Impaired proton pumping in cytochrome c oxidase upon structural alteration of the D pathway

Abstract

Cytochrome c oxidase is a membrane-bound enzyme, which catalyses the one-electron oxidation of four molecules of cytochrome c and the four-electron reduction of O₂ to water. Electron transfer through the enzyme is coupled to proton pumping across the membrane. Protons that are pumped as well as those that are used for O₂ reduction are transferred though a specific intraprotein (D) pathway. Results from earlier studies have shown that replacement of residue Asn139 by an Asp, at the beginning of the D pathway, results in blocking proton pumping without slowing uptake of substrate protons used for O₂ reduction. Furthermore, introduction of the acidic residue results in an increase of the apparent pK_a of E286, an internal proton donor to the catalytic site, from 9.4 to ~11. In this study we have investigated intramolecular electron and proton transfer in a mutant cytochrome c oxidase in which a neutral residue, Thr, was introduced at the 139 site. The mutation results in uncoupling of proton pumping from O₂ reduction, but a decrease in the apparent pK_a of E286 from 9.4 to 7.6. The data provide insights into the mechanism by which cytochrome c oxidase pumps protons and the structural elements involved in this process.

Introduction

Part of the energy-conversion machinery in living cells is confined to membranes where electron transfer from low-potential donors to high-potential acceptors drives transmembrane proton translocation, maintaining a proton electrochemical potential that is used, for example, for synthesis of ATP. The last component of this electron-transfer chain is cytochrome c oxidase (CytcO), which receives electrons from a water-soluble cytochrome c and donates them to an acceptor, dioxygen, i.e. catalyzing the reaction (for recent reviews, see [1–6]):

$$4{\mathrm{e}}^{-}+{\mathrm{O}}_2+4{\mathrm{H}}^{+}\to 2{\mathrm{H}}_2\mathrm{O}$$ (1)

The electrons are donated from the more positive (P) side of the membrane; while the protons are transferred from the opposite, more negative (N) side. As a consequence, the reaction catalysed by CytcO contributes to maintaining the transmembrane electrochemical gradient. In addition, the process is also linked to pumping of one proton across the membrane per electron transferred to O₂ thereby increasing the overall charge-translocation stoichiometry to a total of 8 charges per reduced O₂ (c.f. 4 additional H⁺ in eq. 1). The subject of the study described in this manuscript is the molecular mechanism by which CytcO links the exergonic electron transfer to O₂, to transmembrane proton pumping.

The Rhodobacter (R.) sphaeroides CytcO (cytochrome aa₃) accommodates four redox-active cofactors: Cu_A, which is the primary electron acceptor from cytochrome c, heme a and the catalytic site consisting of Cu_B and heme a₃ [7, 8]. Two proton pathways have been identified in the R. sphaeroides CytcO. One of these, the K pathway, is used for uptake of two substrate protons upon reduction of the catalytic site, while a second, D pathway, is used for uptake of the remaining six protons [9–11]. Thus, the D pathway is used for uptake of all protons that are pumped across the membrane and two substrate protons. The trajectories for the substrate and pumped protons presumably branch at a highly conserved Glu residue, E286 (Figure 1) at the end of the D pathway. The residue is a transient proton donor/acceptor in the pathway [12, 13] and it has a pK_a of 9.4 [14] (see also the Discussion section), i.e. much higher than that of a Glu in solution. Furthermore, it is likely to adopt different positions during turnover, controlling the proton access from the N-side, and towards the catalytic site or an acceptor site for pumped protons [7, 15–19].

Figure 1. The overall structure of CytcO from R. sphaeroides and of the D pathway.

(A) Subunits (SU) I–IV are shown in different colours as indicated. Heme groups are shown in yellow, copper ions as green spheres and water molecules as red spheres. (B) The four redox-active metal centres are shown together with selected amino-acid residues lining the D pathway and water molecules resolved in the X-ray crystal structure (Protein Data Bank entry 1M56). The figure was prepared using the Visual Molecular Dynamics software [53].

