Eur. J. Biochem. 173,437-443 (1988) 0 FEBS 1988

A functional five-enzyme complex of chloroplasts involved in the Calvin cycle Brigitte GONTERO ’, Maria Luz CARDENAS’ and Jacques RICARD’

Centre de Biochimie et de Biologie Moleculaire du Centre National de la Recherche Scientifique, Marseille Departamento de Biologia, Facultad de Ciencias, Universidad de Chile, Santiago

(Received October 15, 1987/January 4, 1988) - EJB 87 11 55

A complex of five different enzymes: ribose-phosphate isomerase, phosphoribulokinase, ribulose-bisphosphate carboxylase/oxygenase, phosphoglycerate kinase and glyceraldehyde-phosphate dehydrogenase, has been purified from spinach chloroplasts. These enzymes catalyse five consecutive reactions of the Calvin cycle, of which the two reactions catalysed by phosphoribulokinase and ribulose-bisphosphate carboxylase/oxygenase are unique to this cycle. The five-enzyme complex has been purified by successive chromatographies on DEAE-Trisacryl, Sephadex G-200 and hydroxyapatite. The homogeneity of the complex has been tested by analytical centrifugation and electrophoresis. Depending on the technique used to estimate its molecular mass, the value obtained varies between 520 kDa and 536 kDa. In addition to the five enzymes mentioned above, the complex contains a 65-kDa polypeptide. The quaternary structure of these enzymes in the complex appears to be different from what has been described for the individual enzymes in the ‘noncomplexed state’.

The five-enzyme complex is functional as glyceraldehyde 3-phosphate is formed from ribose 5-phosphate. Preliminary experiments suggest that channelling of reaction intermediates occurs within the complex.

It has been shown in recent years that, besides tightly associated enzymes forming stable complexes, there exist loose or transitory associations of enzymes forming unstable complexes [l, 21. These loose associations have been described both for proteins of the cytosol[3] and for those of mitochon- dria [4]. Stable or unstable enzyme complexes may represent a functional advantage for the living cell, as compart- mentalization may allow channelling of an intermediate from one active site to another within the same enzyme complex [5, 61. If these associations are loose they may obviously be disrupted in the course of the purification procedure, thus making it difficult to identify the multienzyme complexes. Thus only recently has a large multienzyme complex, the so- called ‘Krebs cycle metabolon’ [4], been shown to occur in mitochondria. No similar complex has ever been reported in the case of enzymes belonging to the Calvin cycle, although some recent reports suggest the possibility of an association between certain enzymes of this cycle [7, 81.

We report in this paper that five enzymes, ribose-phos- phate isomerase, phosphoribulokinase, ribulose-bisphos- phate carboxylase/oxygenase, phosphoglycerate kinase, glyceraldehyde-3-phosphate dehydrogenase, which catalyse five consecutive reactions of the Calvin cycle, exist as a com- plex that may be isolated and purified, and which is functionally active. We also discuss some possible biological implications of the existence of this complex.

Correspondence to J. Ricard, Centre de Biochimie et de Biologie MolCculaire, BP 71, F-13402 Marseille Cedex 9, France

Enzymes. Ribose-phosphate isomerase (EC 5.3.1.6); phospho- ribulokinase (EC 2.7.1.19); ribulose-bisphosphate carboxylase/ oxygenase (EC 4.1.1.39); phosphoglycerate kinase (EC 2.7.2.3); glyceraldehyde-3-phosphate dehydrogenase (EC 1.2.1.13).

MATERIALS AND METHODS

M a ter ials

Spinach plants (Spinacea oleracea var. Giant d’Hiver) were grown in a research garden in Aix-Marseille I1 Univer- sity. Ribulose Sphosphate, ribulose 1,5-bisphosphate and yeast ribose-phosphate isomerase were from Sigma, bovine serum albumin from Calbiochem and the other reagents from Boehringer. Trisacryl was obtained from IBF Biotechnics (France), Sephadex G-200 and Sephacryl S-300 from Pharmacia. Deionized water, further purified with a Millipore super Q system, was used throughout.

