CP12 provides a new mode of light regulation of Calvin cycle activity in higher plants

Abstract

CP12 is a small nuclear encoded chloroplast protein of higher plants, which was recently shown to interact with NAD(P)H–glyceraldehyde-3-phosphate dehydrogenase (GAPDH; EC 1.2.1.13), one of the key enzymes of the reductive pentosephosphate cycle (Calvin cycle). Screening of a pea cDNA library in the yeast two-hybrid system for proteins that interact with CP12, led to the identification of a second member of the Calvin cycle, phosphoribulokinase (PRK; EC 2.7.1.19), as a further specific binding partner for CP12. The exchange of cysteines for serines in CP12 demonstrate that interaction with PRK occurs at the N-terminal peptide loop of CP12. Size exclusion chromatography and immunoprecipitation assays reveal the existence of a stable 600-kDa PRK/CP12/GAPDH complex in the stroma of higher plant chloroplasts. Its stoichiometry is proposed to be of two N-terminally dimerized CP12 molecules, each carrying one PRK dimer on its N terminus and one A2B2 complex of GAPDH subunits on the C-terminal peptide loop. Incubation of the complex with NADP or NADPH, in contrast to NAD or NADH, causes its dissociation. Assays with the stromal 600-kDa fractions in the presence of the four different nicotinamide-adenine dinucleotides indicate that PRK activity depends on complex dissociation and might be further regulated by the accessible ratio of NADP/NADPH. From these results, we conclude that light regulation of the Calvin cycle in higher plants is not only via reductive activation of different proteins by the well-established ferredoxin/thioredoxin system, but in addition, by reversible dissociation of the PRK/CP12/GAPDH complex, mediated by NADP(H).

In recent years, the activities of five enzymes of the Calvin cycle have been shown to be modulated by light driven changes. For phosphoribulokinase (PRK or d-ribulose 5-phosphate kinase; EC 2.7.1.19), fructose bisphosphatase, sedoheptulose-1,7-bisphosphatase (SBPase), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), activation is achieved by reduction of intramolecular sulfhydryl groups via the light regulated ferredoxin/thioredoxin system. In contrast, ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) is activated by carbamylation of a specific lysine residue, mediated by Rubisco activase. Additional control of the activities of various enzymes of the cycle is further mediated by specific metabolites, acting either as inhibitor or stimulator on the neighboring enzymes (for reviews, see refs. 1–5). Another mode of regulating enzyme activity in the cycle was shown for GAPDH, which, depending on its substrate, 1,3-bisphosphoglycerate, or NADP(H) and ATP, can reversibly dissociate from the less active hexadecameric A8B8 form to the highly active A2B2 tetramer (6, 7). GAPDH was also shown to be associated with PRK in a 600-kDa protein complex, made up by the A2B2 and the A4 forms of GAPDH, together with one PRK dimer (8). Further data suggest that GAPDH and PRK could be associated in a complex together with Rubisco (9). These three enzymes together were then proposed to be members of a 600-kDa multienzyme complex, in addition of ribose-5-P isomerase, and 3-phosphoglycerate kinase (PGK) (10, 11). Another multienzyme complex thought to be associated with thylakoids, was proposed to contain SBPase instead of PGK and therefore organized in a stable 900-kDa protein complex, together with ferredoxin-NADP⁺ reductase (12). These multienzyme complexes of Calvin cycle proteins were proposed to have the functional advantage of allowing channelling of reaction intermediates from one enzyme to another, thereby avoiding free diffusion through the stroma.

CP12 is a small nuclear-encoded protein in the chloroplasts of higher plants, consisting of about 75 amino acid residues. Secondary structure predictions suggest that the mature protein contains a central a-helix, flanked by two conserved turns inducing epitopes, which are thought to form peptide loops via disulfide bonds between the flanking cysteine residues (see ref. 13 and Fig. 1A). Because the C-terminal peptide loop is almost identical with a short domain of the C-terminal extension of chloroplast GAPDH subunit GAPB, and this extension was shown to be involved in assembly of four A2B2 tetramers to the hexadecameric A8B8 regulatory form (14, 15), we proposed that the peptide loops of CP12 play a role in protein complex formation. Affinity chromatography of stromal protein extracts to overexpressed and immobilized mature CP12 has shown that CP12, indeed, specifically interacts with chloroplast GAPDH in vitro (13). Attempts to copurify CP12 with the regulatory A8B8 form of GAPDH, however, failed.

