Removal of Ca²⁺ from the Oxygen-Evolving Complex in Photosystem II Has Minimal Effect on the Mn₄O₅ Core Structure: A Polarized Mn X-ray Absorption Spectroscopy Study

Abstract

Ca²⁺-depleted and Ca²⁺-reconstituted spinach photosystem II was studied using polarized X-ray absorption spectroscopy of oriented PS II preparations to investigate the structural and functional role of the Ca²⁺ ion in the Mn₄O₅Ca cluster of the oxygen-evolving complex (OEC). Samples were prepared by low pH/citrate treatment as one-dimensionally ordered membrane layers and poised in the Ca²⁺-depleted S₁ (S₁′) and S₂ (S₂′) states, the S₂′Y_Z^• state, at which point the catalytic cycle of water oxidation is inhibited, and the Ca²⁺-reconstituted S₁ state. Polarized Mn K-edge XANES and EXAFS spectra exhibit pronounced dichroism. Polarized EXAFS data of all states of Ca²⁺-depleted PS II investigated show only minor changes in distances and orientations of the Mn–Mn vectors compared to the Ca²⁺-containing OEC, which may be attributed to some loss of rigidity of the core structure. Thus, removal of the Ca²⁺ ion does not lead to fundamental distortion or rearrangement of the tetranuclear Mn cluster, which indicates that the Ca²⁺ ion in the OEC is not critical for structural maintenance of the cluster, at least in the S₁ and S₂ states, but fulfills a crucial catalytic function in the mechanism of the water oxidation reaction. On the basis of this structural information, reasons for the inhibitory effect of Ca²⁺ removal are discussed, attributing to the Ca²⁺ ion a fundamental role in organizing the surrounding (substrate) water framework and in proton-coupled electron transfer to Y_Z^• (D1-Tyr161).

Graphical abstract

1. Introduction

A fundamental reaction in oxygenic photosynthesis is the light-driven oxidation of water, which takes place at photosystem II (PS II), an integral membrane protein complex.^1–4 The water splitting reaction is catalyzed by the oxygen-evolving complex (OEC), a lumenal protein-bound cluster containing four Mn and one Ca atom that are bridged by oxo/hydroxo-groups (Mn₄O₅Ca, Figure 1A). During the catalytic reaction, the complex undergoes successive advancement through five redox states, S₀ to S₄, before the spontaneous return to S₀, resulting in the release of O₂ (Figure 1A).^5,6 The S-state cycle of the OEC couples the one-electron photo-oxidation of the chlorophyll P₆₈₀ in the PS II reaction center to the four-electron chemistry of water oxidation occurring at the Mn₄O₅Ca cluster. Electron transfer from the Mn₄O₅Ca cluster to P₆₈₀^•+ proceeds via the redox-active tyrosine Y_Z (D1-Tyr161).

Figure 1.

(A) The S-state cycle in native PS II and the S′-state cycle of Ca²⁺-depleted PS II, inhibited after the S₂′Y_Z^• state, as well as a schematic view of the OEC in the S₁ state. (B) S′-state advancement in Ca²⁺-depleted PS II samples including two different methods to generate the S₂′ state, by dark-adaptation and in the presence of 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU), a herbicide that blocks electron transfer from acceptor quinone Q_A to Q_B, thus allowing advancement of the OEC only to S₂ in native PS II or S₂′ in Ca²⁺-depleted PS II.

The general consensus is that a valence configuration of Mn^III₂Mn^IV₂ for the dark-stable S₁ state and Mn^IIIMn^IV₃ for S₂ are consistent with spectroscopic and structural data and theoretical calculations.^8–14 The structure of the OEC has been investigated using X-ray diffraction (XRD) and a variety of spectroscopic methods. Crystal structures of PS II at resolutions of up to 1.9 Å^7,15–17 confirmed the presence of one Ca²⁺ ion, as had been shown previously by extended X-ray absorption fine structure (EXAFS) spectroscopy^18–23 and electron paramagnetic resonance (EPR) spectroscopy.^24–27 At present, the recent X-ray free-electron laser (XFEL) structure of PS II at 100 K provides a radiation-damage-free structure of the native resting state at a resolution of 1.95 Å.⁷ To understand the catalytic mechanism, however, knowledge of the chemical nature of the Mn₄O₅Ca cluster, including the Ca²⁺ binding mode, as PS II advances through the S-state cycle is required. X-ray absorption spectroscopy (XAS), which requires much lower X-ray dose than XRD,²⁸ using synchrotron radiation at cryogenic temperatures allows us to study catalytic intermediates of the reaction under the threshold of radiation-induced changes and can provide higher-resolution distance information on the local environment of the metal ions.^23,29–31

Polarized Mn EXAFS of PS II in oriented thylakoid membranes extracted from spinach yields detailed information about neighboring atoms of the four Mn atoms. Combined with a range-extended EXAFS method, polarized spectra demonstrated that in the S₁ state, there are three di-μ-oxo bridged Mn-(μ-O)₂-Mn vectors, two at ∼2.7 Å (Mn_B3-Mn_C2 and Mn_C2-Mn_D1)⁷ and one at ∼2.8 Å (Mn_A4-Mn_B3)⁷ distance, at an average orientation of ∼60° relative to the membrane normal.³⁰ Furthermore, one mono-μ-oxo bridged Mn-μ-O-Mn vector of 3.2 Å (Mn_B3-Mn_D1) aligned approximately along the membrane plane and Mn-(μ-O)₂-Ca vectors of ∼3.4 Å oriented approximately, on average, along the membrane normal (Mn_B3-Ca and Mn_C2-Ca)⁷ were observed. Sr EXAFS of Sr²⁺-substituted PS II^20,22 and Ca EXAFS²¹ confirmed interactions of Sr and Ca with Mn at a distance of 3.5 and 3.4 Å, respectively, with an angle of 0°–23° relative to the membrane normal.²² Moreover, a total of four Sr–Mn distances of ∼3.5 and ∼4.0 Å were found in all four accessible states, S₀ to S₃, which exhibit changes in the Mn–Mn and Ca/Sr–Mn distances, especially upon the S₂ to S₃ and S₃ to S₀ transitions.²³ Density functional theory (DFT) suggested that in the S₂ state, the OEC may exist in an equilibrium of two valence isomers, an open and a closed cubane conformation of ground spin states S = 1/2 and 5/2,^12,13,32 which give rise to a g = 2 multiline EPR signal (MLS)³³ and signals centered at g ≥ 4.1,^34–37 respectively. It was shown that small alcohols, such as MeOH and EtOH shift the equilibrium in favor of the MLS, suppressing the g ≥ 4.1 signals.

