Purification of a Complex of Alkaline Fructose 1,6=Bisphosphatase and Phosphoribulokinase

have been isolated and purified as a single complex from Rhodospirillum rubrum. During purification proce-dures involving ammonium sulfate precipitation and column isoelectric focusing column. other

chromatography with both DEAE-cellulose and Sephadex, the ratios of activity of one enzyme to that of the other remained approximately the same. The pure complex appeared as a single band on polyacrylamide gels, and the activities of the two enzymes showed the same profiles on an isoelectric focusing column. No other enzyme of the reductive pentose phosphate cycle could be detected in the complex.
The complex was unstable, so that after 48 hours the activity of phosphoribulokinase was reduced by 80% and that of fructose 1,6-bisphosphatase by 50%. This loss of activity was accompanied by the appearance of two bands on the polyacrylamide gels.
The most widespread series of reactions responsible for the fisat,iou of carbon dioside into the early products of photosynthesis are those that constitute the reductive pentose phosphate cycle (the Calvin cycle). Recent,ly, there has been increasing interest in the mechanism by which this cycle is regulated (1). Essent'ially, there have been two approaches to the problem of control of the Calvin cycle. The first involves "physiological" studies of intact organisms or intact rhloroplasts (2-5). The second approach has been the isoltition of enzymes believed to be important in the regulation of l)llotosVlltltetic' carbon dioxide assimilation and attempbs to affect the activity of such enzymes in ways that might relate to their control in z%vo (6)(7)(8)(9)(10)(11)(12)(13)(14) The present work began as an attempt to study the way in which the activity of fructose bisphosphatase in extracts from the photosynthetic bacterium Rhodospirillum rubrum might, be modified bv various effect,ors. However, during the purification of this enzyme the activity of another possible regulatory enzyme of the reductive pentose phosphate cycle, phosphoribulokinase, was found to parallel that of the bisphosphat.ase. The present paper describes t,he isolat.ion and purification of a complex of these t.wo enzymes.

EXPERIMENTAL PROCEDURE
Growth of Organism R. rubrum st'rain Sl was obtained from the collection of one of us (R.C.F.).
The organism was grown nutotrophically by bubbling this basal medium continuously with hydrogen containing 2 to 5% (v/v) carbon dioxide at a rate of 2 liters per hour per liter of medium. The light intensity (provided by a series of BO-watt tungsten filament lamps) at the culture surface was about 850.
foot candles, and the temperat.ure was 30". The bacteria were 4834 harvested 48 hours (about three divisions) after inoculation with a 2.57; inoculum from the midexponential phase of a culture growing photoheterotrophically on medium supplemented with 2.6 g of nL-malnte per liter of medium.
For photoheterotrophic growth on acetate the basal medium was supplemented with 1.8 g of sodium acetate per liter of medium.
Photoheterotrophic growth occurred in bottles filled completely with medium and illuminated at 30" for 36 hours.

Preparation of Cell-free Extracts
Up to 100 liters of organism were harvested in a continuous flow centrifuge, and the pellet was washed twice in ice-cold 0.02 M Tris-HCI buffer (pH 7.5) and resuspended in this buffer. The density was adjusted to about 0.3 g of cells, fresh weight, per ml of buffer, and the suspension was broken by a single passage through a French pressure cell at 10,000 p.s.i. After cooling in an ice bath (addition of commercial ribonuclease at this stage removed the coagulation caused by high concentrations of nucleic acids), the suspension was first centrifuged at 10,000 X g for 10 min, and then the suspension was centrifuged at 100,000 X g for 1 hour. All centrifugations were performed at 4". The pale red supernata.nt was removed carefully and used as the crude extract.
Extracts were usually stored at -20" for up to 4 days; no loss of activity was detected over a period of 2 weeks.
The rate of reaction was measured continuously by following the increase in absorption at 340 nm. This continuous assay showed that the reaction was linear with time, and in some experiments the method was modified to a single point assay. The components of the reaction mixture omitting NADPf, phosphohexose isomerase, and glucose 6-phosphate dehydrogenase, were incubated for 5 min. The reaction was then stopped by placing the reaction mixture in a boiling water bath for 1 min before cooling in an ice bath. The mixture was then incubated with NADP+, phosphohexose isomerase, and glucose 6-phosphate dehydrogenase at 30" for 10 min, and the absorbance at 340 nm was measured.
The activity determined by this method was the same as that by continuous assay.
Two further modifications were made to the above basic assay method for fructose bisphosphatase activity. When precise measurements of fructose hisphosphate concentration were required, excess enzyme was added and the reaction was allowed to proceed to completion.
The concentration of NADP+ reduced (and hence the concentration of fructose bisphosphate) was determined from the millimolar extinction coefficient of reduced pyridine nucleotide of 6.22 at 340 nm with a l-cm light path.
In some experiments fructose bisphosphatase was assayed by measuring the production of inorganic phosphate.
The reaction mixture was the same as that above, except that NADP+ and the linking enzymes were omitted.
After 5 min at 30" the reaction was stopped by adding 1.0 ml of ice-cold 57; (w/v) trichloroacetic acid, and the pH was brought to 4.0 with 4 volumes of 0.1 M sodium acetate (so minimizing hydrolysis of fructose bis-phosphate). After removing any precipitated protein by cent,rifugation, inorganic phosphate in the supernatant was measured by the method of Lowry and Lopez (16). The activity measured by this method was the same as that with the "coupled enzyme" technique.

