Changes in Structural and Functional Properties of Oxygen-evolving Complex Induced by Replacement of D1-Glutamate 189 with Glutamine in Photosystem II

A carboxylate group of D1-Glu-189 in photosystem II has been proposed to serve as a direct ligand for the manganese cluster. Here we constructed a mutant that eliminates the carboxylate by replacing D1-Glu-189 with Gln in the cyanobacterium Synechocystis sp. PCC 6803, and we examined the resulting effects on the structural and functional properties of the oxygen-evolving complex (OEC) in photosystem II. The E189Q mutant grew photoautotrophically, and isolated photosystem II core particles evolved oxygen at ∼70% of the rate of control wild-type particles. The E189Q OEC showed typical S2 state electron spin resonance signals, and the spin center distance between the S2 state manganese cluster and the YD (D2-Tyr-160), detected by electron-electron double resonance spectroscopy, was not affected by this mutation. However, the redox potential of the E189Q OEC was considerably lower than that of the control OEC, as revealed by the elevated peak temperature of the S2 state thermoluminescence bands. The mutation resulted in specific changes to bands ascribed to the putative carboxylate ligands for the manganese cluster and to a few carbonyl bands in mid-frequency (1800 to 1100 cm-1) S2/S1 Fourier transform infrared difference spectrum. Notably, the low frequency (650 to 350 cm-1) S2/S1 Fourier transform infrared difference spectrum was also uniquely changed by this mutation in the frequencies for the manganese cluster core vibrations. These results suggested that the carboxylate group of D1-Glu-189 ligates the manganese ion, which is influenced by the redox change of the oxidizable manganese ion upon the S1 to S2 transition.

n represents the number of stored oxidizing equivalents. In the dark, the OEC is predominantly in a thermally stable S 1 state and advances to the next oxidation state, S 2 , by absorbing a photon. The OEC reaches the highest oxidation state, S 4 , by accumulating four oxidizing equivalents after three flashes and then relaxes to the lowest oxidation state, S 0 , concurrent with the release of an oxygen molecule (2,3).
Site-directed mutagenesis studies using the cyanobacterium Synechocystis sp. PCC 6803 (reviewed in Refs. 4 and 5) have revealed that replacement of amino acid residues by another residue affects the properties of the OEC. These amino acids include Asp-170, Glu-189, His-190, His-332, Glu-333, His-337, Asp-342, and Ala-344 of the D1 protein (6 -11). Some residues are arranged in close proximity to the manganese cluster, as shown by x-ray crystallographic models (12)(13)(14)(15), suggesting that the manganese cluster is coordinated mainly with carboxylate and imidazole groups from these residues in the D1 protein. Several carboxylate and histidine ligands were characterized spectroscopically using a combination of site-directed mutagenesis and isotope labeling. The replacement of D1-Asp-170 with His resulted in marked changes in the low frequency (650 to 500 cm Ϫ1 ) S 2 /S 1 Fourier transform infrared (FTIR) spectrum (16); however, this substitution induced no change in the S 1 and S 2 multiline electron spin resonance (ESR) signals (17) or in the mid-frequency FTIR difference spectra for the S 1 to S 2 (16,18) and other transitions (18). Therefore, it appears that the carboxylate side group of D1-Asp-170 ligates the manganese ion that does not undergo a redox change during the S state cycling (18). Isotopic bands for the 13 C-labeled C-terminal carboxylate of D1-Ala-344 appeared in the mid- (19,20) and low frequency (20) S 2 /S 1 FTIR spectra, indicating the ligation of the D1-Ala-344 carboxylate to the redox-active manganese ion. Consistently, replacement of the methyl side group of D1-Ala-344 with another group induced characteristic changes in OEC properties (21,22). Furthermore, S state-dependent changes of the isotopic bands showed that the manganese ion associated with the D1-Ala-344 carboxylate is oxidized upon the S 1 to S 2 transition and re-reduced during the S 3 to S 0 transition (20). The C-terminal carboxylate was not defined in the electron density map of the diffraction data but was assumed to be feasible to ligate a Ca 2ϩ ion (14). However, FTIR studies showed that neither Ca 2ϩ depletion nor 44 Ca-labeling induced spectral change of the carboxylate bands, including those for the C-terminal carboxylate (23). Furthermore, the ligation of the D1-Ala-344 carboxylate to a redoxactive manganese ion but not to a Ca 2ϩ ion was strongly supported by the FTIR study using Synechocystis in which calcium was biosynthetically replaced with strontium (24). The ligation of histidine to the manganese cluster is suggested based on electron spin echo envelope modulation (25) and FTIR (26) studies for the cyanobacterium Synechocystis sp. PCC 6803 having a histidine containing two 15 N atoms in the imidazole ring.