Ca²⁺ can be depleted from the OEC in PS II membrane preparations³⁸ using two different methods, which either make use of (i) high ionic strength during a salt wash, using 1–2 M NaC^39,40 or (ii) a pH drop to 3 using citric acid, which also functions as a Ca²⁺ chelator simultaneously.⁴¹ Since the NaCl treatment suffers from some drawbacks (Section S1 of the Supporting Information), Ca²⁺ depletions in this work were carried out based on the low pH/citrate method of Ono and Inoue.⁴¹ At pH 3, protons present at a high concentration of 1 mM neutralize negatively charged groups ligating the Ca²⁺ ion, which inhibits Ca²⁺ binding and leads to its release from the cluster. After bringing the thylakoid membranes back to physiological pH, rebinding of Ca²⁺ to the undamaged OEC is possible.

While the Ca²⁺-depleted OEC is unable to complete the reaction cycle and oxidize water,^39–41 the resting S₁′ state can be advanced to higher oxidation states via different illumination procedures.^25,26 Continuous illumination at 4 °C followed by dark-adaptation, a procedure called “annealing” in this paper, advances the OEC to the metastable S₂′ state. This S = 1/2 spin state exhibits a modified MLS with more lines (at least 27) of smaller average line spacing (5.5–6 vs 8.8 mT) than in native S₂, but it does not exhibit any high-spin g ≥ 4.1 signals (although they can be generated by near-infrared illumination).^35,36 During annealing, Ca²⁺-depleted PS II decays rapidly to S₂′ from another S₂-related state, S₂′Y_Z^•, which means that the S₂′ state cannot be further oxidized by the Y_Z^• radical. The interaction between the S = 1/2 spins of Y_Z^• and the S₂′ state generates a split EPR signal characteristic of the S₂′Y_Z^• state, while making the S₂′ signal unobservable in continuous-wave (CW) EPR. The peak-to-peak line width of this Y_Z^• split signal is ∼16 mT in spinach PS II in the presence of the extrinsic polypeptides.

A recent study of the Mn cluster in the S₂′ state using EPR and ⁵⁵Mn electron nuclear double resonance (ENDOR)spectroscopy revealed that the Ca²⁺ ion is not critical for maintaining the electronic structure of the open-cubane conformation,²⁷ which exhibits the same Mn oxidation states and principal electronic exchange-coupling scheme as the native S = 1/2 S₂ state.^42–45 Thus, a role for Ca²⁺ in only tuning the redox or electronic properties of the Mn cluster seems unlikely. While the reason for the inhibition upon removal of the Ca²⁺ ion and thus its precise function remain elusive, the following roles have also been suggested. (i) The Ca²⁺ ion could act as an (intermediate) substrate binding site.^{13,17,23,46–48} (ii) The water binding function of Ca²⁺ could also be that of a “gatekeeper” for H₂O, regulating its access to the cluster and coordinating the specific binding mode to activate it.^26,49 (iii) The fact that Sr²⁺ is the only metal ion able to functionally replace Ca^2+41,50,51 leads to the conclusion that their property as strong Lewis acids with similar electronegativities, giving rise to similar pK_a values of H₂O ligands, may be critical for substrate deprotonation during catalysis⁵² and (iv) proton-coupled electron transfer from the OEC to Y_Z^•.^52,53

More detailed knowledge about the structure of the Ca²⁺-depleted OEC and the effect of Ca²⁺ removal will be useful to more precisely define the ligation mode of the Ca²⁺ ion to the tetranuclear Mn cluster in catalytically relevant states of the reaction cycle and its structural and mechanistic role in photosynthetic water oxidation. Employing polarized Mn K-edge XAS on one-dimensionally ordered Ca²⁺-depleted and Ca²⁺-reconstituted PS II samples, the current study refines the results obtained from XAS on nonoriented samples¹⁹ by providing angular information for Mn–Mn vectors and incorporating the latest knowledge about the OEC structure and function. In this study, the accessible Ca²⁺-depleted states are compared, along with the untreated Mn₄O₅Ca cluster, while Ca²⁺-reconstituted PS II serves as a control for the integrity of the samples after the Ca²⁺ depletion process and the irreversible changes, if any, that are introduced by that process.

2. Experimental Procedures

2.1. Sample Preparation

All work with isolated thylakoid membranes was performed in the dark or under dim green light, and PS II was kept at 4 °C before storage in the dark either at −80 °C or in liquid N₂. PS II-enriched thylakoid membranes were prepared from spinach by detergent treatment using Triton X-100.^38,54

Ca²⁺ depletion and reconstitution based on the low pH/citrate treatment method⁴¹ was achieved as described previously.²⁰ Before use, glassware and plasticware were acid-washed (10% HCl), and Ca²⁺-free buffers were treated with Chelex 100 Resin to remove Ca²⁺ contamination. Untreated, Ca²⁺-depleted and Ca²⁺-reconstituted PS II in the final buffer (0.4 M sucrose, 50 mM 2-(N-4-morpholino)ethanesulfonic acid (MES), 15 mM NaCl, 5 mM MgCl₂, 2.5% EtOH (v/v), pH 6.5 with or without 5 mM CaCl₂ for Ca²⁺-containing and Ca²⁺-free samples, respectively) was centrifuged for 1 h at 23000g to obtain highly concentrated pellets.

PS II pellets were transferred to custom-built Lucite sample holders fitting both into an EPR resonator cavity and the XAS cryostat. For dark state isotropic samples (S₁′ and S₁ states), the cavity of 40 μL sample holders was filled entirely, whereas for light-induced samples (S₂′, S₂′Y_Z^•, and S₂ states), the membranes were spread as a thin film of ∼0.5 mm on the Mylar tape backing of the sample holders. For oriented samples, to preferentially align the PS II membrane normal along the substrate normal, the membranes were painted as a very thin and even layer on the back of the Mylar tape using a paint brush and then dried under a stream of N₂ gas for ∼30 min.⁵⁵ Sufficiently thick samples for XAS measurements were generated by repeating this paint-and-dry process 6 times. All samples were dark-adapted for 1 h to accumulate the S₁′ or S₁ state before freezing them for storage. S state advancement was carried out by illumination of the sample in a nonsilvered dewar from the front and the back with two tungsten lamps employing aqueous 5% CuSO₄ infrared filters to reduce any heating from the illumination. Illumination at 4 °C for 2 min advanced the S₁′ to the S₂′Y_Z^•, when samples were frozen immediately afterward,^25,26 and to the S₂′ state, when samples were dark-adapted for 15–20 min at 4 °C after illumination (Figure 1B).¹⁹ Advancement to the S₂ state to assess the Ca²⁺ content by means of MLS intensities was done by illumination of untreated, Ca²⁺-depleted, and Ca²⁺-reconstituted control samples at 195 K for 20 min.²⁶

2.2. Oxygen Evolution Measurements

Steady-state PS II enzyme activity at 25 °C was determined by the measurement of the O₂ concentration in a PS II-containing assay mixture using a YSI 4004 Clark-type electrode (Yellow Springs Instruments) with a high sensitivity Teflon membrane under continuous illumination with a tungsten lamp through an aqueous 5% CuSO₄ infrared filter. The assay medium was the sample buffer lacking EtOH and with 0.4 mM 2,6-dichloro-p-benzoquinone (DCBQ, 50 mM in dimethyl sulfoxide) added as an electron acceptor. O₂ evolution activities were determined as an average of at least 8 single measurements at a minimum of 2 different chlorophyll concentrations ranging from 5 to 25 μg/mL.

