THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 262, No. 26, Issue of September 15, pp. 12676-12684,1987 Printed in U. S. A.

Identification of a Chloroplast-encoded 9-kDa Polypeptide as a 2[4Fe-4S] Protein Carrying Centers A and B of Photosystem I*

(Received for publication, April 13, 1987)

Peter Bordier HejS, Ib Svendsed, Henrik Vibe SchellerS, and Birger Lindberg MellerS From the $Department of Plant Physiology, Royal Veterinary and Agricultural University, 40 Thorvaldsemvej, DK-1871 Frederiksberg C and the §Department of Chemistry, Carlsberg Laboratory, 10 Gamk Carlsberg Vej, DK-2500 Valby, Denmark

An improved procedure is reported for large-scale preparation of photosystem I (PS-I) vesicles from thy- lakoid membranes of barley (Hordeum uulgare L.). The PS-I vesicles contain polypeptides of molecular masses 82, 18, 16, 14, and 9 kDa in an apparent molar ratio of 4:2:2:1:2. The 18-, 16-, and 9-kDa polypep- tides were purified to homogeneity after exposure of the PS-I vesicles to chaotropic agents. The isolated 9- kDa polypeptide binds 65-70% of the zero-valence sulfur of denatured PS-I vesicles, and the remaining 30-3570 is bound to P700-chlorophyll a-protein 1. The N-terminal amino acid sequence (29 residues) of the 9- kDa polypeptide was determined. Comparison with the nucleotide sequence of the chloroplast genome of Mar- chantiapolyrnorpha (Ohyama, K., Fukuzawa, H., Koh- chi, T., Shirai, H., Sano, T., Sano, S., Umesono, K., Shiki, Y., Takeuchi, M., Chang, Z., Aota, S.-i., Inoku- chi, H., and Ozeki, H. (1986) Nature 322, 572-574) and of Nicotiana tabucum (Shinozaki, K., Ohme, M., Tanaka, M., Wakasugi, T., Hayashida, N., Matsubay- ashi, T., Zaita, W., Chunwongse, J., Obokata, J., Ya- maguchi-Shinozaki, K., Ohto, C., Torazawa, K., Meng, B. Y., Sugita, M., Deno, H., Kamogashira, T., Yamada, K., Kusuda, J., Takaiwa, F., Kato, A., Tohdoh, N., Shimada, H., and Sugiura, M. (1986) EMBO J. 5, 2043-2049) identified the chloroplast gene encoding the 9-kDa polypeptide. We designate this gene psaC. The complete amino acid sequence deduced from the psaC gene identifies the 9-kDa PS-I polypeptide as a 2[4Fe-4S] protein. Since P700-chlorophyll a-protein 1 carries center X, the 9-kDa polypeptide carries centers A and B. A hydropathy plot permits specific identifi- cation of the cysteine residues which coordinate cen- ters A and B, respectively. Except for the loss of the N-terminal methionine residue, the primary transla- tion product of the psaC gene is not proteolytically processed.

P700-chlorophyll a-protein 1 binds 4 iron atoms and 4 molecules of acid-labile sulfide/molecule of P700. Each of the two apoproteins of P700-chlorophyll a- protein 1 contains the sequence Phe-Pro-Cys-Asp-Gly- Pro-Gly-Arg-Gly-Gly-Thr-Cys (Fish, L. E., Kuck, U., and Bogorad, L. (1985) J. Biol. Chem. 260, 1413- 1421). The stoichiometry of the component polypep- tides of PS-I indicates the presence of four copies of this sequence per molecule of P700. Center X may be

* This research was supported in part by grants from the Danish Agricultural Research Council, the Danish Natural Science Research Council, Dansk Investeringsfond, the Thomas B. Thriges Foundation, the Carlsberg Foundation, the Tuborg Foundation, and Stiftelsen Hofmansgave and by a Niels Bohr grant from the Royal Danish Academy of Sciences and Letters. 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 accord- ance with 18 U.S.C. Section 1734 solely to indicate this fact.

composed of two [2Fe-2S] centers bound to the 8 cys- teine residues contained in these four segments.

The photochemical transfer of electrons from reduced plas- tocyanin to ferredoxin is catalyzed by PS-I.’ The complex is known to contain P700, the reaction center chlorophyll of PS-I, and several electron acceptors which become reduced upon illumination. The signals obtained by a large number of spectrophotophysical techniques (e.g. ESR and chemically induced dynamic electron polarization) have revealed the involvement of at least five different electron acceptors de- noted A,,, AI, X, B, and A (1-3). Whereas spectrophotophys- ical techniques have been very useful in detecting these ac- ceptors in complex preparations, the same techniques are of limited value in determining the identity of the chemical structures giving rise to the signals detected. As a result, the chemical identity of P700 and centers A,, and AI still remains unresolved (1-3). Mossbauer and ESR spectrometry indicated that centers B and A are iron-sulfur centers (4-7). Based on microwave power saturation studies, these two centers were further assigned as [4Fe-4S] clusters (8). The ESR spectrum of center X also had some resemblance to those of iron-sulfur centers (9), and Mossbauer spectroscopy suggested that center X could be a [4Fe-4S] center (10). However, microwave power saturation studies indicated that center X was not a typical [4Fe-4S] or [2Fe-2S] center and that the spectrum possibly represented a chlorophyll anion magnetically interacting with iron (8).

An alternative approach to study PS-I is to characterize the structural and functional role of each of the PS-I polypep- tides. Using this strategy, H0j and Merller (11) and Golbeck (2) provided biochemical evidence demonstrating that the 82- kDa polypeptides of P700-chlorophyll a-protein 1 bind an iron-sulfur center, most likely center X. PS-I preparations contain additional polypeptides of lower molecular mass (2, 3,11). Although the number reportedvaries, four polypeptides of approximate molecular masses 18,16,14, and 9 kDa appear to belong to the PS-I core (2,3, 11). The 18-kDa polypeptide has been claimed to carry the iron-sulfur centers A and B (12, 13). However, this identification was based solely on correla- tion between the gradual depletion of the 18-kDa polypeptide and the disappearance of centers A and B as monitored by ESR spectroscopy. Malkin et al. (14) isolated an 8-kDa poly- peptide from chloroplasts of spinach (Spinacia oleracea). The

The abbreviations used are: PS-I, photosystem I; DTT, 1,4- dithiothreitol; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; Mes, 2-N-(morpholino)ethanesulfonic acid; ORF, open reading frame; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel elec- trophoresis; Tricine, N-[2-hydroxy-l,l-bis(hydroxymethyl)ethyl]gly- cine.

12676

A 9-kDa Protein of Photosystem I Carries Centers A and B 12677

preparation showed an ESR spectrum similar to that of a ferrodoxin but different from those of centers, A, B, and X. It remains to be established whether the reported spectra could have been derived from an artificial complex of similar molecular mass and generated from cysteine, sulfide, and iron. As demonstrated by Hej and Meller (ll), the coincidental co- migration of such artificially formed oligomeric complexes with the 18 and 16-kDa polypeptides was previously inter- preted to indicate that these two polypeptides were binding iron-sulfur clusters (15, 16). Lagoutte et al. (17) used in uiuo YS labeling and W-carboxymethylation to demonstrate that a 9-kDa polypeptide is the most cysteine-rich component in PS-I preparations obtained from spinach and showed that this polypeptide contains approximately 8 cysteine residues. Similar results pointing toward the 9-kDa polypeptide as an iron-sulfur protein of PS-I have been obtained in labeling studies with Anacystis nidulans (18). In contrast to these two studies, Bonnerjea et al. (12) found that a 30-40% depletion of the 16-, 14-, and clO-kDa polypeptides from a spinach PS- I preparation did not result in a corresponding decrease in the ESR signals of centers A and B. These contradicting results called for a more direct approach for the identification of the apoprotein(s) of centers A and B. In this study, we have used acid-labile sulfide and zero-valence sulfur as specific markers for intact and degraded iron-sulfur clusters, respec- tively. The 18-, 16-, and 9-kDa polypeptides were purified to homogeneity from PS-I vesicles of barley after exposure of the vesicles to chaotropic agents. From its ability to bind zero- valence sulfur and from amino acid sequencing data, the purified 9-kDa polypeptide is identified as a chloroplast- encoded, 2[4Fe-4S] protein that binds centers A and B.

