Controlled uncoupling and recoupling of proton pumping in cytochrome c oxidase

Abstract

Cytochrome c oxidase (CcO) is the terminal enzyme of the respiratory chain and couples energetically the reduction of oxygen to water to proton pumping across the membrane. The results from previous studies showed that proton pumping can be uncoupled from the O₂-reduction reaction by replacement of one single residue, Asn-139 by Asp (N139D), located ≈30 Å from the catalytic site, in the D-proton pathway. The uncoupling was correlated with an increase in the pK_a of an internal proton donor, Glu-286, from ≈9.4 to >11. Here, we show that replacement of the acidic residue, Asp-132 by Asn in the N139D CcO (D132N/N139D double-mutant CcO) results in restoration of the Glu-286 pK_a to the original value and recoupling of the proton pump during steady-state turnover. Furthermore, a kinetic investigation of the specific reaction steps in the D132N/N139D double-mutant CcO showed that proton pumping is sustained even if proton uptake from solution, through the D-pathway, is slowed. However, during single-turnover oxidation of the fully reduced CcO the P → F transition, which does not involve electron transfer to the catalytic site, was not coupled to proton pumping. The results provide insights into the mechanism of proton pumping by CcO and the structural elements involved in this process.

In the respiratory chains of aerobic organisms electrons are transferred sequentially from low-potential donors to high-potential acceptors, coupled to proton translocation across the membrane. The process maintains a transmembrane electro-chemical proton gradient, which is used, for example, to drive the synthesis of ATP. The final step in the electron transfer chain is catalyzed by the heme-copper oxidases, oxidizing, e.g., cytochrome c and reducing molecular oxygen to water (O₂+4e^-+4H⁺ → 2H₂O). One member of the heme-copper oxidase superfamily is cytochrome c oxidase (CcO). In the CcO from Rhodobacter sphaeroides (cytochrome aa₃) (Fig. 1A) the electrons are donated from the positive side of the membrane to a copper A center (Cu_A), then onto heme a, and finally to the catalytic site consisting of heme a₃ and copper B (Cu_B), where the O₂ reduction takes place (for recent reviews, see refs. 1-5). The protons needed for the O₂-reduction reaction (“substrate protons”) are taken up from the negative side of the membrane through two proton pathways, the K- and D-pathways, where the K-pathway is used for the uptake of one or two protons upon reduction of the catalytic site, and the remaining protons are transferred through the D-pathway (Fig. 1B). The highly exergonic reduction of O₂ to water also drives the translocation of four protons per O₂ molecule across the membrane, from the negative to the positive side. These “pumped” protons are presumably all taken up via the D-pathway, which means that this pathway is used for the transfer of six or seven protons per turnover (3, 6-11). Because both the substrate and pumped protons are taken up through the D-pathway, there has to be a branching-point distributing protons either to the catalytic site or to the acceptor site for pumped protons (“pump site”). Previous studies on mutant forms of CcO indicate that the branching point is located at Glu-286 (E286; refs. 12-15) (if not indicated otherwise, all residues discussed in this work are found in subunit I). The pK_a of E286 is high [9.4 (16); see also refs. 17-19] and presumably finely tuned to deliver protons to the different locations in the correct sequence. In addition, the proton-pumping machinery requires cycling between states with alternating access for protons either to the negative or to the positive side of the membrane (“gating”). The proton gating could, for example, be accomplished by altering the position of the pump site such that it has a high pK_a when in contact with the input (negative) side and a low pK_a when in contact with the output (positive) side (see, for example, refs. 2 and 20 for a more thorough discussion).

Fig. 1.

