R. Scheibe, Plant Physiology, University of Osnabrueck, Osnabrueck D-49069, Germany.
Light/dark modulation of the higher plant Calvin-cycle enzymes phosphoribulokinase (PRK) and NADP-dependent glyceraldehyde 3-phosphate dehydrogenase (NADP- GAPDH-A2B2) involves changes of their aggregation state in addition to redox changes of regulatory cysteines. Here we demonstrate that plants possess two different complexes containing the inactive forms (a) of NADP-GAPDH and PRK and (b) of only NADP-GAPDH, respectively, in darkened chloroplasts. While the 550-kDa PRK/GAPDH/CP12 complex is dissociated and activated upon reduction alone, activation and dissociation of the 600-kDa A8B8 complex of NADP-GAPDH requires incubation with dithiothreitol and the effector 1,3-bisphosphoglycerate. In the light, PRK is therefore completely in its activated state under all conditions, even in low light, while GAPDH activation in the light is characterized by a two-step mechanism with 60–70% activation under most conditions in the light, and the activation of the remaining 30–40% occurring only when 1,3-bisphosphoglycerate levels are strongly increasing. In vitro studies with the purified components and coprecipitation experiments from fresh stroma using polyclonal antisera confirm the existence of these two aggregates. Isolated oxidized PRK alone does not reaggregate after it has been purified in its reduced form; only in the presence of both CP12 and purified NADP-GAPDH, some of the PRK reaggregates. Recombinant GapA/GapB constructs form the A8B8 complex immediately upon expression in E. coli.
Various chloroplast enzymes are subjected to light/dark modulation of their activity, brought about by a redox-modification at specific cysteine residues mediated by the ferredoxin/thioredoxin system . The activity of each of these enzymes is adjusted by fine-tuning, the rates of reduction and/or of oxidation being influenced by specific metabolites . At constant redox conditions, this allows for independent changes in the steady-state activities of each of the enzymes merely by changes in the metabolic state of the chloroplast [3,4]. In some cases, the reversible changes of redox and activation states are accompanied by oligomerization and re-dissociation of transient complexes.
Enzyme aggregations of various compositions have been described repeatedly, their occurrence under in vivo conditions still being under debate [5,6]. But even the actual composition of enzyme aggregates containing NAD(P)-dependent glyceraldehyde-3 P dehydrogenase [NAD(P)-GAPDH] and phosphoribulokinase (PRK) is controversial. Both activities appear to occur in high-molecular-mass forms in darkened chloroplasts, either in homo-oligomers [7–10] or in hetero-oligomers [11,12], the latter involving a small chloroplast protein, CP12, with a high sequence similarity to the C-terminus of subunit B of the unique chloroplast form of GAPDH . Here we describe the presence of both types of aggregations in darkened spinach chloroplasts, one consisting of GAPDH A and B alone, the other consisting of PRK, GAPDH, and CP12. The differential stability of both complexes upon reductive activation, the formation of the A8B8-GAPDH complex from the recombinant subunits in Escherichia coli, the reconstitution of the hetero-oligomeric PRK/GAPDH/CP12 complex from the isolated components in vitro, and coprecipitation from oxidized stroma using antisera against the individual components, support this fact.
Materials and methods
Isolation of intact chloroplasts and preparation of concentrated stroma
Darkened spinach leaves from plants in hydroponic culture were cut and homogenized in isotonic medium consisting of 330 mm mannitol, 30 mm Mops, 2 mm EDTA, pH 7.8, according to . The chloroplast pellet was washed twice and then treated on ice with a glass homogenizer using as little additional medium as possible. All steps were performed in darkness. After centrifugation, the cleared supernatant was filtered through a 0.2-µm filter. The protein content was determined according to  with BSA as a standard. The protein concentration was adjusted to 10 mg per 2 mL sample with column buffer (see gel filtration) either without (dark) or with 20 mm dithiothreitol, and incubated in a darkened vial for 30 min at 25 °C. For light-activation experiments, intact chloroplasts were isolated and incubated in bicarbonate-containing medium as described in .