MATERIALS AND METHODS

Unless otherwise indicated, all procedures were carried out at 4 “C.

Preparation of PS-I Vesicles

Chloroplasts of barley (Hordeurn v&are L.) were isolated and osmotically lysed as described in Ref. 19. The lamellar systems (200- 400 mg of chlorophyll) were resuspended in 20 mM Hepes (pH 6.3), 15 mM NaCl, 5 mhi MgCl, at 2.0 mg of chlorophyll/ml. Triton X-100 (25 mg of 100% Triton X-lOO/mg of chlorophyll) was added, and after stirring for 30 min in the dark, a P700-enriched supernatant was obtained by centrifugation at 48,000 X g for 30 min. The super- natant was immediately frozen at -80 “C or diluted 5-fold with 20 mM Tricine (pH 7.5), 0.2% (w/v) Triton X-100 and applied directly to a column (5 x 20 cm) of DEAE-Sepharose CL-6B (Pharmacia Biotechnology AB, Uppsala, Sweden) equilibrated in the same buffer. The column was washed with 1 column volume of equilibration buffer containing 60 mM NaCl. The PS-I vesicles were-eluted using the eauilibration buffer fortified with a 60-500 mM linear NaCl aradient (2 x 2,000 ml) at a flow rate of 110 ml/h. The P700-containing fractions eluted at an approximate NaCl concentration of 150 mM and were combined and concentrated by ultrafiltration to 30 ml in an Amicon cell (Amicon Corp.) fitted with a PM-30 membrane. The concentrated PSI material was applied to a column (5 X 95 cm) of AcA 34 (LKB-Produkter AB, Bromma, Sweden) equilibrated in 25 mM Mes (pH 6.5), 250 mM NaCl, 0.1% (w/v) Triton X-100 (AcA buffer) and eluted at a flow rate of 3 cm/h (11). The vield of P700 was typically 300 nmol. When needed, the PSI vesicles were concen- trated by ultrafiltration using a PM-10 membrane.

Isolation of Polypeptides

9-/Da Polypeptide-Solid NaSCN (3.2 g) was added to a concen- trated suspension of PSI vesicles (10 ml, 1.32 mg of chlorophyll/ml), followed by gentle shaking for 30 min in the dark. The suspension was diluted to 15 ml with AcA buffer and immediately loaded onto a column (2.6 x 95 cm) of AcA 34 equilibrated in AcA buffer and eluted at a flow rate of 3 cm/h. Fractions (5.5 ml) were analyzed for their content of acid-labile sulfide and zero-valence sulfur (20), and their polypeptide content was monitored by SDS-PAGE, followed by Coo- massie Brilliant Blue R-250 staining. Fractions (70-85) enriched in

the 9-kDa polypeptide but devoid of P700-chlorophyll a-protein 1 were combined and dialyzed against 5000 ml of 15 mM ammonium acetate (pH 5.0) 0.05% (w/v) Triton X-100. After dialysis (conductiv- ity, 1.4 millisiemens), the material was loaded onto a column (0.9 x 30 cm) of CM-Sepharose (Pharmacia Biotechnology AB) equilibrated in 25 mM ammonium acetate (pH 5.5), 0.05% (w/v) Triton X-100 and eluted with a linear gradient of 25 mM ammonium acetate (pH 5.5), 0.05% (w/v) Triton X-100 and 400 mM ammonium acetate (pH 8.0), 0.05% (w/v) Triton X-100 (2 X 120 ml) at a flow rate of 10 ml/ h. The pH of the two buffers were adjusted with HOAc and NH3, respectively. Eluted fractions were again assayed for their content of acid-labile sulfide and zero-valence sulfur (20) and for their polypep- tide content. Fractions rich in the 9-kDa polypeptide were combined, lyophilized, dissolved in a minimal volume of AcA buffer, and finally applied to a column (1.6 x 95 cm) of AcA 34 equilibrated in AcA buffer and eluted with the equilibration buffer at a flow rate of 2 cm/ h. Fractions judged by SDS-PAGE to contain pure 9-kDa polypeptide were combined, lyophilized, and stored at -80 “C.

18- and 16-kDa Polypeptides-PS-I vesicles (146 ml) obtained from the AcA column were concentrated to a volume of 10 ml, diluted with 120 ml of 20 mM imidazole HCI (pH 7.6), and concentrated to 17 ml by ultrafiltration. Solid urea (7 g) was added to the concentrated PS- I vesicles (16 ml, 1.20 mg of chlorophyll/ml), and the material was applied (25 ml/h) to a column (0.9 x 46 cm) of Polvbuffer Exchanger 94-(Pharmacia Biotechnology AB) equilibrated in 20 mM imidazole (pH 7.6), 5 M urea. Washing of the column with 160 ml of equilibration buffer served to elute the major part of the 18-kDa polypeptide originally bound to PSI. The column was subsequently washed-with 75 ml of 25 mM imidazole HCl (oH 7.6) and then with 60 ml of 25 mM imidazole HCl (pH 7.6), 0.1% (w/v) Triton X-100,0.05% Empi- gen BB (Albright & Wilson Ltd., Marchon Works, Whitehaven, Cumbria. United Kinadorn) which eluted some of the 16-kDa DOIV- peptide in a homogenous form. P700-chlorophyll a-protein 1, the 9- and 14-kDa polypeptides, and some of the 16-kDa polypeptides were eluted by applying ‘I-fold diluted Polybuffer 74/HCl (pH 4.5), 0.1% (w/v) Triton X-100 to the column. The fractions containing P700- chlorophyll a-protein 1 were concentrated by ultrafiltration and treated with NaSCN as described above for the 9-kDa polypeptide. These manipulations allowed the isolation of a particle which con- tained only P700chlorophyll a-protein 1 and the 14- and 9-kDa polypeptides.

Polyacrylamide Gel Electrophoresis

Analytical SDS-PAGE was carried out in slab gels at 6 “C according to Fling and Gregerson (21) with a 5% stacking gel (1.5 cm) and an 8-25 or 18% resolving gel (18 cm), both containing 0.1% SDS. Preparative PAGE was carried out in the same system except that the SDS concentration was lowered to 0.033% and the length of the stacking and resolving gel was 1 and 6 cm, respectively. When stated, the cysteine residues of the PS-I vesicle and isolated polypeptides were labeled by S-carbamoylmethylation using iodo[l-“Clacetamide in the presence of 2-mercaptoethanol and 8 M urea (22) before analysis by SDS-PAGE and autoradiography. Apparent molecular masses were deduced from the electrophoretic mobility of the follow- ing standards: catalase, aldolase, bovine serum albumin, ovalbumin, and cytochrome c. Determination of the amount of Coomassie Bril- liant Blue R-250 bound to individual polypeptides after staining and destaining of the polyacrylamide gels was carried out according to Ball (23).

Amino Acid Analysis and Sequencing

Samples were hydrolyzed with 6 N HCl at 110 “C in sealed, evacu- ated tubes for 24 h, and the hydrolysates were analyzed using a Durrum D500 amino acid analyzer. Half-cystine was determined after performic acid oxidation (24) or after S-carbamoylmethylation. S- Carbamoylmethylation of isolated polypeptides was carried out by dialyzing the purified proteins against several hundredfold excess of 20 mM ammonium acetate (pH 6.3), 0.05% (w/v) Triton X-100, followed by lyophilization. The lyophilized protein samnle was dis- solved in 7 M guanidinium chloride, 0.2 M Tris/HCl (pH 8.0), 5 mM EDTA, and the solution was thoroughly flushed with nitrogen. DTT was added to give a final concentration of 10 mM. After incubation at 40 “C for 2 h, iodoacetamide was added to the sample to give a final concentration of 200 mM. After incubation at 25 “C for 2 h, followed by 24 h at 4 “C, the sample was extensively dialyzed against 20 mM ammonium acetate (pH 6.3), 0.05% (w/v) Triton X-100 and lyophilized.

