Pyrophosphate :D-Fructose 6-Phosphate l-Phosphotransferase FUNCTION OF 6-PHOSPHOFRUCTOKINASE*

Abstract The major enzyme of Entamoeba histolytica which produces fructose 1,6-diphosphate from fructose 6-phosphate utilizes inorganic pyrophosphate as phosphate donor. In its reverse reaction the enzyme utilizes orthophosphate and fructose diphosphate to yield inorganic pyrophosphate and fructose 6-phosphate. The enzyme has been purified 100-fold from an amebal homogenate. At pH 7 with 2.5 mm MgCl the following Km values were observed: pyrophosphate, 14 µm; fructose 6-phosphate, 38 µm; fructose 1,6-diphosphate, 18 µm; and orthophosphate, 800 µm. At this pH and magnesium ion concentration the maximum velocity of the reaction was approximately the same in each direction. The average intracellular concentration of pyrophosphate in the ameba is 0.18 mm, a value 13-fold greater than the pyrophosphate Km for the new enzyme. This enzyme raises to three the number of known enzymes in the glycolytic pathway of E. histolytica which utilize or produce pyrophosphate. The trivial name proposed for the new enzyme is 6-phosphofructokinase (PPi).

The assayed activity of the conventional ATP-utilizing 6-phosphofructokinase (EC 2.7

IdentiJication of Two 6-phosphojructokinases
in Ameba-The total activity of the ATP-dependent enzyme in amebal homogenates is influenced by the method of preparation of the homogenate while that of the PPi-dependent enzyme is not. To achieve the maximum yield of the ATP enzyme, 1.25 ml of fresh cells were suspended in 5 ml of 20 mM a-glycerol phosphate, pH 7, containing 2 mM dithiothreitol, and the cells were ruptured by 15 strokes with the close-fitting pestle of a Dounce homogenizer (Kontes Glass Co.). After centrifugation for 30 min at 36,000 x g the supernatant solution contained 3.1 and 28 units of the ATP and PPi enzyme, respectively.
The fractionation of this supernatant solution on a column of Bio-Gel P-300 is shown in Fig. 1. The results suggest that the two enzyme activities belong to distinct proteins.
Evidence That Two Enzyme Activities Represent Distinct Pro- teins-In addition to the evidence shown in Fig. 1 we have made other preliminary observations which indicate that the activities with ATP and PPi are due to distinct proteins.
The ATP enzyme is sedimented by 4 hours centrifugation at 100,000 x g while the PP i enzyme is not. The latter activity is undiminished by lyophilization of whole cells while the former is completely destroyed by this treatment.
The ATP enzyme activity is not present in cell homogenates prepared by vigorous sonication while the PPi enzyme is not diminished by this treatment. When a fraction from a Bio-Gel l-'-300 column similar to Fraction 5 of Fig. 1 was subjected to sonication for 15 s with the microtip of a Bronson Sonifier at power setting 5, the ATP activity was reduced by 84oj, while the PPi activity remained unchanged.
Purification of PP&ilizing Enzyme-Nine-tenths milliliter of fresh cells was suspended in 3.6 ml of the cr-glycerol phosphatedithiothreitol buffer and treated, with cooling, for a total of 30 s with the microtip of a Bronson Sonifier at power setting 5. The ATP enzyme could not be detected in this sonicated homogenate. After centrifugation at 36,000 x g for 30 min the PPi-utilizing enzyme in the supernatant solution was applied to the same Bio-Gel P-300 column ( Fig. 1) and eluted as described above. The elution pattern of this enzyme was sharper than that obtained with the Dounce homogenate, possibly because of the more effective cell disruption provided by sonication. The enzymerich fractions from this column were combined and applied to a column containing hydroxylapatite (Bio-Gel HT), bed volume 35 ml. Enzyme was eluted with a linear gradient formed by placing 50 ml of 0.4 M potassium phosphate in 20 mM imidazole-HCl, pH 7, in the reservoir and 50 ml of the imidazole buffer at the same pH in the mixing chamber.
Enzyme appeared in the effluent solution after 56 ml of the gradient had entered the column.
The peak fractions were combined and dialyzed overnight against two 500.ml portions of the imidazole buffer. The dialyzed solution was applied to a column of DEAE-cellulose, bed volume 12 ml, which had been equilibrated with the buffer. Enzyme was eluted from this column by a linear gradient formed with 50 ml of 0.4 M NaCl in the imidazole buffer in the reservoir and 50 ml of the buffer in the mixing chamber.