Structure and catalytic cycle of CcO. (A) The structure of CcO from R. sphaeroides (50). Subunits I-IV are shown in different colors. Heme groups are shown in blue and the copper ions in purple. (B) Hemes a and a₃,Cu_B, and the residues of the D proton-transfer pathway. Water molecules are shown in red. The residues investigated in this study are highlighted in yellow. (C Upper) A reaction cycle of CcO, where the two- or four-electron reduced enzyme reacts with O₂. For simplicity the Cu_A and heme a sites are not shown. The superscripts indicate the number of electrons transferred to the catalytic site in each state. In the sequence of events during CcO turnover the oxidized CcO receives first two electrons after which O₂ binds to heme a₃ (the outer circle). The remaining two electrons are added one-by-one from an external donor. In the experiments in this study the CcO was initially reduced by four electrons (i.e., also Cu_A and heme a are reduced), and the reaction proceeds through the P³-state (see blue text and arrows). During the P³ → F³ transition there is also fractional electron transfer from Cu_A to heme a (Cu_A ↔ a). (Lower) The P³ → F³ and F³ → O⁴ transitions in the D132N/N139D double-mutant CcO. In the P³ → F³ transition a proton is initially transferred from E286 to the catalytic site (CS) after which E286 is reprotonated by D139. The electron residing at Cu_A equilibrates with heme a (see fraction green color). In the fast F³ → O⁴ transition the electron residing at heme a is transferred to the CS, accompanied by proton transfer from E286. In the slow F³ → O⁴ transition proton transfer to the CS is accompanied by electron transfer from Cu_A, as well as reprotonation of E286 and D139 from solution. The states shown are those obtained after the reactions indicated by arrows have occurred. Protonation and deprotonation of the pumping site is not included in the figure. A and B were prepared by using visual molecular dynamic software (51).

During CcO turnover O₂ binds to heme a₃ when the catalytic site is reduced by two electrons (state R², where the superscript denotes the number of electrons transferred to the catalytic site) forming intermediate A² (Fig. 1C Upper). Then the O-O bond is broken, which results in formation of a ferryl state that is called P². None of these transitions are coupled to proton uptake or pumping. In the next step, proton and electron transfer to the catalytic site results in formation of the ferryl intermediate F³. Finally, another proton and electron are transferred to the catalytic site, resulting in formation of the oxidized enzyme (state O⁴⁽⁰⁾; note that states O⁰ and O⁴ are equivalent because upon transfer of four electrons to O₂ the fully oxidized state of CcO is formed).

If the reaction with O₂ is instead initiated in CcO that is reduced by four electrons, i.e., also Cu_A and heme a are reduced, after O₂ binding to heme a₃ an electron is transferred from heme a to the catalytic site forming a state that is called P³ with a time constant of ≈50 μs (Fig. 1C Upper, first blue arrow). Thus, in contrast to the turnover reaction discussed above, in the P³ state there is excess negative charge as compared with P², and the next P³ → F³ transition involves only proton transfer (τ ≅ 90 μs) to the catalytic site (Fig. 1C Upper, second blue arrow). In the last reaction step, the fourth electron, initially residing at Cu_A, is transferred to the catalytic site concomitantly with proton uptake forming the O⁴ state with a time constant of ≈1.6 ms. Consequently, investigation of the reaction sequence via the P³ intermediate, i.e., starting with the four-electron reduced enzyme, makes it possible to investigate the electron and proton transfer processes separated in time (see A² → P³ and P³ → F³ in Fig. 1C Upper).

Both the P²/P³ → F³ and the F³ → O⁴ transitions are linked to pumping of one proton in each transition (21-24). In addition, the reduction of the oxidized CcO is presumably also associated with pumping (23, 25) maintaining an average proton-pumping stoichiometry of one pumped proton per electron transferred to the catalytic site (26).

In this work we focus on the electron- and proton-transfer control mechanisms used by CcO to maintain a maximum proton-pumping efficiency. We have studied the Asp-132 → Asn and Asn-139 → Asp (D132N/N139D) double-mutant CcO (see Fig. 1B) where two residues in the D-proton pathway have been altered, effectively moving a negative charge a short distance from the entrance of the pathway, up toward the catalytic site. The single-mutant N139D CcO, investigated in previous studies, was fully active, but the O₂-reduction reaction was uncoupled from proton pumping (27-30), which was explained in terms of an elevated pK_a of E286 (from 9.4 to >11) (29). The results from the present study show that in the double-mutant CcO, removal of a negative charge from the D-pathway (the additional D132N mutation) in the N139D mutant CcO restores the pK_a of E286 to the original value, which also results in restored proton pumping. Additionally, even though proton uptake is slowed in the double-mutant CcO, during turnover it pumps protons with the same stoichiometry as the wild-type (WT) CcO; however, in the double-mutant CcO a specific O₂-reduction step, the P³ → F³ transition, is uncoupled from proton pumping.

Results

Multiple Turnover Activity. The turnover activity of the D132N/N139D double-mutant CcO was ≈20% of that of the WT CcO, i.e., ≈250 e^-/sec per CcO molecule at pH 6.5. The proton pumping activity of the double-mutant CcO was the same as for the WT CcO, i.e., four protons pumped per O₂-molecule reduced (data not shown).

