Molecular oxygen modulates cytochrome c oxidase function.

This study sought to determine whether molecular oxygen interacts with cytochrome c oxidase to modify its catalytic activity. Such an interaction could explain the observation that mitochondria incubated under low O2 concentrations exhibit a reversible suppression of State 3 respiration. Oxidized bovine heart cytochrome c oxidase was incubated in oxygen concentrations of <50 μM for 4 h. The enzyme exhibited a reversible decrease in Vmax after incubation, compared with control enzyme incubated at higher oxygen concentrations. This change was accompanied by a small increase in the apparent Km of the enzyme for both cytochrome c and oxygen, although the optical absorption spectra of oxidized, cycling, or reduced enzyme were not affected. Spectroscopy studies after 4 h of incubation revealed that heme a3 was 33% reduced during cycling at [O2] = 25 μM whereas enzyme at [O2] = 135 μM was only 18% reduced, suggesting that the site of inhibition occurred at the electron transfer step between heme a3 and O2. These results provide a mechanistic explanation for the observation that intact cells or mitochondria exhibit a reversible inhibition of respiration during prolonged exposure to [O2] <25 mM, by demonstrating that the catalytic activity of cytochrome c oxidase function is similarly inhibited, possibly through an allosteric effect of molecular O2 on the enzyme.

This study sought to determine whether molecular oxygen interacts with cytochrome c oxidase to modify its catalytic activity. Such an interaction could explain the observation that mitochondria incubated under low O 2 concentrations exhibit a reversible suppression of State 3 respiration. Oxidized bovine heart cytochrome c oxidase was incubated in oxygen concentrations of <50 M for 4 h. The enzyme exhibited a reversible decrease in V max after incubation, compared with control enzyme incubated at higher oxygen concentrations. This change was accompanied by a small increase in the apparent K m of the enzyme for both cytochrome c and oxygen, although the optical absorption spectra of oxidized, cycling, or reduced enzyme were not affected. Spectroscopy studies after 4 h of incubation revealed that heme a 3 was 33% reduced during cycling at [O 2 ] ‫؍‬ 25 M whereas enzyme at [O 2 ] ‫؍‬ 135 M was only 18% reduced, suggesting that the site of inhibition occurred at the electron transfer step between heme a 3

and O 2 . These results provide a mechanistic explanation for the observation that intact cells or mitochondria exhibit a reversible inhibition of respiration during prolonged exposure to [O 2 ] <25 mM, by demonstrating that the catalytic activity of cytochrome c oxidase function is similarly inhibited, possibly through an allosteric effect of molecular O 2 on the enzyme.
Cytochrome c oxidase is the terminal oxidase of cellular respiration and catalyzes the transfer of electrons from ferrocytochrome c to molecular oxygen (1,2). The free energy release associated with this electron transfer is coupled to the translocation of protons from the mitochondrial matrix to the cytosol, generating a proton electrochemical gradient across the inner mitochondrial membrane. This potential is subsequently used to sustain a number of mitochondrial functions including the synthesis of ATP. The enzyme contains four redox-active metal centers, two hemes (a and a 3 ), and two redox-active coppers, Cu A and Cu B (3). Cu A is thought to reside in subunit II of the enzyme and to consist of a binuclear center that accepts electrons from reduced cytochrome c. Electrons are transferred to heme a (4) and then to the binuclear oxygen binding site, heme a 3 and Cu B , in subunit I. Electrons are finally transferred to O 2 , although the details of this process are incompletely understood.
The kinetic behavior of the oxidase has been studied exten-sively (5) Chemicals-All biochemicals were obtained from Sigma. The Pdmeso-tetra-(4-carboxyphyenyl) porphine dye used for phosphorescencequenching measurements of oxygen concentration was obtained from Medical Systems, Inc., Greenvale, NY. The isolated cytochrome c oxidase was a gift from Dr. Martin Horvath.
