Isolation and Characterization of Monomeric and Dimeric CP47-Reaction Center Photosystem II Complexes*

Using the detergents n-dodecyl β-d-maltoside and heptyl thioglycopyranoside, a subcore complex of photosystem II (PSII) has been isolated that contains the chlorophyll-binding protein, CP47, and the reaction center components, D1, D2, and cytochrome b 559. We have found, by using sucrose density centrifugation, that the resulting preparation consisted of a mixture of dimeric and monomeric forms of the CP47 reaction center (RC) complex, having molecular masses of 410 ± 30 and 200 ± 28 kDa, respectively, as estimated by size exclusion chromatography. The level of the dimer in the preparation is significantly higher than the monomeric form. Both the monomer and dimer contain the proteins CP47, D1, and D2 and the α- and β-subunits of cytochrome b 559. Analyses by mass spectrometry and N-terminal sequencing showed that both forms of the CP47-RC complex contain the products of the psbI,psbT c (chloroplast gene), and psbWwith molecular masses of 4195.5, 3849.6, and 5927.4 Da, respectively. In contrast to the monomeric form, the CP47-RC dimer contained two extra proteins with low molecular weights, identified as the products of the psbL and psbK genes having molecular masses of 4365.5 and 4292.1, respectively. It was also found that the dimer contained slightly more molecules of chlorophyll a(21 ± 2.5) than the monomer (18 ± 1.5), a characteristic also observed in the room temperature absorption spectrum by comparing the ratio of absorption at 416 and 435 nm. Of particular note was the finding that the dimer, but not the monomer, contained plastoquinone-9 (estimated to be 1.5 ± 0.3 molecules per RC). The results indicate that the CP47-RC monomer is derived from the dimeric form of the complex, and therefore the latter is likely to represent anin vivo conformation. The PsbTc as well as the PsbI and PsbW proteins are identified as being intimately associated with the D1 and D2 proteins, and in the case of the dimer, importance is placed on the PsbL and PsbK proteins in sustaining plastoquinone binding and maintenance of the dimeric organization. Assuming only one copy of the α- and β-subunits of cytochromeb 559, the monomeric and dimeric forms of the complex would be expected to contain 21 and 23 × 2 transmembrane helices, respectively.

The light-induced splitting of water and the consequential release of dioxygen is a fundamental reaction in photosynthesis taking place in all higher plants, algae, and cyanobacteria. This reaction is catalyzed by a complex known as photosystem II (PSII) 1 which is embedded in the lipid bilayer of the thylakoid membrane. The PSII complex is composed of at least 25 different proteins and binds a large number of pigments (1). At the heart of this complex is the reaction center (RC), consisting of the D1 and D2 proteins, where primary charge separation occurs (2). Closely associated with the D1 and D2 heterodimer are the two chlorophyll a-binding proteins, CP47 and CP43, and a range of small hydrophobic polypeptides including the ␣and ␤-subunits of cytochrome b 559 . CP47 and CP43 act both as an "inner" light harvesting system and as a conduit for the transfer of energy captured by an "outer" light harvesting system composed either of proteins that bind chlorophyll a and b (in the case of higher plants and green algae) or phycobilisomes (in the case of cyanobacteria and red algae) (3). The outer antennae can be removed during isolation procedures to obtain a "core" preparation consisting of CP47, CP43, D1, and D2 proteins and other polypeptides including those that are extrinsically bound to the luminal surface and required for stabilizing and optimizing the water splitting system and its catalytic cluster of manganese atoms (4). Further solubilization of PSII to produce a "CP47-RC subcore" results in the loss of water splitting activity, in part due to the removal of extrinsic proteins and the manganese cluster but may also be due to the stripping away of CP43 (5).
Recently, the structures of several isolated forms of PSII have been reported (6 -9). Using single particle analysis by electron microscopy, the top and side views of a large (725-kDa) PSII particle isolated from spinach have been described at about 25-Å resolution (6). This particle, which is dimeric, contains the PSII core and a complement of chlorophyll a/b-binding proteins (LHCII (Lhcb1 and -2), CP29 (Lhcb6), and CP26 (Lhcb5)), totaling about 100 chlorophyll molecules/RC (7). Since this LHCII-PSII supercomplex is isolated using a mild detergent treatment and maintains high oxygen evolving activity, it seems highly likely that it represents an in vivo form of PSII (7). This would therefore suggest that functional PSII is dimeric in vivo and would account for the fact that the core complex can also be isolated in a dimeric form by stripping away the chlorophyll a/b proteins (8). Indeed the dimeric form of the PSII core complex isolated from spinach is more functionally active and stable than the monomeric form (7).
A PSII structure that has recently been revealed to 8-Å resolution is that of the CP47-RC subcore complex (9). This structure has been obtained by two-dimensional crystallization and cryoelectron microscopy, and the published work was in the form of a projection map. By comparison with x-ray crystallography data of the reaction center of purple bacteria and of photosystem one, it was possible to suggest that some density in the map was due to the transmembrane helices of the D1 and D2 proteins and CP47, while the remaining density results from other proteins present. It was also found that the unit cell in the two-dimensional crystals contained two dimers of the CP47-RC complex and that the size of the dimer was compatible with a central location in the LHCII⅐PSII supercomplex (9).
In order to complement the structural analyses of the twodimensional crystals of CP47-RC, we have undertaken studies to determine the composition and properties of the isolated solubilized CP47-RC complex.

