A chemically explicit model for the mechanism of proton pumping in heme–copper oxidases

Abstract

A mechanism for proton pumping is described that is based on chemiosmotic principles and the detailed molecular structures now available for cytochrome oxidases. The importance of conserved water positions and a step-wise gated process of proton translocation is emphasized, where discrete electron transfer events are coupled to proton uptake and expulsion. The trajectory of each pumped proton is the same for all four substrate electrons. An essential role for the His-Tyr cross-linked species is discussed, in gating of the D-and K-channels and as an acceptor/donor of electrons and protons at the binuclear center.

Introduction

Some three decades ago it was demonstrated that cytochrome c oxidase was a proton pump (Wikstrom and Krab 1979; Wikstrom 1977) and this pumping activity was later shown to be a property of other members of the heme-copper oxidase superfamily (Wikstrom and Verkhovsky 2007; Wikstrom et al. 1994; Wikstrom 2004; Mills and Ferguson-Miller 2003; Gennis 2004; Branden et al. 2006). There have been many reaction schemes postulated for how these enzymes function; however, no model has been able to explain how all the members of this evolutionarily related, but diverse, family of oxidases are able to pump protons. An examination of recent high resolution crystal structures has led us to a focus on the role of water in the facilitation of proton movement (Sharpe et al. 2005), given the absence of many conserved amino acid residues. We have compared the structure and sequence data of a large number of heme-copper oxidases and have proposed a role for conserved water molecules in oxidase function, along with individual amino acid residues. Based on this comparative analysis, and on previous work (Sharpe et al. 2005), we present a model for a common mechanism of proton pumping in the family of heme-copper oxidases. The key features of our model include:

Water based

The protonation and deprotonation of “conserved” water molecules or clusters is a central component of the pumping mechanism. These waters are critical in the function of the recognized K-channel and D-channels (Konstantinov et al. 1997), as well as in additional hydrated cavities: the X-pathway (the eXit pathway), that facilitates the movement of H⁺_(P) (pumped protons) out of the enzyme; and the vestibule, which is found above the heme propionates at the interface of subunits I and II and links the binuclear center (BNC) to the exit channel.

Electrostatically driven

The driving force for the uptake, internal movement and ejection of H⁺_(P) consists of a series of discrete electrostatic potentials. As each of the four electrons and H⁺_(S) (substrate protons) migrates through the enzyme they generate electrostatic potentials, in each case a movement of H⁺_(P) collapses the potential generated. Ours could be described as a ‘Richelian’ mechanism (Rich 1995) initially proposed by Artzatbanov (Artzatbanov et al. 1978) in which entry of a substrate electron into the protein’s low dielectric causes the (electrophoretic) uptake of a H⁺_(P). Subsequent transfer of the electron from heme a to the BNC forms a protonatable anion, inducing (electrogenic) uptake of a H⁺_(S) into the BNC, that causes the expulsion of the pumped proton from the enzyme (Pisliakov et al. 2008; Fadda et al. 2008).

Multiple gated proton pathways

The movements of charges (H⁺_(S) and H⁺_(P)) in response to electron transfer, are strictly controlled by the use of gates that only open and close at the appropriate time. We describe three gating mechanisms, likely common to all the heme–Cu type oxidases, and an additional gate in Cu_A/Mg type oxidases. The necessity of proton gates has long been recognized in electrostatic pumping mechanisms, and our model is consistent with recently reported evidence of location (Pisliakov et al. 2008; Fadda et al. 2008; Busenlehner et al. 2008).

Use of the histidine-tyrosine crosslinked species

The presence of a covalent bond between the two aromatic rings of H284 and Y288 (His-Tyr) provides a Cu_B ligand with some unusual properties (Wikstrom et al. 1994; Bu and Cukier 2005; Buse et al. 1999; Cappuccio et al. 2002; Muramoto et al. 2007; Rauhamaki et al. 2006). These properties are proposed to allow: (1) changes in the His-Tyr geometry that can gate H⁺_(S) movements into the BNC from the K and D-channels (Sharpe et al. 2005; Bu and Cukier 2005); and (2) the change from coordinate to ionic bonding of His-Tyr to Cu_B which facilitates that protonation and deprotonation of H334.