Because during turnover one electron at a time is transferred from cytochrome c to the catalytic site and reduction of O₂ to H₂O requires four electrons, a number of different intermediate states are formed at the catalytic site. These are commonly denoted by one-letter codes and here the superscripts denote the number of electrons transferred to the catalytic site: O⁰ → E¹ → R² → P² (P_M) → F³ → O⁴ (equivalent to O⁰) (recently reviewed in [1, 4]). One approach to study electron and proton-transfer reactions in CytcO is to start from the four-electron reduced state and then initiate the reaction with O₂ by flash-induced dissociation of a blocking CO-ligand at the catalytic site. In this case, after binding of O₂ to heme a₃ a state is formed (τ ≅ 30 μs), which is structurally similar to P² (formed during turnover), but in which one additional electron, transferred from heme a, resides at the catalytic site. Accordingly, this state is denoted P³ (P_R). Because of the excess negative charge at the catalytic site, next, a proton is taken up from the N-side solution, which results in formation of state F³ with a time constant of ~100 μs (at pH 7). The transition from P³ to F³ is also linked to pumping of one proton across the membrane [20, 21]. In the last step of the reaction the electron originally residing at Cu_A is transferred to the catalytic site together with an additional proton from the N-side forming the oxidised (O⁰) enzyme with a time constant of ~1 ms (at pH 7), which is also linked to proton pumping across the membrane.

To investigate the mechanism by which CytcO couples electron transfer from cytochrome c to O₂, to proton translocation, a large number of structural variants of the enzyme have been studied in which specific residues have been altered using site-directed mutagenesis. One class of these structural variants is that in which O₂ is reduced to water, but the catalytic reaction is not linked to proton pumping (often referred to as “uncoupled mutants”). This class of mutant forms is particularly interesting when addressing the proton-pumping mechanism because understanding the origin of the uncoupling at a molecular level would provide valuable information on the structural and functional design of the pumping machinery. In some cases the turnover rate of the uncoupled mutants is dramatically slowed due to impaired proton uptake [22–24]. In these cases the uncoupling is likely to be due to delayed protonation of a “pump site” such that a proton is delivered to O₂ at the catalytic site before transfer of a proton to the pump site can occur [25–31]. However, there is also a class of uncoupled mutant forms of CytcO in which the O₂-reduction rate is essentially unaffected. Such structural variants were first identified in the Paracoccus (P.) denitrificans CytcO [25]. Some of these corresponding mutant forms were investigated in detail in the R. sphaeroides CytcO. In the case of the N139D (Figure 1) mutant CytcO, none of the electron or proton-transfer rates during reaction of the reduced enzyme with O₂ were slowed. However, the pK_a in the pH-dependence of reaction steps involving proton uptake through the D pathway was increased by ~1.5 units [26]. This pK_a was tentatively attributed to residue E286 (see also the “Discussion section”) and it was suggested that this increase in the pK_a was the reason for uncoupling of proton pumping from O₂ reduction. A similar behaviour was also observed upon replacement of Asn207, located near N139 (Figure 1), by an Asp [29]. Because in both mutant forms a negatively charged residue was introduced within the D pathway, one possibility would be that the E286 pK_a is increased due to electrostatic interactions with the introduced additional charge, even though the uncoupling could also be caused by structural alterations within the D pathway [26]. The explanation that electrostatic interactions are responsible for the uncoupling in e.g. the N139D mutant CytcO was questioned based on a theoretical analysis [32]. In an attempt to clarify this issue, in this work we investigated the reaction with O₂ of the N139T mutant CytcO in which a neutral residue is introduced into the D pathway and which displays a significant (~30 %) O₂-reduction activity that is uncoupled from proton pumping. Also in an earlier study with the P. denitrificans CytcO a similar mutation, N139V (N131V in the P. denitrificans CytcO, [25]) resulted in uncoupling of proton pumping, however, that mutant CytcO displayed an activity of only 6 % of that of the wild-type CytcO. Here, we investigated the pH dependence of specific electron and proton transfers upon reaction of the reduced N139T enzyme with O₂ and found that the E286 pK_a was shifted to lower values.

Materials and Methods

The N139T mutant was constructed using the QuikChange site-directed mutagenesis kit (Stratagene) as described in [29]. His-tagged wild-type and mutant CytcOs were purified from R. sphaeroides using histidine affinity chromatography as described previously and the steady-state activity was measured at pH 6.5 as described in [29], and found to be 390 s⁻¹ for the N139T mutant CytcO, i.e. 40% of that of the wild-type CytcO under the same conditions. Proton pumping was measured as described in [29] and no pumping was observed with the N139T mutant oxidase. The liposome-reconstituted CytcO displayed a respiratory-control-ratio (RCR) of >1, i.e. the turnover rate in the presence of a membrane potential was slower than that without a membrane potential.