Isolation of CP47-RC Complexs
CP47-RC complexes were obtained from BBY-type PSII-enriched membranes isolated from market spinach with modifications described by Hankamer and co-workers (6). After washing with 50 mM Tris-HCl (pH 9.0), 1 mg of Chl ml Ϫ1 of BBYs were solubilized with n-dodecyl ␤-D-maltoside (DM; 1% final concentration) and heptyl thioglycopyranoside (HTG; 2.7% final concentration) in the dark with constant stirring for 30 min at room temperature. The nonsolubilized material was removed by a 30-min centrifugation at 40,000 ϫ g at 4°C. For purification the solubilized BBYs were loaded onto a Fractogel ion exchange column (DEAE-Toyopearl 650S, TSK) that had been equilibrated with 50 mM Tris-HCl (pH 7.3), 10 mM NaCl, and 2 mM DM. The sample was washed with 50 mM Tris-HCl (pH 7.3), 2 mM DM, and 40 mM NaCl and was eluted with the same buffer but using a 40 -200 mM NaCl gradient. The ion exchange chromatography was carried out at 4°C under dim light conditions.

Isolation of Monomeric and Dimeric CP47-RC Complexes
Monomeric and dimeric CP47-RC complexes were separated using sucrose gradient centrifugation. The sucrose gradient mix solution consisted of 25 mM Mes, pH 6.5, 0.5 M sucrose, 10 mM NaCl, 5 mM CaCl 2 and 0.03% DM. The sucrose density gradients were prepared after freezing at Ϫ20°C and slow thawing at 4°C. The ion exchange chromatographypurified CP47-RC complex (300 g of Chl) was loaded onto the gradients and centrifuged at 90,000 ϫ g for 16 h at 4°C in a Beckman SW41 swing out rotor and separated into two different fractions.

Room Temperature Absorption Spectra
These were measured using an SLM Aminco (Urbana, IL) model DW2000 spectrophotometer.
HPLC Size Exclusion Analysis-HPLC size exclusion analysis was carried out using a Zorbax GF-450 column 9.4/250-mm (Jones Chromatography). The mobile phase consisted of 50 mM Tris-HCl, pH 7.2, 0.03% DM, 0.3 M sucrose and was passed through the column at a rate of 0.5 ml min Ϫ1 . The 20-l samples injected contained 5 g of Chl, and the profiles were monitored at 418 nm. The molecular weights of the complexes were determined by using a gel filtration calibration kit (Amersham Pharmacia Biotech).

