Precise location of the Cu(II)-inhibitory binding site in higher plant and bacterial photosynthetic reaction centers as probed by light-induced absorption changes.

Light-dependent absorption change at 325 nm, ascribed to QA activity, was strongly reduced in the presence of Cu(II) in oxygen-evolving core complex. This change was much less affected in the presence of the herbicide 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU), indicating that the Cu(II)-binding site is different from that of the DCMU and that Cu(II) blocks QA reduction. Cu(II) did not eliminate the absorption change at 545 nm, ascribed to pheophytin reduction, in Na2S2O4-treated oxygen-evolving core and D1-D2-cytochrome b559 complexes. This indicates that Cu(II) does not affect the electron transport between P680 and pheophytin. Moreover, the activity of the bacterial reaction center probed by the absorption change at 790 nm was inhibited by Cu(II), but the signal at 530 nm, associated to the reduction of bacteriopheophytin in Na2S2O4-treated reaction center, was not inhibited. We conclude that Cu(II) impaired the photosynthetic electron transport between pheophytin and QA in both higher plants and photosynthetic bacteria. Cu(II) would bind to an amino acid(s) highly conserved in non-oxygenic and oxygenic reaction centers, which is(are) necessary for the electron transfer between pheophytin and QA. Based on the atomic structure of the bacterial reaction center several schemes of possible Cu(II) binding are shown.

The high pollution by heavy metals in the biosphere has led to an increasing attention to the effects of these toxic agents on living systems. Toxic levels of some of these heavy metals occur in natural and agricultural soils as a result of environmental pollution due to mining, smelting, manufacturing, agricultural, and waste disposal technologies (1). Among these metals, copper presents a high degree of toxicity (2). It is well known that Cu(I1) inhibits the photosystem * This work was supported by Direccibn General de Investigacibn Cientifica y Tecnica Grant PB-89-0060 and by Consejo Asesor de Investigacibn-Diputacibn General de Aragbn Grant PCB3-89. 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. I( To whom correspondence should be addressed. Tel.: 34-76-576511; Fax: 34-76-575620. (PS)' I1 electron transport in higher plants (3,4), green algae (5), and cyanobacteria (6). However, the precise location of the Cu(I1)-binding site and the underlying mechanism are a subject of debate. Recently, our group (7,8) has made some progress on this matter. We have probed that oxygen evolution by PS I1 membranes was inhibited by Cu(I1) when 2,6dichlorobenzoquinone (DCBQ) or ferricyanide, but not silicomolibdate, were used as electron acceptors. This indicated that Cu(I1) affected the reducing side of PS 11. Moreover, by using trypsin-treated thylakoids we demonstrated that Cu(I1)inhibitory site is located before the QB niche and close to the pheophytin-QA-iron domain of the PS I1 reaction center (RC).
We have also characterized the Cu(I1)-inhibitory mechanism based on measurements of oxygen evolution activity (8). This mechanism resulted non-competitive with respect to DCBQ and DCMU, and competitive with respect to protons. The non-competitive inhibition indicated that Cu(I1)-binding site is different from that of the DCBQ and DCMU sites, the QB niche. On the other hand, the competitive inhibition respect to protons may indicate that Cu(I1) interacts with an essential amino acid group(s) that can be protonated or deprotonated in the inhibitory-binding site.
The endogenous electron acceptors of PS I1 include two "bound" plastoquinone (PQ) molecules, QA and QB, and a pool of "free" PQ molecules that are present in the fluid lipid phase of the photosynthetic membranes (9)(10)(11). Upon illumination, the primary donor chlorophyll (P680) reduces QA via pheophytin, which in turn is re-oxidized by QB. The QB-produced by the transfer of one electron from QA-binds strongly to its binding site but BE2-, generated by a second turnover of QA-, is replaced by a "free" PQ molecule. Crystallization and x-ray studies on the RC from purple bacteria (12,13) together with the isolation of PS I1 RC from higher plants (14) have made possible the comparison between the oxygenic and non-oxygenic reaction centers. A high sequence homology has been revealed between the L and M polypeptides of purple bacterial RC and the D l and D2 polypeptides of the PS I1 RC, respectively (15)(16)(17). This high homology is more apparent in the primary donor and the quinone-iron domains. The photoreduction of QA and pheophytin is known to be accompanied by The abbreviations used are: PS, photosystem; D l and D2, polypeptides of the photosystem I1 reaction center; DCBQ, 2,6-dichlorobenzoquinone; DCMU, 3-(3,4-dichlorophenyl)-l,l-dimethylurea; IR, infrared; L, light polypeptide of the bacterial reaction center; LDAO, lauryl dimethylamine N-oxide; M, medium polypeptide of the bacterial reaction center; MES, 2-N-hydroxyethylpiperazine-N"2-ethanesulfonic acid; OECC, oxygen evolving core complex; OGP, l-o-noctyl-P-D-glucopyranoside; P680, primary donor chlorophyll; PQ, plastoquinone; QA, primary quinone acceptor; QB, secondary quinone acceptor; RC, reaction center; Z, primary electron donor. a bleaching of the absorption bands at 325 and 545 nm, respectively, in oxygen evolving core complex (OECC) (11,18,19) and PS I1 RC (20). In contrast, the photoreduction of bacteriopheophytin can be followed by detecting the absorption change around 530 nm in bacterial RC (21,22).
