|Home | About | Journals | Submit | Contact Us | Français|
Proton matrix ENDOR spectra were measured for Ca2+-depleted and Sr2+-substituted photosystem II (PSII) membrane samples from spinach and core complexes from Thermosynechococcus vulcanus in the S2 state. The ENDOR spectra obtained were similar for untreated PSII from T. vulcanus and spinach, as well as for Ca2+-containing and Sr2+-substituted PSII, indicating that the proton arrangements around the manganese cluster in cyanobacterial and higher plant PSII and Ca2+-containing and Sr2+-substituted PSII are similar in the S2 state, in agreement with the similarity of the crystal structure of both Ca2+-containing and Sr2+-substituted PSII in the S1 state. Nevertheless, slightly different hyperfine separations were found between Ca2+-containing and Sr2+-substituted PSII because of modifications of the water protons ligating to the Sr2+ ion. Importantly, Ca2+ depletion caused the loss of ENDOR signals with a 1.36-MHz separation because of the loss of the water proton W4 connecting Ca2+ and YZ directly. With respect to the crystal structure and the functions of Ca2+ in oxygen evolution, it was concluded that the roles of Ca2+ and Sr2+ involve the maintenance of the hydrogen bond network near the Ca2+ site and electron transfer pathway to the manganese cluster.
Photosynthetic oxygen evolution is the key reaction to support oxygenic life on the earth. The reaction is performed in photosystem II (PSII),2 a sophisticated protein complex that converts light energy into chemical energy via coupled electron and proton transfer reactions through redox-active molecules. When a photon is absorbed by PSII antenna pigments, excitation energy is transferred to P680, the reaction center of PSII consisting of four chlorophylls and two pheophytins, where charge separation occurs. The photoexcited electron is immediately transferred to the accepter side, QB, via pheophytin and QA, whereas the P680+ generated by this charge separation oxidizes a manganese cluster via a redox-active tyrosine residue YZ. The manganese cluster is the core device for water oxidation in PSII, which adopts five different redox states, termed Sn (where n = 0–4). When the manganese cluster is oxidized, the S state is advanced to the next state (namely, from Sn to Sn+1). One or two protons are released in each S state transition except for the transition from S1 to S2. S1 is the most stable state in the dark at room temperature, whereas S4 is a transit state that is difficult to detect experimentally and returns to the S0 state immediately upon excitation of the S3 state and evolution of the molecular oxygen (see reviews in Refs. 1,–3).
Recently, the structure of the S1 state manganese cluster was solved by x-ray crystal structure analysis with 1.9 Å resolution (4), which revealed that the manganese cluster has a distorted cubic structure formed by three manganeses, four oxygens, and one calcium, with another manganese ion (denoted as Mn4) and an oxygen atom connected at the outside of the cubane link the backrest of a chair; thus, the overall structure of the manganese cluster resembled a “distorted chair” form. In this cluster, two water molecules, W1 and W2, are ligated to the outside (dangler) manganese (Mn4), and other two water molecules, W3 and W4, are ligated to the Ca2+ ion (see Fig. 1). Although this structure was obtained at an unprecedentedly high resolution, the manganese cluster may have suffered slight radiation damage from the utilization of strong and continuous synchrotron x-rays. Recently, Suga et al. (5) analyzed the structure of the S1 state manganese cluster using a femtosecond x-ray free electron laser (XFEL), which provided a radiation damage-free structure. The XFEL structure showed that the manganese-manganese distances are shortened by 0.1–0.2 Å, and based on the average distances between the manganese ions and their oxo-bridges and ligands, the manganese valences were assigned as Mn1(III), Mn2(IV), Mn3(IV), and Mn4(III) in the S1 state. The amino acid residues ligated to the manganese cluster were identified unambiguously by the x-ray structures, among which, D1-His332, D1-Asp342, and D1-Glu189 were ligated to Mn1; D1-Asp342, D1-Ala344, and CP43-Glu354 were ligated to Mn2; D1-Glu333 and CP43-Glu354 were ligated to Mn3; D1-Asp170 and D1-Glu333 were ligated to Mn4; and D1-Ala344 and D1-Asp170 were ligated to the Ca2+ ion (4, 5). However, the x-ray crystallographic analysis did not identify the location of protons, and the proton network structure and protonation states of μ-oxo-bridges and water molecules ligated to the manganese cluster are still unclear. The information of proton locations is essential for elucidating the precise structure of the manganese cluster and the mechanism of water splitting, because the proton and/or hydrogen coordinates seriously influence the interpretation of the valences of manganese and manganese-manganese distances, and the site of proton release must be determined before a complete mechanism of water splitting can be elucidated.
