Role of aspartate 132 at the orifice of a proton pathway in cytochrome c oxidase

Abstract

Proton transfer across biological membranes underpins central processes in biological systems, such as energy conservation and transport of ions and molecules. In the membrane proteins involved in these processes, proton transfer takes place through specific pathways connecting the two sides of the membrane via control elements within the protein. It is commonly believed that acidic residues are required near the orifice of such proton pathways to facilitate proton uptake. In cytochrome c oxidase, one such pathway starts near a conserved Asp-132 residue. Results from earlier studies have shown that replacement of Asp-132 by, e.g., Asn, slows proton uptake by a factor of ∼5,000. Here, we show that proton uptake at full speed (∼10⁴ s⁻¹) can be restored in the Asp-132–Asn oxidase upon introduction of a second structural modification further inside the pathway (Asn-139–Thr) without compensating for the loss of the negative charge. This proton-uptake rate was insensitive to Zn²⁺ addition, which in the wild-type cytochrome c oxidase slows the reaction, indicating that Asp-132 is required for Zn²⁺ binding. Furthermore, in the absence of Asp-132 and with Thr at position 139, at high pH (>9), proton uptake was significantly accelerated. Thus, the data indicate that Asp-132 is not strictly required for maintaining rapid proton uptake. Furthermore, despite the rapid proton uptake in the Asn-139–Thr/Asp-132–Asn mutant cytochrome c oxidase, proton pumping was impaired, which indicates that the segment around these residues is functionally linked to pumping.

The transfer of a proton from a donor to an acceptor is one of the most common reactions in biological systems (for review, see refs. 1 and 2). These reactions underpin, e.g., energy conversion, where proton-transfer reactions across biological membranes maintain a proton electrochemical gradient that is used as cellular energy storage. The proton gradient is maintained by membrane-bound proteins that pump or translocate protons against an energy gradient. In these systems, after protons are taken up from solution, they are transferred through specific pathways composed of protonatable residues and water molecules, often coordinated by polar residues. Results from earlier studies have shown that proton uptake from solution is assisted by acidic groups near the orifice of a proton pathway, forming a proton-collecting antenna (1–4). Such groups may act to funnel protons toward the proton-pathway entry point. In addition, e.g., Asp residues may act as selectivity filters for protons (5).

One system in which proton-transfer reactions play a key role is cytochrome c oxidase (CytcO). This membrane-bound enzyme catalyzes reduction of dioxygen to water and uses part of the free energy released in this reaction for proton pumping across the membrane. In the aa₃-type CytcO from Rhodobacter sphaeroides, two proton pathways have been identified in the structure (6, 7) and shown to be used for proton uptake from solution (8–10). One of these pathways (D pathway) starts at Asp-132 close to the membrane surface and ends at Glu-286, from which protons are gated either to the catalytic site of the CytcO or to a loading site for protons that are eventually pumped across the membrane (Fig. 1) (11, 12). The other pathway starts near Glu-101 in subunit (SU) II and leads via a conserved Lys-362 (it is therefore referred to as the K pathway) to the catalytic site.

Fig. 1.

X-ray crystal structure [Protein Data Bank (PDB) ID code 1M56] of CytcO from R. sphaeroides showing redox cofactors involved in electron transfer during oxygen reduction and one of the proton transfer pathways (D) used during oxygen reduction and proton pumping. Electrons are consecutively donated by soluble cytochrome c to Cu_A and are then transferred via heme a to the catalytic site composed of heme a₃ and Cu_B. The D pathway starts with an aspartate residue (Asp-132) near the membrane surface and is connected via water molecules and polar residues to Glu-286. The figure was prepared by using PyMOL (57).

