Impact of Replacement of D1 C-terminal Alanine with Glycine on Structure and Function of Photosynthetic Oxygen-evolving Complex*

The C-terminal alanine 344 (Ala-344) in the D1 protein of photosystem II is conserved in all of the organisms performing oxygenic photosynthesis. A free (cid:1) -COO (cid:2) of Ala-344 has been proposed to be responsible for ligating the Mn cluster. Here, we constructed a mutant having D1 in which D1-Ala-344 was replaced with glycine (Gly) in cyanobacterium Synechocystis sp. PCC 6803. The effects of this minimal change in the side group from methyl to hydrogen on the properties of the oxygen-evolving complex were comprehensively investigated using purified core particles. The mutant grew photoautotrophically, and little change was observed in the protein composition of the oxygen-evolving core particles. The Gly-substituted oxygen-evolving complex showed small but normal S 2 multiline and enhanced g (cid:3) 4.1 electron spin resonance signals and S 2 -state thermoluminescence bands with slightly elevated peak temperature. The Gly substitution resulted in distinct but relatively small changes in a few bands arising from the putative carboxylate

Photosynthetic water oxidation takes place in an oxygenevolving complex (OEC) 1 in which the catalytic center is composed of a tetranuclear Mn cluster located on the lumenal side of the D1/D2 heterodimer. Two water molecules are oxidized to an oxygen molecule through five intermediates labeled S n (n ϭ 0 -4), where n denotes the number of oxidizing equivalents stored. In a dark-adapted sample, a thermally stable S 1 state predominates. The S n state advances to the S nϩ1 state by absorbing each photon to reach the highest oxidation state, S 4 , which spontaneously relaxes to the lowest oxidation state, S 0 , concomitant with the release of an oxygen molecule (1,2).
Studies using chemical modifiers (3) and electron spin echo envelope modulation (4) and FTIR spectroscopy (5,6) suggested that histidine and/or acidic amino acids are involved in the ligation of the Mn cluster. Several residues of the D1 protein have been proposed based on site-directed mutagenesis studies mainly using cyanobacterium Synechocystis sp. PCC 6803 as potential candidates for the ligands to the Mn cluster, (reviewed in Refs. [7][8][9]. They are Asp-170, Glu-189, His-190, His-332, Glu-333, His-337, Asp-342, and Ala-344 (10 -15), some of which were arranged in close proximity to the Mn cluster in x-ray structural models of photosystem (PS) II (16 -19). However, the properties of OEC have not been characterized using isolated PS II preparations with the exception of a few mutants. The studies using the O 2 -evolving PS II core particles from the D1-D170H mutant showed that the mutation leads to little change of the S 1 and S 2 multiline ESR, S 2 multiline electron spin echo envelope modulation signals, and the midfrequency S 2 /S 1 FTIR difference spectrum but does lead to some changes of the low frequency (650 -500 cm Ϫ1 ) S 2 /S 1 FTIR difference spectrum (20,21). The PS II cores from D1-H332E showed no O 2 evolution but retained the Mn cluster with an altered S 2 multiline ESR signal in which the electron spin echo envelope modulation spectrum showed no nitrogen modulation (22,23). The PS II cores from D1-E189D, D1-E189N, D1-E189H, D1-E189G, and D1-E189S showed no oxygen evolution and neither S 1 nor S 2 multiline ESR signal but did reveal a Y Z S 2 -state split signal. In contrast, D1-E189Q and D1-E189L mutants grew photoautotrophically, and their PS II cores showed the normal multiline signals (24).