12678 A 9-kDa Protein of Photosystem I Carries Centers A and B Lyophilized polypeptide material was easily dissolved in 30% (v/

v) HOAc. Amino acid sequences were determined with both a Beck- man 890C spinning cup sequenator (25) and an Applied Biosystems gas-phase sequenator Model 479A using the program provided by the company. Phenylthiohydantoins were identified by reverse-phase high-pressure liquid chromatography as described by Svendsen et al. (26). C-terminal amino acid analysis (27) was carried out by carbox- ypeptidase Y (649 ng) digestion of the S-carbamoylmethylated 9-kDa polypeptide (13 nmol) in a reaction mixture (150 pl) containing 40 mM Mes (pH 6.5). 0.5% SDS, and 14 nmol of norleucine as an internal standard. Aliquots withdrawn at different times were acidi- fied (pH 2.2) to stop the enzymatic reaction and used directly for amino acid analysis.

Acid-labile Sulfide and Zero-valence Sulfur Acid-labile sulfide and zero-valence sulfur were determined as

reported by H0j and Mdller (20) using the methylene blue procedure with EtOAc extraction steps included to avoid interference from chlorophyll. To secure accurate spectrophotometric determination of methylene blue, the absorption spectrum (500-750 nm) of each sam- ple was recorded using an Aminco DW-2c spectrophotometer with a typical full-scale setting of 0.05. The absorbance at 660 nm was determined from the spectra. Barley ferredoxin was isolated essen- tially as described (28) and used as a reference.

Additional Analytical Procedures Thylakoids *S-labeled in vivo were obtained as previously de-

scribed (11). Chlorophyll was determined according to Arnon (29). P700 was quantitated from its ferricyanide-oxidized minus ascorbate- reduced spectrum using an Aminco DW-2c spectrophotometer and an extinction coefficient of 64 mM" cm" (30).

RESULTS

Compared to our previous study (ll), an ion-exchange chromatography step has been introduced in the preparation of the PS-I vesicles from barley. This allows processing of large amounts of starting material and eliminates contami- nation with chloroplast coupling factor (Fig. 1, lane I ) . The isolation procedure here reported has routinely been used to isolate PS-I vesicles from thylakoids containing several hundred milligrams of chlorophyll. The yield of P700 is ap- proximately 1 nmol/mg of chlorophyll of the thylakoids. This corresponds to a yield of 30%. The PS-I preparation used in Ref. 11 had a chlorophyll to P700 ratio of 110, whereas the ratio is 60 in the present preparation. This difference is explained by the prolonged exposure time to Triton X-100 caused by inclusion of the ion-exchange step.

In the PAGE systems previously used, the major PS-I polypeptides were assigned molecular masses of 70 (doublet), 18, 15, 10, and 8 kDa (11). In this study, the polypeptide composition of the isolated PS-I vesicles was analyzed in the high Tris gel system of Fling and Gregerson (21). Although devoid of urea, this system proved superior in focusing the low molecular mass polypeptides of PS-I. Based on the elec- trophoretic mobility of known standards in the high Tris system, the calculated apparent molecular masses of the major PS-I polypeptides were 82 (doublet), 18, 16, 14, and 9 kDa, respectively. Apparent molecular masses of 82 kDa for the two apoproteins of P700-chlorophyll a-protein 1 are in close agreement with the molecular masses predicted from the nucleotide sequence of their genes (31). The band at 105 kDa represents P700-chlorophyll a-protein 1 which has not been converted into the apoprotein. The band at 195 kDa is prob- ably a dimer of P700-chlorophyll a-protein 1. A minor com- ponent migrating just above the 9-kDa polypeptide was ob- served. In some preparations, an additional minor component was observed in the 23-kDa region. This component is thought to represent residual amounts of light-harvesting chlorophyll- protein I (32) and was not detectable in most of the PS-I preparations.

rn ~ 1 2 3 4 5 6 7 8 9 n

195-

105-

23-

9- -

- CBB- Ag CBB 35S FIG. 1. Analysis of the polypeptide composition of purified

PS-I vesicles and isolated polypeptides by SDS-PAGE. Elec- trophoresis was carried out overnight at 6 'C using an 8-25% high Tris gradient gel. Unless otherwise indicated, the gel was stained with Coomassie Brilliant Blue R-250 ( C B B ) . Lane 1 , purified PS-I vesicles; lane 2, isolated 18-kDa polypeptide; lane 3, isolated 16-kDa polypep- tide; lane 4, isolated 9-kDa polypeptide after extensive handling; lane 5, isolated 9-kDa polypeptide; lane 6, thylakoids; lane 7, PS-I vesicles devoid of the 18- and 16-kDa polypeptides as obtained after NaSCN treatment of PS-I vesicles originally treated with urea (the bands were visualized by alkaline silver staining); lane 8, purified PS-I vesicles obtained from plants "S-labeled in uivo; lane 9, autoradiog- raphy of sample in lane 8.

The Coomassie Brilliant Blue R-250 bound to the individual polypeptide bands of the PS-I preparation was eluted from the gels and quantified spectrophotometrically (23). Normal- ization based on their apparent molecular masses and a uni- form binding of Coomassie Brilliant Blue R-250 gave the following stoichiometry for the five main components: 2.01.2:1.2:0.5:1.1, indicating a stoichiometry in the native PS- I complex of 42:2:1:2 for the apoproteins of P700-chlorophyll a-protein 1 and the la-, 16-, 14-, and 9-kDa polypeptides, respectively.

The relative distribution of sulfur amino acids among the PS-I polypeptides was assessed by electrophoresis and auto- radiography of an 3sS-labeled PS-I preparation obtained from barley seedlings grown in the presence of ["S]sulfate (Fig. 1, lanes 8 and 9). In agreement with earlier observations (11,17, 33), the 18- and 14-kDa polypeptides were found to incorpo- rate small amounts of "S label, whereas the 16-kDa polypep- tide was not labeled at all. P700-chlorophyll a-protein 1 and the 9-kDa polypeptide were both strongly labeled. The super- imposition of the labeled band at 9 kDa with that obtained by Coomassie Brilliant Blue R-250 staining (Fig. 1, lanes 8 and 9) shows that the reported (17, 34, 35) inability to visualize the 9-kDa polypeptide by Coomassie Brilliant Blue R-250 staining was not due to an intrinsic property of the protein, but merely reflected the poor characteristics of the SDS-PAGE systems earlier used. Specific assessment of the content of cysteine residues in the individual PS-I polypep- tides was achieved by 14C-S-carbamoylmethylation of the PS-

A 9-kDa Protein of Photosystem I Carries Centers A and B 12679

x 1 2 3 4 5 6 7 0"

-"

195-

105- 0 2 - 0

Y

9 ;gz " - "

14-=

9- - A: CBB stain

1 2 3 4 5 6 7 -

- B: 1 4 ~ pattern

FIG. 2. Polypeptide composition of purified PS-I vesicles and isolated polypeptides as monitored by Coomassie Brilliant Blue R-250 staining ( A ) and autoradiography after I4C-S- carbamoylmethylation ( B ) . Electrophoresis was carried out over- night at 6 "C using an 8-25% high Tris gradient gel. Lane I , thylakoids (not S-carbamoylmethylated) + I4C-labeled molecular mass stan- dards; lane 2, "C-labeled molecular mass standards; lane 3, purified PS-I vesicles; lane 4, isolated 18-kDa polypeptide; lane 5, isolated 16- kDa polypeptide; lane 6, isolated 8-kDa polypeptide; lane 7, purified PS-I vesicles.