Purification of the multienzyme complex

Intact spinach chloroplasts were isolated by the method of Jensen and Bassham [9]. They were suspended in a buffer (15 mM Tris/HCl, 4 mM EDTA, 10 mM cysteine, pH 7.9) called buffer A, which brings about lysis of these organelles. The whole extract was then centrifuged for 60min at 48000 x g to remove membranes. The complex was isolated and purified from the supernatant by successive separations on DEAE-Trisacryl, Sephadex G-200 and hydroxyapatite. The chloroplast supernatant was applied first on a DEAE- Trisacryl column (2.5 x 13 cm) equilibrated in buffer A and washed with buffer A containing 50 mM NaC1. Elution of the column was performed with a gradient of 50 - 300 mM NaCl in buffer A. Fractions containing phosphoribulokinase ac- tivity were pooled, concentrated and submitted to molecular sieve chromatography on a Sephadex G-200 (column 2.6 x 95 cm) using another buffer, B, having the following composition: 30 mM Tris/HCl, 1 mM EDTA, 100 mM NaCl, 10 mM cysteine, pH 7.9. The fractions located in the void volume contained ribose-phosphate isomerase, phosphori-

438

8 EX6 .--I

< 4 3

2

0

1 2 3 0. e

0. 6 - 11% E \ 3

0. 4

0 . 2

0 100 150 200 250 300 350

Elution volume ( m l )

B 0. 25 , 1

0. 20

0. 15 r(

E 1 2 0. 10

0. 05

0 0 200 400 600

Elution volume I m l l Fig. 1. Cvpurificution ofu high-moleculur-mussphosphoribulokinuse and other enzymes,from the Calvin cycle. Elution profiles of Sephadex G-200 and hydroxyapatite. The different enzyme activities were monitored as described in Materials and Methods. Activated phosphoribulokinase corresponds to the enzyme which has been incubated with 50mM dithiothreitol for 1 h at 30°C before being assayed. (0-0) Phosphoribulokiuase, (0---0) activated phosphoribulokinase, (0-0) ribose-phosphate isomerase, ( A-A) glyceraldehyde-3-phos- phate dehydrogenase, (0-0) ribulose-bisphosphate carboxylase/oxygenase. (A) Gel filtration on Sephadex G-200. The arrows indicatc the predicted positions for elution of the isolated enzymes as follows: (1) ribulose-bisphosphate carboxylase/oxygenase, (2) glyceraldehyde-3- phosphate dehydrogenase, (3) phosphoribulokinase, (4) ribose-phosphate isomerase, (5) phosphoglycerate kinase. Column flow rate was 8 ml/h. The activity ratios, calculated in relation to activated phosphoribulokinase activity, can be considered constant within the experimental error value across the peak eluted in the void volume. Thus, the activity ratios for ribose phosphate isomerase, ribulose-bisphosphate carboxylaseloxygenase and glyceraldehyde-3-phosphate dehydrogenase were respectively 0.1 13, 0.064, 0.1 1 at 125.7 ml; 0.08, 0.055, 0.1 3 at 151.30 ml; 0.048, 0.042, 0.096 at 164 ml; 0.062, 0.043, 0.091 at 176.80 ml; 0.098, 0.06, 0.085 at 189.50 ml. (B) Chromatography on hydroxyapatite. Column flow rate was 30 ml/h. The line (-) indicates the potassium phosphate concentration. The inset shows the activity ratios calculated in relation to phosphoribulokinase activity for ribose-phosphate isomerase (0-0) and glyceraldehyde-3-phosphate dehydrogenase (A-A) across peaks A and B

bulokinase, ribulose-bisphosphate carboxylase/oxygenase, phosphoglycerate kinase and glyceraldehyde-phosphate de- hydrogenase activities. These fractions were pooled and dia- lysed against a third buffer, C, made up with 50 mM potass- ium phosphate pH 7.2 and 10 mM cysteine. The resulting extract was then loaded on a hydroxyapatite column (2.5 x 14 cm). After washing with buffer C, the column was eluted with a linear gradient of 50 - 200 mM potassium phos- phate. The fractions containing phosphoribulokinase activity were collected.

