Excitation energy transfer in aggregates of Photosystem I and ...

9 downloads 0 Views 122KB Size Report
between the Photosystem II and Photosystem I peripheral proteins, also play a significant role in their association ... 1986 and Williams and Allen 1987). The first ...
Photosynthesis Research 64: 199–207, 2000. © 2000 Kluwer Academic Publishers. Printed in the Netherlands.

199

Regular paper

Excitation energy transfer in aggregates of Photosystem I and Photosystem II of the cyanobacterium Synechocystis sp. PCC 6803: Can assembly of the pigment-protein complexes control the extent of spillover? Silvina Federman1 , Shmuel Malkin2 & Avigdor Scherz1,∗ Departments of 1 Plant Sciences and 2 Biological Chemistry, The Weizmann Institute of Science, Rehovot 76100, Israel; ∗ Author for correspondence (e-mail: [email protected]; fax: +972-8-9344181) Received 10 December 1999; accepted in revised form 28 June 2000

Key words: cyanobacteria, excitation energy transfer, membrane proteins assembly, Photosystem I, Photosystem II, spillover, state transitions

Abstract The fluorescence profile of Photosystem I/Photosystem II mixtures in different solvent systems shows that both non-hydrophobic and hydrophobic interactions govern their association and control energy transfer from Photosystem II to Photosystem I. The non-hydrophobic interactions lead to a highly efficient excitation energy transfer from Photosystem II to Photosystem I. In view of this, we propose that similar non-hydrophobic interactions, between the Photosystem II and Photosystem I peripheral proteins, also play a significant role in their association in thylakoids that control state transitions in cyanobacteria. Abbreviations: Chla – chlorophyll a; Chlb – chlorophyll b; CMC – critical micelle concentration; DM – n-dodecylβ-D-maltoside; FPSI – fluorescence intensity from Photosystem I; FPSII – fluorescence intensity from Photosystem II; Hepes – N-2-Hydroxypiperazine-N0-2-ethanesulphonic acid; LHC II – light harvesting complex II; OD – optical density; PS I – Photosystem I; PS II – Photosystem II; RC – reaction center Introduction Oxygen evolving photosynthetic organisms function by the joint action of two photosystems: Photosystem I (PS I) and Photosystem II (PS II). Each comprises antenna pigment-proteins and reaction center (RC) pigment-proteins, where solar light energy is harvested and converted to electrochemical potential. In order to obtain the maximum photosynthetic efficiency, the integral excitation of the two photosystems must be equal. Preferential excitation of one of the photosystems leads to dissipation of the excess energy. In order to balance the relative activities of PS II and PS I, the photosynthetic organisms developed longand short-term adaptation mechanisms. A state transition is a short-term adaptation mechanism that regulates the distribution of excitation energy between the two photosystems. Briefly, it reor-

ganizes the two photosystems and peripherial antennas, to allow the redistribution of excess excitation energy. This phenomenon was described by Murata (1969, 1970) and by Bonaventura and Myers (1969), who showed that exposure of the organisms to light predominantly absorbed by PS II resulted in the shuffling of excess excitation energy from PS II to PS I. In contrast, exposure to light predominantly absorbed by PS I brought the photosystems back to an imbalanced state of light distribution in favor of PS II. Two mechanisms were suggested for state transitions in oxygenic photosynthesis (reviewed by Fork and Satoh 1986 and Williams and Allen 1987). The first, known as spillover, controls excitation energy transfer from PS II Chla to PS I Chla. The second mechanism relies on a change of the absorption cross sections of PS I and PS II (Canaani and Malkin 1984), by mobilization of phosphorylated or dephosphorylated antenna

