Mechanism of Oxygen Reduction in Cytochrome c Oxidase and the

Subscriber access provided by Warwick University Library

Article

The mechanism of oxygen reduction in Cytochrome c oxidase and the role of the active site tyrosine Margareta R.A. Blomberg Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.5b01205 • Publication Date (Web): 21 Dec 2015 Downloaded from http://pubs.acs.org on January 8, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

The mechanism of oxygen reduction in Cytochrome c oxidase and the role of the active site tyrosine. Margareta R.A. Blomberg Department of organic chemistry, Arrhenius Laboratory, Stockholm University, SE-106 91, Stockholm, Sweden.

1 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract Cytochrome c oxidase, the terminal enzyme in the respiratory chain, reduces molecular oxygen to water, and stores the released energy through electrogenic chemistry and proton pumping across the membrane. Apart from the heme-copper binuclear center, there is a conserved tyrosine residue in the active site (BNC). The tyrosine delivers both an electron and a proton during the O-O bond cleavage step, forming a tyrosyl radical. The catalytic cycle then occurs in four reduction steps, each taking up one proton for the chemistry (water formation) and one proton to be pumped. It is here suggested that in three of the reduction steps the chemical proton enters the center of the BNC, leaving the tyrosine unprotonated with radical character. The reproprotonation of the tyrosine occurs first in the final reduction step before binding the next oxygen molecule. It is also suggested that this reduction mechanism and the presence of the tyrosine are essential for the proton pumping. Density functional theory calculations on large cluster models of the active site show that, only the intermediates with the proton in the center of the BNC and with an unprotonated tyrosyl radical have a high electron affinity of similar size as the electron donor, which is essential for the ability to take up two protons per electron, and thus for the proton pumping. This type of reduction mechanism is also the only one that gives a free energy profile in accordance with experimental observations for the amount of proton pumping in the working enzyme.

2 ACS Paragon Plus Environment

Page 2 of 33

Page 3 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

1. Introduction Molecular oxygen is reduced to water in the respiratory enzyme cytochrome c oxidase (CcO). The reaction is quite exergonic, and a significant part of the free energy is conserved as an electrochemical gradient across the inner mitochondrial or bacterial membrane. The energy conservation is remarkably efficient in this enzyme, since the gradient buildup occurs not only via electrogenic chemistry, but also via proton pumping. The electron donor, cytochrome c, is located on the P-side of the membrane, and the protons needed for water formation are delivered from the opposite side of the membrane, the N-side, see Fig. 1. Thus, the chemistry corresponds to a charge translocation across the membrane, which is referred to as an electrogenic reaction. This exergonic chemistry is furthermore coupled to the transportation of protons across the entire membrane, from the N-side to the P-side, referred to as proton pumping. Experimental information shows that in the largest family of CcOs, the A-family, four protons are pumped for each reduced oxygen molecule, one per reduction step [1], giving the overall reaction: − + O2 + 8H+ N + 4eP → 2H2 O + 4HP

(1)

The oxygen chemistry takes place in the binuclear active site, the BNC, consisting of a highspin heme a3 and a mono-nuclear copper complex, CuB . The CuB ion has three histidine ligands, one with a cross-linked tyrosine residue, which has been found to be redox active. Electrons are delivered to the BNC from cytochrome c via two metal cofactors, a dicopper complex labelled CuA and a low-spin heme a. The A-family of CcOs has two proton pathways from the N-side to the BNC, see Fig. 1. All four pumped protons and at least two of the protons for the chemistry are taken up via the D-channel, and at least one of the chemical protons is taken up via the K-channel. Molecular oxygen binds to the reduced BNC, and after the O-O bond cleavage the chemistry takes place in four reduction steps, yielding two water molecules. Each reduction step consists of an electron transfer from cytochrome c to the BNC, the uptake of one proton from the N-side of the membrane via one of the proton channels to the BNC, plus translocation of one proton from the N-side of the membrane, via the D-channel, to the P-side of the membrane. Thus, each electron transfer to the BNC is coupled to the uptake of two protons from the N-side of the membrane, and also to the expulsion of one of these protons to the P-side of the membrane, which puts certain constraints on the mechanism for the reduction reaction. Furthermore, the electrochemical gradient over the membrane implies that the protons have to move against a gradient, which complicates the reaction mechanism even further. There are 3 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1: Overview of CcO. The two proton channels indicated, D and K, are present in the A-family different views on the details of the mechanism for the proton pumping. However, most of the mechanisms suggested have some common features [2]. Thus, the general view is that the electron transfer from cytochrome c to heme a (or all the way to the BNC itself) electrostatically triggers the uptake of the proton to be pumped into a pump loading site (PLS) in the vicinity of the BNC, possibly near the propionates on heme a3 . Next the electron will move into the BNC (if it is not already there) which in the following step triggers the uptake of the proton to the BNC for the chemistry. At this point the two newly up-taken protons will repel each other, and the proton in the pump loading site will move out towards the P-side of the membrane. One important feature in this description is that the electron needs to reach the BNC before the proton for the chemistry, and it will be shown below that this puts certain requirements on the properties of the active site and actually determines the reaction mechanism for the oxygen reduction. Obviously there are many details missing in the description of the proton pumping mechanism given, mainly how kinetic barriers govern the protons to move in the right direction at different stages, but these are not the subject for the present investigation.

4 ACS Paragon Plus Environment

Page 4 of 33

Page 5 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

The catalytic cycle of CcO is normally described in the following way. In the resting state, called the oxidized state and labelled O, the oxidation state of the BNC is Fe(III)Cu(II). Before the reaction with molecular oxygen can take place, the metal ions are reduced by two electrons giving state R with Fe(II) and Cu(I). The first of these two electrons is considered to go to copper, giving the so called one-electron reduced intermediate, labelled E, and iron is then reduced in the second step. Oxygen binds reversibly to the ferrous heme a3 , to yield the observed intermediate labelled A. Cleavage of the O-O bond leads to the most oxidized intermediate, PM , with a tyrosyl radical, Fe(IV)=O and Cu(II)-OH. The next two reduction steps are considered to reduce the tyrosyl radical, giving intermediate F and the ferryl, returning to intermediate O. The location of the chemical proton taken up in each reduction step has not been determined experimentally. The entire reduction reaction is exergonic by 51 kcal/mol (2.2 V) with cytochrome c as the electron donor [3]. From a thermodynamic point of view this should be enough energy to allow pumping of four protons in total per oxygen molecule also with the presence of the maximum gradient (200 mV [3]). However, since the four reduction steps involve different reduction processes (TyrO• → TyrO− , Fe(IV) → Fe(III), Fe(III) → Fe(II) and Cu(II) → Cu(I)), corresponding

to different reduction potentials, the free energy is not evenly distributed over the catalytic

cycle. In fact, experimental information indicates that the equilibrium reduction potentials for Cu(II) and Fe(III) are only slightly larger than that of the cytochrome c donor, in the range 0.28-0.35 V [4, 5, 6, 7], as compared to 0.25 V. This means that the two reduction steps from O to E and from E to R should be only slightly exergonic already without the gradient. Both electrogenic chemistry and proton pumping against a gradient makes the reaction more endergonic and thereby increases the rate limiting barriers. At the same time, recent experimental observations show that the A-family CcOs actually pump four protons, one per reduction step, at a reasonable rate, also with a substantial gradient present [8, 9], which thus seems to be in conflict with the low reduction potentials for two of the cofactors in the active site. In a recent computational study the energetics of the catalytic cycle of the oxygen reduction in CcO was investigated [10]. The main purpose was to find an explanation to the above mentioned conflict between the high level of proton pumping at a significant gradient [8, 9] and the low reduction potentials for two of the BNC cofactors (CuB (II) and Fea3 (III)) [4, 5, 6, 7]. The two most important results from that computational study are the following. First, the intrinsic reduction potential of CuB (II) is actually much higher than expected, 5 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 33

His284 Tyr288

His333 CuB

His334 PropA

2.59 Å

1.64 Å 1.72 Å

1.87 Å 2.33 Å

Asp407 2.30 Å

PropD

Heme a3

His419

Arg481

Figure 2: The main model (PropH) used in the present calculations, showing the optimized EH state. The light red circles indicate atoms that are fixed during geometry optimizations. which partly explains the high level of proton pumping. Second, by avoiding protonation of the redox active tyrosine until the very last reduction step before oxygen binding, an activated intermediate labled EH is formed, which leads to a more even distribution of the exergonicity over the reduction steps, and thereby also contributes to the high level of proton pumping [10]. The results just mentioned involve a more important role for the active site tyrosine in the catalytic reaction mechanism than previously realized. The main purpose of the present investigation is to further elaborate on the role of the tyrosine for the oxygen reduction mechanism and for the proton pumping. It is shown that the presence of the tyrosine in the active site is essential, and enables the proton pumping. New calculations are performed using larger models of the active site as compared to the previous study. A limitation in the previous BNC-model was that the heme a3 propionates were omitted, and therefore new models are constructed including the propionates and the most important amino acids and water molecules in their vicinity. The main results reported here are obtained with one of those larger models. As will be described below, the new larger models support the picture obtained with the smaller model, but adds also significant new understanding.