The D1 protein is synthesized with a short C-terminal extension with the exception of Euglena, assembled into the PS II complex (25), and subsequently cleaved on the carboxyl side of Ala-344 by D1 C-terminal-processing protease (26). The processing is prerequisite to the light-dependent assembly of the Mn cluster (27,28), but the mutant with no extension by substituting the stop codon at D1-345 for the amino acid codon (D1-345stop) showed normal photoautotrophic growth and O 2 evolution capability (11,29). Replacement of D1-Ala-344 with Gly, Met, Ser, or Val in the D1-345stop (D1-Ala-344-stop) strain did not affect photoautotrophic growth. However, Tyr or Lys substitution led to a marked decrease in O 2 evolution and loss of the capability for photoautotrophic growth (11). None of D1 C-terminal-truncated mutants evolved oxygen (11). Therefore, the free C-terminal (␣-COO Ϫ ) of the D1 Ala-344 has been proposed to ligate one or more Mn ions (11). The PS II core particles isolated from wild-type Synechocystis cells labeled with L- [1-13 C]alanine showed that several bands in the midfrequency S 2 /S 1 FTIR difference spectrum are affected by the labeling (30). This result indicates that the isotope-affected bands can be ascribed to the ␣-carboxylate group of D1-Ala-344 as a ligand for the Mn cluster, although an indirect structural coupling between the Mn cluster and D1-Ala-344 cannot be excluded. Ligation of the D1-Ala-344 carboxylate to the Mn cluster was proposed based on the 3.6 Å (17) and 3.7 Å (18) x-ray structural model. The C-terminal carboxylate was arranged in close proximity to a Ca ion in a recent 3.5-Å model in which a cubane-like Mn 3 CaO 4 linked to a fourth Mn by a mono--oxo bridge was proposed, although the C-terminal carboxylate was disordered and not visible in the electron density map (19). It may be worthwhile to note in this context that no individual metal ion in the OEC was resolved in any reported x-ray electron density maps.
The D1-A344G mutant has been reported to grow photoautotrophically (11). Although the replacement of alanine with glycine is accompanied with only the change of a side group from methyl to hydrogen, the functional and/or structural properties of the Mn cluster may be affected even by this minimal change if the C-terminal free carboxylate is involved in the ligation of the Mn cluster. In the present study, we have constructed D1-Ala-344-stop (ϭ D1-Ser-345stop) and D1-A344Gstop mutants on a Synechocystis sp. PCC 6803 strain with histidine-tagged CP47 and have isolated active PS II core particles from them. The replacement of the C-terminal Ala with Gly had little effect on biochemical properties of OEC but led to considerable changes in functional/structural properties of the Mn cluster. These results are compatible with the proposal that the C-terminal ␣-COO Ϫ carboxylate of D1-Ala-344 is crucial for maintaining the intrastructure of the Mn cluster, possibly by ligating the Mn cluster.

EXPERIMENTAL PROCEDURES
Construction of a Host Strain for Mutation-The genomic DNA from Synechocystis sp. PCC 6803 was amplified by PCR with a specific primer set corresponding to the psbA1, psbA2, or psbA3 gene. For psbA1, a 1727-bp DNA containing 1080 bp psbA1 plus 414 bp 5Ј and 233 bp 3Ј flanking DNA was cloned into the plasmid pBluescript II SKϩ. A 0.45-kb HincII/NsiI fragment containing 40% psbA1 was replaced by a 1.4-kb fragment from the plasmid pACYC184 conferring resistance to chloramphenicol to generate the plasmid pN⌬A1. For psbA2, a 1972-bp DNA containing 1053 bp 3Ј 97% psbA2 plus 919 bp 3Јflanking DNA was cloned into pUC19 to generate the plasmid pNA2. A 0.9-kb HincII/StuI fragment corresponding to 83% psbA2 and a part of its downstream gene, slr1312, was replaced by a 2-kb fragment of the plasmid pRL463 (a kind gift from Prof. T. Omata, Nagoya University) conferring resistance to spectinomycin to generate the pN⌬A2 plasmid. For psbA3, a 1691-bp DNA containing 1080 bp psbA3 plus 326 bp 5Ј and 285 bp 3Јflanking DNA was cloned into the plasmid pUC19. A 0.4-kb KpnI/ KpnI fragment was replaced by a 1.7-kb fragment of pACYC184 conferring resistance to tetracycline to generate the plasmid pN⌬A3. The plasmids pNA1, pNA2, and pNA3 were successively transformed into the glucose-tolerant wild-type strain of Synechocystis. The resulting strain (N⌬AA) was resistant to chloramphenicol, tetracycline, and spectinomycin and lacked all three psbA genes. A hexahistidine tag was introduced to the C terminus of CP47 of the wild-type and N⌬AA strain as described previously (31) to generate the wild*-type and B-His/N⌬AA strain. The B-His/N⌬AA strain was used for further site-directed mutagenesis.
