H 2 O 2 minus oxidized CcO at pH 7.7 at the indicated times. The spectra are characterized by two peaks at 586 and 610 nm. The progression of changes in the difference spectra indicates that maximum occupancy occurs at 10 min in the reaction. The difference spectra are similar to those obtained for the reaction of oxidized mammalian and Rb. sphaeroides aa 3 -type oxidase with H 2 O 2 (49,50). When this experiment is conducted with the Y280H mutant, the difference spectrum (trace b) is slightly changed with the maxima at 593 and 610 nm. The 610-nm form in both the wild-type and mutant enzyme indicates that its production is independent of the cross-link Tyr-280 -His-276 structure. Mixing of oxygen with the CO-bound MV oxidase yields long-lived oxygenated intermediate(s) that occur after the decay of the oxy-intermediate (41, 50 -52). The visible region optical absorption spectra of the wild-type and the Y280H oxygenated species formed this way are shown in Fig. 1B. The time evolution for the decay of the 608-nm form (trace a) formed in the wild-type MV/O 2 reaction is similar to that reported by others for the mammalian aa 3 MV/O 2 reaction (53). Moreover, the difference spectrum (MV-CO/O 2 minus oxidized) of the Y280H enzyme shown in Fig. 1B (trace b) displays a maximum at 608 nm, albeit with decreased absorbance, which is similar to that of the wild type. If the 608-nm form we detect in the Y280H MV/O 2 reactions has the same structure as that obtained from the oxidized wild-type and mutant enzyme in their respective reactions with H 2 O 2 , then the contribution of the covalently linked tyrosine to its formation is minimal. Fig. 2 shows the pH dependence of the 610-and 586-nm forms. The former is more populated at pH values above 7.7 but is replaced by the 586-nm form as the pH decreases. The effect of pH on the decay of the 610 -586-nm form, shown in the inset, indicates that the transition is controlled by a group with a pK a of 7.5. Since H 2 O 2 is fully protonated in the pH range 5-8.5, we suggest that the protonation state of a residue in the binuclear moiety controls the reaction product. When the group is protonated, the reaction yields the 586-nm form, and when it is deprotonated, the reaction yields the 610-nm form. The 610-nm form weakly absorbs at pH 6.6 and 7.4, and only the 586-nm form is apparent in the spectrum, whereas at pH 7.7-8.1, both forms are present. The accelerated at low pH conversion of the 610-to the 586-nm form, cannot be reversed by restoring the initial pH. Fig. 3 shows the high frequency RR spectra of oxidized enzyme and those of the reaction intermediates with H 2 O 2 in the pH 6.6 -9.6 range at 2 min subsequent to mixing under steadystate conditions at ambient temperatures. It should be noted that the steady-state RR conditions used in this study are similar to those of Proshlyakov et al. (23). Fig. 3 (trace g) shows the high frequency RR spectrum of the Y280H/H 2 O 2 reaction. The RR spectra of the oxidized enzyme at pH 6.6 (trace a), 7.7 (trace c), and 9.6 (trace e) are almost identical, indicating that no unusual pH effects modify either the oxidation state of heme a and a 3 or the heme a 3 ligand binding site. The disappearance of the high spin 2 at 1572 cm Ϫ1 and 10 at 1612 cm Ϫ1 during the reaction indicates the formation of oxygenated species similar to those reported for the bovine aa 3 /H 2 O 2 reaction. The absence of 2 and 10 at lower frequencies does not support the formation of a porphyrin -cation radical. In addition, the increased intensity at 1584 cm Ϫ1 of the 2 mode and that of the 10 mode of heme a 3 at 1638 cm Ϫ1 throughout the pH range indicate that both the 610-and 586-nm forms are low spin heme a 3 adducts. All the modes associated with the high spin heme a 3 3ϩ in the Y280H/H 2 O 2 reaction spectrum (trace g) are very similar to those of the wild-type enzyme. In particular, the high spin 2 at 1572 cm Ϫ1 and 10 at 1612 cm Ϫ1 have lost intensity during the reaction, indicating the formation of oxygenated species similar to those reported for the wild-type aa 3 /H 2 O 2 reaction. The residual high spin 2 at 1572 cm Ϫ1 arises from incompletely reacted heme a 3 . Thus, although the reported activity of the Y280H mutant is 1%, only a fraction of heme a 3 (15-20%) is not converted to six-coordinate low spin species in the peroxide reaction. In addition, the substantial heterogeneity observed in the CO-bound fully reduced mutant enzyme that is a consequence of the different conformations of Cu B is not observed in the peroxide reaction. This is clearly demonstrated by the great similarity in the bandwidth and band position of the 2 mode at 1584 and 10 mode at 1638 cm Ϫ1 in the wild-type (traces b, d, and f) and Y280H/H 2 O 2 (trace g) reactions.

