Heat treatment of intact spinach leaves was found to induce a unique thylakoid membrane association of an approximately 40 kDa stromal protein. This protein was identified as rubisco activase. Most of the rubisco activase was sequestered to the thylakoid membrane, particularly to the stroma-exposed regions, during the first 10 min of heat treatment at 42°C. At lower temperatures (38–40°C) the association of rubisco activase with the thylakoid membrane occurred more slowly. The temperature-dependent association of rubisco activase with the thylakoid membrane was due to a conformational change in the rubisco activase itself, not to heat-induced alterations in the thylakoid membrane. Association of the 41 kDa isoform of rubisco activase occurred first, followed by the binding of the 45 kDa isoform to the thylakoid membrane. Fractionation of thylakoid membranes revealed a specific association of rubisco activase with thylakoid-bound polysomes. Our results suggest a temperature-dependent dual function for rubisco activase. At optimal temperatures it functions in releasing inhibitory sugar phosphates from the active site of Rubisco. During a sudden and unexpected exposure of plants to heat stress, rubisco activase is likely to manifest a second role as a chaperone in association with thylakoid-bound ribosomes, possibly protecting, as a first aid, the thylakoid associated protein synthesis machinery against heat inactivation.
Photosynthesis is one of the most heat-sensitive functions of plants. Exposure of plants to elevated temperatures results in a rapid inhibition of photosynthetic CO2 fixation, oxygen evolution, and photophosphorylation (Berry and Björkman, 1980). Limitation of CO2 fixation by high temperature is accompanied by the inactivation of an enzyme, rubisco activase, leading to a decrease in Rubisco activity (Feller et al., 1998). Of thylakoid-associated processes, on the other hand, the primary target of thermal damage is photosystem II (PS II) (Berry and Björkman, 1980; Havaux and Tardy, 1996). Heat inactivation of PS II induces a release of the extrinsic proteins of the oxygen-evolving complex to the thylakoid lumen (Enami et al., 1994).
Sudden exposure of plants to heat stress, however, also induces mechanisms to minimize the damages. Although there is a general decrease in the synthesis of many constitutively-expressed proteins, the translation machinery keeps active and the synthesis of several heat shock proteins (HSPs) is enhanced (Key et al., 1981; Somers et al., 1989). Induction of HSP synthesis commonly occurs immediately after the shift of plants to an elevated temperature (Vierling, 1991). Plants typically produce more than 10 different small HSPs (sHSPs) that belong to five distinct gene classes encoding proteins targeted to different cell compartments (Vierling, 1991). The chloroplast sHSPs, synthesized and translocated into the soluble stromal fraction of chloroplasts after the first heat shock (Chen et al., 1990), bind reversibly to the thylakoid membrane upon the second heat shock (Adamska and Kloppstech, 1991; Eisenberg-Domovich et al., 1994; Glaczinski and Kloppstech, 1988). Although the synthesis of sHSPs, in general, seems to be important for plants in acquiring thermotolerance, the HSP mechanisms that protect plants during or after heat stress are only poorly understood.
Chloroplast-located sHSPs have been implicated in the protection of the oxygen-evolving complex and electron transport of PSII against heat damage (Downs et al., 1999; Heckathorn et al., 1998). Plants that survive the primary heat stress become resistant to the second stress, which could otherwise be lethal (Lindquist and Craig, 1988; Vierling, 1991). Plant cytosolic HSPs can form complexes containing mostly sHSP, but also HSP70 and RNA (Neumann et al., 1984; Nover et al., 1989). These complexes are supposed to function as transient storage for non-heat-shock mRNA, preventing their degradation during heat stress (Nover et al., 1989). Such complexes, however, have not been found in chloroplasts.