Figure 1.

Analysis of PRK/CP12 interaction in the yeast two-hybrid system. (A) Predicted tertiary structure for the mature CP12. Cysteine residues, proposed to form the N- and C-terminal peptide loops, are numbered. Charged amino acid residues, most probably involved in dimerization of CP12, are marked by circled + and − signs. (B) Expression of the GAL4-BD/CP12 fusion protein in yeast cells. Yeast cells, untransformed (lane 1) or transformed with the recombinant GAL4-BD/CP12 plasmid (lane 2), were selectively grown in a 2-ml liquid culture until saturation. Cells were broken by the glass bead method, and soluble proteins were subjected to SDS/PAGE. After electrophoresis, proteins were blotted onto nitrocellulose and immunodecorated with an antiserum directed against spinach CP12. (C) Verification and topologization of PRK/CP12 interaction. GAL4-AD/PRK plasmids, isolated by screening the pea two-hybrid library, were cotransformed into yeast cells together with GAL4-AD plasmids, expressing the indicated fusion proteins. Transfomation assays were plated onto agar, selecting for both plasmids. From each plate, two colonies were randomly picked and used to inocculate 2-ml liquid cultures. Cultures were grown overnight and equal aliquots, according to measured optical density, were blotted onto nitrocellulose using a Pharmacia slot blot apparatus. Cells were lysed by freezing the membrane in liquid nitrogen and protein–protein interactions were visualized by the colorimetric β-galactosidase assay.

Here we demonstrate that CP12 forms a stable 600-kDa complex with GAPDH and PRK in spinach chloroplasts. In vitro, this complex can be dissociated by NADP and NADPH. Whereas PRK activity of the dissociated complex is shown to be inhibited by NADP, it is significantly stimulated in the presence of NADPH. A model for the topology and the various activation states of PRK and GAPDH of the PRK/CP12/GAPDH complex is presented.

MATERIALS AND METHODS

Materials.

Pea seedlings (Pisum sativum var. Golf) were grown in a greenhouse under 14 h/10 h light/dark regime. Spinach (Spinacia oleracea var. Matador) was grown outdoors.

cDNA-Screening in the Yeast-Two-Hybrid System.

Poly(A)⁺ RNA was prepared from primary leaves of young pea seedlings as outlined in ref. 13. cDNA synthesis and HybriZAPTM two-hybrid library construction were done by Stratagene. Appropriate restriction sites for cloning the cDNA fragments, encoding the precursor and mature forms of pea CP12 in-frame with the DNA-binding domain of GAL4, were introduced into the original cDNA (GenBank accession no. Z72489) by PCR amplification. PCR products were ligated into the pBD–GAL4 Cam phagemid vector and cloned in Escherichia coli strain DH5a. Inserts of both plasmids were reconfirmed by complete DNA sequencing. Recombinant plasmids, pAD–GAL4, containing the pea cDNA library, and pBD–GAL4, containing the bait cDNAs for either precursor or mature CP12, were prepared according to ref. 16. Cotransformation of yeast competent cells, library screening and verification of protein–protein interactions were carried out according to the manufacturer’s protocols.

In Vitro Mutagenesis.

Cysteine for serine exchanges in precursor and mature CP12/GAL4 fusion proteins were achieved by recombinant PCR technique using mutated central and restriction sites added peripheral DNA oligonucleotides. All PCR products were sequenced completely prior to use in the yeast two-hybrid system.

Preparation of Chloroplast Extracts.

About 100 g of spinach leaves were homogenized in a Waring blendor with 250 ml 330 mM sorbitol, 20 mM Mops, 13 mM Tris, 0.1% BSA (wt/vol), and 1 mM MgCl₂. The homogenate was passed through two layers of Miracloth and then centrifuged at 5000 rpm for 30 sec in a Beckman GSA rotor. The sedimented chloroplasts were washed once with 200 ml 330 mM sorbitol, 50 mM Hepes⋅KOH (pH 7.6), and 1 mM MgCl₂ and resuspended in 2 ml 100 mM Tris⋅HCl (pH 8.0) and 1 mM EDTA. Chlorophyll content was estimated by the method of Arnon (17). Chloroplasts were then homogenized and separated into soluble and membrane fractions by three rounds of centrifugation at 15,000 × g for 15 min. The resulting supernatant solution (stroma fraction) was aliquoted according to 2 mg chlorophyll, frozen in liquid nitrogen, and stored at −80°C.