X-ray models showed that the carboxylate side group of D1-Glu-189 serves as a unidentate ligand for the manganese cluster consistently (14,15), but the models were contradictory to each other concerning the presence (15) and absence (14) of the hydrogen bond to Y Z tyrosine . The  PS II cores isolated from D1-E189D, D1-E189G, D1-E189N, D1-E189S, and D1-E189H mutants evolved no oxygen despite possessing the functional reaction centers (27). In these particles, neither an S 1 nor an S 2 multiline ESR signal was observed; however, an S 2 Y Z ⅐ state split signal was apparent under multiple turnover conditions. Furthermore, the accelerated charge recombination of Q A Ϫ and Y Z ⅐ (27) indicated a change in the redox potential of Y Z in these mutants. It was then proposed that D1-Glu-189 is part of a hydrogen bond network around Y Z and the manganese cluster for rapid electron/proton transfer from the manganese cluster to Y Z (27). In contrast, the D1-E189Q, D1-E189R, and D1-E189K mutants grew photoautotrophically with ϳ70 -80% oxygen evolution compared with wild-type (27). They showed normal electron transfer from Y Z to P 680 ϩ⅐ as well as from the manganese cluster to Y Z ⅐ (28), and the PS II cores isolated from E189Q mutants produced normal S 1 and S 2 multiline ESR signals (27). The largely normal OEC properties in these mutants seem to be inconsistent with the proposed role of the D1-Glu-189 carboxylate for ligating the manganese ion as well as hydrogen bonding to Y Z , because the carboxylate group is absent in the mutants, and the negatively charged residue was replaced with a neutral or positively charged residues. Therefore, the structural and functional roles of D1-Glu-189 are still largely unresolved.
In the present study, we constructed the Synechocystis D1-E189Q mutant, which lacks the carboxylate side group but is expected to induce nonspecific structural perturbations minimally. The structural and functional properties of the mutant were examined by means of ESR, thermoluminescence, and FTIR spectroscopy in isolated active PS II core particles. The D1-E189Q mutation induced minimal changes in the magnetic properties of the manganese cluster but resulted in the characteristic changes in the redox potential and ligation geometry of the manganese cluster. Based on these and other findings, the structural and functional roles of D1-Glu-189 are discussed.

EXPERIMENTAL PROCEDURES
Sample Materials-A host strain of Synechocystis sp. PCC 6803 for site-directed mutagenesis (B-His/N⌬AA) was constructed by replacing a portion of psbA1, psbA2, and psbA3 genes with DNA fragments conferring resistance to chloramphenicol, spectinomycin, and tetracycline, respectively, and a hexahistidine tag was attached to the C terminus of the CP47 protein (21). Mutations were introduced into the psbA2 gene of Synechocystis sp. PCC 6803 cloned in pNA219, which retains a DNA fragment conferring resistance to kanamycin downstream of the psbA2 gene, as described elsewhere (21). For construction of the D1-E189Q strain, Glu-189 codon GAG was changed to CAG. A plasmid bearing the mutation was transformed into the host strain of Synechocystis sp. PCC 6803, and single colonies were selected for their photoautotrophic growth ability on solid BG-11 medium containing 5 g/ml kanamycin. For constructing a wild-type control strain, a plasmid bearing the wildtype psbA2 gene was transformed into the host strain. Synechocystis cells were photoheterotrophically grown in liquid BG-11 medium supplemented with 5 mM glucose at 30°C under 30 -50 mol of photons/ m 2 /s. Harvested cells were disrupted using a bead beater (Bio-Spec Products), and the resulting thylakoid membranes were solubilized with n-dodecyl ␤-D-maltoside, and PS II particles were affinity-purified on a nickel-nitrilotriacetic acid column (Qiagen) as described previously (21). The purified core particles were washed with medium A (400 mM sucrose, 20 mM NaCl, 5 mM CaCl 2 , and 20 mM Mes-NaOH, pH 6.0) supplemented with 10% (w/v) PEG6000, and suspended in medium A after extensive washing with medium A. For analysis of protein composition, PS II core particles were solubilized with 1% SDS, followed by separation using SDS-PAGE on a 16 -22% gradient gel containing 7.5 M urea (21).
