Pyrophosphate-dependent phosphofructo-1-kinase complements fructose 1,6-bisphosphatase but not phosphofructokinase deficiency in Escherichia coli

The gene from Propionibacterium freudenreichii for PPi-dependent phosphofructo-1-kinase, an enzyme that is found in some bacteria, in a number of anaerobic protists, and in plants, complements the absence of fructose 1,6-bisphosphatase in Escherichia coli but does not complement the deficiency of the ATP-dependent phosphofructokinase.

PFKs as well as a distinct FbPase. The three activities producing or degrading fructose 1,6-bisphosphate are summarized in Fig. 1. Mertens (7) has recently reviewed the properties of PPi-dependent PFKs and has suggested that their role is that of a glycolytic enzyme adapted to anaerobic conditions. On the other hand, other investigators (3,10) have pointed out the reversibility of the PPi-dependent PFK and have suggested that, at least in plants, it may function in the reverse direction, that is, as an FbPase. The overall free energy change for reaction 1 in Fig. 1 is about 2 kcal (-8 kJ) less than that for the ATP-dependent enzyme (reaction 2), a fact that permits the consideration of reversibility (7). Furthermore, if the ratio of Pi to PPi is maintained at a high level by a pyrophosphatase, the reverse reaction of the PP1dependent enzyme would be favored.
Our laboratory has recently cloned and expressed in (T2R) reL41 pit-10 spoTi (fbp)287] cannot grow on minimal medium containing glycerol and succinate as carbon sources. In addition, it was shown that the Fbp phenotype could be screened on MacConkey-glycerol indicator plates (9). Fbp+ strains form red colonies, whereas Fbp-strains form pale, yellow colonies. Table 1 shows the results of attempts to complement PFK in the strain lacking PFK, DF1020, and to complement FbPase in the strain lacking FbPase, DF657. As shown previously (1), DF1020 could grow on minimal medium supplemented with mannitol when transformed by a vector containing ATP-dependent PFK, by using a construct (pRZ3) containing E. coli PFK (11). In contrast to the results obtained with ATP-dependent PFK, no growth was seen in mannitol minimal medium when the transformation was carried out with pLG1, which consists of pBluescriptII into which has been inserted the gene for PPi-dependent PFK from P. freudenreichii along with 160 bases upstream and 1,100 bases downstream of the coding sequence (5). On the other hand, FbPase-deficient bacteria transformed with a plasmid bearing PPi-dependent PFK but not ATP-dependent PFK were capable of growing on succinate-glycerol medium and produced red colonies on MacConkey-glycerol indicator plates. IPTG (isopropyl-13-D-thiogalactopyranoside), which induces higher levels of PPi-dependent PFK activity with pLG1 transformed cells, did not influence growth on minimal medium (data not shown).
The failure of PP,-dependent PFK to permit growth on mannitol indicates that the reaction does not function efficiently in the forward direction, that is, toward formation of fructose bisphosphate, in E. coli. The most likely explanation for this failure is that the amounts of PPi are too small.  'pBluescriptll is the parent plasmid of pLG1.

While PP1 is produced in a number of biosynthetic steps, it is
presumably maintained in very small amounts by pyrophosphatase, an enzyme present in most, if not all, organisms. While a systematic comparison of pyrophosphatases in organisms with either ATPor PPi-dependent PFKs has not been carried out, a few limited studies suggest that organisms with a PPi-dependent enzyme may have very low pyrophosphatase activity. Trichomonas vaginalis and Entamoeba histolytica are two protists that have a PPi-dependent PFK and that have no detectable metal ion-dependent pyrophosphatase and very little total pyrophosphatase activity (6,8). On the other hand, E. coli has very high levels of pyrophosphatase activity but apparently continues to maintain a significant amount of PP1 (4). Kukko-Kalske et al. (4) have measured PPi in exponentially growing E. coli cells and have reported concentrations of about 0.5 mM. This, one would presume, would be sufficient to permit phosphorylation of fructose 6-phosphate by the PPi-dependent enzyme.
On the other hand, the same investigators have reported that the PPi concentration dropped dramatically to below the detection limit of the assay when glucose was exhausted from the medium. Danchin et al. (2) noted a rapid drop in the concentration of PPi when the energy metabolism of E. coli was altered by the addition of 2-ketobutyrate, suggesting that a constant source of energy is necessary for the main-tenance of PPi levels. It might well be assumed that in the present experiments, the concentration of PPi in E. coli in minimal medium in the presence of mannitol was never sufficient to permit glycolysis.
The data clearly demonstrate that the PPi-dependent PFK can function as an FbPase both in minimal media and on the nutritionally more-complete MacConkey agar. One must assume that the concentration of Pi is very high relative to that of PPi and that, in the absence of any other source of fructose 6-phosphate, reaction 1 (Fig. 1) will operate in the reverse direction. Under these conditions, the kinase, despite an unfavorable equilibrium constant, can perform the role of a gluconeogenic enzyme. This work was supported by Public Health Service grant DK19912 from the National Institute of Diabetes and Digestive and Kidney Diseases.