The influence of pH on the EPR and redox properties of cytochrome c oxidase in detergent solution and in phospholipid vesicles

The EPR signals of oxidized and partially reduced cytochrome oxidase have been studied at pH 6.4, 7.4, and 8.4. Isolated cytochrome oxidase in both non-ionic detergent solution and in phospholipid vesicles has been used in reductive titrations with ferrocytochrome c.The g values of the low- and high-field parts of the low-spin heme signal in oxidized cytochrome oxidase are shown to be pH dependent. In reductive titrations, low-spin heme signals at g 2.6 as well as rhombic and nearly axial high-spin heme signals are found at pH 8.4, while the only heme signals appearing at pH 6.4 are two nearly axial g 6 signals. This pH dependence is shifted in the vesicles.The g 2.6 signals formed in titrations with ferrocytochrome c at pH 8.4 correspond maximally to 0.25–0.35 heme per functional unit (aa₃) of cytochrome oxidase in detergent solution and to 0.22 heme in vesicle oxidase. The total amount of high-spin heme signals at g 6 found in partially reduced enzyme is 0.45–0.6 at pH 6.4 and 0.1–0.2 at pH 8.4. In titrations of cytochrome oxidase in detergent solution the g 1.45 and g 2 signals disappear with fewer equivalents of ferrocytochrome c added at pH 8.4 compared to pH 6.4.The results indicate that the environment of the hemes varies with the pH. One change is interpreted as cytochrome a₃ being converted from a high-spin to a low-spin form when the pH is increased. Possibly this transition is related to a change of a liganded H₂O to OH^? with a concomitant decrease of the redox potential. Oxidase in phosphatidylcholine vesicles is found to behave as if it experiences a pH, one unit lower than that of the medium.     

Some properties of the redox components of cytochrome c oxidase and their interactions.

The electron paramagnetic resonance (epr) properties of cytochrome c oxidase have been examined with special attention to the effect of added ligands and of interactions between the redox components. The fully oxidized preparations have a very small g6 signal which increases greatly as the redox potential is made more negative, a process exactly paralleling the disappearance of the g3 signal. The potential for half appearance or disappearance (E_m), respectively, is 380 mV at pH 7.0 and 300 mV at pH 8.5. This identifies the changes as accompanying reduction of cytochrome a₃ because the E_m of the “invisible copper” is 340 mV and pH independent. Nitric oxide (NO) binds reduced cytochrome a₃ to form a paramagnetic species. This resulting epr signal is strongly dependent on the redox state of cytochrome a, another expression of heme-heme interaction in cytochrome oxidase. The NO compound is also unique in that under the appropriate conditions three of the four redox components (cytochrome a₃, cytochrome a, and the “visible” copper) are epr active. In potentiometric titrations in the presence of azide the formation of the azide compound responsible for the g2.9 signal appears to require reduction of both cytochrome a₃ and the “invisible copper.” An internal sulfur compound is present which, at alkaline pH values, can bind the heme responsible for the g6 signal and change it to a low-spin sulfur compound with a signal at approximately g2.6. Evidence is also presented for the cytochrome c oxidase in situ being an equilibrium mixture of two different conformational states.     

Studies of the orientation of the mitochondrial redox carriers III. Orientation of the gx and gy axes of the hemes of cytochrome oxidase with respect to the plane of the membrane in oriented membrane multilayers

The EPR absorption properties of the hemes of cytochrome oxidase and their liganded derivatives were examined in oriented multilayers from isolated oxidase, mitochondrial membranes and membrane fragments of a bacterium, Paracoccus denitrificans. The hemes of the oxidase in all the systems investigated were oriented normal to the plane of the multilayers. The directions of the g signals corresponding to the g_x and g_y axes of the g tensor were found to be different in low-spin ferric heme in fully oxidized oxidase and in half-reduced liganded oxidase. It is suggested that this different orientation of g_x and g_y in fully oxidized oxidase and half-reduced liganded oxidase arises because the respective EPR signals belong to two different hemes, those of cytochrome a and a₃.     

An EPR signal from the "invisible" copper of cytochrome oxidase.