Generality

The model, with minor changes as to the identity of some semi-conserved amino acids, can be used to describe the mechanism of proton pumping in members of the superfamily regardless of the heme in the BNC, the substrate, or amino acid sequence. The model also explains the phenotypes observed in the large number of site-directed mutants that have been generated in this enzyme.

Molecular consistency

All four H⁺_(P) are taken up and ejected from oxidase in the same manner.

The model

1. Role of water

Table 1 shows the route taken by H⁺_(P) from the inner to outer bulk phases, and the six storage sites at which the H⁺_(P) rests during the course of the reaction cycle. The route taken by the H⁺_(P) is independent of which particular one of the four substrate electrons is being consumed by the enzyme. There are six proposed storage sites, include two conserved residues and four water sites. They are, in order of protonation, ⁽¹⁻⁾D132 (Gennis 2004; Qin et al. 2007; Mills and Hosler 2005; Smirnova et al. 1999, Brzezinski and Adelroth 1998; Fetter et al. 1996), the Protonatable Water Cluster (PWC)(Xu et al. 2007; Xu and Voth 2006), ⁽¹⁻⁾H334 (Ji et al. 2008; Fadda et al. 2005; Popovic et al. 2005; Popovic and Stuchebrukhov 2004), W1_a3(Sharpe et al. 2005), ⁽¹⁻⁾HO-W1_Mg and W1_E254II.

Table 1.

The route taken by H⁺ _(P) in the catalytic cycle

H+(P) Site	1	2	3	4	5	6
Donor	HD132	H+-PWC	HH334	H+-W1a3	W1Mg	H+-W1E254II
Acceptor	(1-)D132	PWC	(1-)H334	W1a3	(1−)HO-W1Mg	W1E254II
Location	D-Channel		Binuclear Center(BNC)		Vestibule	X-Path

There are also two proposed H⁺_(S) binding sites, one located in the K-Channel, H⁺-K362, and one at the apex of the D-channel, HE286. The H⁺-K362 supplies protons (H⁺_(SK)) to the binuclear center following the entry of the 1st and 2nd electrons into the BNC (Mills et al. 2000; Vygodina et al. 1998; Zaslavsky and Gennis 1998; Hosler et al. 1996); it is replenished from the inner bulk. Protons are supplied to the BNC on entry of the 3rd and 4th electrons (Busenlehner et al. 2008), (Vakkasoglu et al. 2006; Olsson and Warshel 2006; Seibold et al. 2005; Heberle et al. 2004; Nyquist et al. 2003; Bailey et al. 2002; Verkhovskaya et al.1997) by HE286 (H⁺_(SD)) and the ⁽¹⁻⁾E286 that is generated is reprotonated by the inner bulk phase.

2. Operation of the Cu_A–Mg gate

The reduction of Cu_A by cytochrome c and its oxidation by heme a is proposed to be linked to the ejection of a H⁺_(P) from the enzyme, providing an extra level of gating and thus efficiency to the oxidases that contain the Mg (Sharpe et al. 2008).

The key to the Cu_A/Mg gate is the common ligand, E254^II, which is bonded to Mg²⁺ so that it is always in the carboxylate state, while the peptidyl carbonyl oxygen is bonded to one of the Cu atoms of the binuclear Cu_A (Fig. 1). Changes in the redox state of Cu_A alter in charge of the Cu_A binding site and affect the ligation of the Mg²⁺ binding site. We propose that upon reduction of Cu_A a pair of neutral waters undergo disproportionation and a H⁺_(P) is transferred from a Mg²⁺ water ligand, leaving a ⁽¹⁻⁾OH–Mg²⁺, to form a hydronium ion that is ionically bonded to E254^II. The reason for this disproportion is the change in the electron density on the carboxylic group of E254^II upon Cu_A redox cycling. When Cu_A becomes reduced, the Cu atom bonded to the carbonyl oxygen of the peptidyl oxygen of E254^II is no longer electron withdrawing and additional electron density is forced onto the carboxylate. Because of its bonding situation, direct protonation is not an option. Instead, this carboxylate forms a salt with a hydronium ion, created by the disproportion reaction to achieve electrostatic neutrality.