Replacement of the negatively charged Asp-132 (13, 14) by, e.g., Asn (D132N) resulted in a dramatically slowed O₂-reduction rate, impaired proton pumping (8, 15), and slowed proton uptake through the D pathway by a factor of ∼5,000 (∼0.5 s compared with ∼100 μs for the wild-type CytcO) (16, 17). Rapid proton uptake and pumping could be partly restored by introduction of an Asp at the site of Asn-139, located ∼7 Å from Asp-132 (N139D/D132N double mutation) (18, 19). Collectively, these results were interpreted to suggest that rapid proton uptake to the D pathway (and proton pumping) cannot be maintained if Asp-132 is removed without introduction of a compensatory negative charge in the vicinity. To test this hypothesis, we investigated the kinetics of proton uptake in the N139T/D132N double-mutant CytcO (20). In this mutant CytcO, the negative charge at Asp-132 is removed, and there is an additional “structural mutation” at Asn-139.

This combination of amino acid residue replacements was investigated because earlier data showed that this CytcO variant displayed full O₂-reduction activity despite the loss of the negative charge at the orifice of the D pathway (Table S1) (20). Furthermore, in the N139T/D132N CytcO, proton pumping was impaired, which raises important mechanistic issues.

Electron- and proton-transfer reactions in CytcO are studied, for example, by measuring absorbance changes of the redox sites and pH-sensitive dyes. The experimental approach to investigate the kinetics of these reactions with a high time resolution involves reduction of all four redox sites, i.e., Cu_A, heme a, heme a₃, and Cu_B. The latter two make up the catalytic site. Carbon monoxide (CO) is bound to the reduced heme a₃, and the CO–CytcO complex is mixed with an O₂-containing solution, after which the CO ligand is removed by means of a laser flash. Upon binding of O₂ to the catalytic site, the O–O bond is broken, concomitantly with electron transfer from heme a to the catalytic site, with a time constant of ∼50 μs forming a state that is called P_R (details of this reaction are described in ref. 21). This reaction is followed in time by proton transfer from Glu-286 (Fig. 1) to the catalytic site with a time constant of ∼100 μs at pH 7, which results in formation of the F state, defined as P_R plus a proton. Over the same time scale, Glu-286 is reprotonated from solution through the D pathway (16, 17). In the final step of the reaction, the fourth electron is transferred to the catalytic site, which is accompanied by proton uptake through the D pathway with a time constant of ∼1 ms at pH 7, yielding an oxidized CytcO (state O).

The results from our studies show that in the N139T/D132N CytcO variant, proton uptake through the D pathway is as rapid as with the wild-type CytcO and is insensitive to changes in pH; i.e., at high pH the rate was faster than with the wild-type CytcO. In addition, because in the double-mutant CytcO the proton uptake rate was unaffected in the absence of Asp-132, we investigated Zn²⁺ inhibition and show that one Zn²⁺ binding site is located near Asp-132. Furthermore, the data indicate that structural changes in the segment around Asp-132 and Asn-139 is functionally linked to the pumping machinery of CytcO.

Results

Reaction Between Reduced CytcO and O₂.

Electron- and proton-transfer reactions associated with the P_R → F and F → O transitions were monitored by measuring absorbance changes at specific wavelengths after flash-induced dissociation of CO from the CO–CytcO complex (Fig. 2 and Table S2). At pH 7.5, the P_R → F reaction (seen as an absorbance increase at 580 nm) with the N139T/D132N variant displayed a rate constant of 1.1 ± 0.1 × 10⁴ s⁻¹ (∼90 μs), which is the same as that observed with the wild-type CytcO (Fig. 2B). The F → O reaction, seen as an absorbance decrease at 580 nm (Fig. 2B) as well as at 445 nm (Fig. 2A), displayed a rate constant of 1,200 ± 100 s⁻¹ (800 μs), which is faster than that with the wild-type CytcO (600 s⁻¹, ∼1.7 ms). The difference between traces at 445 nm in the time range of ∼0.1–0.3 ms (Fig. 2A) is due to a larger extent of heme a re-reduction during the P_R → F reaction (heme a reduction results in an increase in absorbance) with the double mutant than with the wild-type CytcO (see below).

Fig. 2.

