Photosynthesis Research 82: 327–338, 2004. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.

327

Review

Metabolic flexibility of the green alga Chlamydomonas reinhardtii as revealed by the link between state transitions and cyclic electron flow Giovanni Finazzi1,2,∗ & Giorgio Forti2 1 UPR-CNRS

1261 (associ´ee Universit´e Paris 6), Institut de Biologie Physico Chimique, 13 rue Pierre et Marie Curie, 75005 Paris, France; 2 Istituto di Biofisica, Sezione di Milano, via Celoria 26, 20133 Milan, Italy; ∗ Author for correspondence (e-mail: [email protected]; fax: +33-1-58415022)

Received 17 November 2003; accepted in revised form 13 February 2004

Key words: Chlamydomonas reinhardtii, energetic metabolism, linear and cyclic electron flow, state transitions

Abstract In this Review we focus on the conversion of linear photosynthetic electron transport from water to NADP to the cyclic pathway around Photosystem I in the green alga Chlamydomonas reinhardtii. We discuss the strict relationship that exists between the changes in pathways of electron transport and state transitions, i.e., the reversible functional association of light harvesting proteins with one of the two photosystems of oxygenic photosynthesis. Such a link has not been reported in the case of other photosynthetic organisms, where the state transitions do not affect the pathway of electron transport. Rather, they provide a tool to optimise the rate of linear flow. We propose a kinetic-structural model that explains the mechanism of this particular relationship in Chlamydomonas, and discuss the advantages that this peculiar situation gives to the energetic metabolism of this alga. Abbreviations: DBMIB – 2,5-dibromo-3-methyl-6-isopropyl-p-benzoquinone; DCMU – 3-(3,4 -dichlorophenyl)1,1-dimethylurea; PQ – plastoquinone; PQH2 – plastoquinol; PS – photosystem; qE – pH dependent nonphotochemical quenching of fluorescence emission; Qi – plastoquinone reducing site of the cytochrome b6 f complex; Qo – plastoquinol oxidising site of cytochrome b6 f complex

State transitions, phenomenology The phenomenon of state transitions, first discovered in unicellular photosynthetic organisms (Bonaventura and Myers 1969; Murata 1969), involves the reversible transfer of a fraction of the PS II outer antenna to PS I. It is understood as a mechanism for balancing the absorption capacity of the two photosystems under natural illumination conditions, and therefore their photochemical activity. The term ‘state transitions’ can be used to describe the reversible association of the PS II antenna complex (LHC II) with either PS II (state 1) or PS I (state 2). Thus, the transition from state 1 to state 2 decreases the absorption capacity of PS II at the advantage of

that of PS I, while the opposite occurs for the reverse process (reviews in Allen 1992; Gal et al. 1997; Wollman 2001). In plants and green algae, the mechanism of state transitions involves the phosphorylation of LHC II by a membrane bound protein kinase. Several kinase activities associated with thylakoid membranes have been evidenced in past years (reviewed in Wollman 2001). Among them, three related kinases (TAK 1–3, for thylakoid-associated kinases) seem to affect the pattern of LHC II phosphorylation in Arabidopsis thaliana, TAK 1 being a candidate to play an active role in state transitions (Snyders and Kohorn 2001). On the other hand, a kinase different from TAKs (Stt7) is responsible for state transitions in Chlamydomonas reinhardti (Depege et al. 2003).

328 Taken together, these findings raise the possibility that a cascade of phosphorylation events takes place during state transitions. The phosphorylation of LHC II is referred to as state 2 transition, as it is induced by a light mainly absorbed by PS II (light 2). A light mainly absorbed by PS I (light 1) causes its dephosphorylation by a phosphatase, defining a state 1 transition. Upon phosphorylation, a fraction of the LHC II migrates away from the PS II-rich grana region of thylakoids, by lateral diffusion in the thylakoids. This results in the accumulation of LHC II in the stroma unapressed membranes (reviewed in Allen 1992; Gal et al. 1997; Wollman 2001), where PS I is mostly located (reviewed in Albertsson 2001). The detachment of the LHC II in the grana has been interpreted in terms of the electrostatic repulsion generated by the increase of negative region of the thylakoids (reviewed in Allen 1992). Alternatively, conformational changes in the LHC II molecule have been proposed to be responsible for this phenomenon (Nilsson et al. 1997). They might occur either before the phosphorylation, to promote the interaction with the kinase, or after it, to decrease the affinity of LHC II for PS II (reviewed in Wollman 2001). These conformational changes might also be required for the association of LHC II to PS I, which needs a specific docking site provided by one of the small subunits of PS I, psaH (reviewed in Scheller et al. 2001; Wollman 2001, see also Haldrup et al. 2001 for a further discussion of the role of PS I in state transitions). Although the diffusion mechanism for state transitions is widely accepted, there are still some difficulties in the interpretation of some experimental data. For instance, no evidence for the existence of a (even transient) state, where phosphorylated LHC II is disconnected from both PS II and PS I, was reported in a mutant of A. thaliana lacking the docking subunit psaH (Lunde et al. 2000). This is expected in the frame of the diffusion/docking hypothesis mentioned above, since upon phosphorylation LHC II should detach from PS II, diffuse to the stroma domain, and diffuse back to PS II, being unable to bind to PS I. While several different interpretations can be given, it should be recognised that this experimental evidence is consistent with an alternative view of state transitions, where phosphorylation is supposed not to induce lateral migrations of the LHC II, but rather a partial unstacking of the thylakoids, leading to some spillover from PS II to PS I (Georgakopulos and Argyroudi-Akoyunoglu 1994).

