Alternating electron and proton transfer steps in photosynthetic water oxidation

Abstract

Water oxidation by cyanobacteria, algae, and plants is pivotal in oxygenic photosynthesis, the process that powers life on Earth, and is the paradigm for engineering solar fuel–production systems. Each complete reaction cycle of photosynthetic water oxidation requires the removal of four electrons and four protons from the catalytic site, a manganese–calcium complex and its protein environment in photosystem II. In time-resolved photothermal beam deflection experiments, we monitored apparent volume changes of the photosystem II protein associated with charge creation by light-induced electron transfer (contraction) and charge-compensating proton relocation (expansion). Two previously invisible proton removal steps were detected, thereby filling two gaps in the basic reaction-cycle model of photosynthetic water oxidation. In the S₂ → S₃ transition of the classical S-state cycle, an intermediate is formed by deprotonation clearly before electron transfer to the oxidant (). The rate-determining elementary step (τ, approximately 30 µs at 20 °C) in the long-distance proton relocation toward the protein–water interface is characterized by a high activation energy (E_a = 0.46 ± 0.05 eV) and strong H/D kinetic isotope effect (approximately 6). The characteristics of a proton transfer step during the S₀ → S₁ transition are similar (τ, approximately 100 µs; E_a = 0.34 ± 0.08 eV; kinetic isotope effect, approximately 3); however, the proton removal from the Mn complex proceeds after electron transfer to . By discovery of the transient formation of two further intermediate states in the reaction cycle of photosynthetic water oxidation, a temporal sequence of strictly alternating removal of electrons and protons from the catalytic site is established.

In oxygenic photosynthesis, plants, algae, and cyanobacteria oxidize water at the manganese–calcium (Mn₄Ca) complex of photosystem II (PSII) (1–3). This process has shaped the atmosphere by massive O₂ formation (from water) and the biosphere by facilitating the large-scale production of primary biomass and energy-rich carbohydrates (4), but is still insufficiently understood. Improved insight into photosynthetic water oxidation could promote the development of biomimetic systems for direct production of solar fuels (3, 5–9).

In PSII, the absorption of a light quantum results in oxidation of a specific tyrosine (10), Y_Z [redox-active tyrosine residue (Tyr161) in the D1 subunit of PSII], which functions as the oxidant in the redox chemistry of water oxidation (9, 11, 12):

[1]

The actual catalyst facilitating the reaction described by Eq. 1 is the Mn₄Ca complex bound to the proteins of PSII (13–15) (Fig. 1A). Four electrons are removed sequentially from the Mn complex [that is, the Mn₄Ca(μ–O)_n core and its ligand environment] by electron transfer to , resulting in accumulation of four oxidizing equivalents before the onset of O─O bond formation and O₂ liberation, as described by Kok’s classical S-state cycle (16, 17) (Fig. 1B, inner circle of S-states). Four protons are removed by deprotonation of the Mn complex and relocation toward the aqueous phase of the thylakoid lumen (18–20).

Fig. 1.

Photosystem II (A) and reaction cycle of water oxidation (B). In A, crucial redox cofactors and dimensions of the PSII complex are shown (15). Red arrows connect redox cofactors of the ET chain, including the primary electron donor (P₆₈₀), the primary pheophytin acceptor (Phe), the primary (Q_A) and secondary (Q_B) quinone acceptors, and, at the electron donor side, a redox-active tyrosine (Y_Z) and the Mn complex. Water molecules resolved in the crystallographic model (Protein Data Bank entry 3ARC; ref. 15) are shown as red dots; the indicated distances illustrate relevant dimensions. In B, the classical Kok model (16) (inner circle, including states S₄ and S₄′; ref. 22) is extended to describe both oxidation of the Mn complex by ET to the Y_Z radical and proton removal from the Mn complex or its ligand environment by long-distance proton transfer. Coupling of the ET step to local proton shifts is not covered by the shown framework model. The subscripts indicate the number of oxidation equivalents accumulated at the Mn complex; the superscripts indicate the charge relative to the dark-stable S₁-state (+, positive; n, neutral). The proton release steps in the S₀ → S₁ and S₂ → S₃ transitions have not been tracked in time-resolved experiments before, but now these steps are detected in the PBD experiments; the indicated time constants result from the present study.