Gel filtration on Superdex S-200
Preincubated samples were filtered through a calibrated Superdex 200 column (Hiload 16/60) (Pharmacia, Freiburg, Germany) at 1 mL·min−1, collecting fractions of 1 mL which were used for enzyme assays. The column buffer consisted of 10 mm bicine/KOH, 150 mm NaCl, 140 µm NAD, pH 7.8, for ‘dark’ samples, and additionally 2.5 mm dithiothreitol for ‘dithiothreitol’ samples.
Aliquot samples of the fractions were either assayed directly or after preincubation with 50 mm dithiothreitol (for PRK activation) or with 20 mm dithiothreitol and 21 µm 1,3-bisphosphoglycerate in a regenerating system as in  for GAPDH activation. The 1,3-bisphosphoglycerate-system as fivefold stock solution contained 100 mm Tris/HCl, pH 7.8, 8 mm MgSO4, 1 mm EDTA, 4 mm ATP, 9 mm 3-PGA and PGK (3.6 U·mL−1). The PRK assay was as in . NADP-dependent GAPDH activity was determined as in .
PRK was purified from spinach leaves according to  with some modifications. During all purification steps 10 mm dithiothreitol was present. The proteins precipitating between 40 and 55% of the saturation of ammonium sulfate at 4 °C were resuspended and subjected to acid precipitation with acetic acid at pH 5.0. The supernatant was adjusted to pH 6.8 and dialyzed over night. The diluted and clarified solution in 10 mm bicine/KOH, 10 mm dithiothreitol, pH 6.8, was subjected to affinity chromatography on Reactive Red (Merck, Darmstadt, Germany), washed with the same buffer. After a washing step with 10 mm potassium phosphate, 10 mm dithiothreitol, pH 6.9, the PRK was eluted in 10 mm potassium phosphate, pH 7.2, 10 mm dithiothreitol, and 5 mm ATP. The fractions with PRK activity were concentrated by ammonium sulfate precipitation (80% of saturation) and resuspended in 50 mm bicine/KOH, 10 mm potassium phosphate, 1 mm EDTA, 10 mm dithiothreitol, 10% glycerol, pH 8.0, and subjected to gel filtration in 100 mm bicine/KOH, 10 mm dithiothreitol, 1 mm EDTA, pH 8.0. Storage of the active dimeric protein was in 50% glycerol at −20 °C.
GADPH was purified from spinach leaves as described by , slightly modified as in . GapA and GapB were expressed as a combined construct in E. coli BL21DE3 pLys S containing the two clones each unter the control of the T7 promoter, and the kanamycin and the ampicillin resistance gene, respectively, in order to control the presence of both constructs.
CP12 from spinach was produced in E. coli with an N-terminal His-tag and was purified using Ni-chelating chromatography. Elution was achieved with 1 m imidazol, 0.5 m NaCl in 20 mm Tris/HCl, pH 7.9. For reconstitution experiments the His-tag was removed by proteolysis using thrombin according to the manufacturer's protocol (Strategene, Heidelberg, Germany). The CP12 clone for the mature protein was originally described in .
Artificial stroma conditions
Reconstitution experiments with the purified proteins were performed in a solution simulating the high protein conditions in the stromal sample using a modified ‘artificial stroma’. In more detail, the sample for dark conditions consisted of 5 mm MgSO4, 10 mm NaCl, 5 mm KNO3, 5 mm KH2PO4, 25 mm sucrose, 20 mm glucose, 5 mm fructose, 40 µm NAD, 20 mm GSSG, the purified proteins (each about 200 µg) and BSA (defatted) to reach a total protein content of 10 mg per 2 mL sample. Incubation was 30 min at room temperature and then at 4 °C over night. The purified GAPDH was pretreated with 100 µm 1,3-bisphosphoglycerate in a regenerating system and desalted in order to obtain the oxidized A2B2 form prior to the reconstitution assay.