Here we report a novel heat–induced association of a chloroplast 40 kDa protein with the thylakoid membrane. The protein was identified as rubisco activase by microsequencing. Further fractionation of the thylakoid revealed an association of rubisco activase upon the first heat stress with thylakoid-bound polysomes. Recruitment of rubisco activase to the chloroplast stroma occurred upon recovery from heat shock. A possible dual function of rubisco activase, activation of Rubisco under physiological conditions and protection of thylakoid-associated functions under heat stress conditions, is discussed.
Heat shock induced association of a 40-kDa protein with the thylakoid membrane
Heat treatment of spinach leaves for 60 min at 42°C induced a novel protein band in the thylakoid membrane preparation with a molecular mass of approximately 40 kDa (Figure 1a). To test whether the 40 kDa protein was a newly synthesized thylakoid protein induced by the heat shock, we next subjected spinach leaves to pulse labelling experiments with [35S] methionine.
As shown in Figure 1(b,c), the D1 reaction centre protein of PSII was the most rigorously synthesized thylakoid protein under both the control (23°C) and heat shock (42°C) conditions in light, but not in darkness. Several proteins with a molecular mass in a range of 16–25 kDa and one protein of approximately 70 kDa, on the other hand, were synthesized both in light and in darkness, particularly under heat stress conditions. Nevertheless, no labelled protein corresponding to the novel heat-induced thylakoid-associated 40 kDa one was detected. This implied that the 40 kDa protein was not synthesized de novo during the heat shock treatment, but rather the heat treatment had induced an association of the 40 kDa stromal protein with the thylakoid membrane. Such heat-induced membrane association of soluble proteins is not exceptional, as it has been shown previously that the sHSPs, imported into chloroplast stroma during the first heat shock only bind to the thylakoid membrane during the second heat shock (Adamska and Kloppstech, 1991; Eisenberg-Domovich et al., 1994; Glaczinski and Kloppstech, 1988).
Identification of the 40 kDa protein by microsequencing
The novel heat-induced thylakoid associated 40 kDa protein, separated by SDS-PAGE, was subjected to microsequencing. After trypsin digestion and separation of peptides by HPLC, an N-terminal and four internal peptide sequences were obtained (marked with boxes, Figure 2). Based on database searches, the peptide sequences matched with those of rubisco activase. Rubisco activase has a 58 amino acid long transit peptide in the N-terminal end that is cleaved after translocation into the chloroplast (Werneke et al., 1988). An immunoblot with an antibody raised against rubisco activase (from Dr T.J. Andrews, The Australian National University, Canberra, Australia) revealed two isoforms of the enzyme, the 41 and 45 kDa polypeptides, associated with thylakoid membranes isolated from heat-shocked spinach leaves (Figure 3a). As shown in Figure 2, the two isoforms differ from each other at their carboxyl terminus with 37 additional amino acids in the 45 kDa form (Salvucci et al., 1987).
Temperature-dependent association of rubisco activase with the thylakoid membrane
One-hour incubation of leaf discs at 30°C did not lead to any association of rubisco activase with the membrane fraction (Figure 3a). Elevating the temperature to 38°C clearly induced a binding of rubisco activase to the thylakoid membrane and the amount of rubisco activase associated with the thylakoid membrane increased with prolonged incubation. Indeed, both the duration and the temperature of the heat treatment affected the binding of rubisco activase to the thylakoid membrane: similar association was observed whether the leaf discs were incubated at 38°C for 2 h or at 40°C for 1 h. At 42°C, after only 10 min of treatment, nearly all the rubisco activase was detected in the membrane fraction (Figure 3). A loss of rubisco activase from the soluble stroma fraction (Figure 3b) occurred concomitantly with the heat-induced association of rubisco activase with the thylakoid membrane (Figure 3a). The 41 kDa isoform of rubisco activase associated first with the membrane and was followed by association of the 45 kDa isoform. As a control, we also determined the effect of heat treatment on the association of another soluble protein, Rubisco, with the thylakoid membrane. As shown in Figure 3(a), some residual amount of Rubisco always stayed bound to the membrane, but clearly, the amount of bound Rubisco did not change during the heat treatment of spinach leaves.