The location of the Mn complex at the interface between the membrane-intrinsic part of PSII and the extrinsic lumenal proteins (13–15) (Fig. 1A) implies long-distance proton relocation toward the aqueous phase (approximately 30 Å), occurring within tens or hundreds of microseconds along chains of water molecules and ionic residues (21). The interrelation between electron transfer and protonation dynamics (that is, the relocation of protons on various time and length scales) is functionally crucial (1, 22–25). Our study aims to identify the basic sequence of electron transfer (ET) and long-distance proton relocation in the water oxidation cycle.

The electron transfer from the Mn complex to (22, 26–29) and the proton release (i.e., the appearance of protons in the aqueous phase) (18–20), have been investigated extensively. It was found (inter alia) that the observable proton release often does not reflect the removal of a proton from the Mn complex because electrostatically induced deprotonation of residues at the lumenal periphery of the protein masks the protein-intrinsic proton removal (20, 30). Nonetheless, it was possible to determine the “intrinsic proton release pattern” (i.e., the number of protons removed in each of the classical S-state transitions from the Mn complex), which is: 1 H⁺ in S₀ → S₁, 0 H⁺ in S₁ → S₂, 1 H⁺ in S₂ → S₃, and 2 H⁺ in S₃ → S₀ (18, 19, 31). The appearance of protons in the aqueous bulk phase rapidly after formation and prior to its reduction by ET from the Mn complex has been established firmly (19, 20, 30), but likely does not reflect the deprotonation of a chemical group (e.g., a substrate water molecule) at the Mn complex (18, 19, 31). This implies that the time-resolved detection of proton release into the aqueous phase cannot be employed to decide whether the proton removal from the Mn complex precedes the ET to .

The temporal sequence of electron and proton removal steps after rapid formation therefore has remained largely obscure, with one notable exception: Today there is strong experimental support that, after formation of the S₃ state in the S₃ → S₀ transition, a proton is removed from the Mn complex before onset of the electron transfer to (22, 26, 32). This finding has led to an extension of the S-state cycle involving formation of a distinct S₄ state by deprotonation (inner circle of Fig. 1B) before S₄ formation by electron transfer (22, 17). Later, this reaction cycle was extended further to include each of the four protons (24, 33) (outer circle of Fig. 1B). However, the proposed sequence of events has remained hypothetical, in particular because the proton removal from the Mn complex in the and transitions could not be tracked in time-resolved experiments.

To detect proton removal from the Mn complex, we employ a photothermal beam deflection (PBD) experiment exploiting the high sensitivity of the PBD signal to density changes (34–38), which in the following are discussed in terms of apparent volume changes of the protein. Expansion and contraction of the PSII complex were monitored with microsecond resolution at a precision of about 2 Å³ per PSII. For comparison, the average volume of one water molecule in aqueous solution is approximately 30 Å³. In photoacoustic or photothermal measurements, it is usually found that a volume contraction results from the (light-induced) deposition of charges at the electron donor and/or acceptor, as has been shown for synthetic molecules (39, 40) and photosystems (41, 42). Also in PSII, the light-induced formation of the primary quinone acceptor () and of is associated with a volume contraction (37, 43). The decharging of the donor side by removal of a proton results in an expansion that reverts the contribution to the preceding contraction. This enables monitoring of proton removal from the Mn complex by measuring the concomitant volume expansion in the PBD experiment, thereby revealing the temporal sequence of electron and proton removal steps in the classical S₀ → S₁ and S₂ → S₃ transitions.