200 complexes between PS I and PS II (for a review see Allen 1992). Both mechanisms involve dissociation of light-harvesting and reaction center complexes, and protein reassociation, such that the energy distribution between the two photosystems is modified. Eukaryotic red algae and prokaryotic cyanobacteria carry out photosynthesis in a fashion similar to that of higher plants, but differ in their antenna pigments and thylakoid membrane organization (Glazer and Melis 1987). More specifically, LHC II is replaced by phycobilisomes, which are large extrinsic complexes bound to the protoplasmic surface of the thylakoid membrane of cyanobacteria. Grana thylakoid regions do not form in cyanobacteria and, therefore, in contrast to higher plants, PS I and PS II are not partitioned into different regions. Phycobilisomes containing organisms undergo state 1–state 2 transitions similarly to higher plants. However, they involve larger changes in excitation energy distribution and take place much more rapidly (Fork and Satoh 1986). The mechanism of state transitions in these organisms is still a subject of much controversy, and studies supporting both antenna mobilization and spillover mechanisms were reported. Light-dependent phosphorylation of 15 kDa and 18.5 kDa polypeptides accompanied by a change in 77 K fluorescence of cyanobacteria cells was reported by Allen et al. (1985), Sanders et al. (1986) and Harrison et al. (1991). These polypeptides were identified as components of PS II and phycobilisomes, respectively. After these findings, a mechanism for state transitions analogous to higher plants was proposed. The mechanism was supported by picosecond time-resolved spectroscopy of Synechocystis 6301 cells adapted to state 1 and state 2 (Mullineaux et al. 1990). Further support was provided by fluorescence induction measurements (Mullineaux and Holzwath 1990) and laser-induced optoacoustic spectroscopy (Mullineaux et al. 1991). Recently, based on studies of fluorescence recovery after photobleaching (FRAP), Mullineaux et al. (1997) reported that when state transitions occur, phycobilisomes rapidly move, but not PS II. These findings and reports on the interaction between phycobilisomes and PS I (reviewed in Bald et al. 1996), led some researchers to finally conclude that upon phosphorylation, phycobilisomes are disconnected from PS II, subsequently migrating toward PS I and interacting with it. In contrast, strong evidence for the role of spillover was provided in many studies (Ley and Buttler 1980; Bruce et al. 1985, 1986; Malkin et al. 1990), in par-

ticular those by Bruce et al. (1989) and Vernotte et al. (1990), which showed light-induced state transitions and redox-induced state transitions in wild type and phycobilisome-less mutants of cyanobacteria. Their 77 K fluorescence spectra showed efficient redistribution of energy absorbed by Chla. Fluorescence transient measurements showed that state transitions were controlled by changes in the efficiency of energy transfer from PS II to PS I (spillover), rather than by changes in the association of the antenna to PS II (called ‘the mobile antenna model’). Since this model predicts that the distribution of energy absorbed by the phycobilisomes is regulated by state transitions, whereas energy absorbed by Chla is not, it was suggested that there is a conformational change in which the PS II complex moves in relation to PS I, changing the amount of spillover from PS II Chla to PS I Chla. Furthermore, ultrastructural freeze-fracture studies of mutated cyanobacteria showed randomization of PS II particles in thylakoid membranes in state 2 that were similar to wild type (Olive et al. 1986). We studied an in vitro PS I/PS II system in detergent solutions, nearly free of phycobilisomes. We modified the aggregation state of the photosystems either by changing the detergent/protein ratio in the solution or by dilution of the sample. Both modifications led to changes in the relative fluorescence intensities of PS II and PS I at 77 K, after chlorophyll excitation, similar to the changes observed in state transitions in cyanobacteria. We showed that PS I and PS II could spontaneously aggregate and thereby allow for spillover of excitation energy. Moreover, the amount of spillover can be modified by controlling the aggregation state of the complexes. The interaction is partly controlled by the protein environment.

Materials and methods Preparation of PS I/PS II mixtures Pellets enriched with PS II, which contained a significant portion of PS I, were obtained by treatment of thylakoids, as suggested by Oren-Shamir et al. (1995). The pellets were solubilized in a buffer containing 20 mM Hepes pH 6.5, 500 mM mannitol, 10 mM CaCl2 , 10 mM MgCl2, and 0.05% DM to make a final chlorophyll concentration of 500 µg/ml.