The only cation other than manganese ions in the manganese cluster is Ca2+, and its exact function in water splitting is still largely unknown (see reviews in Refs. 6,–8). Ca2+ in higher plant PSII can be depleted by biochemical treatments, resulting in the inhibition of oxygen evolution (9,–19), which can be recovered by the reconstitution of Ca2+ or Sr2+ (10, 14, 16, 20, 21). However, removal of Ca2+ and chemical Sr2+ substitution with the cyanobacterial PSII have been unsuccessful so far, and Sr2+ substitution in cyanobacterial PSII has been performed by cultivating cells in a medium containing Sr2+ instead of Ca2+ (20). Although Sr2+ substitution does not affect the turnover numbers of the catalyst, it does lower its efficiency (TOF), leading to a lower activity (20). Using this method, Koua et al. (21) analyzed the structure of Sr2+-substituted PSII at 2.1 Å resolution.
Extensive spectroscopic studies have been performed on the Ca2+-depleted and Sr2+-substituted manganese cluster to reveal the role of the Ca2+ ion in the oxygen-evolving reaction. Extended x-ray absorption fine structure spectroscopy showed that the distances between manganese ion pairs were not changed by Ca2+ depletion (22). X-ray crystal structure analysis showed that the overall structure of the manganese cluster and the position of the manganese ions were not changed in the Sr2+-substituted PSII, whereas the Sr2+ ion was located toward the outside of the cubane relative to the Ca2+ ion (21). It is notable that the distance between W3 and Ca2+/Sr2+ was elongated by the Sr2+ substitution, whereas the distance between W4 and Sr2+ was similar to that between W4 and Ca2+ (21).
EPR spectroscopy has been used to characterize the manganese cluster (23,–26) extensively, and a typical, S2 state multiline signal centered at g = 2 has been well characterized. The multiline signal showed 19–21 peaks in a range over ~180 mT width, and its properties have been studied by several advanced EPR techniques such as 55Mn-ENDOR (27), ESEEM (28, 29), and pulsed ELDOR (30). The ENDOR results indicated that the S2 state is composed of one Mn(III) and three Mn(IV), where the Mn(III) ion is assigned to the Mn1 ion in the crystal structure (Fig. 1) (28,–30). This assignment is widely accepted at present and is consistent with the oxidation states of the manganese ions in the S1 state assigned by the recent XFEL structure (5), although a combination of three Mn(III) and one Mn(IV) (low oxidation scenario) in the S2 state has also been proposed based on the results of photoactivation (31) and theoretical calculations (32).
Ca2+ depletion has been performed by several chemical treatments such as NaCl/EDTA (9, 18), NaCl/EGTA (15), and citrate (low pH) treatments (10). The S2 state multiline signal in these treatments showed a larger number of lines than in untreated PSII, which is designated as the “modified multiline signal” and is thermally stable at room temperature (11, 12, 16,–18). Accesses to the manganese cluster by some molecules, such as EDTA, EGTA, MES, and pyrophosphate, have been discussed (12, 16, 18, 33, 34). FTIR results showed that chelators interact with carboxylate groups in the Ca2+-depleted PSII (35). However, the dark stability of the multiline signal is not ascribed to the effect of the chelators (19).
Although small magnetic interactions are concealed in the broad EPR line width, information concerning the positions and numbers of protons surrounding the manganese cluster is included in the EPR signal. There are different techniques to extract information regarding the position and number of the nuclei, such as pulsed ENDOR, ESEEM, and ELDOR. Previous studies using Q-band pulsed 55Mn ENDOR spectra indicated that 55Mn hyperfine couplings of untreated PSII and Ca2+-depleted (citrate-treated) PSII are similar (36), even though the latter was in the “modified multiline” state. The spectral resolution of CW-ENDOR is generally higher than pulsed ENDOR/ELDOR; however, the measurement conditions are restricted because of the dependence of the relaxation properties of the electron and nuclei spins (37). ESEEM (HYSCORE) is an alternative technique to obtain the information regarding the nucleus; however, the treatments of blind spots and analysis are more complicated, and therefore careful treatments are required. These techniques should be used to complement each other to gain detailed information on the nuclei surrounding the manganese cluster.