Heat shock treatment induces a structural change in rubisco activase
Heat-induced association of rubisco activase with the thylakoid membrane could be due to a heat-induced structural change either in the rubisco activase protein itself or in the thylakoid membrane. To distinguish between these two possibilities, the thylakoid membrane fraction, isolated from untreated spinach leaves, was first heat shocked for 30 min at 42°C and subsequently combined with an equal volume of fresh stroma fraction isolated from untreated leaves. The thylakoid and stroma fractions were then incubated together at 23°C for 15 min. Subsequently the thylakoid membrane and stroma fractions were again separated by centrifugation (14 000 g, 5 min) and both fractions were analysed for the presence of rubisco activase. Under these conditions all rubisco activase remained in the soluble stroma fraction (supernatant) and no association of rubisco activase with the membrane fraction was observed (Figure 4a, lanes marked 1). The opposite treatment, i.e. combining a heat-treated stoma fraction with untreated thylakoid membranes, resulted in the rubisco activase binding to the thylakoid membrane with a concomitant loss of rubisco activase from the soluble stroma (Figure 4a, lanes marked 2). On the contrary, when the stroma fraction alone was heat-treated at 42°C for up to 30 min, all rubisco activase remained in the supernatant after centrifugation at 14 000 g for 15 min (Figure 4b). This indicated that the association of rubisco activase with the thylakoid membrane upon heat treatment of intact leaves did not result from aggregation of rubisco activase and cosedimentation with thylakoid membranes, which has been reported to occur for sHSPs (Osteryoung and Vierling, 1994). On the contrary, it is plausible that the heat treatment induced a subunit dissociation of rubisco activase as shown earlier (Wang et al., 1993) and a concomitant structural change in rubisco activase, which increased its binding affinity to the thylakoid membrane.
Release of rubisco activase from the thylakoid membrane
To study the stability of rubisco activase association with the thylakoid membrane in heat treated leaves, the thylakoid membranes were washed with 2 m NaCl or 2 m NaBr, or treated for 30 min with 1% Triton X-100. Washing with 2 m NaBr has been shown to remove extrinsic thylakoid proteins including CF1, the stroma-exposed domain of ATP-synthase (Hurt and Hauska, 1981). However, none of these treatments released noticeable amounts of rubisco activase (Figure 5a), indicating a tight association of heat-treated rubisco activase with the thylakoid membrane.
Strong membrane association of rubisco activase raised the question of whether rubisco activase could be partly inserted into the thylakoid membrane or translocated into the lumenal face of thylakoids, even though it does not contain a typical lumen-targeting sequence. To address this question, thylakoids were treated by limited proteolysis with thermolysin. This treatment digests stroma-exposed proteins but lumenal ones are mainly protected against degradation. As shown in Figure 5(b), several distinct digestion fragments of rubisco activase could be observed indicating that rubisco activase was not on the lumenal side of thylakoids. However, we cannot exclude the possibility that some domains of the 41 kDa isoform of rubisco activase were inserted into the thylakoid membrane.
Localization of heat-induced association of rubisco activase in the thylakoid membrane
Heat-treated leaf discs were subfractionated into rightside out (R.O.) and inside out (I.O.) vesicles, corresponding to non-appressed and appressed thylakoid regions, by Yeda press treatment and two-phase partitioning. Rubisco activase was almost exclusively detected in R.O. vesicles demonstrating the binding of rubisco activase to the stroma-exposed membrane regions including the end membranes and margins of the grana (Figure 6a).
To test whether rubisco activase is associated specifically with some distinct thylakoid membrane protein complex upon a heat treatment, the protein complexes were released by N-dodecyl-β,d-maltoside (DM) solubilization and separated by sucrose gradient centrifugation. By this method the main transthylakoid pigment-protein complexes can be separated from each other's (Zhang et al., 1999). Analysis of the sucrose gradient fractions (Figure 6b), however, revealed that rubisco activase was not associated with any of the major thylakoid protein complexes. Instead, rubisco activase was specifically found in fractions towards the bottom of the gradient. These fractions contain ribosomes that are released from the thylakoid membrane by DM solubilization (L. Zhang et al., unpublished results).
