Isolation and Characterization of a Pyrophosphate-dependent Phosphofructokinase from Propionibacterium shermanii*

has from extracts of Propionibacterium shermanii. The catalyzes the transfer of phosphate from pyrophosphate to fructose 6-phosphate to yield fructose-1,6-P, and phosphate. This unique enzymatic activity observed initially in Entamoeba histolytica R. Blytt, and Biol. This is the third pyrophosphate-utilizing enzyme that these two diverse organisms in common. The others are phosphoenolpyruvate and pyruvate The PP,-phosphofructokinase from P. is for fructose-6-P and fructose-1,6-P,, no other phosphorylated sugars were utilized. be replaced by arsenate. The K, are:

. This is the third pyrophosphate-utilizing enzyme that these two diverse organisms have in common. The others are phosphoenolpyruvate carboxytransphosphorylase and pyruvate phosphate dikinase. The PP,-phosphofructokinase from P. shermanii is specific for fructose-6-P and fructose-1,6-P,, no other phosphorylated sugars were utilized. Phosphate could be replaced by arsenate. The K, values are: phosphate, 6.0 x lo-' M; fructose-1,6-P,, 5.1 x 10e5 M; pyrophosphate, 6.9 x 10m5 M; and fructose-6-P, 1.0 x lo-' M. The sZO.,, is 5.1 S. The molecular weight of the native enzyme is 95,000. Sodium dodecyl sulfate electrophoresis of the enzyme showed a single band migrating with an R, corresponding to a molecular weight of 48,000. Extracts of P. shermanii have PP,-phosphofructokinase activity approximately 6 times greater than ATP-phosphofructokinase and 15 to 20 times greater than fructose diphosphatase activities. It is proposed that (a) PP, may replace ATP in the formation of fructose-1,6-P, when the organism is grown on glucose and (b) when the organism is grown on lactate or glycerol the conversion of fructose-1,6-P, to fructose-6-P during gluconeogenesis may occur by phosphorolysis rather than hydrolysis.
The discovery of a pyrophosphate-dependent phosphofructokinase (pyrophosphate; fructose-6-phosphate-l-phosphotransferase) in Entamoeba histolytica by Reeves et al. (1) prompted a search for a similar enzyme activity in Propionibacterium shermanii.
Fructose-6-P + PP, Me2+ fructose-1,6-P, + P, This communication describes the purification and physical and chemical properties of the enzyme from P. shermanii.
The possible role of pyrophosphate in the metabolism of P. shermanii is discussed.  Both pools from Step III have the same specific activities after this procedure.
Some enzyme activity dissolves in the 50 and 45% extractions but has a very low specific activity.

RESULTS
Criteria of Purity-The enzyme from Step 5 of Table  I 1. Polyacrylamide disc gel electrophoresis and sedimentation velocity pattern of the enzyme. A, the gel was prepared using 60 pg of protein as described in Ref. 17 for the standard pH 8.9 system. B, the gel was prepared in sodium dodecyl sulfate as described in Ref. 18 and approximately 30 pg of protein was applied to the gel. C, centrifugation was at 60,000 rpm with a protein concentration of 5.9 mg/ml in 0.1 M TrisCl, pH 7.4. The temperature was 20". The frame was taken 56 min after speed was attained. All techniques were performed on enzyme from Step 5 of Table I. phosphate as a substrate, arsenate will replace phosphate. Although the K, for arsenate appears to be much higher, the V,,, approaches the same value as obtained with phosphate.
Proof that the reaction proceeds as presented in Reaction 1 is given in Table III. There is a one to one ratio for fructose-6-P or PP, utilization and fructose-1,6-P, formation. Reaction Parameters for PP,-phosphofructokinase- The log C versus r2 plot was linear attributing to the purity of the preparation. The molecular weight calculated from the data was 95,000. The subunit molecular weight was determined by electrophoresis in sodium dodecyl sulfate (18). These results are presented in Fig. 2 polypeptide chain. From these observations we conclude that the enzyme is a dimer composed of two subunits of identical molecular weight.
The effect of protein concentration on the sedimentation coefficient was investigated from 100 to 900 pg/ml. There is no effect of protein on the s~,,~ at the concentrations tested. A value of 5.1 was determined for the s&,,,.