SDS-PAGE and Western Blotting
The polypeptide compositions of the isolated PSII preparations were analyzed by gradient SDS-PAGE (10 -17% polyacrylamide) containing 6 M urea, essentially using the method of Laemmli (10). The gels were stained with Coomassie Blue R-250.
The protein profiles resolved by SDS-PAGE were transferred onto nitrocellulose (11) and immunolabeled with CP47-or CP43-specific antibodies raised against electrophoretically purified spinach CP47, CP43, D2 (a kind gift from Dr. R. Barbato), or C-terminal D1-specific antibody (Dupont 304) raised against a synthetic polypeptide, homologous to the 29 amino acids of the C-terminal pea D1 precursor (a kind gift from Dr. P. Nixon). Biotinylated anti-rabbit IgG was used as a secondary antibody and was labeled with Extravidine-alkaline phosphatase conjugate (Sigma). The chromogenic substrates were 5-bromo-4-chloroindolyl phosphate -toluidine salt and nitro blue tetrazolium chloride.

Reverse Phase HPLC Pigment Analysis
Pigment Analysis-Pigments (chlorophyll a, pheophytin a, and ␤-carotene) were extracted into 80% (v/v) acetone/H 2 O at 4°C in dim light conditions. The samples were vortexed for 30 s, centrifuged for 2 min in a microcentrifuge in order to separate out proteinaceous materials, and filtered through a 0.2-m (pore size) membrane (polyvinylidene difluoride; Vatman) before injection (injection volume was 20 l). The pigments were resolved using an ODS-1 Spherisorb column (Anachem) and isocratic elution with methanol/ethyl acetate/water 68:30:2 (v/v/v), at a flow rate of 1 ml/min. We used a variable wavelength detector (Kontron 30), detecting simultaneously at 663 nm (for Chl a and Pheo a) and 453 nm (for ␤-carotene) or 255 nm (for PQ). Chl a and ␤-carotene quantification was performed after calibration of the corresponding peaks with pigment standards (purchased from Sigma), whose concentrations were determined using extinction coefficients 76.79 mM Ϫ1 cm Ϫ1 at 663.6 nm in 80% acetone for Chl a (12) and 139 mM Ϫ1 cm Ϫ1 at 452 nm in 100% hexane for ␤-carotene (13). The Pheo a standard used for calibration was produced by acidifying the Chl a standard with 2 mM HCl (14). PQ-9 (obtained from Sigma) was dissolved in 100% ethanol, and its concentration was determined using the extinction coefficient of 15.2 M Ϫ1 cm Ϫ1 at 255 nm (15).
Separation of Proteins-Partial separation of the protein subunits in the CP47-RC monomer and dimer complexes was carried out for subsequent mass spectrometric measurements and was achieved using a Spherisorb Aquapore RP-300 (220 ϫ 4.6-mm) column fitted to a Kontron HPLC system (Datasystem 450, HPLC pump 420, detector 430 and mixer M800). The PSII preparations were dialyzed against 2 ϫ 2.0 liters of aqueous 5% acetic acid at 4°C for 16 -24 h and were loaded directly onto the column. The components were eluted with a solvent system of A (aqueous 0.1% trifluoroacetic acid) and B (90% acetonitrile in aqueous 0.1% trifluoroacetic acid), using a linear gradient from 100% A to 40% B in 10 min followed by an increase to 100% B over 60 min at a flow rate of 1.0 ml min Ϫ1 . Elution was monitored at 214 and 280 nm, and fractions were collected at 1-min intervals.

Gas Phase Edman Sequencing
Peptides were Edman-degraded using a CI 4000 gas phase protein sequencer, and the standard protein sequencing method was used (16). Samples that failed to give sequence information were treated on the glass fiber disc with 0.6 N HCl at 25°C for 24 h.

Electrospray Mass Spectrometry Analysis
Intact proteins of the CP47-RC complexes and their digestion products were also analyzed by electrospray mass spectrometry. Spectra were obtained by direct injection of 10-l aliquots of HPLC-purified proteins into the ion source of a VG Bio-Q triple quadrupole electrospray mass spectrometer. The instrument was operated using a carrier flow buffer of 90% acetonitrile in aqueous 0.1% trifluoroacetic acid, propan-1-ol and 2-methoxyethanol in a 1:1:1:1 (v/v/v/v) ratio at a flow rate of 5 l min Ϫ1 . The mass spectrometer was calibrated using a solution of horse heart myoglobin (1 pM ml Ϫ1 ) to give an average chemical molecular mass for the signals observed.