We next isolated the thylakoid membrane-bound polysomes and, indeed, a large quantity of rubisco activase from heat-treated leaves was detected in this fraction (Figure 7) which we have earlier shown to be enriched in ribosome nascent chain complexes (Zhang et al., 1999; Zhang et al., 2000).
Association and dissociation of rubisco activase from the thylakoid membrane in vivo
Next we used intact plants to explore the association/dissociation process of rubisco activase with the thylakoid membrane. It has been shown previously that the leaf temperature lags behind the air temperature and reaches the highest level during the second hour of heat stress, being approximately 1°C lower than the air temperature (Chen et al., 1990). Already after a 1 h exposure of intact plants to 42°C, rubisco activase was found to be associated with the thylakoid membrane and the level did not increase with increase in the incubation time (Figure 8). During 10 h and 22 h recovery periods at 23°C, approximately 60% and 70% of membrane bound rubisco activase, respectively, had released from thylakoids (Figure 8). As the total amount of rubisco activase in cells remained constant (data not shown), it is likely that rubisco activase released from the thylakoid membrane to the soluble stroma. A second heat treatment again induced an accumulation of rubisco activase to the thylakoid membrane, but less was bound compared to the first heat treatment (Figure 8). Also, during prolonged heat treatment of up to 4 h, the amount of membrane- bound rubisco activase was clearly diminished from that after 2 h.
We demonstrate a unique association of a stromal protein rubisco activase with the thylakoid membrane upon a heat treatment of intact spinach leaves. Primarily, under physiological conditions, rubisco activase functions in chloroplast stroma in removing inhibitory sugar phosphates from the active site of Rubisco in an ATP-dependent manner (Portis, 1990; Portis, 1992). Rubisco itself is a very thermostable enzyme as revealed by studies with isolated protein (Crafts-Brandner et al., 1997) whereas rubisco activase has been reported to be particularly sensitive to inactivation in situ by elevated temperatures (Crafts-Brandner et al., 1997; Echard and Portis, 1997; Robinson and Portis, 1989). Thus, inactivation of rubisco activase is probably a basis for the inactivation of Rubisco at elevated temperatures in intact plants.
In situ heat-inactivated rubisco activase does not remain soluble in the stroma (Feller et al., 1998; Figure 3b) but undergoes structural changes and becomes strongly associated with thylakoid membranes (Figure 3a). In the absence of thylakoids, however, the rubisco activase, heat-treated either in a crude stromal extract (Figure 4b) or as a purified protein in solution (Crafts-Brandner et al., 1997), remains completely soluble. Heat behaviour of rubisco activase thus differs from that of sHSPs in plants, as the latter have been shown to form large aggregates insoluble in non-ionic detergents (Osteryoung and Vierling, 1994). Indeed, the association of rubisco activase with the thylakoid membrane seems to involve two phases; a heat-induced structural change in rubisco activase and a subsequent specific interaction with the thylakoid membrane.
It has been shown that functional subunit associations, characteristic of active rubisco activase, break down at elevated temperatures (Wang et al., 1993). In Rubisco activation, the functional form of rubisco activase seems to be a high molecular weight complex of the 41 and 45 kDa isoforms (Portis, 1990; Wang et al., 1993). Depending on conditions, the size of the isolated enzyme varies between 140 kDa and 600 kDa, heat-inactivated enzyme having the smallest observed mass of 140 kDa (Wang et al., 1993). Binding of Mg2+ and ATP induces a conformational change in rubisco activase that results in increased subunit association and also in an increase in the activity of ATP hydrolysis (Wang et al., 1993). High temperature, on the other hand, seems to induce a dissociation of rubisco activase subunits, the 45 kDa subunits showing more heat-resistant associations than the 41 kDa isoforms (Crafts-Brandner et al., 1997). Dissociation of the rubisco activase complex apparently leads to an exposition of hydrophobic parts of the protein subunits, which enhance the binding of the protein to the thylakoid membrane.