To obtain information about the transition rates between specific intermediate states, we investigated the reaction of the fully reduced enzyme with oxygen using the flow-flash technique (see introduction).

The P³ to F³ Transition. The F³-state (see Fig. 1C Upper) is formed upon proton transfer from E286 to the catalytic site, which is followed by rapid reprotonation of the Glu residue from bulk solution. The P³ → F³ transition (τ ≅ 90 μs) was investigated by observing absorbance changes at 580 nm (data not shown; see ref. 16). The pH dependence of this rate, measured with the D132N/N139D double mutant and WT CcO, is shown in Fig. 2A. In the double-mutant CcO the pK_a of E286 is 9.7, which is very similar to that of WT CcO (≈9.4; ref. 16). With the double-mutant CcO at high pH the rate did not approach zero as with the WT CcO, but rather leveled out at a value of ≈6.3 × 10³ s^-1.

Fig. 2.

The pH-dependence of the P³ → F³ (A) and F³ → O⁴ (B) rates during reaction of the fully reduced WT (squares) and D132N/N139D double-mutant (circles) CcO with O₂. The solid (WT) and dashed (mutant) lines are fits of the data with standard titration curves (where k₀ is the rate as pH approaches infinity) with pK_a = 9.4 and k₀ = 0 (WT) and pK_a = 9.7 and k₀ = 6,300 s^-1 (mutant) (A), and pK_a (1) = 6.3, pK_a (2) = 8.9 and k₀ = 120s^-1 (B). Experimental conditions were as follows: ≈2 μM reacting enzyme (normalized to 1 μMCcO), 0.1 M Mes (pH 6.0 or 6.5), Hepes-KOH (pH 7.4), Tris·HCl (pH 8.0), 2-(N-cyclohexylamino)ethanesulfonic acid (CHES) (pH 9.0 or 9.5), or 3-cycloexylamino-1-propanesulfonic acid (CAPS) (pH 10.0 or 10.5), 0.1% DDM, and 1 mM O₂ at 22°C.

The maximum rate of the P³ → F³ transition (at pH < 9) increased from ≈1.1 × 10⁴ s^-1 (τ ≅ 90 μs) with the WT to ≈1.6 × 10⁴ s^-1 (τ ≅ 60 μs) with the double-mutant CcO (Fig. 2 A). For simplicity we refer to the P³ → F³ transition as the 100-μs phase for both the WT and the double-mutant CcOs.

The F³ to O⁴ Transition. The next transition, F³ → O⁴, involves transfer of the fourth electron, initially residing at Cu_A, to the catalytic site, coupled to proton uptake from the bulk solution with a time constant of ≈1.6 ms at pH 7.4. The transition rate is determined from absorbance changes, e.g., at 445 nm. With the WT CcO it displays a more complex pH dependence than the P³ → F³ transition; the absorbance changes were fitted with a titration curve displaying two pK_a values of 6.3 and 8.9 (31, 32) (Fig. 2B). In the double-mutant CcO, after the P³ → F³ transition, the absorbance decreases at 445 and 580 nm (data not shown). This decrease, interpreted as the F³ → O⁴ transition, is biphasic with time constants of ≈0.9 ms (≈40% of the amplitude at 445 nm) and ≈30 ms at pH 7.4, referred to as the 1-ms or fast phase and the 30-ms or slow phase, respectively. The rate of the fast phase is pH-independent, whereas the rate of the slow phase decreases with increasing pH (Fig. 2B).

Oxidation of Cu_A. With the WT CcO, the P³ → F³ transition at the catalytic site is associated with fractional (≈50%) electron transfer from Cu_A to heme a, seen as an increase in absorbance at 830 nm with a time constant of ≈100 μs (Fig. 3A). A similar increase in absorbance during the P³ → F³ transition also is observed with the double-mutant CcO. The remaining oxidation of Cu_A takes place during the 1-ms F³ → O⁴ transition with the WT, and during the slow, 30-ms F³ → O⁴ transition with the double-mutant CcO (Fig. 3A).

Fig. 3.

