Complete amino acid sequence of rat liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase

The complete amino acid sequence of 6-phospho-fructo-2-kinase/fructose-2,6-bisphosphatase from rat liver was determined by direct analysis of the S-carboxamidomethyl protein. A complete set of nonoverlapping peptides was produced by cleavage with a combination of cyanogen bromide and specific proteolytic enzymes. The active enzyme is a dimer of two identical polypeptide chains composed of 470 amino acids each. The NH2-terminal amino acid residue of the polypeptide chain was shown to be N-acetylserine by fast atom bombardment mass spectrometry of the purified N-terminal tetradecapeptide isolated after cleavage of the intact S-carboxamidomethylated protein with lysyl endoproteinase (Achromobacter protease I). Alignment of the set of unique peptides was accomplished by the analysis of selected overlapping peptides generated by proteolytic cleavage of the intact protein and the larger purified cyanogen bromide peptides with trypsin, Staphylococcus aureus V8 protease, and lysyl endoproteinase. Four nonoverlapping peptides were aligned by comparison with the amino acid sequence predicted from a partial cDNA clone encoding amino acid positions 166-470 of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (Colosia, A.D., Lively, M., El-Maghrabi, M. R., and Pilkis, S. J. (1987) Biochem. Biophys. Res. Commun. 143, 1092-1098). The nucleotide sequence of the cDNA corroborated the peptide sequence determined by direct methods. A search of the Protein Identification Resource protein sequence database revealed that the overall amino acid sequence appears to be unique since no obviously homologous sequences were identified. However, a 100-residue segment of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (residues 250-349), including the active site histidine residue of the bisphosphatase domain, was found to be homologous to the active site regions of yeast phosphoglycerate mutase and human bisphosphoglycerate mutase.     

Amino acid sequence of the phosphorylation site of rat liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase

6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase from rat liver was phosphorylated by cyclic AMP-dependent protein kinase and [gamma-32P]ATP. Treatment of the 32P-labeled enzyme with thermolysin removed all of the radioactivity from the enzyme core and produced a single labeled peptide. The phosphopeptide was purified by ion exchange chromatography, gel filtration, and reverse phase high pressure liquid chromatography. The sequence of the 12-amino acid peptide was found to be Val-Leu-Gln-Arg-Arg-Arg-Gly-Ser(P)-Ser-Ile-Pro-Gln. Correlation of the extent of phosphorylation with activity showed that a 50% decrease in the ratio of kinase activity to bisphosphate activity occurred when only 0.25 mol of phosphate was incorporated per mol of enzyme subunit, and maximal changes occurred with 0.7 mol incorporated. The kinetics of cyclic AMP-dependent protein kinase-catalyzed phosphorylation of the native bifunctional enzyme was compared with that of other rat liver protein substrates. The Km for 6-phosphofructo-2-kinase/fructose-2, 6-bisphosphatase (10 microM) was less than that for rat liver pyruvate kinase (39 microM), fructose-1,6-bisphosphatase (222 microM), and 6- phosphofructose -1-kinase (230 microM). Comparison of the initial rate of phosphorylation of a number of protein substrates of the cyclic AMP-dependent protein kinase revealed that only skeletal muscle phosphorylase kinase was phosphorylated more rapidly than the bifunctional enzyme. Skeletal muscle glycogen synthase, heart regulatory subunit of cyclic AMP-dependent protein kinase, and liver pyruvate kinase were phosphorylated at rates nearly equal to that of 6-phosphofructo-2-kinase/fructose-2, 6-bisphosphatase, while phosphorylation of fructose-1,6-bisphosphatase and 6-phosphofructo-1-kinase was barely detectable. Phosphorylation of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase was not catalyzed by any other protein kinase tested. These results are consistent with a primary role of the cyclic AMP-dependent protein kinase in regulation of the enzyme in intact liver.     