Sulfide is both an inhibitor and a slow reductant of oxidized cytochrome $\text{c}$ oxidase. When the enzyme is exposed to sulfide for short times (one minute or less) and frozen, the resultant electron paramagnetic resonance (EPR) signals show clearly: low spin heme a, low spin heme a₃, the usual “EPR detectable” Cu²⁺ signal (g_⊥ = 2.17, g_⊥ = 2.03), and a new Cu²⁺ signal superimposed on the same region, with (g_⊥ ～ 2.19, g_⊥ = 2.05). This new signal presumably arises because the antiferromagnetic coupling postulated to exist between the iron atom of heme a₃ and this copper is disrupted when heme a₃ is driven to a low spin state by sulfide. The implications of this result with respect to models of the O₂-binding site and redox geometry of oxidase are briefly discussed.     

Reactions of mercaptans with cytochrome c oxidase and cytochrome c

1. The steady-state oxidation of ferrocytochrome c by dioxygen catalyzed by cytochrome c oxidase, is inhibited non-competitively towards cytochrome c by methanethiol, ethanethiol, 1-propanethiol and 1-butanethiol with K_i values of 4.5, 91, 200 and 330 μM, respectively.2. The inhibition constant K_i of ethanethiol is found to be constant between pH 5 and 8, which suggests that only the neutral form of the thiol inhibits the enzyme.3. The absorption spectrum of oxidized cytochrome c oxidase in the Soret region shows rapid absorbance changes upon addition of ethanethiol to the enzyme. This process is followed by a very slow reduction of the enzyme. The fast reaction, which represents a binding reaction of ethanethiol to cytochrome c oxidase, has a k₁ of 33 M^?1 · s^?1 and dissociation constant K_d of 3.9 mM.4. Ethanethiol induces fast spectral changes in the absorption spectrum of cytochrome c, which are followed by a very slow reduction of the heme. The rate constant for the fast ethanethiol reaction representing a bimolecular binding step is 50 M^?1 · s^?1 and the dissociation constant is about 2 mM. Addition of up to 25 mM ethanethiol to ferrocytochrome c does not cause spectral changes.5. EPR (electron paramagnetic resonance) spectra of cytochrome c oxidase, incubated with methanethiol or ethanethiol in the presence of cytochrome c and ascorbate, show the formation of low-spin cytochrome a₃-mercaptide compounds with g values of 2.39, 2.23, 1.93 and of 2.43, 2.24, 1.91, respectively.     

On the identity of the high spin heme components of cytochrome c oxidase

In oxidized, resting cytochrome c oxidase (EC 1.9.3.1) and under most conditions of partial reduction ? 50% of the heme components are detected by EPR spectroscopy. When the enzyme is fully reduced in the presence of equimolar quantities of cytochrome c, anaerobic reoxidation by an excess of a chemical oxidant (ferricyanide, porphyrexide) produces intense high and low spin heme signals simultaneously. The time range in which maximal high spin signals are observed is 0.1–2 s after mixing. Under these conditions 35–50% of the total heme a is accounted for by the low spin heme signal and 35–40% by the high spin signals, with the rhombic component accounting for 30–35% of the total heme. It is concluded that under these conditions, the major portion of both heme components must be EPR detectable. Thus, if the generally accepted assignment of the low spin signal to cytochrome a is adopted, it follows that in the experiments described, cytochrome a₃ is represented in the rhombic high spin signal. The quantities of heme represented in the axial high spin signal are too small for a definitive assignment; these signals could originate from either heme. When after formation of high spin signals as described, O₂ is admitted, the rhombic signal is eliminated within 4 ms. In the presence of the strongest rhombic high spin signals, the absorption band at 655 nm is only ? 25% developed. The implications of these findings are discussed in the context of present hypotheses concerning the state and interactions of cytochrome c oxidase components during oxidation-reduction.     