Matrix-assisted Laser Desorption Ionization Mass Spectrometry Analysis
Matrix-assisted laser desorption ionization mass spectra were acquired using a Fisons VG ZAB 2SE 2FPD mass spectrometer, fitted with a UV laser (337 nm). The instrument was calibrated using CsI clusters. Data acquisition and processing were performed using VG Analytical Opus software. The samples were dissolved in 60% propan-1-ol in 5% aqueous acetic acid, and 1-l aliquots were loaded onto a probe that had been treated with a saturated solution of 2,5-dihydroxybenzoic acid made up in a 70:30 (v/v) solution of 90% acetonitrile in aqueous 0.1% trifluoroacetic acid and aqueous 0.1% trifluoroacetic acid, respectively.

Isolation of CP47-RC Subcore Complex and Fractionation to Its Monomeric and Dimeric Forms-After
Tris washing (pH 9.0) the PSII-enriched membranes, in order to remove the extrinsic proteins of the oxygen-evolving complex (33-, 23-, and 17-kDa polypeptides), the sample was solubilized with a mixture of two nonionic detergents, DM and HTG (see "Materials and Methods"). The CP47-RC complex was separated from the solubilized PSII components (mainly Lhcb1 to -6 proteins and CP43) using ion exchange chromatography. The elution profile comprised a single peak for CP47-RC (eluted with 120 -130 mM NaCl). Subjecting the isolated CP47-RC to size exclusion HPLC showed that the preparation was heterogeneous. Therefore, for further purification, the ion exchange-purified CP47-RC samples were subjected to sucrose density gradient centrifugation. As Fig. 1 shows, the sample separated into two main fractions (fraction 1 and fraction 2), with fraction 2 being the most dominant. The size exclusion HPLC data showed that both fractions from the sucrose gradient were homogenous with molecular masses estimated to be 200 Ϯ 28 kDa (fraction 1) and 410 Ϯ 30 kDa (fraction 2) (Fig. 2). The SDS-PAGE profiles of the two populations ( Fig. 3) were similar, containing CP47, D1, D2, and cytochrome b 559 . The polypeptide profiles were probed with CP47-and CP43-specific antibodies (Fig. 3). The two bands in the 40 -50-kDa region cross-reacted with anti-CP47. No cross-reaction with anti-CP43 was observed (the antibody was checked for its effectiveness on BBY-type membranes). The presence of D1 and D2 protein was also verified by immunoblotting (data not shown). The subunit composition and molecular mass values of fractions 1 and 2 suggest that the 200-and 410-kDa complexes are monomeric and dimeric forms of the CP47-RC subcore complex.
As clearly seen in Fig. 1, the yield of monomeric CP47-RC complexes was significantly lower than that of dimeric CP47-RC. A second sucrose density gradient centrifugation of the dimeric CP47-RC after solubilization with 2% HTG at 0.5 mg ml Ϫ1 Chl resolved the complex to the same fraction 1 and fraction 2 (data not shown). This result suggested that the monomeric form of the CP47-RC complex is due to solubilization of the dimeric CP47-RC rather than vice versa.
In order to detect and characterize the small molecular weight subunits associated with the complexes, detailed analyses by reverse phase HPLC and mass spectrometry were carried out.
Small Molecular Weight Subunits Associated with Monomeric and Dimeric CP47-RC Subcore Complexes-The analysis of protein components in membrane multisubunit complexes, such as the CP47-RC complex, have traditionally been performed by SDS-polyacrylamide gel electrophoresis (e.g. Fig. 3). While this technique is valuable, its application can be limited due to poor resolution, especially for proteins below 15 kDa. Recently, we reported the development of a complementary reverse phase HPLC purification protocol that provides an effective method for fractionating the low molecular weight subunits of the PSII reaction center complex (16). The products isolated by this procedure are directly amenable to mass spectrometric analyses and N-terminal amino acid sequencing, and together these techniques are usually sufficient to provide information on the primary structure.
The optimized reverse phase HPLC profiles generated for CP47-RC monomer and dimer preparations are presented in Fig. 4. In the monomer samples, there appear to be three main absorbance peaks; the first two components, which elute at retention times of 24.6 and 40.5 min, were separated to a near base-line resolution, while the third UV peak at 53.9 min is only partially deconvoluted. Direct matrix-assisted laser desorption ionization mass spectrometry and electrospray mass spectrometry analyses of the HPLC fractions corresponding to these UV absorbance regions identified the presence of five components with molecular masses ranging from 3849.6 to 9255.1 Da (Table I). The precise molecular masses obtained for three of these small subunits, 4409.1, 9255.1, and 5927.4 Da, which were observed in fractions 24, 40, and 53/54, respec-  Fig. 1). The 20-l samples contained 5 g of Chl a, and the profiles were monitored at 418 nm. The molecular masses of fraction 1 (200 Ϯ 28 kDa) and fraction 2 (410 Ϯ 30 kDa) were estimated by using a gel filtration calibration kit (Amersham Pharmacia Biotech).