We detected six pairs of proton hyperfine signals in the untreated PSII membranes by the proton matrix ENDOR (37, 38), which is similar to the number of protons detected by HYSCORE (39). Curiously, only a small number of the signals was detected in highly resolved CW-ENDOR spectra within ~4 MHz hyperfine couplings, i.e. approximately within 6 Å (37, 38), despite more than 30 ENDOR detectable protons that could be inferred from the crystallographic data (4, 5). This may be ascribed to line broadening of the movable protons, and CW-ENDOR may have detected relatively immovable protons relative to the manganese coordinates. These proton signals, matrix ENDOR, are mainly composed of magnetic dipolar interactions with protons and electrons. Therefore, it is possible to identify the location of the protons from the electrons. By using a point dipole approximation, the locations of the protons are estimated to be 2.7–6.0 Å away from the manganese cluster, corresponding to a hyperfine separation of less than 4 MHz. However, the electrons are largely delocalized on the manganese cluster, and therefore the spin density distribution of individual manganese ions should be used. We previously obtained spin density distribution over the manganese ions using pulsed ELDOR measurements (30). The spin density distribution on the manganese cluster has also been obtained by QM/MM (quantum mechanics/molecular mechanics) calculations (40, 41). These studies showed that the largest spin density is ~2 and distributed on Mn1, and the spin densities for other manganese are ~1, which are in agreement with those obtained from the ESEEM measurements (28). On the other hand, the sign of the spin densities was inconsistent between the pulsed ELDOR results (30) and theoretical studies, because the former showed a spin topology of αβαβ, whereas the latter showed a αββα topology. This inconsistency resulted in slightly different hyperfine parameters for the protons (27, 39, 42). We have assigned a W2 proton to the 4 MHz hyperfine couplings, W1 and W3 protons to 2–3 MHz hyperfine couplings, and W4 protons to 0.7, 1.36 MHz hyperfine couplings, respectively (37), whereas the DFT models assigned the W1 and W2 protons to the 4 MHz hyperfine separations in the presence of isotropic hyperfine couplings, and W3 and W4 protons to 1–2 MHz hyperfine separations (42). In addition, Milikisiyants et al. (39) assigned the hyperfine signals with 4-MHz separation to the His332 proton because it has a very small isotropic hyperfine parameter determined by HYSCORE.
In the present study, we applied the proton matrix ENDOR to the Ca2+-depleted and Sr2+-substituted manganese cluster in the S2 state to investigate the relationship between Ca2+ and water molecules near the manganese cluster. Our results showed that Sr2+ substitution did not affect the ENDOR spectra significantly; however, a pair of peaks was lost upon Ca2+ depletion; these results were discussed in terms of the crystal structure and the role of Ca2+ in oxygen evolution.
PSII-enriched membranes were isolated from market spinach as described previously (43) with slight modifications (9). The isolated membranes were suspended in a medium containing 0.4 m sucrose, 20 mm NaCl, 20 mm Mes/NaOH (pH 6.5) and stored in liquid N2 until use. All of the following procedures were carried out under dark or dim green light.
Before illumination, PSII membranes were dark-adapted for 2–3 h after preillumination to ensure the maximum population in the S1 state. For Ca2+ depletion, the membranes were suspended in a medium containing 0.4 m sucrose, 20 mm NaCl, 10 mm citric acid/NaOH (pH 3.0) at 4 °C for 5 min (10), followed by the addition of a medium containing 0.4 m sucrose, 20 mm NaCl, 0.5 m Mops/NaOH (pH 7.5) and incubation at 0 °C for 10 min. The resulting membranes were washed and resuspended in a medium containing 0.4 m sucrose, 20 mm NaCl, 0.5 mm EDTA-2Na, 20 mm Mes/NaOH (pH 6.5). For Ca2+ reconstitution, the Ca2+-depleted PSII membrane samples were washed with a medium containing 0.4 m sucrose, 20 mm NaCl, 20 mm Mes/NaOH (pH 6.5) and then suspended in a medium containing 0.4 m sucrose, 20 mm NaCl, 20 mm Mes/NaOH (pH 6.5), 20 mm CaCl2 for 10 min. Finally, the Ca2+-reconstituted samples were washed with a medium containing 0.4 m sucrose, 20 mm NaCl, 0.5 mm EDTA-2Na, 20 mm Mes/NaOH (pH 6.5).