Oxidation of Cu_A and proton uptake. (A) Absorbance changes at 830 nm mainly associated with oxidation of Cu_A. Experimental conditions were as follows: ≈2 μM reacting enzyme, 0.1 M Hepes-KOH (pH 7.4), 0.1% DDM, and 1 mM O₂ at 22°C. (B) Absorbance changes at 560 nm (of the pH-sensitive dye) associated with proton uptake during reaction of the fully reduced WT and D132N/N139D double-mutant CcO with O₂. Shown is the difference of traces obtained without and with buffer. Experimental conditions were as follows: ≈2 μM reacting enzyme, 0.1 M KCl or Hepes-KOH (for unbuffered and buffered measurements, respectively), 40 μM phenol red (pH 7.6), 0.1% DDM, and 1 mM O₂ at 22°C.

Proton Uptake by Detergent-Solubilized CcO. Absorbance changes of the pH-sensitive dye phenol red during oxidation of the fully reduced CcO are shown in Fig. 3B, where an increase in the absorbance corresponds to an increase in the pH, i.e., proton uptake from solution. As seen in the figure, with the WT CcO there are two phases of proton uptake with approximately equal amplitudes: one that occurs on the same time scale as the 100-μs P³ → F³ transition and one on the time scale of the 1-ms F³ → O⁴ transition. In the D132N/N139D double-mutant CcO, no proton uptake is observed on the time scale of the P³ → F³ transition, and only one proton-uptake phase is seen with a time constant similar to that of the slow 30-ms F³ → O⁴ transition.

Single-Turnover Proton Pumping. Proton pumping by CcO incorporated into phospholipid vesicles was measured as an absorbance decrease of the dye phenol red on the outside of the vesicles (Fig. 4). With the WT CcO, proton pumping takes place concomitantly with the P³ → F³ and the F³ → O⁴ transitions (21, 22, 24). With the double-mutant CcO there is no proton pumping in the P³ → F³ transition and proton release to the outside of the vesicles is only observed in the next, F³ → O⁴ transition with a time constant of ≈140 ms, somewhat delayed compared to the O₂-reduction reaction, which is most likely due to reconstitution of the CcO in a lipid membrane (see ref. 24).

Fig. 4.

Absorbance changes at 556 nm (of the pH-sensitive dye) associated with proton release during reaction of the fully reduced vesicle-incorporated WT and D132N/N139D double-mutant CcO with O₂. Shown is the difference of traces obtained without and with buffer. Experimental conditions were as follows: ≈1.5 μM reacting enzyme, 0.1 M KCl, 50 μM EDTA, 125 μM phenol red, 25 mM sucrose or Hepes-KOH (for unbuffered and buffered measurements, respectively) (pH 7.6), and 1 mM O₂, at 22°C.

Discussion

We have investigated the function of the D-pathway D132N/N139D double-mutant CcO, which offers information about the control mechanisms of electron and proton transfer in CcO, and the molecular design of the proton pumping machinery of the enzyme.

The P³ → F³ Transition. Upon formation of the P³-state an electron is transferred from heme a to the catalytic site. The excess negative charge at this site is neutralized during the next P³ → F³ transition, which involves proton transfer (33) from E286 to the catalytic site. In the WT CcO, E286 is then reprotonated rapidly from the bulk solution such that in the F³ state E286 is protonated (34, 35). We have previously modeled the observed reaction rate of this transition as the fraction protonated E286 times the proton-transfer rate from E286 to the catalytic site, i.e., the maximum transition rate is determined by the proton-transfer rate between the glutamate and the catalytic site (16). With the double-mutant CcO the maximum rates are increased slightly from 1.1 × 10⁴ to 1.6 × 10⁴ s^-1 for the P³ → F³ transition, and from 0.6 × 10³ to 1.1 × 10³ s^-1 for the F³ → O⁴ transition. The increase in rate may be due to the slight increase of the pK_a of E286 in the double-mutant CcO, or it may be due to an altered water structure around E286 (see also ref. 29).

According to the same model, in the WT CcO the apparent pK_a of E286 is 9.4 (16), whereas in the N139D single mutant CcO it increased to >11 (29). The pK_a increase was explained in terms of either electrostatic interactions between D139 and E286 or a structural reorganization of water molecules within the D-pathway (29). In the D132N/N139D double-mutant CcO the pK_a of E286 (≈9.7) is similar to that of the WT CcO. The reason is presumably the removal of the negative charge at position 132 compared with the N139D single-mutant enzyme, thereby restoring the original electrostatic field or the D-pathway water configuration.