Size Exclusion Chromatography.

Chromatography of chloroplast stroma extracts was performed either on a TSK-G3000SW column (0.75 × 60 cm; Tosohaas, Stuttgart) or on a Sephacryl-S400 column (1.6 × 60 cm; Pharmacia), each preequilibrated with 100 mM Tris⋅HCl (pH 7.0), 75 mM KCl, 1 mM EDTA, and 10% (vol/vol) glycerol using a Milton Roy model HPLC system at 4°C. Prior to analysis, protein solutions were brought to column buffer conditions and fractions of 0.5 or 2.5 ml, respectively, were collected at a flow rate of 0.5 ml/min. For calibration, blue dextran (2,000 kDa), thyroglobulin (670 kDa), ferritin (230 kDa), aldolase (160 kDa), and egg albumin (45 kDa) were passed through both columns.

For stromal protein complex analysis, 1 ml of the adjusted stroma fractions (according to 4 mg chlorophyll) was loaded onto the column. Aliquots (0.5 ml) of the collected fractions were adjusted to 10% (wt/vol) trichloroacetic acid and kept on ice for 30 min. Precipitation assays were then centrifuged at 15,000 × g for 30 min, and the sedimented proteins were resolubilized in appropriate buffer for SDS/PAGE (18). After gel electrophoresis, proteins were blotted onto nitrocellulose and immunodecorated with monospecific antibodies (19). For that purpose, membranes were (after protein transfer) each cut into an upper and a lower half segment and separately incubated with antibodies against PRK (upper half) and CP12 (lower half). After incubation with anti-rabbit IgG or anti-chicken IgY-AP conjugates (Sigma) and color development the upper halfs were then, in a second assay, incubated with antibodies against chloroplast GAPDH.

To analyze the effect of nicotinamide-adenine dinucleotides on the stability of the stromal PRK/CP12/GAPDH complex, the remaining 2 ml of the concerning Sephacryl-S400 fractions (fractions 27 and 31) were pooled and concentrated in an ultrafiltration cell using a YM30 DIAFLO membrane (Amicon). Concentrated fractions were brought up to 1 ml with column buffer, incubated in the presence of 2.5 mM of the indicated nicotinamide-adenine dinucleotides for 1 h at 4°C and subsequently rechromatographed. Proteins of 1 ml of the relevant collected fractions were precipitated with 10% trichloroacetic acid and one-third of the resolubilized proteins were subjected to SDS/PAGE and Western blot analysis.

Immunoprecipitation.

Assays (1 ml) were carried out at room temperature either with stroma or pooled column fractions containing the 600-kDa PRK/CP12/GAPDH complex. After 3 h of incubation with a monospecific antiserum against chloroplast GAPDH, 50 μl protein A-agarose (Boehringer Mannheim) was added and incubation was continued for further 40 min. After centrifugation, the sediments were washed five times with 1 ml 100 mM Tris⋅HCl (pH 7.5) and 1 mM EDTA and finally solubilized in 60 μl of loading buffer for SDS/PAGE. Probes were boiled for 3 min and centrifuged, and the resulting supernatant solutions were loaded onto the gel.

PRK Activity Assays and Thin-Layer Chromatography (TLC).

PRK activity of the pooled column fractions containing the 600-kDa PRK/CP12/GAPDH complex was measured according to ref. 20 with the following modifications. Assay buffer was 50 mM Tris⋅HCl (pH 8.0), 0.5 mM EDTA, 5 mM MgCl₂, 1.4 mM 2-mercaptoethanol, 1 mM ATP, 1 mM d-ribulose 5-phosphate (Ru-5-P), and NAD(P)(H) as indicated. Prior to the activity assays, aliquots of the column fractions were preincubated with the indicated nicotinamide-adenine dinucleotides (2.5 mM) (see Fig. 4A) and from 5 mM to 0.5 mM and vice versa (see Fig. 4B) for 10 min at 15°C. PRK activity assays were then carried out in the presence of 0.2 μCi [γ-³²P]ATP (1 Ci = 37 GBq) for 20 min (Fig. 4A) or 10 min (Fig. 4B) at 15°C. Reactions were stopped by adding 10% acetone on ice.