201 Modification of the detergent/protein ratio Each PS I/PS II mixture in buffer was solubilized in different concentrations of detergent (DM), such that different detergent/protein ratios were obtained, while the Chl content was kept constant ([Chl] = 500 µg/ml). This differential solubilzation was performed by adding different volumes of a stock solution of 9% DM (w/v) in the same buffer. The dilution effect of the added detergent volume (up to 2.75 times in a sample containing 5.75% DM) was examined by adding an equivalent volume of buffer solution + 0.05% DM. Sample concentration The samples were concentrated by spinning Centripreps 10 kDa (Amicon) at 3000 g, using a cooled centrifuge at 4 ◦ C. Spectral analysis The fluorescence spectra at 77 K were recorded using a home built sample holder. The sample was loaded into a 0.1-mm path length cuvette. After aligning inside a transparent dewar to front-face position, the sample was immediately frozen by filling the dewar with liquid nitrogen. The surroundings of the dewar were flushed with dry nitrogen gas. Fluorescence emission spectra, induced by excitation at 432 nm, were recorded by an SLM-8000 spectrofluorimeter (Aminco). Excitation and emission slits were fixed at 8 nm. The fluorescence intensity of PS II (FPSII ) was calculated by integrating the spectrum at 680–695 nm. In order to calculate the intensity of the PS I emission (FPSI ), we subtracted the emission spectrum of pure PS II from the spectrum of the sample and integrated the difference spectrum. We defined a quantity R, such that: R = FPS II /FPS I Because of the front-face positioning, the nature of the measuring cuvette, and the possible precipitation of part of the sample, the fluorescence of each sample was measured many times. Repeated measurements of the same sample could be ended in significant variation of the fluorescence overall intensity. However, the emission lineshape and the corresponding R values for a particular set of protein/detergent and protein content values remained constant (see inset of Figure 2). In fact, the variation of the overall fluorescence intensity (estimated by integrating the complete spectrum of emission) was the same among samples of

Figure 1. (A) The fluorescence spectrum of a PS I/PS II mixture, after excitation at 605 nm. [Chl] = 500 µg/ml in 0.05% DM. (B) The absorption spectrum of a PS I/PS II mixture containing [Chl] = 500 µg/ml in 0.05% DM. Twenty µl of sample were diluted in a 1 ml solution, and scanned in a 1-cm cuvette, therefore the corrected OD is 37.

different values of protein/detergent and protein content as among different measurements of the same sample. The only parameter which changed systematically with the protein/detergent value up to a certain plateau, was R. Hence, the emission spectra were normalized to the same area for measurements of samples with the same pigment/protein/detergent ratio. Absorption spectra, recorded at room temperature and in a 1-cm path length, were scanned on a Spectronic Genesys 2 spectrophotometer (Milton Roy). The scanning resolution was fixed at 1 nm.

Results Figure 1A shows the fluorescence spectrum of the PS I/PS II mixture after excitation at 605 nm. As seen, the fluorescence of phycobilisomes (at λ = 650 nm) is significantly smaller than that of the chlorophyll complexes. Regarding the high quantum yield of the phycobilsome’s fluorescence, this indicates that their

202

Figure 2. The emission spectrum (before normalization) of (A) a sample containing 0.05% DM (solid line), (B) 2.0% DM (dotted line). Inset: the emission spectrum of three PS I/PS II preparations all in 2.5% DM containing buffer. [Chl] = 500 µg/ml in all samples.

Figure 3. Changes in R versus [DM] (%w/v) (triangles). The control samples provide the dependence of R on dilution only (circles).

concentration is small, relative to the concentration of PS I and PS II. Figure 1B presents the absorption spectrum of the PS I/PS II sample. Again, no significant contribution of open tetrapyrroles was observed. Modification of the fluorescence emission of PS I/PS II by changing the detergent/protein ratio (mole/mole), while keeping the protein concentration constant Figure 2 depicts the fluorescence profile of a PS I/PS II mixture in different concentrations of DM, while the protein concentration was kept constant. When the detergent/protein ratio was modified, the relative intensity of the PS II and PS I bands (685 nm and 720 nm, respectively) was changed. At [DM] = 0.05%, the fluorescence of PS II was lower than that of PS I. The fluorescence intensity of PS II increased at the expense of PS I, when [DM] was increased to 2.0%. Figure 3 shows that the ratio between the PS II and PS I fluorescence intensities (R) increases from 0.48 to 1.2 and to 1.8, as [DM] increases from 0.05% to 0.4% and to 1.0%, respectively. No further changes in R were detected up to a detergent concentration of 2.5%. The value of R reaches a plateau around 0.6% DM. In the control experiment, no change in R was detected. The optical density of Chla in the examined samples was 0.37 in a 0.1-mm cell, indicating that there could be some fluorescence reabsorption by the sample. However, this possibility is highly unlikely. First, because the experiment is performed at a front-face position, which reduces the possibility