2.3. EPR Spectroscopy

EPR spectroscopy was used for characterization and quality assurance of the samples before use in XAS studies. Every separate batch of samples was tested for damage of the OEC, Ca²⁺ content, S state composition and, for oriented samples, degree of orientation.

Low-temperature X-band CW EPR spectra at 9.28 GHz were recorded on an E-109 EPR spectrometer (Varian, Inc.) equipped with a standard TE₁₀₂ cavity and a Helitran liquid helium cryostat (Air Products). For oriented samples, the angle between the substrate plane and the direction of the magnetic field was 0°, if not stated otherwise. A Ca²⁺-reconstituted sample for each batch of samples was screened for the release of Mn^II characterized by its six-line signal, centered at g = 2, as an indicator of damage to the Mn₄O₅Ca cluster. The temperature was 8 K, the microwave power 0.5 mW and the modulation amplitude 32 G. None of the batches of Ca²⁺-depleted and Ca²⁺-reconstituted samples contained detectable amounts of Mn^II. In order to determine the mosaic spread Ω as a measure for the degree of disorder (Section S2 of the Supporting Information), EPR spectra of Y_D^• (D2-Tyr160) at g = 2 were collected from one-dimensionally ordered samples at orientations of 0° and 90° relative to the magnetic field at 20 K and 25 μW microwave power with 2.5 G modulation amplitude.^22,56 Spectra of the respective Mn₄O₅Ca cluster signals were recorded for each sample at 8 K with 30 mW microwave power and 32 G modulation amplitude. The various MLS spectra were background-corrected by light-minus-dark subtraction (i.e., the difference between the spectra after and before sample illumination).

2.4. XAS Data Collection

Mn K-edge X-ray absorption spectra were collected at Stanford Synchrotron Radiation Lightsource (SSRL) at an electron energy of 3.0 GeV and an average current of 80–100 mA on beamlines 7-3 and 9-3 using a Si(220) double crystal monochromator. These were equipped with a CF 1208 liquid helium flow cryostat (Oxford Instruments) with exchange gas (He) in the sample space to maintain a sample temperature around 10 K in order to minimize radiation damage. The intensity of the incident X-ray was monitored by an N₂-filled ion chamber (I₀) in front of the sample. The total photon flux on the sample was limited to 1 × 10⁷ photons per μm², which was determined to be non-damaging on the basis of detailed radiation-damage studies of PS II solution samples.²⁸ Spectra were acquired by sweeping the energy of the incident X-ray beam from 6400 to 7108 eV. They were energy-calibrated with respect to the pre-edge peak of KMnO₄, which was placed between two N₂-filled ionization chambers (I₁ and I₂) after the sample, at 6543.3 ev⁵⁷ or PS II sample orientations of 45° and 80° toward the incident beam. For PS II sample orientations of 10°, at which no transmission signal was available, a monochromator crystal glitch at 6460.3 eV was used for energy calibration. Absorption spectra were recorded indirectly via the Mn Kα fluorescence peak using a 30-element solid-state Ge detector (Canberra Instruments) with an energy resolution of ∼150 eV.

2.5. XAS Data Analysis

Energy calibration and averaging of normalized fluorescence data, weighted according to their signal-to-noise ratio, was performed using the EXAFSPAK suite of programs by Graham N. George and Ingrid J. Pickering (University of Saskatchewan). Data reduction was done using the SIXPACK program package⁵⁸ by Sam Webb (SSRL).Second derivatives of the edge spectra were obtained by a third order polynomial fit over a range of ±3 eV around each point. For the conversion of data from E-space to k-space, E₀ = 6561.3 eV was used for the Mn K-edge energy and a spline curve, fit to data above 1 Å⁻¹, was applied to minimize low-frequency background contributions. Fourier transforms were calculated for k³-weighted data ranging from 3.6 to 11.2–11.9 Å⁻¹, truncated at zero crossings in order to minimize apodization distortions. Curve fitting to Fourier isolates back-transformed from R′- to k-space was done according to the EXAFS eq (Section S3 of the Supporting Information) with E₀ and R_j, N_{app, j}, and σ_j for each coordination shell j as fitting variables, using SIXPACK. Ab initio amplitude, phase, and mean-free path functions f_j(π,k,R_j), α_j(k), and λ_j(k) were calculated using FEFF6L (University of Washington).⁵⁹ For the amplitude reduction factor S₀², the empirical value 0.85 from studies on model compounds was used. Relative angles ϕ_j of the absorber-backscatter vectors and isotropic coordination numbers N_iso,j were obtained from the dependence of N_app,j on the sample orientation θ between the substrate normal and the incident X-ray E-vector, given by eq 1, by fitting of the linear correlation between N_app,j and 3 cos²θ−1.

$$N_{\text{app},j}(\theta )=N_{\text{iso},j}+0.5N_{\text{iso},j}(3{cos}^2\theta -1)\times (3{cos}^2{\varphi }_j-1)-I_{\text{ord}}(\mathrm{\Omega })$$ (1)

The order integral I_ord is defined by the mosaic spread Ω as

$$I_{\text{ord}}(\mathrm{\Omega })=0.5\frac{{\int }_0^{\pi /2}sin\alpha \times (3{cos}^2\alpha -1)\times exp(-{\alpha }^{2}ln2/\mathrm{\Omega })d\alpha }{{\int }_0^{\pi /2}sin\alpha \times exp(-{\alpha }^{2}ln2/\mathrm{\Omega })d\alpha }$$ (2)

A more detailed description of polarized EXAFS can be found in refs 22 and 30.

3. Results

3.1. Sample Characterization

We prepared PS II samples in four well-defined states, Ca²⁺-depleted S₁′, S₂′, S₂′Y_Z^•, and Ca²⁺-reconstituted S₁ (Figure 1). The O₂ evolution activities were measured for native, Ca²⁺-depleted, and Ca²⁺-reconstituted samples to ensure sample integrity and quantify the degree of Ca²⁺ depletion. Because the Ca²⁺-depleted OEC fraction cannot complete the S-state cycle and therefore does not evolve oxygen, the O₂ activity serves as a measure for the fraction of Ca²⁺-containing metal clusters in the sample and, thus, for the degree of Ca²⁺ depletion and reconstitution. The degree of reconstitution is an indicator for the integrity of the samples, since irreversible damage of the PS II complex as a result of the treatment with citric acid at low pH can prevent the restoration of O₂ evolution activity. The O₂ evolution rates of different sample preparations were 340–400 μmol O₂/mg Chl/h for native, 50–80 μmol O₂/mg Chl/h for Ca²⁺-depleted, and 250–300 μmol O₂/mg Chl/h for Ca²⁺-reconstituted samples. This corresponds to ∼82% Ca²⁺-free Mn clusters in the Ca²⁺-depleted PS II samples on average, out of which ∼69% could be reconstituted by the addition of Ca²⁺.