Figure 4.

NAD(P)(H)-dependent formation of Ru-1,5-bP by PRK of the PRK/CP12/GAPDH complex. Spinach chloroplast stroma was passed through a TSK-G3000SW size exclusion chromatography column and pooled 600-kDa complex fractions 13 and 14 were preincubated in column buffer containing 2.5 mM of the indicated nicotinamide-adenine dinucleotides for 10 min at 15°C. Then buffer was brought up to assay conditions and incubation was continued in the presence of 0.2 μCi [γ-³²P]ATP for further 20 min in A or 10 min in B. Reactions were stopped with addition of 10% acetone on ice. Aliquots of the assays were mixed with 20 μg Ru-1,5-bP and chromatographed on cellulose thin layers. Plates were developed overnight, stained with Hanes’ reagent, and then exposed to x-ray films, as shown in A and B.

Conditions for TLC of sugar phosphates were taken from ref. 21. Aliquots (2.5 μl) of each activity assay were mixed with 2 μl (10 μg/μl) d-ribulose-1,5-bisphosphate (Ru-1,5-bP) (Sigma) and were then applied onto the thin layer (Polgram CEL 400; Macherey-Nagel, Düren). As a control, 20 μg of Ru-5-P in 4.5 μl assay buffer were spotted onto the cellulose plates. Chromatography was carried out over night in a tank containing t-amyl alcohol/water/p-toluene-sulfonic acid (65/33/2 upper phase, vol/vol/wt) as solvent system. For colorimetric detection of sugar phosphates, the developed plates were dried, sprayed with Hanes’ reagent, and exposed to UV light (366 nm) for 7 min (22). Radiolabeled sugar phosphates were detected after 1 h exposure of the plates to x-ray films at −80°C (with screen).

Miscellaneous Procedures.

DNA cloning, sequencing, and PCR techniques followed standard procedures (23, 24). Protein overexpression in E. coli, as well as preparation and purification of antibodies against CP12, were as outlined in ref. 13.

RESULTS

PRK Binds to the N-Terminal Peptide Loop of CP12.

The yeast two-hybrid system was used to identify additional proteins that interact with CP12 (Fig. 1A). For that purpose, cDNA fragments that encode the mature and the precursor form of pea CP12 were cloned in frame with the GAL4–DNA-binding domain (GAL4-BD). Prior to library screening, expression of the GAL4-BD/CP12 fusion protein in yeast cells was verified by protein extraction and Western blot analysis with antibodies against CP12 (Fig. 1B). Recombinant plasmids, pGAL4-BD/pCP12 and pGAL4-BD/CP12, were each cotransformed into competent yeast cells with those that contain the GAL4 activation domain (GAL4-AD), combined with a complete cDNA library, made from RNA of young pea leaves. Transformants were plated onto selective agar and analyzed by the colorimetric 5-bromo-4-chloro-3-indolyl β-d-galactoside (X-Gal) assay. After the appropriate control transformations of the plasmids of five initially positive colonies, two cDNA clones remained, both coding for PRK. The longer one is made up of 1,313 bp and encodes the complete mature protein with one additional amino acid residue of the C-terminal end of the transit peptide (GenBank accession no. Y11248). Molecular interaction between CP12 and PRK was further verified by screening the pea cDNA library in the yeast two-hybrid system with the isolated PRK cDNA, cloned into the bait plasmid. This reciprocal screening resulted in seven primarily positive clones, of which six encode CP12-sequences of different length. The seventh is most probably a “false positive” (data not shown).

