Phosphofructokinase-1 (Pfk) acts as the main control point of flux through glycolysis. It is involved in complex allosteric regulation and Pfk mutations have been linked to cancer development. Whereas the 3D structure and structural basis of allosteric regulation of prokaryotic Pfk has been studied in great detail, our knowledge about the molecular basis of the allosteric behaviour of the more complex mammalian Pfk is still very limited. To characterize the structural basis of allosteric regulation, the subunit interfaces and the functional consequences of modifications in Tarui's disease and cancer, we analysed the physiological homotetramer of human platelet Pfk at up to 2.67 Å resolution in two crystal forms. The crystallized enzyme is permanently activated by a deletion of the 22 C-terminal residues. Complex structures with ADP and fructose-6-phosphate (F6P) and with ATP suggest a role of three aspartates in the deprotonation of the OH-nucleophile of F6P and in the co-ordination of the catalytic magnesium ion. Changes at the dimer interface, including an asymmetry observed in both crystal forms, are the primary mechanism of allosteric regulation of Pfk by influencing the F6P-binding site. Whereas the nature of this conformational switch appears to be largely conserved in bacterial, yeast and mammalian Pfk, initiation of these changes differs significantly in eukaryotic Pfk.

INTRODUCTION

6-Phosphofructokinase (Pfk; EC 2.7.1.11), is the main regulator of glycolysis and catalyses the Mg-ATP-dependent formation of fructose-1,6-bisphosphate (F16BP) from fructose-6-phosphate (F6P) in prokaryotic and eukaryotic cells. Bacterial Pfk are modulated essentially by two effector molecules: Mg2+-ADP act as an activator, whereas phosphoenolpyruvate (PEP) has an inhibitory effect. In contrast, most eukaryotic Pfk differ significantly in their allosteric regulation from the bacterial enzymes, being influenced by more than 20 effectors [13]. Physiologically most relevant are AMP and fructose-2,6-bisphosphate (F26BP) [4,5] as activators whereas ATP and citrate act as inhibitors [3,6,7]. Mammalian Pfk forms three isozymes: muscle, liver and platelet Pfk (hmPfk, hlPfk and hpPfk for the human enzymes respectively). The subunits form homo- and hetero-meric tetramers, which are the smallest catalytically-active enzyme forms but show a tendency to aggregate to larger but likewise active oligomers [8,9]. In skeletal muscle, only the hmPfk isozyme is expressed. However, in other tissues mostly all three isozymes are present in tissue-specific ratios [10]. It has been demonstrated for various mammalian Pfk that these isozymes differ in their kinetic properties [1115].

Pfk variants are associated with severe diseases including type VII glycogen storage disease (Tarui's disease) and cancer [14,16 and the references therein]. Tarui disease mutations lead to a complete absence and half the normal level of Pfk activity in skeletal muscles and in erythrocytes respectively [17]. Post-translational modification of muscle Pfk in terms of proteolytically truncated C-termini was reported for tumorigenic cell lines [18]. A further post-translational modification of hlPfk, namely glycosylation by β-N-acetylglucosamine at Ser529, was found in several cancer cell lines [19]. This glycosylation reduces Pfk activity and redirects glucose flux through the pentose phosphate pathway resulting in a growth advantage of the cancer cells. Cancer cell proliferation was reduced by blocking Pfk glycosylation.

Structural information on the allosteric switch of bacterial Pfk was obtained from the structures of Bacillus stearothermophilus Pfk (BsPfk) in the R-state (PDB ID: 4PFK) and in the T-state (PDB ID: 6PFK) [20,21]. The enzyme from Trypanosoma brucei was the first eukaryotic representative to be structurally characterized. However, it is more similar to the bacterial diphosphate-dependent Pfk enzymes [22,23]. Despite early reports of the crystallization of rabbit muscle Pfk (rmPfk) [24] structural information on eukaryotic ATP-dependent Pfk became available only a few years ago via the crystal structures of Pichia pastoris Pfk (PpPfk) [25] in a T-state. (The terms ‘R-state’ and ‘T-state’ for eukaryotic Pfk are used to denote activated and inhibited states respectively). The complex regulation of eukaryotic Pfk by different allosteric effectors probably involves more than one R-state as well as T-state conformation) and Saccharomyces cerevisiae Pfk (ScPfk) in an R-state [26]. Structures of mammalian Pfk were obtained for rmPfk [26] and hmPfk [27] so far only in an inactive dissociated dimer state.

In the present study, we report on crystal structures of hpPfk in the physiologically active homotetrameric state to characterize the structure and molecular basis of the enzyme mechanism and allosteric regulation of mammalian Pfk, the subunit interfaces which enable the formation of tissue-specific heteromeric isozymes and the effects of mutations or modifications in Tarui's disease and cancer.

MATERIALS AND METHODS

Cloning, expression and purification

Wild-type (hpPfkwt) and a truncated variant (hpPfkΔNC) were expressed via the pET51b vector (Novagen). hpPfkΔNC lacks 25 and 22 residues from the N- and C-terminus respectively. The genes were introduced between the AatII and NotI restriction sites using the following primers:

  • hpPfkwt

  • 5′-AAAAGACGTCGATGGACGCGGACGACTCC-3′;

  • 5′-AAAAGCGGCCGCGACACTCCAGGGCTGCAC-3′

  • hpPfkΔNC

  • 5′-AAAAGACGTCGGCCATCGGCGTGCT-3′;

  • 5′-AAAAGCGGCCGCTTAGGCCAGGATTTTCATGAGG-3′.

The expressed proteins contain an N-terminal Strep-II-tag followed by an enteropeptidase cleavage site. Overexpression was carried out in the Escherichia coli strain Rosetta(DE3)pLysS (Novagen). Transformed bacterial cells were grown overnight at 37°C in terrific broth containing 100 μg /ml carbenicillin and 33 μg/m chloramphenicol. Cultures were grown at 37°C to a D600 of ∼1.0. Then the temperature was reduced to 27°C and cells were induced with 1.0 mM IPTG for 4—5  h until a D600 value of 8.0 was reached. If not stated otherwise, the protein was kept at 4°C. Cells were disrupted with an APV-2000 homogenizer (APV Products) in buffer A (50 mM potassium phosphate, 5 mM DTT, adjusted with Tris to pH 8.0) containing 0.5 mM PMSF and Roche complete protease inhibitors. The lysate was centrifuged at 100000 g for 1 h. Thereafter, the supernatant was loaded on to a 5 ml StrepTrap HP column (GE Healthcare) equilibrated with buffer-A. For the purification of protein with the Strep-II-tag (hpPfkΔNCtag) buffer B1 (50 mM potassium phosphate, 100 mM KCl, 5% glycerol, 2 mM DTT, 2.5 mM desthiobiotin, adjusted to pH  8.0 with Tris) was used for elution. For the preparation of protein for the removal of the Strep-tag by enteropeptidase cleavage (hpPfkΔNC), buffer B2 (50 mM potassium phosphate, 100 mM KCl, 5% glycerol, 2 mM DTT, 2.5 mM desthiobiotin, adjusted to pH 8.5 with Tris) was used. The elution peak fractions were pooled and 5 mM Tris(2-carboxyethyl)phosphin (TCEP), as well as 5 mM AMP were added to enhance protein stability. The Strep-II-tag was removed by digestion with enteropeptidase in a ratio of 1:1000 (w/w) at room temperature for 3–5 h. Digestion was stopped by adding 25× concentrated stock solution Roche complete protease inhibitors to a concentration of 1× concentrated protease inhibitor. hpPfkΔNCtag and hpPfkΔNC were concentrated from a volume of ∼ 4–7 ml after elution to 1.5–2.0 ml by ultrafiltration (Vivaspin 20, Sartorius).