FTIR Measurements-Mid-frequency (1800 to 1100 cm Ϫ1 ) FTIR spectra were recorded on a Bruker IFS-66v/S spectrophotometer equipped with an MCT detector (EG&G Optoelectronics D316/6) (29). Low frequency (650 to 350 cm Ϫ1 ) FTIR spectra were recorded on a Bomen MB102 spectrophotometer equipped with a silicon bolometer (Infrared Laboratories, HDL-5) (29). PS II core particles suspended in a medium containing 40 mM sucrose, 10 mM Mes-NaOH, 5 mM NaCl, and 5 mM CaCl 2 , pH 6.0, were centrifuged at 100,000 ϫ g for 5 min, and a portion (ϳ40 g of Chl) of the resulting pellet was directly mixed with 0.3 l of sodium ferricyanide solution (0.2 M stock) added as an electron acceptor on either a BaF 2 disk (13 mm inner diameter) for mid-frequency or an AgBr disk (13 mm inner diameter) for low frequency measurements. Sample temperature was maintained at 250 Ϯ 0.03 K using a homemade cryostat (29). Sample cores were illuminated with a flash provided from a frequency-doubled Nd 3ϩ :YAG laser (Spectra Physics INDI-50, 532 nm, pulse width 6 -7 ns) with flash energy of ϳ10 mJ/cm 2 at the sample surface. Single beam spectra were accumulated at 4 cm Ϫ1 resolution in the mid-frequency (150 scans) or low frequency (60 scans) region before and after flash illumination, and light-minusdark difference spectra were calculated. To improve signal-to-noise ratio, 25-29 mid-frequency difference spectra (3750 -4350 scans) or 91-96 low frequency difference spectra (5460 -5760 scans) were averaged.
ESR Measurements-X-band CW and pulsed ESR experiments were performed on a Bruker ESR 580 spectrometer equipped with a cylindrical (ER4122SHQ) or cylindrical dielectric (ER4117DHQ-H) cavity. Sample core particles were suspended in a medium (400 mM sucrose, 5 mM CaCl 2 , 5 mM MgCl 2 , and 50 mM Mes-NaOH, pH 6.0, and 0.1 mM potassium ferricyanide) to be 4 mg of Chl/ml, transferred to a Spracil quartz ESR sample tube, and illuminated at 213 K for 5 s with cold light (Hayashi, LA-150TX). Sample temperature for ESR measurements was controlled using a gas flow temperature control system (Oxford Instruments, ESR900 or CF935). For ESE ELDOR measurements, a threepulse sequence was used as described previously (30) with microwave pulses of 16-, 24-, and 24-ns duration for the 1st, 2nd, and 3rd pulse, respectively. A microwave synthesizer (Hewlett Packard, HP83751B) was used as the second microwave source.
Other Measurements-Thermoluminescence glow curves were measured as described previously (21). Cells and PS II core particles were suspended in medium A at 250 g of Chl/ml in the presence of 0.1 mM DCMU and illuminated at 0°C with a xenon flash. The O 2 evolution activity was measured using a Clark-type oxygen electrode in medium A at 25°C in the presence of 1 mM 2,5-dimethyl-1,4-benzoquinone and 2 mM potassium ferricyanide for cells or 4 mM potassium ferricyanide for core particles as described (21).

RESULTS
TABLE ONE shows the effects of the E189Q mutation on cell growth and O 2 evolution activity. The mutant cells grew photoautotrophically under both high and low light conditions and showed high O 2 evolution activity, although the rates were somewhat lower than those of the wildtype control. The PS II core particles isolated from the mutant cells showed high O 2 evolution activity, which was lower relative to that from the wild-type cells in agreement with the differences in whole cell activ-ity and growth. These properties of the E189Q mutant were basically consistent with those reported previously (27).
