6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase
Created by Jessica Szelc
The breakdown and production of glucose for metabolic energy is essential to nearly all organisms, and glucose homeostasis is maintained by two opposing processes: glycolysis and gluconeogenesis. Glycolysis is a ten-step, catabolic pathway in which one molecule of glucose is broken down to two molecules of three-carbon pyruvate. Glycolysis results in the net production of two molecules of ATP, while pyruvate is further utilized in aerobic catalytic pathways to synthesize additional ATP. Gluconeogenesis is the generation of glucose from non-carbohydrate precursors. It occurs when the cellular energy level is high and results in the storage of glucose (1). It is important to note that gluconeogenesis is not simply the reverse of glycolysis. However, the two pathways are under reciprocal control and must be regulated synchronously in order to avoid the useless cycling of glucose, ATP, NADH, and other metabolites (2).
6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase (1BIF) is a bifunctional enzyme found in the testes of Rattus norvegicus that is indirectly involved in maintaining glucose homeostasis (2). The enzyme has a molecular weight of 53,998.70 Da and an isoelectric point (pI) of 6.07. 6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase (6-PF-2-K/F-2,6-BPase) regulates the availability of the molecule fructose-2,6-bisphosphate (F-2,6-BP), a key allosteric regulator of glycolysis and gluconeogenesis (3). 6-PF-2-K/F-2,6-BPase synthesizes and degrades F-2,6-BP via the opposing kinase and phosphatase activities of the two distinct domains of the protein (4). In glycolysis, the enzyme phosphofructokinase catalyzes the conversion of fructose-6-phosphate to fructose-1,6-bisphosphate. The reverse occurs in gluconeogenesis and is catalyzed by fructose-1,6-bisphosphatase. Fructose-2,6-bisphosphate functions to both activate phosphofructokinase and inhibit fructose-1,6-bisphosphatase, and therefore plays a role in regulating the balance between glycolysis and gluconeogenesis (1).
6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase is a homodimer composed of subunits of 55 kDa each. Each monomer is divided into two functional domains that produce the antagonistic activities of the enzyme. The N-terminal catalytic domain houses the 6-phosphofructo-2-kinase (6-PF-2-K) activity, while the C-terminal domain houses the fructose-2,6-bisphosphatase (F-2,6-BPase) activity. The location of the 6-PF-2-K domain was determined by the localization of mutations that affect kinase activity as well as the presence and location of ATPγS upon crystallization of the enzyme. The location of the F-2,6-BPase domain was indicated by sequence and structural homology with yeast phosphoglycerate mutase and rat acid phosphatase. The 6-PF-2-K domain synthesizes fructose-2,6-bisphosphate from fructose-6-phosphate and ATP, while the F-2,6-BPase domain hydrolyzes fructose-2,6-bisphosphate to form fructose-6-phosphate and inorganic phosphate (2).
The 6-PF-2-K domain extends from residue 1 to residue 250. It consists of a central six-stranded β-sheet with five parallel strands (β1-1 to β1-5) followed by one anti-parallel strand (β1-6) at the outermost edge of the sheet. The β-sheet is surrounded by seven α-helices (α1-α7). The 6-PF-2-K domain has two conserved sequence motifs. A Walker A motif (G-X-X-X-X-G-K-T) is found at the C-terminus of β1-3, residues 45-52. This motif is a nucleotide binding fold and has the sequence Gly-Leu-Pro-Ala-Arg-Gly-Lys-Thr in the kinase domain (2). Secondly, a Walker B motif (Z-Z-Z-D) is found from residues 124-128 of the β1-2 strand with the sequence Val-Ala-Val-Phe-Asp (4). The F-2,6-BPase domain extends from residue 251 to residue 468. It also has a central six-stranded β-sheet; however β-strands β2-1, β2-2, β2-3, β2-4 and β2-6 are parallel while β2-5 is anti-parallel. The β-sheet is surrounded by eight α-helices (α8-α15). Four of the helices form an α-helical subdomain that covers the active site of the molecule (2). Each monomer has both a 6-PF-2-K domain and a F-2,6-BPase domain. A hairpin turn between the β1-5 and β1-6 strands of the 6-PF-2-K domain form hydrophobic interactions with the α7 helix, which in turn interacts with the α13 helix of the F-2,6-BPase domain; these interactions form the interface between the two domains (5). In the dimer, the two 6-PF-2-K domains associate in a head-to-head fashion with the β-sheet of one kinase domain forming an anti-parallel β-strand interaction with the β-sheet of the other kinase domain across the twofold axis of rotation. This connection results in a continuous twelve-stranded, intermonomer β-sheet. In addition, there are several interdigitated side chains resulting from contact between the α7 helix of each monomer. The two F-2,6-BPase domains remain effectively independent, and are connected by a single salt bridge. Two-thirds of the surface area buried in the dimer interface is located between the kinase domains; this implies that the enzyme evolved as a 6-PF-2-K homodimer with independent bisphosphatase domains tethered to the kinase dimer (2).