FIG. 3. Coomassie-stained SDS-PAGE and immunoblots of monomeric CP47-RC (lanes 1) and dimeric CP47-RC (lanes 2).
The loading amount for the SDS-PAGE was 6 g of Chl a, and for the immunoblotting it was 1 g of Chl a. The C-terminal antibody used for detecting the D1 protein was first reported in Ref. 31 and was a kind gift from Dr. Peter Nixon.
tively, are in close correlation (Ϯ0.3 Da) to those calculated from gene derived protein sequences and allowing for the known N-terminal post-translational modifications for the ␤-subunit of cytochrome b 559 , the ␣-subunit of cytochrome b 559 , and the PsbW proteins, respectively.
In an earlier study, we had identified a N-terminal modification entailing the removal of the initiating formyl-methionine modification followed by an acetylation of the amino terminus of the second threonine residue on the ␤-subunit of cytochrome b 559 (16). In the case of the ␣-subunit of cytochrome b 559 and the PsbW proteins, N-terminal amino acid sequencing data had revealed that the initiating formyl-methionine residue of these subunits had been proteolytically processed (16 -19). The presence of these modifications and confirmation of our assignments were obtained by N-terminal amino acid sequencing (Table I). Sequence information initiating at the second residue was obtained for both the ␣-subunit of cytochrome b 559 and the PsbW proteins, whereas in the case of the ␤-subunit of cytochrome b 559 sequence data were only obtained after acid hydrolysis treatment required to unblock the N terminus of this protein.
The two additional components detected in the monomeric form of the CP47-RC complex, eluting predominantly in fractions 52 and 53, have been assigned to the PsbT c (chloroplastencoded) and PsbI proteins, respectively. Both subunits were found to be N-terminally blocked and had to be treated by acid hydrolysis in order to generate the sequence data given in Table I. The molecular mass calculated for the spinach PsbT c from its gene-derived protein sequence is 3821.7 Da; however, the molecular mass measured for this subunit was 3849.6 Da. The 28-atomic mass unit shift detected here can be attributed to a formyl group and is expected to correlate to the previously unidentified N-terminal blocking group detected on this subunit. Unlike the other proteins characterized so far, the psbI gene of spinach chloroplasts has not been fully sequenced. However, a partial sequence (residues 1-20) for this component had been obtained by N-terminal amino acid sequencing (17). Comparisons of protein sequences for the PsbI subunits of higher plants revealed an extremely high homology; in fact, the tobacco, mustard, and wheat sequences are identical. Moreover, the calculated molecular mass attributed to the tobacco, mustard, and wheat PsbI proteins, 4167 Da, differs by 28 atomic mass units from the molecular mass we have obtained for the corresponding spinach protein, 4195.5 Da. This mass shift can be attributed to the formyl N-terminal blocking group, a modification we have previously characterized in the PsbI proteins of pea plants (16).
The HPLC profile obtained with the CP47-RC dimer preparation shown in Fig. 4 indicates the presence of four well resolved absorbance peaks. As can be seen, all of the UV absorbance peaks observed in the monomer samples, eluting at approximately 24, 40, and 52/53 min are also present in the dimer preparation. These UV absorbance components could thus be attributed by reference to the monomer samples to the ␤-subunit of cytochrome b 559 ; the ␣-subunit of cytochrome b 559 ; and the PsbT c , PsbI, and PsbW proteins, respectively. Indeed, mass spectrometric analyses of the relevant HPLC fractions confirms these assignments (Table I). However, in contrast to the monomer samples, the HPLC chromatogram of the dimer samples reveals the presence of an additional UV absorbance peak at 37 min and a further shouldering at the broad UV absorbance at 53/54 min. These UV components were subsequently assigned to the PsbL and PsbK subunits, respectively. The molecular mass attributed to the protein isolated in fraction 37, 4365.5 Da, correlates with that expected for the PsbL protein using its gene-derived protein sequence and the previously elucidated formyl-methionine processing site, 4366.0 Da (18). This assignment was confirmed by N-terminal amino acid sequencing (Table I). The protein sequence obtained for the principle component eluting in fraction 54 was found to correspond with that of the PsbK subunit (Table I) complete psbK gene sequence from spinach chloroplasts has not been elucidated. It is anticipated, however, that there will be a high degree of sequence homology for this subunit among higher plant species. The molecular mass we have obtained for the purified PsbK protein from spinach plants is 4292.1 Da, whereas that calculated from the gene-derived sequence of tobacco plants using the known N-terminal processing site (signal peptide 1-24) would be 4285.2 Da (20). Some of our preliminary data have indicated that this mass discrepancy is likely to correspond to two amino acid changes; the first is anticipated to be an Asn to Ser change at residue 34, and the other is probably an Ile to Phe change at residue 39, although these modifications are presently being confirmed.
Room Temperature Absorption Spectra and Pigment Stoichiometry- Fig. 5 shows the room temperature absorption spectra of CP43/CP47/RC core complex, isolated according to Hankamer et al. The most distinctive difference shown in Fig. 5 is that the absorption spectra of dimeric and monomeric CP47-RC subcore complexes possess more a pronounced 416-nm band than that of the CP43-CP47-RC complex, reflecting a lower Chl a content due to the absence of CP43. Indeed, the ratios of the absorbancies at 416 and 435 nm, which reflect the chlorophyll levels per reaction center, are 0.80, 0.89, and 0.92 for CP43-CP47-RC, dimeric CP47-RC, and monomeric CP47-RC complexes, respectively (see Table II). The same tendency was shown by the HPLC pigment analyses (Table II). CP43-CP47-RC core complex and dimeric and monomeric CP47-RC subcore complex contained 35, 21, and 18 Chl a molecules per reaction center (it has been assumed that each RC binds two molecules of Pheo a).
Since the D1-D2-cytochrome b 559 RC complex contains six Chl