Kinetic studies on cytochrome c oxidase by combined EPR and reflectance spectroscopy after rapid freezing

1. Techniques and experiments are described concerned with the millisecond kinetics of EPR-detectable changes brought about in cytochrome c oxidase by reduced cytochrome c and, after reduction with various agents, by reoxidation with O₂ or ferricyanide. Some experiments in the presence of ligands are also reported. Light absorption was monitored by low-temperature reflectance spectroscopy.2. In the rapid phase of reduction of cytochrome c oxidase by cytochrome c (< 50 ms) approx. 0.5 electron equivalent per hame a is transferred mainly to the low-spin heme component of cytochrome c oxidase and partly to the EPR-detectable copper. In a slow phase (> 1 s) the copper is reoxidized and high-spin ferric heme signals appear with a predominant rhombic component. Simultaneously the absorption band at 655 nm decreases and the Soret band at 444 nm appears between the split Soret band (442 and 447 nm) of reduced cytochrome a.3. On reoxidation of reduced enzyme by oxygen all EPR and optical features are restored within 6 ms. On reoxidation by O₂ in the presence of an excess of reduced cytochrome c, states can be observed where the low-spin heme and copper signals are largely absent but the absorption at 655 nm is maximal, indicating that the low-spin heme and copper components are at the substrate side and the component(s) represented in the 655 nm absorption at the O₂ side of the system. On reoxidation with ferricyanide the 655 nm absorption is not readily restored but a ferric high-spin heme, represented by a strong rhombic signal, accumulates.4. On reoxidation of partly reduced enzyme by oxygen, the rhombic high-spin signals disappear within 6 ms, whereas the axial signals disappear more slowly, indicating that these species are not in rapid equilibrium. Similar observations are made when partly reduced enzyme is mixed with CO.5. The results of this and the accompanying paper are discussed and on this basis an assignment of the major EPR signals and of the 655 nm absorption is proposed, which in essence is that published previously (Hartzell, C. R., Hansen, R. E. and Beinert, H. (1973) Proc. Natl. Acad. Sci. U.S. 70, 2477–2481). Both the low-spin (g = 3; 2.2; 1.5) and slowly appearing high-spin (g = 6; 2) signals are attributed to ferric cytochrome a, whereas the 655 nm absorption is thought to arise from ferric cytochrome a₃, when it is present in a state of interaction with EPR-undetectable copper. Alternative possibilities and possible inconsistencies with this proposal are discussed.     

Thermodynamic and EPR characterization of mitochondrial succinate-cytochrome c reductase-phospholipid complexes

Succinate-cytochrome c reductase can be easily solubilized in a phospholipid mixture (1:1, lysolecithin:lecithin) in the absence of detergents. The resulting solution contains two b cytochromes with half-reduction potentials of 95 ± 10 mV (b₅₆₁), and 0 ± 10 mV (b₅₆₆) and cytochrome c₁ (E_{m 7.2} = +280±5 mV). The oxidation-reduction midpoint potentials obtained by optical potentiometric titrations are identical to those determined by the EPR titrations and are 40–60 mV higher than the corresponding midpoint potentials of these cytochromes in intact mitochondria. In contrast to detergent-suspended preparations, no CO-sensitive cytochrome b can be detected in the phospholipid-solubilized preparation or intact mitochondria. The half-reduction potential of cytochrome b₅₆₆ is pH-dependent above pH 7.0 (?60 mV/pH unit) while that of b₅₆₁ is essentially pH-independent from pH 6.7–8.5, in contrast to its pH dependence in intact mitochondria. EPR characterizations show the presence of three oxidized low-spin heme-iron signals with g values of 3.78, 3.41 and 3.37. The identification of these signals with cytochromes b₅₆₆ (b_T), b₅₆₁ (b_K) and c₁ respectively is made on the basis of redox midpoint potentials. No significant amounts of oxidized high-spin heme-iron are detectable. In addition, the preparation contains four distinct types of iron-sulfur centers: S₁ and S₂ (E_{m 7.4} = ?260 mV and 0 mV), and two iron-sulfur proteins which are associated with the cytochrome b-c₁ complex: Rieske's iron-sulfur protein (E_{m 7.4} = +280 mV) and Ohnishi's Center 5 (E_{m 7.4} = +35 mV).     