In WT CcO there is fractional electron transfer from Cu_A to heme a on the same time scale as the P³ → F³ transition. The results from previous studies have shown that this electron transfer only takes place if E286 is reprotonated after delivering its proton to the catalytic site (refs. 7 and 36; see also ref. 32) such that in the Asp-132 → Asn (D132N) single-mutant CcO there is no proton uptake from solution on the P³ → F³ (100 μs) time scale, and no electron transfer from Cu_A to heme a is observed (37). Although there is essentially no proton uptake from solution with a time constant corresponding to this transition in the double-mutant CcO (see Fig. 3B), fractional Cu_A oxidation is still observed (see Fig. 3A), which suggests that E286 becomes reprotonated by the additional Asp at position 139 and that this residue is initially protonated (see Fig. 1C Lower).

Whereas the rate of the P³ → F³ transition with the WT CcO decreases toward zero with increasing pH, with the double-mutant CcO it levels out at ≈6.3 × 10³ s^-1. This difference is probably due to the additional protonatable group (D139) within the D-pathway of the double-mutant CcO that is able to donate a proton to the catalytic site, similar to what has already been suggested for the bovine heart CcO (38), where the P³ → F³ transition rate levels out at 3 × 10³ to 5 × 10³ s^-1 at high pH (16, 39).

The F³ → O⁴ Transition. The rate of the F³ → O⁴ transition has been modeled as the fraction of the fourth electron residing at the catalytic site multiplied by the proton-transfer rate to the catalytic site (14, 32, 40). With the double-mutant CcO, the F³ → O⁴ transition is biphasic where the fast phase (τ ≅ 1 ms; relative amplitude 40%) displays a pH-independent rate in the range 6-10.5, which indicates that the proton is supplied by an internal donor. Furthermore, because there is no proton uptake from the bulk solution on this time scale or faster, the proton needed for a rapid O⁴-formation, residing transiently on E286, presumably originates from D139 (see Fig. 1C Lower), which is also consistent with an initially protonated D139 (see above).

The amplitude of the fast phase is 40% of the total F³ → O⁴ absorbance change at 445 nm with the WT CcO, which may be explained in terms of a smaller extent of proton transfer (≈0.4 H⁺) from the D-pathway. The extent of proton transfer is diminished presumably because in the previous, P³ → F³ transition, one proton is transferred from the D-pathway to the catalytic site without proton uptake from bulk solution, and it is difficult to fully deprotonate two groups in the D-pathway (both E286 and D139) without charge compensation from solution. The electron needed for fractional O⁴-formation on the 1-ms time scale is transferred from heme a (compare the fractional electron transfer from Cu_A to heme a in the previous, P³ → F³ transition). It is important to note that no additional Cu_A oxidation is observed on this time scale (Fig. 3A; see also Fig. 1C Lower) because of the lack of reprotonation of E286 and excess negative charge in the D-pathway.

The slow F³ → O⁴ transition phase with the double-mutant CcO is associated with both proton uptake from the bulk solution (see Fig. 1C Lower) and proton release to the positive side of the membrane, and the rate increases with decreasing pH (τ ≅ 30 ms at pH 7.4). Thus, the reaction steps are most likely identical to those observed for the F³ → O⁴ transition with the WT CcO. The factor-of-20 slower transition rate is presumably due to the slowed, rate-limiting uptake of protons at the entrance of the D-pathway.

Structural Properties of the D-Pathway Entrance. Another aspect of this study is the comparison between the D132N single-mutant CcO investigated previously (37, 41) and the D132N/N139D double-mutant enzyme. In the first case the entrance of the D-pathway is blocked and the F³ → O⁴ transition is impaired in the major fraction of the CcO population. Introduction of a negative charge ≈6 Å above N132 (the additional N139D mutation) partly reverses the block of the D-pathway entrance, which indicates that the cause of the D-pathway block in the D132N single-mutant enzyme is not only steric but also because of a change in the electrostatic field upon removal of D132. A similar observation was also made with cytochrome bo₃ from Escherichia coli, where proton pumping could be restored in a mutant CcO corresponding to D132N by introduction of a new carboxylate in the vicinity (42).

Proton Pumping. As discussed above, the results from previous studies showed that in the N139D single-mutant CcO the O₂-reduction reaction is uncoupled from proton pumping (27-30). It was proposed that the uncoupling is due to an elevated pK_a of the proton donor E286 from 9.4 to >11, which during CcO turnover results in reprotonation of E286 before protonation of the pump site (29). In this work we show that proton pumping can be restored by introducing an additional mutation in the D-pathway. In this double-mutant CcO the pK_a of E286 is ≈9.7 (see Fig. 2 A), i.e., similar to that of the WT CcO and apparently low enough for protonation of the pump site to take place. Thus, residues at a distance from E286 can presumably modulate the apparent pK_a of the primary proton donor to the catalytic site.