The goal of this work was to precisely locate the Cu(I1)inhibitory binding site. To this aim, we measured the lightdependent absorption changes at 325 and 545 nm in OECC and Dl-D2-cytochrome b559 complex from higher plants, and at 790 and 530 nm in bacterial chromatophores and RC, respectively, in the presence of Cu(I1). Several examples of possible Cu(I1) binding to conserved amino acids in PS I1 and bacterial RCs are also given based on the crystal structure of purple bacterial RC.

MATERIALS AND METHODS
Biological Material-Sugar beet (Beta vulgaris L. cv. Monohill) was grown hydroponically in a growth chamber in half-Hoagland nutrient solution, under 325 pE. m-'. s-' from fluorescent lamps at 25 "C, 80% humidity, and a 16-h light period.
Rhodospirillum rubrum S1 was grown semianaerobically in 2-liter bottles at 32 "C in the medium described by Cohen-Bazire et al. (23). Illumination was provided by two 150-watt lamps, and the bacteria were harvested after 3-4 days at the end of the logarithmic growth phase.
Oxygen Euoluing Core Complex (OECC) Isolation-Photosystem 11 membranes with high rate of oxygen evolution activity (i.e. 500 wmol of 02. h-'.mg chlorophyll" using DCBQ as electron acceptor) were prepared following the method of Berthold et al. (24) with some modifications as described in Ref. 7. Membranes were resuspended in 10 mM NaC1,5 mM MgC12, and 50 mM MES-NaOH (pH 6.0), and centrifuged at 121,000 X g for 40 min. The pellet was solubilized with 35 mM OGP, 0.4 M sucrose, 10 mM CaC12, 0.5 M NaC1, and 50 mM MES-NaOH (pH 6.0) at a chlorophyll concentration of 1.5 mg.ml". After incubation for 15 min at 4 "C, one part of the suspension was NaC1,lO mM CaCI2, and 50 mM MES-NaOH (pH 6.0), and centrifuged mixed with two parts of a solution containing 0.4 M sucrose, 0.5 M a t 40,000 X g for 90 min. The supernatant was desalted by a 90-min dialysis against a solution containing 0.4 M sucrose, 10 mM CaC12, and 50 mM MES-NaOH (pH 6.0) using a 50,000-kDa cut-off dialysis tube (Spectrapor). It was further diluted with the same buffer and subsequently centrifuged at 40,000 X g for 90 min. The pellet was resuspended in 0.4 M sucrose, 10 mM CaC12, and 50 mM MES-NaOH (pH 6.0), and stored at 77 K until use.
Photosystem II Reaction Center (PS II RC) Isolation-The D1-D2cytochrome b559 complex was isolated from highly active P S I1 membranes as in Nanba and Satoh (14) with some modifications (25) and detergent was exchanged from 0.05% Triton X-100 to 2 mM P-dodecyl maltoside (26,27). The material from the first chromatography column was diluted four times with 50 mM Tris-HC1 (pH 7.2), and loaded onto a small (1.6 X 3-cm) Fractogel TSK-DEAE 6 5 0 s anionexchange column (Merck). The column was then washed with 75 ml of 50 mM Tris-HC1 (pH 7.2) containing 2 mM P-dodecyl maltoside, and the Dl-D2-cytochrome b559 complex was eluted with 145 mM NaCl in the same buffer. All steps were carried out at 4 "C in the dark, and the sample was stored at 77 K until used.
Chromatophore Isolation-Chromatophores from R. rubrum S1 were obtained following the method described in Ref. 28. Cells were alumina-ground and centrifuged at 3,000 X g for 5 min to remove alumina and cell debris. The supernatant was subjected to a first differential centrifugation at 20,000 X g for 20 min and then at 100,000 X g for 60 min. The pellet of the latter centrifugation was resuspended in 50 mM Tris-HC1 (pH 8.0) or 50 mM buffer phosphate (pH 7.0).