RESULTS
In Fig. 4 (30 -39). Trace b shows the analogous RR spectra at pH 7.7. In this difference spectrum, two more oxygen-sensitive modes appear.
The first mode appears at 804/762 cm Ϫ1 , and the other appears at 358/342 cm Ϫ1 . Both modes have the same frequency as those observed in the reactions of fully reduced and mixed valence bovine aa 3 /O 2 reaction and oxidized CcO with H 2 O 2 (13, 30 -40). The 804 cm Ϫ1 mode has been assigned to a second heme a 3 Fe(IV)ϭO species, and the 358 cm Ϫ1 mode has been assigned to a His-FeϭO bending mode (23). DISCUSSION For the sake of comparison with earlier work, in the following discussion, we ascribe the 608 -610-and 586-nm forms we have detected in the reactions of P. denitrificans aa 3 3. High frequency resonance Raman spectra of wild type and those of the reaction intermediates with H 2 O 2 at 2 min subsequent to mixing in the pH 6.6 -9.6 range. Trace a, pH 6.6; trace b, the reaction product at pH 6.6; trace c, pH 7.7; trace d, the reaction product at pH 7.7; trace e, pH 9.6; trace f, the reaction product at pH 9.6; trace g, the RR spectrum of the Y280H/H 2 O 2 reaction at pH 7.7. The concentration of the enzymes was 50 M, and that of H 2 O 2 was 3 mM. The excitation laser wavelength was 427 nm, and the incident laser power was 5-6 milliwatt. The total accumulation time was 15 min for each spectrum. Trace a is at pH 6.6 (586-nm form), and trace b is at pH 7.7. The concentration of the enzymes was 50 M, and that of H 2 O 2 was 3 mM. The excitation laser wavelength was 427 nm, and the incident laser power was 5-6 milliwatt. The total accumulation time was 35-45 min for each spectrum. and H 2 O 2 to the species referred to in the literature as having 607-and 580-nm absorbance.
The Origin of the 607-and 580-nm Forms-In the oxidized CcO/H 2 O 2 reaction, the formation of the 607-nm form increases in the pH 7.7-8.1 range and depends on the experimental conditions. Its population increases as the pH increases in the protocol used by the present authors and by Fabian and Palmer (24,26), but not in that used by Proshlyakov et al. (23). The latter authors reported that in the pH 7.4 -10 range, the population of the 607-nm form decreases at higher pH, and the final distribution of the enzyme between the 607-and 580-nm forms at a given pH depends on the enzyme preparation. The pK a of ϳ7.5 we have measured for the 607-580-nm transition is consistent with that reported by Fabian and Palmer (26). However, these authors excluded the involvement of an amino acid residue with a pK a of ϳ7.5 being responsible for the 607-580-nm transition. They found that above neutral pH, their enzyme undergoes autoreduction with a rate similar to that of the decay of the 607-nm form, and the rate of autoreduction increases as the pH is raised. We have not observed such autoreduction in the pH 6.6 -9.6 range, as is evident in the optical and RR data, and thus, we suggest that the observed transition indeed reflects a pK a of ϳ7.5.