Fig. 1.

Shows two R.s. crystal structures a oxidized and b reduced, in which the protonation/redox states of the Cu_A, Mg and associated waters are depicted. Mg⁽²⁺⁾ has three waters in its inner hydration sphere when Cu_A is oxidized; only one of these water molecules, W1_Mg, is able to deprotonate and form a hydroxide, as shown in the reduced Cu_A state. The other two water molecules are hydrogen bonded to the carboxylate group of D229^II and cannot deprotonate. The major difference between the two Cu_A redox states is shown as the formation of an ionic bond between W1_E245II and E254^II. This transition is can be deduced from the contraction of the hydrogen bond between them, from 2.9 Å in the oxidized structure to 2.6 Å in the reduced. The presence of an ionic bond is evidence for the formation of a hydronium ion, H⁺-W1_E245II (magenta) from. Spectroscopic analysis of the Mg(Mn) site is consistent with the formation of H⁺-W1_E245II upon Cu_A reduction, and concomitant deprotonation of W1_Mg, with the formation of ⁽¹⁻⁾HO-W1_Mg (green) and movement of H⁺_(P) as indicated, first to W1_E245II and then to the outside when Cu_A is reoxidized

The basis of this model was a detailed comparison of the high resolution crystal structures of both bovine and R.s. oxidases, to establish the protonation status of the water molecules in the vicinity of Cu_A and Mg²⁺. Figure 1 shows the arrangement and protonation status in the oxidized and reduced states, based on R.s. structures. In the previously published 2GSM structure, Cu_A is oxidized (Fig. 1A), whereas in a reduced structure (Fig. 1B) Cu_A is reduced. Since bond length is a function of bond strength, ionic bonds (<2.7 Å) are shorter than hydrogen bonds (> 2.8 Å). An examination of the distances between E254^II and W1_E254II, in the Cu_A oxidized state gives the distance as 2.9 Å, indicating the presence of a hydrogen bond. In the reduced structure the average distance between E254^II and W1_E254II is only 2.6 Å, indicative of ionic bonding. Similar measurements in the oxidized and reduced structures of bovine oxidase show an even more dramatic change in this bond, >3.7 Å (2DYR, oxidized) versus 2.6 Å (2EIJ, reduced) (Sharpe et al. 2008). The placement of hydrogen and ionic bonds in Fig. 1 is based on the standard geometry of water, hydronium and hydroxide ions, and is fully consistent with the formation of a hydronium/hydroxide pair.

Spectroscopic evidence also supports the existence of ⁽¹⁻⁾OH–Mg²⁺. Using the Mn substituted R.s. oxidase, the Mg/Mn site could be interrogated by a large number of electron spin resonance techniques. We found that the cyanide anion binds to the Mn ion only when Cu_A is reduced, consistent with the displacement of a hydroxide ion but not water. Also consistent with the model are D₂O exchange experiments in combination with ENDOR and ESEEM, which show the reduced enzyme has one less exchangeable inner sphere proton at the Mn site in the reduced, cyanide-bound state (Sharpe et al. 2008).

After Cu_A reduction, the next step in the oxidase cycle is the oxidation of Cu_A by heme a. When Cu_A is oxidized, the E254^II-hydronium salt form is no longer favored. The end result will be the movement of the H⁺_(P) towards the cytochrome c binding site into the bulk aqueous phase (Fig. 1).

3. Operation of the D-gate: Reduction of heme a and H⁺ _(P) uptake

The reduction of heme a by Cu_A is electrostatically linked to the uptake of a H⁺_(P) from the inner bulk phase into the oxidase interior, the energetics of which was described by Rich and others. Simply, forcing an electron into a region with a low dielectric, is thermodynamically costly unless there is the movement of a counter ion. This counter ion flux can consist of a cation entry into the dielectric or the extrusion of an anion out of the dielectric (Rich 1995; Artzatbanov et al. 1978).