The SDS-PAGE profiles of the PS II core particles 5 showed no difference in protein composition between wild-type control and the E189Q mutant in terms of major intrinsic proteins, including CP47, CP43, D2, D1, and ␣ subunit of cytochrome b 559 , as well as three extrinsic proteins including 33 kDa, cytochrome c 550 , and 12 kDa. Therefore, it is conceivable that the lower O 2 activity in the mutant core particles is not ascribed to differences in the protein compositions but reflects some change in the intrinsic properties of the OEC induced by the mutation. Fig. 1 shows the flash-induced S 2 Q A Ϫ thermoluminescence glow curves of the cells (Fig. 1a) and PS II core particles (Fig. 1b) in the wild-type control (dashed curves) and E189Q mutant (solid curves). The wild-type cells showed an S 2 Q A Ϫ band (ϩDCMU) at ϳ9°C, consistent with that reported previously (21). The peak temperature of the S 2 Q A Ϫ band in the mutant cells was elevated to 24°C. The wild-type and mutant PS II core particles showed S 2 Q A Ϫ bands at 20 and 30°C, respectively. Notably, the peak temperature was upshifted in the mutant cores, which agrees with trends observed in the thermoluminescence bands of the whole cells, although the bands appeared at a relatively higher temperature for the core particles than the cells. Presumably, some differences in the salt and pH conditions proximal to PS II between the whole cells and core particles are attributable to the observed difference. The results indicate that the redox potential of the S 2 state manganese cluster was lowered by replacement of D1-Glu-189 with Gln, which may be responsible for the reduced O 2 evolution activity in the mutant, as shown in TABLE ONE. Fig. 2 shows the effects of the mutation on the light-induced S 2 state ESR spectra for the wild-type control (Fig. 2a) and E189Q mutant (Fig.  2b) PS II core particles. Both spectra showed a prominent g ϭ 2 S 2 multiline signal and Fe 2ϩ Q A Ϫ signal at g ϭ 1.9, and a negligibly small g ϭ 4.1 S 2 signal (Fig. 2, inset), although the intensity of the multiline signal in the mutant spectrum was ϳ70% that of the control spectrum. Apparently, the smaller multiline intensity is not ascribed to the conversion of the multiline to the g ϭ 4.1 S 2 signal. No change in the hyperfine structure of the multiline signal by the mutation is consistent with the reported results (27). The results indicated that the intra-magnetic structure of the S 2 state manganese cluster was not affected by replacement of Glu-189 with Gln. The absence of the putative carboxylate ligand for the manganese ion in the E189Q mutant may result in a subtle positional change of the manganese cluster in PS II if the carboxylate is responsible for the ligation of the manganese cluster, even though the intra-cluster magnetic structure was not affected by the mutation. To address this issue, mutation effects on the distance between Y D tyrosine and the manganese cluster were evaluated. Fig. 3 shows the experimental (open circles) and simulated (solid lines) ESE ELDOR spectra for the pair of spins composed of the radical of Y D tyrosine and S 2 multiline center of the manganese cluster in the wild-type control (Fig. 3a) and E189Q mutant (Fig.  3b) PS II core particles. The distance between the two centers was determined to be 26.7 Å with an accuracy of 0.2 Å for the wild-type and the mutant particles, indicating that the mutation does not induce any positional change of the manganese cluster in PS II, although we cannot completely exclude the possibility that the relative arrangement of the manganese cluster on a sphere made by the 26.7 Å radius vector from Y D tyrosine was changed. Notably, the distance obtained was slightly shorter than that obtained in spinach (27.1 Å) (30), suggesting a subtle difference in the relative position of the manganese cluster between cyanobacteria and higher plants. Fig. 4 shows the mid-frequency (1800 to 1100 cm Ϫ1 ) S 2 /S 1 FTIR difference spectra of the wild-type control (Fig. 4a, black) and E189Q mutant (Fig. 4a, magenta) PS II core particles. The wild-type spectrum  Low light-grown cells and PS II core particles from the low light-grown cells were used for assays of oxygen evolution.

FIGURE 1. S 2 Q A ؊ thermoluminescence glow curves of the whole cells (a) and PS II core particles (b) in wild-type control (dashed curves) and D1-E189Q mutant (solid curves) of Synechocystis sp. PCC 6803.