Results

For insight into the individual S-state transitions, PBD measurements were combined with the following laser-flash protocol (Fig. S1): Dark-adapted PSII membrane particles were excited by a sequence of n saturating ns-laser flashes (532 nm, 10 mJ cm^-2). Each flash populated predominantly a specific stable/semistable S-state of Kok’s classical reaction cycle (16, 17), namely S₁ (n = 0, dark-adapted PSII), S₂ (n = 1), S₃ (n = 2), S₀ (n = 3), and again S₁ (n = 4). Subsequently, a single subsaturating ns-flash (0.1 mJ cm^-2) was applied and the PBD signal induced thereby was recorded. The subsaturating flash initiated predominantly the following transitions: S₁ → S₂ (flash 1), S₂ → S₃ (flash 2), S₃ → S₀ (flash 3), and S₀ → S₁ (flash 4). The measured PBD signals were corrected for imperfect advancement in the S-state cycle (S-state mixing) using previously established procedures (26, 44) (Fig. S2). We note that all central conclusions of this work are independent of the details of the correction procedure. Corrected PBD transients for each of the four transitions between semistable S-states are shown in Fig. 2 and discussed below.

Fig. 2.

Flash-induced PBD signals and volume changes: S₁ → S₂ (A, A^′), S₂ → S₃ (B, B^′), (C, C^′), and S₀ → S₁ (D, D^′). Thin lines, experimental data; thick lines, simulations using a step-shaped function for the rapid jump caused by formation and single-exponential functions for the slower signal contributions. (Right) Schematic illustration of volume changes deduced from the analysis of the temperature dependence of the PBD signals (time constants for about 20 °C; see Fig. 3).

The instantaneous rise observed after each flash (at 25 °C) is attributable to the rapid light-induced processes that result in reduction of Q_A and oxidation of the tyrosine donor ( formation) (12). At 12 °C, the magnitude of the instantaneous rise was smaller than at 25 °C, and at 1 °C a decay was observed (Fig. 2B). This behavior results from the temperature-dependent thermal (ΔQ) and temperature-independent nonthermal (ΔV) contributions to the PBD signal (34–36) associated with the formation. Evaluation of the temperature dependence of the rapid phase in comparison to a calorimetric standard (Figs. S3–S5) yielded an apparent volume contraction (ΔV) by about -12 Å³, in reasonable agreement with previous estimates (37, 43). For a possible contribution to the PBD signals associated with interquinone electron transfer, see Figs. S6 and S7 and Table S1. [We note that no ΔV values presented herein were corrected for the effective quantum yield, Φ_eff, of the light-induced transition, which could be as low as 50% (SI Text). This means that the volume changes per PSII complex may be twice as large as the values documented herein.]

In the S₁ → S₂ transition (flash 1), the instantaneous signal (< 10 μs) was followed by an exponential decay with a time constant of about 100 μs at 20 °C (Figs. 2 and 3, and Table 1). The amplitude of the decay phase depended only weakly on the temperature, suggesting that it mostly originates from a volume change (ΔV) of the PSII sample; its negative amplitude (PBD-signal decrease) indicates a contraction. The weak temperature dependence of the signal magnitude suggests a small contribution to the signal from a positive ΔQ (heat release) (Table 1). The decay became only slightly slower at lower temperatures, and the Arrhenius plot of the rate constants (Fig. 3B) revealed a small activation energy (Table 2). From PBD transients measured in D₂O (Fig. 4), a minor H/D kinetic isotope effect (KIE) of 1.3 was determined. (The KIE is the ratio of time constants determined in D₂O or H₂O; KIE = τ_D/τ_H = k_H/k_D.) Time constant, activation energy, and KIE agree well with figures previously determined for the ET from the Mn complex to in the S₁ → S₂ transition (22, 26, 27) (Fig. 1B).

Fig. 3.