Preparations-Bovine heart cytochrome c oxidase concentrations were based on a value of ⑀ ϭ 1.41 ϫ 10 5 M Ϫ1 ⅐cm Ϫ1 at 421 nm for the oxidized enzyme. The enzyme was prepared by the method of Hartzell and Beinert (8) and was stored at a concentration of 0.6 mM at Ϫ80°C until immediately before use. Cytochrome c was reduced as described previously (9). Briefly, bovine cytochrome c (⑀ ϭ 2.12 ϫ 10 4 M Ϫ1 ⅐cm Ϫ1 at 550 minus 540 nm) (10) was dissolved in a mixture containing equal amounts of 20 mM boric acid and 0.2 M sodium borohydride (pH ϭ 10.2) and incubated for 1 h. Cytochrome c reduction was stopped using sodium phosphate (pH ϭ 6.5).
Cytochrome c Oxidase Incubation-Oxidized bovine cytochrome c oxidase (1-10 nM) (11) was incubated in spinner flasks in 100 mM KH 2 PO 4 containing 0.5% dodecyl ␤-D-maltoside at pH ϭ 7.4, 25°C (7). The head space of the flask was continuously flushed with gas containing oxygen ranging from 2.8 to 14.2%, balance nitrogen. The solution was continuously mixed with a magnetic stirrer (60 rpm), and the dissolved oxygen tension was monitored using a polarographic electrode placed in a side port of the flask (12). The flasks were also equipped with glass tubing extending into the media to allow the anaerobic removal of aliquots for analysis.
Measurements of Oxygen Consumption-The apparent K m of cytochrome c oxidase with respect to cytochrome c was assessed in stirred solutions by measuring the decrease in oxygen partial pressure in a water-jacketed respirometer (1.5 ml) equipped with a calibrated polarographic O 2 electrode. Bovine cytochrome c (5 M), TMPD, 1 (500 M), and ascorbate (5 mM) were used as electron donors, and the decrease in oxygen tension over 5-10 min was recorded using a 12-bit analog-todigital converter. However, this method was not suitable for assessing the apparent K m of cytochrome c oxidase with respect to oxygen because the response speed of polarographic electrodes declines at [O 2 ] of less than 10 M. Accordingly, for those studies the oxygen concentrations in solution were measured using the oxygen-dependent phosphorescence quenching of a porphine dye (13). Accuracy of the system was confirmed by comparison of test solutions assessed simultaneously by polaro-* This work was supported by NHLBI Grants HL32646 and HL35440. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. graphic and optical systems under quasi-steady-state conditions. The time constant for phosphorescence decay under anoxia was confirmed in solutions using the glucose oxidase reaction to scavenge O 2 . In both O 2 measurement systems, the turnover number (s Ϫ1 ) for cytochrome c oxidase was calculated by dividing O 2 consumption (nmol O 2 /s) by the enzyme concentration (nM). In studies requiring the use of TMPD and ascorbate, the contribution of autooxidation of ascorbate (14) was small but was nevertheless measured and subtracted from the overall rate of oxygen consumption.
Spectrophotometric Determination of the K m of Cytochrome Oxidase for Cytochrome c-The apparent K m and V max of the enzyme with respect to cytochrome c were determined spectrophotometrically by measuring the oxidation of reduced cytochrome c over time from the absorbance at 550 minus 540 nm. First-order rate constants were calculated from the slopes of the linear plots of ln(C t t Ϫ C oxd ) against time, where C t is the concentration of ferrocytochrome c at time t and C oxd is the concentration of ferrocytochrome c after the addition of ferricyanide, which completes the oxidation of cytochrome c (15). Reaction velocities were calculated from the product of the first-order rate constants and the concentrations of ferrocytochrome c. These were divided by four to obtain the equivalent rates of O 2 uptake. Apparent K m and V max were calculated from the negative reciprocal slope and the intercept with the abscissa, respectively, of plots /S against (Eadie-Hofstee plots), where ϭ turnover number (s Ϫ1 ) and S ϭ ferrocytochrome c concentration (in M).