Sample suspensions included 0.1 mM DCMU. The glow curves were normalized with the peak intensity for the S 2 Q A Ϫ band.

FIGURE 2. Light-minus-dark ESR spectra of the PS II core particles from wild-type control (a) and D1-E189Q mutant (b) cells of Synechocystis sp. PCC 6803.
The inset figure is the expanded view of the spectra. Instrument settings are as follows: temperature, 6 K; microwave power, 0.5 milliwatt; microwave frequency, 9.48 GHz; modulation frequency and amplitude, 100 kHz and 1.6 millitesla, respectively. a.u., arbitrary unit.
showed characteristic S 2 /S 1 vibrational features, such as symmetric (1450 to 1300 cm Ϫ1 ) and asymmetric (1600 to 1500 cm Ϫ1 ) stretching modes from the putative carboxylate ligands for the manganese cluster as well as amide I (1700 to 1600 cm Ϫ1 ) and II (1600 to 1500 cm Ϫ1 ) modes from polypeptide backbones (29). The E189Q spectrum showed S 2 /S 1 vibrational features similar to those of wild-type spectrum in general, but distinctive differences were observed, particularly in the symmetric carboxylate stretching region. Predominant S 2 /S 1 carboxylate stretching bands at 1435(ϩ)/1417(Ϫ) cm Ϫ1 were shifted slightly upward by this mutation, resulting in the appearance of differential bands at 1441(ϩ)/1421(Ϫ) cm Ϫ1 , whereas other symmetric carboxylate bands did not change and occurred at the same positions as those in the wildtype spectrum. Most of the asymmetric carboxylate stretching and amide II bands were not affected by the mutation, with the exception of minor changes near 1520 to 1500 cm Ϫ1 . In contrast, several bands at 1700 to 1600 cm Ϫ1 in the E189Q spectrum were clearly different from those in the wild-type spectrum; intensity of the negative band at 1678 cm Ϫ1 decreased and its peak position was shifted slightly upward to 1680 cm Ϫ1 . To inspect the spectral changes induced by the E189Q mutation more closely, a double difference spectrum (Fig. 4b) was obtained by subtracting the E189Q spectrum from the wild-type spectrum after normalization with respect to peak-to-peak intensity of ferrocyanide (2037 cm Ϫ1 ) and ferricyanide (2115 cm Ϫ1 ) bands (29). The E189Q spectrum contained several symmetric carboxylate stretching bands that were affected, resulting in the bands at 1446(Ϫ), 1427(ϩ), 1417(Ϫ), 1412(ϩ), 1404(Ϫ), and 1392(ϩ) cm Ϫ1 in the double difference spectrum; however, no significant band occurred lower than 1370 cm Ϫ1 , except for minor bands at 1261(ϩ)/1244(Ϫ) cm Ϫ1 . In the 1700 to 1500 cm Ϫ1 region, bands appeared at 1686(ϩ), 1676(Ϫ), 1645(ϩ), 1628(Ϫ), 1520(Ϫ), and 1502(ϩ) cm Ϫ1 , although the bands at 1650 to 1600 cm Ϫ1 may include some ambiguity because of the presence of large amide I and water bands. It is of note that the 1113 cm Ϫ1 band, which was assigned to the ring CN stretching mode of the histidine ligand for the manganese cluster (26), was minimally affected by the E189Q mutation. X-ray models indicate that D1-His-332 ligates the same manganese ion associated with D1-Glu-189 (14,15), and the redox and magnetic properties of both manganese cluster and Y Z are modulated in the D1-H332E mutant (31,32). However, no appreciable difference in a histidine band at 1113 cm Ϫ1 suggests that the interaction between D1-His-332 and the manganese cluster was not affected by