Enzyme activity eluted with the first protein peak. The two fractions with greatest activity were combined and concentrated to 3.5 ml by vacuum dialysis against 150 ml of the imidazole buffer. This solution was then applied to a column containing 112-ml bed volume of Sephadex G-100 and enzyme was eluted with the same buffer. The ratio of enzyme activity to Azso was equal in the two fractions which comprised the enzyme peak. Fractions were assayed for glucose phosphate isomerase throughout the purification and selected for the subsequent step with a view to excluding this activity as far as possible. This resulted in the deliberate discarding of much of the phosphofructokinase in order to achieve low isomerase activity in the final product.
Enzyme from the DEAE-cellulose column contained isomerase activity to about 1% that of the kinase and this contamination was cut in half by the subsequent fractionation on Sephadex.
The purification procedure is summarized in Table I.
Stoichiometric Experiment-The results of an experiment shown in Table II clearly reflect the stoichiometric conversion of fructose 6-phosphate and inorganic pyrophosphate to fructose diphosphate and orthophosphate during incubation. Glucose 6phosphate and triose phosphate were separately assayed in this experiment and their quantities were found to have increased by only 0.1 pmol each during the 15.min incubation.
Kinetic Studies-The effects of substrate concentrations upon the initial velocity of the forward and reverse reactions were de-r termined. The enzyme used for these studies was purified through the DEAE-cellulose column stage, but in this preparation a 36-ml bed volume of DEAE-cellulose was employed. The enzyme solution contained no detectable amount of aldolase, less than 0.5% glucosephosphate isomerase, and its specific activity was 28.
That hyperbolic kinetics was obeyed with each of the forward reaction substrates is indicated by the linear plots in double reciprocal coordinates shown in Fig. 2. From Fig. 2A we calculate the K, for PPi to be 14 C(M and we note that the true I',,,,, is about 0.25 pmol of fructose diphosphate formed per min per ml of the enzyme solution.
From Fig. 2B we calculate the Km for fructose 6-phosphate to be 38 PM.
The same enzyme solution was used to study the kinetics of the reverse reaction.
Again, hyperbolic kinetics was obeyed by each substrate. Fig. 3, A  Reaction conditions were those of the standard assay ("Experimental Procedure") except for substrate concentrations, and that the reaction was started by the addition of fructose g-phosphate instead of enzyme. The fixed micromolar concentrations of fructose g-phosphate are shown by numbers above the respective curves.
The ordinate values are the reciprocal of micromoles of fructose diphosphate formed per min per ml of the enzyme solution. B, the same data with fructose g-phosphate (F6P) as the varied substrate.
The fixed micromolar concentrations of PPi are shown as numbers above each respective curve. The ordinate values are as in A. The broken curves of A and B are plotted from the ordinate intercepts of B and A, respectively.
PPi and when that point was reached reaction ceased. GTP was tested at 0.25 mM concentration, the others at 0.5 and 1 mM. Fructose l-phosphate was not a substrate for the enzyme. Glucose 6-phosphate reacted with crude enzyme which was heavily contaminated by isomerase, but enzyme from the DEAE-or Sephadex columns failed to react appreciably when this substance was substituted for fructose 6-phosphate in the standard assay. In conducting these experiments, a no enzyme control is essential because the assay enzyme system contains a small amount of glucosephosphate isomerase. That added Pi is essential for the reverse reaction is illustrated by the curves or Fig. 4. Curves  The enzyme solution was the same as that used in Fig. 2. A, Pi as the varied substrate.
Cuvettes contained 50 mM imidaaole-HCI, pH 7, 2.5 mM MgC12, 0.5 mM TPN, 4 pg of glucose&ph0sphat.e dehydrogenase, 2 pg of glucosephosphate isomerase, potassium phosphate, pH 7, as indicated on the abscissa, fructose 1,6-diphosphate, and 20 ~1 of enzyme solution in a final volume of 0.4 ml. Reaction was initiated by the addition of enzyme. The micromolar concentrations of fructose diphosphate are indicated by numbers above the respective curves. The ordinate is the reciprocal of micromoles of fructose 6-phosphate formed per min per ml of the enzyme solution.
B, the same data with fructose 1,6-diphosphate @'Pa) as t.he varied substrate. The fixed micromolar concentrations of Pi are shown as numbers above each respective curve. Ordinate values are as in A. The broken curves in A and B are plotted from the ordinate intercepts of the solid line curves of B and A, respectively. which the assayed product was PPi. It is evident that this product was not formed at a significant rate until Pi w&s added to the reaction mixture.
After the addition of Pi, all of the fructose diphosphate contained in cuvette B was recovered as PPi. The slight activity noted before the addition of Pi is believed to be due to traces of Pi in the rather complex assay system. Curves C and D record an experiment in which fructose 6-phosphate was the assayed product.
It was not formed until Pi was added to the reaction mixture.
Almost all of the fructose diphosphate in cuvette D was recovered as fructose 6-phosphate within the time span of the experiment.