After that, the protein was applied to a size exclusion chromatography Superdex 200 16/60 column (GE Healthcare) equilibrated with buffer C1 [50 mM potassium phosphate buffer, 100 mM KCl, 5% glycerol, 1 mM TCEP, 1 mM (NH4)2SO4 adjusted to pH 8.0 with Tris] or C2 [50 potassium phosphate buffer, 100 mM KCl, 5% glycerol, 1 mM TCEP, 1 mM (NH4)2SO4 adjusted to pH 8.5 with Tris] for the tagged construct and the digested protein respectively. Pure fractions were pooled and a buffer exchange was performed by ultrafiltration (Vivaspin 20, Sartorius) against buffer D1 [10 mM potassium phosphate buffer, 100 mM KCl, 5% glycerol, 1 mM TCEP, 1 mM (NH4)2SO4, 5 mM AMP, adjusted to pH 8.0 with Tris] for hpPfkΔNCtag or with D2 [10 mM Tris/potassium phosphate buffer, pH  8.5, 100 mM KCl, 5% glycerol, 1 mM TCEP, 1 mM (NH4)2SO4, 10 mM Mg-ATP, adjusted to pH 8.5 with Tris] for hpPfkΔNC. Finally, for crystallization the protein was concentrated by ultrafiltration to 18–20 mg/ml and stored at 10°C (Vivaspin 500, Sartorius).

Phosphofructokinase activity assay

Kinetic parameters of hpPfkwt and hpPfkΔNCtag were determined according to Akkerman et al. with some modifications [11,12]. The reaction mixture contained 0.2 M Tris/HCl, pH 7.1, auxiliary enzymes (0.45 units/ml aldolase, 4.5 units/ml triosephosphate isomerase and 1.5 units/ml glycerol phosphate dehydrogenase), 0.22 mM NADH, 5 mM MgCl2 and effectors as indicated in the Results section. The influence of effectors on the relative enzyme activity was measured at the following substrate concentrations: hpPfkwt 0.9 mM F6P, hpPfkΔNCtag 0.167 mM F6P (corresponding to the F6P concentrations for half-activation) and 0.5 mM ATP for both enzymes. During ATP inhibition experiments MgCl2 concentration was increased to 15 mM. The reaction was started by adding 3 μl of purified hpPfk (hpPfkwt or hpPfkΔNCtag) diluted in its storage buffer. This buffer contained 10 mM Tris/potassium phosphate, pH  8.0, 100 mM KCl, 5% glycerol and 1 mM TCEP. The NADH-oxidation followed at 25°C was monitored at 340 nm. Curve fittings for kinetic parameters were generated by either Michaelis-Menten or Hill equations using Prism (GraphPad Software). The parameters are defined as follows: KM is the Michaelis-Menten constant; S0.5 is the substrate concentration at half-maximum activity; n is the Hill coefficient; Ki0.5 is the effector concentration at half-maximal inhibition and Kd0.5 is the effector concentration at half-maximal activation.

Crystallization

Initial screening for crystallization conditions was performed with commercially available crystallization screens (Hampton Research) at 292 K using the sitting-drop vapour diffusion technique in three-drop 96-well plates (Greiner). The volume of the reservoir solution was 90 μl. Crystals suitable for diffraction experiments were only obtained via microseed matrix screening. One hundred nanolitre seeds, 200-nl reservoir and 300-nl protein solution were dispensed using a MicroSys eight-channel dispensing system (Zinsser Analytic). Optimizations of the initial hits were performed using the hanging-drop vapour-diffusion method. Therefore, 0.5 μl of seeds, 1.0 μl of reservoir and 1.5 μl of protein solution were mixed and equilibrated against 500 μl of reservoir solution in 24-well trays (Nelipak).

Monoclinic crystals in space group P21 (crystal form I, hpPfkΔNCtag) grew in 0.1 M KCl, 0.025 M MgCl2, 0.05 M sodium cacodylate pH 5.4–6.4 and 11%–17% iso-propanol at 19°C. The protein solution contained 5 mM AMP. Crystals of hpPfkΔNC were obtained in 0.1 M 2-(N-morpholino) ethane sulfonic acid (MES), pH 5.4–6.4, 17%–23% (v/v) MPD at 19°C. The protein solution contained 10 mM Mg-ATP. These crystals belonged to space group P3221.

Before freezing crystals were transferred stepwise to a cryo buffer with increasing PEG 3350 concentrations. Crystals were frozen in liquid nitrogen.

Data collection and processing

Data collection was performed at Bessy II of the Helmholtz-Zentrum Berlin (HZB) (Beamline 14.1) in fine ϕ-slicing mode in order to improve the scaling statistics of the otherwise poorly diffracting Pfk crystals. Initial data processing was done with XDS [28]. The program POINTLESS was used to determine the Laue group [29] and the reflections were scaled and merged with XSCALE [28]. Crystallographic data statistics are listed in Table 1.

Table 1
Statistics of data collection, refinement and validation

Values in parentheses are for the highest resolution shell. *2In chains A and C, residues 83–98 could not be modelled due to disorder. Abbreviations: ASU, asymmetric unit; MPD, 2-methyl-2,4-pentanediol.