the mutation. Fig. 5 shows the low frequency (650 to 350 cm Ϫ1 ) S 2 /S 1 FTIR difference spectra of the wild-type control (Fig. 5a, black) and E189Q mutant (Fig. 5a, magenta) PS II core particles. The wild-type spectrum showed characteristic low frequency S 2 /S 1 vibrational features, such as bands at . ESE ELDOR conditions: first and third pulses, 9.46 GHz; second pulse, 9.55 GHz; , 1200 ns; pulse repetition time, 1 ms; B 0 (external static magnetic field), 0.3438 T; temperature, 6 K. Inset shows the primary ESE field-swept spectra for the PS II core particles from the D1-E189Q mutant cells. The field-swept ESE conditions are as follows: microwave frequency, 9.59 GHz; pulse length, 16 and 24 ns for /2 and pulse; pulse repetition time, 10 ms; temperature, 6 K. a.u., arbitrary unit. FIGURE 4. Mid-frequency (1800 to 1100 cm Ϫ1 ) S 2 /S 1 FTIR difference spectra of the PS II core particles from wild-type control (a, black) and D1-E189Q mutant (a, magenta) cells of Synechocystis sp. PCC 6803 are shown. Double difference spectrum (b) was obtained by subtracting the E189Q spectrum from the wild-type spectrum after normalization with respect to the peak-to-peak intensity of the 2115(Ϫ) cm Ϫ1 ferricyanide and 2037(ϩ) cm Ϫ1 ferrocyanide bands (29). A dark-minus-dark spectrum (c) is presented to show the noise level. FIGURE 5. Low frequency (650 to 350 cm Ϫ1 ) S 2 /S 1 FTIR difference spectra of the PS II core particles from wild-type control (a, black) and D1-E189Q mutant (a, magenta) cells of Synechocystis sp. PCC 6803 are shown. Double difference spectrum (b) was obtained by subtracting the E189Q spectrum from the wild-type spectrum after normalization as described in the text. A dark-minus-dark spectrum (c) is presented to show the noise level. 642(ϩ), 629(ϩ), 621(Ϫ), 605(ϩ), 591(ϩ), 578(Ϫ), 403(Ϫ), and 390(Ϫ) cm Ϫ1 as well as several medium to low intensity bands, consistent with those reported previously (29). The 591(ϩ) and ϳ400(Ϫ) cm Ϫ1 bands were ascribed to the vibrational modes of ferrocyanide and ferricyanide, respectively (29). The E189Q spectrum showed a marked difference from the wild-type spectrum at 640 to 570 cm Ϫ1 ; however, no significant difference was detected below 570 cm Ϫ1 . The effects on the S 2 /S 1 vibrational features induced by the E189Q mutation can be seen clearly in the double difference spectrum (Fig. 5b) obtained by subtracting the E189Q spectrum from the wild-type spectrum after normalization with the factor that was used for the double subtraction of the mid-frequency spectra. The double difference spectrum contained a feature composed of three prominent bands at 623(Ϫ), 604(ϩ), and 581(Ϫ) cm Ϫ1 . The band assignment using multiplex water isotopes (33) demonstrated that this region includes few hydrogen-sensitive modes but many oxygensensitive modes: putative Mn-O-Mn cluster modes at 625 cm Ϫ1 (S 1 state) and 606 cm Ϫ1 (S 2 state) (33,34) as well as skeletal vibrations of the manganese cluster or the Mn-O mode at 577 cm Ϫ1 (S 1 state) (29). Therefore, the marked changes in Fig. 5 indicate that changes in the internal structure of the manganese cluster are directly induced by replacement of a ligand to the manganese ion from D1-Glu-189 to Gln rather than some indirect perturbation through hydrogen bondings.