Absorbance changes associated with reaction of fully reduced wild-type and N139T/D132N CytcOs with O₂. (A) Absorbance changes at 445 nm associated with oxidation of the heme groups. (B) Absorbance changes at 580 nm where the increase in absorbance in the time range 50–200 μs is associated with the P_R → F transition, whereas the decrease in absorbance is associated with the F → O reaction. Experimental conditions were ∼2 μM reacting enzyme in 100 mM Hepes, pH 7.5, and 0.05% DDM at 1 mM O₂ and 22 °C. The traces have been scaled to 1 μM reacting enzyme.

To further investigate the proton-transfer mechanism in the N139T/D132N CytcO variant, the O₂-reduction reaction was studied in D₂O. The P_R → F and F → O reaction rates in D₂O were 5,500 and 180 s⁻¹ (Fig. S1); i.e., they were slowed by about the same factor as with the wild-type CytcO (22, 23).

pH Dependencies of the P_R → F and F → O Transitions.

With the wild-type CytcO, the P_R → F reaction rate titrates with a pK_a of 9.4, which has been attributed to the apparent pK_a of Glu-286 (Discussion). The F → O rate displays a more complex titration behavior, consistent with the rate being determined by the protonation state of two groups with pK_a values of <6.4 and 8.9, respectively (24). With the N139T/D132N double-mutant CytcO, the maximum rate of the P_R → F reaction at low pH was the same as that observed with the wild-type CytcO. No pH dependence was seen in the pH range of 6–10.5 (Fig. 3A); i.e., the rate with the N139T/D132N CytcO was faster than with the wild-type CytcO. Also, the rate of the F → O reaction was pH-independent and faster than that observed with the wild-type CytcO (Fig. 3B).

Fig. 3.

pH dependence of the P_R → F (A) and F → O (B) transitions in the wild-type and N139T/D132N CytcOs. The rates were extracted from measurements at 445 and 580 nm. Data with the N139T and D132N mutant CytcOs are from refs. 49 and 58, respectively, and are shown for comparison. Experimental conditions were 1–2 μM reacting enzyme; 100 mM MES, Hepes, Tris⋅HCl, CHES, or CAPS; and 0.05% DDM at 22 °C and 1 mM O₂.

Proton Uptake.

Proton uptake through the D pathway was investigated by measuring absorbance changes of a pH-sensitive dye [phenol red (pH 7.8) or m-cresol purple (pH 9.1)] (Fig. 4A). As observed earlier (25), with the wild-type CytcO, two kinetic phases are seen, with time constants corresponding to those of the P_R → F (∼45% of the total amplitude) and F → O reactions (∼55%). With the N139T/D132N double-mutant CytcO, the absorbance changes at pH 7.8 displayed two components with rate constants of 8,000 ± 250 s⁻¹ (τ ≅ 130 μs; ∼70% of the total amplitude) and 1,200 ± 100 s⁻¹ (τ ≅ 800 μs; ∼30% of the total amplitude).

Fig. 4.

(A) Absorbance changes of the pH-sensitive dyes phenol red and m-cresol purple associated with proton uptake at pH 7.8 and 9.1, respectively. (B) Absorbance changes at 830 nm associated with electron transfer from Cu_A to heme a. Experimental conditions for A were 1–2 μM reacting enzyme, 100 mM KCl, 40 μM pH dye, 0.05% DDM, and 1 mM O₂ at 22 °C. (B) The conditions were the same as in Fig. 2, except that the pH for the N139T/D132N CytcO variant was 7.8. The traces have been scaled to 1 μM reacting enzyme.

As shown in Fig. 3A, with the double-mutant CytcO, the P_R → F rate displayed pH-independent kinetics. In principle, this reaction can take place upon internal proton transfer from Glu-286 to the catalytic site (16, 17) without reprotonation of Glu-286 from solution (if this reaction were significantly slowed). This scenario would yield pH-independent kinetics because the reaction would only involve internal proton transfer (26). Consequently, we measured proton uptake from solution at pH 9.1 (Fig. 4A) and found that the absorbance changes displayed two components with rate constants of 1.1 ± 0.1 × 10 s⁻¹ (τ ≅ 90 μs; ∼80%) and 1,200 ± 100 s⁻¹ (τ ≅ 800 μs; 20%). At the same pH with the wild-type CytcO the absorbance changes displayed two components with rate constants of 6,300 ± 300 s⁻¹ (τ ≅ 160 μs; ∼50%) and 250 ± 50 s⁻¹ (τ ≅ 4 ms; ∼50%) (Fig. 4A).

