Closing Kok's cycle of nature's water oxidation catalysis

Photosystem II (PSII) catalyzes biological water splitting, a crucial process in photosynthesis. This process is facilitated by the oxygen-evolving complex (OEC), a Mn4CaO5 cluster that cycles through five oxidation states (S0-S4), known as Kok's cycle. The cycle begins with the dark-stable S1 state. Upon illumination, the OEC accumulates oxidizing equivalents until water oxidation occurs during the S3(S4) → S0 transition, yielding molecular oxygen (O2). While significant progress has been made in understanding the earlier steps of Kok's cycle, the final transition, particularly the reconstitution of the S0 state after O2 release, remained elusive. Recent experimental breakthroughs using X-ray free electron lasers (XFEL) have provided structural insights into the S3 state, showing an extra oxygenic ligand (Ox) on Mn1. This Ox is believed to couple with the central µ-O5 to form the O-O bond in the subsequent S4 state. Following O2 release, the S0 state is restored through water insertion and proton expulsion, leaving Mn1 in a pentacoordinate state. However, the precise mechanism of S0 reconstitution and the intermediates involved remained largely unknown, representing a significant gap in our understanding of this fundamental biological process. This research aims to address this gap by employing computational methods, specifically Born-Oppenheimer ab initio molecular dynamics (BO-AIMD) simulations combined with density functional theory (DFT) calculations, to investigate the structural evolution of the OEC cluster after O2 release and elucidate the pathway to S0 state recovery.

Previous theoretical studies have explored the S4 → S0 transition, generally agreeing that W3 (a specific water molecule) refills the vacant site created by O2 release. A stepwise mechanism for O2 release and water insertion is favored over a concerted mechanism. The coupling of O5 and Ox is considered the most viable mechanism for O-O bond formation. However, the cluster evolution after O2 release, including the intermediates and detailed atomic-level transformations leading to S0 reconstitution, has not been fully elucidated. This lack of understanding highlights the need for computational studies to explore potential reaction pathways and intermediates.

This study employs BO-AIMD simulations and DFT calculations to model the water insertion dynamics and Mn4Ca cluster evolution following O2 release. The simulations start from a model of the OEC derived from XFEL data of the S3 state after removal of Ox and O5 to represent the O2-released state (Im0-O2). Two spin states (octet/αααβ and doublet/αβαβ) were considered. The BO-AIMD simulations utilize a large model including the Mn4CaO4 cluster, 20 amino acid residues, 24 crystal water molecules, and a chloride ion, treated with full quantum mechanical DFT calculations. This approach allows for unbiased exploration of the potential energy surface and identification of reaction pathways and intermediates. The long simulation time (at least 30 ps) enables observation of spontaneous, barrierless events. To further investigate longer timescale processes, minimum energy path (MEP) searches were performed using truncated DFT models. Various dispersion-corrected density functionals were employed (e.g., B3LYP, B3LYP*) along with the LanL2DZ and 6-31G* basis sets for Mn/Ca and other atoms respectively, ensuring accuracy. Mulliken spin populations and localized orbital bonding analysis (LOBA) was used to confirm the oxidation states of Mn ions. The reliability and reproducibility of the results were verified by comparing results from different spin states and DFT functionals.

The BO-AIMD simulations reveal a three-stage process for water insertion: (1) W3 insertion towards Mn1, forming a strong hydrogen bond with W2(OH) and triggering a square-pyramidal to trigonal-bipyramidal conversion of Mn4 coordination; (2) proton transfer from W3 to W2 and simultaneous movement of W7 to Mn4, resulting in a closed-cubane structure (Im2); (3) deprotonation of W1, releasing a proton to the lumen. The formation of the closed-cubane structure (Im2) is energetically favorable and occurs on a picosecond timescale. The MEP calculations show that subsequent W2 dissociation and µ3-W3(OH) ligand transfer from Mn1 to Mn4 are facile processes with low activation barriers, leading to the open-cubane structure of the S0 state (S0A). Importantly, the study reveals a structural isomerism in the S0 state, with both closed (S0B) and open (S0A) cubane structures being nearly isoenergetic and readily interconvertible. This isomerization involves switching the orientation of the Jahn-Teller distortion axis of Mn4(III), with the interconversion occurring at a nanosecond timescale. The sensitivity of results to various DFT functionals and model variations was explored and found to not affect the overall conclusions. The observed transitions satisfy kinetic constraints of the S3 → S0 transition.

These findings resolve the missing link in the S4 → S0 transition by identifying the closed-cubane intermediate and subsequent isomerization to the open-cubane S0 state. This structural flexibility, involving closed and open-cubane forms, is a defining characteristic of the OEC, contributing to its high catalytic efficiency. The observation of the closed-cubane intermediate and the reversible S0 isomerization are consistent with experimental data, such as the multiline EPR signal for S0, which may reflect an equilibrium between two closely-energetic isomers. This isomerism may not be detected by current XFEL crystallography, but it plays a crucial functional role in water oxidation catalysis. The computationally proposed structural isomerism complements the existing understanding of the S1 and S2 state isomerisms, painting a more complete picture of the structural dynamics involved in Kok's cycle. This work has implications for the design of bio-inspired water-splitting catalysts.

This study provides a comprehensive computational model of the S4 → S0 transition in PSII, revealing a previously unknown mechanism involving closed-cubane intermediates and a reversible structural isomerism in the S0 state. The facile water insertion, proton transfer, and ligand rearrangements identified in this study account for the rapid recovery of the OEC to its initial state. This work highlights the importance of structural flexibility in the OEC for its remarkable catalytic activity and inspires the development of more efficient biomimetic water-splitting catalysts. Future research could focus on exploring the role of the protein environment in facilitating these transitions and examining the impact of mutations on the structural dynamics of the OEC.

The study focuses primarily on the local structural evolution around the Mn4Ca cluster, neglecting the effects of the broader protein environment and hydrogen-bonding network. The computational models, while large and comprehensive, may still be simplified representations of the complex biological system. While DFT calculations provide reliable qualitative insights, the quantitative energetics might contain some degree of uncertainty due to the inherent limitations of DFT methodology. Although the model used is one of the largest QM models for MD simulations conducted so far on the OEC, the dynamics observed might not fully capture very long timescale events, extending the timescale of the simulations could help verify these conclusions further.