Structure ΔNCtag-unl ΔNCtag-lig ΔNC-ATP 
Ligands bound  ADP, F6P, F16BP ATP 
Crystallization buffer 0.1 M KCl, 0.025 M MgCl2, 0.05 M Na cacodylate, pH 5.4–6.4, 11%–17% iso-propanol, 5 mM AMP 0.1 M KCl, 0.025 M MgCl2, 0.05 M Na cacodylate, pH 5.4–6.4, 11%–17% iso-propanol, 2.5 mM AMP 0.1 M MES, pH 5.4–6.4, 17%–23% (v/v) MPD, 10 mM Mg-ATP 
Cryo/dehydration buffer 0.055 M Na-cacodylate, pH 6.0, 16.5% iso-propanol, 5 mM AMP, 15% MPD, 9% PEG 3350 0.055 M Na-cacodylate, pH 6.2, 16.5% iso-propanol, 10 mM MgATP, 10 mM F6P, 15% MPD, 9% PEG 3350 0.11 M MES, pH 6.0, 24% (v/v) MPD, 10 mM Mg ATP 
Data collection    
 X-ray source BESSY MX 14.1 BESSY MX 14.1 BESSY MX 14.1 
 X-ray detector Pilatus 6M Pilatus 6M Pilatus 6M 
 Wavelength (Å) 0.91841 0.91841 0.91841 
 Crystal form II 
 Space group P21 P21 P3212 
 Cell a, b, c (Å) 79.3, 178.0, 133.1 76.7, 164.5, 133.2 132.9, 399.1 
 Cell α, β, γ (°) 90, 104.5, 90 90, 103.0, 90 90, 90, 120 
 Resolution range (Å) 50.0–2.89 (3.1–2.89) 46.0–2.67 (2.8–2.67) 50.0–2.80 (2.97–2.80) 
 Multiplicity 3.6 (3.5) 5.0 (4.9) 7.9 (8.0) 
 Solvent content (%) 55.2 50.2 59.9 
 Molecules per ASU 
 Wilson B factor (Å269.6 55.8 53.5 
 Mosaicity (°) 0.138 0.299 0.060 
 Total reflections 289757 450502 796444 
 Unique reflections 79810 89992 101291 
 I/σ (I) 13.1 (2.1) 13.6 (2.1) 9.535 (1.75) 
 Completeness (%) 99.3(98.1) 99.1 (97.1) 99.7 (99.5) 
Rmerge (%) 7.8 (60.6) 10.0 (74.5) 22.1 (128.6) 
Rmeas (%) 9.1 (67.6) 11.2 (83.4) 23.6 (137.3) 
 CC1/2 (%) 99.8 (84.9) 99.7 (70.0) 99.1 (58.8) 
Refinement    
Rwork/Rfree (%) 19.7/21.9 21.4 (25.2) 16.4/20.0 
 Reflections in test set 3993 4524 5117 
 Protein residues 25–178*2,182–703,711–762 25–178*2,183–704,710–762 25–705, 711–762 
 Ligands PO4 2 ADP, 2 F6P, 4 F16BP, PO4 4 ATP, PO4 
 Average B factor (Å294.42 64.9 40.0 
 RMSD bond/angles (Å)/(°) 0.01/1.13 0.01/1.48 0.010/1.55 
Ramachandran plot    
 Favoured (%) 96.1 97.2 95.3 
 Additional allowed (%) 3.3 2.3 4.0 
 Number of outliers 17 16 20 
 PDB accession code 4WL0 4XZ2 4U1R 
Structure ΔNCtag-unl ΔNCtag-lig ΔNC-ATP 
Ligands bound  ADP, F6P, F16BP ATP 
Crystallization buffer 0.1 M KCl, 0.025 M MgCl2, 0.05 M Na cacodylate, pH 5.4–6.4, 11%–17% iso-propanol, 5 mM AMP 0.1 M KCl, 0.025 M MgCl2, 0.05 M Na cacodylate, pH 5.4–6.4, 11%–17% iso-propanol, 2.5 mM AMP 0.1 M MES, pH 5.4–6.4, 17%–23% (v/v) MPD, 10 mM Mg-ATP 
Cryo/dehydration buffer 0.055 M Na-cacodylate, pH 6.0, 16.5% iso-propanol, 5 mM AMP, 15% MPD, 9% PEG 3350 0.055 M Na-cacodylate, pH 6.2, 16.5% iso-propanol, 10 mM MgATP, 10 mM F6P, 15% MPD, 9% PEG 3350 0.11 M MES, pH 6.0, 24% (v/v) MPD, 10 mM Mg ATP 
Data collection    
 X-ray source BESSY MX 14.1 BESSY MX 14.1 BESSY MX 14.1 
 X-ray detector Pilatus 6M Pilatus 6M Pilatus 6M 
 Wavelength (Å) 0.91841 0.91841 0.91841 
 Crystal form II 
 Space group P21 P21 P3212 
 Cell a, b, c (Å) 79.3, 178.0, 133.1 76.7, 164.5, 133.2 132.9, 399.1 
 Cell α, β, γ (°) 90, 104.5, 90 90, 103.0, 90 90, 90, 120 
 Resolution range (Å) 50.0–2.89 (3.1–2.89) 46.0–2.67 (2.8–2.67) 50.0–2.80 (2.97–2.80) 
 Multiplicity 3.6 (3.5) 5.0 (4.9) 7.9 (8.0) 
 Solvent content (%) 55.2 50.2 59.9 
 Molecules per ASU 
 Wilson B factor (Å269.6 55.8 53.5 
 Mosaicity (°) 0.138 0.299 0.060 
 Total reflections 289757 450502 796444 
 Unique reflections 79810 89992 101291 
 I/σ (I) 13.1 (2.1) 13.6 (2.1) 9.535 (1.75) 
 Completeness (%) 99.3(98.1) 99.1 (97.1) 99.7 (99.5) 
Rmerge (%) 7.8 (60.6) 10.0 (74.5) 22.1 (128.6) 
Rmeas (%) 9.1 (67.6) 11.2 (83.4) 23.6 (137.3) 
 CC1/2 (%) 99.8 (84.9) 99.7 (70.0) 99.1 (58.8) 
Refinement    
Rwork/Rfree (%) 19.7/21.9 21.4 (25.2) 16.4/20.0 
 Reflections in test set 3993 4524 5117 
 Protein residues 25–178*2,182–703,711–762 25–178*2,183–704,710–762 25–705, 711–762 
 Ligands PO4 2 ADP, 2 F6P, 4 F16BP, PO4 4 ATP, PO4 
 Average B factor (Å294.42 64.9 40.0 
 RMSD bond/angles (Å)/(°) 0.01/1.13 0.01/1.48 0.010/1.55 
Ramachandran plot    
 Favoured (%) 96.1 97.2 95.3 
 Additional allowed (%) 3.3 2.3 4.0 
 Number of outliers 17 16 20 
 PDB accession code 4WL0 4XZ2 4U1R 

Structure solution and refinement

The structure of rmPfk (PDB ID: 3O8N [26]) modified by CHAINSAW [30] served as search model in molecular replacement (MR) trials for structure determination of hpPfkΔNCtag. Initial phases were obtained with PHENIX-AutoMR [31] yielding an electron density map that could be auto-traced with PHENIX-AutoBuild [32]. The resulting model was completed through iterative rounds of rebuilding in COOT [33] and refinement in BUSTER-TNT [34]. PHENIX-AutoBuild was also used for the calculation of an iterative composite omit map [35] in order to reduce model bias during rebuilding. Final refinement steps for structures of crystal form I were carried out with PHENIX-Refine [36].

Afterwards, the structure of hpPfkΔNCtag was used as search model for the hpPfkΔNC structures via MR with PHASER [31]. The crystals appeared to belong to the hexagonal space group P6222 and the Matthews coefficient indicated the presence of one dimer per asymmetric unit, with a solvent content of 60.6%. MR in this space group was not successful and L and H twinning tests with phenix.xtriage [36] indicated that the crystal was perfectly twinned. A perfect merohedral twinning is not possible in P6222 but this twinning in crystal classes P312, P321 or P6 might result in apparent P6222 symmetry [37]. MR trials resulted in an outstanding solution in space group P3221. The resulting model was completed through iterative rounds of rebuilding in COOT [33] and refinement in REFMAC-5 [38].

Up to 98 residues were not included in the structure models, since they were part of flexible regions (residues 83–98, 179–181 and 704–710) that could not be accurately modelled in the electron density maps (Table 1). For all structures, validation with MolProbity [39] was integrated as part of the iterative rebuild and refinement procedure. Molecular diagrams were drawn using PyMOL (Schrödinger).