Temperature dependence of the PBD signals of the four resolved transitions. (A) Temperature dependence of the amplitudes as obtained by an exponential simulation (symbols, experimental data; lines obtained by a fit). The dotted lines show the 1σ error ranges of the fit curves. The bars represent the nonthermal part of the PBD signal (volume change ΔV) that corresponds to the PBD amplitude at -14 °C (T₀ = -14 ± 1 °C; SI Text). (B) Arrhenius plots of the rate constants (k = τ^-1, left y axis; time constants, τ, on right y axis). The symbols indicate the experimentally determined values; the lines are fit curves used for determination of the respective activation energy shown in Table 2.

Table 1.

Time constants (τ at 20 °C), volume changes (ΔV), and heat release (ΔQ) of four transitions resolved in PBD measurements

Flash no.	Transition	τ (μs)	ΔV*, † (A3)	ΔQ † (meV)
1		98 ± 3	-6.1 ± 0.4	160 ± 50
2		29 ± 2	+4.4 ± 1.1	190 ± 120
3	S3 → S0	1,960 ± 90	+15 ± 2.5	60 ± 250
4		94 ± 8	+3.5 ± 1.3	40 ± 140

The given parameters were determined from data shown in Fig. 3; error ranges correspond to uncertainty of the fit result at the 1σ level.

^*The value of ΔV denotes apparent volume changes calculated using SI Text, Eq. S1. Negative or positive signs correspond to contractions or expansions, respectively.

^†The figures given for ΔV and ΔQ were calculated for a single PSII complex after absorption of one light quantum (without correction for nonunity quantum yield).

Table 2.

Kinetic parameters of ET (e^-) and rate-determining proton transfer (PT, H⁺) during the reaction cycle of PSII water oxidation at approximately 20 °C

Kinetic parameters were mostly determined for plant PSII (membrane particles).

^*Activation energies (E_a) were derived from the temperature dependence of rate constants (Fig. 3B).

^†PBD results, this work.

^‡Ref. 26.

^§Ref. 22.

^¶Ref. 27.

^∥Ref. 28.

^**Ref. 66; for thermophilic cyanobacteria, ref. 53.

^††Ref. 46.

^‡‡Ref. 29.

Fig. 4.

Comparison of PBD signals on the four S-transitions measured for PSII membranes in H₂O (black) and D₂O (red) at 20 °C. Thin lines, experimental data; thick lines, simulations with single-exponential functions plus offset, except for the S₀ → S₁ transition, for which a double-exponential function plus offset was used (Fig. S7 and Table S2). The respective rate constant ratio is indicated (KIE = k_H2O/k_D2O). The PBD amplitudes differ because of different thermoelastic properties of H₂O and D₂O.

The large contraction associated with the ET from the Mn complex to is specifically observed in the S₁ → S₂ transition. For S₂ → S₃, such a contraction paralleling the ET step was not detectable (Fig. 2B and Fig. S7). In the S₀ → S₁ transition, a small contraction might be coupled to the ET step (Fig. S4). We note that the contraction associated with the ET in the S₁ → S₂ transition, which likely is reversed in the S₃ → S₀ transition (see below), could reflect an interesting new mode of coupling the ET step to nuclear rearrangements, possibly related to changes in the protein backbone conformation suggested by FTIR data (45).

In the S₂ → S₃ transition (flash 2), the instantaneous rise was followed by an exponentially rising phase with a time constant of about 30 μs (at 20 °C; Figs. 2 and 3). This rise was roughly 10 times faster than the ET from the Mn complex to (30 μs versus 300 μs; Table 2) and is thus assignable to a process that precedes the ET step. Its almost temperature-independent signal amplitude (Fig. 3A) indicates that the rising phase originates mostly from a volume expansion (Table 1). A striking feature of the rapid rise was its large KIE of close to 6 (Fig. 4 and Table 2), facilitating the assignment to a process involving proton movements. We emphasize that at all temperatures, and also in D₂O, the apparent volume expansion clearly preceded the ET step (Table 2). Thus, we conclude that in the classical S₂ → S₃ transition, a proton relocation precedes the ET from the Mn complex to .