A variety of factors are expected to contribute to the differences in absorption maxima between the 607-and 580-nm species. The most likely rationale for the conversion of the 607to 580-nm form involves the hydrogen-bonding differences in the distal environment and variations in the basicity of the proximal to heme a 3 His-411. The large difference in wavelength maximum for the two forms, 25-30 nm, is unlikely to arise from a change in a single hydrogen-bonding interaction. It has been shown, for example, that the difference in a hydrogen-bonded (pH 7) and non-hydrogen-bonded (pH 11) Fe(IV)ϭO species in myeloperoxidase is only 7 nm (54). We suggest, instead, that a change in the distal H-bonded environment of the oxoferryl species can be communicated to the proximal environment, causing a modification of the H-bonding status of the proximal heme a 3 -His-411. This way, the optical transition will shift to higher energy as a result of the increased splitting of the d orbitals and the energy of the antibonding * orbitals. The data for the aa 3 -type oxidase from Bacillus subtilis (55) show that such heme-linked ionization phenomena exist in heme copper oxidases, which also occur for the peroxidase class of enzymes (56), and are consistent with our interpretation that the effect that comes from the unique interactions between the oxoferryl species and the distal heme pocket environment. Thus, the formation of the 607-nm species can communicate events to the proximal environment, causing perturbations in the H-bonding interactions of His-411.
The general consensus that in the CcO/H 2 O 2 reaction, the 607-nm form converts to another form, with a maximum wavelength, which varies between 575 and 593 nm, as a result of changes in the distal hydrogen-bonding interactions, is also consistent with our data (23). Although we detect the 607-nm form in both the wild-type and Y280H mutant enzyme reactions with H 2 O 2 , we observe a 7-nm red shift for the 580-nm transition in the mutant. It is well established that the tyrosine hydroxyl group is hydrogen-bonded to the farnesyl hydroxyl group of heme a 3 and that the cross-link between Tyr-280 and His-276 exerts a certain control on the position of Cu B (9,11,41). Moreover, the Fourier transform infrared studies of the CO-bound form of the Y280H mutant have demonstrated that in the absence of the cross-link Tyr-His, the heme a 3 -Cu B distance has changed, and Cu B is not fixed in a certain position, resulting in different Cu B conformations. Consequently, in the Y280H mutant, Cu B has moved further closer to the CO-bound heme a 3 and exerts stronger steric force on the bound ligands to heme a 3 (41). We suggest that the 593-nm form is an oxoferryl species, similar to the 586-nm form found in the wild-type but lacking the heterogeneous hydrogen-bonding interactions observed in the wild type. The observed 7-nm red shift in the Y280H mutant is consistent with the observed shift between an H-bonded and a non-H-bonded Fe(IV)ϭO species.
The data of the Y280H/H 2 O 2 and Y280H MV/O 2 reactions demonstrate that the 607-nm form is not linked to the redox properties of Tyr-280 and can be produced without its involvement to form a Fe(IV)ϭO/Cu B 2ϩ -His-Y ⅐ species, as recently proposed (13,14,16). It should be noted that Fabian and Palmer (24,25) reported the formation of 7% of a radical species in the oxidized CcO/H 2 O 2 reaction, whereas MacMillan et al. (27) reported a yield of 20% for a tyrosine radical species produced in the same reaction. Interestingly, density functional calculations have shown that in the covalent bond between N ⑀ of His-276 and C ⑀ of Tyr-280, the neutral radical of the tyrosine Y ⅐ could place the spin density of N ␦ of the histidine, and thus, coupling between the copper and the organic radical, Y ⅐ -Cu(II) B 3 Y-Cu B (III) might occur (57). This indicates that in the wild-type enzyme, Tyr-280, which is ideally positioned to act as an acid-base catalyst in both the dioxygen and peroxide reactions, may play a role in the formation of the 607-nm species only if a Y ⅐ -Cu(II) B 3 Y-Cu B (III) structure occurs.