RESULTS AND DISCUSSION

The crystallized hpPfk constructs are permanently activated variants

Crystallization of mammalian Pfk is complicated by the tendency to form aggregates or polymers [24,4046]. To improve the crystallization properties, we prepared several truncated and mutated variants of Pfk. The crystallized truncated variants generated in the present study lack residues 1–25 from the N-terminus and 763–784 from the C-terminal end (see ‘Materials and Methods’; Supplementary Figure S9). Ser25, the first residue of the hpPfkΔNC-construct, is far away from the dimerization or the tetramerization interfaces. In the structures of ScPfk, PpPfk and rmPfk further residues were present at the N-terminus, which were mostly not modelled due to flexibility. This indicates that the N-terminal ∼25 residues of hpPfk are largely flexible and unlikely to affect the allosteric behaviour. In contrast, Ala762 is located close to the putative inhibitory citrate-binding site [47,48] that corresponds to the activating Mg2+/ADP effector-binding site in bacterial Pfk. In the other eukaryotic Pfk structures, three (rmPfk) to ten (ScPfk) additional residues are defined after Ala762 and these residues extend towards the solvent but are located between helices 1 (of the N-terminal half) and 8′ (of the C-terminal half). A deletion of these residues between the two homologous halves of a Pfk chain close to the putative inhibitory citrate is likely to influence the allosteric behaviour. Indeed, truncations and mutations of platelet- and muscle-Pfk revealed that the C-terminal residues and in particular Leu775 and Glu776, which are conserved in all human isozymes, strongly influenced the transmission of the inhibitory ATP signal and abolished citrate binding and inhibition [49]. Whereas these mutations only moderately decreased S0.5F6P a further truncation of the C-terminus increases substrate affinity almost to that of the AMP-activated state. We therefore investigated the kinetic and regulatory properties of the truncated variant used for crystallization.

For hpPfkΔNCtag S0.5F6P (the F6P concentration for half activation) is reduced by approximately one order of magnitude compared with the wild-type enzyme to a value that matches almost the ones in the presence of the activating effector AMP for hpPfkwt (Table 2). Thus, even without activation by AMP, the enzyme is in a permanently activated state and displays high substrate affinity. Concomitantly, the co-operativity between the active sites in the form of the Hill coefficient is significantly reduced to values matching the AMP-activated wild-type Pfk. Also the KM values for the second substrate ATP are reduced for the truncated Pfk variant. Pfk is inhibited by higher concentrations of ATP (>0.5 mM). This inhibiting effect of ATP is more pronounced for the wild-type than for the mutant which is illustrated by its increased Ki0.5ATP value (Table 2). The wild-type enzyme is completely inhibited by 1 mM citrate, whereas the inhibiting effect of this effector is abolished in the truncated enzyme variant. Taken together, these data demonstrate that the truncated variant hpPfkΔNC is permanently activated even in the absence of AMP (Table 2, Supplementary Figure S1) and cannot be inactivated by the addition of citrate.

Table 2
Comparison of kinetic properties of hpPfkwt and hpPfkΔNCtag

Measurements were carried out at pH 7.1 in a buffer containing 5 mM Mg2+ and 0.5 mM ATP. Assays were performed with 0.0167 mM (hpPfkwt) or 0.9 mM (hpPfkΔNCtag) F6P. Mean values of at least three independent measurements are reported. *Kinetic properties of native hpPfk isolated from human [11]. KM is the Michaelis-Menten constant; S0.5 is the substrate concentration at half-maximum activity; n is the Hill coefficient; Ki0.5 is the effector concentration at half-maximal inhibition and Kd0.5 is the effector concentration at half-maximal activation.

Parameter Effector hpPfkΔNCtag hpPfkwt Literature* 
S0.5F6P (μM)  89±12 886±61 450 
n  1.9±0.1 2.6±0.1 2.7 
S0.5F6P (μM) 1 mM AMP 30±9 52±5  
n 1 mM AMP 1.4±0.1 1.7±0.1  
KMATP (μM)  29±4 37±3 50 
Ki0.5ATP (mM)  4.6±0.3 1.6±0.1  
KMATP (μM) 5 mM AMP 76±8 160±9  
Kd0.5AMP (μM)  9±2 13±3  
Parameter Effector hpPfkΔNCtag hpPfkwt Literature* 
S0.5F6P (μM)  89±12 886±61 450 
n  1.9±0.1 2.6±0.1 2.7 
S0.5F6P (μM) 1 mM AMP 30±9 52±5  
n 1 mM AMP 1.4±0.1 1.7±0.1  
KMATP (μM)  29±4 37±3 50 
Ki0.5ATP (mM)  4.6±0.3 1.6±0.1  
KMATP (μM) 5 mM AMP 76±8 160±9  
Kd0.5AMP (μM)  9±2 13±3  

Oligomer structure and relationship to bacterial Pfk

The crystal structures of rmPfk and hmPfk both contain an inactive dissociated dimer state of the enzyme, most probably as a result of the acidic crystallization conditions [26,27]. We examined the crystallization properties of hmPfk and hpPfk as full-length wild-type forms and also variants with N- and C-terminal truncations and point mutations (surface entropy reduction). A notorious problem is the aggregation tendency in particular of liver-Pfk [4042,50]. Diffraction-quality crystals could only be obtained for the hpPfkΔNC variant.

Two crystal forms were used in the present study: crystal form I (space group P21) formed using hpPfkΔNCtag whereas crystal form II (space group P3212) was obtained from hpPfkΔNC. Both crystal forms contain one homotetramer in the asymmetric unit. In crystal form I, we analysed two structures: the unliganded enzyme (designated as ΔNCtag-unl) and a complex structure with ADP, F6P and F16BP (ΔNCtag-lig). In crystal form II the structure ΔNCtag-ATP was analysed with ATP bound to the active sites. Also several phosphate ions were located in the electron density maps, mostly bound to the effector-binding sites. These ions were probably absorbed from the phosphate buffer used during the protein purification.

The homotetrameric hpPfk has D2 symmetry and dimensions of 95×103×144 Å along the three perpendicular 2-fold axes, which are referred to as x, y and z respectively (Figure 1). Based on sequence homology, it has been suggested that the eukaryotic Pfk evolved from an ancestor protein resembling the current bacterial Pfk by a gene duplication and tandem fusion event [51,52] (Supplementary Figure S2). Two of these subunits form a closely interacting homodimer across the z-axis, which corresponds to the homotetramer structure of the bacterial Pfk (Figures 1B and 1C). The N-terminal halves of both chains interact with each other at one side of the homodimer and the C-terminal halves at the other side. The better conserved N-terminal halves contain the active site (termed F and N in the present study for the F6P and nucleotide-binding sites respectively), whereas the former catalytic site of the C-terminal half has developed into new regulatory binding sites (F′ and N′, the prime denotes elements of the C-terminal half). Interactions of the chains at the dimerization interface mainly involves the loop between helix 6 and strand F, the long loop between helix 4 and strand C and helix 10 and strand I. Two homodimers tetramerize within the x, y-plane by interactions between the C-terminal halves. The tetramerization interface is predominantly formed by helices 8′ and 9′. In the resulting homotetramer, the N-terminal halves with the active sites point to the outside, whereas the F26BP allosteric sites point to the tetramerization zone (Figure 1A; Supplementary Figure S3).