6 ACS Paragon Plus Environment

Page 7 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

2. Computational procedure Quantum chemical calculations have been performed on models of the BNC in the Afamily CcO to obtain a detailed description of the catalytic cycle. In the first subsection below the models of the BNC used in the calculations are described, and in the second one the quantum chemical methods are described. In the final subsection the procedure used to obtain the overall energetics of the catalytic cycle is described. More details can be found in ref. [10], where the same computational procedure, including one of the models used in the present investigation, was carefully discussed. 2.1 Model of the BNC The binuclear active site (BNC) in the A-family CcOs consists of a high-spin heme a3 group and a histidine ligated copper complex (CuB ). The models used in the calculations are based on a high resolution crystal structure from Rhodobacter sphaeroides [11], see Fig. 2. The difference between the models used here concerns the description of heme a3 and its surrounding. One model, the smallest one, is the same as the one used in ref. [10]. This model is labelled Noprop, and the heme a3 part of the BNC is modelled by a heme a, keeping the substituents except the 15-carbon farnesyl chain and the two propionate groups, which are replaced by hydrogen atoms. The proximal histidine (His419) is included. The CuB model includes the three histidine ligands (His333, His334 and His284) together with the cross-linked tyrosine (Tyr288). Two larger models are used, which differ from the smaller model by the inclusion of both propionate groups (PropA and PropD) of heme a3 , the protonated arginine (Arg481) that forms hydrogen bonds to PropD, the protonated aspartic acid (Asp407) that forms a hydrogen bond to PropA, plus two water molecules hydrogen bonding to the propionates, see Fig. 2. One of these models has PropA protonated, and it is labelled PropH. The other one has PropA unprotonated, and it is labelled Prop. Since the pump loading site (PLS) is generally considered to be located near the heme a3 propionates, the PropH model may be considered to represent the sitation with the PLS loaded with a proton, and the Prop model would then correspond to an unloaded PLS. The amino acids are truncated at the alpha-carbon, which is fixed to the X-ray coordinates during geometry optimizations to maintain some of the constraints from the surrounding protein. The peptide bonds are replaced by C-H bonds, with the hydrogen atoms fixed. One carbon atom on the porphyrin is also fixed. No water molecules in the

7 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

center of the BNC, apart from the ones formed during the reduction process, are included in the model (see discussion in ref. [10]). The total charge of the models Noprop and PropH is plus one (+1) and the total charge of the model Prop is zero (0). The model Noprop has about 140 atoms (depending on the state), and the models PropH and Prop have about 200 atoms. The main results discussed in the results section are obtained with the PropH model. Using these models the different intermediates in the catalytic cycle are constructed by adding molecular oxygen, electrons and protons in separate steps. Thus, each reduction step is described by two intermediates, one with only an electron added, and one with both an electron and a proton. For the location of the added proton, there are essentially two possibilities, the proton can be placed either on one of the oxygens in the center of the BNC or on the tyrosyl oxygen which was deprotonated in the O-O bond cleavage step. The difference in the properties of these structures turns out to be important, and they are discussed in the results section. Regarding the spin-states of the intermediates, the main issue concerns the spin-state of the heme iron. It has been shown that the DFT functionals used here (see below) stabilize lower spin states for iron too much relative to the higher spin states for heme-models [12, 13, 14, 15]. Therefore, some corrections have to be introduced here. Heme a3 is generally considered to have a high-spin iron in cytochrome oxidase, and therefore the energy of the high-spin iron states are reported for most intermediates. Exceptions are the Fe(IV)=O complexes and the ferric superoxo complex, where the states reported have low-spin iron. Furthermore, in the calculations the high-spin ferrous state is too easily oxidized, resulting in a too large binding energy of molecular oxygen, which corresponds to formation of a superoxo ferric complex. Therefore a correction is introduced, lowering the high-spin ferrous state to give agreement with the experimental binding energy of molecular oxygen in the A intermediate [16]. This correction, corresponding to an increase in the Fea3 (III) reduction potential, becomes somewhat different between the models, varying between 5.0 and 8.2 kcal/mol, see further discussion in the results section. For the structures with an oxo- or a hydroxo-bridge between Fe(III) and Cu(II), there is a magnetic coupling between the two metal ions. Then the calculated DFT-energy of the antiferromagentically coupled state, which is the lowest one, has to be corrected using the Heisenberg spin-Hamiltonian formalism [17, 18].

8 ACS Paragon Plus Environment

Page 8 of 33

Page 9 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

2.2 Computational methods Quantum mechanical calculations were performed on the BNC model employing the dispersion corrected hybrid density functionals B3LYP-D3 [19, 20] and B3LYP∗ -D3 [21], the latter using 15% Hartree-Fock exchange instead of 20% as used in the original functional. All structures were fully optimized, except for some atoms fixed from the crystal structure as mentioned above, using the B3LYP-D3 functional and the basis set labelled lacvp* in the Jaguar program [22], which is a double zeta basis with polarization functions on all second row atoms. Single point calculations were performed in the optimized structures using the large cc-pvtz(-f) basis set plus lacv3p+ for the metal ions with the B3LYP∗ -D3 functional. Solvent effects from the surrounding protein were included using the self consistent reaction field (SCRF) approach as implemented in Jaguar with a dielectric constant of 4.0, in accordance with previous experience [23]. The dielectric calculations were performed for the optimized structures using the B3LYP-D3 functional and the lacvp* basis set. The calculations described so far were performed with the Jaguar 7.6 [22] program. In all optimized structures using the Noprop model the Hessian matrix was calculated using the Gaussian 09 package [24] at the same level of calculation as the geometry optimizations (B3LYP-D3/lacvp*). The Hessians were used to obtain zero-point corrections for the Noprop model, and the same zero-point corrections were used for the two larger models (PropH and Prop). Due to the fixed coordinates in the geometry optimizations, entropy effects can not be taken from the Hessian calculations. The entropy changes within the BNC itself are assumed to be small and therefore neglected. For the gaseous O2 molecule it is assumed that the entropy lost on binding is equal to the translational entropy for the free molecule (10.8 kcal/mol at room temperature). For the binding of a water molecule to bulk water a standard value of 12 kcal/mol is used, which includes explicit zero point effects for the water molecules. The relative energies reported are referred to as free energy values, with entropy effects estimated as described in the previous paragraph. The enthalpy values are obtained from the large basis set calculations using the B3LYP∗ -D3 functional, including zero point and solvent effects. The B3LYP and B3LYP∗ functionals give qualitately similar results, but the energy profiles obtained with the B3LYP∗ functional agree somewhat better with experiments. In the mechanistic discussions in sections 3.1 and 3.2 the results obtained for the PropH model are used, while the results obtained with the different models are compared in section 3.3. 9 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2.3 Overall energetics and barriers To obtain an energy diagram for the full catalytic cycle of CcO the energetics of the reduction steps has to be calculated, which means that the cost for the uptake of electrons and protons from the donors has to be estimated. Since calculated absolute reduction potentials and pKa values for the donors would not be accurate enough, a procedure is used where the overall energy of one catalytic cycle is adjusted to the value obtained from experimental reduction potentials [25, 26, 27, 28, 29, 30]. Using experimental values of 0.25 V for the electron donor cytochrome c, and 0.8 for the electron acceptor, O2 forming water, the exergonicity of one catalytic cycle of CcO without gradient becomes 51.0 kcal/mol (2.2 V) [3]. The combined cost of one electron taken from cytochrome c and one proton from the bulk is set 381.6 kcal/mol, which together with the calculated free energy for the overall chemistry occurring in equation (1), yields a total exergonicity of 51 kcal/mol, in agreement with the experimental value. This value, which corresponds to an X-H bond strength of 67.8 kcal/mol, is determined by the computational level, and it is therefore the same for all models used in the present investigation. Together with the calculated energies of the intermediates, this determines the relative energies for each of the reduction steps in the catalytic cycle. To obtain the individual costs for the electron and the proton a second experimental result for the reduction process, apart from the overall energy, is needed. The rate of electron transfer to the BNC in the PM state is known to be slightly dependent on the presence of a positively charged lysine in the K-channel [31], in a way that indicates that this electron transfer step is close to thermoneutral. Therefore the cost of the electron is chosen to make the PM to PR step only slightly exergonic (about 1 kcal/mol). This parameter, which depends on the model used, is thus a much “softer” parameter than the value used for the combined cost of the electron and the proton, but it is accurate enough for the conclusions to be drawn. The actual values used for the different models will be reported and discussed in the results section. No barriers are calculated in the present investigation. It is expected that the transfer of the pumped protons is rate limiting for each reduction step. The reaction rates for the different reduction steps are known to be in the 100 to 1000 µs range, giving barriers of approximately 13 kcal/mol. The energy profiles presented below concern the situation without the electrochemical gradient, which means that there is no cost associated with proton pumping. Therefore the pumped protons are not included in the profiles, and the barriers concern electron and proton uptake to the BNC. Since these barriers are not rate 10 ACS Paragon Plus Environment