Substrates for Growth and Effects on Enzymes Utilizing Fructose-1,6-P,-The activity of enzymes which utilize fructose-1,6-P, was measured in cells grown on various substrates to determine whether the specific activities of the enzymes varied under different conditions of growth. The results are shown in Table V. In the case of the PP,-phosphofructokinase, the highest activity was observed in glycerol  TABLE IV   TABLE V   Kinetic  constants for PP,-phosphofructokinase  Experiments were performed at 20" as described under "Experimental Procedure" with the following exceptions: fructose-6-P concentrations were varied from 0.014 to 0.14 mM; PP, from 0.014 to 0.042 mM; fructose-1,6-P, from 0.0051 to 0.061 mM; and P, from 0.14 to 1.14 mM. Protein was determined by the biuret reaction. and lactate-grown cells. The activity is somewhat lower in glucose-grown cells. The activity of the ATP-phosphofructokinase and fructose diphosphatase was also investigated. There was very little fructose diphosphatase activity regardless of growth substrates. The activity of the PP,-phosphofructokinase under all conditions was 5-to 6-fold higher than the ATPphosphofructokinase and lo-to 20-fold higher than the fructose diphosphatase.
It is difficult to assess the true activity of these enzymes in crude extracts because of our lack of knowledge concerning the optimum assay conditions for the ATP-dependent phosphofructokinase and fructose diphosphatase. The highest activity of these enzymes was observed at pH 8.0 and 8.5, respectively. The addition of AMP had no effect on the activity of the ATP-phosphofructokinase.
There was no change in the activity of these three enzymes upon sedimentation for 1 hour at 124,000 x g. The ATP-phosphofructokinase in Entamoeba histolytica was found to be membrane-bound by Reeves et al. (1). DISCUSSION The discovery of PPi-phosphofructokinase in P. shermanii brings to four the number of enzymes isolated from this organism that are capable of utilizing PP, as a phosphate donor in reactions in which nucleotide triphosp,hates normally participate. The other three enzymes are carboxytransphosphorylase; pyruvate, phosphate dikinase; and pyrophosphate, Lserine phosphotransferase (22). The reactions are given below. Me'+ Fructose-6-P + PP, f__l fructose-1,6-P, Pyrophosphate is produced in many biosynthetic reactions and it is generally accepted that the PP, thus formed is hydrolyzed rapidly in order to provide the driving force for these endothermic reactions. In most instances this is undoubtedly the case, however, there is mounting evidence for the concept that PP, may serve as a high energy phosphate donor in some organisms. There is increasing evidence concerning the role of PP, in energy conservation in the photosynthetic bacteria, It has been demonstrated that these organisms can couple the synthesis of PP, to light-induced electron transport (23). They can also utilize PP, as an energy source for energy-linked transhydrogenation (24,25), cytochrome reduction (26,27) and succinate-linked NAD+ reduction (28). Extracts of R. rubrum also catalyze a PP, * Pi exchange reaction (29).
Perhaps the strongest case for PP, serving as a phosphate donor for substrate level phosphorylation is provided in E.
histolytica, an anaerobic amoeba, which only grows on glucose. The organism apparently lacks pyruvate kinase (30) but has sufficient pyruvate, phosphate dikinase activity, Reaction ii, to account for its glycolytic rate (2). Recently, Reeves et al. (1) have reported the presence of a PP,-phosphofructokinase in this organism. These authors propose that the enzyme functions in a glycolytic capacity forming fructose-1,6-P,. This hypothesis is based on their observations that the organism does not contain sufficient ATP-phosphofructokinase activity to account for its glycolytic flux. Also, the intracellular level of PP, in the amoeba is 0.18 mM which is 13 times greater than the K, of the PPi-phosphofructokinase for this metabolite. high growth efficiency compared to the other bacteria studied by them and they estimated that a net of at least 6 mol of ATP were formed in the fermentation per mol of fermented glucose. Therefore, they suggested that high energy phosphate must arise not only in the usual steps of glycolysis but also during the formation of propionate.
A high energy phosphate might arise during the reduction of fumarate to succinate as indicated in Box G of Fig. 3. Such electron transport-linked ATP formation has been shown in the obligately anaerobic sulfur bacteria (32) but this has not been shown as yet in propionibacteria.
If the propionic acid fermentation occurred by the following equation, 3 Glucose -4 propionate + 2 acetate + 2 CO, per 3 mol of fermented glucose, 6 mol of ATP would be utilized at A and B (if we assume ATP is used in B), 12 would be formed at C and D, and 2 at E, or a net of 2.66 ATP per mol of glucose, without a contribution by G. If there also is formation of ATP during succinate format,ion (G), one ATP would be formed for each propionate or a net total of four ATP per mol of glucose.
We have considered previously (33,34) that the high growth efficiency of the propionic acid bacteria might result from their utilization of the energy from PP,. The finding of a third enzyme in the metabolic pathway which utilizes PP, re-emphasizes this possibility.