The 6-PF-2-K domain catalyzes the transfer of the γ-phosphate of ATP to the 2-hydroxyl group of fructose-6-phosphate to synthesize fructose-2,6-bisphosphate. Structural homology indicates that the 6-PF-2-K domain is functionally related to the family of nucleoside monophosphate kinases (NMP kinases) as well as to the catalytic core of the G proteins (2). The catalytic mechanism of all three enzymes involves the transfer of the γ-phosphate of a nucleoside triphosphate to an acceptor molecule. In addition, the Walker A and Walker B motifs are conserved among the enzymes (4). In the kinase domain, the Walker A nucleotide binding fold (residues 45-52) functions as the ATP-binding loop. The N-terminus of the α1 helix and the backbone nitrogen atoms of residues 48-53 form an anion hole that serves as a phosphate-binding pocket. Multiple hydrogen bonds are formed between the phosphate oxygen atoms and the peptide backbone within the pocket (2). The conserved threonine (Thr-52) of the Walker A motif and the conserved aspartate (Asp-128) of the Walker B motif coordinate to the Mg2+ ion associated with the β- and γ- phosphates of ATP (4). The adenosine binding pocket mainly involves Van der Waals interactions between adenine and the nonpolar side chains Cys-158, Val-163, Val-220, and Val-246, while the ribose ring interacts with Thr-130 (6). A single hydrogen bond is formed between N6 of adenine and the Oδ oxygen of Asn-167. The location of the fructose-6-phosphate (F6P) binding site was first inferred by superimposition of the substrate pocket of the NMP kinases with the 6-PF-2-K domain. Mutations that affect F6P binding correspond to a region of the 6-PF-2-K structure (Arg-102, Arg-136, and Arg-193) that is similar to the NMP binding site and is adjacent to the ATP binding site of the kinase domain (2). The location of the F6P binding site has been further elucidated by localization of a succinate molecule in the F6P binding loop. The succinate molecule binds to Arg-78, and site-directed mutagenesis has shown that replacement of this residue with histidine or leucine results in a 200-fold increase in the Km for F6P. The binding of octyl β-D-glucopyranoside has also been used to determine the F6P binding site. The glucose ring of β-OG interacts with Arg-102 and Arg-136, and it is proposed that these residues interact with the 6-phosphate of F6P. In the domain, the ATP and F6P binding loops are linked to form a stable structure by way of a salt bridge between Asp-177 of the ATP binding loop and Arg-78 of the F6P binding loop. There is evidence that the interaction between Asp-177 and Arg-78 aids in the reorganization of the ATP binding loop upon ATP binding in a way that promotes F6P binding (4). A final structural similarity between the NMP kinases and the 6-PF-2-K domain is the presence of two mobile segments which close upon substrate binding. The associated conformational change is proposed to function as a way to exclude solvent in order to prevent ATP hydrolysis. These segments of the kinase domain extend from residues 74-100 and residues 159-194 and trap ATP and F6P, respectively (6).
The catalytic mechanism of the NMP kinases and the G proteins is characterized by the lack of a substrate-activating nucleophile and stabilization of the negative charge that accumulates on the β-γ bridge oxygen of GTP/ATP in the transition state (2). In the 6-PF-2-K domain, the positively charged residue Lys-172 is positioned so as to stabilize such a negative charge on the β-γ bridge of ATP (4). This fact along with the lack of a nucleophile near the active site suggests that the kinase reaction operates via transition state stabilization as well (2). Upon ATP binding, Lys-172 is recruited to the active site by its energetically favorable interaction with the β-γ bridge oxygen. This is accomplished by the partial unwinding of the α5 helix. After the γ-phosphate is transferred to F6P, Lys-172 is no longer bound and the helix re-winds. Val-171 is substituted for Lys-172 in the active site, which promotes the dissociation of ADP. Turnover rates for the mechanism are longer due to the lack of an activating nucleophile. However this is acceptable as characterization of the 6-PF-2-K domain as a slow catalyst is consistent with its role as a regulatory enzyme (4).