Page 10 of 33

Page 11 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

F

PM

PR

Cu(II) TyrO OH O Fe(IV) His

+e-

EA = 102.0 kcal/mol

Cu(II) TyrOOH O Fe(IV) His

Cu(I)

TyrO

H 2O

+HD+

+HK+

O Fe(IV) His

Cu(II) TyrOH OH O Fe(IV) His

F'

+eEA = 102.6 kcal/mol

+eEA = 82.6 kcal/mol

Figure 3: Reduction of the PM state yielding two different forms of the F state. The intermediates are described in more detail in Fig. S7-S9 and F13. The electron affinities given in the figure are obtained with the PropH model. limiting they are only sketched in the diagrams.

3. Results and discussion The main results from the present study concern the catalytic reaction mechanism for oxygen reduction in CcO, with particular emphasis on the role of the active site tyrosine for catalysis and for proton pumping. The results and conclusions concerning the catalytic mechanism are presented in the first subsection below. The importance of the tyrosine for the proton pumping mechanism is discussed in the second subsection below. In the third subsection comparisons are made between results obtained with different models. As will be shown in that section, most of the results obtained are quite similar for the different models used. The numeric results mentioned in the first two subsections refer to the model labelled PropH (with a protonated PropA) and shown in Fig. 2. 3.1 The role of the active site tyrosine for the catalytic reaction To elucidate the role of the active site tyrosine in the oxygen reduction reaction in CcO the electronic and geometric structures of the different intermediates need to be determined, together with the energetics of the transitions between them. Each reduction step involves the transfer of one electron and one proton to the active site, and the energetics 11 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of these steps can be estimated based on the calculations. The calculations correspond to the situation without gradient, and then there is no overall energy cost connected with the proton pumping. Furthermore, since the pump protons are not expected to enter the actual BNC active site they are not explicitly treated in the description of the basic mechanism of the reduction process. Since most of the intermediates actually involved during catalytic turnover are quite short-lived, and therefore difficult to study experimentally, quantum chemical calculations can play an important role in providing new information about their properties. The recent computational study mentioned above [10], gave a picture of the reduction reaction occurring during enzyme turnover quite different from the conventional picture, which is based on experimental information obtained under a variety of conditions, and which is described in the introduction above. Interestingly, the calculations revealed some common features of several reduction steps, and the present results show that the similarities between the reduction steps are even more pronounced than realized in the previous study. As also mentioned in the introduction, the new mechanistic picture attributes a larger role for the active site tyrosine than previously anticipated. A natural starting point for the description of the reaction mechanism is the PM intermediate, the immediate product after cleavage of the O-O bond. For this intermediate there is essentially no choice for the electronic structure, it is the most oxidized state, with Fea3 (IV)=O, CuB (II)-OH and a tyrosyl radical, see Fig. 3. The calculated electron affinity for this state is 102.0 kcal/mol (using the PropH model), and when the electron enters the BNC it goes to the tyrosyl radical, forming tyrosinate in the PR state. For the subsequent proton transfer to the active site, leading to formation of the F intermediate, there are two different reaction pathways available, depending on which proton channel is used (compare Fig. 1). Either the proton can be taken up via the K-channel, forming a state with a neutral protonated tyrosine, here labled F’. Or the proton can be taken up via the D-channel, forming a water molecule on CuB . With the proton in the center of the BNC, i.e. with a water molecule on CuB , an electron moves from tyrosinate to copper, resulting in CuB (I)-OH2 and a tyrosyl radical. The most important result at this point is that the two forms of the F state have very different affinities for taking up the next electron, see Fig. 3. The F state with a tyrosyl radical and a proton in the center of the BNC has a calculated electron affinity of 102.6 kcal/mol, very close to that of the PM state, while the state with a neutral protonated tyrosine (F’) has a much lower electron affinity of only 82.6 kcal/mol. Therefore, if the proton goes to the center of the BNC rather than to the tyrosine

12 ACS Paragon Plus Environment

Page 12 of 33

Page 13 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

when forming the F state, the electron affinity stays at a high value, which secures fast electron transfer also in the subsequent reduction step. As it turns out, all reduction steps can be described in a similar way: The initial state has a tyrosyl radical, the reduction step is initiated by electron transfer to the radical yielding tyrosinate, a proton enters the center of the BNC, which triggers a local electron transfer from the tyrosinate to one of the metal atoms, reforming the tyrosyl radical. This result indicates, that with the redox active tyrosine in the active site, it is possible to tune the electron affinities for the different intermediates such that they are all reasonably similar and high, in spite of the fact that the cofactors in the active site have rather different reduction potentials. The high electron affinity of the active site intermediates is not only important for fast electron transfer in each reduction step, but it also secures that the electron comes to the active site before the chemical proton, which is essential for the proton pumping mechanism as will be discussed in the next section. The general mechanism for the reduction steps described above is summarized in the scheme in Fig. 4 for the entire catalytic cycle. As discussed above, when both the electron and the proton has been transferred to the PM state, the tyrosyl radical, originally present in the PM state, is reformed. This means that it is CuB that is reduced in the PM → F step,

the first reduction step after O-O bond cleavage. The next reduction step again starts with a simple electron transfer to the tyrosyl radical forming tyrosinate. Only when the proton enters the BNC to form a hydroxyl group on the heme iron (Fea3 ), the reduction from Fe(IV) to Fe(III) occurs, in concert with the reformation of the tyrosyl radical. This reduction step is thus slightly different from the description given in the previous computational study [10], where a more complicated mechanism was given. The present FR state (see Fig. S10) is found to be lower in energy than the one suggested in [10] (for all models). The product of this reduction step is labelled OH , where the H index is used because it has been suggested, based on experimental results, that the oxidized state involved during catalytic turnover, labelled OH , is different from the experimentally observed resting oxidized state, with the general label O. Thus, the present OH state has CuB (I)-OH2, Fea3 (III)-OH and a tyrosyl radical. The alternative state with the proton on tyrosine rather than in the center of the BNC is here labled OB , using the same notation as in ref. [32], and referring to the bridging nature of the central oxo ligand. Importantly, the calculations show that for both the F state and the OH state it is energetically more favourable to take up the proton to the center of the BNC rather than forming a neutral protonated tyrosine, by

13 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

OH

Cu(I) H 2O

TyrO

+e

OH

Page 14 of 33

Cu(I)

-

H 2O

OH

Fe(III)

EH

H 2O

His TyrO

+H

+

D

+e-

OH2

Cu(I)

ER

OH2 Fe(II)

His

R

TyrO

H 2O

Fe(II)

Cu(I)

OR

Fe(III)

His

Cu(I)

TyrO

+H+

His

K Cu(I)

TyrOH

+O2

H 2O OH2

-2H2O

Fe(II)

TyrOH

A

O2Fe(III) His

His O-O bond cleavage

PM

Cu(II) TyrO OH O Fe(IV) His Cu(I)

F

Cu(II) TyrOOH O Fe(IV)

+e-

His

+H+D TyrO

H 2O O Fe(IV)

Cu(I)

+e-

H 2O

His

OH

TyrO-

O Fe(IV)

Cu(I) TyrO H 2O OH

+H+ +e-

PR

FR

His

D

Cu(I) H 2O

OH

Fe(III)

Fe(III)