Preparation of the fully reduced CO-bound CytcO

The CytcO buffer was exchanged on a PD-10 column (Amersham Biosciences) for 100 mM KCl, 0.1% DDM (n-dodecyl-β-D-maltoside) at pH 7.5, with a final enzyme concentration of 10 μM. The sample was transferred to an anaerobic cuvette, which was repeatedly evacuated on a vacuum line and flushed with N₂ gas. The CytcO was reduced with 2 mM ascorbate and 0.5 μM PMS (redox mediator), and the N₂ was exchanged for CO, which results in formation of the CytcO-CO complex.

Flow-flash kinetic measurement

The CytcO-CO complex was rapidly mixed at a 1:5 ratio with an O₂-saturated buffer-solution in a stopped-flow apparatus. The pH after mixing was set by the O₂-buffer solution, which contained one of the buffers MES, HEPES, Tris-HCl, CHES or CAPS (each at a concentration of 0.1 M), depending on the pH, and 0.1% DDM. About 300 ms after mixing, the CO ligand was dissociated using a 5–10 ns laser flash at 532 nm (Quantel, Brilliant B), and the reaction was monitored by recording the absorbance changes. The locally modified flow-flash apparatus (Applied Photophysics, Leatherhead Surrey, UK) used in these measurements is described in more detail in [11]. The cuvette path length was 1.00 cm. The data were analysed using the PRO-K software (Applied Photophysics).

Proton uptake measurements

For the proton uptake measurement the enzyme solution was supplemented with the pH dye phenol red at a concentration of 40 μM. The reaction was initiated upon mixing the CytcO with an O₂-saturated solution containing 100 mM KCl supplemented with 40 μM phenol red at pH 7.8, and the proton uptake was monitored at 560 nm. The same measurements were then repeated with 100 mM HEPES and the trace obtained with the buffered solution was subtracted from that obtained without buffer to remove any absorbance contributions from the enzyme during the reaction. For a more detailed description of the procedure see [33].

Electrometric measurements

Voltage changes associated with charge movement across CytcO oriented in a membrane were measured using an experimental set-up that has previously been used by Wikström, Verkhovsky and colleagues [34]. The set-up consists of a measuring cell of two ~1 ml compartments, separated by a lipid-impregnated (100 mg Soybean IIS lecithin/ml n-Decane) Teflon film. Liposome-reconstituted CytcO [35, 36] was attached to one side of the Teflon film by adding 15 mM CaCl₂ in 0.1 M MOPS, pH 7.5. After 2–3 hours of incubation excess CytcO-vesicles was removed by gradually exchanging the vesicle-containing solution by pure buffer (a mixture of equal amounts of BisTris propane, CHES and CAPS at a total concentration of 0.1 M). Ag/AgCl electrodes (DRIREF-2SH, World Precision Instruments) were inserted into each of the two compartments to record the membrane potential generated across the Teflon film. The measuring cell was placed in an airtight chamber and the atmosphere in the chamber was exchanged, first to N₂ and then to 100 % CO. The enzyme molecules were reduced, and the chamber was kept anaerobic, by adding 50 mM glucose, 0.12 mg/ml glucose oxidase, 75 μg/ml catalase and 50 μM N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD) as redox mediator, to both compartments. Prior to measuring, the pH on the inside of the liposomes was allowed to equilibrate with the pH on the outside for about 1 hour. An airtight syringe was used to inject 50 μl of O₂-saturated buffer towards the Teflon film about 0.2 s before starting the reaction by photolysis of the CytcO-CO complex with a laser flash. The voltage changes were fitted to a sum of exponentials and the amplitudes were determined using a procedure developed by Siletsky et al. [37].

Results

Reaction of the fully reduced CytcO with O₂

Figure 2 shows absorbance changes after initiation of the reaction of the four-electron reduced N139T mutant CytcO with O₂. For comparison, results with the wild-type CytcO are also included. At 580 nm (Figure 2a) with the wild-type CytcO the increase in absorbance with a time constant of ~130 μs at pH 9.0 is associated with formation of the F³ state. With the N139T mutant CytcO the increase in absorbance was slowed by a factor of ~15 to ~2 ms. The high-pH data are shown in the figure because here the difference between the wild-type and mutant CytcO was most significant. The decrease in absorbance on a slower time scale with the wild-type CytcO is associated with the F³ → O⁴ transition [14].

