The Structure and Unusual pH Dependence of Plastocyanin from the Fern Dryopteris crassirhizoma

Spectroscopic properties, amino acid sequence, electron transfer kinetics, and crystal structures of the oxidized (at 1.7 Å resolution) and reduced form (at 1.8 Å resolution) of a novel plastocyanin from the fern Dryopteris crassirhizoma are presented. Kinetic studies show that the reduced form ofDryopteris plastocyanin remains redox-active at low pH, under conditions where the oxidation of the reduced form of other plastocyanins is inhibited by the protonation of a solvent-exposed active site residue, His87 (equivalent to His90in Dryopteris plastocyanin). The x-ray crystal structure analysis of Dryopteris plastocyanin reveals π-π stacking between Phe12 and His90, suggesting that the active site is uniquely protected against inactivation. Like higher plant plastocyanins, Dryopteris plastocyanin has an acidic patch, but this patch is located closer to the solvent-exposed active site His residue, and the total number of acidic residues is smaller. In the reactions of Dryopteris plastocyanin with inorganic redox reagents, the acidic patch (the “remote” site) and the hydrophobic patch surrounding His90 (the “adjacent” site) are equally efficient for electron transfer. These results indicate the significance of the lack of protonation at the active site of Dryopteris plastocyanin, the equivalence of the two electron transfer sites in this protein, and a possibility of obtaining a novel insight into the photosynthetic electron transfer system of the first vascular plant fern, including its molecular evolutionary aspects. This is the first report on the characterization of plastocyanin and the first three-dimensional protein structure from fern plant.

Plastocyanin is an electron transfer protein having a single copper atom at the active site. Plastocyanin functions as an electron carrier protein between the cytochrome b 6 f complex and P700 ϩ in oxygenic photosynthesis (1). Plastocyanin indicates intense absorption band due to the S Cys 3 Cu(II) charge transfer at visible region and a narrow hyperfine coupling constant in the EPR spectra of the oxidized form (2). The electronic structure of plastocyanin has been reported by several groups (3)(4)(5). Solomon and co-workers (6) gave an implication for the electron transfer reaction mechanisms on the basis of the electronic structure of plastocyanin. The Cu(II) half-occupied highest occupied molecular orbital is the redox active orbital in blue copper proteins and plays a key role in the electron transfer reaction.
Plastocyanin consists of 97-105 amino acid residues. The x-ray crystallographic structures of poplar (7)(8)(9), Enteromorpha prolifera (10), and Chlamydomonas reinhardtii (11) plastocyanins have been determined, and the solution structures of the proteins from Anabaena variabilis (12), parsley (13), French bean (14), and Scenedesmus obliquus (15) have been given by NMR spectroscopy. The copper atom is located at the loop region of eight-stranded ␤-barrel structure and beneath the protein surface at a depth of 5-10 Å, with two histidines, one cysteine, and one methionine as ligand groups.
Two regions on the surface of the plastocyanin molecule have been discussed as potential binding sites for electron transfer partners: an acidic patch (the "remote" site) and a hydrophobic patch (the "adjacent" site). Small inorganic complexes such as [Fe(CN) 6 ] 3Ϫ and [Co(phen) 3 ] 3ϩ (where phen indicates 1,10-phenanthroline) 1 have been used as redox probes. A great deal of evidence from kinetics measurements suggests that the adjacent and remote sites are the sites of electron transfer to [Fe(CN) 6 ] 3Ϫ and [Co(phen) 3 ] 3ϩ , respectively (16). In higher plant plastocyanins, the adjacent site consists of conserved hydrophobic amino acid residues surrounding the solvent-exposed His 87 , and the remote site is an acidic patch comprising Asp 42 , Glu 43 , Asp 44 , Glu 59 , Glu 60 , Asp 61 , and Glu 68 as well as the invariant residue Tyr 83 (9). Electron transfer reaction studies of the nitrated plastocyanin at the remote site Tyr 83 indicated the possible interaction with cytochrome f at the site of Tyr 83 (17). Site-directed mutagenesis studies of pea plastocyanin has demonstrated that the solvent-exposed Tyr 83 is essential for binding and electron transfer reactions (18). From these reaction studies and comparative investigations using the acidic patch mutants, it has been suggested that a physiological electron donor, cytochrome f, might interact with the acidic patch of plastocyanin. Qin and Kostic (19,20) have suggested that the rearrangement of the electron transfer site between * This work was supported by Grants-in-Aid for Science Research on Priority Areas from the Ministry of Education, Science, Sports and Culture, Japan, by the Joint Studies of Program (1997) of the Institute for Molecular Science, and by the Grant-in-Aid for the Ground Experiment for the Space Utilization from the Japan Space Forum and National Space Developments Agency of Japan to Kohzuma. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The atomic coordinates and structure factors (codes 1KDJ and 1KDI) have been deposited in the Protein Data Bank, Brookhaven National Laboratory, Upton, NY.