SDS/PAGE, Western blotting and immunodecoration
Equal volumes of fractions after gel filtration were subjected to SDS/PAGE (12% acrylamide), and the gels were electroblotted and immunodecorated as in .
Protein-A Sepharose was preincubated with either preimmune serum or the indicated antisera in Tris-buffered saline for 30 min at room temperature, washed once and then added to stromal extracts that had been preincubated with 140 µm NAD and 20 mm GSSG. After incubation for 30 min, the supernatant was used to assay for GAPDH and PRK activities after full activation. The antisera against CP12 and NADP-GAPDH have been obtained from rabbits using the spinach proteins purified from E. coli and spinach, respectively. Antiserum against PRK was a kind gift from Fred Hartman, Oak Ridge, USA.
Reversible aggregation of GAPDH and PRK in chloroplasts
When the soluble fraction of isolated darkened spinach chloroplasts was subjected to gel filtration, both GAPDH and PRK were obtained as high-molecular-mass aggregates (Fig. 1A). Enzyme activities in the fractions were detected after full activation. The activity peaks do not coincide completely, GAPDH eluting somewhat earlier than PRK in all cases. This tendency was even more pronounced in maize chloroplasts (Fig. 2), where the GAPDH activity formed a distinct shoulder at 600 kDa in addition to the peak at 550 kDa that coincided with the PRK activity. The high-molecular-mass form of GAPDH is almost inactive due to its decreased affinity for 1,3-bisphosphoglycerate . PRK in darkened chloroplasts when eluted as the 550-kDa form is also inactive; its activity is only retrieved upon preincubation of the fractions with dithiothreitol (Fig. 3A). On the other hand, the dimeric PRK as obtained from illuminated chloroplasts is eluted as fully active enzyme (Fig. 3B). PRK purified in its reduced dimeric state  can be reversibly inactivated by treatment with oxidant , but remains dimeric (Fig. 3C).
Preincubation of spinach chloroplasts with 20 mm dithiothreitol for 30 min resulted in the complete disappearance of high-molecular-mass PRK and in the partial (≈ 50%) dissociation of the GAPDH aggregates (Fig. 1B). From experiments with the purified enzyme it is known that reductive treatment of the 600-kDa A8B8-GAPDH form does not lead to any increase in activity nor to dissociation . From the two-peak elution pattern of the GAPDH activity, we therefore assumed the presence of two types of GAPDH-containing aggregates, namely the A8B8 complex which was still intact after incubation with dithiothreitol alone , and the GADPH/PRK/CP12 complex described by  which, upon reduction, releases GAPDH and PRK as tetramer and dimer, respectively. CP12 and PRK could not be detected by the respective antisera in the high-molecular-mass fractions after dithiothreitol treatment, while all three proteins were detectable with antisera in the peak fractions of the untreated sample (Fig. 1C,D). The high-molecular-mass fraction did not contain other enzyme activities such as phosphoglycerate kinase  or fructose 1,6-bisphosphatase  that had been suggested to also form high-molecular-mass aggregates (data not shown). Furthermore, we never detected any tetramic GAPDH in dark stroma which has been suggested to occur as A4 by .
Differential activation behaviour for GAPDH and PRK upon dithiothreitol and light treatment
Incubation of chloroplast stroma with increasing concentrations of dithiothreitol resulted in 100% activation of PRK even at low dithiothreitol concentrations at pH 8.0 (Fig. 4A). In contrast, even up to 20 mm resulted in only 60–70% for the maximal GAPDH activity. Only in the presence of added ATP and/or 3-phosphoglycerate, thus increasing the 1,3-bisphosphoglycerate concentration in the stroma , 100% activation of GAPDH was reached. This is true also for activation by light, where 100% of PRK and 40–60% of GAPDH activity were reached already at very low light intensities. These levels were unchanged over a wide range of light intensities, only addition of ATP to the isolated chloroplasts increased the level of GAPDH activation to 100% (Fig. 4B). This is in agreement with the fact that both reduction and 1,3-bisphosphoglycerate are required for activation and dissociation of GAPDH from the A8B8 form [9,16].