Figure 2. Absorbance changes associated with the reaction of the fully reduced CytcO with O₂.

(a) At 580 nm the increase in absorbance is associated with the P³ → F³ transition. The data show that at pH 9.0 the transition is slowed in the N139T mutant CytcO. The decrease in absorbance seen with the wild-type (WT) CytcO is associated with the F³ → O⁴ transition. (b) Absorbance changes at 445 nm with the wild-type CytcO at pH 6.5 and 9.0. The slow decrease in absorbance is associated with the F³ → O⁴ transition. (c) Absorbance changes at 445 nm with the N139T mutant CytcO at pH 6.5 and 9.0. (d) Absorbance changes at 830 nm with the N139T mutant CytcO, where the increase in absorbance is associated with oxidation of Cu_A on the time scales of the P³ → F³ and F³ → O⁴ transitions, respectively. Experimental conditions were: 20 mM KCl, 0.1% n-dodecyl-β-D maltoside 0.1 M MES (pH 6.0) or 0.1 M CHES (pH 9.0). The CytcO concentration was 2 μM, but all graphs are normalised to 1 μM reacting enzyme. The temperature was 22 °C.

In Figure 2bc the decrease in absorbance at 445 nm in the time window >0.5 ms is associated with the decay of the F³ state and formation of the oxidised CytcO (F³ → O⁴ transition). With the N139T mutant CytcO the time constants of this transition were 1 ms and 12.5 ms at pH 6 and 9, respectively (Figure 2c). For comparison the corresponding data obtained with the wild-type CytcO are shown in Figure 2b, where the time constants are 1 ms (pH 6) and 3 ms (pH 9).

Figure 2d shows absorbance changes at 830 nm at pH 7.8 with the N139T mutant CytcO, which are associated with reduction-oxidation changes of Cu_A. As seen in the figure, the absorbance increased with time constants of 350 μs (70 % of the total amplitude) and 2 ms. Thus, as with the wild-type CytcO (not shown, [38]), at pH 7.8 Cu_A was oxidised in two distinct phases with time constant corresponding to those associated with the P³ → F³ and F³ → O⁴ transitions.

Figure 3 shows absorbance changes of the dye phenol red associated with proton uptake (increase in absorbance at 560 nm) from solution during reaction of the reduced CytcO with O₂ at pH 7.8. As seen in the figure, protons were taken up with time constants of 350 μs (53 % of the total amplitude) and 2 ms (47 % of the total amplitude), which shows that the P³ → F³ and F³ → O⁴ transitions were both linked to proton uptake from solution. Also these data are qualitatively consistent with those obtained previously with the wild-type CytcO [38], although the time constants are larger than those obtained with the wild-type CytcO at the same pH.

Figure 3. Absorbance changes of the pH-dye phenol red at 560 nm upon reaction of the fully reduced N139T mutant CytcO with O₂.

The increase in absorbance occurs over the same time scale as the P³ → F³ (350 μs, 53 % of the total amplitude) and F³ → O⁴ (2 ms, 47 %) transitions. Experimental conditions: 0.1 M KCl, 0.1% n-dodecyl-β-D maltoside, 40 μM phenol red, 50 μM EDTA, pH ~7.8.

Electrical measurements of charge translocation in the N139T mutant CytcO

Figure 4 shows voltage changes across a membrane containing oriented CytcO molecules after initiation of the reaction of the four-electron reduced CytcO with O₂ [34]. For the wild-type CytcO at pH 6.5 an essentially electrically silent phase was observed during the first ~50 μs after which the membrane potential developed in three phases. The first phase was attributed to charge translocation during the P³ to F³ transition (τ ≈ 170 μs, amplitude 60%) and the remaining biphasic process to the F³ to O⁴ transition (τ ≈ 0.7 ms, amplitude 20%, and τ ≈ 2.9 ms, amplitude 20% at pH 6.5). These data obtained with the wild-type CytcO are in agreement with those obtained previously under similar conditions [21].

Figure 4. Membrane potential generation during the reaction of fully reduced CytcO with O₂.