What could be the physiological role of rubisco activase association with the thylakoid membrane at elevated temperature? Based on database searches and multiple sequence alignments, Neuwald et al. (1999) recently reported that rubisco activase is related to an AAA family of proteins, a class of chaperone-like ATPases associated with a variety of cellular activities. Several motifs conserved in AAA + class proteins were also found in rubisco activase. Proteins in the AAA + class often perform chaperone-like functions that assist in the assembly, operation or disassembly of protein complexes. AAA + modules are also often linked covalently to other protein domains that mediate localization to cellular membranes.
Since the breakdown of functional subunit associations of rubisco activase at elevated temperatures is a likely reason for reversible enzyme inactivation (Wang et al., 1993), it is possible that, upon thylakoid association, rubisco activase regains its ATPase activity and manifests a novel function as a chaperone. Rubisco activase indeed possesses chaperone-like activity as it has been shown in vitro to be capable of restoring the activity of heat-inactivated Rubisco (Sánchez de Jiménez et al., 1995). Given such a chaperone function for rubisco activase raises intriguing possibilities for its metabolic function upon thylakoid association during the heat treatment. Upon a heat shock, the loss of subunit interaction renders rubisco activase inactive in activation of Rubisco and instead, rubisco activase is efficiently sequestered to the thylakoid membrane. Further analysis of rubisco activase association with the thylakoid membrane revealed no stable interaction of rubisco activase with any intrinsic thylakoid protein complex. Instead, rubisco activase was found in the same fraction with ribosome nascent polypeptide complexes. Therefore, we suggest that, under heat shock conditions, rubisco activase manifests a second metabolic role as a chaperone, either in helping to target the ribosome complexes to the thylakoid membrane or in protection of the thylakoid associated translation machinery against heat-inactivation.
Generally, both in eukaryotes and prokaryotes, HSPs have been shown to protect protein synthesis at elevated temperatures. In mammalian cells HSP72 associates with polysomes of thermotolerant cells during a heat shock (Beck and De Maio, 1994). It was speculated that interaction of HSP72 with an unfolded nascent polypeptide maintains the growing polypeptide in solution during the stress, and thereby allows a constant rate of translation. More recently, HSP101 has been reported to function in translational regulation both in plant and yeast cells (Ling et al., 2000; Wells et al., 1998). Similar kinds of functions have also been assigned for sHSPs. In Escherichia coli, HSP15 binds with high affinity to a ribosome 50S subunit, which contains bound nascent chain. HSP15 was suggested to be involved in the recycling of free 50S subunits that still carry a nascent chain (Korber et al., 2000). Also several plant cytosolic HSPs of 15–18 kDa, and some 69–70 kDa HSPs, have been demonstrated to associate with ribosomes during heat stress (Lin et al., 1984). The association of rubisco activase with thylakoid-bound ribosomes upon a heat shock may point to an analogous role for rubisco activase on thylakoid-associated polysomes, particularly in the beginning of the heat shock.
We have recently shown that the heat treatment of thylakoid membranes induces a rapid dephosphorylation of the PSII reaction centre protein D1 (Rokka et al., 2000) and this is followed by a rapid degradation of damaged D1 copies (Figure 1c). To guarantee a constant repair of PSII complexes, and thereby to avoid disassembly and irreversible damage of PSII, it is crucial to maintain active protein synthesis in chloroplasts, despite the transient heat shock. As shown in Figure 1(b), the thylakoid-associated translation machinery remains active even at an elevated temperature and keeps synthesizing D1 in high amounts in light. Therefore, we suggest that the specific association of rubisco activase with thylakoid-bound ribosome nascent chain complexes exposes a chaperone-like function of rubisco activase, which is crucial in maintaining the translation activity at transient elevated temperatures.