His

His

TyrO-

OR

Figure 4: Reaction mechanism during enzyme turnover showing that in every reduction step the initial electron transfer to the BNC results in reduction of the tyrosyl radical into tyrosinate. The subindex on the protons, D or K, indicates which channel is suggested to be used for the proton uptake. Geometric and electronic (spin-populations) structures for all intermediates are shown in Figures S1-S10. 7.6 and 5.2 kcal/mol, respectively. The energetic position of the states with a protonated tyrosine (F’ and OB ) are shown in the energy profile discussed below. The reason for this ordering of states is that the proton coupled reduction potential for CuB (II)-OH is larger than that of the tyrosyl radical in the context of the BNC in CcO. Assuming that the proton transfer barrier in the D-channel is not prohibitively high, the enzyme will follow the low energy path to form the desired intermediates with a high electron affinity during 14 ACS Paragon Plus Environment

Page 15 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

G kcal/mol +e0

-10

+H+D

+H+K

-1.5 OR

OH

Cu(I) TyrO H 2O OH Fe(III)

-20

+H+K

+eEH -11.3

E

Cu(I) TyrOH H 2O OH Fe(III)

-6.4

Cu(I) TyrO H 2O H O 2

Fe(II)

TSO-O -5.1 ER

-2H2O+O2

Cu(II) TyrOH H 2O

-14.7 R A Cu(I) TyrOH Cu(I) TyrOH H 2O H 2O O

-13.7

2

Fe(II)

Fe(III)

+e-19.9 P M

PR

Cu(II) TyrOH HO O

O Fe(IV)

Fe(IV) F'

-32.4

+e-

-40.0

-40

FR

Cu(I) TyrO H 2O

+H+D -41.7

F

-50

O Fe(III) OB

-20.9

Cu(II) TyrO HO

-30

+H+D

O Fe(IV)

Cu(I) TyrO H 2O

OH

-45.8 OB -51.0 OH

Fe(III)

Figure 5: Energy profile for one catalytic cycle obtained with the PropH model. The rate-limiting barriers for transfer of the pump-protons are not included. The barriers for electron transfer and for uptake of chemical protons are indicated in an approximate way. Geometric and electronic (spin-populations) structures for the intermediates are shown in Figures S1-S10. these two reduction steps, referred to as the oxidized part of the cycle. Also in the reductive part of the cycle, the first reduction step, from OH to a state with the general notation E, is quite similar to the previous reduction steps. The electron goes initially to the tyrosyl radical, forming tyrosinate in the OR state, and the proton has two options, forming a neutral protonated tyrosine in a state with low electron affinity, or protonating the center of the BNC yielding a state with high electron affinity. However, in contrast to the situation in the oxidized part of the cycle, the state with a protonated tyrosine and a low electron affinity is now lower in energy than the desired state with a high electron affinity. Therefore, it is suggested that at this point there is a kinetic effect, preventing the proton to reach the state with lowest energy, which means that there should be a high barrier in K-channel and a low barrier in the D-channel, and also a high barrier for proton transfer within the BNC itself at this stage. During enzyme turnover, a state labled EH with the proton in the center of the BNC and with an unprotonated tyrosine 15 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

is then involved. This state has a high electron affinity and it makes this reduction step resemble the two previous ones. The alternative state with the proton on tyrosine rather than in the center of the BNC, here labled E, corresponds to proton coupled reduction of the tyrosyl radical, while formation of the EH state mainly corresponds to reduction of Fea3 (III)-OH to Fea3 (II)-OH2 . The tyrosyl radical has a significantly larger reduction potential than Fea3 (III)-OH, which is one of the problematically low reduction potentials mentioned above, and this is the reason for the low energy of the E state as compared to the EH state. Importantly though, the EH state is not a pure tyrosyl radical state, but the main TyrO• –Fe(II) electronic structure is somewhat mixed with the TyrO− –Fe(III) structure, as elucidated by the spin-populations [10] (see Fig. S3). Clearly, in the final reduction step, yielding the R state, the tyrosyl is fully reduced and protonated, to make the BNC ready for binding and cleavage of the next oxygen molecule. Thus, by protonating the center of the BNC and leaving the tyrosine unprotonated in the EH state, the relatively high reduction energy of the tyrosyl radical is shared with the low reduction energy of Fea3 (III), such that both reduction steps in the reductive part of the cycle become exergonic enough to allow both electrogenic chemistry and proton pumping also at a high gradient. It should be noted that the main aspect of the mechanism presented is that the protons for the chemistry enter the center of the BNC, thereby leaving the tyrosine unprotonated. The present positively charged models (PropH and Noprop) give very clear-cut tyrosyl radicals for the intermediates OH , PM and F (with unpaired spin-populations on the tyrosyl close to 0.9, see Fig. S1, S7 and S9), but it is possible that also for these states (as well as for the EH intermediate) there may be some delocalization of the electron from the metal to the tyrosyl. This is, however, not expected to change the energetics, since a high electron affinity is obtained not only for the tyrosyl radical but also for the metal cofactors that are reduced in these reduction steps when there is an extra proton in the center of the BNC. The model with an uncompensated charge on PropA (Prop) actually shows some electron delocalization for these states, giving unpaired spin-populations on the tyrosyl in the range 0.7 to 0.8. The observation of tyrosyl radical character in calculations on different intermediates in the catalytic cycle of CcO is not new, see for example ref. [33], where, however, completely different and quite unrealistic conclusions were drawn. On the other hand, experimentally it has not been possible to observe a tyrosyl radical in the active site of CcO, not even for the PM state, for which it was suggested to be present already in 1998 [34]. A possible explanation for the lack of experimental evidence of a tyrosyl radical

16 ACS Paragon Plus Environment

Page 16 of 33

Page 17 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

in any of the intermediates in the oxidative part of the cytalytic cycle was presented in a recent computational study [32], where it was shown that all these intermediates (PM , F and OH ) have a rather high affinity for uptake of an extra proton. It was suggested that, if the flow of electrons is stopped, these states will take up an extra proton to the tyrosine, thus eliminating the radical character. This proton uptake is slow enough not to interfere during the catalytic cycle. See also discussion below regarding the low reduction potential for CuB observed in experiments on the resting state. An energy profile for the entire catalytic cycle, as it should appear during catalytic turnover, is shown as the full black line in Fig. 5. The most important result is that the reduction energies in the two steps in the reductive part of the cycle are about equally exergonic, by 6.4 and 7.3 kcal/mol, respectively. As was shown in the previous study [10] this type of free energy profile is compatible with proton pumping in all four reduction steps also in the presence of a significant gradient, in agreement with experimental observations [8, 9]. The dotted gray line in Fig. 5 indicates a reaction path suggested to be avoided during enzyme turnover for kinetic reasons, involving the state labelled E. The formation of the E state from the OH state corresponds to reduction of the tyrosine residue, exergonic by 11.3 kcal/mol, and a transition from the E state to the R state would then become a pure Fe(III) to Fe(II) reduction, exergonic by only 2.4 kcal/mol. This demonstrates clearly how the introduction of the EH state leads to energy sharing between the two reduction steps in the reductive part of the cycle. An interesting aspect of the energy profile in Fig. 5 is that the main exergonicity occurs in the proton uptake part in all reduction steps. Clearly, this is an effect of the chosen value for the cost of the electron, as described in section 2.3. But this choice is based on experimental observations of one of the reduction steps, the A to PR step, indicating that the electron transfer is close to thermoneutral, and should therefore agree with experiments. Together with the mechanism described above, where every electron uptake step actually occurs to the same acceptor (more or less), namely to the tyrosyl radical, this experimental observation leads to close to thermoneutral electron uptake in all reduction steps. However, it should be noted that the proton uptake steps involve an internal electron transfer in most cases, from the tyrosinate to one of the metal atoms, compare Fig 4. The proton uptake steps in the energy profile are therefore not pure proton transfer steps, and the variation in exergonicity in the proton transfer part of the reduction steps corresponds to a variation in the proton coupled (internal) electron transfer. Furthermore, the total exergonicity of each