Enzyme assays

Phosphoribulokinase and ribose-phosphate isomerase ac- tivities were determined at 30°C by a procedure derived from

that described in [lo]. These activities were coupled to NADH oxidation via pyruvate kinase and lactate dehydrogenase. When measuring isomerase activity, the assay cuvette contained 50 mM glycylglycine buffer pH 7.7,lO mM MgC!2, 0.5mM EDTA, 2.75 units lactate dehydrogenase, 1 unit pyruvate kinase, 1 mM phosphoenolpyruvate, 5 mM dithio- threitol, 0.15 mM NADH, 1 mM ribose 5-phosphate and 0.5 mM ATP. The same mixture, supplemented with ribose- phosphate isomerase from the yeast Torula, was used for measuring the kinase activity. Glyceraldehyde 3-phosphate dehydrogenase, phosphoglycerate kinase and ribulose-bis- phosphate carboxylase/oxygenase activities were determined as described in [ l l - 131. The complex activity was measured using the ribulose-bisphosphate carboxylase/oxygenase assay, but without the addition of exogenous phosphoribulokinase,

439

Table 1. Content of enzyme uctivities in peaks A and B of hydroxy- uputite The enzyme activities were assayed as described in Materials and Methods. The total protein content was 712 pg and 541 pg in peak A and B, respectively

Enzyme Peak A Peak B

units

Ribulose-bisphosphate carboxylase/ oxygenase 8.5 6.7

Phosphoribulokinase 6 5" Ribose-phosphate isomerase 3.7 3" Glyceraldehyde-3-phosphate

deh ydrogenase 0.3 1.4 Phosphoglycerate kinase 0.09 0.15

a Assays were also done arter preincubation with 50 mM dithio- thrcitol for 1 h. The activities of both enzymes are enhanced from 5 units to 17 units, and from 3 units to 7.5 units.

ribose-phosphate isomerase, glyceraldehyde-phosphate de- hydrogenase and phosphoglycerate kinase. Estimation of en- zyme activities and absorbance measurements were effected with a six-sample Pye Unicam 8610 spectrophotometer con- nected to an Apple I1 microcomputer. A unit of enzyme ac- tivity is defined as the amount of enzyme which transforms 1 Fmol substrate/min under the conditions of the assay.

Determination of molecular mass and homogeneity of the complex

Molecular mass estimations were effected by molecular sieve chromatography through Sephacryl S-300, by analytical ultracentrifugation and by slab gel electrophoresis. Sephacryl 5-300 columns (2.6 x 90 cm) were calibrated with proteins of known relative molecular mass (thyroglobulin 669 kDa, ferritin 449 kDa, catalase 240 kDa, aldolase 147 kDa, bovine serum albumin 68 kDa, peroxidase 50 kDa, cytochrome c 12.5 kDa, aprotinin 6.5 kDa) using buffer B [14]. Analytical ultracentrifugation experiments were performed with a Beckman ultracentrifuge model E equipped with an in- terferometric system for molecular mass determinations ac- cording to the method of Yphantis [15] and ultraviolet or schlieren optics for sedimentation patterns. Slab gel electrophoresis under denaturing conditions was conducted as in [16]. Under non-denaturing conditions the experimental protocol proposed by Pharmacia was followed. Protein con- centrations were determined by the dye-binding assay of Bradford [17].

RESULTS In the course of elution of DEAE-Trisacryl columns, the

fractions that contain phosphoribulokinase also contain ribose-phosphate isomerase, ribulose-bisphosphate carbox- ylase/oxygenase, glyceraldehyde-phosphate dehydrogenase and phosphoglycerate kinase. Despite the large differences of their molecular mass (spanning from 50 kDa to 550 kDa), enzyme activities coelute in the void volume of the Sephadex G-200 column (Fig. 1 A). A second peak of phosphoribulo- kinase activity is obtained later on, during the process of elution, and does not contain the other enzyme activities. The presence of glyceraldehyde-phosphate dehydrogenase and