Liposomes containing either wild-type (WT, grey) or N139T mutant (black) CytcO were attached to a lipid-impregnated Teflon membrane separating two liquid-containing compartments and charge translocation was measured across electrodes immersed in these compartments as described in the Materials & Methods section. The inset shows the same data on an expanded time scale.

For the N139T mutant CytcO at pH 6.5 (Figure 4) the first phase, associated with the P³ → F³ transition, displayed time constants of ~130 μs and an amplitude of ~35% of the total signal, i.e. the amplitude was significantly smaller that that observed with the wild-type CytcO (note that the total signal corresponds to transfer of less charge in the N139T mutant CytcO than in the wild-type CytcO because the former does not pump protons). The two following, slower phases displayed time constants of 1.1 ms and 5.9 ms (amplitudes of 40% and 25%, respectively).

pH dependence of the transition rates

Figure 5 shows the pH dependence of the P³ → F³ and F³ → O⁴ transition rates as determined from absorbance changes at 445 nm, 580 nm and 830 nm with the wild-type [14, 39, 40] and N139T mutant CytcOs. As seen in Figure 5a at low pH the P³ → F³ rate was approximately the same with both CytcOs (1·10⁴ s⁻¹), which suggests that the D pathway is intact in the mutant CytcO (see also Discussion). At higher pH the rate decreased with increasing pH displaying a pK_a of 7.6 with the N139T mutant CytcO. This value is significantly lower than that observed with the wild-type CytcO (pK_a = 9.4 [14]).

Figure 5. pH dependence of the P³ → F³ and F³ → O⁴ transitions.

The pH-dependence of the P³ → F³ (a) and F³ → O⁴ (b) transition rates with wild-type (grey squares) and N139T (black circles) CytcO. The transition rates were obtained from traces measured at 580 nm and 445 nm. For the N139T mutant CytcO the data were fitted with standard titration curves (solid lines) with a pK_a of 7.6 in both a and b. For the wild-type CytcO the pK_as were previously determined to be ~9.4 (a) and two pK_as of 8.9 and ≤6.7 (b) (see text). Experimental conditions were the same as in Figure 2, except that different buffers, MES, HEPES, Tris-HCl, CHES or CAPS were used (each at 0.1 M) depending on pH.

Qualitatively, the same behaviour was observed for the F³ → O⁴ transition where at low pH the rate was ~10³ s⁻¹ and it decreased with increasing pH with both the wild-type and N139T mutant CytcO (Figure 5b). The decrease occurred at lower pH for the N139T mutant than for the wild-type CytcO.

Discussion

In the wild-type CytcO the P³ → F³ transition (τ ≅100 μs at pH 6) involves only proton uptake from solution and it is not associated with simultaneous electron transfer to the catalytic site (the electron is transferred already upon formation of the P³ state). The rate of this transition is, thus, determined only by proton transfer via the D pathway to the catalytic site. On the same time scale the electron at Cu_A equilibrates with heme a, which results in oxidation of ~50% of Cu_A [38]. The other transition investigated here, F³ → O⁴ (τ ≅ 1 ms at pH 6) involves both electron and proton transfer to the catalytic site, where the proton is also transferred through the D pathway. In the wild-type CytcO both the P³ → F³ and F³ → O⁴ transitions are linked to proton pumping across the membrane.

In the following discussion we focus on the P³ → F³ transition because it is not linked to simultaneous electron transfer and, thus, reflects only proton transfer within CytcO. As seen in Figure 5, significant differences were observed in the pH dependencies of rates measured with the N139T mutant and wild-type CytcOs in the pH range above pH ~7.5. Results from earlier studies indicated that in this transition residue E286 is an internal proton donor to the catalytic site [12], which was also directly demonstrated in recent time-resolved FTIR experiments [13]. We have previously modelled the pH dependence of the P³ → F³ rate, k_PF, in terms of the fraction of protonated E286, α_EH, multiplied by the proton-transfer rate from E286 to the catalytic site, k_H [14]:

$$k_{\text{PF}}={\alpha }_{\text{EH}}{k}_{\mathrm{H}}=\frac{k_{\mathrm{H}}}{1+{10}^{\text{pH}-\mathrm{p}{K}_{\mathrm{E}286}}}$$ (1)

where pK_E286 is the pK_a of E286. It should be noted that this pK_a is determined from the kinetics of the P³ → F³ transition, which means that the pK_a is an apparent value (see detailed discussion below). The assumption is that proton transfer between solution and E286 is rapid (≫10⁴ s⁻¹) and that the rate-limiting step is the proton transfer from E286 to the catalytic site [14]. In the framework of this model, the identical proton-transfer rates at low pH in the wild-type and N139T mutant CytcOs would indicate that the proton-transfer rate from E286 to the catalytic site is unaltered and that the different behaviour of the mutant CytcO is due to a decrease in the pK_a of E286 from 9.4 in the wild-type to 7.6 in the mutant CytcO.