Construction of Site-directed Mutants-A 1.3-kb fragment from the plasmid pUC4K conferring resistance to kanamycin (Km) was inserted into the StuI site located 288 bp downstream of the psbA2 gene in pNA2, thus generating the plasmid pNA219. Mutations were intro-duced into the psbA2 gene in pNA219 using a commercial oligonucleotide-mediated mutagenesis kit (QuikChange site-directed mutagenesis kit; Stratagene). For the Ala-344-stop (ϭ Ser-345stop) strain, the Ser-345 codon TCT was changed to a stop codon TGA. For the A344Gstop strain, the Ala-344 codon GCG and the Ser-345 codon TCT were changed to glycine GGG and the stop codon TGA, respectively. Plasmids bearing the mutations were transformed into the host strain (B-His/ N⌬AA), and single colonies were selected for their photoautotrophic growth ability on solid BG-11 medium containing 5 g/ml kanamycin.
Culture Conditions and Preparation of PSII Core Particles-The Synechocystis cells were photoheterotrophically grown in liquid BG-11 medium supplemented with 5 mM glucose at 30°C under 30 -50 mol photons/m 2 /s in an 8-liter Clearboy (NALGENE), bubbling with air up to 7-8 g of Chl/ml unless otherwise noted. The PS II core particles were prepared as previously described (31). Briefly, the harvested cells were disrupted using a Bead beater (Bio-Spec Products). The resulting thylakoid membranes were solubilized using n-dodecyl-␤-D-maltoside, and then the PS II particles were affinity purified with a nickel-nitrilotriacetic acid column (Quiagen). The purified core particles were washed with a medium (medium A) containing 400 mM sucrose, 20 mM NaCl, 20 mM CaCl 2 , and 20 mM Mes-NaOH (pH 6.0) supplemented with 10% (w/v) polyethylene glycol 6000 and suspended in medium A after extensive washing with medium A.
Protein Composition-The PS II core particles were solubilized using 1% SDS and then electrophoresed in an SDS-PAGE with a 16 -22% gradient gel containing 7.5 M urea (32). A sample corresponding to 0.8 g of Chl was applied to each lane. Peptide bands were visualized by staining with Coomassie Brilliant Blue R-250. The apparent molecular mass of a resolved protein was estimated by comigrating a molecular mass standard (Bio-Rad).
Measurements-Mid-frequency (1800 -1000 cm Ϫ1 ) FTIR spectra were recorded on a Bruker IFS 66v/S spectrophotometer equipped with a mercury cadmium telluride detector (EG&G Optoelectronics D316/6). Low frequency (650 -350 cm Ϫ1 ) FTIR spectra were recorded on a Bomen MB102 spectrophotometer equipped with a Si bolometer (Infrared, HDL-5) as previously described (31). The PS II core suspension (ϳ40 g of Chl) was mixed with 1 l of sodium ferricyanide solution (100 mM stock) as an electron acceptor. The sample suspensions deposited on either a 20-mm BaF 2 disk (mid-frequency) or an AgCl disk (low frequency) were partially desiccated and then rehydrated. The sample temperature was maintained within Ϯ 0.03°C using a homemade cryostat. The 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 for 15 s in the mid-frequency (20 scans) or in the low frequency region (10 scans) before and after excitation, and light Ϫ dark spectra were calculated. 116 -145 mid-frequency difference spectra or 271-284 low frequency difference spectra were averaged.