Analysis of the redox state of the thiol groups of the cysteine residues of purified recombinant CP12, using the DIG Protein Detection Kit (Boehringer Mannheim), has revealed that all the cysteine residues of CP12 are oxidized to disulfide bonds (data not shown) and therefore may form the predicted peptide loops as given in Fig. 1A. To prove whether formation of peptide loops in CP12 is necessary for the interaction with PRK, an in vitro mutagenesis approach was used in combination with the yeast two-hybrid system. Three different CP12 mutants were synthesized by recombinant PCR technique. The first mutation causes an exchange of Cys-1 to serine and thereby prevents the formation of the N-terminal peptide loop of CP12 (see Fig. 1A). The second mutation changed Cys-4 to serine and thereby opens the C-terminal peptide loop. The third mutated CP12 contains both cysteine to serine exchanges, Cys-1 and Cys-4 to serine, and is therefore unable to form any of the two peptide loops. Correct DNA sequences of the CP12 mutants were verified by complete cDNA sequencing. Yeast cells were cotransformed with plasmids that carry either the GAL4-AD/PRK fusion or one of the mutagenized GAL4-BD/CP12 fusions and plated onto agar, selecting for the presence of both plasmids. Of each cotransformation assay, two colonies were randomly picked and used to inoculate 3-ml liquid cultures. Equal aliquots of each culture were slot blotted onto nitrocellulose and tested for protein–protein interaction using the X-Gal assay. As shown in Fig. 1C, positive protein–protein interactions only are observed when the formation of the N-terminal peptide loop of CP12 via the disulfide bond between Cys-1 and Cys-2 is allowed. Destroying the C-terminal peptide loop of CP12 by changing Cys-4 to Ser has no effect on the CP12/PRK interaction. These results clearly demonstrate that the N-terminal peptide loop of CP12 is essential for the molecular interaction between CP12 and PRK.

CP12 Forms a Stable Complex Together with PRK and GAPDH.

To verify that CP12 and PRK indeed interact in higher plant chloroplasts, chloroplast stroma extracts were prepared from spinach leaves and subjected to size exclusion chromatography on a TSK-G3000SW column. Aliquots of the collected fractions were then used for SDS/PAGE and immunoblot analysis with monospecific antisera against PRK, GAPDH, and CP12. Fig. 2A demonstrates that CP12 cofractionates together with PRK and GAPDH at about 600 kDa (fractions 14 and 15) with some uncomplexed proteins being eluted at a later stage (apparent fractions 23 and 24). Silver staining of the same column fractions demonstrates that PRK, GAPDH (GAPA and GAPB), and CP12 are the only proteins having joint maxima in the 600-kDa fractions (fractions 14 and 15), whereas other proteins, represented by two close bands at about 60 kDa and Rubisco (LSU and SSU), are most prominent in fractions 13 and 15–17, respectively (Fig. 2B).

Figure 2.

Identification and characterization of the 600-kDa PRK/CP12/GAPDH complex in spinach chloroplasts. Soluble proteins of isolated chloroplasts, with 4 mg chlorophyll, were passed through a TSK-G3000SW size exclusion column. Fractions were concentrated by trichloroacetic acid precipitation and after SDS/PAGE were analyzed either by immunoblot and subsequent immunodecoration with monospecific antibodies against the indicated proteins (A), or by silver staining of the gel (B). Fraction numbers of the column run are indicated above the lanes of the blot (A) and the gel (B). In C, column fractions 13 and 14 were pooled and subjected to an immunoprecipitation assay using the GAPDH antiserum. Precipitated proteins were separated by SDS/PAGE and analyzed by immunoblotting, using antisera against the indicated chloroplast proteins. Immunochemical reactions were visualized by the colorimetric assay.

Because complex fractions 13 and 14 might also contain hexadecameric GAPDH-A8B8 complexes, immunoprecipitation assays were performed to analyze whether PRK and GAPDH are both complexed by CP12. After SDS/PAGE and transfer onto nitrocellulose, the coprecipitated proteins were identified by immunodecoration with antisera against CP12, GAPDH, and PRK. Fig. 2C shows that all the three proteins were coprecipitated by the GAPDH antiserum, indicating that each CP12 molecule carries indeed both of the two enzymes, PRK and GAPDH. Approaches with rabbit preimmune serum failed to precipitate any of the three proteins (data not shown).

NADP(H)-Dependent Dissociation of the PRK/CP12/GAPDH Complex.