Cyanobacterial PSII dimeric core complexes were isolated from a thermophilic cyanobacterium Thermosynechococcus vulcanus as described previously (44, 45). Sr2+-substituted cyanobacterial PSII was prepared as described previously (21). The samples were suspended in a buffer containing 5% glycerol, 20 mm NaCl, 3 mm CaCl2 (for native PSII) or 3 mm SrCl2 (for Sr2+-substituted PSII), and 20 mm Mes/NaOH (pH 6.1).
Illumination protocols are as followings. To form the S2 state, untreated, Sr2+-substituted, and Ca2+-reconstituted PSII samples at a concentration of over 7 mg Chl/ml (T. vulcanus) and 20 mg Chl/ml (spinach) were illuminated by white light for 5 min at 200 K in an ethanol bath. To form the S2 state of the Ca2+-depleted PSII, white light illumination was applied for 1 min at 0 °C and then incubated for 70 min under dark at 0 °C. EPR and ENDOR measurements were performed using a Bruker ESP 300E ESR spectrometer at 5–6 K with a gas flow temperature control system (CF935; Oxford Instruments, Oxford, UK) (38).
Fig. 2 shows the S2 minus S1 difference EPR spectra of untreated (Ca2+-containing) PSII (Fig. 2A), Ca2+-depleted PSII (Fig. 2B), and Ca2+-reconstituted PSII (Fig. 2C) from spinach and untreated PSII (Fig. 2D) and Sr2+-substituted PSII (Fig. 2E) from T. vulcanus. The S2 multiline signal centered at g = 2 was clearly observed in all of the four samples, which has been attributed to the S2 state manganese cluster with a net spin of S = 1/2, indicating that our illumination protocol was effective in generating the S2 state in all the samples examined. The multiline signals from spinach (Fig. 2A) and T. vulcanus (Fig. 2D) PSII were similar, whereas that from the Ca2+-depleted spinach PSII (Fig. 2B) had more peaks than that from the usual Ca2+-containing PSII, in agreement with that reported previously (46). The Sr2+-substituted PSII from T. vulcanus also exhibited a spectrum (Fig. 2E) similar to that of the Ca2+-containing PSII.
Fig. 3 shows the ENDOR spectra from the four samples corresponding to those used in Fig. 2. The peaks were labeled aa′–ff′ like those used in the previous reports (37, 38, 47). Six pairs of peaks were detected in the T. vulcanus PSII (Fig. 3D), which were the same as those observed in the untreated spinach PSII (Fig. 3A). These results indicated that the location of protons in the S2 state manganese cluster was the same between spinach and T. vulcanus PSII, consistent with the fact that the amino acid residues around the manganese cluster were well conserved from T. vulcanus to spinach. The proton matrix ENDOR spectra of the Sr2+-substituted PSII were also very similar with those from the Ca2+-containing manganese cluster, indicating that the structure of the manganese cluster and their surrounding environment are similar in the S2 state. The crystal structure analysis has shown that the structures of Ca2+- and Sr2+-containing manganese cluster are very similar in the S1 state (4, 21). Although the crystal structure in the S2 state has not been reported yet, our results indicated that the structure of Ca2+-containing PSII in the S2 state is well conserved in the Sr2+-substituted PSII.
Table 1 shows peak separations of each ENDOR signal. It is notable that although the dd′ peaks with a separation of 1.36–1.37 MHz were detected in the Ca2+-containing PSII from spinach and T. vulcanus and the Sr2+-containing PSII from T. vulcanus, they disappeared in the Ca2+-depleted PSII from spinach. The dd′ peaks were recovered by the Ca2+ reconstitution. These results indicated that protons that contributed to the dd′ peaks were strongly correlated with Ca2+. The separation for bb′ peaks was 0.70 MHz for untreated and Sr2+-substituted PSII, whereas it was 0.54 MHz for the Ca2+-depleted PSII, indicating that the bb′ peak separation was affected by Ca2+ depletion but not by Sr2+ substitution. The cc′ peaks had 1.06 MHz hyperfine splitting in both spinach and T. vulcanus PSII and were not affected by Ca2+ depletion or Sr2+ substitution. Peaks e1e1′ and e2e2′ were detected in all of the samples, and the peak e1e1′ was slightly shifted to a narrower separation by Sr2+ substitution or Ca2+ depletion. Peak e2e2′ appeared as a small shoulder and was slightly modified by Ca2+ depletion or Sr2+ substitution. Peak ff′ with the largest hyperfine splitting was detected in all the samples, but the peak separations were slightly different: 4.00 and 4.10 MHz in untreated spinach and T. vulcanus PSII, respectively, whereas the value was 4.20 MHz in Ca2+-depleted PSII and 4.02 MHz in the Sr2+-substituted PSII.