17 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 6: Sketch of one reduction step (F → OH ) including the pump-proton. In a final step, not shown, the pump-proton in the PLS (pump loading site) is expelled to the P-side. reduction step corresponds to the proton coupled reduction potential for the ultimately reduced cofactor in this step, as discussed in the previous computational study [10]. 3.2 The role of the active site tyrosine for the proton pumping An important aspect of the reaction mechanism described above is that in each reduction step the electron enters the active site before the chemical proton, see Fig. 4 and 5. This is different from, for example, the mechanism for NO reduction in cNOR, which is an enzyme belonging to the same heme-copper oxidase family as CcO. cNOR is very similar to CcO, but CuB in the BNC is replaced by a non-heme iron (FeB ), and there is no tyrosine residue in the active site. The reaction mechanism for NO reduction in cNOR was determined using the same approach as in the present study [30], and in the suggested mechanism the proton enters the BNC before the electron in each reduction step (see Fig. 3 in [30]). The cNOR enzyme does not pump protons, which means that in each reduction step, the electron uptake is coupled to the uptake of only one proton. In cytochrome oxidase, on the other hand, each electron transfer to the BNC must be coupled to the uptake of two protons, one for the chemistry and one for the proton pumping. The generally accepted view [2] on how the proton and electron uptake events during one reduction step can be organized is sketched in Fig. 6. The main point here, as shown in the first panel in Fig. 6, is that the first proton uptake occurs to the pump loading site (PLS), probably located in the vicinity of the heme a3 propionates. As indicated in the figure, this proton uptake is 18 ACS Paragon Plus Environment

Page 18 of 33

Page 19 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

coupled to electron transfer from cytochrome c, via the heme a cofactor, to the BNC. The details of this coupled electron and proton transfer are not known, and different schemes have been suggested. One view is that the electron transfer to heme a triggers the uptake of the pump proton to the PLS, and only after that is the electron transferred to the BNC [35, 36, 37, 38]. Another suggestion is that the electron goes all the way to the BNC before, or maybe concerted with the proton uptake to the PLS [39]. Both these scenarios require that the electron affinity in the BNC active site is large enough for electron transfer into the BNC before the chemical proton has arrived in the BNC. In Fig. 6 the calculated electron affinities for the two different forms of the F state are given. Since the results are obtained with the PropH model, they may be considered to correspond to the situation when the pump loading site (PLS) is loaded. The electron affinities are, 102.6 kcal/mol for the state where the proton in the previous reduction step entered the center of the BNC and the tyrosine is unprotonated with a radical character, and 82.6 kcal/mol for the state where the proton instead was delivered to the tyrosine (F’). These values can be compared to the estimated value of 101 kcal/mol for heme a, using the parametrization (for the PropH model) described in the computational details section. It can be concluded that only for the structure with a tyrosyl radical, i.e. with a proton in the center of the BNC, a fast electron transfer to the BNC can occur, as indicated in the Fig. 6. In the next step the proton for the chemistry is taken up to the BNC, as shown in the second panel in the figure. Finally, in a last step, not shown in the figure, the pump loading site proton is expelled to the outside of the membrane, due to the repulsion from the proton in the BNC. The different suggestions for the proton pumping mechanism also have different suggestions for the gating mechanisms, which, however, are not discussed here. In the present context, the most important conclusion from the scheme in Fig. 6 is that only the structure with a proton in the center of the BNC and a tyrosyl radical (first panel) has a high enough electron affinity for fast electron transfer from the heme a cofactor to the BNC. This shows that the presence of the tyrosine in the active site is essential for the proton pumping. 3.3 Comparisons between the models As mentioned above, interesting new conclusions about the reaction mechanism of oxygen reduction in cytochrome oxidase could be drawn on the basis of a previous computational study [10]. A limitation in the model used in that study (here labelled Noprop) is that the propionate groups (PropA and PropD) of heme a3 were not included. Since the 19 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Table 1: Calculated reaction energies in kcal/mol for oxygen bindinga (R → A) and the O-O bond cleavage (A → PM ) steps. For a description of the models Noprop, PropH and Prop, see the beginning of the present section. Ra → A corr. A → PM Fe(II) → Fe(III) Fe(II) Fe(III),Cu(I),Tyr → Fe(IV),Cu(II),Tyr• Noprop(+1) -1.0 -8.2 -5.3 PropH(+1) -1.0 -5.5 -5.2 Prop(0) -1.0 -5.0 -4.8 Model

a. The R → A transition is corrected to the experimental value of -1.0 kcal/mol, corresponding to a lowering of the Fe(II) state, see the column labled “corr.Fe(II)”. Table 2: Calculateda transition energies in kcal/mol for steps in the reductive part of the catalytic cycle that are suggested to be avoided during the catalytic turnoverb OH → E Tyr• → Tyr Noprop(+1) -12.1 PropH(+1) -11.3 Prop(0) -10.1 Model

E → Ra Sum Fe(III) → Fe(II) OH → R -2.8 -14.9 -2.4 -13.7 -3.4 -13.5

a. The R → A transition is corrected to the experimental value of -1.0 kcal/mol, corresponding to a lowering of the Fe(II) state, see Table 1. b. The E state has a neutral protonated tyrosine, see Fig. S12. propionates and their immediate surrounding residues carry several charged groups, they might be important for the reduction processes in the BNC. Therefore new calculations were performed including both the propionates and the surrounding charge compensating residues in the models. As described more in detail in section 2.1 two such models were used, one labelled PropH with a protonated PropA and with the same total charge (+1) as the old model Noprop, and one labelled Prop with an unprotonated PropA and a total charge of 0. It is generally suggested that the proton loading site (PLS) is located near the heme a3 propionates, and therefore the model PropH can be considered to correspond to the situation with the PLS loaded with a pump-proton. As mentioned above, the larger models support the picture obtained with the smaller model. It is still of interest to compare the results obtained with the different models more in detail, which is done in this section. As was mentioned in section 2 certain corrections have to be introduced due to the well known inaccuracy in calculated spin-state splittings for heme-iron using density functional

20 ACS Paragon Plus Environment

Page 20 of 33

Page 21 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Table 3: Calculated transition energies in kcal/mol for the two reduction steps in the oxidative part of the catalytic cycle PM → F Cu(II) → Cu(I) Noprop(+1) -18.7 PropH(+1) -20.1 Prop(0) -20.8 Model

F → OH Sum Fe(IV) → Fe(III) A → OH -11.1 -35.1 -11.0 -36.3 -10.9 -36.5

theory. The main correction concerns the Fe(II)-state, which is the initial state of heme a3 when molecular oxygen binds to the R state forming the A state. From experiments it is known that molecular oxygen binds at most very weakly in the A state. All models used here give a too large binding energy of molecular oxygen, also when the cost of moving the two newly formed water-molecules from the BNC in the R state into bulk water is included. Therefore the Fe(II) state is lowered in energy, such that the R → A step becomes exergonic

by only one kcal/mol. This correction is about 5 kcal/mol for the two larger models, and about 8 kcal/mol for the smaller one, see Table 1. The binding of molecular oxygen is associated with oxidation of the heme a3 iron from ferrous to ferric, forming a superoxide complex. Therefore the correction can also be considered as a correction of the heme a3 Fe(III) reduction potential. The experimentally determined Fe(III) reduction potential in cytochrome oxidase, 0.29-0.35 V, corresponds to an exergonicity of 1-3 kcal/mol for a reaction step reducing Fe(III) to Fe(II) with an electron from cytochrome c. In the present investigation the E → R step corresponds to such a pure Fe(III) reduction, with calculated reaction energies of 2.4-3.4 kcal/mol using the corrections just mentioned, see

Table 2, thus in good agreement with the experimental range. The correction is applied not only to the R state, but also to the ER state, which is also considered to be a pure Fe(II) state. As discussed in section 3.1 above, the EH state is a mixture of TyrO• –Fe(II) and TyrO− –Fe(III) electronic structures, and therefore the Fe(II) correction for this state is applied in proportion to the amount of Fe(II), as measured by the calculated unpaired spin-populations on the tyrosyl. This is slightly different from the procedure used in ref. [10], where a full Fe(II) correction was applied to the EH state. The main conclusions at this point are that the corrections introduced to reproduce the experimental binding energy of molecular oxygen are reasonably similar for all three models (Table 1), and that the Fe(III) reduction potentials become very similar for the three models, and similar to experiments, see Table 2. 21 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 33