plastocyanin and cytochrome c (as a model of cytochrome f ) causes the gating of the electron transfer. Very recently, computational simulation also indicates that the rearrangement is optimized through the -cation interaction between the phenol moiety of the remote electron transfer site Tyr 83 and the ⑀amino group of lysine side chain of cytochrome f to the effective electron transfer interaction mode (21)(22)(23). Recent solution structure analysis of spinach plastocyanin indicates that the complex formation with cytochrome f dominantly occurs through the hydrophobic moiety of plastocyanin (24). Plastocyanin is distributed among wide variety of oxygenic photosynthetic organisms involving cyanobacteria, green algae, and higher plants. The expression level of plastocyanin in a part of cyanobacteria and green algae is regulated by the levels of copper concentration in the culture medium (25). Although the presence of plastocyanin in species of fern has been recognized (26), no plastocyanin from a fern has been characterized previously. A novel plastocyanin has now been isolated and purified from the fern Dryopteris crassirhizoma.
We here report the isolation, spectroscopic properties, amino acid sequence, electron transfer kinetics, and crystal structures of the oxidized (at 1.7 Å resolution) and reduced form (at 1.8 Å resolution) of a novel plastocyanin from the fern D. crassirhizoma. This is the first report on the characterization of plastocyanin from ferns and also the first three-dimensional protein structure from ferns.

EXPERIMENTAL PROCEDURES
Isolation and Purification of Plastocyanin-Leaves of D. crassirhizoma Nakai, collected on a mountain in Nagano Prefecture, Japan, were homogenized in 0.2 M potassium phosphate buffer (pH 8.0) with an automatic mortar. The homogenate was centrifuged to remove debris and subjected to ammonium sulfate fractionation (0.3-1.0 saturation), ion exchange column chromatography with DEAE-cellulose and DEAE-Sephacel (Amersham Pharmacia Biotech), hydrophobic chromatography with Butyl-Toyopearl 650 M (Tosoh), and gel filtration with Bio-Gel P-10 (Bio-Rad). Protein homogeneity was checked with analytical SDSpolyacrylamide gel electrophoresis.
Amino Acid Sequence Determination-The complete amino acid sequence of Dryopteris plastocyanin was determined as described (27), except that carboxymethyl-plastocyanin was subjected to lysyl endopep-tidase digestion (molar ratio of enzyme/substrate ϭ 1/400, in 0.01 M Tris-HCl buffer, pH 8.6, at 37°C for 6 h) or cyanogen bromide cleavage (molar ratio of chemical/protein ϭ 500, in 70% formic acid at room temperature for 20 h), and an Applied Biosystems 473A protein sequencer was used. The sequences of plastocyanin and peptides isolated were confirmed by amino acid analysis.
Sequence Alignment-Amino acid sequence data were obtained from the data base SWISS-PROT release 36 (28), and each sequence was aligned by using CLUSTAL X version 1.64b (29). The putative processing sites are those shown in the data base or the reference cited therein.
Spectroscopic Measurements-UV-visible spectra were determined on a Beckman DU-7500 recording spectrophotometer. Resonance Raman spectra were measured at room temperature with a spinning cell (1800 rpm; 5-mm diameter), and Raman shifts were calibrated to an accuracy of 1 cm Ϫ1 using indene and carbon tetrachloride. Resonance Raman scattering was excited at 607 nm by a rhodamine 6G dye laser (Spectra Physics, model 375B) pumped by an Ar ϩ ion laser (Spectra Physics, model 2017) and an Astromed CCD 3200 detector attached to a single monochromator (Ritsu Oyo Kogaku, DG-1000). The laser power was adjusted to 60 mW at the sampling point.
Kinetic Measurements-All reactions were monitored on an Otsuka Electronics (Osaka, Japan) Photal RA-401 stopped flow spectrophotometer. Electron transfer reactions of plastocyanin with inorganic complexes were monitored at 590 nm at 25°C. All kinetic parameters were calculated using the program IgorPro (WaveMetrics, Lake Oswego, OR). The kinetic data were analyzed using the following equation (16).
where the acid dissociation constant K a and the rate constants are as defined in Equations 2-4.