Structure of the hpPfk tetramer and the ligand-binding sites

Figure 1
Structure of the hpPfk tetramer and the ligand-binding sites

(A) Homotetrameric structure of hpPfk (ΔNCtag-lig). ADP (green) and F6P (yellow) are bound to the active site. The reaction product F16BP (blue spheres) is bound to the activatory F26BP-binding site F′. ADP and F6P in the active sites of chains B and D are depicted in lighter colours as these ligands are superimposed from chains A and C to mark the binding sites, but these ligands are not bound to these subunits in the crystal structure. Furthermore, ligands of the rmPfk structure (PDB ID: 3O8L) and the BsPfk structure (PDB ID: 4PFK) are superposed to mark the (putative) effector-binding sites E, E′, N1 and N2. (B) View along the z-axis. (C) Homotetrameric bacterial Pfk (EcPfk, PDB ID: 1pfk, shown in grey) with F16BP (yellow) and ADP (green) bound to the active site and ADP (white) bound to the effector site E. (D) Chain A of hpPfk in crystal form II (ΔNC-ATP) with ATP (green spheres) bound to the catalytic site. Further ligands have been superimposed from ΔNCtag-lig to mark the binding sites. Flexible regions are marked in black. (E) Scheme of the relationship of mammalian Pfk and an ancestor Pfk.

Figure 1
Structure of the hpPfk tetramer and the ligand-binding sites

(A) Homotetrameric structure of hpPfk (ΔNCtag-lig). ADP (green) and F6P (yellow) are bound to the active site. The reaction product F16BP (blue spheres) is bound to the activatory F26BP-binding site F′. ADP and F6P in the active sites of chains B and D are depicted in lighter colours as these ligands are superimposed from chains A and C to mark the binding sites, but these ligands are not bound to these subunits in the crystal structure. Furthermore, ligands of the rmPfk structure (PDB ID: 3O8L) and the BsPfk structure (PDB ID: 4PFK) are superposed to mark the (putative) effector-binding sites E, E′, N1 and N2. (B) View along the z-axis. (C) Homotetrameric bacterial Pfk (EcPfk, PDB ID: 1pfk, shown in grey) with F16BP (yellow) and ADP (green) bound to the active site and ADP (white) bound to the effector site E. (D) Chain A of hpPfk in crystal form II (ΔNC-ATP) with ATP (green spheres) bound to the catalytic site. Further ligands have been superimposed from ΔNCtag-lig to mark the binding sites. Flexible regions are marked in black. (E) Scheme of the relationship of mammalian Pfk and an ancestor Pfk.

In the previously reported structures of rmPfk and hmPfk [26,27], the physiological tetramer is dissociated to a dimer. This dimer does not correspond to the closely interacting dimer across the z-axis but it is formed by the y-dyad, e.g. the subunits A and C in Figure 1. It has been suggested that the dissociation of the rmPfk and hmPfk tetramers is caused by low pH via protonation of a histidine residue in the dimerization interface that forms contact to the backbone amino group of Gly204 or Gly205 at physiological pH [26,27]. These residues correspond to His208, Gly34 and Asp35 in the hpPfk tetramer. Indeed, His208 forms a hydrogen bonding interaction with the nitrogen backbone of Asp35 and protonation of the histidine would destroy this interaction.

Asymmetry across the dimer interface

Soaking of crystal form I with ATP and F6P resulted in the binding of F6P and the product ADP in the catalytic sites of chains B and D only (Figures 1A and 1B; Supplementary Figure S4A). A superposition of subunits A and B shows that these differences are caused by an asymmetry of the C2 dimer (Supplementary Figure S4B). The rotation around the z-axis is 177.3° instead of 180°, i.e. the A→B and B→A rotations differ by 5.4°. Region 565–574 of helix 6′ is located at the dimer interface and exhibits different conformations. At the side where the subunits are closer together, helix 6′ is longer but has a kink around residue 566 to avoid a clash with the neighbouring subunit. Also the region 673–675 of helix 10′ at the subunit interface near the F26BP-binding site has different conformations. At the corresponding interface of the N-terminal halves, the asymmetry results in differently occupied active sites in crystal form I. At the interface where the subunits are closer together, the active site is occupied. At the other interface residues 83–98 are disordered and the nearby residues 205–208 of helix 6 have very high B-factors and weak electron density indicating significant flexibility. These residues are in contact with the region 33–37 at the subunit interface where binding of ADP and F6P is observed. In chains B and D, residues 83–98 form an important part of the dimerization interface. The disorder of these residues in chains A and C of crystal form I probably prevents substrate binding in these subunits (Supplementary Figure S4). Arg97 binds the α-phosphate group of ATP. Notably, the asymmetry is present in both crystal forms, including the active sites. Although ATP molecules are bound to all four active sites in this crystal form, they have more than 10 Å2 higher B factors in subunits A and C compared with B and C.

The active site and catalytic mechanism

Four ATP molecules are bound to the catalytic sites of all subunits in crystal form II (Figure 1). The ligand-binding modes and co-ordinating side chains are well defined for the resolutions obtained for these crystals (Figure 2; Supplementary Figure S5B). In crystal form I, the density is generally less well-defined despite the comparable resolution of the data and the B-factors of the protein residues and ligands are significantly higher. A superposition of the subunits of the two crystal forms shows differences in subunits B and D in the extended loop region after helix 5 (residues 142–163) and a shift of the close-by helix 4. These differences are probably caused by crystal packing interactions of these exposed peripheral regions of the N-terminal halves. As a result, also helix 5 including the ATP which is co-ordinated by this helix is shifted by up to 1.2 Å further away from the F6P-binding site in crystal form I compared with form II.

Active site structure and substrate-binding mode

Figure 2
Active site structure and substrate-binding mode

(A) Stereo view of F6P and ADP bound to the active site in crystal form I (ΔNCtag-lig). (B) Stereo view of the binding mode and interactions of ATP in the active site in crystal form II (ΔNC-ATP). (C) Scheme of the interactions of the substrates ATP and F6P in the active site as observed in the two crystal forms. The binding mode of the magnesium ion is modelled. The phosphoryl transfer is presumably initiated by a nucleophilic attack of the 1-OH group of F6P on the γ-phosphate of the Mg2+-ATP complex (arrow).

Figure 2
Active site structure and substrate-binding mode

(A) Stereo view of F6P and ADP bound to the active site in crystal form I (ΔNCtag-lig). (B) Stereo view of the binding mode and interactions of ATP in the active site in crystal form II (ΔNC-ATP). (C) Scheme of the interactions of the substrates ATP and F6P in the active site as observed in the two crystal forms. The binding mode of the magnesium ion is modelled. The phosphoryl transfer is presumably initiated by a nucleophilic attack of the 1-OH group of F6P on the γ-phosphate of the Mg2+-ATP complex (arrow).