The ENDOR spectrum of the Ca2+-reconstituted PSII (Fig. 3C) is similar to the native ENDOR spectra of untreated PSII (Fig. 3A). The result indicates that Ca2+ reconstitution recovered the hydrogen bond network structure surrounding the manganese cluster.
The similarity of the four ENDOR spectra among the Ca2+-containing and depleted PSII of spinach and the Ca2+-containing and Sr2+-replaced PSII of T. vulcanus indicates similar geometry and electronic structure among these samples. This is consistent with the results of extended x-ray absorption fine structure and 55Mn-ENDOR studies, which have suggested similar geometry and electronic structure among Ca2+-containing and depleted or Sr2+-substituted PSII (22, 27, 36).
Previously we assigned the dd′ peaks to the proton of a water molecule W4 connecting Ca2+ and YZ based on theoretical studies (37) and the S1 state structure and spin distribution on manganese ions (30). The present ENDOR result is consistent with this assignment, because the W3 and W4 water molecules are ligated to Ca2+, and the loss of the signal upon Ca2+ depletion may be caused by changes in one of these water molecules. Two possibilities may be considered for the loss of the dd′ peak: one is the loss of the hydrogen bond between W4-Tyr161 by the Ca2+ depletion. Based on the QM/MM approach, Saito and Ishikita (48) have proposed a model that the W4 protons are connected to Tyr161 and Gln165 in the Ca2+-containing PSII, and the hydrogen-bonding structure near the manganese cluster is rearranged leading to connections of the W4 protons to W3 and Ala344 in the Ca2+-depleted PSII. The other possibility is the increase in W4 proton mobility. The CW-ENDOR signal would be undetectable because of line broadening, if the distance distributions relative to the manganese coordinates are increased to a level estimated to be more than 0.1 Å by using simple point dipole approximation (37).
The e1e1′ and e2e2′ peaks, arising from exchangeable protons, were assigned to W1 and W3 protons in the previous report (37). The peak position of e1e1′ of the Sr2+-containing manganese cluster of the T. vulcanus PSII was slightly shifted to an inside region compared with that of the Ca2+-containing manganese cluster. However, the ee′ peaks did not disappear upon Ca2+ depletion despite the disappearance of the dd′ peaks. Three possibilities may be considered for this: (i) W3 position is not changed by Ca2+ depletion; (ii) the ee′ peaks in the Ca2+-depleted PSII arise only from the W1 proton and the disappearance of the W3 proton signal is hidden by the W1 proton signal; and (iii) the mobility of W3 is relatively large, making it undetectable by ENDOR in all of the conditions. The possibly large mobility of W3 has been shown by the x-ray crystal structure analysis of Sr2+-substituted PSII where the position of W3 was changed significantly and the bond distance between W3 and Sr2+ was elongated by 0.2–0.3 Å, while leaving the position of W4 unaffected. These results are consistent with the first two possibilities; however, the new XFEL structure of native PSII showed that the position of W3 is similar to the Sr2+-substituted PSII analyzed by XRD (4, 5, 21). Although this also suggests the relatively high mobility of W3 than W4 that led to a larger deviation of the position of W3 in different structures, crystal structure analysis of the Sr2+-substituted PSII using XEFL is required to distinguish the exact differences. On the other hand, QM/MM results concluded that both W3 and W4 are changed by Ca2+ depletion, which supports the second and third possibilities. At present, it is difficult to distinguish the three possibilities experimentally, and further experiments are required to locate the position of the proton(s) affected by the Ca2+ depletion.
Our results showed that Ca2+ depletion resulted in the breakage of the hydrogen-bonding structure around Ca2+. The dark stable S2 state can be formed in Ca2+-depleted PSII (9,–20), and FTIR results suggested that the Ca2+ chelator interacts with the carboxylate group(s) near the manganese cluster (35). It is therefore possible that a molecule such as a chelator gained access to the carboxylate group(s) through the breaking space upon Ca2+ depletion (12, 16, 18, 33, 34). However, no new ENDOR signals were detected in the Ca2+-depleted PSII, suggesting that such an extra molecule is relatively far from the manganese atoms or loosely bound to the manganese cluster.