Bacterial Reaction Center (RC) Isolation-The bacterial RC was obtained following the method described by Vadeboncoeur et al. (28). Chromatophores (Asso = 75) were resuspended (1:l v/v) in 50 mM buffer phosphate (pH 7.0), containing 0.7% lauryl dimethylamine Noxide (LDAO) and incubated in this solution for 1 h at 4 "C in the dark. The LDAO concentration was then brought to 0.1% by dilution with the same buffer, and the preparation was centrifuged at 105,000 X g for 90 min. The supernatant was carefully pipetted out to avoid contamination with the B880 antenna complex and subjected to (NH4),SO4 fractionation with 45% saturation. The precipitate was resuspended in 10 mM Tris-HC1 (pH 8.0) containing 0.03% Triton

Photosynthetic Reaction
Center 1685 X-100, and dialyzed to eliminate the (NH4)2S04. Finally, the RC complex was loaded onto a 8.5 X 0.5-cm Fractogel TSK-DEAE 6 5 0 s anion-exchange column (Merck). The column was washed with buffer 10 mM Tris-HC1 (pH 8.0) containing 0.1% Triton X-100 and the RC eluted with 160 mM NaCl in the same buffer. The temperature was maintained at 4 "C. Inhibition with Cu(II) and DCMU-OECC at a final chlorophyll concentration of 26.9 pg.ml" was resuspended in buffer containing 1 mM OGP, 0.4 M sucrose, 10 mM NaCl, 10 mM CaC12, and 50 mM MES-NaOH (pH 6.5). The Dl-D2-cytochrome b559 complex at an absorbance of 0.1 at 675.5 nm was in buffer 50 mM Tris-HC1 (pH 7.2). Chromatophores (A, = 1.80) were in 50 mM Tris-HC1 (pH 8.0) and bacterial RC (Am = 0.61) in 10 mM Tris-HCl (pH 8.0) and 0.1% Triton X-100. All the samples were preincubated in the dark with appropriate amounts of CuC12 and DCMU (see figure legends) at 4 "C for 10 min, with occasional shaking. After preincubation, the lightdependent absorption changes were measured.
Optical Spectroscopy-Light-induced absorption changes were measured in a Shimadzu 3000 double beam/dual wavelength spectrophotometer using 1-cm pathlength cuvette. Absorbance changes at 325 and 545 nm in OECC and Dl-D2-cytochrome b,,, complex were obtained in the dual wavelength mode using 360 and 560 nm as the corresponding reference wavelengths, respectively. The OECC sample was illuminated through an optical fiber and the Dl-D2-cytochrome b559 complex with a light projector using red-actinic light (filters Schott RG-665 plus KG-3). The photomultiplier tube was protected from scattered light by a 5-mm Schott blue BG-39 filter. The absorbance changes at 530 and 790 nm in bacterial RC and chromatophores were obtained using 555 and 700 nm as the reference wavelengths, respectively. The light-induced absorption changes at 530 nm were carried out with a light projector and those at 790 nm using an optical fiber. The filters used for the measurements at 530 nm with the bacterial RC were the same than that used with the Dl-D2-cytochrome b55, complex. For measurements at 790 nm, a blue-actinic light (filters Schott BG-18 plus KG-3) was used and the photomultiplier was protected by a 5-mm Schott red RG-665 filter. The cuvette was covered with parafilm foil in the measurements made in the presence of Na2S204.

RESULTS
To examine the Cu(1I) effect on the QA activity we measured the light-dependent absorbance change at 325 nm in OECC preparation in the presence of 80 ~L M CuC12 and compared with the control (no addition of inhibitor) (Fig. 1). This  (19). A marked reduction of the absorption change was observed in the presence of CuClz (Fig. 1B) compared to the control experiment (Fig. 1A). In contrast, the signal was less affected in the presence of 10 p~ DCMU (Fig. IC). Data published by Schatz and van Gorkom (11) showed that dif-  (Fig. 1A) was due to the loss of the contribution by the Q B -, that was eliminated in the presence of this herbicide. In this case, the signal decreased by about 45% respect to the control, which is consistent with published data (11). In the case of Cu(I1) the signal decreased by about 81% respect to the control that  (11). Similar residual signal was observed in the presence of CuCI2 in sample pre-treated with DCMU ( Fig. 4 0 ) , indicating again that the Cu(I1)-binding site is different to the DCMU-binding site and that it is located earlier in the electron transfer pathway. The concentrations of inhibitors used in the assays were 421-and 53-fold higher than that of the RC content in the OECC preparation for Cu(I1) and DCMU, respectively, considering 50 chlorophyll/RC in OECC (29). The inhibitor/RC ratio used in these measurements was similar to that used in our previous works based on the Cu(I1)-inhibition effect on the oxygen evolution activity by PS I1 membranes (7).