Resonance Raman Assignment of the Intermediates-The assignment of the vibrational modes in cytochrome aa 3 from P. denitrificans is well established (41). The high frequency RR spectra reported here demonstrate that in the oxidized form of the enzyme and in the pH 6.6 -9.6 range, no autoreduction of either the heme a or the heme a 3 occurs, in contrast to the conclusions reported earlier (26). In the RR spectra of the CcO/H 2 O 2 reactions, the spin state marker bands 2 and 10 are of particular importance as they indicate the formation of low spin heme a 3 and exclude the formation of a porphyrin -cation radical. The structural identification of the 607-and 580-nm species during the CcO reaction cycle has been a matter of considerable debate. The properties of the intermediates that appear in the 785-805 cm Ϫ1 frequency range are complicated, and several interpretations have been presented (13, 22-23, 32-34, 39 -40). The formation of the oxoferryl species during the oxidized CcO/H 2 O 2 reaction shown in Fig. 4 addresses this issue and demonstrates that the 607-nm form gives rise to an oxoferryl species with a Fe(IV)ϭO stretching mode at 804 cm Ϫ1 , in agreement with the results of Kitagawa and co-workers (21)(22)(23), and an oxygen isotope-sensitive mode at 358 cm Ϫ1 (see below). The 586-nm species is a second oxoferryl heme a 3 with a Fe(IV)ϭO stretching mode at 790 cm Ϫ1 . This assignment is also in agreement with the results of Kitagawa (21)(22)(23). The assignment of the 358 cm Ϫ1 mode, however, is still controversial. Proshlyakov et al. (13) considered the Cu B -OH structure as a possible candidate for the 358 cm Ϫ1 mode, and both Kitagawa and co-workers (38) and Rousseau and co-workers (39) have assigned it as a His-FeϭO bending vibration of different distorted His-FeϭO species. Our results indicate that the 358 cm Ϫ1 mode correlates with the 804 cm Ϫ1 mode, giving the impression that they originate from the same species. However, based on the existing inconsistencies (38,39) in the assignment of the 358 cm Ϫ1 species, we cannot assign it at present. However, we do not observe the 358 cm Ϫ1 mode concurrently with the 790 cm Ϫ1 mode, and thus, we conclude, in contrast to the results of Rousseau and co-workers (39), that they arise from different species. In addition, the 358 cm Ϫ1 mode becomes broader only when D 2 O is used as a solvent, and thus, it is highly unlikely that the mode arises from a Cu B -OH species (34).
The Peroxy and Ferryl Species in the MV/O 2 and Oxidized/ H 2 O 2 Reactions-The formation of the 607-nm form in both the MV-wild type and MV-Y280H reactions with O 2 and in the wild-type and Y280H reactions with H 2 O 2 has a number of important implications. First, it establishes crucial links between the two different reactions. Second, in the 607-nm form, which is one electron more oxidized than the form at 580 nm, the additional oxidizing equivalent must be stored at a site other than heme a 3 or Tyr-280. Alternative structures for the 607-nm form include: 1) Fe(V)ϭO/Cu B 2ϩ , 2) Fe(IV)ϭO/Cu B 3ϩ , and 3) Fe(IV)ϭO/Cu B 2ϩ X ⅐ (X ⅐ ϭ W ⅐ or Y ⅐ ). We do not have evidence other than that related to the Y280H to either confirm or reject the presence of a radical species. However, we demonstrate that the 607-nm form produces the 804 cm Ϫ1 species and suggest that it arises from a Fe(IV)ϭO/Cu B 3ϩ species. A Fe(V)ϭO/Cu B 2ϩ structure has been proposed for the 804 cm Ϫ1 mode in the CcO/O 2 reaction, and very recently, evidence supporting the ϩ5 valence state in a non-heme Fe has been reported (58). As suggested by Kitagawa and co-workers (59), in the presence of an electron-withdrawing group attached to heme a 3 (formyl), the a 1u and a 2u level could be lower than the d xz and d yz orbitals. Thus, an electron is taken out of the iron orbital, producing the Fe(V) oxidation state, and the d xz and d yz orbitals may act as antibonding orbitals of the FeϭO bond. Consequently, removal of this electron raises the stretching frequency of the Fe(V)ϭO as compared with the Fe(IV)ϭO frequency. This can account for the appearance of the Fe(V)ϭO at 804 cm Ϫ1 and the Fe(IV)ϭO at 790 cm Ϫ1 . The observation of the 805 cm Ϫ1 mode in the cytochrome bo 3 /H 2 O 2 reaction (60), however, argues against this possibility since the heme O lacks a strong electron-withdrawing peripheral substituent (61). On the other hand, the occurrence of Cu 3ϩ has been demonstrated in model compounds, but supporting evidence in Cu-containing enzymes has not been reported so far (62,63). Given that a small percentage of radical species has been detected when the 607-nm form was produced, we cannot exclude that a radical species is located in the protein nor the possibility that Cu B may be in a Y ⅐ -Cu(II) B 3 Y-Cu B (III) equilibrium intermediate state. If the latter scenario is correct, then the mechanism for the critical O-O scission will be dependent on the equilibrium of Y ⅐ -Cu(II) B 3 Y-Cu B (III).