The LHC II kinase is activated by the reduction of the electron transfer carriers (Allen et al. 1981; Horton and Black 1981). The interplay between the redox state of theses carriers and the occurrence of state transitions has been summarised in a very simple scheme (Allen 1992) that still represents the framework of the present knowledge on this subject. The reduction of the plastoquinone pool (either by the activity of PS II, or by other cellular metabolic processes) activates the kinase. On the contrary, its inactivation is promoted by the oxidation of the PQ pool, i.e., by PS I activity, or other cellular oxidative processes. The central role of the cytochrome b6 f complex in sensing the redox state of the plastoquinone pool was demonstrated by the lack of state transitions in mutants devoid of this complex in both Chlamydomonas and vascular plants (reviewed in Gal et al. 1997; Wollman 2001). Then, transient acidification experiments, as well as the isolation of a cytochrome b6 f mutant of Chlamydomonas unable to fix plastoquinol, showed that the first step in the kinase activation cascade is the formation of the PQH2 –b6 f complex (reviewed in Vener et al. 1998; Wollman 2001). The finding that conformational changes are induced in the cytochrome bc family upon quinol binding (reviewed in Breyton 2000; see also Kurisu et al. 2003 for a discussion) lead to the proposal of a two-step activation mechanism. This hypothesis postulates that the binding and unbinding of the quinol/quinone to the b6 f complex leads to conformational changes within the b6 f complex. These changes are involved in the binding, the activation and the release of the active LHC II-kinase from the cytochrome b6 f (Vener et al. 1998; Wollman 2001; Figure 1). Consistent with this central role of the cytochrome b6 f complex in promoting state transitions is the recent finding that the Antarctic, green alga Chlamydomonas subcaudata, which bears a modified form of the cytochrome b6 f complex (as suggested by the different size of the cytochrome f subunit in the two species), is unable to perform state transitions (Morgan-Kiss et al. 2002). Studies using chimeric mutants of the cytochrome b6 f complex in Chlamydomonas have led to the identification of a possible docking site for the LHC II kinase in a rather peripheral position in the b6 f complex, near to the PetL subunit (Zito et al. 2002). This position corresponds to one of the regions where the occurrence of transmembranal conformational changes has been shown upon substrate binding to the b6 f (reviewed in Breyton 2000). According to

329

Figure 1. Hypothetical model for LHC II-kinase activation. Binding of plastoquinol (PQH2 ) to the quinol binding site (Qo) of cytochrome b6 f complex (A) is expected to induce conformational changes in the complex. These changes would be required to promote the association between the cytochrome and the kinase responsible for state transitions (LHC II-kinase). This might occur via the formation of a kinase–b6 f complex (B). Alternatively, a ternary binding complex might be formed (C) between the cytochrome complex, the kinase and the PS II antenna complex (LHC II). Upon activation, the kinase (and eventually its substrate) would be released from the cytochrome complex, leading to state transitions (D). During activation of the kinase complex, the putative site of regulation of the kinase activity by thioredoxins (black box) might be sequestered within the b6 f –kinase complex. This would limit its accessibility by thioredoxins, and explain their subordinate role in modulation of the kinase activity, with respect to PQH2 .

the recent 3-D structure of the b6 f complex (Kurisu et al. 2003; Stroebel et al. 2003), both the chlorophyll and the carotenoid molecules present in the b6 f complex are exposed to the lipid phase in this area. It is therefore tempting to suggest that this particular pigment arrangement might mimic in some way a similar pigment/protein motif present on the LHC II moiety. Accordingly, the similarity between the two docking sites might provide the rationale for the ability of a relatively small molecule as the LHC II kinase to bind to two very different molecules as the cytochrome b6 f complex and LHC II. Alternatively, it can be proposed that a ternary complex, involving the kinase, LHC II and the cytochrome b6 f complex is transiently formed during activation, the generation of which would be again promoted by the similarity of the pigment binding domains on the LHC II and the cytochrome complex (Figure 1C). State transitions have not only been described in vascular plants and green microalgae, but also in red microalgae and cyanobacteria, i.e., in phycobilisomecontaining organisms. In these organisms, reversible changes in fluorescence emission are observed, which are related to changes in the redox state of the plastoquinone pool (Campbell et al. 1998). However, other mechanisms than state transitions might be responsible for these observations. In cyanobacteria, in

spite of the fact the diffusion of phycobilisomes in thylakoid membranes has been clearly demonstrated (Sarcina et al. 2001), it is still not clear whether the quenching of fluorescence attributed to state transitions represents a true association of the phycobiliprotein antenna complexes with PS I, or a variable spill-over from PS II to PS I. In the case of the red alga Rhodella violacea, the fluorescence quenching commonly associated with state-2 transition seems to be a pH dependent quenching (Delphin et al. 1998), namely a form of qE (Horton et al. 1996). No evidence for such phenomenon has been reported in marine algae such as diatoms (Owens 1986). State transitions and the redox state of the plastoquinone pool Oxygenated cells of Chlamydomonas are locked in state 1, in spite of the fact that their plastoquinone pool is 50% reduced (Wollman 1978). Given the high affinity of the b6 f complex for its substrate PQH2 (reviewed in Cramer et al. 1996), this plastoquinol concentration should be high enough to saturate the Qo site. According to the simple relationship between the binding of PQH2 to the Qo site and the transition to state 2 described above, the cells should therefore be in state 2. This discrepancy suggest therefore the existence of a gap between the saturation profile of the Qo site and the occurrence of state transition in whole cells. This conclusion contrasts the view of Vener and colleagues (1997, 1998), who conclude that the relationship between the occupation of the Qo site the activation of the LHC II-kinase is linear. Several hypotheses can be proposed to explain this controversy: the activity of the phosphatase, which is responsible for the dephosphorylation of LHC II, might be also regulated, as previously suggested (Fulgosi et al. 1998). Since state transitions are the result of the balance between kinase and phosphatase activity, an effect of PQ redox state on the sole LHC IIkinase activity might produce a complex relationship between PQH2 generation and state transitions. Alternatively, it has been proposed that the active state of the kinase is short living, and consequently this enzyme has to interact several times with the b6 f complexes before reaching its substrate LHC II in the active state (see Gal et al. 1997, for a discussion). Accordingly, the non-linear relationship between plastoquinone reduction and LHC II kinase activation would reflect the possibility for a given

330 kinase molecule to encounter at least one b6 f complex in the inactivating conformation (i.e., with a plastoquinone bound) when the pool is even partially oxidised. In this case, no state transitions would occur even in the presence of a large fraction of PQH2 –b6 f complexes (Gal et al. 1997). As a third possibility, it can be suggested that kinase activation requires the concomitant generation of the activating signal within the two b6 f monomers of the dimeric cytochrome complex. In this case, the non-linearity between plastoquinol binding and kinase activation would reflect the need for a simultaneous binding of PQH2 to the Qo site of the two monomers (Figure 1B). A final possibility is based on the recent proposal by Aro and coworkers (Rintamaki et al. 2000), that thioredoxins might be involved in the modulation of state transitions, by exerting a negative control on the kinase itself. In this case the non-linear relationship between the redox state of the plastoquinone pool and the occurrence of state transitions might reflect the different levels of interaction between these two antagonistic regulatory systems of the LHC IIkinase. The ‘thioredoxin hypothesis’ is still under debate. Indeed, although the LHC II-kinase recently isolated in Chlamydomonas presents a putative regulatory site of thioredoxins, the same inhibition of LHC II phosphorylation attributed to the effect of thioredoxins has been observed in isolated thylakoids, i.e., in the absence of thioredoxins, upon increasing the energy of the excitation light (Zer et al. 2003). In any case, it is acknowledged that the putative regulation activity exerted by thioredoxins would act downstream of the one exerted by the plastoquinone pool (Rintamaki et al. 2000). A very simple way to interpret this hierarchy is to assume that the putative regulatory site present on the LHC II-kinase might be hidden by the formation of the b6 f –kinase complex (Figure 1).

