Functions Of Redox Active Metal Sites Of Cytochromec Oxidase

3.1. Spectral Properties

Keilin and Hartree discovered two types of cytochrome containing heme a: one of them oxidizes cytochrome-c and reduces another cytocrome, which is autoxidizable (21). The former is named cytochrome a and the latter, cytochrome a3 (21). However, as shown in X-ray structure, two cyto-chromes are not independent proteins, but the two heme sites are located within the single polypep-tide of subunit I. Thus, the two heme iron sites are called hemes a and a3.

A respiratory inhibitor, cyanide, specifically binds to heme a3 in the oxidized state and stabilizes the oxidized state strongly (22). The cyanide-bound heme a3 cannot be reduced even with an excess amount of dithionite, whereas heme a is unreactive to cyanide and is readily reduced by dithionite. Thus, a difference spectrum of the cyanide-treated enzyme reduced with dithionite versus the cyanide-treated enzyme without the addition of the reductant provides the redox difference spectrum of heme a, provided no interaction exists between the two hemes. On the other hand, a difference spectrum of the fully reduced enzyme against the fully oxidized enzyme gives the sum of the redox difference spectra of heme a and heme a3. Thus, the redox difference spectrum of heme a3 can be determined by subtracting the redox difference spectrum of heme a determined as earlier from the redox difference spectrum without addition of the inhibitor. As stated earlier, these procedures are appropriate for the spectral separation if the spectral changes in the two hemes are independent of each other. The proximity of the two hemes as shown in the X -ray structure of the enzyme (17) strongly suggests some interaction between the two hemes. Nevertheless, the spectral properties of the two hemes thus obtained are consistent with the expected structures of the two hemes. The difference spectrum of heme a shows a sharp intense band in the visible (a-band) region indicating that the reduced heme a is in a six-coordinated ferrous low-spin state. On the other hand, the weak a-band of the difference spectrum of heme a3 indicates a five-coordinated ferrous high-spin state. Both hemes contribute to the Soret band region in almost equal amounts. These spectral analysis were first done extensively by Yonetani (23), followed by Vanneste (24). The extinction coefficients determined by the above procedure have been used for all kinetic investigations of the internal electron-transfer reactions as well as redox titration studies.

The strong absorption spectra of the two hemes mask the absorption spectrum of CuA. In fact, the CuA moiety of one of the subunits (obtained by recombinant methods) shows an absorption spectrum in the visible Soret region of heme, although much weaker than the heme spectra (25). The contribution of CuA seems negligibly small in calculations of the redox difference spectra of the two hemes. On the other hand, CuA contributes significantly to the near-infrared band region (26,27). The redox state of CuA has been monitored by the spectrum in this region. The magnitude of the redox difference absorption at 825 nm is approx 1/30 of those in the a-band peaks (14). However, the contribution of the heme absorption to the near-infrared region should not be ignored (13). No absorption spectrum assignable to CuB has been reported, although reversible one-electron reduction has been established by redox titration studies (4,14,28). Perhaps, the electron paramagnetic resonance (EPR) signal detected under conditions similar to enzymic turnover is the only spectrum of this copper obtained thus far (29).

CuA shows a strong EPR signal near g = 2.0. The signal was assigned to a mononuclear copper center similar to the structure of type-I copper (30). However, about 10 yr ago, similarity between the EPR-detectable copper in nitrous oxide reductase (N2OR) and CuA in cytochrome-c oxidase was reported (31). The similarity between the two copper sites was proposed also by electron spin echo envelope modulation data (32) and magnetic circular dichroism (MCD) and extended X-ray absorption fine structure (EXAFS) spectroscopies (33). The similarity suggests that the CuA site is a mixed-

valence dinuclear copper center in which an electron equivalent is delocalized between the two Cu2+ ions to provide a [Cu(1.5)-Cu(1.5)] center. The structure is confirmed by the multifrequency EPR spectrum of the CuA site in bovine heart cytocrome-c oxidase (34).

The low-spin heme EPR signal is reasonably assigned to heme a in the fully oxidized state as prepared. However, quantitation of each of the two EPR signals indicates that the EPR-detectable component of each metal corresponds to roughly half of each metal contained in the enzyme. No other EPR signal assignable to that of intrinsic metals has been observed in the fully oxidized enzyme. These results suggest that only CuA and heme a in the fully oxidized state are EPR visible. It was proposed as early as in 1969 that CuB and heme a3 in the fully oxidized state are antiferromagnetically coupled with each other to eliminate the EPR spectra of both metal sites (30). The proposal was confirmed by several subsequent investigations (35-37).