Reaggregation of PRK and GAPDH from chloroplast fractions
Purification of PRK according to published procedures [3,17] is always performed in the presence of dithiothreitol, leading to the preparation of the enzyme in its reduced active form. Using this enzyme, the mechanism of reversible redox-modification mediated by thioredoxin has been analyzed in much detail. The redox-active Cys residues have been identified (Cys15 and Cys55) , their redox potential has been determined [3,26], and the interaction with thioredoxin f has been studied .
In order to analyze the structural changes upon redox modification, we have reoxidized the purified, reduced enzyme. Incubation with 50 mm dithiothreitolox or with 25 mm GSSG at pH 8.0 resulted in an almost complete inactivation. This inactivation could be reversed by the addition of reductant (Table 1). However, both the reduced and the oxidized enzyme forms appeared as dimers upon gel filtration (Table 1) (Fig. 3C). This is in contrast to the experiments with chloroplast stroma, where the oxidized dark form of PRK appears as high-molecular-mass form. Such controversial behaviour has been described already [8,21].
Table 1. Activity and aggregation state of purified spinach PRK. Reduced PRK was purified in the presence of 10 mm dithiothreitol. Oxidized PRK was treated with 25 mm GSSG or with 50 mm oxidized dithiothreitol, pH 8.0, for 15 min at 20 °C.
Activity (U·mg protein−1)
Molecular mass (kDa)
In order to investigate the requirement for a small stromal protein (i.e. CP12) for reaggregation of dimeric PRK, we separated the dissociated enzyme (Fraction II: 80–120 kDa in Fig. 5A) from the higher-molecular-mass fraction at 600 kDa (Fraction I) by gel filtration. Then a concentrated fraction containing the smaller stromal proteins (Fraction III: 20–60 kDa) and the enzyme fraction II were incubated together with GSSG. After another step of gel filtration, a new high-molecular-mass fraction became apparent containing PRK activity (after reductive activation of the fractions) (Fig. 5B). This peak contained almost equal activities of GAPDH and PRK.
Reconstitution of GAPDH (A8B8) and PRK/GAPDH/CP12 complexes
In a further approach, we attempted to reconstitute both complexes from the purified components. For these experiments, the purified proteins were kept in an artificial stroma according to  containing various ions, sugars, and amino acids (see Materials and methods) and in addition 140 µm NAD, 10 mm GSSG and BSA (20 mg·mL−1) in order to simulate the conditions of high protein concentration in the chloroplast stroma. Under these conditions, absolutely no aggregation was observed with PRK and CP12 alone (Fig. 6A). In contrast, incubation of PRK with purified GAPDH and CP12 resulted in the aggregation of some of the PRK and all of the GAPDH (Fig. 6B). The low yield of PRK reaggregation was probably due to the rather artificial conditions. Some of the aggregated GAPDH was most likely in its A8B8 form.
Finally, the expression of both GAPDH subunits in E. coli was performed. A high-molecular-mass complex formed immediately in E. coli upon expression of the combined construct for GapA and GapB and eluted as a 600-kDa form (Fig. 6C). The activation characteristics of the recombinant GAPDH are typical for the A8B8 form, which is still present in the 600-kDa peak in dithiothreitol-treated stroma (fraction I). In both cases, activation was obtained only after incubation with both dithiothreitol and 1,3-bisphosphoglycerate, not with dithiothreitol alone (Fig. 7).