The adenine base is stacked between the side chain of Arg102 and helix 5 contacting Gly133. There are no direct polar interactions of the base with the protein (Figure 2B). The ribose is co-ordinated by hydrogen bonds of the OH-groups with the backbone nitrogen and oxygens of Cys98. The phosphate tail of ATP is co-ordinated by Gly34, Arg97, Gly127 and Ser130. Asp175, Glu273, Arg310 and Arg301 are involved in the binding of F6P.

A comparison of the F6P-binding mode with the structures of ScPfk (3O8O), EcPfk (Escherichia coli Pfk; 1PFK) and of the ATP binding mode with the structures of rmPfk (3O8L) shows a good agreement in the binding conformations of the two substrates with the bacterial and yeast Pfk. Since also the key catalytic residues are well conserved, we assume that the homologous Pfk employ an essentially identical catalytic mechanism for phosphoryl transfer from ATP to F6P. To the best of our knowledge, there is no direct structural information on an ATP×F6P-substrate binding mode available, including substrate analogues. The most valuable complex structure in terms of the catalytic mechanism is probably the MgADP×F16BP co-crystal structure of EcPfk (1PFK), in particular since also the Mg2+ ion co-ordinated to the nucleotide could be observed in the present study [53]. In both crystals forms of hpPfk, the divalent metal could not be identified in the electron density maps, probably due to the limited resolution obtained from the crystals. There are three conserved acidic residues located near the phosphate groups of the nucleotide in hpPfk: Asp128, Asp175 and Asp177. It is very likely that some of these aspartates co-ordinate the Mg2+ ion required for catalysis [53]. Based on the Mg2+-co-ordination in 1PFK and the binding modes of F6P in crystal form I (Figure 2A) and ATP in crystal form II (Figure 2B), we constructed a model of the MgATP×F6P×hpPfk ternary complex (Figure 2C). Asp175 is hydrogen bonded to the 1OH-group of F6P and it probably acts as a general base to activate the nucleophile. In EcPfk, an 18000-fold reduction in kcat of the forward reaction has been observed for the mutation of the corresponding residue Asp127 to serine [54]. Whether Mg2+ is co-ordinated directly by Asp128 and Asp177 or via water molecules and its co-ordination to the diphosphate tail could not be analysed due to the lack of an Mg2+×ATP×hpPfk complex structure. All other interactions of ATP and F6P-co-ordination are defined by the structures described in the present study. In addition to the Mg2+ ion, Arg219 and the helix 5 dipole might participate in transition state stabilization.

Effector-binding sites

F′: the F26BP activation site

In the F′ site of the C-terminal half, which corresponds to the F6P site of the N-terminal half, F16BP has bound in structure ΔNCtag-lig (Figures 1A and 3A). Since F16BP was not added to the crystallization setup, it has probably formed from F6P and ATP by the Pfk catalytic activity. F′ has been described as the F26BP activation site of the mammalian Pfk by mutagenesis studies [47,55] and the binding mode of the activating effector was observed in the structure of ScPfk [26]. The F26BP-binding site is located close to the dimer interface and Arg576 and Arg665 of the neighbouring monomer participate in binding the activator (Figure 3A).

Effector-binding sites and allosteric state
Figure 3
Effector-binding sites and allosteric state

(A) F16BP bound to the F26BP activation site in ΔNCtag-lig. The 2Fo-Fc omit map of the ligand is depicted in green. (B) Superposition of F26BP bound to the α-subunit of ScPfk (grey) and F16BP liganded to the A-chain of hpPfk (yellow, chain B orange). (C) Inhibitory citrate-binding site of hpPfk with bound phosphate ion. The C-terminal residues 761 and 762 of the truncated hpPfk construct are coloured green-blue. (D) Subunit rotation of 7° in the allosteric switch between the R-state (BsPfk, PDB ID: 4PFK) and T-state (6PFK) of the bacterial Pfk. (E) Subunit rotations in the comparison of hpPfk (crystal form II) and the R-state of BsPfk. The A-chain of hpPfk was superposed on to the AB-dimer of the R-state of BsPfk (grey subunits in back). This results in a rotational difference of 1.7° for the B-chain of hpPfk (blue) and the bacterial AB-dimer (green). (F) Subunit rotations in the comparison of hpPfk (crystal form II) and the T-state of BsPfk (red). This results in a much larger rotational difference of 8.4°. (GI) Allosteric conformational changes affecting F6P binding in the active site. (G) In the T-state of the bacterial BsPfk (yellow, PDB ID: 6PFK), Glu161 is oriented towards the catalytic site and Arg162 has moved out of the F6P-binding site. In the R-state (green, PDB ID: 4PFK), Arg162 is co-ordinated to the 6-phosphate group, whereas Glu161 points in the opposite direction. (H and I) Superposition of the two crystal forms of hpPfk (chain A in light brown) on to BsPfk (same colouring as in G).

Figure 3
Effector-binding sites and allosteric state

(A) F16BP bound to the F26BP activation site in ΔNCtag-lig. The 2Fo-Fc omit map of the ligand is depicted in green. (B) Superposition of F26BP bound to the α-subunit of ScPfk (grey) and F16BP liganded to the A-chain of hpPfk (yellow, chain B orange). (C) Inhibitory citrate-binding site of hpPfk with bound phosphate ion. The C-terminal residues 761 and 762 of the truncated hpPfk construct are coloured green-blue. (D) Subunit rotation of 7° in the allosteric switch between the R-state (BsPfk, PDB ID: 4PFK) and T-state (6PFK) of the bacterial Pfk. (E) Subunit rotations in the comparison of hpPfk (crystal form II) and the R-state of BsPfk. The A-chain of hpPfk was superposed on to the AB-dimer of the R-state of BsPfk (grey subunits in back). This results in a rotational difference of 1.7° for the B-chain of hpPfk (blue) and the bacterial AB-dimer (green). (F) Subunit rotations in the comparison of hpPfk (crystal form II) and the T-state of BsPfk (red). This results in a much larger rotational difference of 8.4°. (GI) Allosteric conformational changes affecting F6P binding in the active site. (G) In the T-state of the bacterial BsPfk (yellow, PDB ID: 6PFK), Glu161 is oriented towards the catalytic site and Arg162 has moved out of the F6P-binding site. In the R-state (green, PDB ID: 4PFK), Arg162 is co-ordinated to the 6-phosphate group, whereas Glu161 points in the opposite direction. (H and I) Superposition of the two crystal forms of hpPfk (chain A in light brown) on to BsPfk (same colouring as in G).

A superposition of the F16BP- and F26BP-binding modes of hpPfk and ScPfk shows that the two phosphate groups occupy similar positions but the orientation of the ribose is flipped such that 1-phosphate group of F16BP occupies the position of the 6-phosphate group of F26BP (Figures 3A and 3B). The O5 fructose ring oxygens point in approximately the same direction.