DISCUSSION
The present results demonstrate that structural properties of the manganese cluster were characteristically modulated with a little change in function by substituting Gln for Glu at position 189 of the D1 protein. The mid-frequency FTIR spectrum of the E189Q mutant PS II core particles showed changes of the specific bands in the symmetric stretching region at 1450 to 1370 cm Ϫ1 for the putative carboxylate ligands. Apparently, these changes are the direct reflection of the loss of the Glu-189 carboxylate group responsible for manganese ligation or ascribed to some indirect structural perturbation induced by the loss of the Glu-189 carboxylate through the hydrogen bonds occurring in carboxylate group(s), which is not a ligand to the manganese cluster but is located in the surroundings of the cluster. Generally, symmetric stretching bands for the hydrogen-bonded carboxylate show the change in the intensity upon the D 2 O substitution despite no change in the peak position (35). However, the 1450 to 1370 cm Ϫ1 bands specifically affected by the E189Q mutation did not show any change in the intensity as well as in the peak position by the D 2 O substitution 5 (36), suggesting that the indirect perturbation is an unlikely scenario. Consistently, there was no carboxylate ligand that can be structurally coupled with Glu-189 through hydrogen bondings in the x-ray crystallographic models (14,15). Therefore, a more preferable explanation is that the symmetric stretching mode of the Glu-189 carboxylate directly ligating the manganese ion exists in the 1450 to 1370 cm Ϫ1 region, and lack of the mode is responsible for the bands found in the double difference spectrum (Fig. 4b). The E189Q mutation resulted in marked changes of the low frequency S 2 /S 1 bands in the 640 to 570 cm Ϫ1 region (Fig. 5). This region mostly includes oxygen-sensitive but hydrogen-insensitive bands for manganese cluster core vibrations: the Mn-O-Mn stretching mode at 606(ϩ)/625(Ϫ) cm Ϫ1 (33,34) and the skeletal vibration of the manganese cluster or stretching vibrational modes of the Mn-O at 577(Ϫ) cm Ϫ1 (29). Apparently, it is difficult to imagine how the indirect perturbation in the surroundings affects the core structure of the manganese cluster without any changes in the internal magnetic structure of the manganese cluster (Fig. 2). This could be possibly attained if the mutation induces specific change in a specific ligand for the manganese ion. Therefore, the observed changes of the low frequency bands indicate mal S 2 multiline signal (Fig. 2), and a lack of change in the location of the spin center of the S 2 state manganese cluster (Fig. 3) indicate that the manganese associated with the D1-Glu-189 carboxylate is ligated by another amino acid side group or peptide backbone with a position close to the Glu-189 carboxylate in the E189Q mutant. Of particular interest is a differential band at 1686(ϩ)/1676(Ϫ) cm Ϫ1 in the mid-frequency double difference spectrum (Fig. 4b) where the carbonyl stretching mode of glutamine should appear (39). In fact, such modes were assumed to appear at 1692/1659 cm Ϫ1 in the Y D ⅐ /Y D spectrum of PS II core particles from the D2-H189Q mutant (40). Therefore, these results indicate that the carbonyl group of Gln-189 serves as a ligand for the manganese ion in the E189Q mutant. Previous studies have reported that cells showed photoautotrophic growth with O 2 capability in mutants where D1-Glu-189 was replaced with Arg, Lys, Leu, or Ile (27). The peptide carbonyl probably is responsible for manganese ligation in these mutants, except in the E189Q mutant because the side group carbonyl of Gln-189 can lie in a position very similar to the side group carboxylate of Glu-189 to serve as a direct ligand. Another possible role of D1-Glu-189 is as a proton acceptor from Y Z through D1-His-190 either directly or indirectly (41,42). However, no detectable change was induced by the E189Q mutation in the 1790 to 1710 cm Ϫ1 region (Fig. 4b) where the CO stretching mode of the carboxyl group of glutamic acid appears (39). Therefore, protonation of Glu-189 is a less likely scenario, at least during the S 1 to S 2 transition, consistent with high O 2 evolution capability in the E189Q mutant that no longer contains a proton-accepting carboxylate at position 189. It has been proposed that the 1521 or 1254 cm Ϫ1 band in the S 2 /S 1 spectrum can be assigned to the ring CC stretching mode or CO stretching and/or COH bending modes of the Y Z based on the effects of tyrosine isotope labeling (43). Therefore, the double difference bands at 1520(Ϫ)/ 1502(ϩ) and 1261(ϩ)/1244(Ϫ) cm Ϫ1 in Fig. 4b could be attributable to the Y Z modes, which are changed by the E189Q mutation. However, we have not pursued these in more detail in the present study because much ambiguity remained for the assignment of the Y Z bands in the present spectrum.
In conclusion, replacement of D1-Glu-189 with Gln induced changes in the ligation structure of the manganese cluster, as determined by observing shifts in the symmetric carboxylate stretching modes and in manganese cluster core modes. The present results suggested that the D1-Glu-189 carboxylate serves as a direct ligand for the manganese ion, which is influenced by the redox change of the oxidizable manganese ion upon the S 1 to S 2 transition, and does not participate in the proton transfer reaction from redox-active Y Z to the luminal surface of the PS II during the S 1 to S 2 transition.