Isolation and characterization of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase from bovine liver

6-Phosphofructo-2-kinase and fructose-2,6-bisphosphatase activities were copurified to homogeneity from bovine liver. The purification scheme consisted of polyethylene glycol precipitation, anion-exchange and Blue-Sepharose chromatography, substrate elution from phosphocellulose, and gel filtration. The bifunctional enzyme had an apparent molecular weight of 102,000 and consisted of two subunits (Mr 49,000). The kinase had a Km for ATP of 12 microM and a S0.5 for fructose 6-phosphate of 150 microM while the bisphosphatase had a Km for fructose 2,6-bisphosphate of 7 microM. Both activities were subject to modulation by various effectors. Inorganic phosphate stimulated both activities, while alpha-glycerolphosphate inhibited the kinase and stimulated the bisphosphatase. The pH optimum for the 6-phosphofructo-2-kinase activity was 8.5, while the fructose-2,6-bisphosphatase reaction was maximal at pH 6.5. Incubation of the purified enzyme with [gamma-32P]ATP and the catalytic subunit of the cAMP-dependent protein kinase resulted in 32P incorporation to the extent of 0.7 mol/mol enzyme subunit with concomitant inhibition of the kinase activity and activation of the bisphosphatase activity. The mediation of the bisphosphatase reaction by a phosphoenzyme intermediate was suggested by the isolation of a stable labeled phosphoenzyme when the enzyme was incubated with fructose 2,6-[2-32P]bisphosphate. The pH dependence of hydrolysis of the phospho group suggested that it was linked to the N3 of a histidyl residue. The 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase from bovine liver has properties essentially identical to those of the rat liver enzyme, suggesting that hepatic fructose 2,6-bisphosphate metabolism is under the same control in both species.     

Binding of ATP to the fructose-2,6-bisphosphatase domain of chicken liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase leads to activation of its 6-phosphofructo-2-kinase

To understand the mechanism by which the activity of the 6-phosphofructo-2-kinase (6PF-2K) of chicken liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase is stimulated by its substrate ATP, we studied two mutants of the enzyme. Mutation of either Arg-279, the penultimate basic residue within the Walker A nucleotide-binding fold in the bisphosphatase domain, or Arg-359 to Ala eliminated the activation of the chicken 6PF-2K by ATP. Binding analysis by fluorescence spectroscopy using 2'(3')-O-(N-methylanthraniloyl)-ATP revealed that the kinase domains of these two mutants, unlike that of the wild type enzyme, showed no cooperativity in ATP binding and that the mutant enzymes possess only the high affinity ATP binding site, suggesting that the ATP binding site on the bisphosphatase domain represents the low affinity site. This conclusion was supported by the result that the affinity of ATP for the isolated bisphosphatase domain is similar to that for the low affinity site in the wild type enzyme. In addition, we found that the 6PF-2K of a chimeric enzyme, in which the last 25 residues of chicken enzyme were replaced with those of the rat enzyme, could not be activated by ATP, despite the fact that the ATP-binding properties of this chimeric enzyme were not different from those of the wild type chicken enzyme. These results demonstrate that activation of the chicken 6PF-2K by ATP may result from allosteric binding of ATP to the bisphosphatase domain where residues Arg-279 and Arg-359 are critically involved and require specific C-terminal sequences.     

Isolation of a cDNA clone for rat liver 6-phosphofructo 2-kinase/fructose 2,6-bisphosphatase

A cDNA clone for rat liver 6-phosphofructo 2-kinase/fructose 2,6-bisphosphatase was isolated from a lambda gt11 rat liver expression library by antibody screening. The clone was approximately 1100 bases in length and the derived amino acid sequence contained 303 residues at the carboxyl end of the subunit. This derived amino acid sequence corresponded exactly with the actual amino acid sequence of the enzyme determined by direct sequencing of the protein.     

Cloning and expression in Escherichia coli of a rat hepatoma cell cDNA coding for 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase.