The degrees of Ca²⁺ depletion and Ca²⁺ reconstitution were evaluated also on the basis of the S₂ state MLS arising from the fraction of Ca²⁺-containing OEC upon illumination at 195 K. At this temperature, the Ca²⁺-free metal cluster in the S₁′ state, which exhibits no X-band perpendicular-mode EPR signal, cannot be advanced to any other oxidation state.²⁶ The relative MLS intensities in light-minus-dark difference spectra of untreated, Ca²⁺-depleted and Ca²⁺-reconstituted PS II samples normalized to the Y_D^• signal confirmed Ca²⁺ depletion and reconstitution (Figure S1 of the Supporting Information).

Advancement of Ca²⁺-depleted samples from S₁′ to the designated higher oxidized states upon illumination at 4 °C was confirmed by the presence of their characteristic EPR signals, the S₂′ altered MLS, and the S₂′Y_Z^• split signal (Figure 2, middle and bottom, respectively). We note that these spectra also show signals from smaller fractions of other states (consistent with the O₂ assays); the S₂′Y_Z^• sample (not dark-adapted) exhibits residual S₂ state MLS (compare Figure 2, top) from the Ca²⁺-containing PS II fraction, and the S₂′ spectrum contains traces of the S₂′Y_Z^• split signal. The yield of S₂′ state cannot be reliably quantified from the modified MLS spectra. Comparison to spectra from S₂′ state samples generated in the presence of DCMU^19,27 (see Figure 1B) and the known lowered inflection point energy of the Mn K-edge upon annealing¹⁹ (see sections 3.2 and 4.1.1), however, suggest that further reduced S₁′ or even S₀′ states may be present besides S₂′.

Figure 2.

EPR light-minus-dark difference spectra of the native S₂ state and Ca²⁺-depleted samples in the S₂′ and S₂′Y_Z^• states (top to bottom). Due to its inherently smaller signal amplitude, the S₂′ altered MLS is displayed on a four times expanded vertical scale. The overlapping Y_D^• signal (g = 2) was excised for clarity of presentation. At g ≈ 1.84, contributions from the Q_A⁻-Fe²⁺ semiquinone-iron cofactor⁶⁰ were present in all samples. Continuous illumination conditions are as follows. S₂ (top): 195 K, 20 min; S₂′ (middle): 4 °C, 2 min + 15 min dark-adaptation; and S₂′Y_z^• (bottom): 4 °C, 2 min. EPR conditions: microwave frequency, 9.28 GHz; microwave power, 30 mW; modulation amplitude, 32 G; and temperature, 8 K.

For the polarized XAS experiments, one-dimensionally ordered PS II samples in each of the four different states of the OEC investigated were prepared, in addition to randomly oriented ones. The mosaic spread Ω values of the ordered membrane layers were evaluated by means of the orientation-dependent EPR signal of the oxidized Y_D^• radical (Figure S2 of the Supporting Information). For a detailed explanation of the evaluation of orientational disorder within the samples, see section S2 of the Supporting Information. The dichroic ratio of the oriented samples used for XAS experiments ranged from 3.3 to 5.3 with an average of 3.9 ± 0.6. According to the calibration curve of Andrews,⁵⁶ which relates the dichroic ratio to Ω via the angular dependence of the cytochrome b₅₅₉ EPR signal, the mosaic spread was found to be 15–20°. For evaluation of the angle dependence of the apparent coordination numbers N_app,j, derived from polarized EXAFS data (sections 2.5 and 3.5),³⁰ the lower limit of 20° was chosen.

3.2. Polarized XANES

Polarized Mn K-edge XANES spectra of the oriented samples poised in the three Ca²⁺-depleted states S₁′, S₂′, and S₂′Y_Z^•, and the Ca²⁺-reconstituted S₁ state at angles θ of 10° and 80° between the sample normal and the X-ray E-vector are shown in Figure 3 together with the second derivatives. The spectrum of an anisotropic sample collected at the magic angle of ∼54.7°, where the angular-dependent term in eq 1 becomes zero, is identical with the isotropic spectrum of a nonoriented sample. Taking into account the angular dependence of the absorption coefficient μ(E) ∝ 3 cos² θ–1, the isotropic spectrum can be calculated from two polarized spectra measured at different angles θ. The isotropic spectra calculated from the 10° and 80° polarized data are consistent with the spectra obtained from nonordered samples (Figure S3 of the Supporting Information). As Mn K-edge XANES is a direct probe of the electronic structure of the metal complex, this shows that the Ca²⁺-depleted and Ca²⁺-reconstituted OEC is not perturbed by the layering procedure. The isotropic XANES spectra of the different states are compared in Figure 4.

Figure 3.

Mn K-edge absorption spectra (top) and corresponding second derivatives (bottom) of one-dimensionally ordered Ca²⁺-depleted (A) S₁′, (B) S₂′, and (C) S₂′Y_Z^•, and (D) Ca²⁺-reconstituted S₁ state samples at orientations θ of 10° (solid red lines) and 80° (dashed blue lines).

Figure 4.

Isotropic Mn K-edge absorption spectra and corresponding second derivatives from Ca²⁺-depleted S₁′ (solid black lines), S₂′ (dashed red lines), and S₂′Y_Z^• (dotted green lines), and Ca²⁺-reconstituted S₁ state samples (dash-dotted blue lines). The spectra were calculated based on polarized data (10° and 80°, Figure 3) and isotropic spectra from nonoriented PS II samples (54.7°, Figure S3 of the Supporting Information).

The inflection point energy (IPE) of the Mn K-edge was extracted from the zero crossing of the second derivative in the energy region between 6550 and 6554 eV (bottom of Figures 3 and 4 and Figure S3 of the Supporting Information). IPE values for the different states from oriented (10°, 80°) and nonoriented (54.7°) samples (the numbers of which are shown in parentheses) are listed in Table 1. The isotropic values in the last column were calculated by linear fits from the three orientations. The IPE of the S₂Y_Z^• state, made by illumination of the S₁′ state at 4 °C, was shifted to a higher energy by ∼1 eV from that of S₁′. In contrast, the S₂′ state made by dark adaptation of the S₂′Y_Z^• state at 4 °C has a lower IPE than those of S₁′ and the S₂′ state generated by illumination in the presence of DCMU (see Discussion section 4.1.1).¹⁹ As in the native S₁^29,30,61 and S₂ states,²⁹ the dichroism of the spectra is expressed by a higher IPE at 80° compared to 10° in all four cases, with the IPE difference ranging from 0.6 eV in S₁′ state samples to 1.7 eV in S₂′Y_Z^• state samples.

Table 1. IPE Values in eV of Anisotropic Samples at Orientations θ of 10° and 80° and Isotropic Values (from Isotropic Samples, Corresponding to an Angle of 54.7°, as well as Calculated by Linear Regression Analysis of the Experimental Values^a)^b.

