Redox-coupled Proton Transfer Dynamics in Cytochrome c Oxidase

Principal Investigator:
Ville R. I. Kaila

Technical University of Munich

Local Project ID:

HPC Platform used:
SuperMUC of LRZ

Date published:


Biological systems have evolved to effectively capture, store, and transform energy from one form to another. In mitochondria, which function as power plants of the eukaryotic cell, this process is catalyzed by enzymes of the cell respiratory chain that convert the energy from foodstuff into an electrical gradient stored across a biological membrane. Cytochrome c oxidase (CcO) functions as a terminal electron acceptor in all aerobic respiratory chains. It catalyzes the reduction of oxygen from the air into water by using electrons that the enzyme receives from foodstuff. CcO further employs the free energy released from this process to pump protons across a membrane, establishing an electrochemical proton gradient across the membrane, which the cell further employs to drive energy-requiring processes. Nevertheless, despite decades of research, it still remains unclear how CcO pump protons across the membrane and furthermore, what prevents the protons from leaking backwards in the pumping process.

The Cco reaction can be given as:   O2+ 4H++ 4e- → 2H2O

Experiments show that when electrons travel through metal centers in CcO, two types of protons are transferred: the chemical protons are translocated to the active center to complete the oxygen reduction process, whereas the pumped protons are transferred across the membrane. CcO employs two channels for the uptake of these protons. All pumped protons are taken up from the so-called D-channel, whereas chemical protons originate from both the D- and K-channels, for reasons that remain unknown.

Experiments also suggest that the pumped protons are transiently stored at a "proton-loading site" (PLS) before they are ejected across the membrane. The PLS is thus likely to function as an important coupling element in the pumping process.

The aim of this study was to 1) identify the exact location of the PLS, 2) to elucidate how the electron transfer reactions through the metal centers modulate the proton transfer reaction barriers, and 3) to elucidate a molecular mechanism for the channel switching process (Figure 1).

Results and Methods

In order to study the energetics and dynamics of CcO during its catalytic cycle, we performed large-scale atomistic molecular dynamics (MD) simulations on microsecond timescales. To mimic the steps of the catalytic cycle, we modeled the redox-cofactors in different catalytic states based on quantum mechanical calculations, and studied how water dynamics near the activate site depend on the redox states of the enzyme.

To this end, we applied a "travelers’ problem" algorithm, which allowed us to identify shortest pathways along the water wires connecting proton donor and a proton acceptor groups. Computationally we therefore analyzed the MD trajectories using Dijkstra’s algorithm with Fibonacci heaps, where the proton donor (D- or K-channel residues) and acceptor (active center or PLS) were the source and sink of the graph, and water molecules formed the vertices (Figure 2).

Our MD simulations suggest that reduction of an electron-queuing center (heme a) increases the hydrogen-bonded connectivity to the PLS region, whereas reduction of the active site produces an electric field that connected the D-channel with the latter. Our findings thus suggest that water molecules in CcO are sensitive to the changes in redox states of the enzyme and reorganize themselves, providing a prerequisite for the proton transfer process.

To study the energetics of the actual proton transfer reactions, we extracted structures from the classical MD simulations, which were used as a starting point for performing hybrid quantum mechanics/molecular mechanics (QM/MM) MD simulations. QM/MM calculations allow to accurately model bond-formation/bond-breaking process by quantum mechanical (QM) models, while treating the explicit surroundings of the protein by classical models. To this end, we treated the reactive QM part using density functional theory (DFT) calculations, which provides an accurate description of the electronic structure of the systems.

Interestingly, the QM/MM simulations suggested that a proton can be stored in a water cluster above the active site where it remained as a delocalized Zundel cation (H5O2+) (inset Figure 1). Moreover, we found that the reduction of the nearby heme group strongly modulates the proton affinity of the PLS, and reduced the kinetic barriers for its protonation. Our findings thus suggest that that electrostatic effects play an important role in the gating process, i.e. in directing the protons to the right site at the right time.

Activation mechanism of the K-channel

In a second subproject, we studied why the so-called K-channel is activated during the second half of the catalytic cycle. To address this question computationally, we performed microsecond MD simulations in combination with QM/MM calculations in states that are experimentally known to link to activation of the K-channel. Our simulations suggested that the K-channel activates as a response of a specific oxidized intermediate in the active site. This intermediate produces an electric field that increases the amount of water molecules in the K-channel, which in turn lowers the proton transfer barriers.

Interestingly, our simulations also indicated that the connectivity from the D-channel is lost at this step. The molecular basis can be traced back to a dehydration effect, which is in turn induced by the specific structure of the active site that cannot stabilize the water wired-contacts. In order to quantify the kinetics of the proton transfer reaction, we performed QM/MM free energy calculations, in which the computed barriers were found to be close to the experimentally measured rates.

Our multi-scale computational studies performed here thus suggest that water molecules play an important role in the proton pumping process of CcO. Our simulations also identified important protein residues that can be experimentally tested to verify the predicted mechanisms, as well as spectroscopic signatures that are expected to arise during specific steps of the pumping cycle.

On-going Research / Outlook

The HPC power offered by SuperMUC played a crucial role in the realization of this challenging, but highly successful project. To this end, SuperMUC offered unique resources that enabled our large-scale simulations that provided an essential step to derive the mechanistic models. The current data produced with SuperMUC was key for our publications (2 & 3), producing new data that is used for our future research projects.

Research Team

Ana P. Gamiz-Hernandez, Ville R. I. Kaila (PI), Shreyas Supekar (all: Department of Chemistry, Technical University of Munich/TUM)

References and Links

2. Supekar S, Gamiz-Hernandez AP, Kaila VRI (2016) Protonated Water Cluster as a Transient Proton-Loading Site in Cytochrome c Oxidase. Angew Chem Int Ed, 55, 11940.
3. Dewetting transitions coupled to K-channel activation in cytochrome c oxidase. Supekar S, Kaila VRI – Chem. Sci., 2018, 9, 6703. DOI: 10.1039/c8sc01587b

Scientific Contact

Prof. Dr. Ville R. I. Kaila
Department of Chemistry
Technical University of Munich (TUM)
Lichtenbergstr. 4, D-85747 Garching (Germany)
e-mail: ville.kaila [@]

NOTE: This report was first published in the book "High Performance Computing in Science and Engineering – Garching/Munich 2018".

LRZ project ID: pr84gu

December 2019