Modulation of the electron-proton coupling at cytochrome a by the ligation of the oxidized catalytic center in bovine cytochrome c oxidase

Cytochrome a was suggested as the key redox center in the proton pumping process of bovine cytochrome c oxidase (CcO). Recent studies showed that both the structure of heme a and its immediate vicinity are sensitive to the ligation and the redox state of the distant catalytic center composed of iron of cytochrome a₃ (Fe_a3) and copper (Cu_B). Here, the influence of the ligation at the oxidized Fe_a3³⁺–Cu_B²⁺ center on the electron–proton coupling at heme a was examined in the wide pH range (6.5-11). The strength of the coupling was evaluated by the determination of pH dependence of the midpoint potential of heme a (E_m(a)) for the cyanide (the low-spin Fe_a3³⁺) and the formate-ligated CcO (the high-spin Fe_a3³⁺). The measurements were performed under experimental conditions when other three redox centers of CcO are oxidized. Two slightly differing linear pH dependencies of E_m(a) were found for the CN– and the formate–ligated CcO with slopes of −13 mV/pH unit and −23 mV/pH unit, respectively. These linear dependencies indicate only a weak and unspecific electron–proton coupling at cytochrome a in both forms of CcO. The lack of the strong electron–proton coupling at the physiological pH values is also substantiated by the UV–Vis absorption and electron–paramagnetic resonance spectroscopy investigations of the cyanide–ligated oxidized CcO. It is shown that the ligand exchange at Fe_a³⁺ between His–Fe_a³⁺–His and His–Fe_a³⁺–OH⁻ occurs only at pH above 9.5 with the estimated pK >11.0.     

The properties of the mitochondrial succinate-cytochrome c reductase

The cytochromes b and b_T of pigeon heart mitochondria have half-reduction potentials (Em's) of +30 mV and −30 mV at pH 7.2. The midpoint potentials of these cytochromes become more negative by 30–60 mV per pH unit when the pH is made more alkaline. Detergents may be used to prepare a succinate-cytochrome c reductase free of cytochrome oxidase in which the activation of electron transport induced by oxidation of cytochrome c₁ causes the half-reduction potential of cytochrome b_T to become at least 175 mV more positive than in the absence of electron transport. This change is interpreted as indicating that the primary energy conservation reaction at site 2 remains fully functional in the purified reductase. Preliminary electron paramagnetic resonance spectra of the succinate-cytochrome c reductase as measured at near liquid helium temperatures are presented.     

Biochemical and biophysical studies on cytochrome c oxidase XVI. Circular dichroic study of cytochrome c oxidase and its ligand complexes

1. CD spectra of cytochrome c oxidase have been determined both in the absence and presence of the extrinsic ligands CO, NO, cyanide and azide.2. CO and NO affect the CD spectrum of cytochrome c oxidase in a similar way.3. Cyanide and azide also affect the CD spectrum of cytochrome c oxidase in a similar way, but distinctly different from CO and NO.4. From the CD spectra of the oxidized and reduced enzyme, in the presence and absence of extrinsic ligands, CD difference spectra (reduced minus oxidized) are calculated for the so-called cytochrome a and cytochrome a₃ moieties of the enzyme.5. These spectra are largely dependent on the extrinsic ligand used. It is therefore concluded that these spectra do not represent independent cytochrome a and cytochrome a₃ difference spectra, but that heme-heme interactions occur within the cytochrome c oxidase molecule, in such a way that binding of a ligand to one of the heme a groups of cytochrome c oxidase affects the spectral properties of the other heme a group.6. As a consequence, ligand-binding studies cannot give information as to the pre-existence of separate cytochrome a and cytochrome a₃ moieties in the absence of extrinsic ligands.     

Reaction of CO with cytochrome c oxidase. Titration of the reaction site with chemical oxidant and reductant

The number of reducing equivalents required to form the reduced cytochrome a₃-CO compound has been determined for suspensions of submitochondrial particles and for isolated cytochrome c oxidase. Anaerobic preparations were titrated reductively with NADH and oxidatively with O₂ in the presence of high concentrations of CO (0.4 to 0.8 mM) while monitoring reduction of cytochrome a and the formation of the reduced cytochrome a₃-CO compound by their characteristic absorbance changes. Analysis of the titration data show that 2.0±0.3 and 2.1±0.2 reducing equivalents per mol of cytochrome oxidase (per cytochrome a) are required for formation of the reduced cytochrome a₃-CO compound in submitochondrial particles and isolated cytochrome c oxidase, respectively. In each case, the formation of the CO compound is proportional to the number of equivalents accepted by the preparation, indicating that the two equivalents are equal and the effective n value for the reaction is 2.0. Potentiometric titrations of cytochrome c oxidase using the cobalt orthophenanthrolene complex (E_{m, 7.0} = 0.37 V) as mediator give the same half-reduction potential values for cytochrome a and a₃ as those obtained using the ferro-ferricyanide couple. The formation of the reduced cytochrome a₃-CO compound at pH 7.0, in the presence of 0.6 mM CO and with CO-orthophenanthrolene as mediator occurs with a half-reduction potential of 0.45 V and requires two electrons. These data confirm and extend the observation of Lindsay et al. (Arch. Biochim. Biophys. (1975) 169, 492–505) that both the “invisible” copper and cytochrome a₃ must be reduced in order for CO to bind with high affinity.     