		10°	80°	54.7°	isotropic
Ca2+-depleted	S1′	6551.2 ± 0.6 (4)	6551.8 ± 1.1 (4)	6551.3 ± 0.1 (3)	6551.5 ± 0.2
	S2′	6550.3 ± 0.4 (2)	6551.1 ± 0.6 (2)	6550.8 (1)	6550.8 ± 0.2
	S2′YZ•	6551.6 ± 0.1 (2)	6553.3 ± 0.0 (2)	6552.4 ± 0.6 (2)	6552.6 ± 0.1
Ca2+-reconstituted	S1	6550.8 ± 0.4 (2)	6551.7 ± 0.9 (2)	6551.8 ± 0.0 (2)	6551.6 ± 0.2

^aThese averages of the IPE values of the individual samples differ insignificantly from the energies of the zero crossings of the corresponding averaged second derivative curves (Figures 3 and 4 and Figure S3 of the Supporting Information) due to the not completely identical averaging methods.

^bNumbers of samples given in parentheses.

3.3. Polarized EXAFS

k³-Weighted Mn K-edge EXAFS spectra and their Fourier transforms for one-dimensionally oriented samples measured at orientations θ of 10° and 80° in the series of Ca²⁺-depleted and Ca²⁺-reconstituted states are shown in Figure 5. The apparent distance R′ is shorter than the actual distance due to a phase shift induced by the interaction of the respective absorber-backscatterer pair with the photoelectron. The isotropic Fourier transforms calculated from the 10° and 80° polarized data are highly similar to the spectra obtained from isotropic solution samples (Figure S4 of the Supporting Information), confirming structural integrity of the OEC in oriented Ca²⁺-depleted and -reconstituted samples.

Figure 5.

(A) k³-weighted EXAFS spectra and (B) Fourier transforms of the k³-weighted EXAFS spectra of one-dimensionally ordered Ca²⁺-depleted states S₁′, S₂′, S₂′Y_Z^•, and Ca²⁺-reconstituted S₁ state samples (top to bottom) at θ = 10° (solid red lines) and 80° (dashed blue lines). For Fourier transformation to R′-space, the k-space limits were set to the zero crossings of the lowest (∼3.6 Å⁻¹) and highest (11.2–11.9 Å⁻¹) k in this region.

For all states, EXAFS data show clear angular dependence. All Fourier transforms exhibit the same general appearance of three peaks I, II, and III, characteristic for the OEC. Peaks I and II correspond to Mn–O and Mn–N interactions at 1.8–2.0 Å and Mn-(μO)₂-Mn interactions at 2.7–2.8 Å, respectively. These peaks are higher in amplitude at θ = 80° compared to θ = 10°, as also observed for one-dimensionally oriented native samples.^29,30,61 In the Ca²⁺ -reconstituted OEC, peak III is more intense at 10° than at 80°, similar to the native Mn₄O₅Ca cluster in the S₁ and S₂ states, albeit less dichroic.²⁹ In the native Mn₄O₅Ca cluster, Mn-μO-Mn and Mn-(μO)₂-Ca interactions at ∼3.3 Å give rise to peak III.^18-23 Due to the lack of Mn–Ca interactions, this peak has a smaller amplitude for the Ca²⁺-depleted states than for the Ca²⁺-reconstituted state, as has been reported before for isotropic samples.¹⁹ Also, peak II in the Ca²⁺-reconstituted OEC is found to be larger in amplitude than in the Ca²⁺-depleted states. Overall, no major spectral changes are observed between the Ca²⁺-depleted and reconstituted samples, indicating that the spatial arrangement of the tetranuclear Mn cluster remains the same.

3.4. EXAFS Curve Fitting

In order to extract quantitative information from the Mn K-edge EXAFS spectra describing the local structure of the OEC in the investigated states and to reveal subtle differences between them, curve fitting to Fourier isolates (0.9 Å < R′ < 3.5 Å) in k-space was performed. Features at R′ > 3.5 Å and R′ < ∼1 Å, which result from high-frequency noise components in the k-space spectrum and imperfect background removal respectively, are filtered out by Fourier isolation. As in previous studies of the native²⁹ as well as theCa²⁺-depleted²⁹ OEC, the same three major paths corresponding to the individual peaks were used for the fits. For the Ca²⁺-containing S₁ state, the Mn-(μO)₂-Ca interactions contributing to peak III were not explicitly considered.^19,29 The Debye–Waller factors σ_j, which are highly correlated to the coordination number N_j, were fixed to reasonable numbers chosen empirically based on previous studies^{18,19,23,29–31} in order to compare the N_j values and thus the number of interactions in the different states and determine their orientations. SIXPACK returns the fit error parameters χ and χ², the latter of which takes into account the number of fit variables and independent data points. Absolute fit quality was chosen based on a combination of criteria, which include the above parameters and physical and chemical feasibility of the resulting parameters.

Curve fitting of k-space data of an oriented and a nonoriented S₁′ state sample are shown in Figure 6. All averaged absorber-backscatterer distances R_j and apparent coordination numbers N_app,j from fits to Fourier isolates are reported in Table 2. It can be seen that the EXAFS spectra of Ca²⁺-depleted samples can be fit well without incorporation of a Mn–Ca interaction, whereas this interaction needs to be accounted for in the Ca²⁺-reconstituted samples. The coordination numbers found reflect the dichroism seen in the EXAFS spectra.

Figure 6.

Fourier isolates (0.9 Å < R′ < 3.5 Å, solid colored lines) and corresponding EXAFS curve fitting results (dashed black lines) of a one-dimensionally ordered S₁′ state sample at θ = 10° (top) and 80° (bottom), and a nonordered S₁′ sample (θ = 54.7°, middle).

Table 2. Polarized EXAFS Fit Parameters R_j and N_app,j of the Coordination Shells j = 1–3 from Fits to Spectra of Individual Samples at the Different Orientations θ^a,b.

state	θ	shell 1 Mn–O		shell 2 Mn–Mn		shell 3 Mn–Mn(Ca)
		R1 (Å)	Napp,1	R2 (Å)	Napp,2	R3 (Å)	Napp,3
S1′	10°	1.84(1)	2.3(1)	2.70(1)	0.8(1)	3.27(4)	0.2(1)
	54.7°	1.85(2)	2.8(3)	2.72(2)	1.2(2)	3.24(4)	0.3(1)
	80°	1.87(1)	2.8(2)	2.75(1)	1.3(1)	3.31(5)	0.2(1)
S2′	10°	1.85(1)	2.4(1)	2.71(0)	0.9(1)	3.29(2)	0.3(0)
	54.7°	1.85	2.8	2.72	1.1	3.27	0.2
	80°	1.86(0)	2.8(3)	2.73(0)	1.1(1)	3.27(0)	0.3(1)
S2′YZ•	10°	1.87(0)	2.8(1)	2.73(0)	0.9(0)	3.29(1)	0.3(0)
	54.7°	1.87(2)	3.2(0)	2.74(3)	1.2(1)	3.30(2)	0.3(0)
	80°	1.88(1)	3.2(1)	2.76(1)	1.3(1)	3.32(3)	0.3(0)
S1	10°	1.83(0)	2.4(1)	2.70(0)	1.0(0)	3.27(1)	0.6(0)
	54.7°	1.85(0)	2.6(1)	2.73(1)	1.3(0)	3.25(0)	0.4(2)
	80°	1.86(1)	2.6(2)	2.74(0)	1.3(0)	3.25(1)	0.5(0)

^aStandard deviations are reported in parentheses as the uncertainty in the least significant digit, except for the parameters from the single nonordered S₂′ state sample.