Fig. 2. Gradient polyacrylamide gel slab: 4-30%. From left to right: track 1, 10 pg thyroglobulin; track 2, 10 pg complex A eluted from hydroxyapatite chromatography; track 3, 10 pg complex B eluted from hydroxyapatite chromatography; track 4, mixturc of 10 pg catalase, LO pg thyroglobulin, 10 pg ferritin; track 5, 10 pg RNA polymerase; track 6, 10 pg ferritin; track 7 , 10 pg catalase. Electro- phoresis was done at 70 V for about 20 min, until the protein moved into the gel, then for 16 h a t 150 V. Electrophoresis buffer was 90 mM Tris, 80 mM boric acid, 2.5 mM NazEDTA pH 8.4

phosphoribulokinase activities in the void volume of the Sephadex G-200 column, is consistent with the previous finding that these enzymes may occur under a state of high molecular mass [ll, 18 -231. The fractions of the void volume that contain the five enzyme activities generate, upon hydroxyapatite chromatography, two peaks, A and B, each of which contains the same five enzyme activities (Fig. 1 B). The ratios of the various enzyme activities remain approx- imately constant across each of the two activity peaks. Two of these activity ratios are given as an inset in Fig. 1 B and some other values are given in the legend of the figure. The phosphoglycerate kinase activity, however, is too low to be determined in every fraction across the elution profile. The reason for this low activity is unknown at the moment but might be due to a poor accessibility of the enzyme's active site within the complex to the bulk substrate.

The results of the hydroxyapatite chromatography are reminiscent of the findings of Nicholson et al. [24], which showed the existence of two fractions of phosphoribulokinase activity extracted from the green alga Scenedesmus obliquus. Table 1 shows the respective amounts of the various enzyme activities present in the fractions of the two peaks A and B.

In order to test the purity of these multienzyme complexes, electrophoresis was performed in polyacrylamide gradient gels (4-30%) or under native conditions. As shown in Fig. 2, peak A yields a single band with a molecular mass of 449 kDa, but fractions of peak B contain two molecular entities with molecular mass of 449 kDa and 669 kDa respectively. These results strongly suggest that peak B contains the same multienzyme complex as peak A plus an additional polypep- tide. The homogeneity of the multienzyme complex A may be verified by analytical centrifugation. Results of Fig. 3A show that this complex is indeed homogeneous. A schlieren profile of this fraction is also symmetrical thus giving additional support to this conclusion (Fig. 3 B).

In order to define their composition, electrophoresis of these protein complexes was performed in 9% SDS/poly- acrylamide gels (denaturing conditions). The results obtained are shown in Fig. 4.

Fractions of peak A show the two subunits of ribulose- bisphosphate carboxylase/oxygenase, the subunits of phos- phoribulokinase and of glyceraldehyde-phosphate dehydro- genase. The amount of ribose phosphate isomerase and

440

B

I 50 50. 5 51 51. 5

r2 (crnI2

Fig. 3. ( A ) Equilibrium sedimentation analysis of complex A at p H 7.5. ( B ) Sedimentation patterns (schlieren optics) of the jive-enzyme complex. (A) The protein concentration was 1.2 mg/ml. The rotation speed was 9700 rev./min. Column lengths were 3 mm. Temperature 20°C and experiments were run for 48 h. J, the fringe displacement. Buffer used was 100 mM phosphate, 2 mM EDTA, 0.15 M NaCI, 10 mM cysteine, 10% glycerol pH 7.5. The molecular mass obtained was 491.3 + 5.7 kDa. (B) Migration is from right to left. Bar angle is 60°C. Photographs were taken every 4 min. Protein concentration was 5.8 mg/ml in buffer B containing glycerol 5%. Photographs show a single symmetrical peak

phosphoglycerate kinase is such that no bands pertaining to those enzymes are detected. For ribose-phosphate isomerase this is not surprising, since we have found that on polyacrylamide gel, very large amounts (above 30 pg) of this enzyme are needed in order to detect a band by Coomassie blue. Fractions of peak B display basically the same pattern as those of peak A, except that a band of 65 kDa molecular mass is present in the fractions of peak B, but not of the peak A. This 65-kDa protein presents the characteristic of not being stained by silver dyes (data not shown). Although we still do not know the function of this protein, analysis of the literature suggests that it may correspond to a manganese- protein possibly involved in oxygen evolution [25]. It thus appears quite likely that the complex present in the fractions of peak B is also present in the fractions of peak A, but associated with the 65-kDa polypeptide.