Taken together, our results suggest a temperature-dependent dual function for rubisco activase. A switch of the protein from one function to another according to the needs of the cell under fluctuating environmental conditions represents a great design, not only for cell protection, but also for saving of metabolic energy, particularly if such environmental fluctuations are only transient. An analogous situation was recently shown for the HSP DegP, whose function switches from chaperone at low temperature to protease with an increase in temperature (Spiess et al., 1999). As for rubisco activase, at optimal temperatures it has an essential function in releasing inhibitory sugar phosphates from the active site of Rubisco, thereby allowing continuous CO2 fixation (Portis, 1992). A second function, deduced from the observed association of rubisco activase with thylakoid-bound polysomes upon a heat stress, is probably to maintain translation of essential thylakoid proteins, even during a sudden and unexpected exposure of plants to heat stress. The fast spatial segregation of rubisco activase to the thylakoid membrane upon a heat stress strongly suggests a role for rubisco activase as a chaperone in maintaining and protecting the thylakoid associated translation apparatus. With prolonged heat stress and upon repetition of the stress, actual de novo synthesized heat shock proteins are already available to rescue thylakoid functions and less rubisco activase was found to be associated with thylakoid protein synthesis machinery.
Spinach (Spinacia oleracea, L.) was grown hydroponically under a photon flux density (PFD) of 400 µmol m−2 s−1 with a 10 h light/14 h dark rhythm at 23°C. Mature leaves of plants were used for experiments.
Discs, 3 cm in diameter, were punched from spinach leaves and exposed to heat treatments at different temperatures while floating on preheated distilled water. After the heat treatment, the leaf discs were rapidly frozen in liquid nitrogen and stored at −80°C.
Intact spinach plants were used in experiments to study the reversibility of heat shock responses. Heat shock treatments were given to plants at 42°C under a PFD of 200 µmol m−2 s−1 at 90% relative humidity for 2 h and recovery was followed for 22 h at 23°C under growth light using normal light/dark rhythm starting with a light period.
Isolation and subfractionation of chloroplasts
For in vitro experiments, intact chloroplasts were isolated in percoll gradient (Cline, 1986) and, after an osmotic shock, the thylakoid membrane and stroma fractions were further separated by differential centrifugation. Thylakoids were suspended in a buffer containing 10 mm HEPES-NaOH, pH 7.5, 100 mm sucrose, 5 mm NaCl and 10 mm MgCl2. Stoma fraction was concentrated with a Centricon concentrator (cut off 10 kDa, Millipore) and supplemented with 5 mm MgCl2 and 2 mm ATP.
For thylakoid subfractionation, isolated thylakoids were passed twice through a Yeda press (Linca, Lamon Instrumentation Co. Ltd, Tel Aviv, Israel) followed by separation of the rightside out (R.O.) and inside out (I.O.) vesicles by aqueous polymer two-phase partition (Andersson and Anderson, 1980). I.O. vesicles represent the appressed membranes of the grana stacks, whereas R.O. vesicles are derived from non-appresed thylakoid regions including grana margins and end membranes.
Sucrose gradient (5–35%) fractionation of thylakoid membranes was performed according to Zhang et al. (1999). Thylakoids (0.5 mg chlorophyll ml−1) were solubilized with 1% (w/v) N-dodecyl-β,d-maltoside (DM, Sigma, St Louis, MO, USA), applied on the top of the sucrose gradient and spun at 180 000 g for 26 h. After centrifugation, the sucrose gradients were divided into 20 fractions of equal volume, and the proteins in each fraction were precipitated with acetone. Protein pellets were then washed with methanol and suspended in 50 µl of electrophoresis sample buffer.