Low temperature X-band ESR spectra were measured using a Bruker E580 spectrometer equipped with an Oxford-900 cryostat and a temperature controller (Oxford, ITC4). The sample cores (4 mg of Chl/ ml) in a Spracil quartz ESR sample tube were illuminated at 213 K for 3 s with a cold light (Hayashi, LA-150TX) passing through a long pass filter (Ն680 nm). Thermoluminescence was measured using a homemade apparatus (33). The cells were suspended in medium A at 250 g of Chl/ml in the presence or absence of 0.1 mM DCMU. The sample suspension was illuminated at 0°C with a saturating xenon flash. The O 2 evolution activity was measured using a Clark-type oxygen electrode in medium A at 25°C under saturating light conditions supplemented with exogenous electron acceptors, 1 mM 2,5-dimethyl-1,4-benzoquinone, and 2 mM potassium ferricyanide for cells or 4 mM potassium ferricyanide for core particles. Table I, the histidine-tagged wild-type (wild*-type) and the C-terminal extension-truncated Ala-344-stop cells grew photoautotrophically under both low and high light conditions and evolved oxygen at rates similar to the wild-type cells. The affinity-purified PS II core particles from wild*-type and Ala-344-stop cells showed high O 2 evolution capability, although the activity of the Ala-344-stop cores was slightly lower than that of the wild*-type cores, following the trend observed for O 2 evolution in the cells. The A344G-stop cells grew photoautotrophically under low light conditions as previously reported (11) and evolved oxygen at ϳ90% of the rate of the Ala-344-stop cells but did not grow photoautotrophically under high light conditions. The PS II core particles from the low light-grown A344G-stop cells preserved high O 2 evolution activity, but it was relatively lower (ϳ60%) than that from the Ala-344-stop cells. The activity was not enhanced by the further supplementation of the Ca 2ϩ and/or Cl Ϫ to medium A. Fig. 1 shows SDS-PAGE profiles of the PS II core particles from wild*-type (lane a), Ala-344-stop (lane b), and A344G-stop (lane c). All the core particles showed very similar protein composition in terms of the major intrinsic proteins, including CP47, CP43, D2, D1, and ␣ subunit of cytochrome b 559 , and the three extrinsic proteins, including 33 kDa, cytochrome c 550 , and 12 kDa. The particles also showed very similar profiles in the low molecular mass-region bands (Ͻ10 kDa), although each band was diffusive and could not be individually defined. The results demonstrated that neither deletion of the C-terminal extension nor replacement of the C-terminal alanine with glycine affects the protein composition of OEC. A protein band at ϳ15 kDa (lane c, asterisk) was sometimes detected with a slightly higher amount in the A344G-stop than the other cores, although the identity of the band was not defined in this study.

Physiological and Biochemical Properties-As shown in
Redox Properties of OEC-The effects of mutations on the redox properties of OEC were studied by measuring the thermoluminescence glow curve in the presence (panel A) or absence (panel B) of DCMU as shown in Fig. 2. The wild-type cells (a) showed a 37°C (ϪDCMU) and a 10°C band (ϩDCMU) because of the charge recombination of the S 2 Q B Ϫ and S 2 Q A Ϫ pair, respectively. Very similar thermoluminescence glow curves were observed in the wild*-type (b) and Ala-344-stop (c) cells. The results indicate that the attachment of the histidine tag and the deletion of the C-terminal extension do not affect the redox potential of the S 2 -state OEC and, thus, do not affect the Mn cluster. In contrast, the peak temperatures of the respective bands in the A344G-stop cells (d) were upshifted by ϳ5°C for the S 2 Q B Ϫ and S 2 Q A Ϫ pair, respectively, indicating that the redox potential of the Mn cluster for the S 2 /S 1 couple in the A344G-stop OEC is lower than that of the control OEC.
Structural Properties of the Mn Cluster- Fig. 3 shows the light-induced ESR spectra in the PS II core particles from the Ala-344-stop (a) and A344G-stop (b) cells. The sample cores were illuminated at 213 K, a temperature at which the accumulation of the S 2 state is allowed. The Ala-344-stop particles showed a prominent g ϭ 2 S 2 multiline, a much smaller g ϭ 4.1 S 2 signal, and a Fe 2ϩ Q A Ϫ signal at g ϭ 1.9. The spectral features of the observed S 2 ESR signals were very similar to those in the normal OEC. The A344G-stop spectrum showed a smaller multiline and much larger g ϭ 4.1 signals compared with the Ala-344-stop spectrum, although the hyperfine structure of the multiline signal and the width and position of the g ϭ 4.1 signal were almost identical to those of the Ala-344-stop spectrum. These results indicate that the Gly substitution in-