Figure 4. The effect of dilution with buffer alone and with buffer + 0.05% DM. (A) Sample before dilution (full line), (B) sample diluted 40 times with buffer alone (dashed line), and (C) sample diluted with buffer + 0.05% DM (dotted line).

of distortions in the fluorescence spectrum resulting from reabsorption by a concentrated sample (Lakowicz 1983, pp 44–47). Second, the control experiment showed that the same dilution with buffer only (while maintaining the same detergent/protein ratio) did not affect R (Figure 3, circles). Therefore, the modifications in the emission spectra could not reflect changes in the chromophore concentration. Some diversity in the R/[DM] relationships within samples isolated from different cell cultures reflect the variability of the PS I/PS II stoichiometry. This could be due to a long-term adaptation mechanism of these organisms, resulting from changes in the growth conditions (Murakami 1997). Modification of the fluorescence emission of PS I/PS II by sample dilution When the sample containing [Chl] = 500 µg/ml and 0.05% DM was diluted with buffer without DM to

203 Discussion

Figure 5. The reversibility of the dilution effect. Samples containing (A) [Chl] = 500 µg/ml, 2.5% DM (full line). (B) same but diluted 50-fold (dotted line), and (C) after reconcentration (dashed line).

Figure 6. The dependence of 1/R on 1/[DM].

[Chl] = 12.5 µg/ml and 0.00125% DM (40 times), a modification was observed in the emission profile (Figure 4). Here the detergent/protein ratio was not modified. Interestingly, the same dilution using buffer + 0.05% DM, led to a dramatic decrease in the energy transfer. Because of this form of dilution, the detergent/protein was also increased by a factor of 40. The reversibility of the changes in the PS I/PS II organization, as a result of dilution, which led to the observed variations in the emission spectrum, was checked by a 50-fold dilution of a sample containing [Chl] = 500 µg/ml in 2.5% DM and subsequent reconcentration such that, the [Chl] and [DM] reached the initial values of 500 µg/ml and 2.5% DM, respectively. Figure 5 shows that the fluorescence of PS I, relative to that of PS II, in the sample diluted 50-fold is small. However, after the sample was concentrated, the fluorescence profile was very similar to that of the sample before dilution.

The fluorescence spectrum of the PS II-enriched particles undergoes profound changes, upon extensive dilution and a concomitant increase in the detergent/protein ratio. Under these conditions, the PS I/PS II sample fluoresce mainly at 680–695nm (Figures 4 and 5), indicating that PS II particles are the major component of the isolated membrane complexes. Hence, the relative enhancement of PS I fluorescence in samples with a low detergent/protein ratio (Figure 2) may be explained by two alternative mechanisms: (1) reabsorption of energy emitted from PS II by the long wavelength-absorbing PS I complexes, and (2) energy transfer from PS II to PS I. Since: (i) the PS I fluorescence substantially decreased relative to the PS II fluorescence, whereas the protein concentration remained nearly constant and, (ii) the overall fluorescence remained fairly constant upon adding detergent (Figure 2), the first explanation was ruled out. This leaves energy transfer from PS II to PS I as the more likely explanation for the high fluorescence at 720 nmm, before adding DM or dilution of the sample by buffer. Here, R was significantly enhanced.1 The extent of energy transfer could be controlled by changing the detergent/protein ratio (Figures 2 and 3) or by dilution of the protein (Figures 4 and 5). In each case, the emission of PS I at 720 nm decreased, and the emission of PS II at 685 nm increased. When the PS I/PS II particles were exposed concomitantly to the two modes of perturbations, a synergistic effect was observed, and R reached a value close to 6 (Figure 5). These effects were reversible. We propose that the overall effect of DM on the PS I/PS II complex is given by: PS I − PS II + 2D  PS I − D + PS II − D