State transitions in C. reinhardtii and the commutation between linear and cyclic flow While the basic mechanism of state transitions is very similar in vascular plants and Chlamydomonas, their amplitude varies to a large extent. In vascular plants, ca. 15% to 20% of the LHC II is involved in state transitions (Allen 1992). Owing to the different absorption

of the two photosystems (PS I being enriched in chlorophyll forms absorbing at longer wavelengths), this small change is effective in balancing the two photochemical reactions in series. On the contrary, in the case of Chlamydomonas, state transitions involve the reversible transfer of up to 85% of LHC II (Delosme et al. 1996), as well as a significant fraction of the cytochrome b6 f complex toward the stromal, unapressed region of the thylakoids (Vallon et al. 1991). In this case, therefore, the extent of the antenna migration is too large to fulfil the role of re-adjustment of the absorption properties. In state 2, the ratio between the absorption capacity of PS I and PS II is of ∼5. This would lead to a large waste of the absorbed energy if the two photosystems were operating in series, as expected in the case of linear electron flow. As a rationale for state transitions in Chlamydomonas, it has been therefore proposed that state transitions correspond to a switch from linear to cyclic electron flow (Vallon et al. 1991). This hypothesis has been demonstrated by the finding that electron injection into the cytochrome b6 f complex was insensitive to the addition of the PS II inhibitor DCMU in state-2 conditions. This process was, however, inhibited by the b6 f inhibitor DBMIB (Finazzi et al. 1999). The sensitivity to DCMU could be restored only upon adaptation of algae to state 1 promoting conditions. This indicates, as expected for the linear electron flow, that in state 1 electron injection into the b6 f complex may be inhibited with an identical efficiency by blocking either PQH2 oxidation by the b6 f complex or PQ reduction by PS II. On the contrary, the situation observed in state 2 is consistent with the operation of cyclic flow, where cytochrome b6 f reduction does not require PS II activity (Figure 2). Another independent evidence for this correlation was provided by the study of thylakoid swelling induced by illumination of Chlamydomonas mutants devoid of the ATP-synthase CF0 –F1 . This swelling, which reflects the generation of the transthylakoid pH, was inhibited by DCMU addition in state-1, but not in state-2 conditions (Majeran et al. 2001). These results indicate that cells are able to build a pH, by linear flow in state 1, and cyclic flow around PS I in state 2. The existence of a strict relationship between state transitions and the commutation between linear and cyclic flow was indicated by the fact that DCMU inhibits electron flow in state 2 promoting conditions

331

Figure 2. Commutation between linear and cyclic electron flow in Chlamydomonas upon state transitions. The supramolecular organisation of the thylakoids in state 1 is optimised to allow the functioning of linear electron flow where both PS I and PS II photochemistry operates in series resulting in ATP generation, NADPH production and ultimately CO2 assimilation. In state 2 the large association of the LHC II to PS I would strongly favour its photochemical activity at the expense of that of PS II. In addition, cytochrome b6 f movements that follow transition to state 2 would increase the concentration of electron carriers in the PS I-enriched membranes, reducing at the same time the functional connection between PS II and the cytochrome b6 f complex. These two concomitant events would promote cyclic electron flow around Photosystem I (PS I), and the over-reduction of the PQ pool connected to PS II. In addition, a possible reversible association of the FNR enzyme with either PS I or the b6 f complex, as well as electron flow through the additional c heme of the b6 f complex (star) might contribute to the cyclic commutation is state 2.

in a LHC II-kinase-less mutant (Finazzi et al. 2002), which is thus locked in state 1, because of the lack of LHC II phosphorylation. More recently, the concept of the strict correlation between state transition and the occurrence of cyclic electron flow was extended to physiological conditions. In these conditions, it was possible to show that the cells are in an intermediate state between state 1 and state 2, and both linear and cyclic flow operate in the same chloroplast (Forti et al. 2003).

Mechanism of the shift between linear and cyclic flow in Chlamydomonas Dynamic versus structural models for cyclic flow commutation The existence of this tight relationship between state transitions and commutation between linear and cyclic electron flow rises the question of the intimate mechanism(s) that relates so precisely the occurrence of

332 state transitions to the appearance and disappearance of cyclic flow. The simplest way to explain the observed correlation between the association of LHC II with PS I and the appearance of cyclic flow would be to assume a dynamic competition between PS II and PS I for the reduction of the PQ pool. Under state-1 conditions, the ability of PS I to reduce the PQ pool is lower than that of PS II. Upon state-2 transition, however, the large increase of relative absorption capacity of PS I would increase its activity at the expense of that of PS II. This would lead to a situation where cyclic flow is prevailing for simple kinetic reasons. A corollary of this hypothesis is that even in state 2, the contribution of linear flow might increase as a function of light intensity, because of the progressive compensation of the PS II absorption deficiency. This is indeed the case under intermediate conditions (Forti et al. 2003), where the contribution of cyclic flow is largely decreased under saturating light. Besides the absorption changes, structural factors might also advantage cyclic flow in state 2. It has been proposed (Lavergne and Joliot 1991) that the PQH2 diffusion is restricted to relatively small domains in the membranes, which are generated mainly by LHC II contacts in the thylakoids. This is the result of the uneven distribution of photosynthetic complexes between the grana (PS II-rich) and stroma (PS I-rich) domains of the thylakoids, and of the large protein concentration in the membranes (reviewed in Albertsson 2001). Within the same domain, plastoquinol connects PS II and the cytochrome b6 f complex in a time range (ms) compatible with its physiological function, while its diffusion between different domains is a comparatively slow process (in the second time range). On the other hand, less restrictive conditions for plastocyanin diffusion seem to take place. This would mean that PS II and cytochrome b6 f complexes involved in linear electron flow are present within one diffusion domain, whereas longer distances would be allowed between PS I and the b6 f complexes. During transition to state 2, the progressive depletion of LHC II and b6 f of the grana, and their concomitant accumulation in the stroma lamellae would reduce the fraction of domains where both PS II and the cytochrome b6 f complex are present at the same time. This would reduce the efficiency of linear flow in favour of cyclic flow. This effect of protein compartmentation on cyclic flow is not expected to be dependent on the light intensity. Therefore, this hy-