FIG. 2. Thermoluminescence glow curves of wild-type (a), wild*-type (b), Ala-344-stop (c), and A344G-stop (d) cells of Synechocystis sp. PCC 6803 in the presence (A) or absence (B) of 0.1 mM DCMU for generating S 2 Q A
؊ or S 2 Q B ؊ pair. Wild*-type represents a wild-type strain with a histidine tag. duces changes in the Mn cluster that facilitate the g ϭ 4.1 state formation. Fig. 4 shows the mid-frequency (1800 -1000 cm Ϫ1 ) S 2 /S 1 FTIR difference spectra of the PS II core particles from the wild*-type (a), Ala-344-stop (b, blue line), and A344G-stop (b, red line) cells. The wild*-type and Ala-344-stop spectra showed largely identical S 2 /S 1 vibrational features, including the symmetric (1450 -1300 cm Ϫ1 ) and asymmetric (1600 -1500 cm Ϫ1 ) stretching modes from the putative carboxylate ligands for the Mn cluster as well as the amide I (1700 -1600 cm Ϫ1 ) and II (1600 -1500 cm Ϫ1 ) modes from polypeptide backbones (5,31). In contrast, the A344G-stop spectrum (red line) showed a small but distinctive difference from the Ala-344-stop one (blue line) as can be clearly observed in the double difference spectrum (c) obtained by subtracting the A344G-stop from the Ala-344-stop spectrum. Several of the bands in the symmetric carboxylate stretching region (1450 -1300 cm Ϫ1 ) were shifted in the A344Gstop S 2 /S 1 spectrum to yield the bands at 1403(Ϫ), 1394(ϩ), 1360(ϩ), and 1345(Ϫ) cm Ϫ1 , in spite of no change of the S 2 /S 1 bands at 1437(ϩ) and 1417(Ϫ) cm Ϫ1 . In the 1600 -1500-cm Ϫ1 region, the bands were yielded at 1565(Ϫ), 1552(ϩ), 1521(Ϫ), and 1503(ϩ) cm Ϫ1 , which are ascribed to the changes in the asymmetric carboxylate stretching modes and/or amide II modes of the polypeptide backbones. The S 2 /S 1 bands at 1652(ϩ), 1642(Ϫ), and 1620(ϩ) cm Ϫ1 in the amide I region decreased in their intensity by Gly substitution to yield the 1653(ϩ)-, 1642(Ϫ)-, and 1623(ϩ)-cm Ϫ1 bands in the double difference spectrum. The negative band at 1113 cm Ϫ1 assigned to the CN stretching mode of the putative histidine ligand for the Mn cluster (6) was considerably enhanced in the A344Gstop spectrum, as indicated by a distinct positive band in the double difference spectrum. Notably, no mode for the Y D tyrosine was included in the double difference spectrum, indicating little contribution of the Mn-depleted PSII to the A344G-stop spectrum. Fig. 5 shows the low frequency (670 -350 cm Ϫ1 ) S 2 /S 1 FTIR difference spectra of the PS II core particles from the Ala-344stop (a, blue line) and A344G-stop (a, red line) cells and the double difference spectrum (b). The Ala-344-stop S 2 /S 1 spectrum showed prominent bands at 629(ϩ), 617(Ϫ), 606(ϩ), 590(ϩ), 577(Ϫ), 403(Ϫ), and 388(Ϫ) cm Ϫ1 as well as many other medium to low intensity bands. The 590(ϩ)-and ϳ400(Ϫ)-cm Ϫ1 bands were ascribed to the vibrational modes of ferrocyanide and ferricyanide, respectively (31). The spectrum was most comparable with the previously reported wild*-type spectrum measured under different sample conditions, but bands at 660(ϩ), 652(Ϫ), 642(ϩ), 374(ϩ), 368(Ϫ), and 359(ϩ) cm Ϫ1 had not been well resolved because of the large absorption of water (31). The spectrum of A344G-stop (a, red line) was markedly different from that of Ala-344-stop. The double difference spectrum (b) showed prominent bands at 663(ϩ), 648(ϩ), 625(Ϫ), 615(Ϫ), 606(ϩ), 384(ϩ), 368(Ϫ), and 361(ϩ) cm Ϫ1 as well as several minor bands at 532 cm Ϫ1 and in the 500 -420-cm Ϫ1 region. Based on a study using 18 O water, it has been suggested that the Mn-O-Mn cluster modes in the S 2 and S 1 states are responsible for the 606(ϩ)-and 625(Ϫ)-cm Ϫ1 bands (34). Therefore, the results indicate that the interactions between Mn ions in the cluster are considerably influenced by the Gly substitution. However, the Gly substitution was not observed to affect the bands at 577(Ϫ), 565(ϩ), 554(ϩ), and 542(ϩ) cm Ϫ1 . The 577(Ϫ)-cm Ϫ1 band has been ascribed to the skeletal vibration of the Mn cluster or the Mn-ligand(oxygen) interaction (31). Therefore, the primary structure of the Mn cluster is thought to be little affected by the Gly substitution. Double difference spectrum (c) was obtained by subtracting the A344Gstop spectrum from the Ala-344-stop spectrum after normalization with respect to the peak-to-peak intensity of the 2115(Ϫ)-cm Ϫ1 terricyanide and 2034(ϩ)-cm Ϫ1 ferrocyanide bands (31). A dark Ϫ dark spectrum (d) is presented to show the noise level.