EPR study of the effect of formate on cytochrome C oxidase

The addition of formate to oxidized cytochrome $\text{c}$ oxidase (ferrocytochrome $\text{c}$: oxygen oxidoreductase, EC 1.9.3.1) causes the appearance of a high spin heme signal at g = 6 and a splitting of g = 3 signal to g = 2.98 and 3.07. When formate-cytochrome $\text{c}$ oxidase is reduced, the g = 2.98 signal decreases significantly. The spectrophotometric studies showed that formate is a specific ligand to cytochrome $\text{a}$₃. Data suggest that binding of formate to oxidized cytochrome $\text{c}$ oxidase produces a ligand-$\text{a}$₃ interaction leading to the splitting of g = 3 signal hitherto considered as due to cytochrome $\text{a}$. Thus both cytochrome $\text{a}$ and $\text{a}$₃ contribute to the resonance of g = 3 signal of cytochrome $\text{c}$ oxidase.     

EPR studies of the photodissociation reactions of cytochrome c oxidase-nitric oxide complexes

Three complexes of NO with cytochrome c oxidase are described which are all photodissociable at low temperatures as measured by EPR. The EPR parameters of the cytochrome a²⁺₃-NO complex are the same both in the fully reduced enzyme and in the mixed-valence enzyme. The kinetics of photodissociation of cytochrome a²⁺₃-NO and recombination of NO with cytochrome a²⁺₃ (in the 30–70 K region) revealed no differences in structure between cytochrome a²⁺₃ in the fully reduced and the mixed-valence states. The action spectrum of the photodissociation of cytochrome a²⁺₃-NO as measured by EPR has maxima at 595, 560 and 430 nm, and corresponds to the absorbance spectrum of cytochrome a²⁺₃-NO. Photodissociation of cytochrome a²⁺₃-NO in the mixed-valence enzyme changes the EPR intensity at g 3.03, due to electron transfer from cytochrome a²⁺₃ to cytochrome a³⁺. The extent of electron transfer was found to be temperature dependent. This suggests that a conformational change is coupled to this electron transfer. The complex of NO with oxidized cytochrome c oxidase shows a photodissociation reaction and recombination of NO (in the 20–40 K region) which differ completely from those observed in cytochrome a²⁺₃-NO. The observed recombination occurs at a temperature 15 K lower than that found for the cytochrome a²⁺₃-NO complex. The action spectrum of the oxidized complex shows a novel spectrum with maxima at 640 and below 400 nm; it is assigned to a Cu²⁺_B-NO compound. The triplet species with Δm_s = 2 EPR signals at g 4 and Δm_s = 1 signals at g 2.69 and 1.67, that is observed in partially reduced cytochrome c oxidase treated with azide and NO, can also be photodissociated.     

Cytochrome c interactions with membranes. A photoaffinity-labeled cytochrome c.

The aryl azide, 2,4-dinitro-5-fluorophenylazide, was reacted with horse heart cytochrome c to give a photoaffinity-labeled derivative of this heme protein. The modified cytochrome c, with one to two dinitroazidophenyl groups per mole of the enzyme, has a half-reduction potential the same (± 10 mV) as native cytochrome c. The dissociation constant for the modified cytochrome c from cytochrome c-depleted mitochondrial membranes and the apparent K_m for the reaction with cytochrome c oxidase were each five to six times greater than the values for native cytochrome c. Irradiation of cytochrome c-depleted mitochondrial membranes supplemented with an excess of photoaffinity-labeled cytochrome c resulted in covalent binding of the derivative to the mitochondrial membranes. Fractionation of the irradiated mitochondria in the presence of detergents and salts followed by chromatography on agarose, Bio-Gel A, showed that labeled cytochrome c was bound covalently to cytochrome c oxidase in a 1:1 molar complex. The covalently linked cytochrome c-cytochrome c oxidase complex was active in mediating the electron transfer between N,N,N′,N′-tetramethyl-p-phenylenediamine/ascorbate and the oxidase.     