3.2. Redox Properties

Redox titrations of this enzyme have been done by various methods, including changing the potential of the enzyme solution and additions of reductants and oxidants (14,28,38-43). In the potentio-metric method, various electron-transfer mediators must be added for equilibration between the enzyme and the electrode, because the redox active sites of the enzyme are located inside the protein. For titrations with chemical reagents, NADH with a catalytic amount of phenazinemethosulfate, dithionite, ferricyanide, ferrocyanide, and O2 have been used. The results so far obtained are not consistent with each other.

Redox titration experiments of cytochrome-c oxidase are influenced by many factors. Because of the extremely high reactivity with O2, a trace amount of contaminant O2 affects the titration results. On the other hand, complete removal of O2 from cytochrome oxidase in a mixed-micelle state is very difficult and tedious. Attainment of the equilibrium state is usually assessed monitoring the spectrum to an asymptotic level. Slight damage to the conformation of the enzyme could influence the kinetic properties of the redox active metal sites to give no apparent spectral change with time without attaining the equilibrium state. Thus, both the stability and integrity of the enzyme preparation are critical for the redox titration of cytochrome-c oxidase. Another important point often neglected is the effect of electron-transfer mediator. This point is especially important for the redox titration of cytochrome-c oxidase because the electron-transfer mediators stimulate autoreduction (reduction of the active site metals by electrons from amino acids in the protein moiety) (14). For example, it has been shown that phenazinemethosulfate under anaerobic conditions influences the reductive titration curve (14). Furthermore, significant amounts of contaminant metals are included in the cytochrome-c oxidase preparation if the enzyme is not purified by crystallization (14). The contaminant metals could influence the titration curve by redox interactions with the metal sites of cytocrome-c oxidase.

An anaerobic reductive titration using an anaerobic titration system designed for detergent-stabilized redox enzymes is given in Fig. 1, where bovine heart cytochrome oxidase purified by crystallization was used (14). Dithionite was used as the reductant without additional electron-transfer mediators. The spectral changes in Soret, visible, and near-infrared regions proceed essentially in parallel, as shown in the absorption changes at fixed wavelengths given in the insets. The results indicate that the oxidation state of all four redox active metal sites are essentially identical at any overall oxidation state, indicating that the redox potentials of all four metal sites are essentially identical. The insets show that six electron equivalents are required for complete reduction of the fully oxidized enzyme.

The initial 1- to 2-electron equivalents provide the absorbance change with shallow slopes as shown in the inset of Fig. 1. The shallow slopes are not the result of incomplete anaerobiosis because increasing the number of evacuation-equilibration cycles for removing O2 from 3-10 times did not remove the initial shallow portion of the titration curve. Incomplete anaerobiosis (one cycle of the evacuation-equilibration treatment) provided a two-phase curve similar to those given in the insets

Fig. 1. A reductive titration of the crystalline bovine heart cytochrome-c oxidase with dithionite. Absolute spectra for each oxidation states are shown for Soret (A) and visible (B) regions. The difference spectra against that in the fully reduced state are given for the near-infrared region (C). The insets show titration curves against the electron equivalent per enzyme. The reaction mixture contained 7.5 ^M bovine heart cytochrome-c oxidase in 0.1 M sodium phosphate buffer (pH 7.4). The enzyme preparation was stabilized with a synthetic nonionic detergent, CH3(CH2)„(OCH2CH2)8OH. The light path was 1 cm.

Fig. 1. A reductive titration of the crystalline bovine heart cytochrome-c oxidase with dithionite. Absolute spectra for each oxidation states are shown for Soret (A) and visible (B) regions. The difference spectra against that in the fully reduced state are given for the near-infrared region (C). The insets show titration curves against the electron equivalent per enzyme. The reaction mixture contained 7.5 ^M bovine heart cytochrome-c oxidase in 0.1 M sodium phosphate buffer (pH 7.4). The enzyme preparation was stabilized with a synthetic nonionic detergent, CH3(CH2)„(OCH2CH2)8OH. The light path was 1 cm.

of Fig. 1, but both slopes were shallower than those given in the insets. The above results are consistent with the X-ray structure of fully oxidized bovine heart cytochrome-c oxidase in which peroxide is bridged between Fea33+ and CuB2+. The initial two-electron equivalents are mainly used for cleavage of the O-O bond of the bridging peroxide to generate the fully oxidized form with no bridging ligand.