Absorbance Spectra-Optical spectra (Hewlett-Packard 8453A diode array spectrophotometer) were recorded in stirred solutions at 25°C in a cuvette that was filled via a capillary chimney (1.5 mm diameter, 4 cm length). The spectrophotometer was interfaced to a computer that was used to control data acquisition and to store results. Spectra of the enzyme were recorded at 10-s intervals with an integration time of 1 s in 1-nm wavelength steps. Steady-state turnover of cytochrome c oxidase (1-2 M) was initiated by adding cytochrome c (5 M), TMPD (400 M), and ascorbate (5 mM). As the cuvette solution reached anaerobicity, the cytochrome c oxidase became fully reduced. The anaerobicity of the solution in the cuvette was verified using the phosphorescence quenching method. ] of less than 50 M produced an inhibition of the enzymatic turnover number that was rapidly reversed if the enzyme was subsequently exposed even briefly to oxygen concentrations greater than 50 M.
Conceivably, the oxygen dependence of the cytochrome c oxidase turnover number could have been a consequence of an increase in the apparent K m for cytochrome c and/or an increase in the apparent K m for O 2 . The apparent K m with respect to cytochrome c was therefore determined from Eadie-Hofstee plots using polarographic measurements of O 2 consumption in the presence of different concentrations of ferrocytochrome c, with TMPD and ascorbate added to reduce ferricytochrome c (Fig. 2, A and B). nM) was measured spectrophotometrically from the change in ferrocytochrome c concentration over time. In these studies, TMPD and ascorbate were not added, so the concentration of ferrocytochrome c decreased with time. As shown in Fig. 3, the oxidase exhibited first-order kinetics, and the first-order rate constant decreased as the initial cytochrome c concentration was increased from 0.8 to 20 M after 4 h incubation, regardless of whether the [O 2 ] was 135 M (Fig. 3A) or 25 M (Fig. 3B). Enzyme incubated at [O 2 ] ϭ 135 M yielded a V max and apparent K m for cytochrome c of 57 s Ϫ1 and 1.53 M, respectively (Fig. 3C), under the conditions of this assay. The V max and the apparent K m for cytochrome c of enzyme incubated at [O 2 ] ϭ 25 M were 38.5 s Ϫ1 and 2.8 M, respectively (Fig. 3D). As expected, these values differed from the corresponding measurements made in the presence of TMPD plus ascorbate (see Fig.  2) due to the ability of TMPD to reduce ferricytochrome c that is still bound to cytochrome oxidase (1). Collectively, these results demonstrate that prolonged exposure to [O 2 ] Ͻ50 M results in a reversible transition of the enzyme from its native state to a relatively inhibited (conformed) state associated with a decrease in V max and an increase in the apparent K m for oxygen and cytochrome c.
To determine whether the redox state of the enzyme during cycling was altered after prolonged exposure to [O 2 ] ϭ 25 M, absorbance at 444 nm was measured in the oxidized state, during cycling in the presence of excess cytochrome c, and in the reduced state after reaching anaerobicity. As shown in Cytochrome c, TMPD, and ascorbate then were added to initiate cycling of the enzyme, and spectra were again recorded (18). After the solution reached anaerobicity, the enzyme became fully reduced and spectra were again recorded.

DISCUSSION
Collectively, our data suggest that molecular oxygen directly modifies the V max and the apparent K m of cytochrome c oxidase, leading to a reversible inhibition of the enzyme after it is exposed to low oxygen concentrations (Ͻ50 M) for several hours. To interpret our data we employed Minnaert's mechanism IV, which applies first-order kinetics to explain cytochrome c oxidase kinetics (19). Typically, the first-order rate constants of a simple bimolecular collisional mechanism remain constant with increasing substrate concentrations until V max is reached at finite substrate concentration (20). However, although cytochrome c oxidase kinetics show that the enzyme remains strictly first-order even at high cytochrome c concentrations, the rate constant decreases with increasing cytochrome c concentration (17). Minnaert explained this phenomenon by proposing that ferro-and ferricytochromes have equal binding capacity, with ferrocytochrome forming a productive complex while ferricytochrome can act as a competitive inhibitor. Subsequently, Yonetani and Ray (21) demonstrated that under conditions where first-order kinetics are observed, the K m for ferrocytochrome c was the same as the dissociation constant (K i ) for ferricytochrome c (21).