In the wild-type CytcO during the P_R → F transition at the catalytic site, the electron at Cu_A equilibrates with heme a, resulting in a partial oxidation of Cu_A. This electron transfer is seen as an absorbance increase at 830 nm (Fig. 4B). In the next step, F → O, Cu_A becomes fully oxidized. The relative amplitudes (i.e., fraction-oxidized Cu_A) of the two components were ∼57% and ∼43%, respectively. In the N139T/D132N CytcO variant, the two phases displayed rate constants of 8,000 ± 200 s⁻¹ (τ ≅ 130 μs) and 1,100 ± 100 s⁻¹ (τ ≅ 900 μs), with relative contributions of 70% and 30%, respectively. Thus, the extent of electron transfer from Cu_A to heme a in the P_R → F reaction was larger with the double mutant than with the wild-type CytcO, which is also seen as a lag phase (increase in absorbance due to heme a reduction) referred to above in the 445-nm data (Fig. 2A). The data in Fig. S2 show that the amplitude ratio of the two components at 830 nm was essentially pH-independent. Because the extent Cu_A → heme a electron transfer in the P_R → F reaction reflects the extent of proton uptake (27), these data are consistent with a rapid proton uptake also at high pH (see above).

Effect of Zinc on Proton Uptake.

Results from earlier studies have shown that the activity of wild-type CytcO is inhibited by zinc ions (Zn²⁺) (28–35). One of the suggested Zn²⁺ binding sites is located near Asp-132, consistent with structural data of the mitochondrial CytcO (36). It has been difficult to functionally test whether Asp-132 is involved in Zn²⁺ binding in the R. sphaeroides CytcO (and if binding here has any functional effects) because removal of Asp-132 results in dramatically slowed proton uptake. However, the N139T/D132N CytcO is ideal for testing the above-mentioned hypothesis because this structural variant displays rapid proton uptake in the absence of Asp-132. The data show that, whereas with the wild-type CytcO the first proton-uptake phase was slowed by a factor of ∼7 in the presence of 200 μM zinc, no effect was observed with the double-mutant CytcO (Fig. 5; see also Fig. S3).

Fig. 5.

Absorbance changes of the dye phenol red (at pH 7.8) upon reaction between CytcO and O₂ in the presence or absence of zinc. Experimental conditions were ∼1 μM reacting enzyme, 100 mM KCl, 0.05% DDM, 0 or 200 μM ZnSO₄, and 40 μM phenol red at pH 7.8.

We also compared the effect of Zn²⁺ addition on the steady-state activities of the N139T/D132N and wild-type CytcOs. Because proton uptake through the D pathway presumably is not rate limiting for this activity, any effects of Zn²⁺ addition may reflect binding to other sites (32–34). We observed the same decrease in activity upon addition of Zn²⁺ to the N139T/D132N as to the wild-type CytcO, which indicates that zinc binding to the other sites was not affected by the D pathway mutations.

Effect of Chloride on O₂-Reduction Kinetics.

In earlier structural studies of the Asp-132–Ala (D132A) mutant CytcO, a chloride ion was found to replace the missing carboxyl group (37). Because proton uptake is strongly inhibited in the D132A mutant CytcO [by a factor of ∼2,000 (38)], this Cl⁻ ion apparently cannot replace the role of the carboxyl group as a proton acceptor/donor near the D-pathway orifice. Nevertheless, we tested the effect of Cl⁻ removal on the O₂-reduction kinetics with the N139T/D132N CytcO and did not observe any effects.