In liver, the 470-residue bifunctional enzyme 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (PFK-2/FBPase-2) catalyses the synthesis and degradation of fructose 2,6-bisphosphate, a potent stimulator of glycolysis. In rat hepatoma (HTC) cells, this enzyme has kinetic, antigenic, and regulatory properties, such as insensitivity to cyclic AMP-dependent protein kinase and lack of associated FBPase-2 activity, that differ from those in liver. To compare the sequence of the HTC enzyme with that of the liver enzyme, we have cloned the corresponding fully-coding cDNA from HTC cells. This cDNA predicts a protein of 448 residues in which the first 32 residues of liver PFK-2/FBPase-2 including the cyclic AMP target sequence have been replaced by a unique N-terminal decapeptide. The rest of the protein is identical with the liver enzyme. An N-terminally truncated recombinant peptide of 380 residues containing the PFK-2 and FBPase-2 domains was expressed in Escherichia coli as a beta-galactosidase fusion protein. It was recognized by anti-PFK-2 antibodies but its enzymic activities were barely detectable. In contrast, a cDNA fully-coding for the HTC enzyme could be expressed in E. coli as a beta-galactosidase-free peptide that exhibited both PFK-2 and FBPase-2 activities. This peptide had those PFK-2 kinetic properties of the HTC enzyme that differ from the liver enzyme. These data, together with immunoblot experiments, suggest that the lack of associated FBPase-2 activity in HTC cells results from a post-translational modification of the enzyme rather than from the difference in amino acid sequence. As well as this peculiar type of PFK-2/FBPase-2 mRNA, HTC cells also contained low concentrations of the liver-type mRNA. Unlike in liver, neither mRNA was induced by dexamethasone in these cells.     

Inactivation of liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase by o-phthalaldehyde.

The two activities of chicken liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase were inactivated by o-phthalaldehyde. Absorbance and fluorescence spectra of the modified enzyme were consistent with the formation of an isoindole derivative (1 mol/mol of enzyme subunit). The inactivation of 6-phosphofructo-2-kinase by o-phthalaldehyde was faster than the inactivation of fructose-2,6-bisphosphatase, which was concomitant with the increase in fluorescence. The substrates of 6-phosphofructo-2-kinase did not protect the kinase against inactivation, whereas fructose-2,6-bisphosphate fully protected against o-phthalaldehyde-induced inactivation of the bisphosphatase. Addition of dithiothreitol prevented both the increase in fluorescence and the inactivation of fructose-2,6-bisphosphatase, but not that of 6-phosphofructo-2-kinase. It is proposed that o-phthalaldehyde forms two different inhibitory adducts: a non-fluorescent adduct in the kinase domain and a fluorescent isoindole derivative in the bisphosphatase domain. A lysine and a cysteine residue could be involved in fructose-2,6-bisphosphate binding in the bisphosphatase domain of the protein.     

A rat gene encoding heart 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase.

There are at least 3 isozymes of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase, a bifunctional enzyme which catalyzes the synthesis and degradation of fructose 2,6-bisphosphate. A 22-kb rat gene that encodes the heart isozyme has been identified and compared with the 55-kb rat gene encoding the liver and muscle isozymes which had been described earlier. Although these 2 genes include 12 successive similar exons, they contain dissimilar exons at both ends, consistent with the occurrence of different regulatory domains at the N- and C-termini in the 3 isozymes.     

Characterization of distinct mRNAs coding for putative isozymes of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase

Three distinct clones encoding full-length 6-phosphofructo-2-kinase (PFK-2)/fructose-2,6-bisphosphatase (FBPase-2) were characterized from a rat liver cDNA library. Clone 22c was 1859 bp long and coded for the 470 amino acids of the bifunctional subunit of the liver homodimer. This polypeptide is phosphorylated on serine 32 by cyclic-AMP-dependent protein kinase. Clone 4c (2681 bp) had a coding region identical to that of clone 22c but it included a putative intron of 959 bp. In clone 5c (1750 bp), the sequence upstream from amino acid 33 differed from that in clone 22c and coded for a unique N-terminal portion of 10 amino acids. Poly(A)-rich RNA from rat tissues was hybridized with cDNA probes corresponding to the unique N-terminal portions of clones 22c and 5c. Dot and Northern blots showed signals indicative of three distinct PFK-2/FBPase-2 mRNAs. There were a 6.8-kb mRNA typical of cardiac tissue, a 2.1-kb mRNA typical of liver, corresponding to clone 22c, and a 1.9-kb mRNA typical of skeletal muscle, corresponding to clone 5c. Primer extension analysis showed that clones 22c and 5c were nearly complete since their respective 5'-untranslated sequences were at most 96/97 bp and 44 bp shorter than the corresponding mRNAs. These data provide a molecular basis for the existence of PFK-2/FBPase-2 isozymes.     