Table 4: Calculateda transition energies in kcal/mol for the two reduction steps suggested to take part in the reductive part of the catalytic cycle during catalytic turnover OH → EH Fe(III) → Fe(II) (Tyr• → Tyr− ) Noprop(+1) -7.1 PropH(+1) -6.4 Prop(0) -9.1 Model

b

EH → R Tyr → Tyr− (Fe(III) → Fe(II))b -7.8 -7.3 -4.4 •

a. The R → A transition is corrected to the experimental value of -1.0 kcal/mol, and the corresponding Fe(II)-correction used for the OH → EH transition is scaled according to the spin-populations in the EH state. b. The reductions occurring during the OH → EH and E → R are somewhat mixed with each other, see text. Turning to the directly calculated values, i.e. those that are not dependent on the Fe(II) correction discussed above, the last column in Table 1 shows that the oxygen bond cleavage step, A → PM , is very similar in all models, exergonic by 4.8-5.3 kcal/mol. The energetics

for the two reduction steps in the oxidative part of the catalytic cycle are described in Table 3. The PM → F step, corresponding to proton coupled reduction of Cu(II) (Cu(II)OH →

Cu(I)OH2 ) varies between 18.7 and 20.8 kcal/mol. Similarly, the F → OH transition,

corresponding to proton coupled reduction of Fe(IV) (Fe(IV)=O → Fe(III)OH), is almost

identical for the three models, 10.9-11.1 kcal/mol. Thus, the oxidative part of the catalytic cycle, A → OH , has a total exergonicity that varies between 35.1 and 36.5 kcal/mol among the different models (Table 3). Furthermore, for the reduction step described by the OH

→ E transition, which corresponds to proton coupled reduction of the tyrosine (TyrO•

→ TyrOH), all three models give similar energies, 10.1-12.1 kcal/mol, as shown in Table 2. To summarize the results in Tables 1 to 3, the effect on the calculated proton coupled

reduction potentials of increasing the size of the model, going from the Noprop model to PropH with the same total charged is less than 1 kcal/mol in almost all cases. Also, changing the total charge of the model, going from PropH to Prop, has a similar small effect on the calculated proton coupled reduction potentials. It is therefore concluded that all three models give the same picture in terms of proton coupled reduction potentials for all reductions involved in the entire catalytic cycle, as well as for the O-O bond cleavage step. The most difficult part in the calculated energy diagram for the catalytic cycle is the position of the EH state. As discussed above, this state is a mixture of electronic config22 ACS Paragon Plus Environment

Page 23 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

urations, TyrO• –Fe(II) and TyrO− –Fe(III), and clearly the mixing will be sensitive to the details of the model used. On top of that, one of the states involved, the Fe(II) state, is subject to a correction, also discussed above, which should further increase the uncertainty in the calculated energy level. Still, as is shown in Table 4, the variation in the calculated energetics for the OH → EH transition among the models, 6.4-9.1 kcal/mol, is only slightly

larger than the variations in the previously discussed reduction steps. In particular, the two

positively charged models give very similar and reasonable results, 6.4 and 7.1 kcal/mol, respectively. The result for the neutral model, though, 9.1 kcal/mol, is a bit too large, since it leads to a somewhat too small value for the next step, 4.4 kcal/mol. However, taking into account that several approximations are involved, these results should still be considered to support the suggested reduction mechanism. As described in section 2.3, by using another piece of experimental information, apart from the overall reaction energy, it is possible to describe the electron and proton uptake steps separately, see Fig. 5. It is expected that the calculated electron and proton affinities, i.e. properties that involve a change of the total charge, are more sensitive to the model used than the properties discussed above where the charge is not changed. This is also the case, as shown by the Tables 5 and 6, reporting electron affinities for different intermediates as calculated with the different models. In Table 5 the calculated electron affinities for the states suggested to be involved during catalytic turnover are given, i.e. the states described in the scheme in Fig. 4. As already discussed in section 3.1, the electron affinities, within each model, are quite similar for all four intermediates, and the reason for this is that the electron in all four steps goes to a tyrosyl radical. A comparison of the results obtained for the two models with the same charge (Noprop with values 101-105 kcal/mol and PropH with values 100-103 kcal/mol), shows that increasing the model decreases the calculated electron affinity by 2 to 4 kcal/mol. This means that the increased model stabilizes the initial, positively charged, state more than the final neutral state, which is reasonable. Since the positive charge is mainly located in CuB , and the model increase occurs on the heme and quite far from the copper ion, the effect on the calculated electron affinity is small (about 2 kcal/mol). However, in the OH state there is a hydroxyl bridge between Cu(I) and Fe(III), which means that the positive charge is partly delocalized to the heme iron. This may explain the somewhat larger effect (about 4 kcal/mol) from the increased heme model for the OH state. As expected, changing the total charge of the model, from +1 in the PropH model to 0 in the Prop model decreases the calculated electron affinity.

23 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Table 5: Calculated electron affinities in kcal/mol for the different intermediates involved during catalytic turnover, corresponding mainly to Tyr• → Tyr− . Model OH EH PM F Noprop(+1) 106.3 101.4 104.0 104.8 PropH(+1) 102.5 99.7 102.0 102.7 Prop(0) 98.1 88.4 97.2 95.6

The effect varies quite a bit between the intermediates, from 4 to 11 kcal/mol, with the largest effect for the EH which is not as pure a tyrosyl radical as the other intermediates, as discussed in section 3.1. It is noted that the introduction of an uncompensated negative charge on PropA has several effects on the intermediates, including changes in the charge and spin distribution. For example, the amount of radical character on the tyrosine in the different intermediates is somewhat unstable, and sensitive to the addition of charges. Therefore the spin-populations on the tyrosyl are generally decreased in the Prop model (from about 0.9 in the positively charged models to 0.7-0.8 in the neutral model). It can also be noted that it is generally believed that the electron transfer to the BNC is coupled to a proton uptake to the PLS, see Fig. 6, and therefore are the results for the electron affinities obtained with the Prop model most likely not as relevant as those obtained with the positively charged models. The main conclusion from the preceding paragraph is that all models actually give the same picture of the catalytic cycle during turnover, also when the individual steps of electron and proton transfer are taken into account. The parameter for the cost of the electron from cytochrome c (or heme a) should be chosen slightly different for the different models, but if this parameter is chosen in the same way, all models would give similar energy profiles. For the main model PropH the cost of the electron is put to 101 kcal/mol, giving a close to thermoneutral electron transfer for the PM to PR step, in accordance with experiments. This gives the energy profile in Fig. 5. For the results from the smaller model Noprop the corresponding value is chosen to 103 kcal/mol, resulting in the energy profile shown in Fig. S15. The latter energy profile corresponds (only slightly modified) to the energy profile in ref. [10]. If a similar procedure is applied to the results from the neutral model Prop, a value for the electron cost would be 94 kcal/mol, and a reasonably similar energy profile is obtained, see Fig. S16. It can be noted that the experimental reduction potential of 0.25 V for the electron donor cytochrome c, corresponds to 104.5 kcal/mol (using 4.28 V for the

24 ACS Paragon Plus Environment

Page 24 of 33

Page 25 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Table 6: Calculated electron affinities in kcal/mol for alternative intermediates with a protonated tyrosinea . Geometric and electronic (spin-populations) structures for these intermediates are shown in Figures S11-S13. Model

OB E F’ Cu(II)-OR→Cu(I)-OR Fe(III)-OH→Fe(II)-OH Fe(IV)=O→Fe(III)-O Noprop(+1) 85.0 87.6 87.3 PropH(+1) 89.7 86.2 82.6 Prop(0) 78.1 77.3 74.5 a. Unfortunately there is no common notation for these states.

standard hydrogen electrode [40]). Interestingly, the models corresponding to the situation with the pump loading site protonated (PropH and Noprop) has parametrized values for the donor of 101 - 103 kcal/mol, quite close to the experimental value, while the model without the pump proton, Prop, has a significantly lower value of 94 kcal/mol. Although the calculated absolute reduction potentials are not expected to be very accurate, these results support the view that a proton in the PLS is needed to stabilize the electron in the BNC. The results also indicate that a comparison between the two larger models give a rough estimate of the energetic effect of the pump proton. In Table 6 the calculated electron affinities for the intermediates with a protonated tyrosine are summarized. It should be noted that the electron affinities by themselves do not give a meassure of the corresponding proton coupled reduction potentials. As already discussed above these electron affinities are significantly smaller than the values for the states which instead has the proton in the center of the BNC and the tyrosine unprotonated, shown in Table 5, in many cases by as much as 20 kcal/mol. The main reason for the low electron affinities for these states is that due to the lack of a proton in the center of the BNC, unstable intermediates are formed when the electron enters, Fe(III)-O, Cu(I)-OR or Fe(II)-OH. In all cases another ligand, proton or R-group, should be present on the oxygen to stabilize the reduced state of the metal. The effect of increasing the size of the model, comparing model Noprop with model PropH, varies between the cases. Similar to the results in Table 5, a decrease of the electron affinity is expected since the increased model stabilizes the initial positively charged state more than the final neutral state. This is true for the two cases where the reduction occurs on the heme iron (F’ and E), see Table 6. However, the copper reduction (state OB ) is affected in the