MetaboZite Concentrations in Amebu-Cells harvested from growth medium were taken up in medium lacking serum, incubated anaerobically 30 min at 37", sedimented, and quickly washed twice with a balanced salt buffer which had been bubbled with nitrogen containing 5% CO*. Supernatant fluid was re- Cuvettes contained 50 mM imidazole-HCl, pH 6.8,5 mM MgCle, 1.25 mM phosphoenolpyruvate, 1 mM AMP, 20 mM ammonium chloride, 0.2 mM DPNH, 13 pg of lactate dehydrogenase, 0.3 unit of pyruvate, orthophosphate dikinase, enzyme, and water to a final volume of 0.4 ml. Cuvette B contained, in addition, 15 nmol of fructose diphosphate. After 4 min (heavy arrow), 10 ~1 of 0.2 M potassium phosphate, pH 7, were added to A and B. Curves C and D record an assay for fructose 6-phosphate formation.
After 4 min (heavy arrow), 10 ~1 of the potassium phosphate solution were added to C and D. The curves of this figure have been corrected to allow for dilution by the added potassium phosphate solution.
The ordinate represents total nanomoles of DPNH or TPNH in each cuvette.
moved from the final cell pellet and cold 6% perchloric acid was stirred into the pellet with a thin glass rod. The cells were disrupted by two 15-s treatments, with cooling, with the microtip of a Bronson Sonifier at power setting 5. After centrifugation for 20 min at 36,000 X g the supernatant fluid was neutralized and assayed for metabolites by methods described under "Experimental Procedure." The assayed concentrations were corrected to intracellular concentration by taking the value 0.80 to represent the ratio of intracellular fluid to packed cell volume. Average intracellular concentrations from four or five assays on each metabolite are as follows, in millimolar concentration with standard deviation: PPi, 0.18 =t 0.02; fructose diphosphate, 0.13 f 0.04; fructose 6-phosphate, 0.16 f 0.06; and ATP, 0.65 f 0.14. The average of two closely agreeing determinations on orthophosphate (in cells washed twice with phosphate-free buffer) was 2.8 mM.

DISCUSSION
The pyrophosphate-dependent enzyme was first observed acting in its reverse direction in a phosphate-containing buffer. We saw evidence of what appeared to be a specific fructose diphosphate 1-hydrolase with great affinity for its substrate.
This was puzzling for two reasons: ameba grow only when glucose is supplied and have no apparent need for a gluconeogenic pathway; secondly, this anaerobic organism could ill afford the energy dis-sipation implicit in such great hydrolase activity.
Later work proved that the supposed hydrolase activity was that of the new enzyme which, in the reverse direction, produced PPi and fructose 6-phosphate from orthophosphate and fructose 1,6-diphosphate.
This enzyme is now called 6-phosphofructokinase (PPi) to distinguish it from the ATP-utilizing 6-phosphofructokinase also present in ameba. ma1 activity. 2 These findings together with the observed intracellular metabolite levels imply that the physiological activity of the ATP enzyme would be much lower than the glycolytic flux in ameba. On the other hand, the 6-phosphofructokinase (PPi) could account for the required net synthesis of fructose diphosphate.
Among the substances tested this enzyme is specific for fructose 6-phosphate and PPi in the forward reaction.
Each of the four common nucleoside triphosphates failed to show activity as phosphate donors.
In the reverse reaction both Pi and fructose diphosphate are essential in order for the reaction to proceed.
We have not found prior reference to a phosphofructokinase which employs PP; as phosphate donor.
Bar-Tana and Cleland (6) recently found that MgPPi and MgPPPi are not substrates, but are inhibitors of rabbit muscle phosphofructokinase.
They were competitive against ATP and had Ki values of less than 1 rnM.
This enzyme raises to three the number of reversible enzymes utilizing inorganic pyrophosphate that are known to operate in the glycolytic pathway of E. histolylica (8)(9)(10).
The other two are pyruvate, orthophosphate dikinase and phosphoenolpyruvate carboxykinase (pyrophosphate) (EC 4.1.1.38). The repeated substitution of pyrophosphate for a nucleoside triphosphate suggests that amebal metabolism may have diverged very early from the evolutionary pattern of development followed by most other investigated organisms.
The physiological significance of these pyrophosphate-utilizing enzymes is emphasized by our observation that pyrophosphate is present in ameba at a level comparable to that of other metabolites. Wood, Davis, and Lochmtiller (7) studied the free energy of hydrolysis of the PPi bond. Their results show the apparent value, based on sums of all ionic species, to vary markedly with magnesium ion concentration.
They report AFlanai to be -8.0 f 0.3 Cal per mol in the absence of magnesium and -4.5 f 0.3 Cal per mol in the presence of 5 mM magnesium ion. In view of this it is probable that the reaction catalyzed by the enzyme described here will be profoundly influenced by magnesium ion concentration, particularly the ratio of rates of reaction in the forward versus reverse directions.