So far three roles for the Ca2+ in OEC have been proposed, namely, control of the electronic structure (case I), the redox potential (case II), and proton transfer from substrate water (case III). The present results showed that the electronic structure was not significantly different between Ca2+-containing and Ca2+-depleted PSII, consistent with the 55Mn ENDOR results reported previously (36). Therefore, Ca2+ is hardly involved in the control of electronic structure (case I) and redox potential (case II), which is supported by the observation of the multiline signals in the Ca2+-depleted PSII (19, 35) and recent QM/MM calculation results (49). The present results indicate that control of proton transfer (case III) is one of the essential roles of Ca2+ in OEC. Based on the assignment of the dd′ peaks to the W4 proton where A = 0.70 MHz and A‖ = 1.36 MHz (37), the disappearance of the dd′ peaks in the Ca2+-depleted PSII suggested that the hydrogen bond network connecting Ca2+ and YZ was lost. It has been proposed that the His190 and Asn298 residues and surrounding water molecules are a part of the hydrogen-bonding network connecting Ca2+ to the outside of the PSII protein matrix, which is one of the candidates for the proton release pathways (Fig. 4) (4, 50,–52). The Ca2+-depleted manganese cluster cannot undergo the transition from S2 to normal S3, during which one proton must be released. The inhibition of oxygen evolution by Ca2+ depletion can therefore be ascribed to the inhibition in the proton release during the S2-S3 transition because of the disruption of the hydrogen-bonding network connecting Ca2+ to the YZ site. These results are consistent with recent theoretical calculations (48, 53). Note that Ca2+ can also play another important role in electron transfer (case IV), because the electron transfer rate in Ca2+-depleted PSII became very slow in the order of seconds (54, 55). Because electron transfer is mediated through bonds between the molecular groups participating in it (56), a close connection between the molecular groups is essential. In Ca2+-containing PSII, the closest pathway from the manganese cluster toward YZ is YZO-W4-Ca-O-Mn (~7 Å). Assuming that only Ca2+ and W4 were lost from the S1 structure upon Ca2+ depletion, electron transfers involve a longer distance, resulting in a slower transfer rate. Another role of Ca2+ would therefore be to mediate the electron transfer effectively from the manganese cluster to YZ through W4. The back reaction would also be suppressed by the lack of the electron transfer path, consistent with the results that spin properties of the manganese cluster in Ca2+-depleted PSII are very similar to those in normal PSII (36), and the S2 state is remarkably dark stable in Ca2+-depleted PSII (19). Recently, Yang et al. (53) showed that the highest occupied molecular orbital contribution on Mn1 in the S2 state decreased with Ca2+ depletion, where Mn1 is expected to be oxidized in the transition from S2 to S3 state. The change would contribute to lowering the electron transfer rate, concertedly with the elongation of the electron transfer pathway from the manganese cluster. We have previously detected a higher spin state beyond the S2 state in Ca2+-depleted PSII (57). The spin state was observed after longer illumination conditions, which indicates that manganese may be oxidized beyond the S2 state with a low efficiency, supporting the calculated results (53).
Fig. 5 shows that there are slight differences between the crystal structure of Ca2+-PSII and Sr2+-PSII, among which, a water molecule W543 close to Mn1 was not found in Sr2+-substituted PSII, indicating that W543 has been lost upon Sr2+ substitution and therefore has a larger mobility than other water molecules. Mobile water protons would be undetectable in our ENDOR measurements because protons detectable by the proton matrix ENDOR must have a very small mobility (less than 0.1 Å) (37). In our previous report, we excluded the possibility that the W543 proton is the candidate for the ff′ peaks, which is consistent with the present ENDOR results showing that the ff′ peaks were not changed in the Sr2+-substituted PSII.
In the structure of the Sr2+-substituted PSII, the bond length of Sr-O(W4) is 2.3 Å, which is shorter than the Sr-W3 distance. Terrett et al. (58) proposed that the W4 water is dehydrogenated, and the rest of the hydrogen-bond is connected with YZ or Gln165. The present ENDOR results showed that the W4 proton is connected with YZ, even if W4 is dehydrogenated. On the other hand, Rapatskiy et al. (42) have reported different hyperfine couplings of the protons based on DFT calculations, where the W3 proton might be assigned to dd′ peaks, and W4 might be undetected in our ENDOR spectra. The protons in the W3 and W4 molecules are in close proximity to each other, and the location of the protons in the DFT calculations is strongly dependent on the structural models employed (42, 48, 59, 60).