As with the wild-type CytcO, in the N139T mutant the P³ → F³ transition was linked to proton uptake from solution (Figure 3) and Cu_A was oxidised (to ~70 %) (Figure 2d). Results from earlier studies showed that this electron transfer takes place only if E286 is reprotonated from solution after the initial proton transfer from E286 to the catalytic site [12, 41]. Furthermore, because the sign of the voltage changes measured with the N139T mutant CytcO was the same as that observed with the wild-type CytcO, the results indicate that also in the mutant CytcO protons are taken up from the N-side of the membrane (see comment in [3]).

Results from earlier studies showed that in various mutant CytcOs changes in the pK_a of E286 are correlated with reduced pumping stoichiometries or impaired proton pumping. In these studies, mutant CytcOs were investigated in which charged residues were introduced near or in the D pathway. For example, replacement of Asn139 in the lower part of the D pathway (see Figure 1) by an Asp (N139D mutant CytcO) resulted in an increase in the E286 pK_a from 9.4 to ~11 [26]. Qualitatively, the same behaviour (i.e. an increase in the E286 pK_a) was observed upon replacement of the nearby Asn207 residue (see Figure 1) by an Asp resulting in a non-pumping N207D mutant CytcO [29]. Additionally, when, in the N139D mutant CytcO, a second mutation was introduced where an Asp residue at position 132 (see Figure 1) was replaced by Asn (i.e. a N139D/D132N double mutant CytcO), the E286 pK_a dropped from ~11 to 9.7 and proton pumping was restored [42].

Furthermore, when E286 was removed and a Glu residue was introduced at about the same spatial location, but on the opposite side of the D pathway (E286A/I112E double mutant CytcO), the pK_as in the P3 → F³ and F³ → O⁴ transitions decreased significantly and the pumping stoichiometry was reduced to ~0.5 H⁺/e⁻ [43, 44]. Yet another example is the non-pumping S197D mutant CytcO, which also displayed a decrease in the observed pK_a in the F³ → O⁴ pH dependence [45]. However, in this case the situation was more complicated because the acidic residue was introduced relatively close to E286 (see Figure 1, ~7 Å) and could presumably itself act as a proton donor [45].

Collectively, the above-discussed results suggest that there is a link between an altered pK_a of E286 and a reduced pumping stoichiometry. Such a pK_a shift may result in an altered pumping stoichiometry because of altered relative proton-transfer rates (or extents of proton transfer) from E286 to the catalytic site and/or to an acceptor for pumped protons. The question is then why the pK_a is shifted in the above-discussed mutant CytcOs. For example, in the cases of the N139D, N207D and S197D mutant CytcOs, charged residues were introduced within the D pathway. Even though the distance between D139 or D207, and E286 is relatively large (~20 Å), in principle the modulation in the pK_a could be explained in terms of electrostatic interactions between D139 or D207, and E286. Such an explanation would also conveniently explain why in the N139D/D132N double mutant CytcO the pK_a was restored to essentially the original value (see also [46], for similar results with the E. coli cytochrome bo₃). On the basis of results from a recent theoretical study, Olsson and Warshel [32] suggested that the uncoupling of proton pumping in the N139D mutant CytcO is due to stabilization of the proton at a water molecule adjacent to E286, which would increase the energetic barrier for proton transfer to the acceptor of pumped protons. This effect would result in selectively slowing the rate of proton transfer to the pump acceptor compared to the rate of proton transfer to the catalytic site, which would lower the stoichiometry of proton pumping. These results were obtained using a dielectric constant (ε) of 10 for the interactions of the charged Asp139 with the water molecules, which is close to the value (ε=8) that would account for the 1.6 units shift in pK_a of E286 upon introduction of a charge at position 139. In another theoretical analysis of the effect of the N139D mutation Xu and Voth [47] noted that when Asp139 was deprotonated (during the proton transfer through the D pathway), an energetic barrier for proton transfer to E286 was created at a “proton trap” region [48, 49] between these two residues.