HmPfk and hpPfk both bind F16BP with a low micromolar Kd value [56], but only hmPfk is strongly activated by the reaction product [57]. Protection against thermal denaturation suggests that F16BP also binds to hpPfk at an allosteric site [56]. The hpPfk×F16BP structure now confirms that F16BP and F26BP share the same effector-binding site but induce different conformational changes as suggested previously [5759]. The differences in the orientation and hydrogen bonding patterns of the furanose rings of F16BP and F26BP could be the reason for the different effects of these molecules on the activation of hpPfk such that F26BP forms energetically more optimal interactions and shifts the equilibrium to an activated state.

E and E′: the inhibitory citrate effector site

The bacterial effector sites evolved to putative effector-binding sites in the N- and C-terminal halves of eukaryotic Pfk (designated E and E′; Figures 1C and 1D). Since the bacterial E sites are located at the subunit interfaces of the tetramer, both halves of the eukaryotic Pfk contribute to sites E and E′. In hpPfk, the structure of sites E and E′ has diverged significantly from the bacterial effector-binding sites but polar-binding pockets are maintained. Phosphate ions are bound to all E and E′ sites in both crystal forms, except for the E′ sites in chains A and C in crystal form I, which are affected by the disorder of the region 83–98 in these two chains. These findings are in agreement with the structures of the yeast Pfk and rmPfk, where phosphate or sulfate ions are bound to E and E′ sites.

Site-directed mutagenesis showed that mutations in E′ (residues 48, 775 and 776) abolished the inhibition by citrate in rabbit Pfk type C [47]. Polar residues line a pocket, which corresponds to the phosphate-binding site of the ADP and PEP effector molecules and it has been implicated in binding the inhibitory citrate in mammalian Pfk [47,48]. This appears plausible also for hpPfk based on the size and positive charge of the binding pocket. In addition to Arg48, Arg44, Ser83, Lys567, Lys625 and Arg632 also co-ordinate the phosphate ion in the E′ site and might be involved in citrate binding (Figure 3C).

The phosphate ion in the E site is co-ordinated by the side chains of Lys264, Arg430, Arg434 and by the main chain carbonyl of Gly233 and nitrogen of Gly468. Mutagenesis of two arginines of mouse Pfk (corresponding to Arg430 and Arg434 in hmPfk) reduced or abolished the inhibition by ATP [48]. However, the identification of novel nucleotide-binding sites in the structures of PpPfk and rmPfk and the fact that no nucleotide binding was observed in site E in the present and previous crystallographic studies of eukaryotic Pfk suggest that further experiments are required to characterize the relevance of this site for the binding of ATP and transition of the allosteric signal.

N: additional nucleotide effector-binding sites

In the crystal structures of PpPfk, ScPfk and rmPfk nucleotide-binding sites were characterized that do not correspond to the effector or active sites of bacterial Pfk. The ADP-binding site N2 found in the rmPfk structure within the centre of the subunit between the N- and C-terminal halves (Figure 1D) is conserved but not occupied by a ligand in any of the hpPfk structures. The functional relevance of this binding site has not yet been clarified.

A novel binding site N1 was characterized in the PpPfk and rmPfk structures in the N-terminal half (Figure 1D). In rmPfk, ATP as well as ADP could be bound to this site [26]. In the PpPfk structure, an ATP has bound at this site; however, the nucleotide base is flipped by 180° in this structure such that the phosphates interact with a different protein region. It was proposed that N1 is the inhibitory ATP-binding site of eukaryotic Pfk [60]. In the hpPfk structure, this binding site is not occupied by any ligand. Superposition with rmPfk shows that the binding site is generally intact but the linker between the two Pfk halves (residues 401–409) has a different conformation and helix 8 is shifted. It appears plausible that the binding of ATP between the linker region and helix 8 influences helix 9 and strand I at the subunit interface and thus the active site structure as discussed above. We assume that the active state conformation of the crystallized hpPfk variant does not allow for ATP binding at this site. Notably, the ATP-binding site is not highly conserved in sequence between rmPfk and hpPfk and further studies are needed to characterize this putative ATP inhibitory site. Interestingly, D543A is one of the mutations that leads to Tarui's disease (see analysis below) and this residue is hydrogen bonded to the adenine N6 in the inhibitory ATP site. This finding indicates the functional relevance of this site in regulating Pfk activity.

Allosteric state

We applied three structural criteria to characterize the allosteric state of hpPfk in the two crystal forms obtained in this study. In the R to T-transition, BsPfk undergoes a 7° rigid body rotation of the AB-dimer relative to the CD-dimer around the p-axis [21] (Figure 3D). This corresponds to a rotation of chain A relative to chain B in the human enzymes. A comparison of crystal forms I and II of hpPfk with the two states of BsPfk shows that all structures of hpPfk correspond to the R state of BsPfk (Figures 3E and 3F). A second criterion is the distance between the symmetry-related β-strands I as well as I′ at the dimer interface. In the R-state structures of BsPfk the two strands are separated by water molecules whereas they form direct hydrogen bonding interactions in a continuous β-sheet in the T-state [21]. In both crystal forms of hpPfk, a distance of 4.5–5.4 Å between the β-strands I and I′ is characteristic of an R-state conformation (Supplementary Figure S6). A third criterion is the conform-ation of the loop between helix 6 and strand F, in particular the orientation of the arginine residue (Arg162 in BsPfk, Arg210 in hpPfk) co-ordinating the phosphate group of F6P in the active site in the R-state whereas the preceding residue (Glu161 in BsPfk, Gln209 in hpPfk) occupies this position in the T-state. In both crystal forms of hpPfk, the conformation of the corresponding residues Gln209 and Arg210 is indicative of an R-state structure (Figures 3G-3I).

Taken together, these structural features clearly indicate that the hpPfk structures in both crystal forms obtained in this study represent activated states. Whereas this is in agreement with the presence of the activator AMP in the crystallization of crystal form I, the presence of high concentrations of MgATP in the crystallization buffer of crystal form II should induce an inhibited state of the wild-type enzyme. However, the truncations at the C-terminus used in the present study for the purpose of obtaining crystals keep the enzyme in an activated state even in the presence of high concentrations of ATP as demonstrated in the kinetic analysis (Table 2).

Allosteric changes

The structures of hpPfk described in the present study show the enzyme in an active state. Binding of ATP or citrate inactivates the enzyme by reducing the affinity for the substrate F6P. Only on a much longer timescale of minutes does the active tetramer dissociates into inactive dimers [61]. The crystal structures of the dimeric rmPfk and hmPfk [26,27] probably correspond to this state, which is characterized by dissociation of the AB dimer (Figure 1) and thus a loss of the F6P-binding site. Structural information on the effects of citrate or ATP binding on the tetramer structure of mammalian Pfk is currently not available. We do not know if a structure similar to the T-states of BsPfk and PpPfk exists for mammalian Pfk, in which strands I and I' of the AB dimer interface move closer together for direct hydrogen bonding interaction thus affecting the position of Arg301 of the F6P-binding site. Alternatively, ATP or citrate binding might cause a loosening of the AB dimer interface resulting in larger flexibility and an intermediate towards complete dissociation of the subunits. The asymmetric homotetramer structure observed in an electron microscopy study of hmPfk, in which one AB dimer interface is partially dissociated whereas the other appears intact, may be such an intermediate of this dissociation process [62]. A different kind of asymmetry is observed for the C2 z-dyad in hpPfk but also in the yeast Pfk [25] and it likewise affects the F6P-binding site. One binding site is in a catalytically competent structure for substrate binding and catalysis of phosphoryl transfer, whereas the other active site is in an open state. We assume that this asymmetry also exists in solution and the active sites may switch between open and closed states by a counter-rotating motion of the subunits around the z-axis. This motion enables substrate binding and product leaving from the open state, similar to many enzymes exhibiting domain rotations. An opening motion of the active site has also been described for the bacterial 1Pfk within the R-state [53]. Although a T-state structure of mammalian Pfk is not yet available, all current data suggest that allosteric effectors control F6P binding by affecting the AB dimer interface. These changes are induced from completely different effector-binding sites, namely the F26BP F′-site close to the dimerization and tetramerization interfaces, the citrate E′-site at the dimer interface, the ATP-binding site N1 in the N-terminal halves and by the putative ADP activatory site N2 between the N- and C-terminal halves.