The F-2,6-BPase domain is responsible for the degradation of fructose-2,6-bisphosphate to fructose-6-phosphate and inorganic phosphate. The bisphosphatase domain shares significant structural homology as well as a common catalytic mechanism with two other families of enzymes: the cofactor-dependent phosphoglycerate mutases and the acid phosphatases. Each of these enzymes has a central β-sheet surrounded by α-helices, in addition to an α-helical subdomain that functions as a lid over the active site. This subdomain consists of residues 325-363 and residues 443-468 and contributes side chains that are involved in substrate binding. The catalytic center of all three enzymes is defined by a catalytic histidine residue located in a conserved sequence motif: Z-X-R-H-G-E/Q-X-X-X-N. This motif is seen in residues 255-262 of the F-2,6-BPase domain. The 2-phosphate of fructose-2,6-bisphosphate (F-2,6-BP) binds to the catalytic histidine, while the 6-phosphate of F-2,6-BP is bound to side chains that are part of the α-helical subdomain, including Arg-350, Lys-354, Gln-391, Tyr-336, Tyr-365, and Arg-395. In the reaction mechanism, the catalytic histidine His-256 acts as a nucleophile and attacks the 2-phosphate of F-2,6-BP causing an inversion of the phosphate geometry. In the crystal structure, Nε of His-256 is shown to be directly in line with a P-O bond. The resulting transition state, a covalent phosphohistidine intermediate, is oriented and stabilized by salt bridges as well as hydrogen bonds involving the side chains of three conserved residues: Arg-255, Asn-262, and Arg-305 (2). Arg-255 and Arg-305 interact with the reactive phosphate group on the 2-carbon of F-2,6-BP; Arg-255 stabilizes the phosphate group in the ground state while Arg-305 stabilizes the phosphate group in the transition state (7). A second histidine residue His-390 exists in a protonated state and acts to stabilize the transient negative charge on the leaving group oxygen of the transition state, allowing the product fructose-6-phosphate to dissociate in its unprotonated form. Finally, a water or hydroxide molecule attacks the phosphohistidine intermediate to regenerate protonated His-390 and produce inorganic phosphate (Pi) (2). It has been proposed that Glu-325 serves as a general acid and maintains the protonated state of His-390; however the crystal structure shows that these two residues are too far apart to support this conclusion. Instead, it was found that Glu-325 caps the α14 helix and could only interact with His-390 if substrate binding resulted in a structural rearrangement (7). Finally, the role of the conserved asparagine residue (Asn-262) has not been determined experimentally, however its proximity to the 2-phosphate binding site suggests that it aids in substrate binding or transition state stabilization (2).
The use of online bioinformatic resources such as the BLAST search engine and the Dali server can further elucidate sequence and structural homologies between proteins. A BLAST search is used to compare the primary structures of two proteins, and returns a quantitative E value that indicates sequence homology. The smaller the E value, the greater the homology; an E value of less than 0.05 is considered significant. The Dali server compares intramolecular distances to determine similarities between protein tertiary structures. It measures homology as a quantitative Z score; a Z score of greater than 2 is considered significant. These two tools were used to analyze sequence and structural homology between 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase and related proteins. Two comparison proteins are highlighted due to their significant E values and Z scores and the information they contribute to understanding the relationship between the structure and function of 6-PF-2-K/F-2,6-BPase.
6-Phosphofructo-2-kinase/fructose-2,6-bisphosphatase was first compared to phosphoglycerate mutase (5PGM) from Saccharomyces cerevisiae. This comparison protein had an E value of 4 x 10-59 and a Z score of 19.5. Knowledge of the structure and function of this enzyme was used to develop an understanding of the catalytic mechanism of the F-2,6-BPase domain. Phosphoglycerate mutase catalyzes the transfer of the phosphoryl group of 3-phosphoglycerate to the 2-carbon of the same molecule to produce 2-phosphoglycerate in the eighth reaction of glycolysis (1). It is a tetrameric protein with 246 residues per subunit. Despite the sequence differences between phosphoglycerate mutase and the bisphosphatase domain, the two enzymes share multiple structural similarities including two catalytic histidine residues in the active site (His-256 & His-390 for F-2,6-BPase; His-8 & His-181 for PM), which also leads to a shared catalytic mechanism involving a phosphohistidine intermediate. In both enzymes, the region around the catalytic histidine residues is abundant in positively charged residues that attract the negatively charged phosphate ligand and direct it to the catalytic site. Unlike the bisphosphatase domain, phosphoglycerate mutase requires t co-factor, 2,3-bisphosphoglycerate; a small amount of 2,3-bisphosphoglycerate must be available to phosphorylate His-8 before the mechanism can proceed. Phosphorylation of the enzyme at His-8 and binding of a monophosphoglycerate at the active site initiates a round of catalysis. His-8 then phosphorylates the monophosphoglycerate to form a 2,3-bisphosphoglycerate intermediate. This intermediate must be able to reorient itself within the active site without dissociating from the enzyme in order to transfer its other phosphate group to the histidine residue and regenerate the active form of the enzyme (8). It is suggested that the His-181 residue acts an acid catalyst during the transfer of the phosphate group to His-8, similar to one of the functions of His-390 of the F-2,6-BPase domain which protonates the leaving group to resolve the transition state (2,8).