Isolation of thylakoid membrane bound ribosome fraction
Thylakoids (equivalent to 1.0 mg of chlorophyll) were solubilized with 1% (w/v) DM in Medium A (50 mm HEPES-KOH, pH 8.0, 5 mm MgOAc, 50 mm KOAc, 250 µg ml−1 chloramphenicol, 0.5 mg ml−1 heparin, 2 mm DTT, 2 µg ml−1 antipain and leupeptin, and 0.1 mg ml−1 pefablock) and unsolubilized material was removed by 5 min centrifugation at 14 000 g. Ribosomes were collected by centrifugation through a 1 m sucrose cushion in Medium A at 270 000 g for 1 h and finally suspended in electrophoresis sample buffer.
Salt washing and thermolysin treatments of isolated thylakoids
Thylakoid membrane (2 mg chlorophyll ml−1), isolated from heat-treated leaf discs, was washed with 2 m NaCl or 2 m NaBr in 10 mm HEPES-KOH, pH 7.8, 5 mm MgCl2 or incubated for 30 min on ice in the presence of 1% (v/v) Triton X-100. Thylakoids were also incubated in the presence of thermolysin (0.4 mg ml−1, Sigma) for 2–12 min on ice in 50 mm Tricine, pH 7.8, 100 mm sorbitol, 5 mm NaCl, 10 mm MgCl2 and 2 mm CaCl2. Digestion was stopped by adding EDTA to a final concentration of 50 mm. After treatment, thylakoids were collected by 10 min centrifugation with full speed in an Eppendorf centrifuge.
Pulse labelling of thylakoid proteins with [35S] methionine
Leaf discs were first pressed gently against coarse sandpaper to facilitate incorporation of [35S] Met while floating on solution containing 6.7 µCi ml−1 of [35S] Met in 0.4% (v/v) Tween 20. Labelling was performed under a PFD of 50 µmol m−2s−1 or in darkness both at 23 or 42°C. After pulse periods of 1.5 and 3 h, leaf discs were washed with 0.4% (v/v) Tween 20 and thylakoid membrane was prepared for autoradiography
SDS-PAGE, immunoblotting and autoradiography
Samples were solubilized in the presence of 6 m urea and separated by SDS-PAGE (Laemmli, 1970) using 15% (w/v) acrylamide gels with 6 m urea. Thylakoid membrane fractions were loaded on chlorophyll basis (1 µg chlorophyll) and stroma fractions on protein basis (5 µg protein). Chlorophyll and protein concentrations were determined according to Porra et al. (1989) and Bradford (1976), respectively. For immunoblotting, polypeptides were transferred to a polyvinylidene fluoride (PVDF) membrane (Millipore, Saint-Quentin, France) and proteins were detected with antibodies raised against Rubisco activase (T.J. Andrews, The Australian National University, Canberra, Australia), Rubisco large subunit (M.A. Parry, IACR-Rothamsted, Harpenden, UK), or the D1 protein (Research Genetics, Huntsville, AL, USA). For autoradiography, the gels were stained, dried and exposed to X-ray films.
Thylakoid membrane proteins were separated by SDS-PAGE and transferred to a PVDF membrane. N-terminal amino acid sequence analysis was performed with an Applied Biosystems model 477 A protein sequencer equipped with an on-line Applied Biosystems model 120 A phenylthiohydantoin amino acid analyser. For internal amino acid sequence analysis, the protein was first digested with 2 µg of sequencing grade trypsin (Boehringer Mannheim GmbH, Germany) for 16 h at 37°C, and fragments were separated by a Vydac C18 reverse phase column. The separated fragments were analysed by mass spectrometry and protein sequencing.
We thank Professors T.J. Andrews (The Australian National University, Australia) and M.A. Parry (IARC-Rothamsted, UK) for generous gifts of Rubisco activase and Rubisco large subunit antibody, respectively. This work was supported by grants from The Academy of Finland, Nordiskt Kontaktorgan för Jordbruksforskning and the European Community grant (ERB IC15-CT98–0126).
EMBL/GenBank/DDBL database accession number J03610.