Even though in the D132N/N139D double-mutant CcO the pK_a of E286 is similar to that of the WT CcO and proton pumping during turnover is fully restored, there is no pumping during the P³ → F³ transition. The reason is probably the slowed proton transfer into the D-pathway due to the replacement of D132 by an Asn. After proton transfer from E286 to the catalytic site forming F³ (Fig. 1C Lower), the D-pathway is left with an excess negative charge such that the pK_a of the second proton donor is significantly raised, preventing proton transfer to the pump site. Alternatively, during the P³ → F³ transition the proton from E286 may initially be transferred to the pump site after which E286 is reprotonated by D139, and this second proton at E286 is transferred to the catalytic site. However, because of the slow proton uptake from solution the proton bound at the pump site would not be released to the outside, but rather would end up at E286 rendering the same final protonation pattern as that shown for F³ in Fig. 1C.

As seen in Fig. 4 proton pumping is observed during the slow F³ → O⁴ phase. Because this reaction is associated with electron transfer from Cu_A to the catalytic site, it is presumably representative of every reaction step associated with proton pumping during steady-state turnover. On this time scale the D-pathway is in equilibrium with the bulk solution, and, although significantly slowed, the reaction presumably proceeds along the same sequence of events as in the WT CcO. Thus, the results show that even if the proton transfer from the outside is slowed by a factor of 20 in the double-mutant CcO, full proton pumping is maintained. Furthermore, the results from previous studies showed that proton pumping is maintained even when electron transfer into the catalytic site is slowed because of an increased midpoint potential of Cu_A [the Met-263 → Leu (M263L) mutation in subunit II (36, 43)]. Together, these results show that the electron- and proton-transfer rates are orchestrated such that one cannot take place without the other, independently of the rates of these reactions, which maintains a proton-pumping stoichiometry of ≈1 proton per electron. Furthermore, the results are consistent with a scenario where the electron and proton transfer to the catalytic site is controlled by the release of the pumped protons (44). As discussed above, uncoupling of proton pumping from electron transfer during the F³ → O⁴ transition or during turnover occurs when the pK_a value(s) of an internal group(s) is modified such that the internal proton-transfer rates are altered. In this context it is relevant to further discuss why proton pumping does not take place during the P³ → F³ transition in the double-mutant CcO. As described above, the transition is not associated with any electron transfer to the catalytic site and is likely to be unique for the reaction of the fully reduced CcO with O₂ (cf. ref. 45). Because the electron resides at the catalytic site already before the transition, proton pumping is critically dependent on a rapid proton delivery through the D-pathway.

The scenario outlined in this section was previously summarized in the context of a molecular mechanism for proton pumping by the heme-copper oxidases (46), which is shown in Fig. 5 also to visualize the data obtained from the present study.

Fig. 5.

Schematic picture of a proposed molecular model for the function of CcO. CS, catalytic site; PS, pump site. Proton and electron transfers are shown with red and green arrows, respectively. The first proton is transferred from E286 to the high-pK_a acceptor at the catalytic site (1). Deprotonation of E286 induces structural changes, making the pump site accessible to the N-side surface (the “inside”). In this conformation the pK_a of the pump site is high, and a proton is delivered to PS from the inside (2). E286 is reprotonated from the inside (3), and the structure relaxes back to the original conformation. Protonation of the pump site raises the midpoint potential of the catalytic site, and an electron is transferred to CS via Cu_A and heme a (4). In this way the pumped proton controls the electron transfer. With E286 reprotonated, the pump site is now in contact with the P-side surface (the “outside”) and has a low pK_a; therefore, the proton is expelled to the outside (i.e., pumped) (5). If the pK_a of E286 is altered, the O₂-reduction reaction is uncoupled from proton pumping. If electron transfer to the catalytic site is slowed (M263L mutant CcO) or proton uptake from solution is slowed (D132N/N139D double-mutant CcO), proton pumping is maintained.