The arrangement of amino acids, including alcohols and the carbonyl oxygen atoms of peptide bonds, at the top of the D-channel allows water molecules to form a protonated water cluster (H⁺-PWC). Modeling studies suggest that, in vivo, the proton in the H⁺-PWC is more delocalized and more Eigen-like rather then Zundel-like. In a previous paper we and colleagues conclude from computational analysis that oxidase is capable of using water molecules to stabilize a proton at the top of the D-channel, but did not explicitly state how this capacity might be linked to oxidase turnover (Xu et al. 2007; Xu and Voth 2006). The most obvious linkage is to the reduction of heme a, where electron entry into the low dielectric interior can be electostatically compensated by the movement of a H⁺_(P) into the D-channel from HD132, with the formation of the H⁺-PWC, while ⁽¹⁻⁾D132 is rapidly reprotonated.

4. Reactions at the binuclear center

A key difference of our model, in comparison with previous models, concerns the eventual fate of each of the four substrate electrons of the oxygen reduction reaction. We postulate that the 1st/2nd and the 3rd/4th electron pairs take part in different chemistry; this allows the overall catalytic cycle to be split into two distinct subcycles, the metal reduction phase and the oxygen reduction phase. In the metal reduction phase, the 1st/2nd electrons sequentially enter the BNC, via Cu_A and heme a, reducing Cu_B and heme a₃. In the oxygen reduction phase, the reduced BNC reacts with molecular oxygen forming a ferryl heme a₃ and a protein radical. Reduction of the ferryl and the protein radical by the 3rd/ 4th electrons, completes the cycle.

• i) Initial State Before we can describe the movements of protons and electrons within oxidase we must describe the starting state of the oxidase, including the occupancy of both H⁺_(P)/H⁺_(S) donor and acceptor sites within the enzyme. Our starting point is the fully oxidized R.s. oxidase that has just completed a reaction cycle and is in the state we define as O_H; pre-primed to pump protons. There are likely a large number of oxidized oxidase states, given the number of protons sites that may or may not be occupied, but only one is fully catalytically competent (Brand et al. 2007; Belevich and Verkhovsky 2008; Jancura et al. 2006; Wrigglesworth et al. 1988; Brunori et al. 1981, 1979; Antonini et al. 1977). This state is defined by Wikstrom as O~ or O_H (Wikstrom 2004; Verkhovsky et al. 1999; Bloch et al. 2004).

The two initial H⁺_(S) acceptors of O_H are two hydroxides in the BNC. The two oxygen atoms in the BNC of “resting” oxidase (the “O” or “resting” state) were initially identified as a peroxy-bridge (Mochizuki et al. 1999; Yoshikawa et al. 1998), but that remains controversial (Aoyama et al. 2008). O_H has one less proton in its BNC than does “O”, so the BNC of O_H contains a pair of hydroxides bonded to each other. O_H has a short half-life because of the ease at which this dihydroxide can protonate and form a hydroxide hydrate. This protonation status, where the BNC in the O_H state contains a dihydroxide, has been recently proposed independent of ourselves (Fadda et al. 2008).

The internal H⁺_(S) donors are H⁺-K362, HE286, both protonated in O_H. In O_H, an H⁺_(P) is also on the water molecule coordinated to the Mg²⁺ ion, deposited there during the last catalytic cycle and on HD132.

Below is an outline of the redox reactions that occur in the BNC involving the 1st and 2nd electrons of the metal reduction phase (Fig. 2A) and the 3rd and 4th electrons of the oxygen reduction phase (Fig. 2B).