Using the specific antisera against the components of the two complexes, namely GAPDH, PRK and CP12, GSSG-oxidized stromal extracts were fractionated. In order not to disturb the existing complexes, the immunoglobulin fraction was bound to immobilized Protein A and washed with 50 mm bicine/KOH, pH 8.0, before it was exposed to oxidized stroma containing 140 µm NAD which stabilizes the complexes and also the enzyme activities. The enzyme composition in the supernatant was quantified from the activities obtained after complete activation. The results are shown in Table 2. With antiserum against NADP-GAPDH about 82.5% of the PRK activity could be coprecipitated with 91.5% of the GAPDH activity, indicating that only 17.5% of the PRK was not associated with GAPDH in a complex. On the other hand, the antiserum against PRK removed only 42.9% of the GAPDH from the solution, together with 96.1% of the PRK. This again indicates the presence of an independent GAPDH-A8B8 complex apart from the PRK/GAPDH/CP12-mixed complex. The fact that the antiserum raised against recombinant CP12 only removed a small proportion of the enzymes from the solution, is most likely due to inaccessibility of the CP12 in the native complex, because the serum recognized native soluble CP12 in reduced stromal extracts (data not shown).
Table 2. Percentage coprecipitation of PRK-, GAPDH- and CP12-containing complexes from oxidized stroma. The activities remaining in the supernatant were determined after full activation.
In general, evidence is increasing that cellular contents are well organized in microcompartments due to protein–protein interactions between partners of a metabolic pathway or of a signal transduction cascade. In particular for chloroplasts, there have been various attempts to show the presence of bi- or multi-enzyme complexes of Calvin-cycle enzymes (reviewed in ); however, there are intrinsic technical problems when trying to confirm any of the interactions unambiguously. The critical step is always the breakage of the cell or the organelle, since changes of the protein concentration and of the low-molecular-mass components of the soluble medium will occur.
In order to avoid any new formation of complexes, ammonium sulfate precipitation has been omitted in our procedure and stromal fractions were applied directly to the gel filtration column. This leads to reproducable elution profiles on the Superdex 200 column in the presence of 150 mm NaCl, with two distinguishable peaks at 600 and 550 kDa eluting well after the void volume. It should be pointed out, however, that the protein concentration of the stroma sample for gel filtration is critical to obtain the described results .
Rubisco is eluted at the same position as PRK (550 kDa), but the subunit composition (L8S8) of the former suggests that further enzymes are not associated in the same complex, although this assumption is in contrast to the results of Rault et al.  who purified a five-enzyme complex of 540 kDa containing Rubisco, GAPDH, PRK, phosphoriboseisomerase, and 3-phosphoglycerate kinase. Other activities, such as FBPase that had been described to be subject to aggregation  or had been seen as part of the so-called photosynthetosome  could not be detected in this range. It cannot be excluded, however, that weaker interactions are involved in such supramolecular organization observed by other groups that are therefore not stable under the conditions applied in this study.
Previously we had identified and characterized the A8B8-GAPDH form present in darkened chloroplasts [9,16]. Later, we showed by limited proteolysis  and using truncated constructs expressed in E. coli that the unique C-terminal sequence extension of GapB is responsible for aggregation and inactivation of GAPDH in the dark. Likewise, expression of the complete subunit B in E. coli had led to aggregated forms [30,31]. Dissociation and activation of the A8B8 complex requires 1,3-bisphosphoglycerate [9,16,32].
Analyzing PRK activation, there was an obvious discrepancy between in vitro results with the purified enzyme and results obtained after gel filtration of chloroplast stroma, a fact also observed by . Therefore, it was assumed that PRK had lost a component enabling aggregation in vivo, and re-addition of a low-molecular-mass stromal fraction to the fraction of dissociated enzymes indeed allowed reaggregation (see Fig. 5).