The Cu(I1) inhibition seems to be a reversible process. The optical absorption change at 325 nm, markedly reduced in the presence of Cu(I1) (Fig. 2 A ) compared to the change with no addition of inhibitor (Fig. lA), was recovered after the sample was centrifuged and washed twice with buffer to eliminate the Cu(I1) (Fig. 2 B ) . This result confirmed previous data reported by others (3).
It is well established that pheophytin presents a lightinduced reversible absorption change around 540 nm in the Dl-D2-cytochrome bSs9 complex (20) and OECC (11). In the later complex, QAand Q B -also present a maximum at 545 nm, however no appreciable contributions of Zhave been described (11). Fig. 3A shows the light-induced absorption change at 545 nm in OECC. A certain decrease of this change was observed in the presence of DCMU (Fig. 3B). Additional decrease but not complete elimination of the signal was observed with 80 p~ CuC12 (Fig. 3C). From Fig. 3 ( B and C) we calculated the intensity of the absorbance changes which corresponds to Q A -and Q B -and determined the absorbance ratio AA(QB-)/AA(QA-), resulting 0.32. This value corresponds well with that of 0.25 published by Schatz and van Gorkom (11). This result indicated that Cu(I1) eliminated the contributions by Q A -and QBat this wavelength. The remaining absorption change must be due to the reduction of pheophytin. In that sense, the absorption change at 545 nm was measured in OECC preincubated with Na2S204. This treatment induced chemically the reduction of both plasto-

FIG. 2. Effect of Cu(I1) on the light-induced absorption change at 325 nm by red actinic light at room temperature in OECC preparation pre-incubated with 80 PM CuClz (A) and after eliminating CuClz by washing twice with buffer ( B ) .
The buffer content, chlorophyll concentration and illumination conditions are described in Fig. 1. quinone molecules, thus eliminating the contributions by Q Aand QB-. The signal intensity (Fig. 4A) was strongly reduced by the treatment with Na2S204 (Fig. 4B). Interesting, addition of 80 p M cuCl2 to the pretreated sample did not result in a further decrease of the absorption change (Fig. 4C), as well as in the presence of DCMU (Fig. 4 0 ) . Note that the signal obtained in NazS2O4 treated sample (Fig. 4B) was similar to that observed with CuClz alone (Fig. 3C). The ratios AA(Na2S2O4+CuCl2/AA(Na2S2O4)andAA(Na2S2O4+DCMU)/ AA(NazS204) have a value close to 1. Therefore, these results confirmed that both plastoquinone molecules, Q A and Q B , were completely reduced by NazS2O4 and indicated that the residual signal obtained (Fig. 4, B-D) corresponds to that of reduced pheophytin. Moreover, this signal was not modified by Cu(II), indicating that this inhibitor has no effect on the reduction of pheophytin. Table I summarizes the intensity of all the light-induced absorption changes obtained with the OECC preparation.
To eliminate the possible interference of the chemical treatment on the Cu(I1)-inhibitory effect the absorption change at 545 nm in the Dl-D2-cytochrome b559 complex isolated from sugar beet was measured. This preparation is devoid of the secondary acceptors, Q A and Q B (14,30). Fig. 5 shows that the absorption change at this wavelength was not modified by Cu(I1). These results match well with the above mentioned data and confirm again that Cu(I1) inhibits the electron    Since a high homology between PS I1 and bacterial RCs has been described (15)(16)(17), we have also studied the Cu(I1) inhibition effect in chromatophores and RC preparation from R. rubrum. Fig. 6 depicts the effect of Cu(I1) on the light-induced absorption change at 790 nm in chromatophore preparation. This optical change due to a shift of the 800-nm band and associated with the RC activity (22) decreased with increasing amounts of CuC12. This result indicated that Cu(I1) inhibits the bacterial RC activity and eliminated the oxidizing side of PS I1 as possible Cu(I1)-inhibitory binding site, as suggested by Yruela et al. (7). From the inhibition curve, a value of 62.6 p M for the I,, was calculated. Note that the I,,/ RC ratio was 260, which corresponded well with that obtained for PS I1 membranes (15,/RC = 228) (7). To calculate the RC concentration in chromatophores and PS I1 membranes, we considered 28 bacteriochlorophyll/RC (31) and 250 chlorophyll/RC (24,29), respectively. Furthermore, Cu(I1) has no effect on the absorption change at 530 nm, associated to the reduction of bacteriopheophytin (21) in bacterial RC pretreated with Na2S204 (Fig. 7), indicating that both PS I1 and bacterial RCs present similar behavior in respect to the Cu(I1) inhibition effect.