If fractions contained in the void volume of Sephadex G-200 filtration are submitted to Sephacryl S-300 chro- matography, two peaks are obtained upon elution (Fig. 5). In peak 1 the five enzyme activities are detected as a molecular entity of 536 kDa. Moreover, under these conditions the ac- tivity of phosphoribose isomerase, phosphoribulokinase and glyceraldehyde-phosphate dehydrogenase are greatly en- hanced by preincubation with dithiothreitol. A second protein peak is also detected on the elution profile that pertains to a molecular mass of 190 kDa and contains ribose-phosphate isomerase and phosphoribulokinase but lacks ribulose-bis- phosphate carboxylase/oxygenase, phosphoglycerate kinase and glyceraldehyde-phosphate dehydrogenase activities.

The multienzyme complex of 536 kDa appears homo- geneous on analytical centrifugation. By equilibrium sedimen- tation one finds a molecular mass of 520 kDa, which is consis- tent with the value of 536 kDa given above. In sedimentation velocity studies the pattern appears symmetrical (Fig. 3 B). Analysis of subunit composition of this complex allows one to detect the subunits of ribulose-bisphosphate carboxylase/ oxygenase, glyceraldehyde-phosphate dehydrogenase, phos- phoribulokinase and the polypeptide of 65 kDa.

Dithiothreitol has been reported to promote the depoly- merization of a ‘high-molecular-mass form of phosphoribulo- kinase [21, 221. This reductant has, therefore, been used to promote dissociation of ribulose-bisphosphate carboxylase/ oxygenase from the rest of the complex and to show whether the enzyme obtained in that way may have a molecular mass different from that of the so-called ‘native’ enzyme, namely 550 kDa. Upon treatment of the complex with dithiothreitol and molecular sieving through Sephacryl, several fractions are obtained (Fig. 6). The first contains a complex of lo3 kDa totally devoid of activity. The second protein peak has ribulose-bisphosphate carboxylase activity, but the corre- sponding molecular mass is 420 kDa instead of 550 kDa. The 65-kDa polypeptide is also detected in the same fraction.

The rather low molecular mass of this form of ribulose- bisphosphate carboxylase/oxygenase strongly suggests that this enzyme may exist with a structure different from the classical ABBB quaternary structure. This 420-kDa form con- taining the 65-kDa polypeptide resembles, however, another

44 1

0 0.5 R f

1

Fig. 4. ( A ) 9% SDS/poIyacrylarnide gel; (B) gelcalibration curve. (A) From left to right: track 1, 10 pg phosphorylase 6 ; track 2, 10 pg rabbit muscle glyceraldehyde-3-phosphate dehydrogenase; track 3, 30 pg spinach phosphoriboisomerase; track 4, 7 pg lactate dehydrogenase; track 5, 10 pg pyruvate kinase; track 6, 15 pg pure spinach phosphoribulokinase; track 7, 50 pg complex B; track 8, mixture of 20 pg phosphoriboisomerase, 10 pg glyceraldehyde-3-phosphate dehydrogenase, 10 pg phosphoribulokinase; track 9, 45 pg complex A. (B) SDS/ polyacrylamide gel calibration curve: phosphorylase b (I), rabbit muscle glyceraldehyde-3-phosphate dehydrogenase (4), lactate dehydrogenase ( 5 ) , pyruvate kinase (2), spinach ribose-phosphate isomerase (6), spinach phosphoribulokinase (3) were used as markers. Arrows indicate the position of polypeptides found in complex A (b, c, d, e) and complex B (a, b, c, d, e). We have found for a, b, c, d and e molecular masses of 65,55,40,38 and 36 kDa. In the bottom of the gel for complexes A and B, a small polypeptide of 14 kDa, corresponding to the small ribulose- bisphosphate carboxylase/oxygenase subunit, was also found. This polypeptide is not represented in the calibration curve because of the poor resolution of small-molecular-mass polypeptides in 9% SDS gel