25 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

opposite way by the larger heme model, with an increased electron affinity. The single ligand on the copper bound oxygen is denoted R in Table 6, and R corresponds to the entire heme a3 complex with surrounding residues. Apparently the increased size of the R ligand improves the stabilization of the unsaturated reduced copper complex, and thereby increases the electron affinity. The extra negative charge in the Prop model, as compared to the PropH model decreases the calculated electron affinities by about 10 kcal/mol. The main conclusion from this paragraph is that all models yield a significantly lower electron affinity for the states with a protonated tyrosine (Table 6) than the states with the proton in the center of the BNC yielding an unprotonated tyrosyl radical (Table 5), validating the importance of the tyrosine residue in the active site. Finally, in contrast to the inherent high proton coupled reduction potential for CuB obtained in the calculations [10], titration experiments on the resting oxidized state give a very low copper potential [4, 5, 6, 7]. To explain the low experimental potential it was suggested that when the flow of electrons has stopped, a relaxed resting state is formed by a slow uptake of an extra proton to the active site [32]. The protonated oxidized state was labled O+ P , and it has a protonated tyrosine plus a hydroxo bridge between Cu(II) and Fe(III) (see Fig. S14). It was shown by quantum chemical calculations that such a proton uptake results in a significantly lower proton coupled reduction potential [32]. Compared to the Cu(II) reduction potential corresponding to the PM → F transition shown in Table

3, the calculated proton coupled reduction potential for the O+ P state is 0.6 V lower using the Noprop model, which is the same as in ref. [32]. The corresponding effect is 0.8 V with the PropH model, and 0.7 V using the Prop model. Thus, all three models support the suggestion that a resting state with an extra proton, on top of the charge compensating protons taken up in every reduction step, has a low reduction potential, in agreement with equilibrium experiments.

4. Conclusions Using large cluster models, 140-200 atoms, of the active site in cytochrome c oxidase, hybrid density functional theory calculations have been performed to elucidate the details of the enzymatic reaction mechanism for oxygen reduction. The present results obtained with the larger models agree with previously published results using a smaller model [10, 32]. Together the computational studies give the following picture of the oxygen reduction

26 ACS Paragon Plus Environment

Page 26 of 33

Page 27 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

mechanism and the proton pumping, with particular emphasis on the role of the conserved active site tyrosine. The resting oxidized state is suggested to have an extra proton in the active site, as compared to the situation during catalytic turnover, and the uptake of this extra proton occurs slowly when there are no more electrons supplied. This state, labled O+ P , has a neutral protonated tyrosine, Cu(II) and Fe(III) with a bridging hydroxyl ligand. The proton coupled Cu(II) reduction potential for this state is found to be low, in agreement with equilibrium experiments of 0.28-0.35 V. Due to the extra proton there should be no proton pumping coupled to the initial reduction of this state. In contrast, the oxidized state during catalytic turnover, labled OH , has an unprotonated tyrosine with a tyrosyl radical, Cu(I) and Fe(III), which means that in the two reduction steps from the oxidized to the reduced state, OH → R, it is the tyrosyl radical and Fe(III) that are reduced. A major

result from the computational studies is that the intermediate state, which is referred to as the one electron reduced state, and labled EH , also has an unprotonated tyrosine, i.e. the proton has entered the center of the BNC. This state is a mixture of two electronic configurations, the main one with a tyrosyl radical and Fe(II) and a secondary one with tyrosinate and Fe(III). In this way the reduction energy of the high potential tyrosyl radical is shared with the low reduction energy of heme a3 Fe(III), and both reduction steps become exergonic enough to allow proton pumping also with a significant gradient. In agreement with the generally accepted reaction mechanism, molecular oxygen binds reversibly to the reduced state with Fe(II) and Cu(I), and after the O-O bond cleavage there is a tyrosyl radical, Cu(II)-OH and Fe(IV)=O (PM state). According to the computational studies, the first reduction step after the O-O bond cleavage involves reduction of Cu(II) to Cu(I) and formation of a water molecule, and it leaves a tyrosyl radical also in the F state. At this stage the proton coupled reduction potential of CuB is high, on the order of 1 V. In the next reduction step Fe(IV) is reduced to Fe(III), and there is still an unprotonated tyrosyl with radical character in the OH state. In these two reduction steps, PM → F and F → OH , it is found to be thermodynamically more favourable to leave the tyrosine

unprotonated and let the proton go to the center of the BNC. In contrast, in the OH

→ EH reduction step it has to be assumed that there is a kinetic effect that causes the

avoidance of the lower lying E state with a protonated tyrosine. It is suggested that the kinetic effect is due to a high barrier in the K-channel, forcing the proton to enter via the D-channel ending in the center of the BNC, implying that three of the chemical protons

27 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

are taken up via the D-channel. The K-channel, ending at the active site tyrosine, is only used in the EH → R step, in which the tyrosine is reprotonated. This is in agreement with

experimental observations on the A-familiy oxidases, indicating that at least one proton is taken up via the K-channel during the reductive part of the cycle. In this context it can be noted that the B- and the C-family oxidases have only one proton channel, corresponding to the K-channel in the sense that it ends at the active site tyrosine, and should therefore not be able to avoid protonating the tyrosine. Interestingly, these forms of CcO are known to have a significantly lower proton pumping than the A-family. Another major result from the calculations is that the presence of the tyrosine in the active site, and the reduction mechanism presented here with an unprotonated tyrosine with radical character in all intermediate steps, are essential for the proton pumping. Most likely the uptake of the pump-proton to a temporary loading site, the PLS, is driven by an electrostatic coupling to the electron transfer into the active site. Therefore it is important that the product of each reduction step has a large enough affinity for uptake of the next electron to the active site already before the chemical proton has arrived. The calculations show that the intermediates with a proton in the center of the BNC and the tyrosine unprotonated (radical) secures such a high electron affinity, essentially the electron goes initially to a tyrosyl radical in all reduction steps. The alternative form of the intermediate states where the proton instead is delivered to the tyrosine are found to have a significantly lower electron affinity, most likely requiring uptake of the chemical proton before electron transfer to the active site, as has been suggested for the NO reductase reaction. Thus, the presence of the tyrosine in the active site enables the uptake of two protons per electron in cytochrome c oxidase. The initial proton uptake occurs to the pump loading site, and due to electrostatic repulsion to the second, chemical proton, the pump proton is expelled to the outside. It is finally suggested that, the tyrosyl radical is not easily observed experimentally, because it is erased by the uptake of an extra proton to the active site when the flow of electrons is stopped.

Acknowledgements The author is grateful to Per Siegbahn, Peter Brzezinzki, M˚ arten Wikstr¨om and Liao Rongzehn for valuable discussions.

28 ACS Paragon Plus Environment

Page 28 of 33

Page 29 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Supporting Information available at (http://pubs.acs.org): Geometric and electronic (spin-populations) structures for all intermediates obtained with the main model (PropH are shown in Figures S1-S14. Cartesian coordinates for the corresponding structures are given at the end. Energy profiles for the two other models (Norop and Prop are given in Figures S15-S16.

Author information E-mail: [email protected]. Phone: +46-8-16 26 16 Funding This work was generously supported by a grant from the Swedish Research Council (2012-2408). Computer time was provided by the Swedish National Infrastructure for Computing.