To get an idea about the physiological function of the identified PRK/CP12/GAPDH complex in higher plant chloroplasts, the effect of different enzymatic cofactors on the stability of the complex was tested. For this purpose, spinach chloroplast stroma was passed over a Sephacryl-S400 size exclusion column. Fractions 27–31 containing the peak of the 600-kDa complex were pooled, concentrated by ultrafiltration, and incubated with either water (as a control) or nicotinamide-adenine dinucleotides (2.5 mM, 1 h at 4°C) as indicated and subsequently rechromatographed on the same column. After SDS/PAGE of the collected fractions and protein transfer onto nitrocellulose, the distribution of the complex proteins was analyzed by immunodecoration with antisera against PRK, GAPDH, and CP12. Although immunoblot analysis of the rechromatographed water control showed partial dissociation of the complex, incubation with the different nicotinamide-adenine dinucleotides resulted in specific, reproducible distribution patterns (Fig. 3). Whereas incubation of the complex with NAD (Fig. 3A) and NADH (Fig. 3C) exhibited no effect on complex stability, incubation with NADP (Fig. 3B) as well as with NADPH (Fig. 3D) caused almost complete dissociation of the PRK/CP12/GAPDH complex. The apparent molecular weights of the dissociated components were estimated to 160 kDa (GAPDH), 90 kDa (PRK), and 70 kDa (CP12). Complete dissociation of the complex was also achieved by incubation with 50 mM DTT for 1 h at 30°C (conditions as in ref. 10), most probably due to destruction of the peptide loops of CP12 by reduction of its disulfide bonds (data not shown). Incubation with 2.5 mM ATP or ADP resulted in the same pattern, as in the water control and therefore they appear to be not the effectors for complex dissociation (data not shown). The results presented in Fig. 3 further demonstrate that the molecular association of PRK and GAPDH with CP12 is not due to covalent interaction, but most probably to conformational and electrostatical protein–protein interactions.

Figure 3.

NADP(H)-dependent dissociation of the 600-kDa PRK/CP12/GAPDH complex. Spinach chloroplast stroma was passed through a Sephacryl-S400 size exclusion chromatography column. Peak fractions for the 600-kDa complex (fractions 27–31) were pooled and concentrated by ultrafiltration. Aliquots of the concentrated protein solution were then incubated for 1 h at 4°C either with 2.5 mM NAD (A), NADP (B), NADH (C), or NADPH (D) and subsequently rechromatographed on the same column. Proteins were precipitated, separated by SDS/PAGE, transferred onto nitrocellulose and immunochemically analyzed with antisera against the indicated proteins. (E) As a reference for the uncomplexed CP12, in E. coli overexpressed mature CP12 was run through the Sephacryl-S400 column and immunochemically detected as above. Blot lanes are numbered according to the collected column fractions. Lane S, stromal proteins before purification. Fractions containing the highest PRK and CP12 concentrations are marked at the top and the bottom of the blots, respectively.

NADP Inhibits and NADPH Stimulates PRK Activity in Vitro.

Assays were performed to analyze PRK activity of the pooled 600-kDa complex fractions 13 and 14 from the TSK-GS3000SW column. Because NADP and NADPH were both able to dissociate the PRK/CP12/GAPDH complex in vitro, we investigated, whether the four different nicotinamide-adenine dinucleotides also have a direct effect on the enzymatic activity of PRK (Fig. 4A). The results demonstrate that there is little Ru-1,5-bP formation in the water control, as well as in the presence of NADH, most probably due to weak dissociation of the complex caused by dilution in the assay mixture, whereas the other nondissociating dinucleotide, NAD, inhibits Ru-1,5-6P formation. Whether this is due to a complex stabilizing effect of NAD, as demonstrated for the high molecular weight GAPDH complex (7) remains to be shown. The complex dissociating reagents, NADP and NADPH, exhibited a contrary effect on PRK activity. In contrast to NADP, which inhibited Ru-1,5-bP formation in the assay, NADPH caused a strong increase in the production of radiolabeled Ru-1,5-bP.

Further analysis of the pooled complex fractions for PRK activity, testing both components, NADP and NADPH, together in one assay but in different countercurrent ratios (Fig. 4B), shows that PRK is significantly activated by a ratio of 1:3 (NADP/NADPH). Decreasing the ratio further to 1:10 (NADP/NADPH) results in a more pronounced increase of activity. Because both NADP and NADPH have been shown to be able to dissociate the PRK/CP12/GAPDH complex (Fig. 3), the enzymatic activity of PRK seems to be further regulated by the light balanced ratio of accessible NADP/NADPH.