Induction of rat liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase mRNA by refeeding and insulin

The effects of fasting/refeeding and untreated or insulin-treated diabetes on the bifunctional enzyme 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase and its mRNA in rat liver were determined. Both enzymatic activities fell to 20% of control values with fasting or streptozotocin-induced diabetes and were coordinately restored to normal within 48 h of refeeding or 24 h of insulin administration. These alterations in enzymatic activities were always mirrored by corresponding changes in amount of enzyme as determined by phosphoenzyme formation and immunoblotting. In contrast, mRNA for 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase did not decrease during starvation or in diabetes, but there was a 3-6-fold increase upon refeeding a high carbohydrate diet to starved rats or insulin treatment of diabetic rats. The decrease of the enzyme in starved or diabetic rats without associated changes in mRNA levels suggests a decrease in the rate of mRNA translation, an increase in enzyme degradation, or both. The rise in enzyme amount and mRNA for 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase with refeeding and insulin treatment suggests an insulin-dependent stimulation of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase gene expression. Northern blots of RNA from heart, brain, kidney, and skeletal muscle probed with restriction fragments of a full-length cDNA from liver showed that only skeletal muscle contained an RNA species that hybridized to any of the probes. Skeletal muscle mRNA for 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase was 2.0 kilobase pairs but in contrast to the liver message (2.2 kilobase pairs) was not regulated by refeeding.     

The sugar phosphate specificity of rat hepatic 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase

The sugar phosphate specificity of the active site of 6-phosphofructo-2-kinase and of the inhibitory site of fructose-2,6-bisphosphatase was investigated. The Michaelis constants and relative Vmax values of the sugar phosphates for the 6-phosphofructo-2-kinase were: D-fructose 6-phosphate, Km = 0.035 mM, Vmax = 1; L-sorbose 6-phosphate, Km = 0.175 mM, Vmax = 1.1; D-tagatose 6-phosphate, Km = 15 mM, Vmax = 0.15; and D-psicose 6-phosphate, Km = 7.4 mM, Vmax = 0.42. The enzyme did not catalyze the phosphorylation of 1-O-methyl-D-fructose 6-phosphate, alpha- and beta-methyl-D-fructofuranoside 6-phosphate, 2,5-anhydro-D-mannitol 6-phosphate, D-ribose 5-phosphate, or D-arabinose 5-phosphate. These results indicate that the hydroxyl group at C-3 of the tetrahydrofuran ring must be cis to the beta-anomeric hydroxyl group and that the hydroxyl group at C-4 must be trans. The presence of a hydroxymethyl group at C-2 is required; however, the orientation of the phosphonoxymethyl group at C-5 has little effect on activity. Of all the sugar monophosphates tested, only 2,5-anhydro-D-mannitol 6-phosphate was an effective inhibitor of the kinase with a Ki = 95 microM. The sugar phosphate specificity for the inhibition of the fructose-2,6-bisphosphatase was similar to the substrate specificity for the kinase. The apparent I0.5 values for inhibition were: D-fructose 6-phosphate, 0.01 mM; L-sorbose 6-phosphate, 0.05 mM; D-psicose 6-phosphate, 1 mM; D-tagatose 6-phosphate, greater than 2 mM; 2,5-anhydro-D-mannitol 6-phosphate, 0.5 mM. 1-O-Methyl-D-fructose 6-phosphate, alpha- and beta-methyl-D-fructofuranoside 6-phosphate, and D-arabinose 5-phosphate did not inhibit. Treatment of the enzyme with iodoacetamide decreased sugar phosphate affinity in the kinase reaction but had no effect on the sensitivity of fructose-2,6-bisphosphatase to sugar phosphate inhibition. The results suggest a high degree of homology between two separate sugar phosphate binding sites for the bifunctional enzyme.     