FIG. 5. Low frequency (670 -350 cm Ϫ1 ) S 2 /S 1 FTIR difference spectra of the PS II core particles from the Ala-344-stop (a, blue line) and A344G-stop (a, red line) cells of Synechocystis sp. PCC 6803. The double difference spectrum (b) was obtained by subtracting the A344G-stop spectrum from the Ala-344-stop spectrum after normalization with respect to the peak-to-peak intensity of the ferricyanide band near 400(Ϫ) cm Ϫ1 and the ferrocyanide band at 590(ϩ) cm Ϫ1 (31). A dark Ϫ dark spectrum (c) is presented to show the noise level.

DISCUSSION
The present results showed that the A344G-stop mutant grew photoautotrophically with a high O 2 evolution activity and revealed no biochemical difference in the isolated PS II core complex. However, the mutant cells could not grow photoautotrophically under high light conditions. The elevated peak temperature of the S 2 -state thermoluminescence bands in the A344G-stop cells indicates that the oxidation potential of the S 2 -state Mn cluster is lower in the mutant than in the control cells. This may account for the increased susceptibility of the A344G-stop cells to high light because this change lowers the efficiency for the oxidation of the substrate water and, consequently, leads to the increase in the charge recombination probability in PS II, promoting the generation of the reactive oxygen species.
The Ala-344-stop Ϫ A344G-stop double difference spectrum (Fig. 4c) showed the bands at 1403-1345 and 1565-1503 cm Ϫ1 in the carboxylate symmetric and asymmetric stretching regions, respectively. The band change induced by the Gly substitution is explained by assuming a shift of a single carboxylate group. As shown in Fig. 6, the observed double difference spectrum was reproduced by assuming the downshift of an Ala-344-stop S 2 /S 1 band pair at 1360/1403 cm Ϫ1 (a, blue line) to 1346/1395 cm Ϫ1 (a, red line) or the downshift of an S 2 /S 1 band pair at 1392/1404 cm Ϫ1 (b, blue line) to 1345/1360 cm Ϫ1 (b, red line). The latter (scheme b) is less likely because the presumed S 2 bands were much smaller relative to the presumed S 1 bands. This is, however, quite difficult to rationally explain. Alternatively, the observed band change was reproduced by the upshift of an S 2 /S 1 band pair from 1395/1346 (or 1360/1345) cm Ϫ1 to 1403/1360 (or 1404/1392) cm Ϫ1 , although the presumed shift opposes the expected trends of the ligand (5,35). It was indicated that the symmetric stretching modes of the ␣-carboxylate group of the D1 C-terminal Ala-344 appear at ϳ1356 cm Ϫ1 for the S 1 state and at ϳ1320 cm Ϫ1 for the S 2 state (30), respectively. The downshift of the Ala-344 band has been interpreted to indicate the ligation of C-terminal carboxylate to the Mn ion, which is oxidized during the S 1 to S 2 transition. The lack of the C-terminal bands in the double difference spectrum shown in Fig. 4 indicates that the Gly substitution induces little change in the modes for the C-terminal carboxylate but results in a change of the other carboxylate ligand.