In the transition S₃ → S₀ + O₂ (flash 3), a prominent millisecond rise (τ around 2 ms at 20 °C) that resulted mostly from a volume expansion by about 15 Å³ was visible (Figs. 2 and 3A, and Table 1). The moderate activation energy and small KIE of the millisecond phase (Figs. 3 and 4, and Table 2) were similar to the respective values for the ET step () and the concomitant dioxygen formation () (26, 27, 46, 47).

Using time-resolved X-ray spectroscopy and near-UV measurements to monitor the oxidation state of the Mn complex, it has been found that an apparent lag phase of an approximately 200-μs duration precedes the millisecond rise of O₂ formation (22, 26, 27, 32). A similar lag phase was not discernable in the PBD transients. Instead, a rising phase with a similar τ value to the previously observed lag phase could be detected by simulation of summed PBD transients (Fig. S5). However, this phase was not sufficiently well-resolved for analysis of its temperature dependence and quantitative determination of ΔQ and ΔV. Conservatively, we conclude that the PBD data are compatible with a volume expansion associated with proton release in the transition. Thus, we propose that the overall extent of the expansion in the S₃ → S₀ transition is explainable by three additive contributions—namely, proton removal from the Mn complex prior to the ET, reversal of the contractions associated with previous manganese oxidation (in the S₁ → S₂ transition), and removal of a second proton from the Mn complex after the ET (more quantitative considerations appear in SI Text).

In the S₀ → S₁ transition (flash 4), the initial signal increase was followed by an exponentially rising phase with a time constant of approximately 100 μs (at 20 °C). The amplitude of this signal rise was almost temperature-independent, indicating a volume expansion (Fig. 3A), similar to the signal rise in the transition. Also, further parameters of this phase in the S₀ → S₁ transition were similar—namely, its large activation energy, its large KIE (approximately 3), and the ΔV magnitude (Tables 1 and 2, and Table S2). The 100-μs rise in the PBD signal was slower than the ET from the Mn complex to (approximately 40 μs in refs. 22 and 29) (Table 2), implying that the underlying process occurred after the ET step—that is, after the transition (Fig. 1B).

We note that vastly different values have been reported for the ET rate constant of the S₀ → S₁ transition, ranging from about 40 μs to 300 μs (at approximately 20 °C) (27–29, 32, 48). For the same PSII samples used herein, we have previously determined a value of approximately 40 μs (22), suggesting that the ET step in the S₀ → S₁ transition is faster than the proton removal step already at room temperature, and even more so at lower temperatures and in D₂O (27, 28). In conclusion, the volume expansion observed in the classical S₀ → S₁ transition results from a proton relocation (the transition in Fig. 1B) that takes place after manganese oxidation by ET to .

Discussion

In the course of the classical S₂ → S₃ transition, a volume expansion precedes the ET step and can be assigned to the removal of a proton from the Mn complex in the transition of the extended S-state cycle shown in Fig. 1B. This is a central finding of our investigation because it directly supports the proton-first ET in the S₂ → S₃ transition and establishes a close analogy to the first steps () in the S₃ → S₀ + O₂ transition (22). This finding implies that a previously undetected intermediate state () is transiently formed by deprotonation. Why this proton removal step is associated with a volume expansion is explained straightforwardly: The removal of a positive charge (H⁺) from the donor side of PSII reverts the preceding volume contraction caused by charging of the donor side by formation. Detection of this proton removal step (t_1/2 of only approximately 20 μs at pH 6.2) by analysis of electrochromic absorption changes of the PSII pigments may be feasible, but has not been achieved yet (28, 29, 32), presumably because of technical limitations.

In the course of the classical S₀ → S₁ transition, a volume expansion occurs after the ET step and is assigned to proton removal from the Mn complex in the transition. In analogy to the transition, this expansion is explained by removal of a positive charge (H⁺) from the donor side of PSII, thereby reversing the contraction caused by formation. As opposed to the transition, the deprotonation in the course of the classical S₀ → S₁ transition proceeds after oxidation of the Mn complex (after formation). This result directly supports the extended S-state cycle scheme (Fig. 1B) by revealing transient formation of the intermediate.