Discussion

As described above, results from earlier studies have shown that introduction of negative charges near the D pathway orifice in the D132N/A mutant oxidases could restore proton pumping and accelerate proton uptake (18, 19, 39). These data were interpreted to suggest that slowed proton uptake in the D132N single mutant is caused by the removal of negative charge (40). Proton uptake in the D132A CytcO also could be partly restored upon removal of SUIII, which eliminates a significant part of the residues surrounding the D-pathway orifice and presumably allows water molecules to access the proton pathway (38). Even more dramatic effects were observed for the N139C/D132N and N139D/D132N double-mutant CytcOs, where removal of SUIII resulted in restoration of full CytcO activity (19). The data from the present study indicate that the proton-uptake rate through the D pathway could be accelerated by a factor of 5,000, to reach the maximum observed value (∼100 μs), upon introduction of a single structural mutation, N139T in the D132N mutant CytcO.

Proton transfer through intraprotein pathways is believed to require hydrogen-bonded chains of water molecules and protonatable amino acid residues (41). The D pathway is defined by such a chain, but it may be kinetically interrupted near Asn-139—i.e., ∼7 Å “above” Asp-132 (42). This conclusion was also reached on the basis of an analysis of the water connectivity within the D pathway (Fig. 1) (43), and the results from these calculations indicated that transient orientation changes of the Asn-139 side chain may result in opening of the pathway. Kinetic studies of proton transfer through the D pathway have shown that the functional effects of mutations at the “lower part” of the pathway, which often result in uncoupling of the proton pump, are complex, and their explanation requires detailed calculations of the activation barriers (44–47), where water molecules within the proton pathway play an important role (48).

If proton transfer across the protein segment around site 139 would be rate limiting for proton transfer to Glu286, the overall transfer rate would be determined by the probability for pathway opening around site 139. With the D132N mutant CytcO, the proton-transfer rate would be determined by the product of the probabilities for pathway opening via Asn-132 and -139, which would yield a slow proton uptake. Our molecular dynamics simulations showed that in one configuration of the Thr-139 side chain in the N139T/D132N double-mutant CytcO, there is a chain of water molecules spanning the distance across the 139 site (Fig. 6A). This structural arrangement would presumably yield a rapid proton transfer via Thr-139 such that the overall proton-transfer rate would be determined only by the probability for pathway opening via Asn-132. This interpretation is supported by a comparison of the maximum proton-transfer rates in the N139T and N139T/D132N variants (Fig. 3A) (49). Although this maximum rate remained at a value of ∼10⁴ s⁻¹ upon removal of Asp-132 in the N139T CytcO (to yield the N139T/D132N double mutant), removal of Asp-132 in the wild-type CytcO (i.e., the D132N single mutant) resulted in slowing the rate by a factor of 5,000. These observations indicate that the N139T mutation keeps the pathway open, even upon removal of Asp-132.

Fig. 6.

(A) Representative snapshots from molecular dynamics simulations (carried out as described in ref. 59) of the wild-type and N139T/D132N mutant CytcOs with one of two possible configurations of Thr-139. The enzyme is displayed in cartoon, and selected amino acid residues are shown as sticks. (B) Effect of the N139T/D132N mutations on the protein surface electrostatics. The amino acid substitutions were introduced in the R. sphaeroides CytcO structure (PDB ID code 1M56). The electrostatic potential surfaces and isocontours were calculated, excluding the cofactors, by using PDB2PQR (60) with the Adaptive Poisson–Boltzmann Solver (61) using default settings and visualized in PyMOL (57). The area around Asp-132 is encircled.

However, as noted, the N139T/D132N CytcO does not pump protons (20). Results from earlier studies of mutant CytcOs in which proton pumping is uncoupled form O₂ reduction showed that impaired proton pumping is typically correlated with a shift in the pK_a in the pH dependence of the proton-transfer rate through the D pathway [interpreted as a shift in the apparent pK_a of Glu-286 (50)]. To interpret these observations in the framework of the CytcO structure, we suggested a model in which the Glu side chain is assumed to adopt two configurations (see also ref. 12) with different pK_a values (see ref. 51 for an alternative model). Proton transfer from the Glu to the catalytic site would take place only from one of these configurations. The observed apparent pK_a would then reflect the equilibrium constant of the two configurations and their pK_a values (50, 52). Any changes in the observed pK_a as a result of mutations—for example, around the 139 site—were explained in terms of changes in the energy profile for proton transfer through the D pathway and changes in the equilibrium constant for the two configurations of Glu-286. These changes result, in turn, in changes in the relative rates for proton transfer to an acceptor site for pumped protons and the catalytic site, respectively. It should be noted that in the N139T/D132N double-mutant CytcO, proton transfer from solution to Glu-286 is presumably not slowed because at pH values of >9, it is even faster than in the wild-type CytcO (Fig. 3).