A possible candidate of the Gly-affected modes is the carboxylate side chain mode of D1-Asp-342. This residue was assigned as a ligand for a Mn ion in the 3.5-Å x-ray structural model (19) in which the CH 3 group of Ala-344 is arranged toward the Mn cluster. Therefore, the replacement of CH 3 (Ala) with H (Gly) can influence the strength of the coordination bond between the Asp-342 carboxylate and a Mn ion. The absence of the Gly substitution effect on the C-terminal carboxylate modes may be related to the finding that the C-terminal carboxylate was disordered and not visible in the x-ray electron density map (19). Possibly, the C-terminal ligand is relatively flexible to be able to move to some extent with little change in the ligation geometry. It is of note in this context that the observed frequency difference between the asymmetric and symmetric modes (ϳ160 cm Ϫ1 ) for the putative Asp-342 carboxylate is similar to the value empirically found for the bidentate coordinate (5,30,38), although the x-ray model indicated a unidentate coordinate of Asp-342 (19). According to Fig. 6a, the putative Asp-342 carboxylate band is downshifted by 43 cm Ϫ1 upon the S 2 -state formation, suggesting oxidation of the ligating Mn ion or a change in ligation structure during the S 1 to S 2 transition. Taking into account the possible ligation of the Ala-344 C-terminal carboxylate to the Mn ion that changes in oxidation state during the S 1 to S 2 transition (30), it may be assumed that the carboxylate side chain of Asp-342 and the ␣-carboxylate of Ala-344 participate in the ligation of the Mn ion that is oxidized upon the S 2 formation.
Most of the bands in the low frequency S 2 /S 1 difference spectrum were changed upon Gly substitution as seen in Fig. 5. The affected bands possibly included the S 2 /S 1 modes of the Mn-O-Mn stretching vibration of the Mn cluster at 606(ϩ)/ 625(Ϫ) cm Ϫ1 (34), Mn-COO Ϫ bending modes of putative carboxylate ligands (36), the ring torsion mode of histidine residues (37), the amide IV (40% O ϭ C-N bending) and amide VI (C ϭ O bending) modes of polypeptide backbone (38), and ligand-dependent Mn-O modes at 532-420 cm Ϫ1 (31,39). The Gly substitution resulted in marked changes in the bands in the 400 -350-cm Ϫ1 region, frequencies that include the stretching vibrations between the Mn ion and the ligands (O and/or N) (40). These marked spectral changes upon Gly substitution suggest a gross structural change in the Mn cluster. Nevertheless, there are relatively insignificant changes in the midfrequency region of the spectrum except for the putative Asp-342 mode. Interestingly, the D1-D170H OEC Synechocystis showed prominent changes in the low frequency S 2 /S 1 spectrum (660 -500 cm Ϫ1 ) but limited changes in the mid-frequency (1800 -1200 cm Ϫ1 ) spectrum (20). A possible explanation for the different manifestation of the Gly substitution in the midand low frequency spectra is that the mid-frequency bands include the modes, which are not the direct ligands to the Mn cluster. Such indirect structural coupling may be less sensitive to intrastructural changes of the Mn cluster. The observed structural changes of the Mn cluster induced by the Gly substitution suggest that Ala-344 is crucial for maintaining the intrastructure of the Mn cluster as a ligand. Nevertheless, the present FTIR results do not preclude the possibility that the C-terminal Ala-344 carboxylate is located close to the Ca ion of the Mn 3 CaO 4 cubane-like cluster core as proposed by the recent x-ray structural model (19). In this case, the present results suggest that Ca may participate in controlling the structural changes of the Mn cluster during the S 1 to S 2 transition and the C-terminal Ala is involved in this Ca-dependent function, although it is not clear how it is achieved at present. As shown in Fig. 3, ϳ70% of the total A344G-stop OEC existed in the S ϭ 5/2 g ϭ 4.1 state, whereas more than 90% OEC existed in the S ϭ 1/2 multiline state in the control Ala-344-stop. Some difference in the electronic structure within the Mn cluster has been proposed to be responsible for the appearance of these two states (41,42). The change in the intrastructure of the Mn cluster may induce considerable alterations in the low frequency modes of the Mn cluster with much smaller or little change in the mid-frequency modes indirectly coupled with the Mn cluster. The Sr 2ϩ -substituted OEC with an enhanced g ϭ 4.1 signal showed the normal-like mid-frequency S 2 /S 1 difference spectrum (43) and the markedly altered low frequency spectrum (34). Furthermore, little change of the mid-frequency S 2 /S 1 spectrum has been reported upon conversion from the multiline to the g ϭ 4.1 state by IR illumination (44). A possible difference between these two S 2 ESR states is a valence exchange between strongly antiferromagnetic Mn(IV) and Mn(III) (41,42). The valence exchange may alter the Mn-Mn and/or Mn-ligand interactions that affect the low frequency modes but scarcely influence the mid-frequency modes.