form of this enzyme isolated from blue-green algae, which shows a molecular mass of 450 kDa [26], and another form from Rhodopseudomonas palustris and sphaeroides that shows a molecular mass of 360 kDa [27, 281. The other fractions contain a complex of 130 kDa molecular mass which has phosphoribulokinase activity but has no ribulose-bis- phosphate carboxylase/oxygenase. It thus appears probable that the complex present in the first peak is an artefactual molecular aggregate resulting from a thiol - disulfide inter- change. The interesting result that emerges from these data is that, after a dithiothreitol treatment, ribulose-bisphosphate carboxylase may occur with a molecular mass and an aggregation state different from that of the so-called native enzyme.

To show that the complex of these five enzymes of the Calvin cycle represents a functional entity, ribose 5-phos- phate, the substrate of the first reaction, was mixed with the complex in the presence of ATP, bicarbonate and NADH, substrates required for later steps, and the progress of the final reaction was monitored by measuring the production of NAD+. The results (Fig. 7) show that NADH is continuously oxidized by the complex, thus implying that the sequence of the five enzyme reactions was indeed taking place.

Moreover, the steady-state velocity of product appearance at saturating concentrations of ribose phosphate, ATP and NADH was identical, within experimental error, to the rate of oxidation of NADH at saturating glycerate 3-phosphate, coenzyme and ATP concentrations. This result suggests the existence of channelling within this multienzyme complex.

DISCUSSION An important question raised by the present results is that

the molecular mass of the complex, determined by different

techniques, is smaller than the value (904 kDa) one would expect from the sum of the molecular masses of the individual enzymes. A likely explanation of this apparent discrepancy is that the quaternary structures of these enzymes within the complex are different from the ones they have when isolated in free solution. A number of arguments seem to support this contention. It is well known [29, 301 that the large subunit of ribulose-bisphosphate carboxylase/oxygenase is highly hy- drophobic and insoluble in the absence of the small subunit. It is probable that, depending on its microenvironment, different interactions with different enzyme molecules may stabilize this hydrophobic subunit. The result that dithiothreitol treatment promotes a partial depolymerization of the complex, which releases an enzyme form that retains its activity, is consistent with this view. A complex made up of four pairs of heavy and light subunits of ribulose-bisphosphate carboxylase/oxyge- nase (54 kDa+ 14 kDa) instead of eight, one subunit of phosphoribulokinase (42 kDa) instead of two, one pair of subunits (39 kDa + 36 kDa) of glyceraldehyde-phosphate de- hydrogenase instead of two, one molecule of phospho- glycerate kinase (50 kDa) and one subunit of ribose-phos- phate isomerase (30 kDa) instead of two, would nicely fit the observed molecular mass if the complex is associated with the 65-kDa polypeptide. Then the overall molecular mass would be 542 kDa, which is quite compatible with the observed values of 536 kDa and 520 kDa. The presence of a 65-kDa polypeptide chain could also be a point of interest as it re- sembles a manganese-protein from the thylakoid membrane [25] . Thus it is attractive to speculate that the 65-kDa polypep- tide corresponds to this thylakoid protein, in which case it could serve as an anchor for the complex we have described. If this were the case, the chloroplast multienzyme complex would be attached to the thylakoid, essentially as the Krebs

442

2. 0

1. 5

A

Ti.0 3

0. 5

0

A 1 I

1

150 200 250 300 350

Elution volume ( m l l

2. 5 B

2. 0 K

3. 5 4 4. 5 5 5. 5 6

Log ( M r )