Activity
The activity of this enzyme was measured by two methods. In the first, ADP produced in the reaction was measured by following the oxidation of NADH in the presence of excess pyruvat.e kinase (ATP :pyruvate phosphotransferase, EC 2.7.1.40) and lactate dehydrogenase (L-lactate oxidoreductase, EC 1.1.1.27) (17). The reaction mixture contained in 1 ml: 100 pmoles of Tris-HCl (pH 8.25), 5 pmoles of IVIgClt-6H&, 0.3 pmole of ATP, 0.3 ,umole of phosphoenolpyruvate, 0.2 pg of lactate dehydrogenase, 0.1 pg of pyruvate kinase, and 0.01 to 0.1 ml of extract. After 5 min at 30" the reaction was started by adding 0.5 pmole of ribulose 5-phosphate.
With some crude extracts the presence of NADH oxidase activity made this assay unusable; this was not the case with the partially purified preparations.
The second method of measuring phosphoribulokinase activity was by following the fixation of sodium[14C] bicarbonate into acid-stable products when the phosphoribulokinase reaction was coupled to a homogeneous preparation of ribulose bisphosphate carboxylase purified from R. rubrum (after (12)). The reaction mixture was the same as that above, except that the phosphoenolpyruvate, NADH, pyruvate kinase, and lactate dehydrogenase were omitted.
After 5 min the reaction was stopped by placing the reaction mixture in a boiling water bath for 1 min and then cooling rapidly in an ice bath.
After 10 min (preliminary experiments showed that all ribulose bisphosphate was convert,ed to 3-phosphoglycerate in 2 to 3 min), 0.2 ml of the mixture was added to 1 ml of glacial acetic acid. After drying, scintillant (6 g of butyl-PBD, 750 ml of toluene, and 250 ml of methanol) was added, and the radioactivity was determined with a Packard Tri-Carb Scintillation Spectrometer, model 3320. In the above assays absorption at 340 nm was measured with a Cary (model 14) or a Unicam S.P. 800 or S.P. 1800 spectrophotometer.
Protein was measured by the method of Lowry et al. (18), and by absorbance at 280 nm.

Enzyme Purification Method
The final procedure presented in the results section involved the following techniques.
Dialysis-Adsorbed metal ions were removed by boiling t,he dialysis tubing in 0.02 M EDTA.
EDTA was removed by boiling the tubing twice in Pyrex-distilled water. Ext,racts were dialyzed against 200 times their volume of buffer with a.t least one change of buffer.
Ammonium Sulfate Precipitation-The estract was stirred at 4' and solid ammonium sulfate was added as one addition (recovery of the enzyme was not increased by using a saturated solution of ammonium sulfate).
Stirring continued for 30 min after the ammonium sulfate had dissolved.
Column Chromatography on Sephadex and DEAE-cellulose-Sephadex G-25 and G-200 were prepared according to the manufacturer's instructions.
Sephadex G-200 columns (45 x 2.5 cm) were run with upward flow of buffer.
Previously swollen micro-granular DEAE-cellulose (DE 52) was prepared according to the manufacturer's (Whatman) instructions. After degassing the cellulose and equilibrating it with the required buffer, the column (30 x 0.9 cm) was filled through an extension tube with a concentrated slurry of 1 wet settled volume of equilibrated DEAEcellulose and 0.5 volume of buffer. Buffer was immediately pumped through the column at 75 ml per hour per cm2 until the bed height reached a constant level.