Fig. 6B shows the electrostatic field mapped on the water-accessible surface of the CytcO around the D-pathway entrance. This field changes only slightly upon removal of a negative charge at Asp-132, which indicates that the effect of the mutation on the proton-collecting properties of the surface is small. In reality, the effect may be even smaller than that seen in Fig. 6B because Asp-132 forms a tight hydrogen bond with His-26, which may be at least partly ionized (13, 53). Qualitatively, this effect was also observed by Olkhova et al., although they obtained a lower pK_a of His-26 (14).

Asp-132 is highly conserved within the oxidase family, and it is clear that upon removing this residue, proton transfer is dramatically slowed. However, the data from this study indicate that neither the presence of a protonatable site nor a negative charge at position 132 is critical for maintaining a rapid proton uptake. We reasoned that the role of Asp-132 near the D pathway orifice could be to act as a selectivity filter for protons, similar to the situation observed with a voltage-gated proton channel (5). In other words, the Asp-132 side chain would sterically block the D pathway and at the same time provide a carboxyl group that is able to shuttle protons. To test whether external negatively charged ligands could block the D pathway when the Asp-132 is removed, we measured the kinetics of CytcO oxidation in the presence of a wide range of molecules that potentially could be found in a cell. As seen in Fig. S4 and Table S1, relatively small effects (less than a factor of two change in the rates) were observed when comparing the wild-type to the N139T/D132N mutant CytcO. The only significant difference was found with Zn²⁺ (Fig. 5), but a regulatory effect of free Zn²⁺ is excluded because of the negligible concentration of free zinc in a cell.

In summary, the results show that rapid proton uptake could be accelerated by a factor of 5,000 by a structural mutation in CytcO when a proton acceptor/donor near the D pathway orifice is removed. However, the results also show that rapid proton delivery through the D pathway does not necessarily maintain proton pumping. In other words, structural changes in the protein segment around Asp-132/Thr-139 may impede proton pumping without affecting the observed proton-transfer rate (even if proton transfer is faster than in the wild-type CytcO at high pH; see above). Because the branch point for the pumped protons and those that are delivered to the catalytic site presumably is located near or at Glu-286, the observation of an unchanged proton-transfer rate indicates that the impaired proton pumping in the N139T/D132N double-mutant CytcO originates from changes of the energy profile for proton transfer closer to Glu-286 rather than at residues 132 and 139. In future studies, it would be interesting to address this issue from a theoretical point of view to find a link between structural changes near the orifice of the D pathway, changes in the energy profile for proton transfer through the pathway, and the protonation dynamics of the acceptor site for pumped protons.

Materials and Methods

Expression and Purification of CytcO.

The R. sphaeroides strain carrying mutations that resulted in the N139T/D132N structural alteration was prepared, expressed (20), and purified (54, 55) as described. The CytcO was stored in 100 mM Hepes, pH 7.5, and 0.05% n-dodecyl-β-maltoside (DDM) at −80 °C. The steady-state oxygen consumption was measured by using a Clark-type O₂ electrode. A buffer containing 50 mM K⁺-phosphate, 0.05% DDM, and 1.1 mg/mL asolectin (pH 6.5) was added to the reaction chamber of the O₂ electrode to a final volume of 1.5 mL. Sodium ascorbate (6 mM), N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD, 0.7 mM), and cytochrome c (30 μM) were added to the chamber, and depending on conditions, the solution was supplemented with 200 μM ZnSO₄ and/or 300 μM EDTA. To initiate the reaction, 10 μL of the CytcO solution (1 μM, 100 mM Hepes, and 0.05% DDM, pH 7.5) was added.