In summary, the results show that the uncoupled N139D single-mutant CcO can be recoupled by the additional D132N mutation, ≈35 Å from the catalytic site, which indicates that the effects of the mutations are of electrostatic nature. One of the key characteristics of the enzyme is the pK_a of E286, which should be ≈9.5. Any deviation from this pK_a (or an alteration in the dynamics of E286) results in uncoupling of the proton pump. However, the proton-pumping machinery is extremely robust, and the maximum stoichiometry is maintained even if the rates of electron or proton delivery from the outside are diminished.

Materials and Methods

Steady-State Activity Measurements. Histidine-tagged CcO was purified from R. sphaeroides as described in ref. 47. The catalytic activity was determined from the oxygen consumption rate by using an oxygen-concentration meter (Model 53, YSI, Yellow Springs, OH), where a buffer containing 50 mM potassium phosphate, 0.1% dodecyl maltoside (DDM), 1.1 mg/ml asolectin, 2.8 mM ascorbate, 0.55 mM N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD), and 40 μM cytochrome c (pH 6.5) was mixed with 0.5-4 nM CcO.

Preparation of the Fully Reduced CO-Bound CcO. The CcO buffer was exchanged on a PD-10 desalting column (Amersham Pharmacia Biosciences) for 100 mM KCl, 0.1% DDM, 0.2 μM phenazine methosulfate (PMS), and 0.1 mM Hepes-KOH (pH 7.4), with a final enzyme concentration of 10 μM. The sample was then transferred to an anaerobic cuvette, which was repeatedly evacuated on a vacuum line and flushed with N₂. The CcO was reduced with 5 mM ascorbate, and the N₂ was exchanged for CO (≈1 mM).

Flow-Flash Kinetic Measurements. The CcO-CO complex was rapidly mixed at a 1:5 ratio with an O₂-saturated buffer-solution in a stopped-flow apparatus (Applied Photophysics, Surrey, U.K.) (48). The pH after mixing was set by the buffer solution, which contained 100 mM of either Mes, Hepes-KOH, Tris·HCl, 2-(N-cyclohexylamino)ethanesulfonic acid (CHES), or 3-cycloexylamino-1-propanesulfonic acid (CAPS), depending on the pH, and 0.1% DDM. Approximately 300 ms after mixing, the CO ligand was dissociated by using a short laser flash at 532 nm (Brilliant B, Quantel, Santa Clara, CA), and the reaction was monitored by recording the absorbance changes at a number of single wavelengths. The data were analyzed by using the pro-k software (Applied Photophysics).

Proton-Uptake Measurements. The CcO was prepared as described above, with the exception that buffer was excluded, and 40 μM of the pH-sensitive dye phenol red was added before reduction with ascorbate. After CO-binding to CcO, the pH was adjusted to 7.6. The kinetics of proton uptake was followed at 560 nm, as described in ref. 7.

Proton-Pumping Measurements. Vesicles containing CcO were prepared as described in ref. 49. The enzyme was diluted to 0.5 μM in 50 μM Hepes-KOH, 45 mM KCl, 44 mM sucrose, 1 mM EDTA, and 100 μM phenol red (pH 7.6), and the reaction was initiated upon mixing at a 1:1 ratio with a reduced cytochrome c solution (16 μM, in the same buffer as above) in a stopped-flow apparatus (Applied Photophysics). Proton release was measured in the presence of 2 μM valinomycin and proton uptake in the presence also of 5 μM carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP). The absorbance changes of phenol red were recorded at 558.7 nm, which is an isosbestic point for cytochrome c oxidation.

Single-Turnover Proton-Pumping Measurements. Detergent-solubilized CcO was reconstituted into small unilamellar vesicles by slow detergent removal with Bio-Beads and a PD-10 desalting column according to the protocol described in ref. 24. The buffer was exchanged for 100 mM KCl, 25 mM sucrose, and 50 μM EDTA (pH 7.6), with a final CcO concentration of 3 μM, and the air was removed and replaced by N₂-gas. The sample was reduced with 0.5 μM hexa-amine ruthenium and 3 mM ascorbate, and finally the N₂ gas in the anaerobic cuvette was replaced by CO. The reduced CO-CcO complex was mixed rapidly in the flow-flash apparatus at a 1:1 ratio with an oxygen-saturated buffer containing 100 mM KCl, 50 μM EDTA, 125 μM phenol red, and 25 mM sucrose or Hepes-KOH (for unbuffered and buffered measurements, respectively) (pH 7.6). Proton release during single turnover of vesicle-incorporated CcO was monitored at 556 nm.