DISCUSSION

One strategy to estimate the physiological function of CP12 in higher plant chloroplasts was to look for proteins that can interact with it. Previously we have shown by affinity chromatography that CP12 interacts with chloroplast GAPDH in vitro (13). Later incubations of the nitrocellulose membranes of this experimental series with antibodies against PRK have shown that PRK has the same behavior of elution from the CP12 affinity columns, as has GAPDH. To analyze protein–protein interactions of CP12 in vivo, we decided to screen a pea cDNA library in the yeast two-hybrid system using CP12 as a bait protein. After control transformations of five originally isolated plasmids, two clones remained positive, both containing cDNAs for PRK. None of the isolated cDNAs encoded GAPA or GAPB sequences, indicating that CP12 can interact only with the heterotetrameric A2B2 form of GAPDH. This would be in good agreement with the inability of GAPA homotetramers to assemble into higher molecular weight aggregates. Detection of interaction between CP12 and the GAPDH heterotetramer, A2B2, seems unlikely because this would presume stable cotransformation of pGAL4-AD/GAPA and pGAL4-AD/GAPB out of the library with pGAL4-BD/CP12 into one yeast cell. On the other hand, interaction of CP12 with GAPDH homotetramers might be prevented by the GAL4 fusion domains, which may have hindered correct assembling of the GAPDH subunits to tetramers, or interaction of them with CP12 in the two-hybrid system. N-terminal GAL4 fusions might also have prevented the detection of CP12/CP12 interaction in this system. An additional, reciprocal screening of the library with the isolated PRK cDNA, as the bait protein, led to the isolation of seven primarily positive clones, of which six contain cDNAs, coding for CP12, sustaining the specificity of PRK/CP12 interaction.

Further analysis of protein–protein interactions between PRK and mutagenized CP12 proteins, which were unable to form either the N-terminal or the C-terminal peptide loop, has shown that PRK binds to the N-terminal peptide loop of CP12 (Fig. 1C). As a result of what is mentioned above and also the high degree of identity of the C-terminal loop of CP12 to the complex formation mediating C-terminal extension of GAPB (13–15), we propose that interaction between CP12 and GAPDH occurs at the C-terminal peptide loop of CP12. This suggestion is further supported by coimmunoprecipitation of all the three proteins, using a monospecific antiserum against only one (GAPDH) of them, showing that in this complex, each CP12 molecule is assembled with both of the other enzymes, PRK and GAPDH (Fig. 2C). Size exclusion chromatography of stromal protein extracts shows that PRK, GAPDH (GAPA and GAPB), and CP12 cofractionate at ≈600 kDa (Fig. 2A). Silver staining of the pherographed column fractions (Fig. 2B) demonstrates that, although the complex peak fractions 14, and 15 are contaminated with other proteins, especially CPN60, Rubisco, and most probably the A8B8/GAPDH complex, there are no other proteins visible, which have their maximum in these two fractions. Therefore, we propose that only PRK, GAPDH, and CP12 form this 600-kDa complex by dimerization of two CP12 molecules (24 kDa), each loaded with one PRK dimer (90 kDa) on its N-terminal and one GAPDH heterotetramer (160 kDa) on its C-terminal peptide loop. Dimerization of CP12 might be achieved via electrostatical interactions between charged amino acid residues located in its N-terminal domain in front of the first peptide loop (Fig. 1A). Primary structure analysis of this domain from CP12 proteins of three different plant species have revealed that each N terminus contains a cluster of at least two positively and two negatively charged residues, able to form salt bridges, when a second N terminus is placed in an antiparallel position. Another hint for CP12 dimerization is given by the unusual high molecular weight fractionation (60–70 kDa) of uncomplexed CP12 during size exclusion experiments (Figs. 2 and 3), in part, most probably also reflecting the stretched tertiary structure of the oligomerized protein.