K1 , (1)

where D represents a detergent ring (such mode of membrane protein solubilization was found for bacterial RC (Roth et al. 1989, 1991). In fact, Loach et al. (1985) showed that antenna protein/bacteriochlorophyll complexes of B-870 were isolated with several tens of β-octylglucoside molecules that probably do not form regular micelles. PS I–D and PS II–D represent the PS I and PS II complexes with D, respectively. PS I–PS II represents a complex of PS I and PS II in an aqueous solution with no detergent.

204 We further suggest that: PS I + D  PS I − D PS II + D  PS II − D

(2)

K2

(3)

K3

PS I − PS II  PS I + PS II K4 PS I − PS II + 2D  (PS I − PS II)D2

(4) K5 (5)

(PS I − PS II)D2 PS I − D + PS II − D K6 .

(6)

The equilibrium constant for Equation (1) can be written as: [PS I − D][PS II − D] = K1 [PS I − PS II][D]2

(7)

Note that the fluorescence intensity at 685 nm is a result of emission from PS II alone and is proportional to [PS II-D]. In contrast, the fluorescence at 720 nm originates in PS I and its intensity is proportional to [PS I–PS II] + [PS I–D]. Using the definition of R (R = FPS II /FPS I ): R=

[PS II − D] · [PS I − D] + [PSI − PSII]

(8)

Using Equations (2–6) for substituting [PS II–D], [PS I–D] and [PS I–PS II] in Equation (8), we obtain the dependence of R on the detergent and protein concentrations: K2 [PS I] 1 FPS I = + = • FPS II R K3 [PS II] 1 2K2 · • [PS I] • K1 [D]

(9)

The dependence of 1/R on 1/[DM] is depicted in Figure 6. PS I–PS II interactions in the hydrophobic domain (detergent concentration above the CMC) The effect of the detergent/protein ratio on 1/R Neutron scattering measurements of RC crystals from Rhodobacter viridis and sphaeroides (Roth et al. 1989, 1991) showed that detergent molecules are ordered as rings that surround the hydrophobic parts of the transmembrane α-helices. Here, the detergent–protein interactions are hydrophobic, and they replace the

protein interactions with the lipids of the native membrane. In our system, at low detergent/protein ratios and high protein concentrations, PS I trimers and PS II dimers probably exist in the solution. Since there is a structural similarity between the photosystems (Schubert et al. 1998 and recently Hankamer et al. 1999) mixed aggregates of PS I and PS II can also coexist with the former ones. Mixed complexes of PS I and PS II with different compositions were indeed isolated in the course of this study throughout the perfusion chromatography at different detergent/protein ratios (data not shown). Equation (9) indicates that 1/R should approach an asymptotic value proportional to PS I/PS II, as the detergent/protein ratio is increased (i.e. when [PS I]/[D] is decreased). This prediction agrees with the experimental data depicted in Figure 6. However, further dilution of the solution should not change the value of 1/R, when [PS I]/[DM] is maintained constant. In contrast, 1/R continued to decrease when the solution containing 2.5% DM was diluted (Figure 5). To account for this discrepancy, we suggest that several D form large cylinders at high detergent concentrations, such that Mn−1 + D  Mn ,

(10)

where n is an integer. We proposed that D is an elementary micelle (i.e. a simple ring of dodecylmaltoside molecules such as found for bacterial RC). An increase of [D] causes the formation of larger Mn micelles, probably by cooperative aggregation (Fisher et al. 1990) or simple accumulation around the PS II/PS I bodies. Under these conditions, one may find several PS I and PS II particles within the same ‘extended’ micelle. Then, the hydrophobic interactions between PS I and PS II, which control their association in low detergent concentrations, can not drive the formation of the residual PS I/PS II complexes. Rather, non-hydrophobic interactions that optimize packing, hydrogen bonding, and Van-der-Waals forces determine the association within the detergent hydrophobic milieu as reflected in the energy transfer between the two photosystems. Notably, bacteriochlorophyll– protein complexes from B-873 show similar behavior in the presence of β-octyl-glucoside (Loach et al. 1985). Inversely, dilution of a particular solution with a fixed protein/detergent value, causes the dissociation n P of Mn micelles to smaller ones while the overall Mi i=1

remains nearly constant; at the same time, [PS I] decreases. Thus, following Equation (9), 1/R should also