The actual pK_a of D139 in the N139D mutant CytcO is not known, and the calculated pK_a depends on the assumed value of the dielectric constant [50]. It is important to note that even if Asp139 were protonated, this residue could act as a transient proton donor/acceptor and would carry a charge when transiently deprotonated. This is not likely to be the case for Thr139, which would not act as an internal proton donor/acceptor. Hence, explanations for the behaviour of the N139D mutant CytcO discussed above, either with Asp139 deprotonated or protonated, do not apply to N139T. If the decoupling resulting from both N139T and N139D mutant CytcOs has the same molecular explanation, that explanation is different from those proposed previously for the N139D mutant CytcO. As noted above, the pK_a in the pH dependence of the P³ → F³ transition is an apparent pK_a value determined from the kinetics of the transition. If, for example, the proton transfer from E286 requires a reorientation of the E286 side chain (see Introduction) then the pK_a value would also reflect the equilibrium constant between the two orientations of the residue. In other words, the apparent pK_a value will be sensitive not only to the electrostatic field, but also to the structural environment of the residue (see detailed discussion below). Indeed, in a recent FTIR study Vakkasoglu et al. [51] noted that the E286 environment was altered in the N139D and N207D mutant CytcOs. Even though the observed shift of 2 cm⁻¹ was too small to account for the free energy corresponding to the experimentally observed pK_a shift of 1.5 units [51], a change in the E286 environment may alter the apparent pK_a by altering the equilibrium constant between the two positions of the residue as indicated above. Conceivably, the influence of the N139D and N139T mutations on E286 might originate from changes in the structure of the internal water molecules.

Measurements of time-resolved voltage changes during the reaction of the reduced enzymes with O₂ provides further information that is useful in comparing the wild type and the N139T mutant CytcOs. With the wild-type CytcO, the net voltage observed upon reaction of the fully reduced CytcO with O₂ is the result of the uptake of approximately two substrate protons, electron transfer from Cu_A to heme a and pumping of two protons across the membrane (note that electron transfer from heme a to heme a₃ is essentially parallel to the membrane surface and does not contribute) [21, 34]. The measured voltage is proportional to the transferred charge times the component of the distance that the charge travels perpendicular to the membrane surface. For the wild-type CytcO the total charge is determined as follows [52]: Assuming that heme a and the catalytic site are located a distance of (2/3)d from the N-side surface (d is the thickness of the membrane), then the two substrate protons contribute with 2q · (2/3)d (q is the unit charge), electron transfer from Cu_A to heme a contributes with 1q · (1/3)d and the two pumped protons contribute with 2q · 1d. The total voltage is thus proportional to (11/3) · qd, i.e. ~3.7qd.

Because the N139T mutant enzyme does not pump protons, the measured voltage is the result of only the uptake of substrate protons and electron transfer from Cu_A to heme a, i.e. it is proportional to 2q · (2/3)d + 1q · (1/3)d, which gives 1.67qd. The relative amplitude of the rapid phase (τ ≅ 130 μs) of 35 % for the N139T mutant CytcO, coincident with the P³ → F³ transition, gives 0.35 · (1.67)qd ≅ 0.58qd. To estimate the contribution of proton transfer to this voltage change we note that during the P³ → F³ transition Cu_A is oxidised to ~70% (Cu_A oxidation at 830 nm, Figure 2d), i.e. the contribution from the electron transfer to the rapid phase is 0.7 · (1/3)qd ≅ 0.23qd (the charge travels 1/3 of the membrane thickness). The remaining voltage change is assumed to be due to proton transfer and corresponds to 0.58qd − 0.23qd = 0.35qd.

For the next step of the reaction with O₂, the F3 → O⁴ transition, the voltage corresponds to 0.65 ·1.67qd ≅ 1.1qd (65 % of the total amplitude). Because during this transition the remaining fraction of 30 % Cu_A is oxidised, the contribution from proton transfer is 1.1qd − 0.3 · (1/3)qd = 1qd. Thus, the total voltage attributed to proton transfer is distributed between the two transitions such that 0.35/(0.35+1) = 26% takes place during the P³ → F³ transition and the remaining 74 % during the F³ → O⁴ transition. At first sight, these data are not consistent with the proton-uptake data in Figure 3, which show that equal numbers of protons are taken up during the two phases.