The kinetic parameters determined for proton removal from the Mn complex relate to the slowest step in the sequence of all elementary proton transfer steps of the long-distance relocation of a proton from the Mn complex toward the aqueous phase. Presently, the site and physicochemical nature of the rate-determining steps are unknown. We determined similarly large activation energies and KIE values for proton removal from the and intermediates, pointing to the formation of a transition state of relatively high energy and proton tunneling across a sizeable energetic barrier; a more elaborate analysis of the thermodynamic and kinetic parameters may provide deeper insight (49, 50). Interestingly, activation energy and KIE for proton removal from the intermediate are smaller, suggesting that site and/or nature of the rate-determining proton transfer may differ. Identification of pathways and modes of proton transfer within PSII could be approached experimentally (e.g., by time-resolved infrared spectroscopy) (51, 52).

In extension of the classical S-state cycle model, we have suggested a basic reaction-cycle model (24, 33) that involves nine intermediate states and describes the temporal sequence of light-induced Y_Z oxidation, electron transfer to , proton removal from the Mn complex, and O₂ formation (Fig. 1B). However, with respect to the proton removal steps, the supporting evidence remained largely circumstantial. Only the transition (or S₃ → S₄; ref. 22) had become detectable in time-resolved experiments (22, 26, 27, 32, 53, 54). Now, we have followed the charge-compensating proton removal in the classical S₀ → S₁ and S₂ → S₃ transitions and have determined the kinetic parameters of these essential reaction steps (see Table 2).

On these grounds and based on earlier investigations (reviewed in refs. 2 and 24), we propose the basic reactions sequence for the transitions from to the state outlined in the following (Figs. 1B and 5): (i) Starting in the most reduced semistable S-state, , absorption of a photon by PSII induces rapid Y_Z oxidation (< 1 μs) followed by ET from the Mn complex to (40 μs in Fig. 1B), resulting in formation. Mn oxidation in the transition lowers the pK of a Mn ligand, possibly a bridging hydroxide (44, 55, 56), to a value around 3.3 (57). The proton is removed from the Mn complex and relocated toward the lumen only after formation in the transition. (ii) The next absorbed photon induces oxidation followed by a Mn^III → Mn^IV oxidation in the transition. The oxidation of the Mn complex lowers the pK values of ligand groups, but not to an extent sufficient for deprotonation. (iii) Oxidation of the Mn complex without any charge-compensating chemical change raises its redox potential to a level that prohibits a second oxidation by (1, 2, 7). In the state, this redox-potential problem initially prevents oxidation of the Mn complex by . However, drives the removal of a proton from the Mn complex, resulting in formation of the state within about 30 μs. We propose that the proton is removed from the cluster of water molecules indicated in Fig. 5. Thereby created is a proton vacancy that is effectively delocalized within the water cluster (on the μs time scale), but likely resides mostly on the water molecule close to , which is coordinated to the Ca ion of the Mn complex. In the subsequent ET step (), Mn^III→IV oxidation of Mn1 (58–60) is directly coupled to proton transfer to the previously deprotonated water cluster. Such a concerted electron–proton transfer is in line with a comparably large H/D isotope effect for this ET step, and is well-suited to solve the redox-potential problem mentioned above. In the proton removal step preceding the ET, deprotonation of D1–Asp₆₁ can be excluded because the S₂ → S₃ transition is not severely affected in the Asp–Asn mutant (61, 62). A central role of CP43–Arg₃₅₇ (63) is unlikely because this residue is close to neither Y_Z nor Mn1. Thus, we consider deprotonation of the water cluster interconnecting Y_Z and the Mn complex to be the most plausible option, which is supported by the importance of the pK value of the water molecules coordinated to the Ca ion (64). (iv) In the transitions, the basic temporal sequence of proton removal and electron transfer is similar to the one described above for the transitions. However, the key players are others. As opposed to the S₂ → S₃ transition, the proton removal step likely involves D1–Asp₆₁ (62). Therefore, and in line with the clear differences in H/D isotope effect (KIE) and activation energy (Table 2), we propose that proton removal in the transition proceeds along a path that includes D1–Asp₆₁ (14, 21), whereas proton removal in the transition proceeds along another path starting close to D1–Tyr₁₆₁/His₁₉₀ (15, 65). In comparison to the ET of the transition, the transition may be coupled to more extensive chemical changes suitable to initiate O─O bond formation. Gaining insight into the presently merely hypothetical and intermediates represents a central challenge in future research on photosynthetic water oxidation.