In this paper, we have attempted to characterize such a PRK-containing complex, having in mind the postulated existence of a mixed GAPDH–PRK complex containing CP12 as described previously . Therefore it was necessary to separate the A8B8 form of GAPDH and such mixed complex. This could be achieved by various means due to their diverging properties upon activation and dissociation. In addition, reconstitution and coprecipitation experiments using specific antisera helped to confirm the presence of two different complexes. Unfortunately, the anti-CP12 serum was not able to reach this protein when part of the complex due to sterical hindrance. Taken together, it could be established that GAPDH activity is present in two different complexes reacting with distinct sensibility towards dithiothreitol and the effector 1,3-bisphosphoglycerate leading to dissociation and activation. On the other hand, all PRK activity was present in a mixed complex forming only in the presence of both GAPDH and CP12.
In order to understand the differential regulation of Calvin-cycle enzymes, their midpoint redox potentials have been determined and compared to physiological data obtained with the intact system . The apparent discrepancies with respect to GAPDH could be easily explained with the presence of two differently responding GAPDH-containing aggregates: (a) the early evolved system, namely PRK/GAPDH/CP12, which is activated merely by reduction and is already present in cyanobacteria ; and (b) the A8B8 form that evolved with the appearance of multicellular green organisms when GapB emerged in Characeae for the first time (R. Cerff and J. Peterson, Institut für Genetik, Tu Braunschweig, Germany, personal communication).
The occurrence of both aggregates in plants indicates their special role for optimal photosynthesis under all conditions. A fast-responding system for complete PRK activation and activation of a portion of GAPDH is required for CO2 assimilation under rather constant conditions, and is thus available in all photosynthetic organisms. Green plants with less constant environments (evolution of land plants) acquired in addition a larger fraction of GAPDH activity that, however, remains latent until a certain level of the substrate 1,3-bisphosphoglycerate indicates the demand to increase the flux specifically at this point. For the situation in higher plants this means that there is a two-level activation of GAPDH, reaching 50–60% of the total activity even under low light and low reductant concentrations and 100% activation only with at increased 1,3-bisphosphoglycerate levels. Full activation of PRK is achieved under all conditions (see Fig. 4). Calculating the stoichiometry of the enzyme subunits engaged in the two complexes using the specific activities of the purified enzymes (GAPDH 200, PRK 400 U·mg protein−1) and the activities present in the stroma (700 and 300 µmol·mg Chl−1·h−1). The ratio of 2 GAPDH/PRK/CP12 complexes to 1 A8B8 complex supports the finding of a ratio of 1 : 1 of the total GAPDH activity in both complexes.
From our experience, even very dim light during chloroplast isolation or preparation of extract leads to a significant level of PRK activity and to some basic GAPDH activity that is often described as ‘dark activity’. In contrast, strict darkness during these steps will result in complete inactivation. The residual GAPDH activity obtained in vitro is rather the result of the high 1,3-bisphosphoglycerate concentration (21 µm) in the standard assay leading to some activity of the inactivated enzyme (low affinity for substrate 1,3-bisphosphoglycerate) as opposed to the in vivo situation with a stromal concentration of 1–2 µm 1,3-bisphosphoglycerate [9,16].
Regulation of the activity at the PRK step is exclusively achieved by noncovalently acting inhibitors, with ribulose 1,5-P2 (Ki = 0.7 mm), 3-phosphoglycerate (Ki = 2 mm), inorganic phosphate (Ki = 4 mm) and ADP (Ki = 40 µm), to list only some of them , since even mild reduction results in complete dissociation of PRK from the GAPDH/PRK/CP12 complex and thus PRK activation.
The authors thank Dr Carsten Sanders, and Dr Simone Holtgrefe, Osnabrück, for initial experiments in the preparation of the clones and some chloroplast experiments. Thanks are also due to Dipl.-Biol. Elisabeth Baalmann and Dr Jan E. Backhausen, Osnabrück, for helpful discussion and to Prof Dr R. Cerff, Braunschweig, for initial encouragement and advice. Finally, the financial support given by the Deutsche Forschungsgemeinschaft to Renate Scheibe (Sche 217/8) is gratefully acknowledged.