pothesis seems to better explain the situation observed in fully adapted state-2 cells, where no linear flow was observed, even at saturating light intensities (Finazzi et al. 1999). It seems therefore, that the two alternative phenomena (kinetic competition and protein segregation) contribute differentially to the onset of cyclic flow in intermediate conditions and in fully established state-2 conditions. The structural hypothesis can explain the complete commutation to cyclic flow observed in state 2 only if a complete physical separation between the ‘linear’ and the ‘cyclic’ chains is assumed. In vascular plants, it has been proposed that this physical separation is due to the formation of tightly bound ‘cyclic’ supercomplexes involving ferredoxin, FNR, plastocyanin, the cytochrome and the PS I complexes (Joliot and Joliot 2002). In the case of Chlamydomonas, no evidence for such supercomplexes exist. Indeed, a complete commutation to cyclic activity upon transition to state 2 was observed in a mutant where the cellular cytochrome b6 f concentration is reduced by (at least) a factor of 10 (Finazzi et al. 2002). Obviously, in this mutant the formation of stoichiometric supercomplexes between the PS I and the cytochrome b6 f complex should be largely prevented. In Chlamydomonas, therefore, it is necessary to imagine a complete segregation between PS II and the cytochrome b6 f complex to interpret the conversion to cyclic flow in state 2 according to a structural hypothesis. It is equivalent to assume the accumulation of the entire population of the b6 f in the stroma lamellae. Such a hypothesis, however, is not consistent with the experimental observation, where only a fraction of the cytochrome b6 f complex is involved in the reversible migration between the stromal and the grana lamellae (Vallon et al. 1991). It is possible, however, that this fraction is largely underestimated. Indeed, in the case of LHC II, immunological labelling assessments (Vallon et al. 1991) greatly underestimate the fraction of complexes involved in the reversible migration with respect to the functional data (Delosme et al. 1996). As an alternative possibility, the existence of some partially unstacked, PS I-enriched domains within the grana stacks of Chlamydomonas can be proposed. These domains, where the b6 f and the LHC II complexes might accumulate during state-2 transitions, could not be distinguished from grana membranes in immunolabelling experiments, owing to their reduced size. This would explain the apparent contradiction between structural and functional data discussed above. The role of these PS I-rich granal domains

333 might be essential to promote cyclic flow, but also linear electron transport. Indeed, the structure of Chlamydomonas thylakoids is rather different from that of vascular plants: the grana stacks are very extended and form a reduced number of stacked membranes (see, e.g., Vallon et al. 1991). Therefore, in the absence of such ‘PS I-rich’ patches in the grana, the average distances between PS II and PS I would be very large, and hardly compatible with a fast diffusion of soluble electron carriers during photosynthesis. Other partners involved in cyclic flow Besides PS I and the cytochrome b6 f complex, cyclic flow in Chlamydomonas likely involves ferredoxin and plastocyanin, as in vascular plants (reviewed in Bendall and Manasse 1995). Another likely candidate is ferredoxin-NADP reductase (FNR). Its involvement in cyclic flow was proposed in spinach chloroplasts, where an antibody raised against purified FNR inhibited DCMU resistant endogenous photophosphorylation (Forti and Zanetti 1969). This enzyme strongly associates with either the cytochrome b6 f complex (Zhang et al. 2001), or with PS I. The latter interaction is likely weaker than that with the b6 f complex, and is mediated by the PsaE subunit (Scheller et al. 2001). By its reversible binding ability, this enzyme might provide a link between cyclic flow and state transitions. Because FNR cannot bind to the stacked membranes of the grana (Jennings et al. 1979), its interaction with the cytochrome b6 f complex might occur only when the latter is present in the stroma lamellae. The accumulation of the b6 f in the stroma membranes upon state-2 adaptation might therefore promote a preferential binding of FNR with this complex. This would enhance plastoquinone reduction, rather than NADP reduction by PS I, mainly for kinetic reasons (Figure 2). Although no direct evidence for such a b6 f – FNR association has been provided in the case of Chlamydomonas, to our knowledge no attempts to compare the FNR binding to cytochrome b6 f complexes isolated from state-1 and state-2 adapted cells has been performed so far. Its relevant role in the commutation between linear and cyclic flow is nevertheless clearly underlined in cyanobacteria, where salt stress resulted in the over-expression of FNR, which binds to the thylakoid membranes (van Thor et al. 2000). At the same time, this stress treatment enhanced the activity of cyclic flow, which has normally

a limited contribution to the overall electron flux in physiological conditions (see, e.g., Yu et al. 1993). Another element of potential relevance in the process of cyclic flow is the additional c type heme that has been identified in the three-dimensional structure of the cytochrome b6 f complex of Chlamydomonas (Stroebel et al. 2003) and Mastigocladus laminosum (Kurisu et al. 2003). This heme is indeed located in the plastoquinone reducing site (Qi), where it occupies an intermediate position between the bh heme and the plastoquinone itself. It is also directly accessible by the stroma, through a region where contacts with hydrophilic partners (FNR?) could take place. Obviously, it is tempting to associate this heme with Lavergne/Joliot’s ‘soluble’ G cytochrome that was identified in green algae (reviewed in Cramer et al. 1996) and already proposed to participate in cyclic electron flow. As a final alternative hypothesis, the role of a nucleotide-driven plastoquinone reductase in the cyclic flow in Chlamydomonas should also be considered. The involvement of such activity has already been proposed in the chlororespiratory pathway (reviewed in Peltier and Cournac 2002), as well as in the cyclic electron flow of cyanobacteria (Yu et al. 1993) and vascular plants (reviewed in Peltier and Cournac 2002). According to this alternative hypothesis, commutation to cyclic flow in state 2 would occur because of the different activation of the reductase depending on the redox state of the PS I soluble acceptors (reviewed in Peltier and Cournac 2002), which is obviously different in oxidising (state 1) or reducing (state-2) conditions. Although the nature of this reductase remains unknown in Chlamydomonas, it is clear that the ability to inject electrons into the plastoquinone pool at the expense of cellular reducing power is particularly efficient in the case of this alga, with respect to vascular plants (reviewed in Peltier and Cournac 2002). One possible clue about the nature of this enzyme in Chlamydomonas is the finding that the PGR5 gene is present in its genome (contig 20021010.5695.1). The precise function of the protein encoded by this gene is not yet understood. Even if it has been proposed to play a role in the direct electron transfer to the cytochrome b6 f complex (Munekage et al. 2002, 2003), its absence in an A. thaliana mutant clearly decreases the efficiency of NADPH (and/or Fd?) mediated reduction of the plastoquinone pool in isolated chloroplasts (Munekage et al. 2002, 2003). This might suggest its direct participation in the so-called an-