TABLE I
Analysis of the small polypeptides of monomeric and dimeric forms of the CP47-RC complex The N-terminal amino acid sequences and molecular masses of the HPLC-purified components are shown. The calculated molecular mass values were determined from the gene-derived protein sequences and known post-translational modifications. The sequences marked by an asterisk correspond to peptides that were N-terminally blocked and had to be treated by acid hydrolysis (0.6 N HCl for 24 h) to obtain sequence information; the components in boldface-type represent the differences observed between the monomer and dimer preparations. a molecules per two Pheo a molecules (14,(21)(22)(23), CP47 and CP43 should contain 12-15 Chl a molecules each. These numbers are in good agreement with the data published by Barbato et al. (24). The slightly different chlorophyll levels associated with monomeric and dimeric CP47-RC could be due to a removal of some peripheral Chl a molecules during the dimer to monomer conversion, which may be linked with the loss of the PsbL and PsbK proteins. The ␤-carotene levels, determined by reverse phase HPLC and calculated on the basis of two Pheo a molecules were 6.5, 5.3, and 4.6 molecules for CP43-CP47-RC cores and the two forms of CP47-RC, respectively (Table II). The most striking difference between the monomeric and dimeric CP47-RC was the absence in the former and the presence in the latter of plastoquinone (Table II). The dimeric CP47-RC complexes contained 1.5 Ϯ 0.3 molecules per two Pheo a.