^bThe Debye–Waller factors σ_j used for all samples were 0.006 Ų for the Mn–O/N interactions of peak I, 0.003 Ų for the Mn–Mn interactions of peak II, as well as for the Mn–Mn interaction of peak III in Ca²⁺-depleted samples, and 0.004 Ų for the Mn–Mn(Ca) interactions of peak III in Ca²⁺-reconstituted samples owing to the increased number of ligand atoms in the same shell.

3.5. Determination of N_iso,j and ϕ_j

The apparent coordination numbers N_app at different measurement angles θ depend on the angle ϕ of the absorber-backscatterer vector toward the membrane normal as given in eq 1. In order to obtain angular information for individual shells and the averaged isotropic coordination number N_iso,j, we carried out linear regression analysis of N_app,j with respect to θ. For all fits, the approximate average mosaic spread Ω of 20° assessed by EPR measurements was used.

Figure 7 shows linear plots in which the experimental N_app,j values are plotted against 3 cos² θ − 1. The ordinate intercepts of the linear fits correspond to N_iso, while the slope contains contributions from both N_iso and ϕ. Table 3 lists N_iso,j and ϕ_j from the linear regression analysis, as well as distances R_j of the backscattering atoms in the various states investigated, averaged over the samples of all orientations. Additionally, numbers for the native OEC in the S₁ state³⁰ reported previously are listed for comparison.

Figure 7.

Linear plots of the apparent coordination numbers N_app vs 3 cos² θ − 1 and corresponding fit curves for shells j = 1–3 of Ca²⁺-depleted S₁′, S₂′, S₂′Y_Z^•, and Ca²⁺-reconstituted S₁ state samples (top to bottom) both from spectra of individual samples (black) and from averages thereof (green). θ = 10° and 80° correspond to x = 1.9 and −0.9, respectively, in this representation. The magic angle θ ≈ 54.7° corresponds to x = 0, so that the intercept of the y axis equals N_iso.

Table 3. N_iso,j and ϕ_j for the Various Sample States from Linear Regression Analysis, as well as R_j Averaged over All Samples in a Particular State^a.

state	shell 1 Mn–O			shell 2 Mn–Mn			Shell 3 Mn–Mn(Ca)
	R1 (Å)	Niso,1	ϕ1 (°)	R2 (Å)	Niso,2	ϕ2 (°)	R3 (Å)	Niso,3	ϕ3 (°)
S1′	1.85(2)	2.7(1)	58	2.72(2)	1.2(0)	64	3.28(5)	0.2(0)	57
S2′	1.86(1)	2.7(1)	58	2.72(1)	1.1(0)	60	3.28(2)	0.3(0)	54
S2′YZ•	1.87(1)	3.1(0)	58	2.74(2)	1.2(0)	63	3.30(2)	0.3(0)	58
S1	1.85(1)	2.6(1)	57	2.72(2)	1.2(0)	59	3.26(1)	0.5(0)	50
native S1b				2.74(2)	1.3(3)	63(5)	3.20(2)b	0.4(1)b	>70b

^aStandard deviations are given in parentheses.

^bParameters for the native S₁ state Mn₄O₅Ca cluster from ref 30. The numbers given for shell 3 represent the Mn-μO-Mn contribution only, extracted by range-extended Mn K-edge EXAFS.

4. Discussion

We have studied the structural role of Ca²⁺ in the OEC using XAS by comparing the Ca²⁺-depleted and Ca²⁺-reconstituted samples. Previous studies by several groups including ours have reported that Ca²⁺ is a critical element for the catalytic activity of the OEC; without Ca²⁺, the S₂ to S₃ transition is blocked and Ca²⁺-depleted PS II loses its ability to evolve O₂.^19,25–27 The enzymatic activity can be mostly recovered by adding Ca²⁺ to the Ca²⁺-depleted samples.^39–41 To date, the precise reason for this inhibition in the absence of Ca²⁺ has still not been identified. While this study also does not provide a final answer to this question, the structural analysis of the Ca²⁺-depleted and Ca²⁺-reconstituted OEC on the basis of interatomic distances and binding angles in the following is able to rule out possible roles of Ca²⁺ in photosynthetic water oxidation.

4.1. Effects of Ca²⁺ Depletion and Reconstitution on Structure and S′-State Transitions

4.1.1. Mn Oxidation States

The IPE of an X-ray absorption K-edge indicates the difference between the ionization threshold and the 1s orbital energy. Thus, it is generally sensitive to the charge density at the absorber atom, which is most affected by the formal oxidation state. The Mn K-edge XANES spectra of the Ca²⁺-depleted S₁′ state and the Ca²⁺-reconstituted S₁ state show similar rising edge positions and spectral features (Figure 3, panels A and D, and Figure S3, panels A and D, of the Supporting Information), with IPE values of 6551.5 ± 0.2 eV and 6551.6 ± 0.2 eV, respectively. The large variation in the IPE of the native S₁ state in the literature (6551.5 ± 1.2 eV),^{8,10,31,62–69} which arises mostly because of the different data processing methods used, makes a comparison difficult. However, the Ca²⁺-depleted S₁′ and Ca²⁺-reconstituted S₁ state IPE values, matching the untreated S₁ IPEs reported by our group using the same method, indicate that the overall Mn oxidation state, assigned as Mn^III₂Mn^IV₂ for the S₁ state,^8–11,13,14 remains unchanged by Ca²⁺ depletion and reconstitution. This is in line with the EPR/⁵⁵Mn ENDOR-spectroscopic analysis of the Ca²⁺-depleted S₂′ state,²⁷ which reveals the same Mn valence state distribution as in the native S₂ state, Mn_D1^IVMn_C2^IVMn_B3^IVMn_A4^III. While identical IPEs for the native and Ca²⁺-reconstituted S₁ states have been observed in this as well as some of the previous studies,^18,70 we note that two studies reported 0.4–0.8 eV lower edge positions for S₁′ compared to Ca²⁺-containing S₁ states.^68,70 We think this discrepancy may result from different methods used in these previous studies, namely the use of half-height energies instead of IPEs⁷⁰ and Ca²⁺-depletion by a NaCl salt wash,⁶⁸ which could potentially destabilize the OEC by washing out some of the extrinsic polypeptides. With regard to S′-state advancement of the Ca²⁺-depleted OEC (Figure 1B), we observed a phenomenology described before: (i) The IPE of the S₂′Y_Z^• state is markedly higher than that of the S₁′ state (1.1 eV), consistent with oxidation of the Mn atoms during transition from S₁′ to S₂′Y_Z^•.^19,68,71,72 (ii) The IPE of S₂′ state samples, generated by annealing of the S₂′Y_Z^• state, is ∼0.7 eV below the edge position of the S₁′ state despite the one-electron oxidation, suggesting a modification of the electronic state during the dark-adaptation. This may indicate vulnerability to reduction of the Ca²⁺-free tetra-Mn complex as proposed before,¹⁹ even over the rather short times used for dark adaptation (15–20 min). The following findings suggest a relatively slow Mn reduction process in annealed S₂′ by a reductant accumulating during illumination of the inhibited PS II to be the cause: (i) After shorter annealing times (5–10 instead of 15–20 min), a considerably larger IPE lying between the S₁′ and the S₂′Y_Z^• state has been measured.⁶⁸ (ii) When the S₁′ state is advanced to the S₂′ state in the presence of DCMU^19,70 or by one laser flash,⁷¹ in both cases without a subsequent annealing step, the IPE is higher than that of the S₁′ state but slightly lower than that of S₂′Y_Z^•. It is noted that there are neither indications of Mn^II (EPR, XANES¹⁹) nor of considerable structural changes in the EXAFS, such as Mn-O and Mn-Mn distance elongations characteristic of Mn^II formation.^28,73 Additionally, no intensity changes were observed in the pre-edge region in the Ca²⁺-depleted and -reconstituted states (Figure 3 and Figure S3 of the Supporting Information), suggesting that no symmetry changes occur in their Mn ligand spheres.