334 timycin sensitive cyclic electron flow around PS I (reviewed in Bendall and Manasse 1995). Metabolic implications of state transitions in Chlamydomonas Besides the mechanism that underlines the strict relationship between state transitions and the commutation between linear and cyclic flow, another interesting aspect of this peculiar commutation is the metabolic one. State transition as a shift from oxygenic to bacterial photosynthesis From an energetic point of view, state transitions in Chlamydomonas represent a shift from an oxygenic type of photosynthesis (that generates both reducing power and ATP, state 1) to a bacterial photosynthesis, where only ATP is synthesised (state 2, Figure 2). This commutation may therefore represent an advantage in terms of capacity of adaptation to environmental changes. Indeed, because a large fraction of LHC II antenna migrates to PS I, the photosynthetic quantum yield remains high in cyclic flow conditions, because most of the absorbed photons can be used photochemically by PS I. The rationale of this form of adaptation would be therefore to use the absorbed light to maintain a high photosynthetic ability to synthesise ATP under conditions where photosynthetic CO2 assimilation could not occur. This would allow the maintenance of other vital processes, and enable the cells to re-start more rapidly the photosynthetic activity, once the stress is over. Consistently, it has been observed that a systematic transition to state 2 is induced in Chlamydomonas under nutrient deprivation (reviewed in Davies and Grossman 1998). In all these conditions, a systematic decrease of PS II activity is also observed, leading to a consequent decrease of the linear electron flow. In the case of another abiotic stress, high light, a decrease in the ability to perform linear flow was also observed (Finazzi et al. 2001), likely because of photoinhibition. On the contrary, no consequences were found on the ability to perform cyclic flow, and thus to maintain high ATP synthesis. The tight relationship between state transitions and the PS II repair cycle that follows photoinhibition has been recently reviewed by Kruse (2001). Interestingly, the same signalling pathway (changes in the redox state of the PQ pool) involved in the reversible transition to cyclic electron flow, is also

employed in a more general process, the ‘chloroplast redox signalling’ (reviewed in Pfannschmidt 2003). This phenomenon refers to a series of regulatory processes, which mediate a long-term adaptation of the photosynthetic apparatus to changes in the external conditions, by relating changes in the plastoquinone redox-state to gene expression. Energetic consequences of the linear-cyclic commutation under intermediate conditions Under more physiological conditions, i.e., intermediate conditions between state 1 and state 2, cyclic flow contributes only partially to photosynthesis. This is what one would expect, since cyclic flow is expected to provide the ATP required by the Calvin–Benson cycle in excess of that produced by the linear electron transport. It is indeed well known that linear flow generates ATP and NADPH in a 1:1 stoichiometry (Berry and Rumberg 1999), while 3 ATP and 2 NADPH are required per CO2 assimilated. It seems therefore that the existence of the strict relationship between state transitions and commutation to cyclic flow allows the precise matching between cyclic flow and cellular needs. Because of the substantial occurrence of cyclic flow in steady-state photosynthesis, CO2 assimilation in Chlamydomonas is not limited by ATP availability, but essentially by the rate of NADPH production (Forti et al. 2003). This is an important metabolic difference compared to vascular plants, where CO2 assimilation is limited by both ATP and NADPH generation and consumption (reviewed in Noctor and Foyer 2000). This suggests that in vascular plants, the implication of cyclic electron flow in steady state conditions is less relevant. On the other hand, the occurrence of cyclic flow can be clearly evidenced under ‘stress’ conditions, where PS II activity is largely reduced (reviewed in Bendall and Manasse 1995). Cyclic flow is also very active during a dark–light transition (Joliot and Joliot 2002), i.e., when ATP synthesis without NADP reduction is required to promote activation of the Calvin–Benson cycle. Under more physiological conditions, other mechanisms seem to modulate the efficiency of ATP generation in plants. For example, the linear electron transport of the Mehler-ascorbate system (reviewed in Asada 2000) is very active. Accordingly, it has been shown that O2 is continuously reduced during photosynthesis at 25 to 30% of the rate of CO2 assimilation (Egneus et al. 1975). This process is coupled to ATP formation with the same ATP/electron stoichiometry

335 as the NADP reduction system (Forti and Elli 1995). Thus, the correct ATP/NADPH stoichiometry required for CO2 assimilation might be obtained through the sole linear electron transport, well balanced in terms of antenna size by the state transitions. This system might be efficiently feedback regulated, depending upon the utilisation of ATP and NADPH for carbon assimilation. Indeed, the alternative reduction of NADP or O2 -ascorbate-free radical is regulated by the availability of the electron acceptor NADP, i.e., by the ATP/NADPH ratio in the stroma. Energetic consequences of the thioredoxins modulation of state transitions In light of this peculiar significance of state transition in Chlamydomonas, i.e., the modulation of ATP and NADPH synthesis, the involvement of thioredoxins in the regulation of state transitions assumes a particular relevance. Thioredoxins are reduced by the acceptors of PS I, under conditions where the flow of reducing equivalents to the Calvin–Benson cycle is limited. By reducing specific thiol bridges, they activate the ATPsynthase CF0 –F1 , some of the key enzymes of the Calvin–Benson cycle itself (reviewed in Knaff 1989), and of the NADP malate-dehydrogenase pathway, which is implicated in the export of reducing equivalents from the chloroplast (see a recent discussion in Lemaire et al. 2003). Consequently, they directly control the rate of ATP synthesis, and that of ATP and NADPH consumption. Their negative control on the LHC II-kinase, downstream of the plastoquinone pool regulation, might provide the fine-tuning of the extent of state transitions. This might be very important in light of the metabolic consequences of state transitions discussed above. By preventing excessive commutation to cyclic flow under physiological conditions, thioredoxins might exert a key regulatory function on the rate of NADPH synthesis as well. Therefore, they might play a central role in the regulation of the energetic processes related to CO2 assimilation (Figure 3).