205 decrease. Note that diluting the PS I/PS II system does not change the value of the first term on the right hand side of Equation (9). The effect of protein concentration on 1/R Equation (9) and Figure 4 show that when the detergent concentration is constant, the extent of energy transfer from PS II to PS I depends on the protein concentration. Note that when [DM] is kept constant, dilution results in an increase of the DM/protein ratio. However, a solution of [Chl] = 500 µg/ml in 2.5% DM and a solution of [Chl] = 10 µg/ml in 0.05% DM contain the same detergent/protein ratio. Still, as depicted in Figure 5, in the diluted solution we observe a much smaller fluorescence band at 720 relative to the 685 nm one, than in the concentrated solution. In order to explain this effect, we suggest that interactions other than those that are hydrophobic, contribute to the PS I–PS II association which allows for efficient energy transfer. These may include electrostatic interactions as suggested by MacKenzie and Engelman (1998). Here dilution of the protein is thought to enhance dissociation. The involvement of electrostatic interactions are in line with the reversible dilution effect on the fluorescence lineshape (Figure 5), and with the effect of salt on PS I trimerization reported by Kruip et al. (1994 and recently 1999). The PS I–PS II complex is completely disrupted only by a concomitant dilution with buffer and detergent. Under these conditions, [PS I]/[D] in Equation (9) rapidly approaches zero. PS I–PS II interactions in an hydrophilic environment When the PS I/PS II solution containing [Chl] = 500 µg/ml and 0.05% DM (Figure 4, full line) was diluted 40-fold, [DM] was brought under the CMC, and we observed a precipitation of PS I and PS II. Yet R significantly increased. Under these conditions, the solvent is hydrophilic. Studies on Rhodobacter viridis RCs and on bacteriorhodopsin, dissolved in detergent solutions below the CMC, showed lower binding of DM molecules to the protein, and the formation of large protein aggregates (Møller and le Maire 1993). In our experiment, the PS I emission at 720 nm decreased significantly compared with the PS II emission at 685 nm. Eventually, the complexes precipitate. This precipitation probably reflects the formation of large aggregates. However, this aggregation disturbed the energy transfer from PS II to PS I. In other words,

we observed a situation where PS I–PS II dissociate as the solution becomes more hydrophilic. Hence, we have concluded that here, the PS I–PS II particles disintegrate and instead the PS I–PS I and PS II–PS II large aggregates are probably formed. In conclusion, there probably exists two modes of assembly in PS I/PS II mixtures, depending on the environment: (a) Hydrophobic aggregation in an hydrophilic environment (in a solution with a low content of protein and under the CMC). This aggregation involves extensive PS I–PS I and PS II–PS II complex formation, a small energy transfer and eventually, precipitation of the large complexes. (b) Specific, non-hydrophobic aggregation that mostly involves the PS I–PS II complex formation, and results in highly efficient energy transfer. This interaction is highly effective in a hydrophobic environment, where protein–protein non-specific hydrophobic interactions are replaced by protein– detergent interactions, and therefore are possibly applied to the thylakoid membrane. The observed changes in the 77 K fluorescence emission spectrum of the isolated PS I/PS II mixture resemble those obtained in state transitions of cyanobacteria (Bruce et al. 1989; Vernotte et al. 1990). On the basis of previous reports on the randomization of PS II complexes in the membrane that enables PS II to approach PS I (Olive et al. 1986; Vernotte et al. 1990) and the results of in vitro measurements, we concur with the notion that state transitions of Chla excitation energy is performed by the spillover of the energy to PS I, and in addition, believe that the transfer extent is at least partly controlled by the association of the two Photosystems. It is possible that in vivo, detachment of the phycobilisomes by phosphorylation leads to a conformational change that enables such a specific association. Recently, El Bissati et al. (1998) showed that in a Synechocystis 6803 mutant with a more rigid membrane, state transitions are not detected.