4.1.2. Geometric Structure

In order to interpret the coordination numbers N_iso,j in a multinuclear system in terms of the number of absorber-backscatterer interactions I_j in one shell j, the total number of absorbing atoms A in the system and the number of absorbers B_j in the respective shell have to be considered according to eq 3:

$$N_{\text{iso},j}=I_j\times B_j/A$$ (3)

To determine I_j in the various shells for OEC, the values of N_iso,j from the fits have to be normalized by the four Mn atoms (by multiplication with A = 4). For interactions involving two absorbing atoms (i.e., for the Mn–Mn shells), they additionally have to be divided by B = 2.

Due to the strong correlation with the Debye–Waller factors, which are rather flexible over a certain range, and the large number of interactions in shell 1, their N_iso and absorber-backscatter angle ϕ_j cannot be interpreted quantitatively on this basis. However, by keeping σ_j constant under the assumption of a similar degree of distance disorder, N_iso can be compared for different states of the OEC (Table 3). The Ca²⁺-reconstituted S₁ and the Ca²⁺-depleted S₁′ and S₂′ states exhibit a similar N_iso of 2.6–2.7, indicating no changes in Mn coordination spheres and identical numbers of Mn–O/N ligands. The small increase in the intensity of peak I in the S₂′Y_Z^• state (Figure 5B, see also ref 72) results in a slightly larger N_iso of 3.1. This increase in intensity could be interpreted as an additional ligand to Mn in the S₂′Y_Z^• state. Another possible interpretation is a more uniform distribution of Mn-ligand distances in the interactions contributing to this Fourier peak in the S₂′Y_Z^• state if σ_j is not constant.

All four states investigated exhibit coordination numbers of ∼1.2 for shell two, similar to the native OEC (Table 3, refs 29 and 30). Thus, the peak II region, with an N_iso systematically smaller than 1.5, can be explained by three Mn-(μO)₂-Mn interactions at average distances between 2.7 and 2.8 Å: Mn_A4-Mn_B3, Mn_B3-Mn_C2, and Mn_C2-Mn_D1.^{7,13,32,74–76} While it also can be simulated by two shorter (∼2.7 Å) and one longer (2.8–2.9 Å) Mn-(μO)₂-Mn distances (not shown) as in native S₁ and S₂,^30,77 this arrangement cannot be proven for the Ca²⁺-depleted and reconstituted OECs due to the resolution limit of non-range-extended EXAFS. In terms of the peak III region, the coordination numbers in the Ca²⁺-depleted OEC are consistently lower than in the Ca²⁺- containing Mn₄O₅Ca clusters, immediately reflecting the absence of the Mn–Ca interactions and thus the absence of the Ca²⁺ ion in the Ca²⁺-depleted samples. The N_iso = 0.2–0.3 in the S′ states arises from a long Mn–Mn interaction around 3.3 Å,¹⁹ which is most likely the Mn_B3-μO-Mn_D1 interaction (Figure 8).^{7,13,32,74–76} The dominating contribution from peaks I and II in the k-space spectra, as well as the lower signal-to-noise level of peak III, lead to numbers smaller than expected. By comparison with both conventional and range-extended EXAFS on the native OEC previously reported by our group,^29,30 we conclude that there are also at least two Mn-(μO)₂-Ca interactions at ∼3.3 Å in the Ca²⁺-reconstituted S₁ state. Coordination numbers and distances for the S′ states are in good agreement with recent DFT-based structures of the Ca²⁺-depleted OEC.^74–76

Figure 8.

Structural model of the Ca²⁺-depleted OEC in the S₁′ state with a possible H₃O⁺ ion in place of the Ca²⁺ ion present in the native OEC.

In terms of distances and angles, no fundamental rearrangement of atoms upon Ca²⁺ removal or reinsertion is found in any of the investigated states (Table 3). However, small changes of internuclear distances seem to accompany these processes. Most pronounced is the lengthening of the Mn_B3-μO-Mn_D1 interaction in peak III from 3.20 Å in native S₁³⁰ to ∼3.3 Å upon Ca²⁺ depletion. The superposition of interactions in the small peak III of the Ca²⁺-containing OEC does not allow for interpretation of the exact distances and angles of the shell 3 fit to conventional EXAFS data. For all states, the other distances are highly similar to the native S₁ state (see also ref 11). Small elongations (0.02 Å) of distances in all shells seem to occur during the advancement of the Ca²⁺-depleted OEC to the S₂′Y_Z^• state. Evaluation of the dichroism in the polarized EXAFS spectra shows basically no changes in the angles (∼60°) of the Mn-(μO)₂-Mn vectors of shell 2 relative to the membrane normal, and are consistent with the ones found in the native S₁ and S₂ states.^29,30 The fact that their orientations are constant throughout all of the states investigated (S₁ and S₁′, S₂ and S₂′, or S₂′Y_Z^•) is in agreement with the above discussion of the interatomic distances and shows that only minor structural changes occur in the Ca²⁺-depleted OEC, as also suggested by recent computational models.^74–76