/ I 32x , t I I

0 5 10 15 20 0 5 10 1 20 pl PS I VESICLES pl PSI VESCLES

FIG. 3. Quantitation of the amount of zero-valence sulfur associated with the homogeneous preparation of the 9-kDa polypeptide as compared to the content of acid-labile sulfide in native PS-I particles containing an identical amount of the 9-kDa polypeptide. A, standard curve indicating a linear relation- ship between the amount of PS-I vesicles (0.27 mg of chlorophyll/ ml) assayed and the total amount of acid-labile sulfide and zero- valence sulfur detected; B, standard curve indicating the linear rela- tionship between the amount of PS-I vesicles assayed and the amount of Coomassie Brilliant Blue R-250 bound to the 9-kDa polypeptide of the PS-I vesicle after SDS-PAGE. A and B, aliquots of a homoge- neous preparation of the 9-kDa polypeptide were subjected to analyses identical to those described above. In one such experiment, the 9- kDa polypeptide was found to bind Coomassie Brilliant Blue R-250 corresponding to an absorption of 0.033 at 595 nm and to contain zero-valence sulfur producing an absorption of 0.0022 at 660 nm. As illustrated on A and B, this particular experiment demonstrated that the isolated 9-kDa polypeptide binds zero-valence sulfur correspond- ing to 32% of the amount of acid-labile sulfide present in PS-I vesicles containing an identical amount of the 9-kDa polypeptide.

I preparation prior to electrophoresis (Fig. 2). The 9-kDa polypeptide was by far the most labeled component, with less label in P700-chlorophyll a-protein 1, little in the 18-kDa polypeptide, and none in the 16- and 14-kDa polypeptides. I t should be noted that the denaturing conditions used for the reductive alkylation of PS-I resulted in a partial loss of the 14-kDa polypeptide which seems to aggregate more easily than the rest of the PS-I components (Fig. 2, lane 3 ) . The presence of I4C label in the P700-chlorophyll a-protein 1 polypeptides is in contrast to the results obtained with a PS-

I preparation from spinach (17), but is in agreement with the published nucleotide sequences of the genes encoding the two apoproteins (36).

An isolated intact iron-sulfur protein can be detected by its property to release acid-labile sulfide, whereas a denatured iron-sulfur protein may retain acid-labile sulfide in the form of zero-valence sulfur (37). Incubation of the PS-I preparation with 3.4 M NaSCN and subsequent gel filtration on AcA 34 allowed collection of fractions containing varying amounts of the 18-, 16-, and 9-kDa polypeptides as monitored by SDS- PAGE and Coomassie Brilliant Blue R-250 staining of each individual fraction. The fractions also contained zero-valence sulfur, whereas no acid-labile sulfur was detectable. The elu- tion of zero-valence sulfur correlated with that of the cysteine- rich 9-kDa polypeptide, but not with that of the 18- and 16- kDa polypeptides. Fractions rich in zero-valence sulfur were combined and dialyzed. Subsequent chromatography on CM- Sepharose CL-GB followed by a final gel filtration step on AcA 34 resulted in a homogeneous preparation of the 9-kDa polypeptide. From both columns, the elution of the 9-kDa polypeptide and zero-valence sulfur coincided. The homoge- neity of the isolated 9-kDa polypeptide was assessed by SDS- PAGE, followed by Coomassie Brilliant Blue R-250 staining (Fig. 1, lane 5 ) and autoradiography after I4C-S-carbamoyl- methylation (Fig. 2, lane 6). The zero-valence sulfur contained in the isolated 9-kDa polypeptide was stable to dialysis a t pH 5 and to lyophilization and was therefore bound covalently to the polypeptide backbone, most likely as a trisulfide (37). The isolated polypeptide lacked chromophores absorbing around 420 nm. Such chromophores are present in native ferredoxins (37). Extensive handling of the isolated polypeptide, e.g. by repeated lyophilizations, tended to generate traces of a com- ponent of slightly faster electrophoretic mobility (Fig. 1, lane 4 ) .

The procedure developed to purify the 9-kDa polypeptide also provided homogeneous preparations of the 18- and 16- kDa polypeptides. However, large amounts of these two poly- peptides were more easily obtained after treatment of the PS- I preparation with 6 M urea. When such an extract was applied to a column of Polybuffer Exchanger 94 equilibrated in 20 mM imidazole HCl (pH 7.6), 5 M urea, the 18-kDa polypeptide did not bind to the column and could be collected in a homogeneous form (Fig. 1, lane 2). Upon subsequent washing of the column with 25 mM imidazole HCl (pH 7.6), 0.1% Triton X-100, 0.05% Empigen BB, a proportion of the 16- kDa polypeptide was released in a homogeneous form (Fig. 1, lane 3 ) . A preparation containing P700-chlorophyll a-protein 1 and low molecular mass polypeptides was obtained by elu- tion with Polybuffer 74 (pH 4.5). Subsequent treatment of this preparation with NaSCN followed by gel filtration re- sulted in a preparation containing P700-chlorophyll a-protein 1 and the 14- and 9-kDa polypeptides (Fig. 1, lane 7). A specific association of acid-labile sulfide or zero-valence sulfur with the 18- or 16-kDa polypeptides was not observed under any of the isolation procedures tested. Even when purified, the 16-kDa polypeptide was not reactive toward i ~ d o - l - [ ' ~ C ] acetamide (Fig. 2, lune 5), whereas a weak labeling was obtained with the 18-kDa polypeptide (Fig. 2, lane 4 ) . A condition for the successful application of the purification procedures reported here to obtain homogeneous preparations of the 18-, 16-, and 9-kDa polypeptides of PS-I is the use of a highly purified preparation of PS-I vesicles as the starting material.

The amount of zero-valence sulfur bound to the isolated 9- kDa polypeptide was quantitated with native PS-I as a stan- dard (Fig. 3). Increasing amounts of PS-I were subjected to

12680 A 9-kDa Protein of Photosystem I Carries Centers A and B

SDS-PAGE. After electrophoresis, the Coomassie Brilliant Blue R-250 specifically bound to the 9-kDa polypeptide band of each gel lane was quantitated spectrophotometrically (Fig. 3B) (23). Identical amounts of the PS-I preparation were assayed for acid-labile sulfide and zero-valence sulfur (Fig. 3A), and two standard curves were constructed. In an analo- gous manner, the Coomassie Brilliant Blue R-250 bound to the purified 9-kDa polypeptide after SDS-PAGE and the corresponding content of zero-valance sulfur were deter- mined. The yield of zero-valence sulfur obtained by assaying the 9-kDa polypeptide from two different preparations was 32 and 37% of that obtained when assaying an amount of native PS-I containing exactly the same amount of 9-kDa PS- I. However, when the native PS-I vesicle was subjected to a NaSCN treatment analogous to that used to isolate the 9-kDa PS-I, the recovery of acid-labile sulfide was less than 60% even after incubation with DTT. It is therefore evident that the native 9-kDa polypeptide must bind a major part of the acid-labile sulfide of the native PS-I particle.

To determine the relative distribution of zero-valence sulfur between P700-chlorophyll a-protein 1 and the 9-kDa polypep- tide under identical experimental conditions, the purified PS- I vesicles were subjected to preparative SDS-PAGE. After electrophoresis, the gel was cut into narrow horizontal seg- ments. The content of acid-labile sulfide and of zero-valence sulfur was determined after DTT incubation as described (20). Small, equally sized parts of each segment were used for re-electrophoresis to establish the polypeptide composition in the segments and to spectrophotometrically quantify the poly- peptides present (23) . When an 8-25% gradient gel was used, between 30 and 35% of the recovered acid-labile sulfide was associated with P700-chlorophyll a-protein 1, whereas 65- 70% was associated with polypeptides in the low molecular mass region (Fig. 4). To separate the 9-kDa polypeptide more efficiently from those at 18, 16, and 14 kDa, a similar exper- iment using an 18% resolving gel was performed (Fig. 5). This experiment demonstrated that the acid-labile sulfide re- covered after incubation with DTT was derived from the 9- kDa polypeptide. Thus, after separation of the polypeptides of the PS-I vesicle by SDS-PAGE, 65-70% of the acid-labile sulfide recovered after DTT treatment is associated with the 9-kDa polypeptide, whereas 30-35% resides in the large poly-

I I 0,008 'BAD"

8 2 a 16ADa 142Da

l l i I i 0-

1 3 5 7 0 1 1 1 3 GEL SLICE NUMBER

FIG. 4. Relative distribution of zero-valence sulfur between P700-chlorophyll a-protein 1 and the low molecular mass polypeptides. Electrophoresis was carried out for 5.5 h at 6 "C using an 8-25% high Tris gradient gel. The bars indicate the polypeptide distribution between the different gel slices (5 mm) as monitored by re-electrophoresis.

* In m

i - m

; 3 -

- 4 0 1 s

-0,010

-0,005

-0

GEL SLCE NUMBER

FIG. 5. Relative distribution of zero-valence sulfur between the low molecular mass polypeptides. Electrophoresis was carried out for 5.5 h at 6 "C using an 18% high Tris gradient gel. After SDS- PAGE, the gel was cut into 5-mm segments which were analyzed for their content of zero-valance sulfur (So, 0). The polypeptide content of each segment was analyzed by re-electrophoresis, and the amount of Coomassie Brilliant Blue R-250 bound to each polypeptide was quantitated. In this particular experiment, the recovery of the 9-kDa polypeptide is low.

peptides of P700-chlorophyll a-protein 1. The apoproteins of P700-chlorophyll a-protein 1 barely enter the 18% resolving gel, which is therefore not suitable for determination of the relative distribution of zero-valence sulfur between P700- chlorophyll a-protein 1 and the low molecular weight poly- peptides.

It was of interest to obtain sequence information on the isolated 9-kDa polypeptide. The S-carbamoylmethylated 9- kDa polypeptide was therefore subjected to Edman degrada- tion in a liquid-phase spinning cup sequenator (-10 nmol, 29 cycles, repeated three times, 94% repetitive yield) as well as in a gas-phase sequenator (-1 nmol, 92% repetitive yield). The sequence for the 29 N-terminal residues of the isolated 9-kDa polypeptide is shown on Fig. 6. The spacing of the 4 identified cysteine residues is strongly indicative of a [4Fe- 4S] protein (38). Amino acid analysis of the 9-kDa polypeptide after treatment with performic acid (24) revealed the presence of 8 cysteine residues/78 amino acids (Table I), indicating that the protein might be a 2[4Fe-4S] protein. Very recently, the complete nucleotide sequences of chloroplast DNA from the liverwort Marchantia polymorpha (39) and from tobacco (Nicotiana tabacum) (40) have been determined. Ohyama et al. (39) pointed out that the M. polymorpha sequence con- tained two ORFs denoted frxA and frxB in which the periodic appearance of cysteine residues resembled that of [4Fe-4S] ferredoxins. When the N-terminal sequences predicted from these two ORFs were compared with the N-terminal amino acid sequence of the 9-kDa polypeptide isolated from PS-I vesicles of barley, the homology with frxA, but not with frxB, was striking and leaves no doubt that the isolated PS-I polypeptide is coded for by the corresponding ORF on the barley chloroplast genome (Fig. 6). To accord with the no- menclature used by Ohyama et al. (39) and Gray e t al. (41) for the previously identified chloroplast genes encoding mem- brane proteins catalyzing light reactions of photosynthesis, we designate this gene as psaC. The psaC of the M. polymor- pha chloroplast genome is located between the ndh41 and ndh4 genes (39). An initial search on the chloroplast genome of tobacco (40) for a similar ORF coding for an N-terminal

A 9-kDa Protein of Photosystgm I Carries Centers A and B 12681

FIG. 6. Partial amino acid sequence for the isolated 9-kDa 2[4Fe-4S] polypeptide of PS-I from barley compared with the amino acid sequences deduced from the corresponding gene psaC on the chloroplast genomes of tobacco (40) and M. pol- yrnorpha (39). Identity between codons and amino acid residues of the 9-kDa polypeptide from different species is indicated with a plus. Differences between the two nucleotide sequences causing amino acid substitutions are indicated by dashed bores. The cysteine residues chelating the [4Fe-4S] clusters are bored. The nucleotide inserted into the tobacco sequence (40) to restore identity with the N-terminal amino acid sequence of the 9-kDa polypeptide from barley is indicated with a question mark.

TABLE I Amino acid composition of the 9-kDa photosystem Ipolypeptide

carrying the two I4Fe-4SI centers A and B Amino acid H. vulgare N. tabacurn" M. polymorphab

Cys' 8.1 Asx 5.6 Thr 6.1 Ser 7.0 Glx 7.2 Pro 3.6

Ala 6.3 Val 4.8 Met 1.2 Ile 3.5 Leu 4.1 TY r 2.5 P he 1.8 His 1.3 LYS 4.0 Arg 4.7 Trp NDd

GlY 6.2

Total 78.0

8 6 6 7 7 4 6 6 5 1 4 4 3 2 1 4 5

ND

79

9 5 7 6 6 4 4 6 6 3 4 4 3 1 2 4 5 2

81

9 7 7 5 6 4 5 6 6 3 4 4 3 1 1 3 6 1

81 -

a Deduced from the nucleotide sequence identified on the genome of N. tabacurn (40) after insertion of a missing nucleotide (see "Re- sults").

* Deduced from the nucleotide sequence identified on the genome of M. polymorph (39).

Determined as cysteic acid after performic acid oxidation (24). ND, not determined.

part of the polypeptide homologous to the barley 9-kDa poly- peptide was unsuccessful. Search at the nucleotide level, how- ever, revealed extensive homology between the psaC region of M . polymorpha and the corresponding region in tobacco. It turned out that the ORF found in M. polymorpha was not listed in the analysis of the tobacco chloroplast genome, most probably because a single base pair had been missed in the sequencing of the tobacco chloroplast genome. Thus, insertion of a thymine residue in the noncoding strand of the tobacco sequence at a position corresponding to the wobble position

of codon 20 in the psaC ORF of M. polymorpha restored a nucleotide sequence which coded for a protein identical to the N-terminal part of the barley 9-kDa protein (Fig. 6). The amino acid composition of the isolated 9-kDa PS-I polypep- tide of barley resembles that predicted from the psaC genes of M. polymorpha and tobacco (Table I). The identity between the partial sequence of the 9-kDa PS-I of barley and the sequence deduced from the psaC gene of tobacco strongly indicates that the sequence of the remaining part of the barley protein will be very homologous to the corresponding se- quences in tobacco and M. polymorpha.

The C-terminal amino acid of the 9-kDa polypeptide was determined by carboxypeptidase Y digestion (27) and was found to be tyrosine. Thus, apart from the removal of the N- terminal methionine residue, the primary translation product of the 9-kDa polypeptide is not proteolytically processed. X- ray crystallographic analyses of the soluble 2[4Fe-4S] ferre- doxin from Peptococcus aerogenes had established the identity of the cysteines bound to the two [4Fe-4S] clusters (42). One [4Fe-4S] cluster is bound to Cys-X-X-Cys-X-X-Cys (where X represents amino acid) in the N-terminal half of the protein and to Cys-Pro in the C-terminal half. The second cluster is bound to Cys-X-X-Cys-X-X-Cys in the C-terminal half and to Cys-Pro in the N-terminal part of the protein. The partial amino acid sequence of the isolated 9-kDa polypeptide and the deduced amino acid sequence for the corresponding pro- teins in tobacco and M. polymorpha reveal identical cysteine- containing segments, thereby identifying these proteins as 2[4Fe-4S] proteins.

A hydropathy plot according to Hopp and Woods (43) is shown in Fig. 7. It is interesting to note that the 4 cysteine residues at positions 10, 13, 16, and 57 anchoring one of the [4Fe-4S] centers (42) are positioned in relatively hydrophobic regions, whereas the cysteine residues at positions 20, 47, 50, and 53 coordinating the second center are located in more hydrophilic stretches of the molecule. The positioning of the cysteine residues in the bacterial 2[4Fe-4S] ferredoxins (38) suggests that the cysteine residue at position 33 is not involved in anchoring an iron-sulfur center.

The amino acid composition of the isolated 16- and 18-kDa

r I l l I I I l l 1

1 i0 20 30 40 50 60 70 &J AMINO ACID NUMBER

FIG. 7. Hydropathy plots of the 9-kDa 2[4Fe-4S] polypep- tides of tobacco and M. polyrnorpha carrying centers A and B of photosystem I. The amino acid sequences were derived by identification of the gene psaC for the two polypeptides on the completely sequenced chloroplast DNA of tobacco (40) and M. poly- morphs (39). The hydropathy plots were calculated according to Hopp and Woods (43).

12682 A 9-kDa Protein of Photosystem I Carries Centers A and B SIWC. Reriduer

P. elrdenli 11-281

P . elldenti iz9-541

FIG. 8. Comparison of the amino acid sequences of the 9- kDa 2[4Fe-4S] PS-I polypeptides of barley and tobacco with that of a 2[4Fe-4S] ferredoxin of P. ehdenii (38). To indicate the presumed gene duplication (47), the N- and C-terminal halves of the proteins have been aligned with respect to the positioning of the cysteine residues chelating the two [4Fe-4S] clusters. Identical resi- dues are indicated with bores. Each amino acid is indicated by the standard single-letter code.

polypeptides was also determined.' The 16-kDa polypeptide contained neither methionine nor cysteine, in accordance with the earlier reported absence of label in this protein after I4C- carbamoylmethylation and after labeling with 35S in vivo (11). This polypeptide also lacks histidine. The 18-kDa polypeptide contained small but significant amounts of cysteine (-1 resi- due/molecule).2 Thus, the 16- and 18-kDa polypeptides are not iron-sulfur apoproteins. Both the 18- and 16-kDa poly- peptides contained very high levels of alanine and proline. This was reflected in the partial amino acid sequences which we have obtained for these two proteins.*

DISCUSSION

Partial amino acid sequencing of the 9-kDa polypeptide isolated from PS-I vesicles of barley permitted identification of its corresponding gene psaC on the chloroplast genomes of tobacco (40) and M . polymorph (39) and identification of the protein as a carrier of two [4Fe-4S] clusters (Fig. 6). The spacing of the cysteine residues does not fit that of soluble [ZFe-ZS] proteins which also lack the Cys-Pro segment (38, 44-46). The soluble 2[4Fe-4S] ferredoxins reveal a strong internal homology presumably due to a gene duplication (47). The deduced sequence for the 9-kDa polypeptide of tobacco displays a similar internal homology between residues 3-25 and 40-62 (Fig. 8).

The availability of analytical procedures (20) permitting fast and reliable determination of acid-labile sulfide and zero- valence sulfur was essential in the development of the proce- dure which resulted in isolation of the 9-kDa polypeptide from barley as a partially denatured 2[4Fe-4S] protein. Partial denaturation was evidenced by the lack of absorption around 420 nm and by the inability of the polypeptide to release acid- labile sulfide without prior reduction. The amount of zero- valence sulfur bound to the isolated 9-kDa polypeptide was quantified by two different procedures. After SDS-PAGE of the purified PS-I vesicle, the amount of zero-valence sulfur associated with the 9-kDa polypeptide was twice the amount found to be associated with P700-chlorophyll a-protein 1 (Fig. 4). In its purified state, the amount of zero-valence sulfur associated with the 9-kDa polypeptide was one-third of the amount of acid-labile sulfide present in a native PS-I prepa- ration containing an identical amount of 9-kDa polypeptide (Fig. 3). This is explained by the less than 60% yield of acid- labile sulfide obtained from denatured PS-I vesicles after DTT reduction. Incomplete conversion of zero-valence sulfur into acid-labile sulfide upon reduction may be one reason for the lower recovery of acid-labile sulfide from the isolated 9- kDa polypeptide. Thus, the yield obtained with cysteine tri- sulfide as a standard was 77% (37). In addition, Petering et

* H. V. Scheller, P. B. Hej, I. Svendsen, and B. L. Mller, manu- script in preparation.

al. (37) observed that the zero-valence sulfurs bound in the oxidatively denatured bacterial 2[4Fe-4S] ferredoxins of Mi- crococcus lactylyticus, Clostridium pasteurianum, and Pepto- streptococcus elsdenii were only 42, 48, and 63% of the acid- labile sulfide found in the native proteins, respectively. Of these soluble ferredoxins, that of P. elsdenii shows the highest degree of structural homology with the psaC gene product (Fig. 8) (38). The low recoveries led Petering et al. to conclude that the zero-valence sulfur of these proteins is bound mainly in a cysteine trisulfide structure (37). Quantitative retainment of the acid-labile sulfide originally present in native [4Fe-4S] proteins would require the formation of a cysteine tetrasulfide structure (37). However, the recovery of zero-valence sulfur from oxidatively denatured [2Fe-2S] proteins was also low (37). In contrast to the results of Petering et al. (37) and to the results obtained in this study, Golbeck and Kok (48) have reported a 100% recovery of acid-labile sulfide following de- naturation of PS-I particles and regeneration of acid-labile sulfide from zero-valence sulfur by DTT treatment. Using recoveries of 63 and 77% for the formation of zero-valence sulfur and the regeneration of acid-labile sulfide, respectively, the amount of acid-labile sulfur calculated to correspond to the amount of zero-valence sulfur detected on the isolated 9- kDa PS-I polypeptide corresponds to 70% of the acid-labile sulfide in the native PS-I particle. This value is in close agreement with the relative distribution of zero-valence sulfur between P700-chlorophyll a-protein 1 and the 9-kDa polypep- tide as determined after the denaturing conditions of SDS- PAGE (Fig. 4) by which acid-labile sulfide is converted into zero-valence sulfur (11, 20, 34, 35).

Native PS-I vesicles are generally found to contain 12 molecules of acid-labile sulfide for each molecule of P700 (1, 2, 3, 11). From the results presented here, we can assign 8 of the molecules of acid-labile sulfide to the 9-kDa polypeptide and the remaining 4 to the apoproteins of P700-chlorophyll a-protein 1. H0j and Maller (11) and Golbeck and Cornelius (49) have recently demonstrated that P700-chlorophyll a- protein 1 carries center X. Centers A and B are therefore identified as the two [4Fe-4S] clusters of the chloroplast- encoded 9-kDa 2[4Fe-4S] polypeptide. Experiments based on ESR spectrometry have revealed a differential sensitivity of centers A and B toward oxidative denaturation (48) and toward reactivity with mercurials (50), with center B as the most sensitive cluster. Similarly, center B has been demon- strated to be sensitive to the membrane-impermeant probe p - diazonium benzene sulfonate (51). A hydropathy plot of the 9-kDa polypeptide reveals that the [4Fe-4S] cluster coordi- nated by cysteine residues 10, 13, 16, and 57 is buried in the membrane, whereas the cluster coordinated by cysteine resi- dues 20,47,50, and 53 is more external (Fig. 7). We therefore conclude that these two [4Fe-4S] clusters represent centers A and B, respectively. Selective destruction of one [4Fe-4S] center had also been reported in the soluble 2[4Fe-4S] ferre- doxin I of Azotobacter uinelandii (52). The localization of centers A and B on the same polypeptide chain is in agreement with the strong interaction observed between these two cen- ters by ESR spectroscopy ( 5 , 7, 54).

One important aspect regarding the biosynthesis of the 2[4Fe-4S] holoprotein is the formation of the iron-sulfur cluster. Denatured soluble [2Fe-2S] ferredoxins are easily reconstituted (55). Although N- and C-terminal analyses of the isolated 9-kDa polypeptide established that the primary gene product of psaC is not post-translationally cleaved except for the loss of the N-terminal methionine residue, it has not yet been possible by reconstitution experiments to regenerate the ESR signals of the iron-sulfur centers from denatured PS-

A 9-kDa Protein of Photosystem I Carries Centers A and B 12683

I vesicles (56). Takahashi et al. (57) have recently demon- strated that sulfur atoms of the iron-sulfur cluster of chloro- plast ferredoxin are derived from cysteine and that a soluble stroma enzyme is involved in the cluster formation. It will be interesting to test the activity of this enzyme toward the isolated 9-kDa apoprotein.

P700-chlorophyll a-protein 1 was shown in the present study to carry 4 of the 12 molecules of acid-labile sulfide associated with the native PS-I vesicle per molecule of P700. We have previously reported that P700-chlorophyll a-protein 1 binds 4.3 iron atoms/molecule of P700 (11). Using experi- mental conditions where P700-chlorophyll a-protein 1 had been functionally detached from the lower molecular mass polypeptides, Golbeck and Cornelius (49) obtained absorbance transients at 698 nm, indicating that the iron-sulfur center associated with P700-chlorophyll a-protein 1 was center X. Mossbauer spectroscopy identified center X as a [4Fe-4S] center (10). However, extended x-ray absorption fine-struc- ture measurements (58) and core extrusion studies (59) indi- cated that PS-I contains [2Fe-2S] clusters as well as [4Fe-4S] clusters. In this study, centers A and B have been identified as [4Fe-4S] centers. This points toward center X as a [2Fe-2S] center. With 4 iron and 4 sulfur atoms present on P700- chlorophyll a-protein 1 per molecule P700, this would permit the location of two [2Fe-2S] centers on this protein. Recently, Bonnerjea and Evans (60) have provided evidence for a cor- responding heterogeneity of the signal associated with center X.

If center X is composed of two traditional [2Fe-2S] centers, this would require the availability of 8 cysteine residues. P700- chlorophyll a-protein 1 is composed of two apoproteins with approximate molecular masses of 83 kDa and which are present in near equimolar amounts (31). Both apoproteins are chloroplast-encoded, and their genes (psaA and psaB) have been sequenced in maize (36), tobacco (40), spinach (61), pea (62), M . polymorpha (39), Euglena (63), and Synechococ- cus (64). In all these species, both genes specified the following completely conserved stretch of 12 amino acids: Phe-Pro-Cys- Asp-Gly-Pro-Gly-Arg-Gly-Gly-Thr-Cys. Besides the 2 cys- teines of this segment, no other cysteines were found in the psaB gene product. The psaA gene of tobacco, maize, and pea specified 2 additional cysteines, whereas that of Euglena and Synechococcus specified only 1 of these residues. Fish et al. (53) speculated that the 2 cysteine residues of the conserved segment indicated above were connected by a disulfide bond in the native protein. We propose, in agreement with specu- lations of Golbeck et al. (59), that these conserved cysteine residues from two psaA polypeptides and from two psaB polypeptides constitute the 8 cysteine residues necessary to coordinate the two [2Fe-2S] clusters of P700-chlorophyll a- protein 1. Thus, four large polypeptides would be required to bind 1 molecule of P700 and the two [2Fe-2S] clusters repre- senting center X. Assuming a uniform binding of Coomassie Brilliant Blue R-250, the stoichiometry of the polypeptides of the native PS-I vesicle was calculated to be 42:2:1:2 for the apoproteins of P700-chlorophyll a-protein 1 and the 18-, 16-, 14-, and 9-kDa polypeptides, respectively. These ratios are in close agreement with those obtained with I4C-labeled prepa- rations from Synechococcus (56). It has not escaped our atten- tion that these ratios predict the presence of two 9-kDa polypeptides/molecule of P700, which would give a calculated content of 20 eq of iron and acid-labile sulfide/molecule of P700. The experimentally found ratio is 12 (1, 2, 3, 11). At present, we are not able to discriminate between the possible explanations for this discrepancy.

During the course of this work, the 18- and 16-kDa poly-

peptides were also purified to homogeneity. The function of these two polypeptides remains unknown. However, partial amino acid sequencing indicated that the two polypeptides are related.*

Acknowledgments-Hanne Linde Nielsen, Inga Olsen, Bodil Cor- neliussen, Lone Sbrensen, and Pia Breddam are thanked for skillful technical assistance. Drs. Birte Svensson and David Simpson are thanked for helpful discussions. Professor Knud W. Henningsen and Drs. Jan Lembeck, T. G. Petersen, and C.-E. Olsen are thanked for performing the computer analyses.

1.

2. 3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20. 21.

22.

23. 24. 25.

26.

27.

28.

29. 30.

31.

32.

33.

REFERENCES

Rutherford, A. W., and Heathcote, P. (1985) Photosynth. Res. 6 ,

Golbeck, J. H. (1987) J. Membr. Sci., in press Meller, B. L., H@j, P. B., Halkier, B. A., Olsen, I., Nielsen, H. L.,

Evans, E. H., Rush, J. D., Johnson, C. E., and Evans, M. C. W.

Malkin, R., and Bearden, A. J. (1971) Proc. Natl. Acad. Sci.

Evans, E. H., Cammack, R., and Evans, M. C. W. (1976) Biochem.

Evans, M. C. W., Telfer, A., and Lord, A. V. (1972) Biochim.

Rupp, H., Rao, K. K., Hall, D. O., and Cammack, R. (1978)

Evans, M. C. W., Sihra, C. K., and Cammack, R. (1976) Biochem.

Evans, E. H. Dickson, D. P. E., Johnson, C. E., Rush, J. D., and

H0j, P. B., and Mbller, B. L. (1986) J. Biol. Chem. 2 6 1 , 14292-

Bonnerjea, J., Ortiz, W., and Malkin, R. (1985) Arch. Biochem.

Golbeck, J. H., Parrett, K. G., and Root, L. L. (1987) in Progress in Photosynthesis Research (Biggins, J., ed) Vol. I, pp. 253-256, Martinus Nijhoff, Dordrecht, The Netherlands

Malkin, R., Aparicio, P. J., and Arnon, D. I. (1974) Proc. Natl. Acad. Sci. U. S. A. 71 , 2362-2366

Hiller, R. G., Mdler, B. L., and H@yer-Hansen, G. (1980) Curls- berg Res. Commun. 45 , 315-328

Mdler, B. L., H@yer-Hansen, G., and Hiller, R. G. (1981) Proc. Int. Congr. Photosynth. Res. 3,245-256

Lagoutte, B., Setif, P., and Duranton, J. (1984) FEBS Lett. 174 ,

Guikema, J., and Sherman, L. (1982) Biochim. Biophys. Acta

Mdler, B. L., and Hplj, P. B. (1983) Carkberg Res. Commun. 48 ,

H0j, P. B., and Mdler, B. L. (1987) Anal. Biochem. 2 4 0 , in press Fling, S. P., and Gregerson, D. S. (1986) Anal. Biochem. 155,

Crestfield, A. M., Moore, S., and Stein, W. H. (1963) J. Biol.

Ball, E. H. (1986) Anal. Biochem. 155, 23-27 Hirs, C . H. W. (1967) Methods Enzymol. 11, 197-199 Johansen, J. T., Overballe-Petersen, C., Martin, B., Hasemann,

V., and Svendsen, I. (1979) Carlsberg Res. Commun. 44 , 201- 217

Svendsen, I., Martin, B., and Jonassen, I. (1980) Carlsberg Res. Commun. 45,79-85

Martin, B., Svendsen, I., and Ottesen, M. (1977) Carlsberg Res. Commun. 42,99-102

Buchanan, B. B., and Arnon, D. I. (1971) Methods Enzymol. 2 3 , 413-440

Arnon, D. I. (1949) Plant Physiol. 2 4 , 1-14 Hiyama, T., and Ke, B. (1972) Biochim. Biophys. Acta 2 6 7 , 160-

Fish, L. E., and Bogorad, L. (1986) J. Biol. Chem. 2 6 1 , 8134-

Lam, E., Ortiz, W., and Malkin, R. (1984) FEBS Lett. 168 , 10-

Nechushtai, R., andNelson, N. (1981)J. Bid. Chem. 256,11624-

295-316

and Madsen, A. (1987) Dan. J . Agron. 132,5-21

(1979) Biochem. J. 182,861-865

U. S. A. 68, 16-19

Biophys. Res. Commun. 6 8 , 1212-1218

Biophys. Acta 2 6 7 , 530-537

Biochim. Biophys. Acta 537 , 255-269

J. 158 , 71-77

Evans, M. C. W. (1981) Eur. J . Biochem. l l 8 ,81 -84

14300

Biophy~. 2 4 0 , 15-20

24-29

681,440-450

161-185

83-88

Chem. 238,622-627

171

8139

14

11628

12684 A 9-kDa Protein of Photosystem I Carries Centers A and B 34. Mdler, B. L., Halkier, B. A., and H#j, P. B. (1987) in Progress in 48. Golbeck, J. H., and Kok, B. (1978) Arch. Biochern. Biophys. 188,

Photosynthesis Research (Biggins, J., ed) Vol. 11, pp. 49-52, 233-242 Martinus Nijhoff, Dordrecht, The Netherlands 49. Golbeck, J. H., and Cornelius, J. M. (1986) Biochim. Biophys.

~ho to~yn thes i s Research (Biggins, J. , ed) V O ~ . 11, pp. 53-56, 50. Kojima, y., Hiyama, T., and Sakurai, H. (1987) in Progress in Martinus Nijhoff, Dordrecht, The Netherlands Photosynthesis Research (Biggins, J., ed) Vol. 11, pp. 57-60,

37. petering, D., F ~ ~ , J. A., and palmer, G. (1971) J , ~ i ~ l , then. 52. Morgan, T. V., Stephens, P. J., Devlin, F., Burgess, B. K., and Stout, C. D. (1985) FEBS Lett. 183, 206-210

38. Yasunobu, K. T., and Tanaka, M. (1980) Methods Enzyrnol. 6 9 , 53. Fish, L. E., Kuck, U., and Bogorad, L. (1985) in Molecular Biology of the Photosynthetic Apparatus (Steinback, K. E., Bonitz, S.,

39. Ohyama, K., Fukuzawa, H., Kohchi, T., Shirai, H., Sano, T., Arntzen, C. J., and Bogorad, L., eds) pp. 111-120, Cold Spring

S a m S., Umesono, K., Shiki, Y., Takeuchi, M., Chaw, z., 54. A ~ ~ ~ , R., ~ ~ ~ ~ ~ t ~ o ~ , J., and Vhnngird, T. (1981) ~ i ~ ~ h i ~ . ~ i ~ - Harbor Laboratory, Cold Spring Harbor, NY

Aota, S.-i., Inokuchi, H., and Ozeki, H. (1986) Nature 3 2 2 ,

40. Shinozaki, K., Ohme, M., Tanaka, M., Wakasugi, T., Hayashida, 55. Palmer, G. (1975) in The Enzymes (Boyer, P. D., ed) Vol. 12, pp.

1-56, Academic Press, New York N., Matsubayashi, T., Zaita, N., Chunwongse, J., Obokata, J., 56, Lundell, D. J., Glazer, A. N., Melis, A., and Malkin, R. (1985) J . Yamaguchi-Shinozaki, K., Ohto, C., Torazawa, K., Meng, B. Biol. Chem. 260,646-654 Y., Sugita, M., Deno, H., Kamogashira, T., Yamada, K., Ku- 57. Takahashi, Y., Mitsui, A., Hase, T., and Matsubara, H. (1986) suda, J., Takaiwa, F., Kato, A., Tohdoh, N., Shimada, H., and Proc. Natl. Acad. Sci. U. S. A. 83, 2434-2437 Sugiura, M. (1986) EMBO J. 5 , 2043-2049 58. McDermott, A. E., Yachandra, V. K., Guiles, R. D., Britt, R. D.,

41. Gray, J. C., Blyden, E. R., Eccles, C. J., Dunn, P. P. J., Hird, S. Dexheimer, S. L., Sauer, K., and Klein, M. P. (1987) in Progress M., Hoglund, A.-S., Kaethner, T. M., Smith, A. G. , Willey, D. in Photosynthesis Research (Biggins, J., ed) Vol. I, pp. 249-252, L., and Dyer, T. A. (1987) in Progress in Photosynthesis Re- Martinus Nijhoff, Dordrecht, The Netherlands search (Biggins, J., ed) Vol. IV, pp. 617-624, Martinus Nijhoff, 59. Golbeck, J. H., McDermott, A. E., Jones, w. K., and Kufiz, D. Dordrecht, The Netherlands M. (1987) Biochirn. Biophys. Acta 891,94-98

Chem. 248,3987-3996

35. H0j, P. B., Halkier, B. A., and Mdler, B. L. (1987) in Progress in Acta 849,16-24

36. Fish, L. E., Kuck, U., and Bogorad, L. (1985) J . Biol. Chern. 260, Martinus Nijhoff, Dordrecht, The Netherlands 1413-1421 51. Malkin, R. (1984) Biochirn. Biophys. Acta 764 , 63-69

246,643-653

228-238

572-574 p h y ~ . Acta 637, 118-123

42. Adman, E. T., Sicker, L. C., and Jensen, L. H. (1973) J. Bioi. 60. Bonneiea, J. R.7 and Evans, M. c. w. (1984) Biochim. BioPhYs. Acta 767, 153-159

43, H ~ ~ ~ , T. p., and Woods, K, R. (1981) proc, Natl, ~ ~ ~ d , sei. 61. Kirsch, W., Seyer, P., and Herrmann, R. G. (1986) Curr. Genet.

44. Bruschi, M. H., Guerlesquin, F. A., Bovier-Lapierre, G. E., Bon- 62. Lehmheck, J., Rasmussen, 0. F., Bookjans, G. B., Jepsen, B. R., icel, J. J., and Couchoud, P. M. (1985) J. Biol. Chern. 260,

Stummann, B. M., and Henningsen, K. W. (1986) Plant Mol. Biol. 7, 3-10

45. Takahashi, Y., Hase, T., Wada, K., and Matsubara, H. (1983) 63. Cushman, J. C., Hallick, R. B., and Price, C. A. (1987) in Progress

Plant Cell Physiol. 2 4 , 189-198 in Photosynthesis Research (Biggins, J., ed) Vol. IV, pp,

46. Hase, T., Matsubara, H., Koike, H., and Kat& s. (1983) 64. Bryant, D. A., Lorimier, R. D., Guglielmi, G., Stirewalt, V. L., 667-670, Martinus Nijhoff, Dordrecht, The Netherlands

Biochirn. Biophys. Acta 7 4 4 , 46-52 Cantrell, A,, and Stevens, S. E., Jr. (1987) in Progress in 47. Tanaka, M., Nakashima, T., Benson, A., Mower, H., and Yasu- Photosynthesis Research (Biggins, J., ed) Vol. IV, pp. 749-755,

nobu, K. T. (1966) Biochemistry 5, 1666-1681 Martinus Nijhoff, Dordrecht, The Netherlands

IJ. S. A. 78, 3824-3828 10,843-855

8292-8296