Mechanisms of oxidation of the electron transport chain causing the state-2 to state-1 transition The transition from state 2 to state 1 upon illumination of dark anaerobic cells adapted to state 2 has been reported to occur after a variable lag of a few minutes, depending on light intensity. During this

lag, no O2 evolution is observed in Chlamydomonas (Finazzi et al. 1999), in agreement with cyclic electron flow activity, as discussed above. It has been proposed that the ATP/ADP ratio may control state transitions in Chlamydomonas (discussed in Wollman 2001). It has been demonstrated, however, that the level of ATP rises very rapidly (seconds), while the lag lasts longer (i.e., tens of seconds to minutes) and depends on the light intensity (Finazzi et al. 1999; Forti et al. 2003). This suggests that the triggering factor in the state1 transition is not only ATP generation, but also the balance between processes that lead to the reduction and the oxidation of the electron transport chain, and particularly of PQH2 . Oxidation of the chain occurs in the light and in the presence of an electron acceptor at the reducing side of PS I. Among the possible acceptors, O2 is the most effective, but oxaloacetate is also active (G. Forti et al., unpublished). Its influence on the lag is likely mediated by the NADP-dependent malatedehydrogenase present in the stroma (Scheibe 1987). The activity of the malate-oxalacetate shuttle leads to the oxidation of NADPH, and therefore to the oxidation of PQH2 . However, oxaloacetate has no oxidative effect in darkness, in agreement with the idea that the photochemical activity of PS I is required to oxidise the intersystem chain. The effect of oxygen in the light is therefore likely mediated by PS I-dependent reduction of O2 (Mehler reaction). Indeed, while O2 is effective in promoting state-2 to state-1 transition in the dark as well, the rate and extent of this transition are much lower than in illuminated cells (G. Forti et al., unpublished). This suggests that the oxidation of the PQ pool by either mitochondrial respiration (indirectly, through the shuttle of electrons from the chloroplast to the cytoplasm) or by the chlororespiratory pathway can oxidise PQH2 at low rate with respect to the Mehler reaction itself, consistently with the estimates obtained in vivo (reviewed in Peltier and Cournac 2002). The involvement of the Mehler reaction in the transition from state 2 to state 1 would explain the autocatalytic kinetics observed (G. Forti et al., unpublished). Indeed the initial generation of O2 by PS II is expected to increase O2 concentration, which is the limiting factor of the Mehler reaction. The stimulatory effect of O2 on the state-2 to state-1 transition is saturated at the concentration of 15–20 µM (G. Forti et al., unpublished), which might represent the affinity of this process for oxygen in vivo: such an affinity is 3 to 4 times lower than that of the mitochondrial respiration. On the other hand, the observation that ascorbate,

336

Figure 3. Flexible modulation of linear and cyclic flow commutation, owing to a double control mechanism of state transitions. Upon illumination of thylakoid membranes with PS II-absorbed light, reduction of the plastoquinone pool might induce state-2 transition, and consequently the commutation between the linear and cyclic electron flow (B). Due to the temporary inactivation of the Calvin–Benson enzymes, reduction of thioredoxins might also take place (B) leading to partial inhibition of the LHC II kinase, and the activation of both the Calvin–Benson enzymes and the ATP synthase CF0 –F1 . This would lead to the generation of an intermediate state where both linear and cyclic activity operate within the same chloroplast leading to an optimised CO2 assimilation (C). Upon illumination with PS I-enriched light, inactivation of the kinase would be induced by the combined oxidation of the plastoquinone pool and the reduction of the thioredoxin pool, leading to a complete reversion to state 1 (D).

or rather the free radical produced by its oxidation, is not a terminal electron acceptor in the chloroplast of Chlamydomonas (Forti et al. 2003) accounts for the rather poor ability of ascorbate to restore state-1 conditions (G. Forti et al., unpublished). In Chlamydomonas, the role of the ascorbateperoxidation might be the protection against the generation of reactive oxygen species (reviewed in Asada 2000) rather than the contribution to the energy requirements of photosynthesis. Consistent with this conclusion is also the finding that the properties of the ascorbate peroxidase – the key enzyme of the Mehler-ascorbate system – of Chlamydomonas are very different from those of higher plants (reviewed in Asada 2000).

Conclusions The tight relationship that exists between state transitions and the commutation from linear to cyclic electron flow appears to provide a very fast and efficient way to adapt the photosynthetic activity to modifications of the external milieu. Although rather specific to the case of Chlamydomonas, it is interesting to note that in plants as well, the occurrence of cyclic electron flow has been clearly identified in stress conditions, which closely resemble those inducing state 2 in Chlamydomonas. This raises the question of the real specificity of this phenomenon: is it possible that some of the phenomena presented here for Chlamydomonas are also shared by plants, although to a lesser

337 extent? In this case, it would be much easier to interpret the common interplay between ATP changes, redox poise in the chloroplast, and signal transduction for genetic regulation between chloroplast and nucleus – and within the chloroplast itself, which is largely shared between plants and algae (reviewed in Goldschmidt-Clermont 1998; Wollman 2001). Independently of its role in Chlamydomonas, the possibility of modulating different pathways of electron transfer simply by performing a segregation of diffusing carriers in the membranes is of particular interest, as it constitutes a very rapid and low-cost system to adjust the energetic metabolism to the cellular needs. As most of the membrane systems involved in cellular energetics are densely packed, it is tempting to assume that this process is not confined to Chlamydomonas thylakoids, but also to other membrane systems. This might be typically the case in mitochondria, where a strict compartmentation of the electron carriers between the inner membrane and the cristae membranes seems to be induced by a particular three dimensional structure (reviewed in Mannella et al. 2001). In agreement with this possibility is the finding that in yeast mitochondria, changes in the external growth conditions modify the compositions of the supercomplexes formed between the various components of the respiratory chain (reviewed in Shagger 2002). This likely reflects deep changes in terms of the respiratory activity. Acknowledgements We thank Pierre Joliot and Daniel Picot for discussions and Cécile Breyton for discussions and critical reading of the manuscript. References Albertsson PÅ (2001) A quantitative model of the domain structure of the photosynthetic membrane. Trends Plant Sci 6: 349–354 Allen JF (1992) Protein phosphorylation in regulation of photosynthesis. Biochim Biophys Acta 1098: 275–335 Allen JF, Bennett J, Steinback KE and Arntzen CJ (1981) Chloroplast protein phosphorylation couples plastoquinone redox state to distribution of excitation energy between photosystems. Nature 291: 25–29 Asada K (2000) The water–water cycle as alternative photon and electron sinks. Phil Trans R Soc London B Biol Sci 355: 1419–1431 Bendall DS and Manasse RS (1995) Cyclic photophosphorylation and electron transport. Biochim Biophys Acta 1229: 23–38 Berry S and Rumberg B (1999) Proton to electron stoichiometry in electron transport of spinach thylakoids. Biochim Biophys Acta 1410: 248–261 Bonaventura C and Myers J (1969) Fluorescence and oxygen evolution from Chlorella pyrenoidosa. Biochim Biophys Acta 189: 366–383

