Fig. 6. Possible proton transfers through hydrogen bonds. (A) A possible proton-transfer pathway through a hydrogen bond between the two groups, both of which can be either a hydrogen donor or an acceptor. (B) A possible irreversible proton transfer with migration of a hydronium ion.

Proton transfer through a long hydrogen-bond network may be promoted by an arrangement of disconnecting points in the network where protons are transferred irreversibly with the aid of confor-mational changes. For example, a water molecule fixed by a hydrogen bond at a hydroxyl group as given in Fig. 6B, on receiving a proton from the hydroxide group, moves to other OH groups to form a new hydrogen bond. Then, a proton on the water molecule could be transferred to the OH group on the right-hand side. The migration of water could be irreversible if it is controlled by a conforma-tional change inducing the polarity decrease in the environment of the hydronium ion, hydrogen-bonded to the left-hand hydroxyl group, down to the level significantly lower than that of the right-hand hydroxyl group. This irreversible conformational change provides an irreversible protontransfer step from the left-hand hydroxyl group to the right-hand hydroxyl group. If this conforma-

tional change is coupled to the oxidation state of a metal center, a redox coupled proton transfer is facilitated by this system. The migration of water can be replaced with that of ionizable amino acid side chain. Proton transfers inside the protein both for water formation and for pumping protons must include such irreversible steps. In other words, protons for making water molecules must also be actively transported (pumped).

6.2. Proposed Mechanisms of Redox Coupled Proton Transfer

The first proposal for the proton pumping element of cytochrome-c oxidase included the redox coupled change in the coordination of the CuA site (59). Unfortunately. this mechanism was not widely accepted, because the proton-pumping activity of Escherichia coli quinol oxidase, which does not have CuA site, was shown experimentally. Many researchers believe that cytochrome-c oxidase and quinol oxidase pump protons by an identical mechanism, although evidence to support this proposal is lacking.

Another proton-pumping element including the CuB site was proposed in 1994 (60). This mechanism proposes that protons are pumped directly by a redox coupled coordination change in the CuB site. The key point of this proposal is that redox coupled movement of one of the three hisidine imidazole group responsible for pumping protons. This histidine cycle mechanism was quite widely accepted because all of the amino acids included in this mechanism are completely conserved in all biological species with terminal oxidases in the heme-copper terminal oxidase superfamily. Furthermore, an X-ray structure of Parraccocus cytochrome-c oxidase in the fully oxidized azide bound state lacks one of the histidine imidazoles bound to CuB, which suggests the mobility of the imidazole group (7). Many researchers accepted that this X-ray structure supports the histidine cycle mechanism. However, X-ray structures of the bacterial enzyme in both fully oxidized and fully reduced states showed that all three histidine imidazoles are coordinated to CuB (61). Furthermore, X-ray structures of bovine heart cytochrome-c oxidase in the fully oxidized, fully reduced, fully oxidized azide-bound and fully reduced CO-bound states clearly showed that three histidine imidazoles are bound to CuB (9,54).

A drawback of this proposal is that no mechanism for sorting protons for pumping from those for making waters is given. If protons to be pumped are introduced to the O2 reduction site, they would be readily used for water formation. Then, the protons cannot be pumped to the intermembrane side. One would state that the O2 reduction and the proton pumping can occur in a separated timely manner. However, experimental results show that both the proton pumping and the proton transfer for water formation are coupled tightly with the electron transfer for the O2 reduction (62,63). Thus, the mechanism for the time sharing would not be simple. No experimental evidence supporting the time sharing has been obtained.

6.3. Redox Coupled Conformational Change in Bovine Heart Cytochrome-c Oxidase

Figure 7A shows a comparison of X-ray structures of bovine heart cytochrome-c oxidase in the fully oxidized and reduced states at 2.30 A and 2.35 A resolution, respectively (9). A fairly large change in conformation is detectable in a loop region between helices I and II on the intermembrane side. No significant conformational change is detectable in other protein moiety of this enzyme. The conformational change includes the peptide backbone. Only the conformational change in Asp51 induces the change in the accessibility to the bulk water phase on the intermembrane side. In the fully oxidized state, Asp51 is completely buried within the interior of the protein, and water molecules in the bulk water phase are not able to access the amino acid side chain. On reduction of this enzyme, Asp51 moves toward the molecular surface to be exposed to the bulk water phase in the intermembrane side. On the other hand, as shown in Fig. 7B, Asp51 in the fully oxidized state has an effective accessibility to the bulk water phase on the matrix side, because Asp51 is connected to Arg38 via a hydrogen-bond network that includes a peptide bond. The Arg38 is located on the wall of a large cavity located close to heme a plane, which holds mobile water molecule. The cavity is connected to the matrix space via a water path. Thus, Arg38 is equilibrated with the bulk water phase on the matrix side, so that Asp51 can take up protons from the matrix side. In this sense, Asp51 is accessible to the matrix side. In the reduced state, the hydrogen bond between Asp51 and the peptide NH group of Ser441 is broken and Asp51 loses the accessibility to the matrix space. Thus, the redox coupled movement of Asp51 strongly suggests that this is the site for proton pumping.