Effect of sulfhydryl modification on the activities of rat liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase

Alkylation of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase with p-mercuribenzoate caused a rapid stimulation of the kinase and an inhibition of the bisphosphatase. At later times, the kinase activity also became inhibited. In contrast, treatment with N-ethylmaleimide abolished kinase activity but had no effect on the bisphosphatase. Selective modification of residues involved in the kinase reaction was also seen with iodoacetamide, which caused a 10-fold stimulation of the kinase Vmax without affecting the bisphosphatase. The stimulatory effect of carboxyamidomethylation was seen when the kinase was assayed in the presence of inorganic phosphate, an allosteric activator of the enzyme. The iodoacetamide-treated enzyme had a 10-20-fold higher Km for fructose 6-phosphate than the native enzyme and the Ki for fructose 2,6-bisphosphate was also increased. However, the adenine-nucleotide site did not seem to be affected since there was no change in the Km for ATP, the Ki for ADP, or the adenine-nucleotide exchange. There was also a direct correlation between the incorporation of [14C]acetamide into the enzyme and activation of the kinase. The residues modified by iodoacetamide were shown to be cysteines by the exclusive appearance of carboxymethylcysteine in protein hydrolysates. Activation was associated with alkylation of 2 cysteines/subunit, of the 12 which could be alkylated after denaturation/reduction. Iodoacetamide-activated kinase was inhibited by ascorbate/Fe3+, which has been shown to modify sulfhydryl groups in the native enzyme, with concomitant loss of kinase activity.     

Contribution of different protein phosphatases to the dephosphorylation of 6-phosphofructo-1-kinase and 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase in rat liver.

The nature of rat liver protein phosphatases involved in the dephosphorylation of the glycolytic key enzyme 6-phosphofructo-1-kinase and the regulatory enzyme 6-phosphofructo-2-kinase/fructose 2,6-bisphosphatase was investigated. In terms of the classification system proposed by Ingebritsen & Cohen [(1983) Eur. J. Biochem. 132, 255-261], only the type-2 protein phosphatases 2A (which can be separated into 2A1 and 2A2) and 2C act on these substrates. Fractionation of rat liver extracts by anion-exchange chromatography and gel filtration revealed that protein phosphatase 2A is responsible for most of the 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase phosphatase activity (activity ratio 2A/2C = 4:1). On the other hand, 6-phosphofructo-1-kinase phosphatase activity is equally distributed between protein phosphatases 2A (2A1 plus 2A2) and 2C. In addition, the possible role of low-Mr compounds for the control of purified protein phosphatase 2C was examined. At near-physiological concentrations, none of the metabolites studied significantly affected the rate of dephosphorylation of 6-phosphofructo-1-kinase, 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase, pyruvate kinase or fructose-1,6-bisphosphatase.     

Molecular cloning, sequence analysis, and expression of a human liver cDNA coding for fructose-6-P,2-kinase:fructose-2,6-bisphosphatase

A cDNA coding for 378 amino acids from the C-terminus of the human liver bifunctional enzyme, Fructose-6-phosphate,2-kinase:Fructose-2,6-bisphosphatase was isolated, sequenced, and expressed in E. coli K38. The expressed protein, identified by specific immunoassay, showed Fru 2,6-bisphosphatase activity but no Fru 6-P,2-kinase activity, demonstrating directly that the Fru 2,6-bisphosphatase activity resides in the C-terminal region. The Km for Fru 2,6-P2 was 4.3 microM. Fru 6-P was a noncompetitive inhibitor (Ki = 2.9 microM), and formed a phosphorylated intermediate when incubated with Fru 2,6[2-32P]P2. The subunit Mr of the enzyme was 36,600, and the active enzyme showed Mr = 37,000 by gel filtration.     

Hormonal regulation of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase: kinetic models

A graphical method to reveal the so-called 'critical fragments' in schemes of biochemical systems is considered. These fragments produce multiple steady states or self-oscillations in systems. As an example, the bifunctional enzyme, 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase, regulated by glucagon through enzyme phosphorylation, is discussed. It is shown that this enzyme may act as a metabolic switching mechanism in discontinuous or oscillatory regimes, depending on the specific structure of its kinetic scheme. The boundaries of concentrational and parameter domains for these critical phenomena are also predicted.