Fig. 5. ( A ) Molecular sieving of the ‘void-volume Sephadex G-200 ,fraction’on Sephacryl S-300. ( B ) Estimation uf the molecular mass of the complex. (A) The different enzyme activities were monitored as described in Materials and Methods. (0-0) Activated phos- phoribulokinase, (0-0) ribose-phosphate isomerase, (A-A) glyceraldehyde-3-phosphate dehydrogenase, (0-0) ribulose-bis- phosphate carboxylasc/oxygenase. The arrows indicate the predicted positions for elution of the isolated enzymes as follows: (1) ribulose- bisphosphate carboxylase/oxygenase, (2) glyceraldehyde-3-phosphate dehydrogenase, (3) phosphoribulokinase, (4) ribose-phosphate iso- merase, ( 5 ) phosphoglycerate kinase. Column flow rate was 15 ml/h. The activity ratios, calculated in relation to activated phospho- ribulokinase, can be considered constant within the expcrimental error value across peak I. Thus, the activity ratios for ribose-phos- phate isomerase at 224 ml, 234 ml and 244 ml are respectively 0.16, 0.16,0.33; for glyceraldehyde-3-phosphate dehydrogenase at 214 ml, 234 ml and 244 ml thcy are respectively, 0.2, 0.29, 0.21; for ribulose- bisphosphate carboxylase/oxygenasc at 224 ml, 234 ml and 244 ml they are, respectively, 0.16,0.31,0.36. (A) Estimation of the molecular mass was performed by molecular sieving through a calibrated Sephacryl S-300 column by following the technique of Whitaker [14]. Aprotinin (l), cytochrome c (2), peroxidase (3), bovine serum albumin (4), aldolase (9, catalase (6), ferritin (7), thyroglobulin (8) were used as markers. The arrow indicates the position of Sephacryl S-300 peak I, which corresponds to a molecular mass of 536 kDa

cycle multienzyme complex is attached to the inner mitochondria1 membrane [4].

Some results presented above are quite consistent with previous proposals. For instance, ribulose-bisphosphate car- boxylase of pea is associated with ribose-phosphate isomerase and phosphoribulokinase activities [7]. Moreover the forms of phosphoribulokinase and glyceraldehyde-phosphate de- hydrogenase of high molecular mass (550 kDa [21, 221 and 600 kDa [18,20] respectively) that have been called ‘regulatory forms’ may either represent an aggregated state of the same enzyme molecules or a multienzyme complex similar to the one studied in this report.

0.6 0.25

150 200 250 300 3 50

Elution volume (ml)

Fig. 6. Dissociating effect of dithiothreitol. Gel filtration of Sephacryl S-300 peak I on Sephacryl S-300 after dithiothreitol treatment. A fraction of concentrated S-300 peak I was reduced with 50mM dithiothreitol for 1 h at 30°C and then submitted to a second gel filtration using buffer B containing 5 mM dithiothreitol instead of 10 mM cysteine. Ribulose-bisphosphate carboxylase/oxygenase ac- tivity (+-+) and phosphoribulokinase activity (0-0) were monitored as described in Materials and Methods. The arrows indi- cate the predicted positions as indicated in Fig. 1 A

1.3 ‘ I 0 5 10 15 20 25

Time(rnin)

Fig. I . Functionality of the 520-kDa complex. The multienzyme complex was preincubated for 15min in the presence of 50mM glycylglycine, 0.5 mM EDTA, 10 mM MgZt, 5 mM dithiothreitol and 50 mM sodium bicarbonate. After this prcincubation, the activity of thc complex was measured at pH 7.7 by adding the following (final concentrations): 1 mM ribose Sphosphate, 1 mM ATP, 0.25 mM NADH, 5 units triose-phosphate isomerase and 11 units gly- cerophosphate dehydrogenase in a final volume of 1 ml. The multienzyme complex concentration used was 0.3 pM. A control, in which ribose 5-phosphate was omitted, gave no decrease of absorbance indicating no interfering NADH oxidase activity (other details are as described in Materials and Methods)

The main point of this study is related to the functionality of the multienzyme complex. The existence of a complex of this sort in chloroplasts may represent a functional advantage as it allows chanelling of reaction intermediates of the Calvin cycle from one site to another, thus preventing diffusion of these intermediates in the bulk phase. Results presented in this report suggest that this release of intermediates does not occur to a significant extent, as the velocity of the last step of the reaction sequence at saturating substrate concentration is identical to that of the overall process at saturating ribose- phosphate concentration. Although the existence of chan-

443

nelling cannot yet be taken as proved, it is suggested by these results. Further studies are now under way to investigate it further and to study its dynamics.