A discrepancy in the redox titration for the samples with and without purification by crystallization is detectable in the relative distribution of electron equivalents in the four redox active sites (14,28,38-43). Various uneven distribution have been reported (38-43). It should be noted that crystalline enzyme preparation shows an even distribution of electron equivalents added (14,28). The integral enzyme in mitochondrial membrane is the most likely to show an even distribution. The structures of the four redox active metal centers are quite different from each other. However, it is not surprising that different structures of metal sites have an identical redox potential, because small structural changes in bond length or bond angles could induce large changes in the affinity of the metal for electrons. The equipotential property of the four redox active metal sites is expected to facilitate facile electron transfers among the four sites, as well as multiple-electron transfers to the bound O2.

3.3. Electron Transfer Mechanism

The kinetics of internal electron transfer during the course of O2 reduction have been extensively studied by following the absorption changes in visible, Soret, and near-infrared regions. At ambient temperature, O2 reduction by this enzyme is complete within 0.1 ms or so. A conventional stopped-flow apparatus with a dead time of 1 ms or longer cannot be used for the kinetic investigation. Thus, a flow-flash method developed by Gibson and Greenwood (44) has long been used for kinetic investigations. This method is outlined as follows: The fully reduced enzyme solution saturated with CO is mixed with O2-containing buffer using a stopped-flow apparatus to introduce the mixture into an observation cell designed for effective flash irradiation of the mixture. When the flow stops, the flash is turned on to photolysing CO from the enzyme and triggers the initiation of the reaction between the fully reduced enzyme with O2. The affinity of CO to the enzyme is not high enough for complete protection of the enzyme from auto-oxidation (oxidation by O2), although CO inhibits the reaction between the enzyme and O2 for several milliseconds. Thus, the flash-photolysis system equipped with a stopped-flow apparatus is required to observe the O2 reduction kinetics. Since the historical work by Gibson and Greenwood in 1963 (44), the reaction was extensively studied with various instrumental improvements for about 30 yr. In 1994, the electron-transfer pathway from cytochrome-c to O2 was determined as cytochrome-c to CuA to heme a to [heme a3-CuB] (45).

3.4. Chemistry of O2 Reduction

As shown in Fig. 2, the one-electron reduction process of O2 in the ground state has a negative oxidation-reduction potential; all the other steps have positive potential (13). In other words, a one-electron reduction of O2 is energetically unfavorable, but the other steps are strongly favorable. This unfavorable one-electron reduction of O2 contributes greatly to the stability of oxygenated forms of hemoglobins and myoglobins because, in the oxygenated forms, Fe2+-O2 is isolated within the protein moiety so that a second electron for the two-electron reduction of O2 is not available. On the other hand, ferrous heme in aqueous solution is readily auto-oxidizable, because the second electron is available from another ferrous heme iron. In the two-electron reduction, a p-peroxo complex (a peroxide compound with two heme irons on both ends) is formed as an intermediate species. In fact such a p-peroxo intermediate has been identified in an auto-oxidation reaction of an amine cobalt compound (46). It was well known long before publication of the X-ray structure of the O2 reduction site of cytochrome-c oxidase that, in contrast to hemoglobins and myoglobins, the penta-coordinated heme iron in the O2-binding site is located near one of the copper sites, CuB (4,47). The structure suggests an effective O2 reduction process including the p-peroxo intermediate between ferric heme

Fig. 2. Standard oxidation-reduction potentials for the steps involved in the conversion of O2 to water at 25°C and pH 7.0.

a3 iron and cupric CuB ion. The mechanism was proposed by Caughey et al. 25 yr ago (47), and was widely accepted. An important corollary of this mechanism is a negligibly low level of the oxygenated intermediate (Fea32+-O2) during the course of the O2 reduction. The formation of the ^i-peroxo intermediate (Fea33+-O-O-CuB2+) from the oxygenated form (Fea32+-O2'CuB1+) is rate limited by an electron transfer from CuB1+ to the oxygen atom of the O2 closest to CuB. The electron-transfer rate through such a short distance (approx 2 A) could be on the order of 1 ps. On the other hand, formation of the oxygenated form is rate limited by migration of O2 through an O2 transfer pathway, which is limited by the protein dynamics. Thus, formation of the oxygenated form is likely to be much slower than the transition to the ^-peroxo form. The oxygenated form is therefore undetectable during the course of O2 reduction.