6.4. Proton Transfer Through a Peptide Bond

As shown in Fig. 7B, the hydrogen-bond network between Asp51 and Arg38 includes a peptide bond. This peptide bond is likely to facilitate a unidirectional proton transfer to Asp51. It has been well established that H/D exchange in a peptide NH group proceeds through an imidic acid intermediate (-C(OH)=N+H-) because the peptide C=O is much more basic than the peptide NH (64). The imidic acid releases the proton to give the enol form of the peptide. The enol form tautomerizes back to the keto form of the peptide, which is much more stable than the enol form. In this step, H/D exchange in NH group occurs in solution. In the enzyme, once the peptide carbonyl is protonated, the positive charge migrates to the nitrogen atom, which can then donate protons to the deprotonated carboxyl group of Asp51 hydrogen-bonded to the peptide NH, leaving the enol form of the peptide. The enol form will be readily transformed to the more stable keto form by migration of the proton on the enol OH to =N-. When the peptide is in the enol form, the reverse reaction to form the imidic acid is possible. However, once the enol form tautomerizes back to the keto form, the reverse reaction is unlikely to occur. Even if the NH group is protonated to give N+H2, the imidic acid intermediate cannot be formed from this form. The proton will then be readily taken up by the nearby COO- group. Thus, the stability of the keto form versus the enol form contributes to the unidirectionality in the proton transfer. Furthermore, a conformational change of Asp51 also strongly contributes for the unidirectinality of the proton pump. The third factor contributing to the unidirectionality could be that oxidation of heme a decreases significantly the p^ of the propionate groups. Dissociated protons from one of the propionates shown in Fig. 7B could be transferred to the peptide carbonyl to form an imidic acid. The positive charge on Arg38 may prevent the protons from approaching the large cavity in which water molecules are equilibrated with bulk solvent. The electrons taken from heme a are transferred irreversibly to the O2-reduction site. When heme a is reduced again to increase the p^ value of the dissociated propionate, the propionate will be reprotonated by water molecules in the large cavity via Arg38.

6.5. FTIR Studies on the Redox Coupled Conformational Change

The redox coupled conformational change in Asp51 is clearly detectable by comparison of the X-ray structure of the fully oxidized enzyme at 2.3 A resolution with that of fully reduced enzyme at 2.35 A resolution (9). However, the resolution of the X-ray structure is not high enough for detection of the protonation state of Asp51. For this purpose, infrared spectroscopy is very useful. An infrared positive band near 1750 cm-1 assignable to the COOH group of Asp or Glu was observed in the difference spectrum of the oxidized versus reduced forms of bovine heart cytochrome-c oxidase. This band was not observed in the redox difference spectrum of Paracoccus cytochrome-c oxidase, which does not have aspartate in the corresponding position in the amino acid sequence (65). Thus, the infrared band is likely to be the result of the redox coupled protonation change of Asp51 (COOH in the oxidized state and COO- in the reduced state).

Recently, it has been claimed that the conformational change in Asp51 is controlled only by electrostatic attraction between Asp51 COO- and CuA, which are located only 6 A apart in the oxidize state (66). However, these Fourier-transform infrared (FTIR) results do indicate that Asp51 does have a redox coupled protonation change. Furthermore, in the oxidized state, the carboxyl group is protonated so that no electrostatic interaction is possible for the -COOH group. On the other hand, in the reduced form, the carboxyl group is deprotonated, but the CuA site in the reduced form does not have any net charge. Thus, no electrostatic interaction is possible between them. However, during the

Fig. 7. Redox coupled conformational change in a loop between helices I and II of subunit I. A stereoview (A) and a schematic representation of the hydrogen-bond network connecting Asp51 with the matrix space (B). (A) The molecular surface on the intermembrane side is shown by small dots. Red and green sticks represent the structures in the fully oxidized and reduced states.

Fig. 7. Redox coupled conformational change in a loop between helices I and II of subunit I. A stereoview (A) and a schematic representation of the hydrogen-bond network connecting Asp51 with the matrix space (B). (A) The molecular surface on the intermembrane side is shown by small dots. Red and green sticks represent the structures in the fully oxidized and reduced states.

Fig. 7. (B) Dotted lines show hydrogen bonds. The rectangle represents a cavity near heme a. The two dotted lines connecting the matrix surface and the cavity represents the water path. The dark balls show the positions of fixed water molecules.

oxidation process of the reduced form of the enzyme, the oxidized form of CuA could trigger the conformational change of Asp51 COO- by an electrostatic attraction.

6.6. Proton-Pumping Mechanisms in Cytochrome-c Oxidase in Various Biological Species

The key amino acid in the proton-pumping process of bovine heart cytochrome-c oxidase is conserved only in the animal kingdom. Plant cytochrome-c oxidase, bacterial cytochrome-c oxidase, and other enzymes in the heme-copper terminal oxidase superfamily do not have Asp51. Thus, the above proton-pump mechanism including Asp51 has not been accepted widely. An interpretation for the incomplete conservation of Asp51 is as follows: Reduction of O2 without release of active oxygen species is an extremely complex and specific reaction. For this purpose, a heme-copper system is optimal. Thus, amino acids ligand to the metals must be conserved. On the other hand, proton pumping is much simpler chemically. Many amino acids, cofactors, and even water molecules can transfer protons. Thus, it is not surprising that Asp51 is not completely conserved. In fact, the Paracoccus enzyme contains a structure similar to that of the proton-pumping system in the bovine heart enzyme, although no COOH-containing amino acid is located at the position corresponding to that of Asp51 of the bovine heart enzyme. The structure suggests that the Paracoccus enzyme also pumps protons using a system that is closely similar to that of bovine heart enzyme, which does not include the

O2-reduction site. Thus, the proton transfers for pumping is completely separated from the proton transfer for water formation.

According to the recent mutagenesis works for the bacterial channel-like structure corresponding to the possible proton pumping channel of the bovine heart enzyme, mutations of amino acids located on the wall of the water channel in the bottom half of the channel has no significant effect on the enzyme activity (67,68). On the other hand, a mutation of Arg38 to Met in bovine number, which is at the starting point of the hydrogen-bond network, kills the enzyme (68,69). These mutation results are consistent with the structure of the channel composed of the water channel and the hydrogen-bond network in the bottom and upper halves, respectively, with Arg38 at the interface.

0 0

Post a comment