The vibrational properties of oxygenated intermediates in histidine-containing peroxidase catalysis have been explained by the push-pull reaction mechanism (64,65) in which the distal histidine acts as a base to accelerate the binding of H 2 O 2 by deprotonating the substrate to give a Fe 3ϩ -O-O-R species. The pull aspect of the mechanism involves donation of a proton to the terminal oxygen of the hydroperoxide by the same residue acting as an acid. Concurrently, in the push component of the mechanism, an amino acid in the proximal pocket hydrogen-bonds to the proximal histidine and imparts significant anionic character to this residue. In the absence of Tyr-280 in the distal site, Glu-278 or a deprotonated Cu B -histidine can play a role as a general acid-base catalyst. It is noteworthy that the involvement of Glu as an acid-base catalyst has been demonstrated in chloroperoxidases where Glu-186 is located at the position corresponding to the distal histidine of horseradish peroxidase (66,67). Similarly, the basicity of the proximal His-411 can be significantly enhanced by local protein forces and can explain the observed structural differences between the two oxoferryl species we observe at 790 and 804 cm Ϫ1 . This interpretation finds support from earlier RR studies in which it has been shown that the properties of the trans ligand of oxoferryl complexes can affect the bonding between the iron and the oxygen, and thus, can affect its vibrational frequency (68); the vibrational frequency decreases as the ligand becomes more electron-donating to Fe(IV)ϭO. Along these lines, a high Fe-His mode arises from a population in which the hydrogen bond is intact, whereas a lower frequency of the Fe-His vibration reflects a form of the enzyme lacking the hydrogen bond (55). Based on the crystal structure, the proximal His-411 is capable of forming a hydrogen bond with Gly-378, and thus, the H-/non H-bonding interaction of the proximal His-411 with Gly-378 is the most likely contributor to the vibrational differences between the two oxoferryl species.
The disparity in the Fe(IV)ϭO frequency of the oxoferryl species we have detected is illustrated in Fig. 5. The 804 cm Ϫ1 species represents a His-Fe(IV)ϭO structure in which the proximal His-411 is not hydrogen-bonded to Gly-378. We postulate that conformational changes of the 804 cm Ϫ1 Fe(IV)ϭO species by the heme a 3 /Cu B pocket are communicated to the proximal environment, as has been observed in the peroxidase class of enzymes, to facilitate the H-bonding interaction of the proximal His-411, producing the 790 cm Ϫ1 species. In the presence of an H-bonded His-411, electron density is pushed into the antibonding * orbitals of Fe(IV)ϭO, weakening the bond, and thus, shifting the Fe(IV)ϭO stretching frequency from 804 to 790 cm Ϫ1 . The details on how the 804 -790 cm Ϫ1 oxoferryl transition is specifically tuned to initiate the H-bonding interaction and whether it is coupled to the heme a 3 -propionates-Asp-H 2 O protonic connectivity, in a similar manner to that found in cytochrome ba 3 from T. thermophilus (69), remain to be determined.