Lakshmi and co-workers (39, 61) reported HYSCORE results of the S2 state multiline signal, where similar proton signals to those of the proton matrix ENDOR were detected. However, the assignments of these signals are significantly different from those in the ENDOR assignments (37, 39, 61). The causes underlying these experimental differences and the assignments of these signals are unknown. Table 2 shows the comparison of detected signals and these assignments, where ENDOR peaks were correlated with the HYSCORE parameters. There are two main experimental differences: (i) the dd′ ENDOR peaks that disappeared with Ca2+-depleted treatment were not detected in HYSCORE, and (ii) other ENDOR peaks in Ca2+/Sr2+ replacements did not disappear, whereas some peaks in HYSORE disappeared with the Ca2+/Sr2+ substitution (39, 61). These differences could have been caused by an overlapping in the unresolved spectrum. Fig. 3 (D and E) show that Sr2+ substitution modifies the e2e2′ peak shape and decreases the intensity of the bb′ peaks, corresponding to HI and HIII peaks in HYSCORE, which have been previously assigned to the overlapping of W3 and W4 peaks. Lakshmi and co-workers assigned these proton signals to water ligands of manganese ions (either W1 or W2) based on large isotropic hyperfine parameters. On the other hand, we assigned the ff′ ENDOR peaks to the W2 proton, whereas the corresponding HYSCORE signals have been assigned to the ring protons of His332 close to Mn1. Note that the ring protons of His332 are not exchangeable, whereas the ff′ peaks have been confirmed as exchangeable protons (37, 38, 47). We have previously assigned the protons in the His332 side chain to the bb′ and dd′ peaks, because the orientation dependence of these nonexchangeable protons showed that they are directed to the perpendicular axis to the membrane normal from manganese cluster (37). Our previous peak assignments are based on the pure dipole interaction-derived structural information, which is mostly consistent with the DFT calculations (42), except for the difference caused by the spin projections on manganese atoms. On the other hand, it does not include isotropic hyperfine constants. Anisotropic hyperfine is related to the distance, whereas isotropic hyperfine is proportional to the spin density on nucleus. It is important to evaluate isotropic parameters. However, theoretical treatment is difficult for interpretation. Actually, isotropic hyperfine of a proton is not necessarily related directly to the metal ligation (62). Although anisotropic hyperfine parameters are formally derived in the HYSCORE experiments, the evaluation would be difficult under the overlapping of similar signals such as those in manganese cluster.
In summary, we have reported the first ENDOR spectra arising from the S2 state manganese cluster of Ca2+- and Sr2+-containing PSII from T. vulcanus. No significant differences were observed in the ENDOR spectra between spinach and T. vulcanus PSII. In the Ca2+-depleted PSII, dd′ peaks disappeared and were therefore assigned to the water molecule W4 ligated to Ca2+ in the crystal structure. This indicates the rearrangement of the hydrogen-bonding network near the Ca2+ site upon its depletion. The ENDOR spectra were not significantly changed by Sr2+ substitution, in agreement with the results of crystal structure analysis showing that the structure of the manganese cluster was not significantly changed in Sr2+-substituted PSII. Based on the present results, the roles of Ca2+ and Sr2+ could involve the maintenance of the hydrogen bond network near the Ca2+ site and electron transfer pathway to the manganese cluster.
H. N., J.-R. S., and H. M. designed the study and wrote this paper. H. N. and H. M. prepared PSII samples from spinach and performed EPR and ENDOR measurements. Y. N. and J.-R. S. prepared PSII samples from T. vulcanus.
*This work was supported by a Program for Leading Graduate Schools “Integrative Graduate Education and Research in Green Natural Sciences,” MEXT (Ministry of Education, Culture, Sports, Science and Technology), Japan; Grant-in-Aid for JSPS (Japan Society for the Promotion of Science) Fellows 26011113 (to H. N.); JSPS Grant-in-Aid for Specially Promoted Research 24000018 (to J.-R. S.); and MEXT/JSPS Grant-in-Aid for Exploratory Research 26620003 (to H. M.). The authors declare that they have no conflicts of interest with the contents of this article.
2The abbreviations used are: