Indian Journal of Biochemistry & Biophysics Vol. 37, December 2000, pp. 405-417

An insight into the assembly and organization of £hotosystem I complex in the- thylakoid membranej of the lhermophilic cyanobacterium, gastigocla us Zaminasus

_ ¥1· ��:sh;, Zengyong He, �, era aras ahana, Lirnor inai�liYvVerchovsk; Dorit IC ae I and�

roepartment of Plant Sciences, The 1 erman Institution of Life Science, The Hebrew University of Jerusalem, L-- Jerusalem 9 1904, Israel � Accepted 30 May 2000-

The present study�racterizes the assembly and organization of Photosystem I (PSI) complex, and its individual subunits into the thylak�embranes of the thermophilic cyanobacterium, Mastigocladus laminosus. PSI is a multi protein complex that contains peripheral as well as integral subunits. Hence, it serves as a suitable model system for understanding the formation and organization of membrane protein complexes. In the present study, two peripheral cytosol facin�subunits of PSI, namely, PsaD and PsaE were overexpressed in E. coli and used for assembly studies. The gene encoding PsaK, an integral membrane spanning subunit of PSI, was cloned and the deduced amino acid sequence revealeqsaK to ave two transmembrane a-helices. The characterization of the in vitro assembly of the peripheral subunits, PsaD and PsaE, as well as of the integral subunit, PsaK, was performed by incubating each subunit with thylakoids isolated from Mastigocladus laminosus. All three subunits studied were found to assemble into the thylakoids in a spontaneous mechanism, showing no requirement for cytosolic factors or NTP's (nucleotide 5'-triphosphate). Nevertheless, further characterization of the assembly of PsaK revealed its membrane integration to be most efficient at 55°C.

The associations and protein-protein interactions between different subunits within the assembled PSI complex were directly quantified by measurements performed using the B IACORE technology. The preliminary results indicated the existence of specific interaction between PsaD and PsaE, and revealed a very high binding affinity between PsaD and the PSI electron acceptor ferridoxin (Kd= 5.8 x 1O·1 lM). PsaE has exhibited a much lower binding affinity for ferridoxin (Kd = 3 . 1 x 1 0.5 M), thereby s���ng the possibility of PsaE being one of the subunits responsible for the dissociation of ferridoxin from the PSI com/ � 6 V 'r!Yf �

Introduction r' Photosystem I (PSI) is a photochemically active In general, the knowledge available on the assembly membrane protein complex present at the reducing

and organization of membrane proteins and membrane site of the photosynthetic electron transfer chain. It protein complexes, like PSI, lags behind as compared functions as plastocyanin-ferridoxin oxido-reductase to that available on soluble/globular proteins. Although in the thylakoids of plants, algae and cyanobacteria. during the last decade a wealth of information has In cyanobacteria, plastocyanin, the PSI electron donor accumulated, the principles and mechanisms dictating is occasionally replaced by a c type cytochrome the formation of a membrane protein complex and the (cytochrome C6) and during iron deficiency, regulation of this process are not yet fully understood. ferridoxin, the PSI electron acceptor, is replaced by Studies on the assembly and organization of membrane flavodoxin6•7 • protein complexes from different origins have greatly contributed to our understanding of these complex processes1.5. Among the different membrane systems studied were the thylakoid protein complexes. The inner membrane network of the chloroplast in which photosynthesis takes place comprises the following four multi-subunit membrane protein complexes: photo system II (PSII), cytochrome b6f, photosystem I (PSI) and the ATP-synthase5.

* Author of correspondence

The PSI complex of cyanobacteria consists of 1 1 different subunits namely PsaA to PsaF (6 subunits) and PsaI to PsaM (5 subunits). PsaA and PsaB with 1 1 membrane spanning helices each, form the hydrophobic core of the complex to which all the photosynthetic pi�ments and most of the electron carriers are bound .9. The two final electron acceptors of PSI, the iron-sulfur (Fe-S) clusters, FA and FB, from where the electrons are transferred to ferridoxin (or to flavodoxin) are located in PsaC which is a peripheral

406 INDIAN J B IOCHEM. B IOPHYS . , VOL. 37, DECEMBER 2000

subunit facing the cytosol (the stroma in plants). PsaC is shielded by two other peripheral subunits, PsaD and PsaElo, both of which are involved in the docking of ferridoxin 1 1 - 1 8 and flavodoxin 1 9 , and in strengthening the binding of PsaC to PSI I I , 1 2.20.2 1 . Cross-l inking experiments and mutagenesis studies have revealed that PsaD functions as the docking site for ferridoxin l l - 1 7 . A specific interaction between PsaE and ferridoxin has been observed only in barle/2. However, PsaE of cyanobacteria is also postulated to be invol ved in the docking and binding of ferridoxinmavodoxin to PSI I 7. 1 8 . Barth ef al. 1 7 have suggested that PsaE is responsible for the dissociation of ferridoxin from PSI. Another function related to PsaE is its involvement in the cycl ic electron flow occulTing around the PSI complex23.24.

PsaF is the only subunit located on the lumenal side of the thylakoid membrane in cyanobacteria and it has been shown to contain at least one tran membrane a_hel ix20,2s. While in eukaryotes, PsaF was shown to function as the docking site for plas .ocyanin26.:n, the cyanobacterial PsaF is not involved i n the docking of cytochrome C6cS.28 and its specific function is st i l l unknown.

The other subunits of PSI (PsaI-PsaM) traverse the thylakoid membrane. Apart from PsaL, all these subunits of PST are low molecular weight subunits. PsaJ has been shown to be involved in strengthening the binding of PsaL and PsaM to the PSI complex29,30. PsaJ was found to stabilize the binding of PsaE to the PSI complex' i . The psaJ gene is monocystronic with the psaF gene. Hence deletion of psaJ gene leads to 80% reduction in the PsaF content in PST28,32. PsaL is the subunit responsible for PSI trimer formation in cyanobacteria29,33 while the functions of PsaK and PsaM are not yet known.

As mentioned above, a multi-protein membrane complex such as PSI, that contains peripheral as wel l as transmembrane subunits is an attractive model system for understanding the principles and mechanisms dictating the organization and assembly of proteins and complexes in biological membranes. These processes include the assembly of the peripheral and transmembrane proteins in the l ipid bilayer and the association of pigments and co-factors with the proteins, besides the translocation of proteins across the membrane to their functional location and finally the interactions among the different subunits to comprise the fully-assembled functional PSI complex.

The peripheral subunits of PSI facing the

cytosoVstroma have been shown to assemble into the thylakoids spontaneously wi thout the need of any cytosolic factors or the presence of NTP' S3 1 .34. As for transmembrane thylakoid proteins, several mechanisms have been characterized for their integration into the membrane. Among these are the Sec-dependenr mechanism reqtl 1nng A TP, the SRP-dependent mechanism requiring GTP and the �pH-dependcnt pathway in which there is a requirement for the . f l ' " 1 3 '; assistance 0 t 1e proton moti ve lorce " - .

The present study aims at the characterization and organization of the PSI complex and its indiv idual subunits. Two complimentary experimental approaches were used. The integration of different PSI subunits to the thylakoids was fol lowed by ill vitro thylakoid assembly systems in which a specific PSI subunit was incubated with the isolated thylakoids3 J .3.uu.37. The interactions between different subunits as well as between the entire PSI complex and its electro n acceptor, fenidoxin, were measured directly by biophysical method.

The results indicate that within the same complex (PSI), the membrane integration of proteins belonging to different classes uti l ize the same spontaneous mechanism. Although the mechanism for the assembly of the integral membrane subunit PsaK was found to be similar to the assembly of the peripheral subunits, PsaD and PsaE, high temperature during assembl y increased the efficiency of PsaK integration.

Materials and Methods

MastigocladIts laminosus growing conditions Cel ls were grown in medium D of Castenholz38 at

pH 8 .2 with constant stirring at 55°C. The cul ture was bubbled with water-saturated air supplemented with 5% CO2• The cel l s were i l luminated with white l ight at an intensity of 105 erg cm-2 sec- I .

Isolatioll of thylakoid membranes from Mastigocladlls

laminosus About 2-5 l itres of cells grown for 3-4 days were

pelleted by centrifugation at 4°C for ! O min at 9000 rpm in a Sorvall GSA rotor. The pellet obtained was washed once in buffer A (50 ruM MES-NaOH pH 6.0, 25% g lycerol (v/v), 5 mM CaCh and 5 ruM MgCI2) and pelleted again by centrifugation as described above. The cells were resuspended in buffer A containing protease inhibitors (benzamidine, amino caproic acid and PMSF at a concentration of 1 mM each) and broken in a bead-beater chamber (Biospec

408 INDIAN J BIOCHEM. BIOPHYS., VOL. 37, DECEMBER 2000

carried out for 45 min. After this, the cells were harvested, broken and centrifuged as described above.

Cloning of psaK gene The psaK gene from Mastigocladus laminosus was

cloned by heterologous hybridization. The probe used was the Synechocystis sp. PCC 6803 psaKI gene coding region amplified by PCR from Synechocystis sp. PCC 6803 genomic DNA. The PCR products were purified on agarose gel and labeled by Digenxin using the High-Dig Random DNA Labeling Kit (Boheringer). The probe was used to screen a Mastigocladus laminosus HindlII genomic library which was constructed by a complete restriction enzyme digestion of Mastigocladus laminosus ' genomic DNA. The HindlII digested genomic DNA was ligated to pBluescript SK( +) (Stratagen, La Jolla, CA) which was digested by HindlII and dephosphorylated. Hybridization and detection were performed following the manufacturer' s instructions. One positive clone with an insert of -4kb was isolated after three rounds of screening. Following sequencing, the psaK gene was identified, isolated and subcloned into pET 22b vector to form the plasmid pET 22b-psaK.

In vitro translation of PsaK The plasmid harbouring psaK (pET 22b-psaK) was

linearized by Bpu1 102I which is located -60bp downstream to the target gene. The linear DNA of the psaK gene was transcribed in vitro using the T7 RNA polymerase. The resulting mRNA was then translated in a wheat germ extract in the presence of eSS]methionine41 . In order to remove small molecules like dNTP's, the translation reaction was chromatographed through G-25 column.

Assembly reaction into Mastigocladus laminosus thylakoids

A typical assembly reaction included 200,000 cpm of pure labeled protein (PsaD, PsaE) or translation product (PsaK) and thylakoids (an amount equivalent to 50 J..lg or 25 J..lg chlorophyll). The insertion reaction of PsaK included 10 mM Mg-ATP and 1 0 mM methionine. Thylakoids were washed prior to the assembly reaction with the assembly buffer. For PsaD and PsaE the assembly buffer HM contained 30 mM HEPES-NaOH pH 8 .0, 1 mM MgCI2. Whereas, for PsaK assembly, a buffer containing 1 00 mM Tricine­NaOH pH 8.0 was used. A typical reaction was carried out for 30 min at 25°C or 55°C, unless

indicated otherwise. Following the incubation, the samples were treated either with proteases (thermolysin or Proteinase K) or washed with high salt, 2 M NaBr. Subsequently, thermolysin (0.2 mglrnl) was added to the reaction mixture in the presence of 5 mM CaCh and the mixture was incubated for 30 min at 25°C. Proteinase K was added to a final concentration of 50 J..lglrnl, and the mixture was incubated as with thermolysin. NaBr was added to a final concentration of 2 M, and the incubation was carried out for 20 min on ice. Following these treatments, the thylakoids were washed three times in buffer, as indicated above. Since in the assembly of PsaE, aggregates of the protein were found to accumulate with the thylakoid pellets, whenever assembly reaction was performed with PsaE, the thylakoids were extracted in the presence of 1 % DM for 10 min at 25°C, and the extract was loaded on the gel . This extraction proved to be efficient in isolating only PsaE embedded in the thylakoids. However the aggregated protein could not be solubilized.

Measurement of equilibrium constant (subunit association)

The binding constant between different PSI subunits as well as between the entire PSI complex and its electron acceptor ferridoxin were determined by surface plasmon resonance using the BIACORE system (BIACORE®3000 from B IACORE AB). The binding measurement of PsaD-PsaE and that of PsaD­ferridoxin were performed with the CM5 chip. Pure PsaD (4 J..lg) was immobilized on the chip and the analyte at each experiment (i.e. ferridoxin or PsaE) was passed through the chip at a flow rate of 5 J..ll/min, at 4-5 different concentrations ranging from 0.2 J..lM to 1 J..lM. The same procedure was carried out in the experiment measuring the binding of PsaE-ferridoxin, with PsaE being the l igand and ferridoxin being the analyte. For the binding measurement of PSI with ferridoxin, the NT A chip was used. Ferridoxin (His­tagged at its C-terminus) was immobilized on the chip. PSI was used as analyte and passed through the cell at a flow rate of 5 J..ll/min, in 5 different concentrations ranging from 8 nM to 40 nM. Miscellaneous techniques

Chlorophyll concentration was determined as described by Amon42.

Protein concentration was determined by B io-Rad protein assay kit following the manufacturer' s instructions or by measuring the absorbance of the

LUSHY et at.: AN INSIGHT INTO THE ASSEMBLY AND ORGANIZATION OF PHOTOSYSTEM I COMPLEX 407

Products) in the presence of pre-chilled 0. 1 mm diameter glass beads. The breakage was carried out following 70 min incubation of the cells mixed with the beads on ice, using 8 pulses of 20 sec each with 5 min cooling intervals. Unbroken cells and residual beads were then removed by centrifugation for 5 min at 6000 rpm. Thylakoid membranes were then pelleted by centrifugation at 38000 rpm for 20 min in a Beckman 50.2 Ti rotor and the supernatant containing the cyanobacterial cytosol was collected for further use. In order to remove phycobiliproteins, the obtained pellet was suspended and washed 3 times in buffer B (buffer A containing 5 mM CaClz). After the last wash, the thylakoids were suspended in buffer B to a chlorophyll concentration of 1 mg/m!.

Purification of PSI complex from Mastigocladus laminosus thylakoids

Thylakoids ( 1 mg chl/ml) in buffer B were stirred gently with 0. 1 % (w/v) of n-dodecyl [3-D-maltoside (OM) for 10 min in dark and then pelleted by centrifugation at 43000 rpm for 30 min (Beckman 70 Ti rotor). This step was repeated with 0.6% DM. The extract was loaded on semi-preparative anion exchange column filled with DEAE-TOYOREARL TSK650S (TOSOH Corporation, Japan) ( l mg chI: 1 ml resin) after filtration through 0.2 �m cellulose mixed ester membrane. The columri was previously washed with 2 column volumes of buffer C (50 mM HEPES-NaOH pH 7.2, 0. 1 M MgS04, 0.03% (w/v) OM, 20% (v/v) glycerol) and with 2 column volumes of buffer D (buffer C containing 10 mM MgClz and 1 2 mM CaCh). Following the application of thylakoid OM extract, the column was washed (5 ml/min) with 6 column volumes of buffer D and 7 column volumes of buffer E (50 mM HEPES-NaOH pH 7.2, 5 mM MgCI2, 12 mM CaCh, 0.03% (w/v) DM, 20% (v/v) glycerol) in order to eliminate most of the free pigments (chlorophyll, carotene and phycobilies). Elution of photosynthetic complexes was then performed by a multi-stage linear gradient of 5%-1 8% buffer C in buffer E at a flow rate of 4 ml/min. Since PSI was contaminated with small amounts of PSII (as revealed by non-denaturing gel electrophoresis), it was necessary to carry out further purification. The additional PSI purification was carried out on a column similar to that used in the first step ( l mg chI: 2-3 ml resin). The column was pre-equilibrated with 2 column volumes of buffer F (20 mM HEPES-NaOH pH 7.2, 0. 1 M MgS04, 0.03% (w/v) DM), followed by 2 column volumes of buffer G (buffer F excluding

MgS04) . The peak fraction containing PSI was loaded on the second column and washed with one column volume of buffer G (4 mlImin). Elution was performed by a multi-stage l inear gradient using 33.5%-50% of buffer F in buffer G at a flow rate of 2 ml/min. The eluted fractions were analyzed by non-denaturing gel and SDS-PAGE. No bands other than the ones representing PSI components could be detected. Since the experiments performed in this study utilized only monomer PSI, separation of monomer from the highly purified trimer was performed as previously descri bed 39.

Overexpression and purification of ferridoxin, PsaD and PsaE

Overexpression of PsaD from Mastigocladus laminosus was performed36 with minor changes of induction with 0.4 mM IPTG (isopropyl [3-D­thiogalactopyranoside, Sigma). Overexpression of PsaE was carried out with pET 22b as the host overexpressing vector (Novagen) which was used to transform the E. coli strain HMS 1 74(DE3). The IPTG concentration used for induction was 1 mM. The ferridoxin petF gene from Mastigocladus laminosus was cloned into pET 20b (Novagen) and the plasmid was used to transform the E. coli strain BL21 (DE3). Induction of ferridoxin expression was performed by the addition of 0.6 mM IPTG. After 3 hr (PsaD and PsaE) or 5 hr of shaking (ferridoxin) at 250 rpm, the cells were broken by sonication. Following centrifu­gation ( 13000 rpm for 20 min in SS-34 rotor) the soluble ' fraction was loaded on Ni-Agarose column (OIAGEN) pre-washed with 20 mM Tris-HCI pH 7.9 and 0.5 M NaC!. The PsaD and PsaE proteins were eluted with elution buffer (20 mM Tris-HCI pH 7.9, 0.3 M imidazole, 0.5 M NaCl). The eluted proteins were kept at -20°C till used.

Radiolabeling of PsaD and PsaE PsaD was labeled with e5S]methionine (40 �Ci/ml)

and PsaE was labeled with 1 �Cilml of C4C]valine. The labeled amino acids were supplied by Amersham. Bacterial cells harbouring the desired gene were grown in A Salts medium40 until reaching an A600 value of -0.6 OD. Then, IPTG was added and bacteria were shaken at 37°C for additional 1 .5 hr (PsaD) or 3 hr (PsaE). After 1 .5 hr rifampicin (200 �g/ml) was added to the bacterial culture expressing PsaD and was returned to the incubator for another 1 .5 hr. Following this, the radiolabeled amino acid was added to the solution and incubation was further

LUSHY et al. : AN INSIGHT INTO THE ASSEMBLY AND ORGANIZATION OF PHOTOSYSTEM I COMPLEX 409

protein at 280 nm, whenever the extinction coefficient was known. SDS-PAGE was performed according to the method of Schagger and Von Jagow43. For the determination of the N-terminal sequence of PsaK, the PSI proteins were blotted onto a PVDF membrane and stained with 0. 1 % Amido B lack in 50% methanol. The N-tenninal sequence was determined on an ABI 494A (Applied Biosystems) at the Protein Sequencing Facility in Technion, Israel Institute of Technology. Western blot analysis was performed by using polyclonal antibodies raised against spinach PsaD44. Polyclonal antibodies against Mastigocladus laminosus PsaE were obtained in 6-week old rabbits. Pure PsaE ( 1 mg) was mixed with the same volume of Freund' s complete adjuvant (Sigma) was injected to rabbits. Two boosts of 0.5 mg PsaE in incomplete Freund's adjuvant (Sigma) were injected at a month' s interval, about 4-5 weeks after the initial injection. Serum was obtained and treated45, and the antibodies were used for Western blot analysis. HRP (Horse Radish Peroxidase) anti-rabbit was used as a secondary antibody in all Western blot analysis.

Results As described above, PSI contains subunits

belonging to all the categories of membrane proteins: peripheral-cytosoVstromal facing proteins, peripheral­lumenal facing proteins, and integral proteins containing transmembrane helices embedded within the lipid bilayer. To characterize the membrane organization of representatives of these different classes of proteins in the PSI complex, the following experiments were conducted.

Assembly of the peripheral- cytosol facing subunits PsaD and PsaE of the PSI complex

Previous studies have shown that the peripheral­cytosoV stromal facing subunits of PSI, PsaD and PsaE do not require NTP' s or cytosolic factors for their proper assembly into the PSI complex31 ,34,46. These studies have also indicated that information concerning the stable assembly of these subunits into the thylakoids, resides within the polypeptide itself, Moreover, PsaD has been shown to integrate specifically into PSI-its final functional destination34,4 1 . Nevertheless, all the reported studies done in the past were performed with extremely small amounts of subunits used for characterization of their membrane assembly. In order to find out whether PsaD and PsaE undergo similar biogenesis pathways if substantial amounts (�g quantities) of these

subunits are used, which better resemble the in vivo situation, the two subunits were cloned and overexpressed in E. coli. Following purification, the proteins were incubated with thylakoid membranes isolated from the thermophilic cyanobacterium, Mastigocladus laminosus. Proteolytic digestion and treatment with chaotrophic agents which serve as the commonly used assays for characterizing the

bl f . . b 47-49 d assem y 0 protems mto mem ranes , were use

as our two criteria to determine proper assembly. High concentration of NaBr (2 M), and thermolysin proteolytic digestion were used to detennine the assembly of the overexpressed PsaD and PsaE. The results indicated that in solution, in the absence of thylakoids, PsaD and PsaE are both sensitive to proteolytic digestion by thermolysin. In both cases, the proteins became -3 kDa shorter following treatment with thermolysin and there was almost no trace of the bands representing the fully mature PsaD and/or PsaE proteins (Fig. l A). Following their thylakoid integration, both subunits became resistant to this thermolysin proteolysis (Fig. IB). To confirm that the newly assembled PsaD and PsaE became organized within the thylakoids in a similar way to the PsaD and PsaE subunits present in situ, isolated thylakoids from Mastigocladus laminosus has undergone the same treatments, and the detection of PsaD and PsaE was performed by Western blot analysis using polyclonal antibodies raised against PsaD and PsaE. The results showed that the PsaD and PsaE present in situ were resistant to proteolysis by thermolysin (Fig. l C), similar to the resistance observed following the in vitro assembly reaction with newly integrated PsaD and PsaE (Fig. lB) , Thus, it can be assumed that the overexpressed PsaD and PsaE were assembled in the thylakoid membranes in a conformation similar to that of the PsaD and PsaE existing in situ.

The organization of PsaD and PsaE within the reducing site of PSI and direct measurement of their association

It is well -documented that both PsaD and PsaE serve as important components of the reducing site of the PSI complex and are both involved in the binding and docking of the PSI electron acceptors, ferridoxinlflavodoxin to the complexl l.1 8. To directly reveal the interactions formed between PsaD and PsaE as well as their interactions with ferridoxin, the binding affinity (Kd) between the different proteins was detennined. The availability of overexpressed

4 10 INDIAN J BIOCHEM. BIOPHYS . . VOL. 37. DECEMBER 2000

The effect of proteolytic digestion and chaotrophic treatment on PsaD and PsaE:

A. In solution B. After in vitro assembly into C. Westem blot analysis of the

isolated thylakoids in situ present PsaD and PsaE

COD NaBr Therm COD NaBr Therm COD NaBr Therm

. -

PaaE �I"''' ... ... . . . ·. 1 . . .. .

. . . ........ . " .

1 ..... · ·· � .

:. -:':" - . : "'�.··".'· ' . I . . . ,- ," . " . - , - , . ' - , . ' . ''' ' . . - ' .

Fig. I - The assembly of PsaD and PsaE into thylakoids isolated from Mastigocladus laminosus. [To characterize the effects of proteolytic digestion (thermolysin) and washings with high concentration of chaotrophic agent (2 M NaBr), PsaD and PsaE were exposed to these treatments under different conditions. In order to confirm that the newly assembled subunits organized within the thylakoids in a similar manner to the PsaD and PsaE present in situ, thylakoids isolated from continuous culture were treated identically and the effects on the PsaD and PsaE present in situ was determined. (A): Overexpressed PsaD or PsaE ( 10,000 cpm each) in HM buffer (total volume of 50 Ill) were incubated for 30 min at 25°C in the absence of thylakoid membranes, i.e. in solution. FolIowing incubation, the samples were divided into 3 tubes: control (Con); treatment with 2 M NaBr (NaBr); treatment with 0.2 mg/ml thermolysin (Therm), (as described in Materials and Methods). Immediately after treatments the samples were denatured in a denaturing buffer, boiled for 5 min and electrophorized on SDS-PAGE on to which half of the reaction mixture was loaded. The gel was dried and autoradiographed. (B): Overexpresscd PsaD or PsaE (200,000 cpm each) were incubated with isolated, washed thylakoids (50 Ilg chlorophylI) for 30 min at 25°C. FolIowing incubation, the samples were washed 3 times in HM buffer and divided into 3 tubes: control (Con), wash with 2 M NaBr (NaBr) and thermolysin treatment (Therm). FolIowing treatments, the samples were re-washed 3 times in HM buffer (or in HM buffer containing 10 mM EDTA for the thermolysin treatment), and analyzed by SDS-PAGE ( 1 0 Ilg chI loaded on each lane), and the gel was autoradiographed. (C): Isolated thylakoids (50 Ilg chlorophylI) were washed 3 times in HM buffer. and then treated as described above. The samples were analyzed by SDS-PAGE ( 1 0 Ilg chlorophylI loaded on each lane). FolIowing transfer to PVDF, immunodetection with polycIonal antibodies against PsaD or PsaE was performed as described in Materials and Methods].

PsaD, PsaE and ferridoxin as well as the entire purified PSI complex, enabled the direct measurement of protein-protein interactions by using the B IACORE®3000 technology.

This preliminary study is the first to directly measure the binding constant (Kd) between PsaD and PsaE. PsaD has been previously shown to interact with PsaAIB50, PsaCI 2.2 1 , PsaE31 .34 and PsaL33.5 1 but no direct measurements between these subunits and PsaD had been performed. Our novel measurement of the PsaD-PsaE interaction has indicated a Kd value of 6.4 x lO

'8M (Table 1) . Such Kd value indicates that the interaction between the two subunits is specific, especially in light of the fact that both PsaD and PsaE are highly basic at pH 8.052,53, and even at this pH they still exhibit such high association (Kd = 6.4x lO-8M).

Since PsaD and PsaE are the reducing site components that play a major role in the binding of ferridoxin to PSI"' 1 8, we aimed at determining the specific contribution of each of these subunits to the binding of ferridoxin to PSI via the measurement of their direct association with ferridoxin. Regarding the Kd measured for PsaD or PsaE-ferridoxin binding was 5.8 x 10'" M and 3. 1 x 10-5 M respectively (Table l) .

The measured values are in good agreement with the hypothesized functions of PsaD and PsaE. In our direct measurement, PsaD which had previously been indicated as being essential for ferridoxin binding to PSI" , I 6, showed high affinity to ferridoxin (5.8 x 1 O-1 IM). Our Kd measurements for ferridoxin­PsaE interaction (3. 1 x 10'5 M) showed lower affinity of PsaE to ferridoxin than PsaD-ferridoxin. These findings are in good agreement with the previous

LUSHY et al. : AN INSIGHT INTO THE ASSEMBLY AND ORGANIZATION OF PHOTOSYSTEM I COMPLEX 4 1 1

Table 1 - The protein-protein interactions within the reducing site of PSI and with the PSI electron acceptor- ferridoxin

[The Kd value of protein-protein interactions was measured by BIACORE®3000. For each study the specified ligand was immobiiized to a chip (e.g. PsaD and PsaE were bound to the CM5 chip and ferridoxin was bound to the NTA chip). following the m�uf�cturer�s instructions. The binding of analyte to the immobilized ligand was performed at 4-5 different concentrations of analyte, as mdlcated m the table. Each binding reaction was repeated twice for each concentration measured. The Kd is presented in molar (M); the Chi2 value represents the statistical compatibility to the theoretical calculated value for each measurement as predicted by the BIACORE software. The Chi2 value obtained for the binding of the entire PSI complex and ferridoxin is very much higher than the value obtained in the other experiments. This high value is due to the huge difference in the molecular weight of ferridoxin and the entire PSI complex (see manufacturer's instructions)]

Binding reactions Analyte Ligand (immobilized on the chip) (passed through the chip)

Measured equilibrium constant, Kd (Kd=Kofl/KoIl) Name Range of concentration�

PsaD-PsaE PsaD-Fd PsaE-Fd Fd-PSI

PsaD PsaD PsaE Fd

PsaE Fd Fd PSI

studies in which it was suggested that PsaE is responsible for the dissociation of ferridoxin from PSI17.

To characterize the interactions and associations formed between the entire PSI complex (including PsaD and PsaE) and its electron acceptor ferridoxin, we measured the association of PSI with ferridoxin. The data obtained showed a Kd value of 2.8 x 10-9 M (Table 1 ). This Kd exhibits two orders of magnitude stronger association between PSI and its acceptor ferridoxin than the value reported previously (5 x 1O-7M)1 7. One should note, however, that Barth et at. 17 measured the two forms of ferridoxin, i .e. the oxidized as well as the reduced form and the result reported was the average of these two measurements. Here, only the association of PSI with oxidized ferridoxin was measured. The oxidized form exhibits stronger association with PSI than the reduced form does, thus showing lower Kd value.

Moreover, the binding affinity between PSI and ferridoxin is two orders of magnitude lower than the affinity measured for the binding of PsaD with ferridoxin (2.8 x 10-9 M vs 5.8 X 10-1 1 M). This suggests that some of the other PSI subunit(s) are responsible for lowering the binding affinity of ferridoxin to the entire PSI complex. The Kd measured for the binding of PsaE with ferridoxin (3. 1 x 10-5 M) indicates that the decrease in the binding affinity of ferridoxin to the entire PSI complex (2.8 x 10-9 M), vis-a-vis the stronger affinity of PsaD-ferridoxin binding (5.8 x 10-1 1 M), might be attributed to PsaE.

measured

0.2- 1 � 6.4 x 10-8 M 0. 165 0.2- 1 � 5.8 x 10-1 1 M 0. 1 67 0.2- 1 � 3 . 1 x 10-5 M 0. 141 8-40 nM 2.8 x 10-9 M 55.4

The assembly of PsaK, a membrane transversing subunit of PSI

In contradistinction to the knowledge available on the assembly of the peripheral PSI subunits facing the cyanobacterial cytosol, very little is known about the assembly of cyanobacterial integral membrane subunits of PSI. There is more information on the membrane integration of PSI integral subunits in plants. The best-characterized protein in this category is the chlorophyll alb binding protein (LHCI) which has been shown to integrate into the thylakoids via the SRP-dependent pathway54. As for the plant core complex (CCI) subunits, PsaF has been shown to utilize the Sec-dependent mechanism for its membrane translocation55.56, and PsaK has been characterized to assemble into barley thylakoids in a spontaneous way56.

We wanted to compare the membrane integration of an integral membrane trans versing subunit of the cyanobacterial PSI to our findings for the peripheral PsaD and PsaE subunits of the same PSI complex. As a first step, we cloned the gene encoding PsaK, our choice as model system of membrane trans versing subunit. Mastigocladus laminosus PsaK was cloned as described under 'Materials and Methods' . The open reading frame was found to encode a hydrophobic precursor protein with a calculated molecular mass of 1 3 kDa. The initial codon is ITO instead of ATO, with a standard ribosome binding site (AOOA) 1 3bp upstream of the starting codon. The psaK gene was identified by the similarity of its encoded polypeptide to PsaK from other cyanobacteria or chloroplasts (Fig. 2A). The psaK gene encodes an 87 amino acid

412 INDIAN J BIOCHEM. BIOPHYS., VOL. 37, DECEMBER 2000

protein predicted to have two membrane transversing helices (Fig. 2B), with a predicted molecular mass of 9321 .88 Da and pI of - 10.44.

acid sequence (MLTSTLLA-Fig. 2A) were possibly removed after translation.

As a first step in the characterization of the assembly of Mastigocladus laminosus PsaK, we examined the requirement for the presence of cytosolic factors. PsaK was incubated with thylakoids

The N-terminal sequence of PsaK present in the PSI complex was determined as AATTPLQWSPT. Thus, the first eight amino acids of the deduced amino

A PSAK JM.SLA. 1 PSA.xPA VA. I FSA.XJ'TNEN I S. ,Ion II Ills 1 PSA.¥C:rm' J I PSA. _CHLRE 1 PSA.XJ/ORVU 1 PSAK U4.S LA. 13 PSA.xPA VA. 5 PSA.X_S'TNEN 8 S. tlonllills 4 PSA.X S:rm' J 13 PSA.X:CH J.RZ 30 PSA.X_HORYU 45 PSAK.JM.S LA. 53 PSA.X.fi1'/A. VA. 0 PSA.X_S'TNEN 50 S. ,lonllills 46 PSA.XJ:rm' J 53 PSA.X CH J.RZ 15 PSA.KJ/ORVU 89

B 1 . 2

� 9 . 9 .. ... .. ... .II " 9 . 6 .Q 0 '-Go

9 . 4

9 . 2

8

· . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . fU T S ·T L L IT T,

12 · . . ' . . . • . . . . . . . . . . . . • . . . • • • . . . . . . . • . . . . . . . T·T 4 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · IIIV ' ·

, L I' 1 · , , , , , , , , , , , , , , , , , . , , , , , , , , , , , , , , , , , , , , , , , T ,x. r: 3 : : : : MQAiATitl : : S A i it: : : : l TKAA : : : : : : �:�:�:3'�: �� MAS Q L S AM T S V! Q r XG-L RT Y S S ! RS MAT L I'S L RRR I.S Q e l R .CD · J' 44

: : : : : ii E ' ' : ' t Ei' " ' T il'G" . ;. , . ' YSI' iQY�Ai : : : : : " , • , Lr B B L L' T S r ' & ' vs 't , , · · , " ' , · , , . . . , ri' o :l L 'TAS 'f . I G I L AlIli " , , . G-VI S liD! J;.VD G-j; G 1 v t KG I At " , , · G- t Q T G- D I" WADI c ia. 1 M I VI: K.I L D Q I I G-

t r ii n s rA l1 m b r ii n e - _

i n s i d� -

I I -\ f

I ! l I ('

i 21 3 9

��rll1-1 .�\

49 SI pos i t i on

I

I l

I I I I I 68 78

Fig. 2 - The PsaK subunit of Mastigocladus laminosus PSI [The Mastigocladus laminosus HindIII genomic library was screened with a probe of Synechocystis sp. PCC 6803 psaKI coding sequence. A 259bp fuJI coding region of psaK was identified according to its similarity to other psaK genes. (A): The deduced amino acid sequence of the isolated psaK gene from Mastigocladus laminosus (PSAK_MASLA) was compared to deduced amino acid sequences of psaK from other organisms: PSAK_ANA V A· Anabaena variabilis (68% identity); PSAK_SYNEN­Synechococcus elongatlls naegeli (41 % identity); PSAK_SYNY3- Synechocystis sp. PCC 6803 (45% identity); PSAK_CHLRE­Chlamydomonas reinhardtii; PSAK_HORVU- Hordeum vulgare; S. elongatus- the sequence of PsaK from Synechococcus elongatus. (B): Hydropathy plot of the PsaK polypeptide sequence. The plot was predicted using the TMHMM 1 .0 software57].

LUSHY et aL.: AN INSIGHT INTO THE ASSEMBLY AND ORGANIZATION OF PHOTOSYSTEM I COMPLEX 4 1 3

in the presence or absence of Mastigocladus laminosus cytosol extract. Following the washing of the thylakoids, the membranes were treated with proteinase K (Fig. 3). The results strongly suggest that for its proper assembly into the thylakoid membranes, PsaK does not require the presence of any cytosolic factors. To unravel whether NTP's are required for the stable assembly of PsaK, we performed the assembly reaction in the presence of apyrase. Apyrase abolishes residual NTP' s remaining from the in vitro translation of the protein. The results (Fig. 4) clearly indicate that the presence of apyrase has no effect on the efficiency of PsaK assembly into the thylakoids. To verify the use of Sec-dependent or �pH-dependent pathways and to eliminate the possibility of having residual cytosol fraction associated with the isolated thylakoids, the assembly of PsaK was performed in

the presence of azide or valinomycin. Azide is the inhibitor of the SecA protein involved in the Sec pathway while valinomycin abolishes the �pH, in case proton gradient exists across the thylakoid membrane. The results (Fig. 4) showed that the presence of either of these inhibitors had no effect on the assembly of PsaK as indicated by the fact that in all experiments presented in Fig. 4, the amount of PsaK in the thylakoids, following proteolysis, was practically the same. Taken together, the different experiments strongly favor the fact that Mastigocladus laminosus PsaK, like the barley PsaK56, assembles into the thylakoid spontaneously, i.e. not the Sec-dependent, nor the �pH-dependent neither the SRP-dependent mechanisms are involved.

The three subunits characterized in this study PsaD, PsaE and PsaK exhibit a similar spontaneous mechanism

PsaK After in vitro assembly into in solution isola ted thylakoids

I I I Cytosol

I I

+

I

+

I

+

I

ProteinaseK + + + +

1 2 3 4 5 6 MW

Fig. 3 - The assembly of Mastigocladus laminosus PsaK into isolated thylakoids. [ln vitro translated PsaK (200,000 cpm) was incubated without (lanes 1 -2) or with (lanes 3-6) isolated thylakoids (an amount equivalent to 25 ).lg chlorophyll) pre-washed with 0. 1 M Tricine­NaOH pH 8.0. The assembly reaction was carried out for 30 min at 55°C (at a total volume of 50 ).ll), in the presence (+) or absence (-) of Mastigocladus laminosus cytosol fraction (3 mg/mI). Following incubation, the samples were divided into two tubes: one was treated (+) with 50 ).lg/ml proteinase K (as described in Materials and Methods), and the other was the untreated control (-). Following the proteolytic treatment, the thylakoids were washed with buffer containing 10 mM EDTA and the membrane integration analysis was performed by SDS-PAGE (5 ).lg chlorophyll loaded on each lane) and autoradiography.]

4 14 INDIAN J BIOCHEM. BIOPHYS., VOL. 37, DECEMBER 2000

Buffer Apyrase Valino Azide MW I I I I I Tp

+ I + I + I + I Proteinase K I 1 2 3 4 5 6 7 8 9

;� .

Fig. 4 - The assembly of PsaK into thylakoids isolated from Mastigocladus laminosus in the presence of different inhibitors. [In vitro translated PsaK (200,000 cpm of translation product (Tp)- lane 1) was incubated with isolated washed thylakoids (25 Ilg chlorophyll) in the presence of 1 0 mM methionine. The incubation was carried out for 30 min at 55°C. Immediately after incubation the sample was divided into two tubes: half was kept as untreated control (lane 2), and the other half was treated with 50 Ilg/ml proteinase K (lane 3). A similar reaction was carried out in the presence of different inhibitors: 1 U of apyrase was added to the reaction mixture under the same incubation conditions (Apyrase). Half of the apyrase treated sample was loaded before (lane 4) or after proteinase K treatment (lane 5). Valinomycin (5 mM valino) was present during the membrane assembly reaction. Following this incubation, the thylakoids were treated without (lane 6) or with (lane 7) proteinase K. Sodium azide ( 1 0 mM NaNr Azide) was added to the reaction and incubation was performed as described. The azide incubated membranes were then treated without (lane 8) or with (lane 9) proteinase K. The total volume of all the reaction mixtures was 50 III and the membrane integration was analyzed by SDS-PAGE (5 Ilg chlorophyll loaded on each lane) fol lowed by autoradiography].

for integration into the thylakoid membrane. Since PsaK is an integral protein buried within the bilayer, we aimed at analyzing whether the membranal state affects the protein integration. Since it is well documented that the membrane fluidity is highly influenced by temperature, and especially in light of Mastigocladus laminosus being thermophilic (grown at 55°C), we performed the assembly reaction with in vitro translated PsaK, under different temperatures.

Temperature dependent assembly of the membrane transversing PsaK subunit

The in vitro assembly of the PsaK subunit was investigated under different temperatures (Fig. 5) . PsaK was integrated into the thylakoids at all temperatures tested but the assembly was maximal at 55°C (Fig. 5) . When the assembly was performed with in vitro translated PsaL, the same phenomenon was observed as the integration of PsaL was maximal at 55°C (data not shown). The fact that the integral membrane subunits exhibit a more efficient membrane integration at higher temperature may

suggest that this type of protein needs the membrane to be at high fluidity state for its efficient integration.

Discussion The present study aimed at the molecular

characterization of the in vitro assembly of different PSI subunits into thylakoids isolated from Mastigocaldus laminosus. In order to characterize the organization of these subunits within the PSI complex, attempts were made to unravel, on a quantitative level, the interactions/associations between these PSI subunits as well as between them and the PSI electron acceptor-ferridoxin.

Characterization of assembly of overexpressed peripheral PsaD and PsaE subunits indicated that the assembly of substantial amounts (Ilg quantities) of these proteins resembles the pathway characterized earlier with in vitro translated proteins. No requirement for NTP's or cytosolic factors was identified (Fig. 1 ). The newly assembled PsaD and PsaE were resistant to digestion by thermolysin, similar

LUSHY et al. : AN INSIGHT INTO THE ASSEMBLY AND ORGANIZATION OF PHOTOSYSTEM I COMPLEX 4 1 5

25 37 55 75 90°C ......... " ��' ' �" " � ' ,��·;;��;;kZ '

"31; >".x' ",

1 000 c:: 900 V

... c 800 V 0 ! .. 700 V c e ::s Q. 600 V 0 E " 41 500 V IV :Q CD 400 > E 300 V +l 41 IV II) r--e II) 200 V IV 9 1 00 V

0 / 25 37 55 75 90 temperature during assembly

Fig. 5 -The optimum temperature for the thylakoid integration of PsaK is 55°C [In vitro translated PsaK (200,000 cpm) was incubated with isolated, washed thylakoids (amount equivalent to 25 llg chlorophyll) in the presence of 1 0 mM methionine. The incubation was carried out for 30 min at different temperatures. After the incubation, all samples were treated with 50 IlglmI proteinase K and subjected to SDS-PAGE after denaturation. The gel was dried and exposed to Fuji imaging plate for 1 2 hr (upper panel). The quantitation and determination of the relative amount of PsaK assembled in the thylakoids was performed by scanning the autoradiogram in Bio-imaging analyzer (FUJIX BAS 1 000) . The TINA version 2. 1 0 g software was used for calculating the relative amount of protein by the net absorbance of each band (graph)].

to the subunits present in the Mastigocladus laminosus thylakoids in situ (Figs. 1 B and IC). This resistance resembles the proteolytic resistant forms of the mature PsaD and PsaE subunits in the , thylakoids of plants37,58. The in vitro thylakoid assembled precursor of spinach PsaD was found to be susceptible to thermolysin proteolysis. Following its processing to mature-PsaD, the protein became resistant to proteolysis. Only the -2 kDa N-terminus extension of the eukaryotic PsaD was removed by thermolysin37. The remaining resistant form of PsaD (mature-less-extension) in the spinach thylakoids, which exhibited resistance to thermolysin proteolysis, is similar to the protected mature PsaD and PsaE found in Mastigocladus laminosus thylakoids (Figs. lB and l C). This may further indicate that only the evolutionary conserved sequence of the mature PsaD is important for the functionality and associations of the protein with other members of the PSI core complex (CCI).

To better characterize the interactions formed between the peripheral subunits, PsaD and PsaE, as well as between them and the PSI electron acceptor

ferridoxin, quantitative measurements were performed with the BIACORE technology (Table 1) . Previous studies have shown the existence of interaction between PsaD and PsaE31 ,34. However the present study, for the first time, determined these interactions quantitatively. The Kd value obtained (6.4 x 1O�8 M) implies that the association between these two subunits is highly specific, especially in light of the fact that both subunits are highly basic at pH 8 .052,53. The results obtained from the measurement of the PsaD association with ferridoxin (5 .8 x 10-1 1 M) indicated a very high binding affinity, thereby supporting the postulated function of PsaD being the site for ferridoxin binding to PSI. One should note, though, that this Kd value is higher than the one measured for the association of the entire PSI and ferridoxin (2.8x1O-9 M). Thus, it is reasonable to assume that other subunits of the PSI complex reduce the binding affinity of PSI and ferridoxin, thereby allowing the dissociation of ferridoxin from the complex. The Kd value obtained for the interaction of PsaE with ferridoxin (3. 1 x 10-5 M) indicated a low affinity between these two proteins, supporting the

4 1 6 INDIAN J BIOCHEM. BIOPHYS., VOL. 37, DECEMBER 2000

possibility of PsaE being one of the subunit(s). responsible for the dissociation of ferridoxin from PSI, as suggested before1 7•

In contradistinction to available infonnation on the assembly of the peripheral PSI subunits, very little is known about the assembly and organization of the cyanobacterial integral membrane trans versing subunits of the PSI complex. PsaK was chosen as our model protein for the characterization of a PSI integral membrane spanning subunit.

The psaK gene was cloned and characterized (Fig. 2). The deduced amino acid sequence from the isolated gene showed an extension of eight amino acids (MLTSTLLA) at the N-terminus of the protein, as compared to the N-terminal sequence of the in situ present PsaK subunit in isolated PSI complexes. This extension resembles the bacterial signal sequences of secreted proteins59, characterized by hydrophobic core followed by the sequence A-X-A that usually serves as the peptidase recognition signal. The function of PsaK N-tenninal extension needs further characterization, yet it is highly possible that this extension serves as thylakoid targeting signal.

To characterize the integration of Mastigocladus laminosus PsaK into the thylakoid membranes, in vitro translated PsaK was incubated with isolated thylakoids. The results indicated that PsaK, similarly to Mastigocladus laminosus PsaD and PsaE, assembled into the thylakoids in a spontaneous way (Figs. 3 and 4). Such spontaneous assembly pathway has been previously described for peripheral thylakoid proteins3 1 •34. Moreover, integral thylakoid proteins that contain a single transmembrane domain, like the CFo subunit II (CFoII) of the ATP synthase complex and the PsbX and PsbW subunits of PSII were also found to integrate into the thylakoids by this spontaneous mechanism60-62. One should note, though, that all these integral proteins posses a single transmembrane domain. The present study is the first report indicating that a protein containing multiple membrane trans versing domains can integrate into the thylakoids spontaneously (Figs. 3 and 4). PsaK is predicted to have two transmembrane a helices (Fig. 2B). This finding contradicts the general assumption that the nature of a protein, peripheral or integral, with one or multi-transmembrane domains, dictates the mechanism by which the protein is assembled into the membrane.

We further characterized the assembly of the PsaK subunit, trying to determine whether the state of the

membrane influences the integration of this multi­membrane spanning protein into the thylakoids. Since it is well documented that higher temperature ensures higher membrane fluidity and in light of Mastigocladus laminosus being thermophilic cyanobacteria, with optimal growth temperature at 55°C, we perfonned the PsaK assembly reaction under different temperatures (Fig. 5) . Our experiment revealed that PsaK was assembled into the thylakoids at all temperatures tested, but the optimal, most efficient assembly was observed at 55°C (Fig. 5). The high temperature probably secured the fluidity of the membrane to a degree allowing an easier integration of the protein that needs to transverse the bilayer.

The sequence analysis of different Mastigocladus laminosus PSI subunits indicate them to be more closely related to another filamentous cyanobacteria, rather than to another thermophilic cyanobacteria (besides PsaK, the amino acid sequences of PsaAIB63, PsaD64, PsaE65 and PsaL66 are known). This fact strongly suggests that the basis for thennostability resides in the secondary and/or teritiary structures of the protein rather than in the primary structure, i.e. the sequence of the polypeptide itself. Alternatively, the thennostability of the PSI complex from Mastigocladus laminosus may result from the surrounding environment of the subunits like the composition of its surrounding lipids and the state of the bilayer.

The novel molecular and biophysical methodologies used in the present study allowed a better understand­ing 6f the membrane assembly of subunits belonging to the same membrane protein complex, PSI from Mastigocladus laminosus. Direct quantitative determination of protein-protein interactions provided additional insight into the organization within the membrane complex and with its electron acceptor.

Acknowledgement This work was supported by the NIH grant No.

l RO I GM53 104 and by the Magnet Foundation for the DA' AT Consortium from the Israeli Ministry of Industry and Trade. We thank Maisie Berman for her advice on English syntax.

References I Schatz G & Dobberstein B ( 1 996) Science 27 1 , 1 5 19- 1 526 2 Lill R, Nargang F E & Neupert W ( 1 996) Curr Opin Cell Biol

8, 505-5 1 2 3 Corsi A K & Schekman R ( 1 996) J Biol Chern 27 1 , 30299-

30302 4 Wickner W & Leonard M R ( 1 996) J Biol Chern 27 1 , 295 1 4-

295 1 6

LUSHY et aL.: AN INSIGHT INTO THE ASSEMBLY AND ORGANIZATION OF PHOTOSYSTEM I COMPLEX 4 17

5 Wollman F A, Minai L & Nechushtai R ( 1 999) Biochim Biophys Acta 141 1 , 2 1 -85

6 Golbeck J H ( 1 993) Proc Natl Acad Sci USA 90, 1642- 1646 7 Chitnis P R, Xu Q, Chitnis V P & Nechushtai R ( 1 995)

Photosynth Res 44, 23-40 .

8 Fromme P ( 1996) Curr Opin Struct Bioi 6, 473-484 9 Chitnis P R ( 1 996) Plalll Physiol I l l , 661-669

1 0 Klukas 0, Schubert W D, Jordan P, Krauss N, Fromme P, Witt H T & Sanger W ( 1999) J Bioi Chern 274, 735 1 -7360

I I Hanley J, Setif P, Bottin H & Lagoutte B ( 1 996) Biochemistry 35, 8563-857 1

12 Chitnis V P, Jung Y S, Albee L, Golbeck J H & Chitnis P R (1996) J Bioi Chern 27 1 , 1 1 772- 1 1 780

13 Zanetti G & Merati G ( 1 987) Eur J Biochem 1 69, 143-146 14 Zilber A L & Malkin R ( 1 988) Plant Physiol 88, 8 1 0-8 14 15 Lelong C, Setif P, Lagoutte B & Bottin H ( 1 994) J Bioi Chern

269, 10034- 10039 16 XU Q, Jung Y S, Chitnis V P, Guikema J A, Golbeck J H &

Chitnis P R ( 1994) J Bioi Chern 269, 2 1 5 12-2 1 5 1 8 1 7 Barth P, Lagoutte B & Setif P ( 1 998) Biochemistry 37, 16233-

16241 1 8 Rousseau F, Setif P & Lagoutte B ( 1 993) EMBO J 12, 1 755-

1765 19 Muhlenhoff U, Zhao J & Bryant D A ( 1996) Eur J Biochem

235, 324-33 1 20 Kruip J, Chitnis P R, Lagoutte B, Rogner M & Boekema E J

( 1997) J Bioi Chern 272, 1706 1 - 1 7069 2 1 Li N, Zhao J, Warren P V, Warden J T, Bryant D A &

Golbeck J H ( 199 1 ) Biochemistry 30, 7863-7872 22 Andersen B, Koch B & Scheller H V ( 1 992) Physiol Plalll 84,

154- 16 1 23 Y u L, Zhao J , Muhlenhoff U, Bryant D A & Golbeck J H

( 1993) Plant Physiol 103, 1 7 1 - 180 24 Zhao J, Snyder W B, Muhlenhoff U, Rhiel E, Warren P V,

Golbeck J H & Bryant D A ( 1 993) Mol Microbiol 9, 1 83-194 25 Chitnis P R, Purvis D & Nelson N ( 199 1 ) J Bioi Chern 266,

20 146-2015 1 26 Hippler M , Reichert J , Sutter M, Zak E , Altschmied L,

Schroer U, Herrmann R G & Haehnel W ( 1996) EMBO J 15 , 6374-6384

27 Farah J, Rappaport F, Choquet Y, Joliot P & Rochaix J D ( 1995) EMBO J 14, 4976-4984

28 XU Q, Yu L, Chitnis V P & Chitnis P R ( 1 994) J Bioi Chern 269, 3205-32 1 1

29 Schluchter W M, Shen G, Zhao J & Bryant D A ( 1996) Photochem Photobiol 64, 53-66

30 XU Q, Hoppe D, Chitnis V P, Odom W R, Guikema J A & Chitnis P R ( 1995) J Bioi Chern 270, 1 6243- 1 6250

3 1 Cohen Y, Chintis V P, Nechushtai R & Chitnis P R ( 1993) Plant Mol Bioi 23, 895-900

32 XU Q, Odom W R, Guikema J A, Chitnis V P & Chitnis P R ( 1 994) Plant Mol Bioi 26, 29 1 -302

33 Chitnis V P & Chitnis P R ( 1 993) FEBS Lett 336, 330-334 34 Chitnis P R & Nelson N ( 1992) Plant Physiol 99, 239-246 35 Cline K & Henry R ( 1 996) Annu Rev Cell Dev Bioi 12, 1 -26 36 Jin P, Sun J & Chitnis P R ( 1999) Biochim Biophys Acta

14 10, 7- 1 8

3 7 Minai L , Cohen Y , Chitnis P R & Nechushtai R ( 1 996) Proc Natl Acad Sci USA 93, 6338-6342

38 Castenholz R W ( 1 967) Nature, London 2 1 5 , 1285- 1 286 39 Almog 0, Shoham G, Michaeli D & Nechushtai R ( 1991 )

Proc Natl Acad Sci USA 88, 53 1 2-53 1 6 40 Davies B D & Mingioli E S ( 1950) J Bacteriol 60, 17-28 41 Cohen Y, Steppuhn J, Herrmann R, Yalovsky S & Nechushtai

R ( 1 992) EMBO J I I , 79-85 42 Amon D ( 1949) Plant Physiol 24, 1 - 1 4 4 3 Schagger H & Von Jagow G ( 1987) Anal Biochem 1 66, 368-

379 44 Nechushtai R & Nelson N ( 1 985) Plant Mol BioI 4, 377-384 45 Harlow E & Lane D ( 1 988) Antibodies - A laboratory

manual, pp 1 17-1 19, Cold Spring Harbor Laboratory, USA 46 Cohen y, Nelson N, Chitnis P R & Nechushtai R ( 1 995)

Photosynth Res 44, 1 57-164 47 Chintis P R, Nechushtai R & Thornber J P ( 1 987) Plant Mol

BioI 10, 3 - 12 48 Reed J , Cline K, Stephens L , Bacot K & Viitanen P ( 1990)

Eur J Biochem 194, 33-42 49 Yalovsky S, Paulson H, Michaeli D, Chitnis P R &

Nechushtai R ( 1 992) Proc Natl Acad Sci USA 89, 5616-5619 50 Xu Q & Chitnis P R ( 1 995) Plant Physiol 108, 1067-1075 5 1 XU Q, Armburst T S, Guikema J A & Chitnis P R ( 1 994)

Plant Physiol 1 06, 1 057- 1 063 52 Chitnis V P, Ke A & Chitnis P R ( 1 997) Plant Physiol 1 1 5,

1 699- 1 705 53 Falzone C J, Kao Y H, Zhao J, Bryant D A & Lecomte J T

( 1 994) Biochemistry 33, 6052-6062 54 Walter P & Johnson A E ( 1 994) Annu Rev Cell Bioi 10, 87-

1 19 55 Kamauchov I, Cai D. Schmidt I, Herrmann R G & Klosgen R

B ( 1994) J BioI Chern 269, 3287 1 -32878 56 Mant A, Nielsen V S, Knott T G, Moller B L & Robinson C

( 1 994) J Bioi Chern 269, 27303-27309 57 Sonnhammer E L L, Von Heijne G & Krogh A ( 1 998) in Proc

Of Sixth Int Conf On Intelligent Systems for Molecular Biology (Glasgow, J, Littlejohn, T, Major, F, Lathrop, R, Sankoff, D & Sensen, C, Eds), pp 1 75- 1 82, Menlo Park, CA: AAAI Press, USA

58 Zilber A L & Malkin R ( 1 992) Plant Physiol 99, 901 -9 1 1 59 Halpin C, Elderfield P D, James H E, Zimmermann R, Dunbar

B & Robinson C ( 1 989) EMBO J 8, 3917-392 1 60 Herrmann R G, Steppuhn J, Herrmann G S & Nelson N

( 1 993) FEBS Lett 326, 192- 198 6 1 Michl D, Robinson C, Shackleton J B, Herrmann R G &

Klosgen R B ( 1 994) EMBO J 13, 1 3 1 0- 1 3 1 7 62 Kim S J , Robinson C & Mant A ( 1998) FEBS Lett 424, 105-

108 63 Sun J, He Z Y, Nechushtai R & Chitnis P R ( 1 998) Plant

Physiol 1 16, 1 192 (accession No AF038558) 64 Jin P, Sun ], Nechushtai R & Chitnis P R ( 1997) Plant Physiol

1 15, 1288 (accession No U975 1 8) 65 He Z Y, Chitnis P R & Nechushtai R ( 1 998) Plant Physiol

1 2 1 , 1057 (accession No AF093820) 66 He Z y, Chitnis P R & Nechushtai R ( 1998� Plant Physiol

1 1 6, 868 (accession No AF030003) I