Powls and coworkers have described 560-kDa protein complexes in the chloroplast stroma of the green alga, Scenedesmus obliquus (25, 26) and spinach (8), both having latent PRK and GAPDH activities. Incubation of these complexes with DTT and NADPH induced depolymerization and marked stimulation of the latent enzyme activities. They have also shown that, if the alga is grown heterotrophically in the dark, all PRK was assembled with GAPDH. In contrast, the complex was dissociated during photoheterotrophic growth in the light, implicating that DTT and NADPH in vitro may simulate light in vivo. To investigate the effect of nicotinamideadenine dinocleotides and other potential dissociating reagents on the stability of our PRK/CP12/GAPDH complex, we have incubated the pooled complex fractions with either 2.5 mM NAD, NADH, NADP, and NADPH (Fig. 3), as well as ATP and ADP, for 1 h at 4°C, or with 50 mM DTT for 1 h at 30°C (data not shown), and rechromatographed them on the same column. In contrast to NAD, NADH, ATP, and ADP, which showed no effect compared with the water control, incubation with NADP, NADPH, and DTT caused almost complete dissociation of the complex. Further investigation of the effects of the different dinucleotides on PRK activity of our 600-kDa complex (Fig. 4A) demonstrates that only NADPH can stimulate Ru-1,5-bP formation, whereas NAD and NADP inhibit PRK activity. These results, taken together with the data presented and discussed above, suggest that the spinach PRK/GAPDH complex, published by Powls and coworkers (8), is identical with our PRK/CP12/GAPDH complex. Because CP12 is not simply detectable by gel electrophoresis of chloroplast extracts and protein staining, and nothing was known about the existence of CP12 at the time they isolated the complex, they might have failed to notice it.

Ricard and coworkers (10) have described a functional five-enzyme complex also from spinach chloroplasts, which was proposed to consist of ribose-5-P isomerase, PRK, Rubisco, 3-phosphoglycerate kinase, and GAPDH. The PRK activity of the isolated complex was shown to increase 6- to 7-fold by incubation with 50 mM DTT for 1 h at 30°C, whereas the activity of the free PRK in the lower molecular weight fractions could only be increased 0.25-fold. They further demonstrated that this treatment leads to total dissociation, either of PRK activity from the complex or of the complete complex. We conclude that activation of PRK in their assays is due to complex dissociation and is not, or only to a minor degree, due to to reduction of sulfhydryl groups in PRK. For the reasons described above, we propose that also in this complex PRK and GAPDH were assembled with CP12 and that the nonphysiological DTT treatment led to the reduction of the peptide loop forming disulfide bonds in CP12 and thereby to dissociation of the PRK/CP12/GAPDH complex. In refs. 27 and 28, the authors have shown that PRK is much more rapidly activated by reduced thioredoxin when assembled in the complex, than as the free enzyme. In addition, they have demonstrated that, in the absence of reduced thioredoxin, the free enzyme spontaneously becomes oxidized, whereas the complexed PRK remains stable. From these and our results we conclude that PRK activity in higher plants is regulated by two successive light-dependent mechanisms. First, a change in activation state of the complexed PRK (and most probably GAPDH) is achieved by reduced thioredoxin f. The second step is the NADP(H) mediated dissociation of the PRK/CP12/GAPDH complex, producing free dimeric PRK and tetrameric GAPDH. The enzymatic activity of PRK seems then to be further regulated by the ratio of accessible NADPH to NADP. In the dark, both enzymes may become spontaneously oxidized and reassemble into the inactive 600-kDa PRK/CP12/GAPDH complex (Fig. 5). We propose that the molecular (co)evolution of CP12 and the homologeous C-terminal extension of GAPB (13, 29) has enabled higher plants to directly link PRK and GAPDH activity in the stroma with the light-driven electron flux in the thylakoids not only via reduced thioredoxin f, but also via reversible protein complex dissociation, mediated by the other reductive product, NADPH.

Figure 5.

Model for the topology and the various activation states of PRK and GAPDH of the 600-kDa PRK/CP12/GAPDH complex (see text).

ABBREVIATIONS

PRK

phosphoribulokinase

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

Rubisco

ribulose-1,5-bisphosphate carboxylase/oxygenase

GAL4-BD

GAL4–DNA-binding domain

GAL4-AD

GAL4 activation domain

X-Gal

5-bromo-4-chloro-3-indolyl β-d-galactoside

Ru-1

5-bP, d-ribulose-1,5-bisphosphate

GAPA and GAPB

subunits A and B of GAPDH