Fig. 2.

a The mechanism of H⁺_(P) uptake and ejection from the binuclear center (BNC), upon entry of the 1st electron in the metal reduction phase. In the oxidized enzyme, the charge of Cu_B ⁽²⁺⁾ is neutralized by two counter ions, the anionic ligands, ⁽¹⁻⁾OH and ⁽¹⁻⁾H334 (1). Reduction of Cu_B ⁽²⁺⁾ to CuB⁽¹⁺⁾ by the 1st electron causes a localized charge imbalance on CuB so that one of its ionically bonded ligands must protonate (2). Proton gating of the D- and K-channels regulates proton movements into the BNC. The K-Gate is closed, so that electron entry into the BNC must be accompanied by ^(1-)H334 protonation, by H⁺_(P) from the H⁺-PWC. Formation of HH334 triggers a change in the configuration of His-Tyr and opens the K-gate (3). This opening of the K-gate allows a H⁺_(SK) to travel from H⁺-K362 into the BNC, protonating Cu_B ⁽¹⁺⁾-OH⁽¹⁻⁾ and forming Cu_B⁽¹⁺⁾… OH₂, and also closing the K-gate. The formation of CuB⁽¹⁺⁾…OH₂ has deprived CuB⁽¹⁺⁾ of an anionic ligand. The charge balance of CuB⁽¹⁺⁾ is restored by HH334 deprotonation, restoring the ionic ligand ^(1-)H334. The H⁺_(P) is ejected from the BNC and is transferred to W1a₃ (4). The subsequent reprotonation of K362 from the inner bulk generates an electrostatic potential that forces the H⁺_(P) from H⁺-W1a₃ into the ‘proton hole’ previously generated in the vestibule, ^(1-)HO-W1_Mg. Before entry of the 2nd electron into the BNC there is a reorganization reaction, whereby an electron and proton are transferred from Cu_B to heme a₃: ^(1-)HO-a₃⁽³⁺⁾ + ⁽¹⁻⁾H334-Cu_B⁽¹⁺⁾…OH₂ → H₂O…a₃ ⁽²⁺⁾ + ^(1-)H334-Cu_B ⁽²⁺⁾-OH^(1-), the pumping form of CuB in the metal reduction phase (1). The input of the 2nd electron then proceeds as the 1st. Fig. 2 b shows the mechanism of H⁺_(P) uptake and ejection from the BNC upon entry of the 3rd electron in the oxygen reduction phase, initiated by the reaction of the two-electron reduced BNC with molecular oxygen. This generates a ferryl (a₃ ⁽⁴⁺⁾=O⁽²⁻⁾), a radical (His-•Tyr) and CuB in the form ^(1-)H334–Cu_B ⁽²⁺⁾-OH⁽¹⁻⁾ (1). Reduction of His-•Tyr to ⁽¹⁻⁾His-Tyr by the 3rd electron causes a localized charge imbalance on CuB so that one of its ionically bonded ligands must protonate (2). Gating of the D- and K-channels ensures that electron entry is accompanied by the protonation of ⁽¹⁻⁾H334 with a H⁺_(P) from the H⁺-PWC of the D-channel. Formation of HH334 triggers a change in the configuration of ⁽¹⁻⁾His-Tyr and opens up a pathway connecting HE286 to ⁽¹⁻⁾His-Tyr, allowing a H⁺_(SD) to protonate ⁽¹⁻⁾His-Tyr (3). The protonation of ⁽¹⁻⁾His-Tyr deprives Cu_B⁽²⁺⁾ of an anionic ligand, and so HH334 undergoes deprotonation so as to restore the charge balance of Cu_B ⁽²⁺⁾. The deprotonation of HH334 results in H⁺_(P) ejection from the BNC, transferring it to W1a₃ (4). The subsequent reprotonation of ⁽¹⁻⁾E286 by a H⁺_(SD) from the inner bulk generates an electrostatic potential that forces the H⁺_(P) from H⁺-W1a₃ into ^(1-)HO-W1_Mg. Before entry of the 4^th electron into the BNC there is a reorganization reaction, whereby an electron and proton are transferred from His-Tyr to heme a₃: a₃⁽⁴⁺⁾=O⁽²⁻⁾ +(His-Tyr)Cu_B⁽²⁺⁾→ ⁽¹⁻⁾HO-a₃⁽³⁺⁾ + (His-•Tyr)Cu_B⁽²⁺⁾, the pumping form of Cu_B in the oxygen reduction phase (1). The input of the 4th electron then proceeds as the 3rd.