The second comparison protein was human (Homo sapiens) liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (1K6M). The human liver form of the enzyme has an E value of 0.0 and a Z score of 56.4. Human liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase is a head-to-head homodimer with 432 amino acids per monomer. It has 74% sequence identity to the testis isozyme and the two enzymes exhibit structural, folding pattern, and mechanistic similarities (5). As fructose-2,6-bisphosphatase plays an important role in all glycolysis-dependent tissues it has been found along with the 6-PF-2-K/F-2,6-BPase enzyme in virtually every eukaryotic tissue or cell examined, including plants and yeast (3). The enzyme plays an important role in maintaining glucose homeostasis not only within rats but within all mammals, and not only in testis tissue but in five different tissues. To date, five mammalian isozymes of 6-PF-2-K/F-2,6-BPase have been discovered and isolated: liver, skeletal muscle, heart, testis, and brain (2). However, research has yet to entirely address the question of how individual tissue isozymes differ in structure and function in order to optimize glucose metabolism according to the physiological role of the tissue. Comparison of rat testes 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (1BIF) to human liver 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (1K6M) yields results that are a first step towards providing an answer.
First, while ATP binding has been shown to be a prerequisite for fructose-6-phosphate (F6P) binding in the 6-PF-2-K domain of the testes isozyme, loose coupling of the ATP and F6P binding loops suggest that F6P binding is independent of ATP binding for the liver isozyme (4,5). In the testes isozyme, the last turn of the α5 helix unwinds when bound to ATP which allows Lys-172 to stabilize the bridge oxygen between the β and the γ phosphates of ATP. This conformational change is transmitted to the coupled F6P binding loop which is then able to bind F6P (4). This is termed the “switch-on” conformation. When bound to ADP, the testes isozyme exhibits the “switch-off” conformation. The liver isozyme, however, does not show the “switch-off” conformation, even when bound to ADP. This structure-function difference could be related to the different physiological goals of the two isozymes. The liver regulates blood glucose levels which requires sensitivity to blood F6P levels, while the testes maintain a steady level of energy production, which requires sensitivity to ATP levels. Therefore each isozyme is sensitive to the molecule whose concentration acts as a key piece of information for that tissue (5).
Secondly, a domain twist results in increased stability of the F-2,6-BPase domain of the human liver isozyme compared to the rat testes isozyme. In both the testis and the liver form, a hairpin turn between the β1-5 and β1-6 strands of the 6-PF-2-K domain interacts with the α7 and α13 helices at the interface between the two domains in each monomer. However, in the liver isozyme the hairpin turn is rotated by 3° due to a positive charge at Arg-225 which pulls the turn upward and causes new hydrogen bonds to form. These changes cause the α7 helix to shift in order to maintain hydrophobic interactions, which results in the shifting of the α13 helix as well. This shift is propagated throughout the structure, and results in an overall twist of 5° in the F-2,6-BPase domain of the liver isozyme. The twist causes an increase in dimeric interactions between the two F-2,6-BPase domains of the isozyme, as shown by an increase in the buried surface area between them. Overall, the stability of the F-2,6-PBase domain of the liver isozyme is increased with respect to the testes isozyme. This change, along with sequence differences in the fructose-2,6-bisphosphate (F-2,6-BP) binding loops, enhances the affinity of the human liver F-2,6-BPase domain for fructose-2,6-bisphosphate. The concentration of fructose-2,6-bisphosphate is a key indicator of fluctuations between glycolysis and gluconeogenesis in the liver, so it is reasonable that the liver enzyme would have increased sensitivity to the concentration of F-2,6-BP. Meanwhile, the bisphosphatase domain of the testes isozyme has a lower affinity for F-2,6-BP, indicating that it acts more to regulate the rate of glycolysis alone rather than the interplay between glycolysis and gluconeogenesis (5).