The expert technical assistance of Mrs Mireille Rivikre and Mr Paul Sauve is deeply appreciated. Thanks are due to Mr A. Ribas for having grown the spinach. This work was supported in part by Biologic. Muleculuire Egbtale du Centre National de la Recherche Scientijique, Action Thematique Programmbe.

REFERENCES 1. Fulton, A. B. (1982) Cell 30, 345-347. 2. Clegg, J. S. (1984) Am. J . Physiol. 246, R133-151. 3. Tompa, P., Bar, J. & Batke, J. (1986) Eur. J . Biochem. 159, 117-

4. Robinson, J. B. & Srere, P. A. (1985) J . Biol. Chem. 260,10800-

5. Welch, G. R. & Clegg, J. S. (eds) (1987) The organization of cell

6. Keleli, T. & Ovadi, J. (1987) Curr. Top. Cell. Regul., in the press. 7. Sainis, J. K. & Harris, G C. (1986) Biochem. Biophys. Res.

Commun. 139,947-954. 8. Anderson, L. E. (1987) FEBS Lett. 212,45-48. 9. Jensen, R. G. & Bassham, J. A. (1966) Proc. Natl Acad. Sci. USA

10. Porter, M. A,, Milanez, S., Stringer, C. D. & Hartman, F. C.

11 . Pawlizki, K . & Latzko, E. (1974) FEES Lett. 42,285-288. 12. Lavergne, D. & Bismuth, E. (1973) Plant. Sci. Lett., 229-236. 13. Pierce, J., Lorimer, G. H. & Reddy, G. S. (1986) Biochemistry 25,

124.

10805.

metabolism, Plenum Press, New York.

56,1095-1101.

(1986) Arch. Biochem. Biophys. 245,14-23.

1636- 1644.

14. Whitaker, J. R. (1963) Anal. Chem. 35, 1950-1953. 15. Yphantis, D. A. (1964) Biochemistry 3, 297-317. 16. Laemmli, U. K. (1970) Nature (Lond.) 227, 680-685. 17. Bradford, M. M. (1976) Anal. Biochem. 72,248-252. 18. Yonuschol, G. R., Ortwerth, B. J. & Koeppe, 0. J. (1970) J . Bid.

19. Pupillo, P. & Piccari, G. G. (1973) Arch. Biochem. Biophys. 154,

20. Pupillo, P. & Piccari, G. G. (1975) Eur. J . Biochem. 51,475-482. 21. Ward-Aswapati, O., Kemble, R. J. & Bradbeer, J. W. (1980) Plant

Physiol. 66, 34 - 39. 22. Bradbeer, J. W., Riiffer-Turner, M. E., Deferre, E. J., Wara

Aswapati, 0. & Kemble, R. J. (1981) Photosynthesis IV, regu- lation of carbon metabolism (Akoyunoglou Balaban, G., ed.) pp. 389 - 394, International Science Services, Philadelphia.

23. Wolosiuk, R. A. & Buchanan, B. B. (1978) Arch. Biochem. Bio-

24. Nicholson, S., Easterby, J. S. & Powls, R. (1987) Eur. J . Biochem.

25. Spector, M. & Winget, G. D. (1980) Proc. Nut1 Acad. Sci. USA

26. Tabita, F. R., Stevens, S. E. Jr & Gibson, J. (1974) J . Bacteriol.

27. Anderson, L. E., Price, G. B. & Fuller, R . C. (1968) Science

28. Gibson, J. L. & Tabita, F. R. (1977) J . Biol. Chem. 252, 943-

29. Jordan, D. B. & Chollet, R. (1985) Arch. Biochem. Biupizys. 236,

30. Andrews, T. J. & Lorimer, G. H. (1985) J . Bid. Chem. 260,4632-

Chem. 245,4193-4198.

324 - 33 1.

phys. l # Y , 97-101.

162,423-431.

77,957 - 959.

125, 531 - 539.

(Wash. DC) 161,482-484.

949.

487 - 496.

4636.