• ii) The 1st and 2nd electrons In O_H, Cu_B ²⁺ is charge balanced by two ionically bonded ligands, ⁽¹⁻⁾OH and ⁽¹⁻⁾H334, both of which are capable of protonation. The ability of H334 (H291 in bovine) to flip between the unprotonated imidazolate (⁽¹⁻⁾H334) and neutral, protonated, imidazole (HH334) states in response to changes in Cu_B redox/ligation state was proposed by ourselves and others. Cu_B ²⁺ reduction is accompanied by the protonation of one of its ligands, keeping this metal electroneutral. However, gating (K-gate closed) ensures that ⁽¹⁻⁾H334 is the Cu_B ¹⁺ ligand that undergoes protonation, ultilizing a H⁺_(P) from the D-channel; the H⁺_(P) donor is the protonated water cluster, H⁺–PWC.

Following the rapid uptake of H⁺_(P) by ⁽¹⁻⁾H334, ⁽¹⁻⁾OH–Cu_B ¹⁺ is protonated by H⁺_(SK) entering the BNC from the K-channel, from H⁺–K362. As a result of the entry of H_(SK), H⁺_(P) is pushed from HH334 and deposited into the interface between the BNC and the vestibule H⁺–W1a₃, Fig. 2A.

Preceding the entry of the second electron into the BNC, a rearrangement occurs involving the coordinated transfer of both a proton and electron from Cu_B to heme a₃, generating ferrous aqua heme a₃ and regenerating Cu_B as ⁽¹⁻⁾HO–Cu²⁺ to act again as the electron acceptor for the 2nd electron. The events which follow the entry of the second electron into the BNC are essentially the same as occur following the 1st, so that proton pumping occurs in an identical manner. The only difference in the chemistry of the BNC for uptake of the first and second electrons is that in the former, heme a₃ is present as ⁽¹⁻⁾OH–Fe³⁺ state and in the latter a Fe²⁺–OH₂ state.

• iii) His-Tyr/Farnesyl-OH interaction: K-channel gating

Our K-Gate model for the regulation of proton access from the K-channel into the BNC is based on two critical observations. Firstly, the universal presence of the very strong hydrogen bond between the farnesyl hydroxide and the phenol group of Y288 precludes the entry of H⁺_(SK) into the BNC and represents the K-gate in the normal, ground state, closed position. Secondly, it has been shown by EXAF examinations of the oxidase that Cu_B may sometimes become three coordinate, bound only to two histidines and one aqueous ligand (Fann et al. 1995; Powers et al. 1994).

We postulate that reduction of Cu_B results in the release of the H284-Y288 ligand, and its movement, resulting in breaking the strong hydrogen bond between the farnesyl hydroxide and Y288, opening the K-gate. This opening allows a proton to move from H⁺-K362 into the BNC, protonating the cuprus hydroxide. Figure 2A shows the proposed positions of His-Tyr, Cu_B and Farn-OH during the K-gate opening cycle.

The interaction of the H284–Y288 pair with Farn-OH is a central feature of the metal reduction phase of our proposed model. However, Y288 and Farn-OH are both absent in the b₃ class of oxidases and may not be universally present in oxidases with unusual hemes in their BNCs. The apparent paradox is resolved by comparative sequence analysis, modeling studies and more recently by site-directed mutagenesis studies on a b₃-type oxidase. In the non-classical b₃ type oxidases, the roles we postulate for Y288 and Farn-OH are undertaken by two different and fully conserved alcohol groups. In the b₃ oxidases, Y327 (helix 7) replaces Farn-OH and Y395 (helix 9) replaces Y288. Our modeling studies indicate the functionality of the Y327/Y395 K-gate in b₃ oxidases.

• iv) The third and fourth electrons

Soon after it binds to the fully reduced BNC, molecular oxygen is reduced and split, with two electrons taken from heme a₃ (generating a ferryl), one electron taken from Cu_B, and a H⁺/e⁻ is abstracted from His-Tyr, to generate a ferryl/radical state which is generally called P_M, Fig. 2B. There are two major differences in proton pumping in the oxygen reduction phase, when compared to the metal reduction phase: H⁺_(S) are supplied from HE286 at the top of the D-channel and not from H⁺-K362 of the K-channel; and the radical form of the H284-Y288 crosslinked pair, His-•Tyr, is the initial acceptor of both electrons and H⁺_(S).