Minnaert mechanism IV: where [S] ϭ ferrocytochrome c, [P] ϭ ferricytochrome c; (K m ϭ K i ), and k 1 , k 2 , k 3 ϭ electron transfer rates. To apply this model to the interpretation of kinetic data obtained after incubation at low oxygen concentrations, it is necessary to demonstrate that these conditions hold true during both the native and the conformed states of the enzyme. Fig. 3 shows that cytochrome c oxidase exhibits first-order kinetics after incubation at [O 2 ] ϭ 135 M (Fig. 3A) or at [O 2 ] ϭ 25 M (Fig. 3B), thus confirming this assumption. Furthermore, the first-order rate constant was observed to decrease with increasing cytochrome c concen- tration under both incubation conditions. Thus, the assumption for mechanism IV regarding the cytochrome c dependence of enzyme kinetics described by Minnaert remain valid for the enzyme in its conformed and native states.
How might the oxidase V max decrease and the apparent K m for cytochrome c and oxygen increase during prolonged exposure to [O 2 ] Ͻ50 M? Based on Minnaert's mechanism IV, the observed changes could have resulted either from changes in k 1 , k 2 , or k 3 . The rate constant k 3 represents the electron transfer step from a 3 . k 3 is decreased).
These observations are consistent with the hypothesis that molecular oxygen interacts with cytochrome c oxidase to modify its kinetic behavior. The need for prolonged (Ͼ1 h) incubation at low oxygen concentrations to induce conformance, and the rapid reversibility of the inhibition when the enzyme is exposed even briefly to oxygen concentrations greater than 50 M, suggests that a second binding site for oxygen exists on the enzyme, which is distinct from the catalytic binuclear site. When oxygen is bound to this regulatory site the enzyme would assume its normal catalytic activity. However, when incubated at lower oxygen concentrations, O 2 would gradually dissociate from this putative site, causing the enzyme to revert to its inhibited (conformed) state. Exposure to higher oxygen concentrations would result in the rapid reassociation of O 2 , returning the enzyme to its native state. While supportive of this hypothesis, the data of our study do not directly demonstrate the existence of a regulatory site capable of interacting with oxygen. However, the enzyme did appear to revert to its conformed state more quickly when incubated at lower O 2 concentrations, consistent with the notion that oxygen must dissociate from the enzyme to revert to the conformed state. Cytochrome c oxidase has been the focus of extensive analysis over the past few decades, and a large number of different factors are known to influence its kinetic behavior (2,3,5). However, a regulatory effect of oxygen on the catalytic function of the enzyme has not been identified previously. The reason for this may relate to the necessity for prolonged exposure to low oxygen concentrations, and the rapid reversibility of the conformed state upon exposure to higher O 2 concentrations. Classical studies have shown that cellular respiration remains independent of oxygen concentration above a critically low [O 2 ] of 3-5 M (13,22,23). However, because of technical considerations those studies were carried out using hypoxia of short duration, which would not be expected to reveal effects that require more prolonged exposure to elicit. Indeed, the present study represents an extension of our previous studies of prolonged hypoxia in isolated hepatocytes (12) and cultured cardiomyocytes (24). Those studies demonstrated that prolonged exposure to lowered oxygen concentration was associated with a reversible inhibition of cellular respiration, an effect that was similarly evident during TMPD respiration in isolated mitochondria incubated at low oxygen concentrations (6). The previous observation that intact cells of diverse phenotype demonstrate an O 2 -dependent regulation of respiration, combined with the present observation that an inhibition of cytochrome c oxidase activity occurs during prolonged exposure to lowered oxygen concentrations, suggests that this response may represent a component of an adaptive response to lowered oxygen availability. A regulatory effect of molecular oxygen on cytochrome c oxidase would provide a mechanism allowing cells to adjust their respiratory rates in accordance with the ambient oxygen concentrations and could explain how specialized cells within intact tissues are able to respond to changing oxygen concentrations over the physiological range (25). Further elucidation of the significance of this response will require additional studies.