The smaller P³ → F³ voltage change associated with proton transfer in the N139T mutant CytcO can be explained in terms of a shorter distance for the charge transfer. Because a total of two protons (i.e. two positive charges) are transferred from solution to the catalytic site during the reaction from P³ to O⁴ (P³ → F³→ O⁴), the 26 % voltage change during the P³ → F³ transition corresponds to transfer of a positive charge ~52 % (2 · 0.26) of the distance from the entrance of the D pathway and the catalytic site. Considering that the F³ state is formed, the proton must be transferred from within the D pathway to the catalytic site. The identity of the proton donor is not known, but one candidate is a water cluster surrounded by residues S197, S200, S201 and F108 [48, 49]. Another way of viewing the problem is that the negative charge formed after deprotonation of E286 in the wild-type CytcO upon F³ formation, in the N139T mutant CytcO is distributed such that it effectively resides further down the D pathway. The data suggest that the internal proton donor is not reprotonated on the time scale of the P³ → F³ transition in the N139T mutant CytcO presumably because of a slowed proton transfer via the introduced T139 residue. A similar scenario was previously proposed by Wikström and Verkhovsky for the N139D mutant CytcO in which proton transfer to E286 was suggested to be slowed, although not so much as to decrease the turnover activity of the enzyme [3]. In the N139T mutant CytcO the water cluster would be re-protonated on the same time scale as the slower F³ → O⁴ transition, simultaneously with the uptake of the second proton. A question then arises why we observe proton uptake from solution with the same time constant as that of the P³ → F³ transition if proton transfer via the T139 site is slowed. It is clear that this proton uptake has to be non-electrogenic, i.e. it is not transferred perpendicular to the membrane surface, which would imply that the proton is bound at the surface of the protein, presumably due to electrostatic interactions between the water cluster discussed above and residues around the orifice of the D pathway (for example, there are several His residues in this region).

The above-described scenario may also explain the altered pK_a observed in the P³ → F³ and F³ → O⁴ transition because in these transitions the proton is not in rapid equilibrium between E286 and the N-side water solution, but is rather effectively taken from a site further down the D pathway. Furthermore, the scenario may also explain the lack of proton pumping because a slowed proton uptake through the D pathway (see also [3]) would result in protonation of the catalytic site before a proton is transferred to a “pumping element” above the hemes or before relaxation of structural changes required for proton translocation [18].

As noted in the Materials and Methods section, the respiratory-control-ratio of the membrane-reconstituted N139T mutant CytcO was >1. In other words, the turnover rate in the presence of a membrane potential was slower than that without a membrane potential. This result indicates that even in the presence of a membrane potential protons are only taken up from the N-side and not from the P-side of the membrane, as observed previously in other D-pathway structural variants of CytcO [24].

In conclusion, although the N139D and N139T mutant CytcOs manifest the same remarkable phenotype of completely decoupling proton pumping from oxidase activity, it may be that the molecular explanations are distinctly different for the two mutations. Certainly, this is the case if the explanation for decoupling by the N139D mutation is based on an electrostatic effect, such as by shifting the apparent pK_a of E286 or stabilizing an internal proton within the D channel. On the other hand, it may be that the loss of proton pumping is the result of each mutation slowing the rate of proton transfer near the entrance of the D pathway, thus eliminating the prompt protonation of the acceptor in the pump site. If this is the primary origin of the decoupling effect of these mutants, then changes in the apparent pK_a of E286 may be an independent phenomenon and not directly related to the uncoupling action of the mutants.

Abbreviations

CytcO

cytochrome c oxidase

R²

CytcO with a two-electron reduced catalytic site

P³ (also P_R)

the “peroxy” intermediate formed at the catalytic site upon reaction of the 4-electron reduced CytcO with O₂.

F3

“oxo-ferryl” intermediate

O⁰

fully-oxidised CytcO

N-side

negative side of the membrance

P-side

positive side of the membrane; Time constants are given as (rate constant)⁻¹. If not otherwise indicated, amino-acid residues are numbered according to the Rhodobacter sphaeroides CytcO sequence and the residues are found in subunit I