DISCUSSION
Recently, we have probed that Cu(I1) impaired the photosynthetic electron transport on the reducing side of PS I1 at the level of the pheophytin-&-iron domain of the RC ( 7 ) , and we have characterized its inhibitory mechanism (8) mainly based on measurements of oxygen evolution activity. The aim of the present paper was to define precisely the location of the Cu(I1)-binding site. We provide data on the effect of CuC12 on the light-induced absorption changes associated with the reduction of the secondary acceptors, QA and QB, and pheophytin in higher plant OECC and D1-D2cytochrome b559 preparations, and in bacterial chromatophores and RC samples. Our results indicated that Cu(1I) blocks the reduction of QA and QB, but has no effect on the (bacteri0)pheophytin reduction. These data are consistent with our previous results ( 7 , 8 ) and imply that Cu(I1)-binding site is located after the (bacterio)pheophytin and before the quinone QA-binding site in both the PS I1 and the bacterial RCs. Moreover, the value of ISO/RC ratio was similar in both RCs, which indicates a high homology between the Cu(I1)inhibitory binding site in both types of RCs.
The presence of an essential amino acid(s) than can be protonated or deprotonated in the Cu(I1)-inhibitory binding site in the PS I1 RC with which Cu(I1) might interacts has been suggested by Yruela et al. (8). It is well established that His and Trp have a high affinity to bind copper. This is the basis for the immobilized metal-affinity chromatography for protein purification (32). This fact is also being exploited in protein engineering of metal-dependent enzyme activity (33).
It has been shown that Cu(I1) inhibits some proteases by coordinating to His residues in the active site (33). The presence of these amino acid residues close to QA site has been described in spinach PS I1 (17, 34) and bacterial (35)(36)(37) RCs. Protein modeling based on the crystal structure of Rhodopseudomonas viridis has suggested that some amino acid residues of QA-binding site in spinach PS I1 RC have homologues in the bacterial RC (17). Among these, it has been found that D2-Trp254 is homologous to M-TrpZ5' and some His residues close to QA site are present in both RCs (L-His"' and M-His217 in Rps. viridis, and D1-His215 and D2-His215 in spinach). The x-ray crystal structure of the RC from Rps. viridis has shown that M-TrpZ5' residue is positioned within van der Waals contact distance of both bacteriopheophytin and QA (35,36). The location of this amino acid residue with respect to bacteriopheophytin and QA has prompted speculation that it might be involved in promoting both tight binding of the quinone to the apoprotein (36) and fast electron transfer from bacteriopheophytin to QA by enhancing the electronic overlap between both molecules (38). This amino acid residue has also been found in other bacteria (M-TrpZ5' in Rhodobacter capsulatus and in Rb. sphaeroides) (37). Recently, works based in site directed mutagenesis in Rb. capsulatus have been done to dilucidate these questions and have probed that this amino acid residue affects the QA binding and the speed of its reduction which would imply that this residue is involved in the photosynthetic electron transport (39). The latter hypothesis has been proposed to explain the rapid rate of the electron transfer between bacteriopheophytin and QA, considering that the indole group of M-TrpZ5' might act as a bridge between the orbitals of the primary electron donor, bacteriopheophytin, and the secondary electron acceptor, QA (37).
From these data and our results, we conclude that the Cu(I1)-inhibitory binding site is similar in oxygenic and nonoxygenic RCs and propose that Cu(I1) might bind to a specific amino acid(s), His and/or Trp, located between (bacterio)pheophytin and QA. The metal ion could bind in such a way that disrupts the local conformation of the (bacterialpheophytin-QA domain and thus inhibiting the photosynthetic electron transport. Fig. 8 shows some possible Cu(I1)inhibitory binding sites in the bacteriopheophytin-QA domain of the RC based on the crystal structure of the RC from Rps. uiridis (16). involvinz the conserved amino acid residues M-M-ThrZz0, M-Trp250, and L-G1u'O4.