Breyton C (2000) The cytochrome b6 f complex: structural studies and comparison with the bc1 complex. Biochim Biophys Acta 1459: 467–474 Campbell D, Hurry V, Clarke AK, Gustafsson P and Oquist G (1998) Chlorophyll fluorescence analysis of cyanobacterial photosynthesis and acclimation. Microbiol Mol Biol 62: 667–683 Cramer WA, Soriano GM, Ponomarev M, Huang D, Zhang H, Martinez SE and Smith JL (1996) Some new structural aspects and old controversies concerning the cytochrome b6 f complex of oxygenic photosynthesis. Annu Rev Plant Physiol Plant Mol Biol 47: 477–508 Davies JP and Grossman AR (1998) Responses to deficiencies in micronutriens. In: Rochaix JD, Goldschmidt-Clermont M and Merchant S (eds) The Molecular Biology of Chloroplasts and Mitochondria in Chlamydomonas, pp 613–635. Kluwer Academic Publishers, Dordrecht, The Netherlands Delosme R, Olive J and Wollman FA (1996) Changes in light energy distribution upon state transitions: an in vivo photoacoustic study of the wild type and photosynthesis mutants from Chlamydomonas reinhardtii. Biochim Biophys Acta 1273: 150–158 Delphin E, Duval JC, Etienne AL and Kirilovsky D (1998) Delta pH-dependent photosystem II fluorescence quenching induced by saturating, multiturnover pulses in red algae. Plant Physiol 118: 103–113 Depege N, Bellafiore S and Rochaix JD (2003) Role of chloroplast protein kinase Stt7 in LHC II phosphorylation and state transition in Chlamydomonas. Science 299: 1572–1575 Egneus H, Heber U, Matthiesen U and Kirk M (1975). Reduction of oxygen by the electron transport chain of chloroplasts during assimilation of carbon dioxide. Biochim Biophys Acta 408: 252–268 Finazzi G, Furia A, Barbagallo RP and Forti G (1999) State transitions, cyclic and linear electron transport and photophosphorylation in Chlamydomonas reinhardtii. Biochim Biophys Acta 1413: 117–129 Finazzi G, Barbagallo RP, Bergo E, Barbato R and Forti G (2001) Photoinhibition of Chlamydomonas reinhardtii in state 1 and state 2: damages to the photosynthetic apparatus under linear and cyclic electron flow. J Biol Chem 276: 22251–22257 Finazzi G, Rappaport F, Furia A, Fleischmann M, Rochaix JD, Zito F and Forti G (2002) Involvement of state transitions in the switch between linear and cyclic electron flow in Chlamydomonas reinhardtii. EMBO Rep 3: 280–285 Forti G and Elli G (1995) The function of ascorbic acid in photosynthetic phosphorylation. Plant Physiol 109: 1207–1211 Forti G and Zanetti G (1969) The electron pathway of cyclic photophosphorylation. In: Metzner H (ed) Progress in Photosynthesis Research, Vol III, pp 1213–1216. Metzner Publisher, Tübingen, Germany Forti G, Furia A, Bombelli P and Finazzi G (2003) In vivo changes of the oxidation–reduction state of NADP and of the ATP/ADP cellular ratio linked to the photosynthetic activity in Chlamydomonas reinhardtii. Plant Physiol 132: 1464–1474 Fulgosi H, Vener AV, Altschmied L, Herrmann RG and Andersson B (1998) A novel multi-functional chloroplast protein: identification of a 40 kDa immunophilin-like protein located in the thylakoid lumen. EMBO J 17: 1577–1587 Gal A, Zer H and Ohad I (1997) Redox-controlled thylakoid protein phosphorylation. News Views Physiol Plant 100: 869–885 Georgakopulos JH and Argyroudi-Akoyunoglu KH (1994) On the question of the lateral migration of LHCCI upon thylakoid protein phosphorylation in isolated pea chloroplasts: the stroma lamellar fraction separated from phosphorylated chloroplasts is not homogeneous. Biochim Biophys Acta 1188: 380–390 Goldschmidt-Clermont M (1998) Coordination of nuclear and chloroplast gene expression in plant cells. Int Rev Cytol 177: 115–180