Formamide probes a role for water in the catalytic cycle of cytochrome c oxidase

Formamide is a slow-onset inhibitor of mitochondrial cytochrome c oxidase that is proposed to act by blocking water movement through the protein. In the presence of formamide the redox level of mitochondrial cytochrome c oxidase evolves over the steady state as the apparent electron transfer rate from cytochrome a to cytochrome a₃ slows. At maximal inhibition cytochrome a and cytochrome c are fully reduced, whereas cytochrome a₃ and Cu_B remain fully oxidized consistent with the idea that formamide interferes with electron transfer between cytochrome a and the oxygen reaction site. However, transient kinetic studies show that intrinsic rates of electron transfer are unchanged in the formamide-inhibited enzyme. Formamide inhibition is demonstrated for another member of the heme-oxidase family, cytochrome c oxidase from Bacillus subtilis, but the onset of inhibition is much quicker than for mitochondrial oxidase. If formamide inhibition arises from a steric blockade of water exchange during catalysis then water exchange in the smaller bacterial oxidase is more open. Subunit III removal from the mitochondrial oxidase hastens the onset of formamide inhibition suggesting a role for subunit III in controlling water exchange during the cytochrome c oxidase reaction.     

Reduction of oxygen-pulsed cytochrome c oxidase by cytochrome c and other electron donors

1. Stopped-flow experiments were performed in which solutions containing dithionite were mixed with air-saturated buffer. Cytochrome c oxidase present in the dithionite-containing syringe is fully oxidized within the mixing time and the oxygen-pulsed form of the oxidase is produced.2. The reduction of this form by dithionite, by dithionite plus cytochrome c and by dithionite plus methyl viologen or benzyl viologen was followed and compared with the corresponding reduction reactions of the ‘resting’ oxidized enzyme. Reduction by dithionite is relatively slow, but the rate of reduction is greatly increased by addition of cytochrome c or the viologens, which are even more effective than cytochrome c on a molar basis.3. Profound differences between the transient kinetics of the reduction of the two oxidized oxidase derivatives were observed. The results are consistent with a direct reduction of cytochrome a followed by an intramolecular electron transfer to cytochrome a₃ (k^obs₁ = 7.5 s^?1 for the oxygen-pulsed oxidase).4. The spectrum of the oxygen-pulsed oxidase formed within 5 ms of the mixing closely resembles that of the ‘oxygenated’ compound, but there were small differences between the two spectra.     

Effect of crosslinking cytochrome c oxidase

Bovine heart cytochrome c oxidase and rat liver mitochondria were crosslinked in the presence and absence of cytochrome c. Biimidate treatment of purified cytochrome oxidase, which results in the crosslinkage of all of the oxidase protomers except subunit I when ? 20% of the free amines are modified, inhibits ascorbate-N,N,N′,N′-tetramethyl-p-phenylene diamine oxidase activity. Intermolecular crosslinking of cytochrome oxidase molecules, which results in the formation of large enzyme aggregates displaying rotational correlation times ? 1 ms, does not affect oxidase activity. Crosslinking of mitochondria covalently binds the cytochrome bc₁ and aa₃ complexes to cytochrome c, and inhibits steady-state oxidase activity. Addition of cytochrome c to purified cytochrome oxidase or to cytochrome c-depleted mitoplasts increases this inhibition slightly. Cytochrome c oligomers act as competitive inhibitors of native cytochrome c; however, crosslinking of cytochrome c to cytochrome c-depleted mitoplasts or purified cytochrome oxidase results in a catalytically inactive complex. These experiments indicate that cytochrome c oxidase subunit interactions are required for activity, and that cytochrome c mobility may be essential for electron transport between cytochrome c reductase and oxidase.