Electrochemical Measurements-Cyclic voltammetry was carried out using a BAS model CV-27 Voltammograph (Bioanalytical Systems Inc.). Modification of gold electrode was described in the previous report (30). 2,2Ј-Diethylaminoethanethiol (Aldrich) was used for the modification of the electrode as a promoter. A single-compartment electrochemical cell was used with an Ag/AgCl reference electrode (Bioanalytical Systems Inc.), and a platinum wire counter electrode was separated by a vicor glass tip from the working solution. The electrode potential was calibrated with the potential of [Co(phen) 3 ] 2ϩ/3ϩ couple. Oxygen was removed from the working compartment by passing humidified O 2 -free argon through the electrochemical cell for 15 min.
X-ray Crystallography-The hanging drop vapor diffusion method was applied for the crystallization of plastocyanin using ammonium sulfate as the precipitant. A hanging drop comprising a 6-l droplet of a 10 mg/ml protein solution in 0.1 M sodium acetate buffer (pH 4.5) and 32% saturated ammonium sulfate was suspended above 500 l of a reservoir solution comprising the same buffer, 1 mM sodium azide and 64 -65% saturated ammonium sulfate. Single crystals with maximum dimensions 0.4 ϫ 0.4 ؋ 0.2 mm 3 were obtained. All processes were carried out at 20°C. The reduced form was prepared by soaking a crystal in the mother liquid containing 10 mM sodium ascorbate. The  characteristic blue color disappeared after 10 min. Intensity data for both the oxidized and reduced forms were collected from single crystals on a Rigaku RAXIS-IIc imaging plate, using CuK ␣ radiation from a rotating anode x-ray generator, Rigaku RU-300 with fine-focused beam and ␤-filtered (40 kV, 100 mA). The unit cell was determined to be hexagonal with a ϭ b ϭ 73.15 Å and c ϭ 31.10 Å by autoindexing software on the RAXIS. The data were reduced in Laue group 6/m, and reflections with (I/) Ͼ 1 were accepted. The asymmetric unit in space group P6 1 includes one plastocyanin molecule (molecular mass ϭ 10,000 dalton) with a V m value of 2.4 Å 3 /dalton (31). From the value of V m the estimated solvent content is 49%. Intensity data sets were initially collected up to 2.0 Å resolution. A second intensity data set for reduced form was collected up to 1.8 Å resolution. Among 28,202 accepted observations up to 1.8 Å resolution, 8,478 independent reflections were obtained with an R merge of 6.1% and a completeness of 93.5%. A second set of intensity data for the oxidized form was collected at ϭ 1.00Å with synchrotron radiation at the Photon Factory using Sakabe's Weisenberg camera for macromolecules (32). Among 69,003 accepted observations up to 1.7 Å resolution, 10,220 independent reflections were obtained with an R merge of 8.9% and a completeness of 96%. The crystal structure of oxidized form was solved by the molecular replacement method with the program of AMoRe (33) in CCP4 program package (34). The molecular structure of plastocyanin from poplar (35) was used as the starting model. Model rebuilding was performed with the program FRODO (35). Refinement of the oxidized structure was carried out at 1.7 Å resolution by the simulated annealing refinement method (X-PLOR (34)). Water molecules were added in three steps by using the WATPEAK program in CCP4 (34). The current structure of oxidized plastocyanin includes 758 protein atoms (nonhydrogen), one metal ion and 47 water molecules. After rebuilding and several cycles of refinement, the R-factor and free R-factor are 24.1 and 28.8% for 9,943 unique reflections in the range of 6.0 to 1.7Å. The structure of reduced plastocyanin including 38 water molecules is refined up to 1.8 Å resolution with an R-factor and a free R-factor of 20.7 and 22.6%, respectively.

RESULTS AND DISCUSSION
Plastocynain from D. crassirhizoma consists of 102 amino acid residues. The complete amino acid sequence of the plastocyanin has been determined (Fig. 1). The amino acid sequence of D. crassirhizoma plastocyanin is much different from those of the seed plant proteins so far reported (Fig. 2). Percentage divergences between the fern sequence and each of the seed plant sequences are 62-67%; in contrast, those among the seed plant sequences are 2-37%. Because the sequence for D. crassirhizoma is identical to that for another fern, Polystichum longifrons, except for 4 amino acid residues, 2 fern plastocyanin sequences might be similar to each other.