338 Haldrup A, Jensen PE, Lunde C and Scheller HV (2001) Balance of power: a view of the mechanism of photosynthetic state transitions. Trends Plant Sci 6: 301–305 Horton P and Black MT (1981) Light-dependent quenching of chlorophyll fluorescence in pea chloroplasts induced by adenosine 5 -triphosphate. Biochim Biophys Acta 635: 53–62 Horton P, Ruban AV and Walters AV (1996) Regulation of light harvesting in green plants. Annu Rev Plant Physiol Plant Mol Biol 47: 655–684 Jennings RC, Garlaschi FM, Gerola PD and Forti G (1979) Partition zone penetration by chymotrypsin and the localization of the chloroplast flavoprotein and Photosystem II. Biochim Biophys Acta 546: 207–219 Joliot P and Joliot A (2002) Cyclic electron transfer in plant leaf. Proc Natl Acad Sci USA 99: 10209–10214 Knaff DB (1989) The regulatory role of thioredoxin in chloroplasts. Trends Biochem Sci 14: 433–434 Kruse O (2001) Light-induced short-term adaptation mechanisms under redox control in the PS II–LHC II supercomplex: LHC II state transitions and PS II repair cycle. Naturwissenschaften 88: 284–292 Kurisu G, Zhang H, Smith JL and Cramer WA (2003) Structure of the cytochrome b6 f complex of oxygenic photosynthesis: tuning the cavity. Science 302: 1009–1014 Lavergne J and Joliot P (1991) Restricted diffusion in photosynthetic membranes. Trends Biochem Sci 16: 129–134 Lemaire SD, Collin V, Keryer E, Quesada A and Miginiac-Maslow M (2003) Characterization of thioredoxin y, a new type of thioredoxin identified in the genome of Chlamydomonas reinhardtii. FEBS Lett 543: 87–92 Lunde C, Jensen PE, Haldrup A, Knoetzel J and Scheller HV (2000) The PS I-H subunit of Photosystem I is essential for state transitions in plant photosynthesis. Nature 408: 613–615 Majeran W, Olive J, Drapier D, Vallon O and Wollman FA (2001) The light sensitivity of ATP synthase mutants of Chlamydomonas reinhardtii. Plant Physiol 126: 421–433 Mannella CA, Pfeiffer DR, Bradshaw PC, Moraru II, Slepchenko B, Loew LM, Hsieh CE, Buttle K and Marko M (2001) Topology of the mitochondrial inner membrane: dynamics and bioenergetic implications. IUBMB Life 52: 93–100 Morgan-Kiss RM, Ivanov AG and Huner NP (2002) The Antarctic psychrophile, Chlamydomonas subcaudata, is deficient in state I–state II transitions. Planta 214: 435–445 Munekage Y, Hojo M, Meurer J, Endo T, Tasaka M and Shikanai T (2002) PGR5 is involved in cyclic electron flow around Photosystem I and is essential for photoprotection in Arabidopsis. Cell 110: 361–361 Murata N (1969) Control of excitation transfer in photosynthesis. I. Light-induced change of chlorophyll a fluorescence in Porphyridium cruentum. Biochim Biophys Acta 172: 242–251 Nilsson A, Stys D, Drakenberg T, Spangfort MD, Forsen S and Allen JF (1997) Phosphorylation controls the three-dimensional structure of plant light harvesting complex II. J Biol Chem 272: 18350–18357 Noctor G and Foyer CH (2000) Homeostasis of adenylate state during photosynthesis in a fluctuating environment. J Exp Bot 51: 347–356 Owens TG (1986) Light-harvesting function in the diatom Phaedactylum tricornutum. II Distribution of excitation energy between the two photosystems. Plant Physiol 80: 739–746 Peltier G and Cournac L (2002) Chlororespiration. Annu Rev Plant Biol 53: 523–550

Pfannschmidt T (2003) Chloroplast redox signals: how photosynthesis controls its own genes. Trends Plant Sci 8: 33–41 Rintamaki E, Martinsuo P, Pursiheimo S and Aro EM (2000) Cooperative regulation of light-harvesting complex II phosphorylation via the plastoquinol and ferredoxin–thioredoxin system in chloroplasts. Proc Natl Acad Sci USA 97: 11644–11649 Sarcina M, Tobin MJ and Mullineaux CW (2001) Diffusion of phycobilisomes on the thylakoid membranes of the cyanobacterium Synechococcus 7942. J Biol Chem 276: 46830–46834 Schagger H (2002) Respiratory chain supercomplexes of mitochondria and bacteria. Biochim Biophys Acta 1555: 154–159 Scheibe R (1987) NADP +-malate dehydrogenase in C3 plants: regulation and role of light activated enzyme. Physiol Plant 71: 393–400 Scheller HV, Jensen PE, Haldrup A, Lunde C and Knoetzel J (2001) Role of subunits in eukaryotic Photosystem I. Biochim Biophys Acta 1507: 41–60 Snyders S and Kohorn BD (2001) Disruption of thylakoidassociated kinase 1 leads to alteration of light harvesting in Arabidopsis. J Biol Chem 276: 32169–32176 Stroebel D, Choquet Y, Popot JL and Picot D (2003) An atypical haem in the cytochrome b6 f complex. Nature 426: 413–418 Vallon O, Bulté L, Dainese P, Olive J, Bassi R and Wollman FA (1991) Lateral redistribution of cytochrome b6 /f complexes along thylakoid membranes upon state transitions. Proc Natl Acad Sci USA 88: 8262–8266 van Thor JJ, Jeanjean R, Havaux M, Sjollema KA, Joset F, Hellingwerf KJ and Matthijs HC (2000) Salt shock-inducible Photosystem I cyclic electron transfer in Synechocystis PCC6803 relies on binding of ferredoxin:NADP(+)reductase to the thylakoid membranes via its CpcD phycobilisome-linker homologous N-terminal domain. Biochim Biophys Acta 1457: 129–144 Vener AV, van Kan PJ, Rich PR, Ohad I and Andersson B (1997) Plastoquinol at the quinol oxidation site of reduced cytochrome bf mediates signal transduction between light and protein phosphorylation: thylakoid protein kinase deactivation by a singleturnover flash. Proc Natl Acad Sci USA 94: 1585–1590 Vener AV, Ohad I and Andersson B (1998) Protein phosphorylation and redox sensing in chloroplast thylakoids. Curr Opin Plant Biol 1: 217–223 Wollman FA (1978) Determination and modification of the redox state of the secondary acceptor of Photosystem II in the dark. Biochim Biophys Acta 503: 263–273 Wollman FA (2001) State transitions reveal the dynamics and flexibility of the photosynthetic apparatus. EMBO J 20: 3623–3630 Yu L, Zhao J, Muhlenhoff U, Bryant DA and Golbeck JH (1993) PsaE is required for in vivo cyclic electron flow around Photosystem I in the cyanobacterium Synechococcus sp. PCC 7002. Plant Physiol 103: 171–180 Zer H, Vink M, Shochat S, Herrmann RG, Andersson B and Ohad I (2003) Light affects the accessibility of the thylakoid light harvesting complex II (LHC II) phosphorylation site to the membrane protein kinase(s). Biochemistry 42: 728–738 Zhang H, Whitelegge JP and Cramer WA (2001) Ferredoxin:NADP+ oxidoreductase is a subunit of the chloroplast cytochrome b6 f complex. J Biol Chem 276: 38159–38165 Zito F, Vinh J, Popot JL and Finazzi G (2002) Chimeric fusions of subunit IV and PetL in the b6 f complex of Chlamydomonas reinhardtii: structural implications and consequences on state transitions. J Biol Chem 277: 12446–12455