Sample Preparation.

The enzyme buffer was exchanged for 1 mM Hepes, pH 7.5, and 0.05% DDM by consecutive concentration and redilution using concentration tubes with a 100-kDa cutoff (Amicon Ultra; Millipore). Finally, the CytcO was diluted to a concentration of 10 μM. To remove O₂ from the sample, the enzyme solution was transferred to an anaerobic cuvette, and the air in the cuvette was exchanged for N₂. Ascorbate (2 mM) and the redox mediator hexamine-ruthenium chloride (1 μM) were added to fully reduce the CytcO. The redox state of the enzyme was monitored by using UV-visible spectroscopy, and when the fully reduced state was obtained, the atmosphere in the cuvette was exchanged for CO. In the measurements of the pH dependence of the reaction kinetics, the reduced CytcO was mixed with O₂-containing buffer solutions (100 mM) of different composition, depending on pH (sets the final pH value): MES, pH 6.0, Hepes, pH 7.0–8.0, 2-(N-cyclohexylamino)ethane sulfonic acid (CHES), pH 9.0, or N-cyclohexyl-3-aminopropanesulfonic acid (CAPS), pH 10.0–10.5, and 0.05% DDM. For control experiments at pH 7.5 and 10.5, the enzyme was kept in buffers containing 100 mM Hepes (pH 7.5) or CAPS (pH 10) and 0.05% DDM. For studies of the effect of zinc addition on the O₂-reduction kinetics, CytcO was repeatedly washed (using concentration tubes; Amicon Ultra; Millipore) with 100 mM Hepes, 0.05% DDM, and 20 mM EDTA to remove any metal traces. The buffer was then exchanged for 100 mM Hepes, pH 7.5, and 0.05% DDM, and the CytcO was diluted to a final concentration of 10 μM. For measurements of proton uptake, the buffer of the enzyme solution was exchanged for a solution containing 100 mM KCl (pH 7.8 or 9.1) and 0.05% DDM. A pH-sensitive dye (phenol red, pH 7.8, or m-cresol purple, pH 9.1) was added to a final concentration of 40 μM, and the enzyme solution was transferred to an anaerobic cuvette. The fully reduced CytcO–CO complex was then prepared as described above. To investigate the possible effect of Cl⁻ on the O₂-reduction kinetics, the CytcO solutions were diluted to a concentration of ∼7 μM in 50 mM K₂SO₄ and 0.05% DDM (pH 7). The samples were then dialyzed (5 × 1 L) in 50 mM K₂SO₄ and 0.05% DDM (pH 7) at 4 °C for ∼2 d. Finally, the buffer was exchanged for 100 mM Hepes and 0.05% DDM (pH 7.5).

Measurements of the Oxygen-Reduction Kinetics and Proton Uptake.

The fully reduced CO-bound CytcO was rapidly mixed at a 1:5 ratio with an O₂-saturated buffer in a stopped-flow apparatus (Applied Photophysics). Approximately 200 ms after mixing, the CO ligand was dissociated by a 10-ns laser flash (Quantel Brilliant Nd-YAG; 532 nm), which enabled O₂ to bind to heme a₃ and react (56). The reaction was followed in time as absorbance changes at specific wavelengths. The kinetic data were analyzed by using the Pro-K software (Applied Photophysics). In measurements of the effect of addition of Zn²⁺ ions on the oxygen-reduction kinetics, the buffer contained 100 mM Hepes, pH 7.5, 0.05% DDM, and ZnSO₄. The proton uptake measurements were performed as described above, except that the absorbance changes of the pH dyes phenol red or m-cresol purple were monitored in a buffer-free solution containing 100 mM KCl, pH 7.8 or 9.1, 0.05% DDM, and 40 μM dye. To remove any contribution from the enzyme at these wavelengths, the measurements were repeated with a buffered solution containing 100 mM Hepes or CHES, 0.05% DDM, and 40 μM pH dye. These traces were then subtracted from those obtained with the unbuffered solution.