DISCUSSION
The results presented here show that CP43 is more readily removed from the PSII core than CP47 despite the fact that both proteins are likely to be structurally rather similar based on the homologies between their amino acid sequences (3). Of particular importance is that we have discovered by our isolation procedure that the resulting CP47-RC complex is mainly in a dimeric state with a molecular mass estimated to be 410 Ϯ 30 kDa. We present several lines of evidence that the monomeric form of the complex is derived by dissociation of the dimer. This dissociation gives rise to the loss of PsbL and PsbK proteins and some chlorophyll. Of particular note is that the dimer to monomer conversion also results in the loss of bound plastoquinone-9.
The fact that the CP47-RC dimer was significantly more abundant than the monomeric form in our preparation, contained more protein subunits, and bound plastoquinone suggests that it represents a more native form of the complex. This conclusion is consistent with the finding that the largest and functionally most active forms of isolated PSII from spinach are dimeric (7). Moreover, the most up to date mapping of subunit positioning within PSII places CP43 toward the outer side of the central core dimer (4,8). Thus, its removal need not disturb the dimeric subcore complex composed of CP47-RC. It is noteworthy that a dimeric organization of CP47-RC was also observed in two-dimensional crystals, which have been analyzed recently by electron microscopy to 8 Å (9).
We have shown by a combination of reverse phase HPLC, mass spectrometry, and N-terminal sequencing that the monomeric and dimeric forms of CP47-RC contain a number of small polypeptides in addition to CP47 and the D1 and D2 proteins. Present in both forms of the complex are the ␣and ␤-subunits of cytochrome b 559 and proteins that are the products of the psbI, psbT c , and psbW genes. The presence of cytochrome b 559 polypeptides and the PsbI protein is expected, given that these subunits occur in the isolated reaction center complex, which is depleted of CP47 (16,17,25). That PsbW would be present is also expected based on the work of Schroeder and colleagues (19,26). This is a nuclear encoded gene not found in the genome of Synechocystis 6803. The psbT c gene is located on the same operon as the psbB gene (which encodes CP47) in the chloroplast genome. It had previously been known as the ycf8 gene, and its identification as a PSII gene has only recently been made using antibodies raised against its product (27). Here we confirm that this protein is indeed a PSII component and further show that it is in the CP47-RC complex and therefore located close to the D1-D2 heterodimer.
The fact that the dimer also contains the products of the psbL and psbK genes but that the monomer does not could be important for explaining the difference in their ability to bind functionally active plastoquinone as well as accounting for the slightly higher chlorophyll content. It has previously been argued that PsbL is involved in the functioning of Q A , possibly by stabilizing its binding to the D2 protein (28). In terms of the recent structural analysis of the CP47-RC complex (9), it is possible to estimate the number of transmembrane helices likely to be present. Assuming only one copy of the ␣and ␤-subunits of cytochrome b 559 (29) and given that the small subunits, PsbI, PsbT c , PsbW, PsbK, and PsbL, are all predicted to have one transmembrane helix, the monomer would contain 21 transmembrane helices and the dimer would contain 23. These numbers rely on the likelihood that the D1-D2 heterodimer and CP47 contain 10 and 6 transmembrane helices, respectively (1,4). If there are two copies of the cytochrome b 559 subunits as some advocate (30), then the total number of membrane-spanning segments would be 23 for the monomer and 25 for the dimer. In the two-dimensional crystals of CP47-RC recently analyzed to 8 Å by electron crystallography (9), the complex seems to be dimeric within the unit cell. Whether the dimer in the crystal is identical to that studied here is not certain, but if it is then the projection map obtained for the CP47-RC complex would be expected to contain at least 23 transmembrane helices.