Fig. 5.

Sequence of events in the classical S₂ → S₃ transition of photosynthetic water oxidation. The Mn₄CaO₅ cluster, the redox-active tyrosine (Tyr₁₆₁), and the key groups of the surrounding hydrogen-bonded network (15) are shown. All indicated amino acid residues are from the D1 subunit of PSII, with exception of CP43–Arg₃₅₇. (Water molecules, H_xO, are indicated as red spheres; putative H-bonds as broken lines that connect H-bond donor and acceptor. Of all the protons, only the phenolic proton is shown as a grey sphere.) The grey mesh outlines a water cluster that includes 4 H_xO in the first coordination sphere of manganese (Mn4), as well as the calcium (Ca) and three second-sphere water molecules. Within less than 100 ns after absorption of a photon and oxidation of the primary chlorophyll donor of PSII (P680), Tyr₁₆₁ (Y_Z) is oxidized by P680⁺ (“1st”). formation results in a rearrangement of the shown H-bonded network (completed within less than 1 µs), likely involving a shift of the phenolic proton to His₁₉₀ and lowering of pK values for deprotonation of the water molecules in the outlined cluster (grey mesh). A proton is removed from the Mn complex/Y_Z environment within about 30 µs, as evidenced by the PBD results presented herein, and a proton vacancy is supposedly created within the outlined water cluster (“2nd”). In the ET to (about 300 µs), Mn oxidation is directly coupled to a proton transfer step involving the previously created proton vacancy of the water cluster (concerted electron–proton transfer) (“3rd”).

In conclusion, the basic sequence of events in the reaction cycle of water oxidation has now been established for the six transitions leading from to . This sequence is characterized by the strictly alternating removal of electrons and protons from the Mn complex. Three intermediate states of the catalytic metal center (, , ) and its immediate ligand environment are transiently formed. On these grounds, strategies can be developed for characterization of the new reaction intermediates at the atomistic level. The extended S-state cycle model (Fig. 1B) may serve as a framework for the future design of experimental and theoretical investigations on photosynthetic water oxidation. It may also be worthwhile to scrutinize synthetic systems for catalysis of water oxidation within a conceptual framework that involves redox-potential leveling and local accumulation of four oxidizing equivalents by alternating electron and proton removal from the catalytic site.

Materials and Methods

PSII membrane particles prepared from spinach were resuspended in a buffered solution (80 μg chlorophyll per mL; pH/pD of 6.2), and 20 μM of 2,6-dichloro-p-benzoquinone (DCBQ) were added as artificial electron acceptor. The PBD experiments were carried out and analyzed as described previously (37, 38). In the PBD experiments presented herein, 100 mL of PSII suspension were kept in a dark reservoir on ice (gently stirred) and pumped first through a thermostated laboratory-built heat exchanger and then into the sample compartment of the thermostated flow-through cuvette (3-mm optical path). Then, the respective laser-flash protocol was applied (0–4 saturating ns-laser flashes plus a single subsaturating flash; 532 nm; see Fig. S1). The PBD signal induced by the nonsaturating laser flash was recorded. In the next pump cycle, the flow-through cuvette was filled again with a fresh sample of dark-adapted PSII. Extensive signal averaging was applied; about 4,000 measurements were needed for obtaining the data in Fig. 2 (SI Text).