Dryopteris plastocyanin has a significantly different electronic absorption spectrum from the usual higher plant plas-2 Y. Nagai and F. Yoshizaki, unpublished results. and the intensity of the absorption band at 463 nm is higher. This might reflect a different electronic structure of the active site. Fig. 4 shows a resonance Raman spectrum of Dryopteris plastocyanin obtained by excitation at 590 nm. Resonance Raman spectra of blue copper proteins including plastocyanin excited in resonance with the S Cys Ϫ 3 Cu 2ϩ charge transfer band characteristically have one or two Raman bands in the 250 cm Ϫ1 region and multiple Raman bands in the 330 -490 cm Ϫ1 region. The former and latter are associated with Cu-N His and Cu-S Cys moieties, respectively (36,37). Very recently, the lower frequency band at 267 cm Ϫ1 has been assigned to the Cu-N His in poplar plastocyanin by Dong and Spiro (38). Dryopteris plastocyanin has a unique spectrum in the Cu-S Cys region, with two strong peaks at 426 and 381 cm Ϫ1 , and additional small peaks at 491, 447, 410, 392, 367, and 339 cm Ϫ1 . According to the assignment in poplar plastocyanin (38), the 426 and 381 cm Ϫ1 bands may arise from the coupling of the modes of (Cu-S Cys ), ␦(C ␤ C ␣ N), and ␦(C ␤ C ␣ S), and remains are also assignable to internal modes of Cys residue and/or the coupling with (Cu-S Cys ) (38). There is also an isolated Cu-N His mode close to 262 cm Ϫ1 . This might be contributed from the stretching between copper and coordinated imidazole nitrogen atom of solvent exposed His 90 . The overall similarity of the resonance Raman spectrum to those of blue copper proteins suggests that the coordination at the copper site of Dryopteris plastocyanin is generally similar to that in higher plant plastocyanins (38,39).
Two regions on the surface of the plastocyanin molecule have been discussed as potential binding sites for electron transfer partners: an acidic patch (the remote site) and a hydrophobic patch (the adjacent site). A great deal of evidence from kinetics measurements suggests that the adjacent and remote sites are the sites of electron transfer to [Fe(CN) 6 ] 3Ϫ and [Co(phen) 3 ] 3ϩ , respectively (16). The calculated second-order rate constants (pH 7.5) for the oxidation of Dryopteris plastocyanin by [Fe(CN) 6 ] 3Ϫ and [Co(phen) 3 ] 3ϩ are 1.87 (Ϯ 0.02) ϫ 10 5 and 1.69 (Ϯ 0.04) ؋ 10 3 M Ϫ1 s Ϫ1 , respectively. The electron transfer rate constants for the reaction with [Fe(CN) 6 ] 3Ϫ and [Co(phen) 3 ] 3ϩ seem to be almost identical to the values for higher plant plastocyanins. The crystal structure of Dryopteris plastocyanin has been determined at a resolution of 1.7 Å (oxidized) and 1.8 Å (reduced). The root mean square difference between the backbone structures of poplar and Dryopteris plastocyanins is 0.74 Å. The structure of Dryopteris plastocyanin demonstrates the migration of the acidic patch (but not the invariant Tyr 86 ) toward the adjacent site (Fig. 5). The disappearance of the acidic residues from the remote site reduces the electrostatic repulsion for negatively charged [Fe(CN) 6 ] 3Ϫ , thus facilitating electron transfer to [Fe(CN) 6 ] 3Ϫ at the remote site. The electron transfer rate constants, similar to those for higher plant plas-tocyanins (16) despite the significant structural changes, support the hypothesis that the remote site including Tyr 86 (corresponding to Tyr 83 for usual higher plants) and the adjacent site including His 90 (His 87 for usual higher plants) are the electron transfer site with [Fe(CN) 6 ] 3Ϫ and [Co(phen) 3 ] 3ϩ , respectively. The kinetic results also indicate that the remote and adjacent sites in Dryopteris plastocyanin are equally efficient for electron transfer, as predicted by the electronic structure analysis (3,40).
The pH dependence of the rate constants for the reactions of various plastocyanins with [Fe(CN) 6 ] 3Ϫ and [Co(phen) 3 ] 3ϩ complexes have been reviewed (16). The variation with pH of the second-order rate constants (k sec ) for the oxidation of Dryopteris plastocyanin by [Fe(CN) 6 ] 3Ϫ and [Co(phen) 3 ] 3ϩ complexes is illustrated in Fig. 6. The acid dissociation constants, pK a , associated with oxidation by [Fe(CN) 6 ] 3Ϫ and [Co-(phen) 3 ] 3ϩ are 5.9 Ϯ 0.1 and 6.2 Ϯ 0.3, respectively. Studies of the electron transfer kinetics and crystal structure of plastocyanin as functions of pH have suggested that electron transfer is inhibited at low pH by the protonation of the active site His 87 (16). In contrast, Dryopteris plastocyanin does not become redox-inactive under acidic conditions. This suggests that in Dryopteris plastocyanin the active site His 90 (His 87 in higher plants) does not become protonated. In the structure of Dryopteris plastocyanin, the imidazole ring of the coordinated His 90 is stacked at a van der Waals' contact distance against the phenyl group of Phe 12 (Fig. 7). We suggest that thestacking interaction stabilizes the Cu-N(His 90 ) bond, inhibits the rotation of the His 90 imidazole ring, and thus prevents the imidazole group from becoming protonated at acidic pH.