Acknowledgements We thank Mr Dror Noy from the Department of Plant Sciences for stimulating discussions. This study was supported by the Wilsttaeter-Avron Minerva Foundation for Photosynthesis Research and by a D.F.G. grant A.Z. SCHE 140/18-2.

206 Note 1 The effects of small changes in the protein concentration, as a

result of the added detergent solution, were controlled by adding buffer alone (Figure 3).

References Allen JF (1992) Protein phosphorylation in regulation of photosynthesis. Biochim Biophys Acta 1098: 275–335 Allen JF, Sanders CE and Holmes NG (1985) Correlation of membrane protein phosphorylation with excitation energy distribution in the cyanobacterium Synechococcus 6301. FEBS Lett 193: 271–275 Bald D, Kruip J and Rögner M (1996) Supramolecular architecture of cyanobacterial thylakoid membranes: How is the phycobilisome connected with the photosystems. Photosynth Res 49: 103–118 Bonaventura C and Myers J (1969) Fluorescence and oxygen evolution from Chlorella pyrenoidosa. Biochim Biophys Acta 189: 366–383 Bruce D, Biggins J, Steiner T and Thewalt M (1985) Mechanism of the light state transition in photosynthesis. IV. Picosecond fluorescence spectroscopy of Anacystis nidulans and Prophyridium cruentum in state 1 and state 2 at 77 K. Biochim Biophys Acta 806: 237–246 Bruce D, Hanzlik C, Hancok LA, Biggins J and Knox RS (1986) Energy distribution in the photochemical apparatus of Porphyridium cruentum: Picosecond fluorescence spectroscopy of cells in state1 and state2 at 77 K. Photosynth Res 10: 283–290 Bruce D, Brimble S and Bryant DA (1989) State transitions in a phycobilisome-less mutant of the cyanobacterium Synechococcus sp. PCC 7002. Biochim Biophys Acta 974: 66–73 Canaani O and Malkin S (1984) Distribution of light excitation in an intact leaf between the two photosystems of photosynthesis. Changes in absorption cross-sections following state 1–state 2 transitions. Biochim Biophys Acta 766: 513–524 El Bissati K, Murata N, Etienne AL and Kirilovsky D (1998) In: Garab G (ed) Photosynthesis: Mechanisms and Effects, Vol III, pp 1835–1838. Kluwer Academic Publishers, Dordrecht, The Netherlands. Fisher JRE, Rosenbach-Belkin V and Scherz A (1990) Cooperative polymerization of photosynthetic pigments in formamide-water solutions. Biophys J 58: 461–470 Fork DC and Satoh K (1986) The control by state transitions of the distribution of excitation energy in photosynthesis. Ann Rev Plant Physiol 37: 335–361 Glazer AN and Melis A (1987) Photochemical reaction centers: structure, organization and function. Ann Rev Plant Physiol 38: 11–45 Hankamer B, Morris EP and Barber J (1999) Revealing the structure of the oxygen-evolving core dimer of photosystem II by cryoelectron crystallography. Nat Struct Biol 6(6): 560–564 Harrison MA, Tsinoremas NF and Allen JF (1991) Cyanobacterial thylakoid membrane proteins are reversibly phosphorylated under plastoquinone-reducing conditions in vitro. FEBS Lett 282: 295–299 Kruip J, Bald D, Boekema E and Rögner M (1994) Evidence for the existence of trimeric and monomeric Photosystem I complexes in thylakoid membranes from cyanobacteria. Photosynth Res 40: 279–286