It is noteworthy that the amount of ATP-phosphofructokinase found in crude extracts of P. shermanii (0.05 unit/mg of protein, Table V) is very low compared to that of other enzymes of P. shermanii which function in the direct metabolic pathway. Such enzymes are usually present in crude extracts at more than 0.5 unit/mg of protein (15,33). In view of the small amount of the ATP-phosphofructokinase, it appears that the PP,-phosphofructokinase may account for a significant proportion of the glycolytic flux with PPi replacing ATP as the phosphorylating agent. Since there is very little fructose diphosphatase present in these bacteria the PP,-phosphofructokinase also is indicated in Box B to function in the conversion of fructose-1,6-P, to fructose-6-P when the organism is grown on three carbon compounds.
There are several problems which need to be considered in relation to this suggestion. One relates to the source of the PPi for the reaction. Very little PP, would be formed by the pyruvate, phosphate kinase reaction (Box D) which has an activity of about 0.04 unit/mg of protein in crude extracts (35). This low activity is to be expected if its function is primarily anaplerotic as indicated in Fig. 3. The enzyme is particularly important for cells growing on lactate or pyruvate, since the P-enolpyruvate required for anabolic purposes must be synthesized from pyruvate in this case. It is to be noted that extracts of P. shermanii contain pyruvate kinase in sufficient quantity to account for the formation of pyruvate from P-enolpyruvate at the required glycolytic rate. The carboxytransphosphorylase reaction (Box F) likewise serves an anaplerotic function, since the oxalacetate of the main metabolic pathway is supplied by the transcarboxylase reaction and the carboxytransphosphorylase is only required to replenish C,-dicarboxylic acids which are withdrawn from the cycle when succinate is a fermentation product and for synthesis of aspartate and other compounds. The crude extract contains 0.1 unit of carboxytransphosphorylase/mg of protein (36). Thus, the scheme shown in Fig. 3 does not provide a major source of PP, unless there is formation of PP, during the reduction of fumarate to succinate (Box G of Fig. 3). It is possible that PP, may be generated in this step by an electron transport-coupled phosphorylation linked to the reduction of a flavoprotein by NADH and the reduced flavoprotein may then be reoxidized in the reduction of the fumarate to succinate. A coupled synthesis of PP, during light-induced electron transport of photosynthetic bacteria (23) provides some precedence for such consideration.
It is to be noted that the PPi-phosphofructokinase reaction provides a means of salvaging the bond energy of the PP, which arises during the synthesis of fats, carbohydrates, proteins, and nucleic acids. This may, in part, account for the highly efficient growth of the propionic acid bacteria as compared to some other microorganisms (31). A second problem relative to the proposal that PP,-phosphofructokinase may act in both the synthesis and breakdown of fructose 1,6-diphosphate is the mechanism of control. In mammalian cells, the ATP-phosphofructokinase and fructose diphosphatase reactions provide different routes for glycolysis and gluconeogenesis which are controlled reciprocally by allosteric effecters (37). This control prevents futile cycles which would result in an ATPase-like activity by the two enzymes. Thus far, no allosteric inhibitors of PP,-phosphofructokinase have been found. In fact, if the PP,-phosphofructokinase serves in both directions, allosteric inhibition would not be expected. It is conceivable that the concentration of metabolites and the thermodynamic properties of the reaction are such as to permit the necessary control and to prevent futile cycling. The equilibrium of the PP,-phosphofructokinase reaction favors the formation of fructose-1,6-P, and is strongly dependent on divalent metals. 1 There probably are controls of the propionic acid fermentation at other points (9). If there were no specific controls of the PP,-phosphofructokinase the flux through fructose-1,6-P, would depend upon the kinetic properties of the enzyme and the intracellular concentrations of the substrates and products of the reaction.
The PP,-phosphofructokinase from P. shermanii does not appear to have any structural similarities either with the mammalian or bacterial fructose diphosphatases (38) which are generally larger and composed of four identical subunits nor with the ATP-phosphofructokinases (39). It will be inter-esting to compare the mechanism of action of the PP,-phospbofructokinase with these other enzymes. The PP,-phosphofructokinase from propionibacteria may be very similar not only chemically but also structurally with the corresponding enzyme isolated from E. histolytica.
Such comparisons must await further investigations of these enzymes. It is tempting to speculate about the evolutionary significance of finding such unique enzyme activities as pyruvate, phosphate dikinase, carboxytransphosphorylase, and now PP,phosphofructokinase in such diverse organisms as an amoeba and a bacteria. It is possible that these two organisms evolved from a common ancestor, one that utilized PP, instead of nucleotide triphosphates as its high energy phosphate donor. Alternatively, these organisms might have evolved independently but in similar environment in which PP, was abundant and hence the enzyme systems were developed for its utilization. Although the possibility of comparing the amino acid sequences of these enzymes from the two organisms is not presently feasible, a comparison of their immunological properties might provide evidence concerning the evolutionary hypotheses.