3.5. Resonance Raman Studies on O2 Reduction by Cytochrome-c Oxidase

As described in Section 3.3., the reaction of the fully reduced enzyme with O2 has been investigated extensively for so many years by the flow-flash method by following the spectral changes in the visible-Soret and near-infrared regions (4,45). The method is very effective for investigations of the kinetics of the redox reaction between the metal sites. However, the electronic spectra provide essentially no information about the chemical structure of the ligand on Fea3. An important approach for identification of the chemical structure of the intermediate species is the trial for trapping the oxygenated form at low temperature by Chance and co-workers (48). They followed the reaction by visible spectroscopy. The oxygenated form was identified by the very weak photosensitivity of the intermediate in addition to the characteristic visible spectrum at 590 nm closely similar to that of the CO compound. Subsequently, a similar approach was done by Gibson's group in which they followed the disappearance of the oxygenated form by its photosensitivity to laser flashes (49). These works have contributed significantly to improvements of our understanding of the mechanism of O2 reduction by the emzyme. However, they have never succeeded in obtaining direct chemical structural evidence for the oxygenated form.

Vibrational spectroscopy is amenable to examination of the chemical structures of the ligands on heme a3 in the intermediate species. Time-resolved resonance Raman spectroscopy was applied effectively for identification of the intermediate species during the course of O2 reduction. Three research groups identified almost at once the initial intermediate as an oxygenated form (Fe2+-O2) decisively (50-52). The identification of Fe2+-O2 is extremely astonishing considering the above discussion. Figure 3 shows the resonance Raman band of the initial intermediate during the course of the reaction between the fully reduced enzyme with O2, determined by one of the three groups. The results were taken by using the extensively improved instrumentation given in the article by Ogura et al. (53). Because of the weak intensity of the resonance Raman band, the bands were identified only in the difference spectrum between the spectrum obtained for the naturally abundant 16O2 and that for an isotopically labeled 18O2, as shown in Fig. 3. To examine the mode of binding of O2 to Fea32+, a

Fig. 3. Resonance Raman spectra of the Fe2+-O2 stretching frequency region of bovine heart cytochrome-c oxidase at 0.1 ms after initiation of the reaction of the fully reduced enzyme with O2. Spectra on the left-hand and right-hand sides are the observed spectra and the calculated spectra with the differences of the observed vs calculated spectra, respectively. Spectrum d is obtained by the following calculation: (Spectrum b-Spectrum c)/2. (e) Simulated bands for Fe-16O2 (1), Fe-16O18O (2), Fe-18O16O (3), and Fe-18O2 (4). The peak intensity ratio is 6:6:5:5. All bands have the Gaussian band shape with a half-maximal bandwidth of 12.9 cm-1.

Fig. 3. Resonance Raman spectra of the Fe2+-O2 stretching frequency region of bovine heart cytochrome-c oxidase at 0.1 ms after initiation of the reaction of the fully reduced enzyme with O2. Spectra on the left-hand and right-hand sides are the observed spectra and the calculated spectra with the differences of the observed vs calculated spectra, respectively. Spectrum d is obtained by the following calculation: (Spectrum b-Spectrum c)/2. (e) Simulated bands for Fe-16O2 (1), Fe-16O18O (2), Fe-18O16O (3), and Fe-18O2 (4). The peak intensity ratio is 6:6:5:5. All bands have the Gaussian band shape with a half-maximal bandwidth of 12.9 cm-1.

terminally labeled 16O18O was used. If O2 binds to Fea3 in a bent end-on fashion, two bands with equal intensity are expected at a band position near but slightly lower than that of 16O2 and at a position slightly higher than that of 18O2. On the other hand, for a side-on fashion, a single band is expected at a position halfway between those of 16O2 and 18O2. If only an oxide binds to the heme iron to produce a ferryl oxide (Fe4+=O2-), the terminally labeled 16O18O provides two bands strictly at the positions obtained for 16O2 or 18O2. As shown in Fig. 3D, the spectrum obtained for 16O18O is not identical with the average of 16O2 and 18O2. The difference spectrum is able to be simulated by the four bands given in Fig. 3D, which indicates a bent end-on type coordination of O2 to Fea32+. Furthermore, the band at 571 cm-1 is in the range of the Fe-O2 stretch band of various hemoglobins and myoglobins. This is one of the most fundamental results for elucidation of the reaction mechanism of this enzyme, which is inconsistent with the conclusion given in Section 3.4.

The next intermediate exhibits resonance Raman band at 785 cm-1, followed by the band at 804 cm-1, which is the result of the third intermediate. Similar analysis of the second intermediate using 16O2, 18O2, and 16O18O showed, again, astonishingly that the 785-cm-1 intermediate is the ferryl oxide (53). That was the case for the 804-cm-1 band (53). The implication of these results will be discussed next in relation to the O2 reduction mechanism.

0 0

Post a comment