The structure of the reduced protein indicates only 0.15 Å root mean square deviation from the oxidized conformation. The bonding parameters for the oxidized (pH 4.5) and reduced (pH 4.5) forms strongly support the lack of protonation at the active site of Dryopteris plastocyanin in contrast to higher plant plastocyanins where significant structural changes are caused by the dissociation of His 87 (His 90 for Dryopteris) from the copper atom (Table I). A stacking interaction between the coordinated imidazole ligand of histamine and the phenyl group of phenylalanine was demonstrated in the previous studies of the stability and structure of copper complexes (41). Yamauchi and co-workers (42) have pointed out that the stacking interaction induces stronger imidazole-copper bonding due to the delocalization of electron density on the copper center. By the comparison of electronic absorption spectra of usual seed plant plastocyanins, the electronic absorption spectrum of Dryopteris plastocyanin indicates the 7 nm blue-shifted charge transfer band, and an additional absorption band at 410 nm is recognized in the electronic absorption spectrum (see above). The resonance Raman excitation profile of poplar plastocyanin  has suggested that the charge transfer band from His 1 to Cu 2ϩ should be lying into the most intense absorption band around at 600 nm (38). On the other hand, the charge transfer band between stacked Phe 12 and His 90 residues would be expected around 300 -400 nm region (42). The differences of electronic absorption spectra between Dryopteris and usual higher plant plastocyanins may reflect the stacking structure on the active site electronic structures. Cyclic voltammetry was performed for Dryopteris plastocyanin in the potential range ϩ600 -0 mV versus normal hydrogen electrode, and showed a well defined quasi-reversible Faradaic response with a peak-to-peak separation ⌬E p of 95mV at a diethylaminoethanethiol modified gold (DEAET/Au) electrode at pH 7.0 (data not shown). The redox potential of Dryopteris plastocyanin has been determined to be 387 mV versus normal hydrogen electrode at pH 7.0. This value is approximately 20 mV higher than the values for most other plastocyanins (360 -370 mV versus normal hydrogen electrode), although it was expected that a lower redox potential due to negatively charged acidic amino acid residues surrounding the hydrophobic part of Dryopteris plastocyanin would be observed as a decrease in the redox potential in the mutant plastocyanins (43) and azurin (44) substituted with acidic amino residues near the copper center. Despite the several acidic amino acid residues located near the copper center (adjacent site) in the case of Dryopteris plastocyanin, the redox potential appears at a higher potential. The remarkable positive value of the redox potential may reflect the reduction of electron density on the copper center through thestacking interaction between the coordinated imidazole ligand of His 90 and the phenyl group of Phe 12 . The pH dependence of the reduction potential of the protein displays at least one acid-base equilibrium (data not shown). At the lower pH values the behavior is different from that observed for other plastocyanins (16,45) and pseudoazurin (30), which indicate protonation of solvent-exposed histidine residue located at the active center. In the usual higher plant plastocyanins, the differences of redox poteintial, ⌬E between pH 5 and pH 7 have been evaluated to be 50 -60 mV, but the corresponding ⌬E value of Dryopteris plastocyanin has been estimated to be only 18 mV. It is most likely that the independence of the redox potential on pH also reflects the stacking interaction between Phe 12 and His 90 preventing the protonation of the active site His 90 .
Sykes suggested that the pK a values obtained from the kinetic experiments using [Co(phen) 3 ] 3ϩ are due to contributions arising from both the acidic patch and the active site protonation (16). The acid dissociation constant, pK a ϭ 6.2 Ϯ 0.3, obtained from the kinetic studies of Dryopteris plastocyanin would reflect the pure protonation process at the acidic patch, because the redox reaction proceeds without active site protonation. The affinity of cationic [Co(phen) 3 ] 3ϩ for Dryopteris plastocyanin should be decreased by the partial neutralization of the negative charge on the protein molecule, which may explain the lowering of the activity at low pH (Fig. 6).
These results indicate the significance of the lack of protonation at the active site of Dryopteris plastocyanin, the equivalence of the two electron transfer sites in this protein, and a possibility of obtaining a novel insight into the photosynthetic electron transfer system of the first vascular plant fern, including its molecular evolutionary aspects.