Transfer of the 3rd electron into the BNC reduces His-•Tyr, forming a protonatable anionic Cu_B ²⁺ ligand, ⁽¹⁻⁾His-Tyr. Electron entry into the BNC is charge compensated by H⁺_(S) uptake, gating ensures that ⁽¹⁻⁾H334 is protonated via the H⁺–PWC. There now follows a slower internal movement of protons: HE286, of the D-channel, passes a H⁺_(S) to ⁽¹⁻⁾His-Tyr, consequentially Cu_B ²⁺ requires an additional ionic bond and so HH334 is forced to deprotonate, ejecting a H⁺_(P) out of the BNC.

We propose that prior to the arrival for the 4th electron into the BNC there is a second regeneration reaction. The F-state decays oxidizing His-Tyr and regenerating the proton pumping acceptor, His-•Tyr (Blomberg et al. 2000a, b; Shi et al. 2000; Proshlyakov et al. 1998, 2000). Thus His-•Tyr is again the acceptor of the 4th electron and 4th substrate proton from HD286.

• v) The overall movements of charges and counter charges through oxidase.

Figure 3 provides an overview of our stepwise, sequential model of H⁺_(P) pumping. The center section shows a cutaway section of oxidase that is color coded to show some of the key amino acids, cofactors and water molecules that are important for H⁺_(P) pumping.

Fig. 3.

Is a color coded, diagrammatic representation of the electrostatically coupled movements of substrate electrons, substrate protons from the K-channel (H⁺_(SK)), substrate protons from the D-channel (H⁺_(SD)) and pumped protons (H⁺_(P)), into, through and out of the oxidase. The figure is divided into three parts: The central panel is a cutaway diagram of the R.s. oxidase and shows the route taken by protons, during both phases of the catalytic cycle, and key amino acids, metal centers and water molecules. The panel on the left shows the counter charge movements of H⁺_(P), in numbered sequence, that occur during the consumption of each of the four electrons in the overall metal/oxygen reduction phases. The panel on the right shows how discrete, electrostatic potentials are generated by the movement of substrate electrons and protons, in both the metal (MR) and oxygen (OR) reduction phases (first/second and third/fourth electrons respectively)

Electrogenic movements of substrate electrons and H⁺_(S), are shown on the right-hand-side.

The route and storage sites of H⁺_(P) are shown on the left-hand-side.

Summary

Pumped protons move through cytochrome oxidase from the inner to the outer bulk phases along the same route, in a stepwise fashion utilizing six acceptor/donor sites. Only two of the acceptor/donor sites are amino acids. The other four acceptor/donor sites consist of water molecules or clusters of water molecules (Fig. 3): a protonateable water cluster (PWC), near the top of the D-channel; W1a₃ between the heme a₃ propionates; W1_Mg associated with the Mg ion; and W1_E254II associated with Cu_A/Mg²⁺. Storage of a proton as part of a water cluster, near the top of the D-channel, has previously been suggested on the basis of computational analysis (Xu et al. 2007; Xu and Voth 2006). The model proposed for proton pumping is dependent on the coupling of electron movements to proton movements outside the BNC as well as the chemistry that occurs within it. The chemistry that occurs in the BNC is only responsible for the movement of a H⁺_(P) from the H⁺–PWC to H⁺–W1a₃, at the BNC interface, which is ≈ 30% of the total distance between H⁺_(P) input and output sites. Recycling in the BNC is a key component of this pumping model. The BNC undergoes a rearrangement following the entry of the first and third electrons and their associated substrate protons, to allow the regeneration of Cu_B to its common pumping state. Generally, in common with many proposed pumping mechanisms for cytochrome oxidase, protons move through the protein along water chains, utilizing Grotthuss-type mechanisms, but these movements are constrained by gates, which open and close at specific parts of the cycle.