Kruip J, Karapetyan NV, Terekhova IV and Rögner M (1999) In vitro oligomerization of a membrane protein complex. J Biol Chem 274: 18181–18188 Lakowicz JR (1983) Principles of Fluorescence Spectroscopy. Plenum Press, New York Loach PA, Parkes PS, Miller JF, Hinchigeri S and Callahan PM (1985) Structure–function relationships of the bacteriochlorophyll-protein light-harvesting complex of Rhodospirillum rubrum. In: Steinback KE (ed) Molecular Biology of the Photosynthetic Apparatus, pp 197–208. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York Ley AC and Buttler WL (1980) Energy distribution in the photochemical apparatus of Porphyridium cruentum in state I and state II. Biochim Biophys Acta 592: 349–363 MacKenzie KR and Engelman DM (1998) Structure-based prediction of the stability of transmembrane helix–helix interactions: The sequence dependence of glycophorin A dimerization. Proc Natl Acad Sci USA 95: 3583–3590 Malkin S, Herbert SK and Fork DC (1990) Light distribution, transfer and utilization in the marine red alga Porphyra perforata from photoacoustic energy-storage measurements. Biochim Biophys Acta 1016: 177–189 Møller JV and le Maire M (1993) Detergent binding as a measure of hydrophobic surface area of integral membrane proteins. J Biol Chem 268: 18659–18672 Mullineaux CW and Holzwath AR (1990) A proportion of Photosystem II core complexes are decoupled from the phycobilisome in light-state 2 in the cyanobacterium Synechococcus 6301. FEBS Lett 260: 245–248 Mullineaux CW, Bittersman E, Allen JF and Holzwarth AR (1990) Picosecond time-resolved fluorescence emission spectra indicate decreased energy transfer from the phycobilisome to Photosystem II in light-state 2 in the cyanobacterium Synechococcus 6301. Biochim Biophys Acta 1051: 231–242 Mullineaux CW, Griebenow S and Braslavsky SE (1991) Photosynthetic energy storage in cyanobacterial cells adapted to light-states 1 and 2. A laser-induced optoacoustic study. Biochim Biophys Acta 1060: 315–318 Mullineaux CW, Tobin MJ and Jones GR (1997) Mobility of Photosynthetic complexes in thylakoid membranes. Nature 390: 421–424 Murakami A (1997) Quantitative analysis of 77 K fluorescence emission spectra in Synechocystis sp. PCC 6714 and Chlamydomonas reinhardtii with variable PS I/PS II stoichiometries. Photosynth Res 53: 141–148 Murata N (1969) Control of excitation transfer in photosynthesis. I. Light induced change of chlorophyll a fluorescence in Porphydium cruentum. Biochim Biophys Acta 172: 242–251 Murata N (1970) Control of excitation transfer in photosynthesis. IV. Kinetics of chlorophyll a fluorescence in Porphyra yezoensis. Biochim Biophys Acta 205: 379–389 Olive J, M’Bina I, Vernotte C, Astier C and Wollman FA (1986) Randomization of the EF particles in thylakoid membranes of Synechocystis 6714 upon transition from state I to state II. FEBS Lett 208: 308–312 Oren-Shamir M, Maruthi Sai PS, Edelman M and Scherz A (1995) Isolation and spectroscopic characterization of a plantlike Photosystem II reaction center from the cyanobacterium Synechocystis sp. 6803. Biochemistry 34: 5523–5526 Roth M, Lewit-Bentley A, Michel H, Deisenhofer J, Huber R and Oesterhelt D (1989) Detergent structure in crystals of a bacterial reaction centre. Nature 340: 659–662 Roth M, Arnoux B, Ducruix A and Reiss-Husson F (1991) Structure of the detergent phase and protein–detergent interactions

207 in crystal of the wild-type (strain Y) Rhodobacter sphaeroides photochemical reaction center. Biochemistry 30: 9403–9413 Sanders CE, Holmes NG and Allen JF (1986) Membrane protein phosphorylation in the cyanobacterium Synechococcus 6301. Biochem Soc Trans 14: 66–67 Schubert WD, Klukas O, Saenger W, Witt HT, Fromme P and Krauss N (1998) A common ancestor for oxygenic and anoxygenic photosynthetic systems: A comparison based on the structural model of Photosystem I. J Mol Biol 280: 297–314

Vernotte C, Astier C and Olive J (1990) State 1–state 2 adaptation in the cyanobacteria Synechocystis PCC 6714 wild type and Synechocystis PCC 6803 wild type and phycocyanin-less mutant. Photosynth Res 26: 203–212 Williams WP and Allen JF (1987) State 1/State2 changes in higher plants and algae. Photosynth Res 13: 19–45.