Altogether, Ca²⁺ depletion neither led to release of Mn from the cluster nor to a change in Mn oxidation states, Mn coordination geometries or the stoichiometry of Mn–Mn interactions in the individual ligand shells, or to rebinding of backscatterer atoms in place of the Ca²⁺ ion. The impact of the removal of the Ca²⁺ ion from the OEC is not an immediate distortion of the metal complex but rather leads to a marginal loosening of the cluster, evident from the increase (<0.1 Å) of the longest Mn–Mn distance, constituting peak III. Thus, Ca²⁺ removal slightly destabilizes the cuboidal core structure, likely affecting the Mn_B3-μO-Mn_D1 interaction. This provides a rationale for the fact that the Ca²⁺-depleted OEC in the S₂′ state does only adopt the low-spin, proposed open-cubane conformation,^12,13 as apparent from its EPR spectrum, featuring the altered S = 1/2 MLS but lacking the g = 4.1 signal with a proposed closed-cubane conformation. The Ca²⁺ ion can be rationalized as a requirement for maintaining the contracted, closed cubane of the high-spin conformation with a shorter Mn_B3-Mn_D1 distance of ∼2.9 Å,^12,13,32 or for advancement to an S₃ state, where a closed cubane-like structure has been proposed from EXAFS studies of native PS II.³¹

4.2. Role of the Ca²⁺ Ion in Photosynthetic Water Oxidation

The EXAFS results show that the Ca²⁺ ion is not replaced by other metal ions. Instead, loss of the Ca²⁺ ion is probably compensated by the binding of protons⁷⁵ and/or hydroxonium ions (Figure 8). The structural changes in distances, angles, and coordination numbers observed for the Ca²⁺-depleted S′ states reflect some loosening of the tetranuclear Mn complex. The retention of structural integrity in the Ca²⁺-depleted Mn cluster is the prerequisite that the Ca²⁺ ion can be exchanged and that reconstitution can reactivate the catalyst. Thus, the oxygen-bridged tetranuclear Mn open-cubane structure possesses an inherent stability even in the absence of the Ca²⁺ ion. Its nondestructive removal demonstrates that Ca²⁺ is not strongly bound to the cluster. This could be a consequence of some structural flexibility of the Mn₄O₅Ca cluster, as also seen in the S₂ state, in which the O5 ligand (Figure 1) of Ca²⁺ has been proposed to switch between two positions.^12,13 This is consistent with the observation that Ca²⁺ depletion is facilitated under illumination.^24,78 In the native S₂ state, the Ca²⁺ exhibits a more asymmetric, distorted ligand geometry in the open than in the closed conformation.¹² Stabilization of the Ca²⁺ coordination sphere by increasing its symmetry would then contribute to the low energy of the closed structure, only 1 kcal mol⁻¹ higher than the open one. As there is no such energetic gain in the absence of the Ca²⁺ ion, the open cubane is energetically favored over the closed form in the Ca²⁺-depleted S₂′ state.

Lately, a number of theoretical studies investigated the nature of the S₂Y_Z^• state, with varying outcomes.^13,48,53,79 Very recently, the first pulsed EPR-spectroscopic study of the OEC in the S₃ state suggested in combination with DFT computations that the cluster is in an open-cubane Mn^IV₄ state with an additional hydroxide bound to the hexacoordinate Mn_D1.⁸⁰ The group of Guidoni proposed the S₂-to-S₃ transition to proceed via a closed-cubane structure in the S₂Y_Z^• state^48,79 (see also ref 13), which could be envisaged as a rationale for the inhibition of the Ca²⁺-depleted OEC at this stage, unable to advance to an S₃′ state. However, the absence of a corresponding split signal in native PS II preparations involving the high-spin g = 4.1 signal (see refs 53 and 79) and lack of a pathway that describes the (back-)conversion of the closed (S₂Y_Z^•) into the open (S₃) conformation leaves this question unresolved at this point.

Previous EXAFS studies^11,23,31 showed that the formation of the S₃ state is accompanied by structural changes of the cluster, which involves an elongation of Mn–Mn distances, specifically one of the shorter Mn–Mn interactions of ∼2.8 Å, albeit not of the Mn–Ca vectors. Since such a displacement would likely not be impeded in the less rigid Ca²⁺-depleted OEC, the role of the Ca²⁺ ion in this transition could instead be the ordering of the nearby water matrix/H-bonding network. Such a possibility includes (i) an involvement in coordination and transfer of the second (late binding) substrate to Mn_A4, as suggested in refs 13 and 48 or Mn_D1, see refs 80–82, by (a) direct substrate binding [e.g., W3 (Figure 1)], (b) positioning/directing a substrate coming from the bulk matrix, and/or (c) facilitating deprotonation to yield OH⁻.^13,51,52 The Lewis acidity of Ca²⁺ and the correlation of proposed H⁺ release and Ca²⁺-dependence of the individual S-state transitions suggest that the H⁺ could indeed come from a Ca²⁺-bound H₂O.⁸³ (ii) It could act as a gatekeeper, restricting access of water to the Mn open coordination site at too early a stage of the cycle,^26,49 timing its binding exactly after reduction of Y_Z^•. The similar spectral changes during advancement to the S₃ and the S₂′Y_Z^• state [i.e., the Mn–Mn distance elongations (Table 3 and refs 11 and 23)] could be a result of water binding in S₂′Y_Z^•, arrested in an energetically stabilized configuration. On the other hand, the resulting coordination of the water ligand to Mn^III instead of Mn^IV makes this seem less likely. (iii) A crucial role of the Ca²⁺ ion in proton-coupled electron transfer from the OEC to the Y_Z^• residue is strongly supported by recent studies.^52,53,74,76 (iv) A direct adverse effect of Ca²⁺ depletion on the oxidation of Mn_D1 during the S₂′-to-S₃′ transition has recently been put forward in theoretical studies.^75,76 (v) Experimental^45,82,84 and computational⁸⁵ evidence that the first (early binding), slowly exchanging substrate is O5 suggests that the Lewis acidity of the Ca²⁺ ion may also be critical for providing the specific electropositive environment for this ligand discussed above. Withdrawing electron density from O5 could make it susceptible to a nucleophilic attack or to direct oxidation conferring radical character to it.

Understanding the function of the Ca²⁺ ion in photosynthetic water oxidation is highly important as it is linked to several critical steps in the mechanism of catalysis. This study shows that this role is not so much the provision of a structural framework of the tetranuclear Mn complex itself, at least in the S₁ and S₂ states, but rather to fulfill a distinct catalytic task, for example by organizing the functionally critical water environment. Further knowledge about the details of this process is necessary to fully understand the way nature converts solar energy into a chemical form and could help to find a way to make improved usage of the sun as our primary energy source.

Abbreviations

CW

continuous wave

EPR

electron paramagnetic resonance

EXAFS

extended X-ray absorption fine structure

IPE

inflection point energy

MLS

multiline EPR signal

OEC

oxygen-evolving complex

PS II

photosystem II

XANES

X-ray absorption near-edge structure

XAS

X-ray absorption spectroscopy

XFEL

X-ray free-electron laser

XRD

X-ray diffraction