References [1] Bloch D, Belevich I, Jasaitis A, Ribacka C, Puustinen A, Verkhovsky MI, Wikstr¨om M ( 2004) The catalytic cycle of cytochrome c oxidase is not the sum of its two halves. Proc. Natl. Acad. Sci. USA 101, 529-33. [2] Brzezinski, P., Gennis, R.B. (2008) Cytochrome c oxidase: exciting progress and remaining mysteries, J. Bioenerg. Biomembr. 40, 521-531. [3] Brzezinski, P. (2004) Redox-driven membrane-bound proton pumps, Trends Biochem. Sci. 29, 380-387. [4] Babcock, G.T., Wikstr¨om, M. (1992) Oxygen activation and the conservation of energy in cell respiration, Nature 356, 301-309. [5] Wikstr¨om, M., Krab, K., Saraste, M. (1981) Cytochrome oxidase, A Synthesis, Academic Press. [6] Jancura, D., Berka, V., Antalik, M., Bagelova, J., Gennis, R.B., Palmer, G., Fabian, M. (2006) Spectral and kinetic equivalence of oxidized cytochrome c oxidase as isolated and “activated” by reoxidation, J Biol Chem 281, 30319-30325. [7] Kaila, V.R.I., Verkhovsky, M.I., Wikstr¨om, M. (2010) Proton-Coupled Electron Transfer in Cytochrome Oxidase, Chem. Rev. 110, 7062-7081. 29 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

[8] Ferguson, S.J. (2010) ATP synthase: From sequence to ring size to the P/O ratio, Proc. Natl. Acad. Sci. USA 107, 16755-16756. [9] Wikstr¨om, M., Hummer, G. (2012) Stoichiometry of proton translocation by respiratory complex I and its mechanistic implications, Proc. Natl. Acad. Sci. USA 109, 4431-4436. [10] Blomberg, M.R.A., Siegbahn, P.E.M. (2015) How cytochrome c oxidase can pump four protons per oxygen molecule at high electrochemical gradient, Biochim. Biophys. Acta. 1847, 364-376. [11] Qin, L., Hiser, C., Mulichak, A., Gavarito, R.M., Ferguson-Miller, S. (2006) Identification of conserved lipid/detergent-binding sites in a high-resolution structure of the membrane protein cytochrome c oxidase, Proc. Natl. Acad. Sci. USA 103, 16117-16122. [12] Blomberg, M.R.A., Siegbahn, P.E.M. (2014) Proton pumping in Cytochrome c oxidase: energetic requirements and the role of two proton channels, Biochim. Biophys. Acta. 1837, 1165-1177. [13] Radon, M., Pierloot, K. (2008) Binding of CO, NO, and O2 to Heme by Density Functional and Mulitireference ab Initio Calculations, J. Phys. Chem. A 112, 1182411832. [14] Vancoillie, S., Zhao, H., Radon, M., Pierloot, K. (2010) Performance of CASPT2 and DFT for Relative Spin-State Energetics of Heme Models, J. Chem. Theory Comput. 6, 576-582. [15] Olah, J., Harvey, J.N. (2009) NO bonding to Heme Groups: DFT and Correlated av Initio Calculations, J. Phys. Chem. A 113, 7338-7345. [16] Verkhovsky, M.I., Morgan, J.E., Wikstr¨om, M. (1994) Oxygen binding and activation in the reaction of oxygen with cytochrome c oxidase, Biochemistry 33, 3079-3086. [17] Noodleman, L., Case, D.A. (1992) Density-functional theory of spin polarization and spin coupling in iron-sulfur clusters, Adv. Inorg. Chem. 38, 423-470. [18] Siegbahn, P.E.M., Blomberg M.R.A. (2010) Bond-Dissociation Using Hybrid DFT, Int. J. Quant. Chem. 110, 317-322.

30 ACS Paragon Plus Environment

Page 30 of 33

Page 31 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

[19] Becke, A.D. (1993) Density-functional thermochemistry. III. The role of exact exchange, J. Chem. Phys. 98, 5648-5652. [20] Grimme, S., Anthony, J., Ehrlich, S., Krieg, H. (2010) A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu, J. Chem. Phys. 132, 154104. [21] Reiher, M., Salomon, O., Hess, B.A. (2001) Reparameterization of hybrid functionals based on energy differences of states of different multiplicity, Theor. Chem. Acc. 107, 48-55. [22] Jaguar 7.6 (2009) Schr¨odinger, LLC, New York, NY. [23] Blomberg, M.R.A., Siegbahn, P.E.M., Babcock, G.T. (1998) Modeling electron transfer in biochemistry: A quantum chemical study of charge separation in Rhodobacter sphaeroides and photosystem II, J. Am. Chem. Soc.120, 8812-8824. [24] Gaussian 09, (2010) Revision C.01, Gaussian Inc., Wallingford, CT. [25] Siegbahn, P.E.M., Blomberg, M.R.A., Blomberg, M.L. (2003) A Theoretical Study of the Energetics of Proton Pumping and Oxygen Reduction in Cytochrome Oxidase, J. Phys. Chem. B 107, 10946-10955. [26] Blomberg, M.R.A., Siegbahn, P.E.M. (2006) Quantum Chemistry Applied to the Mechanisms of Transition Metal Containing Enzymes - Cytochrome c Oxidase a Particularly Challenging Case, J. Comp. Chem. 27, 1373-1384. [27] Siegbahn, P.E.M., Blomberg, M.R.A. (2008) Modeling of Mechanisms for Metalloenzymes Where Protons and Electrons Enter or Leave, in Computational Modeling for Homogeneous Catalysis and Biocatalysis, p57-81, edited by K. Morokuma and J. Musaev, Wiley-VCH, Germany. [28] Siegbahn, P.E.M., Blomberg, M.R.A. (2010) Quantum chemical studies of protoncoupled electron transfer in metalloenzymes, Chem. Rev. 110, 7040-7061. [29] Siegbahn, P.E.M. (2013) Water oxidation mechanism in photosystem II, including oxidations, proton release pathways, O-O bond formation and O2 release, Biochim. Biophys. Acta 1827, 1003-1019.

31 ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

[30] Blomberg, M.R.A., Siegbahn, P.E.M. (2012) Mechanism for N2 O Generation in Bacterial Nitric Oxide Reductase: A Quantum Chemical Study, Biochemistry 51, 5173-5186. ¨ [31] Br¨and´en, M., Sigurdson, H., Namslauer, A., Gennis, R., Adelroth, P., Brzezinski, P. (2001) On the role of the K-proton transfer pathway in cytochrome c oxidase, Proc. Natl. Acad. Sci. USA 98, 5013-18. [32] Blomberg, M.R.A., Siegbahn, P.E.M. (2015) Protonation of the binuclear active site in cytochrome c oxidase decreases the reduction potential of CuB , Biochim. Biophys. Acta. 1847, 1173-1180. [33] Sharma, V., Karlin, K.D., Wikstr¨om, M. (2013) Computational study of the activated OH state in the catalytic mechanism of cytochrome c oxidase, Proc. Natl. Acad. Sci. USA 110, 16844-16849. [34] Proshlyakov, D.A., Pressler, M.A, Babcock, G.T. Dioxygen activation and bond cleavage by mixed-valence cytochrome c oxidase (1998) Proc. Natl. Acad. Sci. U.S.A. 95, 8020-8025. [35] Wikstr¨om, M., Verkhovsky, M.I., Hummer, G. (2003) Water-gated mechanism of proton translocation by cytochrome c oxidase, Biochim. Biophys. Acta 1604, 61-65. [36] Belevich, I., Bloch, D.A., Belevich, N., Wikstr¨om, M., Verkhovsky, M.I. (2007) Exploring the proton pump mechanism of cytochrome c oxidase in real time, Proc. Natl. Acad. Sci. USA 104, 2685-2690. [37] Siegbahn, P.E.M., Blomberg, M.R.A. (2008) On the Proton Pumping Mechanism in Cytochrome c Oxidase, J. Phys. Chem A 112, 12772-12780. [38] Blomberg, M.R.A., Siegbahn, P.E.M. (2012) The Mechanism for Proton Pumping in Cytochrome c Oxidase from an Electrostatic and Quantum Chemical Perspective, Biochim. Biophys. Acta. 1817, 495-505. [39] Stuchebrukhov, A.A., Hammes-Schiffer, S. (2010) Theory of coupled electron and proton transfer reactions, Chem. Rev. 110, 6939-6960. [40] Marenich, A.V., Ho, J., Coote, M.L., Cramer, C.J., Truhlar, D.G. (2014) Computational electrochemistry: prediction of liguid-phase reduction potentials, Phys. Chem. Chem. Phys. 16, 15068-15106. 32 ACS Paragon Plus Environment

Page 32 of 33

Page 33 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

For Table of Contents Use Only: The mechanism of oxygen reduction in Cytochrome c oxidase and the role of the active site tyrosine. Margareta R.A. Blomberg F PM

PR

Cu(II) TyrO OH O Fe(IV) His

+e-

EA = 102 kcal/mol

Cu(II) TyrOOH O Fe(IV) His

Cu(I) TyrO

D-channel

+H+

x

+H+

K-channel

H 2O O Fe(IV) His Cu(II) TyrOH OH O Fe(IV) His

F'

33 ACS Paragon Plus Environment

+eEA = 103 kcal/mol

x

+e-

EA = 83 kcal/mol