Analytical
Polyacrylamide Gel Electrophoresis-Polyacrylamide gels were prepared by the method of Davis (19). Protein samples (containing 20 to 50 Fg of protein per gel) in 10% sucrose were layered onto the surface of the gels. A constant current of 2 ma per gel was applied from a direct current power supply until the tracking dye was within 0.5 cm of the base of the gel. Protein was stained with lo{> (w/v) naphthalene black in 7'; (v/v) acetic acid and destained with 7% acetic acid.
Preparative Polyacrylamide Gel Electrophoresis-The gels were prepared by the method of Davis (19) and the apparatus used was the 1Zuchler apparatus (Poly Prep).
The buffer in the upper chamber was 0.01 M l'ris-base and 0.525 M pjycine (the same as the elution buffer); in the lower chamber the buffer was 0.5 M Tris-HCl (pII 9.0). The proteill sample in 5%, sucrose was pumped onto tlhe surface of the gel, and a constant current of 50 ma was applied.
The voltage increased from 200 to 400 volts during the run, fractions being collected every 30 min.
Isoelectric Focusing-AI1 LKlZ 8101 column was used. Carrier ampholyte (LKB Anll)holine) of pH range 3.0 to 6.0 (constant voltage, increased stepwise from 300 to 700 volts during the first 2 to 3 hours) were used. After 3 days of isoelectric focusing tile column was drained and 2-ml fractions were collected each minute.

RESULTS
Studies with Crude Cell-free Extracts-Most photosynthetic organisms have two fructose bisphosphatases with different pH optima.
The enzyme with the alkaline pH optimum appears to be that of the reductive pentose phosphate cycle, and that with the more neutral pH optimum is associated with gluconeogenesis (20). However, no data were available for R. rubrum and it therefore seemed desirable to determine whether this organism also possessed two fructose bisphosphatases, and whether the ellzyme associated with photosynthetic carbon dioxide assimilation was, in fact, the one we were studying.
Two pieces of evidence suggested that R. rubrum did contain both enzymes and that the one with the more alkaline pH optimum was the photosynthetic one. Firstly, the pH profile of fructose bisphosphatase activity in crude cell-free extracts became modified after dialysis (Fig. 1). The activity of the dialyzed preparation showed a more rapid loss of activity with decreasing pH below 8.5, suggesting a reduction in the activity of an enzyme with a pfl optimum between 7.5 and 8.5. This loss of activity at neutral pH values was also observed by Smillie (20) in extracts of Buglena gracilis.
A second piece of evidence came from an examination of the effects of photoheterotrophic growth on the activity of fructose bisphosphatase at various pH values. Anderson and Fuller (21) showed that photoheterotrophic growth on malate or acetate reduced the activity of the reductive pentose phosphate cycle in photosynthetic carbon dioxide assimilation by R. rubrum. Table I shows that photoheterotrophic growth had a most marked effect on the bisphosphatase activity at the more alkaline pH values, indicating  that the enzyme with the alkaline p1-I optimum was associated with the photoassimilation of carbon dioxide in this organism. Therefore, pH 8.5 was used for all further nssnys of fructose bisphosphatase activity described in this paper.
Purification Procedure for Fructose Bisphosphatase-Phosphoribulokinase Complex-After the pH of the crude extract was adjusted (where necessary) to pH 6.0 with lactic acid, solid ammonium sulfate was added, and the fraction precipitated between 25 and 40% saturation was retained (no activity was detected in the other ammonium sulfate fractions).
The precipitate was dissolved in 0.02 M Tris-HCl buffer (pH 8.0) and the ammonium sulfate was removed by passage through Sephadex G-25.
The protein solution (about 25 ml, still highly pigmented) was pumped on to the DEAE-cellulose column (pH 8.0) at the rate of 50 ml per hour.
Most of the pigment was not absorbed and both fructose bisphosphatase and phosphoribulokinase appeared in the same fractions.
The fractions of peak activity were pooled, and ammonium sulfate was added.
The precipitate from the 25 to 400/, fraction was dissolved in 0.02 M malonate buffer (pH 6.0) and the solution was desalted by passage through Sephadex G-25. The protein solut'ion (about 15 ml, still slightly pigmented) was pumped on to a second DEAE-cellulose column (pH 6.0) at a rate of 30 ml per hour. After washing with 30 ml of buffer (removing some of the pigment) the protein was eluted with 300 ml of a 0.1 to 0.25 M linear gradient of KC1 in 0.02 M malonate buffer (pH 6.0). Again the peaks of activities of both enzymes coincided.
The protein from the pooled peak fractions was precipit'ated by adding a saturated solution of ammonium sulfate to a final value of 507, saturation.
After dissolving the precipitate in 1 ml of 0.02 M malonate buffer (pH 6.0), the solution was applied to a Sephadex G-200 column and eluted with the same buffer. Again the peaks of activity of the two enzymes coincided (Fig. 2). Also, in the eluate from this column the protein peak was well defined and coincided with that of the two enzyme activities (Fig. 2).
The two enzymes could not be separated on an isolectric focusing column (Fig. 3). Furthermore, Fraction 10 from the G-200 column (Fig. 2) appeared as a single prot.ein band on analytica polyacrylamide gels (Fig. 4). The purification procedure is summarized in Table II,    b, complex stored for 48 hours at 4" before being pIIt, on the column. ~hosl,horibulokill~~se activity remained constant throughout the various stages in the procedure.
p:lrtial sepamtiou of the t.\vo runymes n-as xlso achieved by preparative polyncr~lamidc gel rlwtropl~oresis (Fig. 5). It seemed likely that' tllc origill:il c~~nlples Imd broken dowvll to its compoue~~t parts. 01~ l)iece of evidence in favor of this idea ~vas the rapid movement of t,hc two euzyrnes through this prepnrative gel; it appears Iullikrly that an enzyme that was esc.du&d from Sephadrs G-100 would move through a 7.5C; gel so rapidly.
Also, the activity of the phosphoribulokinuse was much less than that of the fruct,ose bisphosphntase, again suggesting that the comples had broken down to its two components.
Further evidence for the instnbiMy of the comples came from :I comp:trison of tile elution profile from Seplmdes G-200 of the freshly prepared pure complex (Fig. 6Ll) with that of the same protein solution aft,er 48 hours of storage at 4" (Fig. 6B). Again, this experiment showed a loss of a&iv&y of both enzymes (this being particularly marked for phosphoribulokinase) and a change in the position of the peaks in the eluate.
The pure complex did not show any activity of the following enzymes of the reductive pentose phosphate cycle: glycerate 5-phosphate kinnse, glycernldehyde a-phosphate dehydrogenase, trios:e phosplmte isomerase, aldolnse, pentose phosphate iso-by guest on March 23, 2020 http://www.jbc.org/ merase and ribulose bisphosphate carboxylase (all assayed by the methods described by Anderson and Fuller (21)).

Possible
Regulatory Characteristics of Enzyme Complex-Roth phosphoribulokinase and fructose bisphosphatase have been implicated as possible control points in the regulation of photosynthetic carbon dioxide assimilation. We have examined various aspects of the isolated complex in an effort to understand the way in which the activities of the component enzymes might be regulated.
Phosphoribulokinase activity was inhibited by AMP, with Ki values between 1.2 and 1.8 mM (depending on whether the concentration of ribulose 5-phosphate or of ATP is varied) and was activated by 0.05 rnlT NADH. These results agree with those obtained with other autotrophic bacteria (9, 10, 12). The most potent inhibitors of fructose biphosphatase were MgATP2-, 2.6 mM giving 50% inhibition, and magnesium pyrophosphate (MgPz0T2-), 3.3 MM giving 50yc inhibition.
These results agree with those of Morris (8) with the fructose bisphosphatase from spinach chloroplasts. However, our results do not agree with those of Morris (8) and Preiss,Biggs,and Greenberg (7) in that the relationship between reaction rate and fructose bisphosphate concentration is hyperbolic and not sigmoidal as observed with the enzyme from spinach chloroplasts.
Also, the pH optimum of the chloroplast fructose bisphosphatase is altered by increasing Mg++ concentration (7) but there is no such change with the enzyme from R. rubrum.

DISCUSSION
Two enzymes of the reductive pentose phosphate cycle in the photosynthetic bacterium R. rubruwz have been isolated and purified as a complex.
The ultraviolet absorption spectrum of the pure complex shows a peak at 280 nm and is similar to bovine serum albumin.
Also, it appears as a single band on analytical polyacrylamide gels. Furthermore, two enzyme activities remain associated (and show a constant ratio of activities) during several different fractionation techniques. These various observations suggest that the complex is a single protein.
However, we have not made physicochemical measurements of this protein.
Thus, although it appears likely that t,he complex consists of two subunits, each of which is associated with the activity of one of the comporent enzymes, the size of the complex and the precise nature of the association between the subunits remains unknown.
Indeed, its instability would make analysis with (for example) the analytical ultracentrifuge difficult.
Whatever the precise nature of this link between the two enzyme activities, an association between two enzymes not catalyzing consecutive reactions would appear to be of interest. A "multienzyme complex" has been defined as "an orderly association (not involving peptide linkages) of various enzymes that catalyze successive steps in a reaction sequence" (definition of Henning, quoted by Gaertner and De Moss (22)).
The complex described here does not fit in with this concept since the two enzymes do not catalyze successive reactions in the reductive pentose phosphate cycle. Rather, the association reported here more closely resembles that between aspartokinase and homoserine dehydrogenase in Escherichia coli K-12 (23, 24). These two enzymes catalyze the first and third steps in the sequences leading to the synthesis of methionine and threonine. The activities of both aspartokinase and homoserine dehydrogenase are regulated by the end products of the biosynthetic pathways, and it seems probable, therefore, that an association between them has some regulatory function.
Possibly, the association between fructose bisphosphatase and phosphoribulokinase reported here also has some kind of regulatory function. Certainly, both enzymes have been implicat'ed in the control of the reductive pent.ose ph0sphat.e cycle (l-14).