Disease implications: Tarui's disease and cancer

The crystal structures of hpPfk facilitate an understanding of how Tarui's disease mis-sense mutations may interfere with proper enzyme function. Alignment of the sequences and superposition of the structures of hmPfk [27] and rmPfk [26] on to the structure of ΔNCtag-lig allowed the identification of the positions of these point mutations within the tetrameric hmPfk structure (Supplementary Figure S7). Based on this analysis, we were able to rationalize the effects of several mutations (Supplementary Table S1 and Supplementary Information). Interestingly, none of these mutation sites is located at the dimerization or tetramerization interfaces (Figure 1; Supplementary Figure S3).

Pfk is glycosylated by β-N-acetylglucosamine at Ser529 (determined for hlPfk) in several cancer cell lines resulting in an inhibited Pfk activity and redirected glucose flux through the pentose phosphate pathway and a growth advantage of the cancer cells [19]. Cancer cell proliferation was reduced by blocking Pfk glycosylation. Liver and platelet Pfk isoforms were found to be most sensitive to glycosylation. hlPfk Ser529 corresponds to Ser540 in hpPfk and it is a conserved residue of the F26BP-binding site (Figure 3A). Glycosylation at Ser540 will block the F26BP-binding site and impede activation by the effector. Furthermore, glycosylation might promote a dissociation of the tetramer into inactive dimers as the F26BP-binding site is located at the dimer interface. Furthermore, Smerc et al. [18] suggested a relevance of post-translational truncation of Pfk in connection with tumour growth. This truncated form loses the inactivating ATP and citrate-binding sites, but also the F26BP-binding sites (see Supplementary Information and Supplementary Figure S8 for a model and further analysis).

Calmodulin-binding sites

Muscle-Pfk exists in different oligomeric states, which depend on the effector composition, pH and protein concentration [6,63,64]. Calmodulin promotes dissociation of Pfk into inactive dimers but it can also activate the tetrameric Pfk or stabilize the dimeric Pfk in an active conformation under certain conditions [65,66]. The muscle Pfk binds two calmodulin molecules per subunit with Kd values of 3 nM and 1 μM [6668].

Based on the sequence of calmodulin-binding peptides isolated from a cyanogen bromide (CNBr) digest of phosphofructokinase, two calmodulin-binding fragments were identified [68]. A low affinity peptide M22 corresponded to residues 715–775 in hmPfk (724–784 in hpPfk, Figure S9) whereas the second peptide M11 corresponding to residues 380–404 in hmPfk (389–417 in hpPfk) has a high binding affinity of 11.4 nM. Both peptides form predominantly helices at the surface of the Pfk tetramer (Figure 4). Peptide M11 contains the linker region between the N- and C-terminal Pfk halves. Therefore, both sites are accessible for calmodulin binding which may cause structural changes inducing tetramer dissociation or influence the allosteric equilibrium.

Calmodulin-binding sites Met11 and Met22 in hpPfk

Figure 4
Calmodulin-binding sites Met11 and Met22 in hpPfk

The positions of the calmodulin-binding peptides Met22 (blue-green) and Met11 (green) are highlighted in the structure of hpPfk.

Figure 4
Calmodulin-binding sites Met11 and Met22 in hpPfk

The positions of the calmodulin-binding peptides Met22 (blue-green) and Met11 (green) are highlighted in the structure of hpPfk.

In summary, the structures of hpPfk provide detailed insight into the architecture of human phosphofructokinases and the evolution of these complex regulated enzymes from an ancestor Pfk resembling the current bacterial enzymes. This includes the effector-binding sites and the oligomer structure. Our structural model demonstrates that human Pfk bind F16BP and F26BP at the same allosteric site which evolved from a former bacterial active site by domain and gene amplification. The hpPfk structure in an activated state provides a model for the active site and the Michaelis complex with F6P and ATP. The allosteric conformational changes induced by effectors at different binding sites probably all induce changes at the interface between the A and B chains influencing F6P binding. Mutations causing Tarui's disease could be rationalized as well as cancer mutations, in particular the glycosylation at a serine in the F26BP-activation site. The Pfk structure may also serve as a guide to develop modulators of Pfk activity as potential anti-tumour drugs. It also forms a molecular basis to further study the allosteric mechanism of the human enzymes by biochemical and structural studies, in particular concerning T-state structures and mutational studies on the newly characterized effector-binding sites for ATP inhibition and ADP activation.

AUTHOR CONTRIBUTION

Marco Kloos generated the expression constructs, purified and crystallized the protein, conducted the crystal structure analysis and kinetic characterization and wrote the manuscript. Antje Brüser and Jürgen Kirchberger supported and advised the kinetic analysis of hpPfk. Torsten Schöneberg and Norbert Sträter conceived and supervised the study. All authors contributed to the evaluation and interpretation of the data and manuscript preparation.

The authors thank Uwe Müller and his team for assistance during data collection at the HZB-MX beam lines at the Helmholtz-Zentrum Berlin. The Helmholtz-Zentrum Berlin is also acknowledged for travel support.

FUNDING

This work was supported by the Deutsche Forschungsgemeinschaft [grant number B8 of Sfb 610] to N.S. and T.S.

Abbreviations

     
  • BsPfk

    Bacillus stearothermophilus Pfk

  •  
  • EcPfk

    Escherichia coli Pfk

  •  
  • F16BP

    fructose-1,6-bisphosphate

  •  
  • F26BP

    fructose-2,6-bisphosphate

  •  
  • F6P

    fructose-6-phosphate

  •  
  • hlPfk

    human liver Pfk

  •  
  • hmPfk

    human muscle Pfk

  •  
  • hpPfk

    human platelet Pfk

  •  
  • hpPfkwt

    wild-type hpPfk

  •  
  • hpPfkΔNC

    truncated hpPfk

  •  
  • hpPfkΔNCtag

    truncated hpPfk with N-terminal Strep-II-Tag

  •  
  • MES

    2-(N-morpholino) ethane sulfonic acid

  •  
  • MR

    molecular replacement

  •  
  • PEP

    phosphoenolpyruvate

  •  
  • Pfk

    6-phosphofructo-1-kinase

  •  
  • PpPfk

    Pichia pastoris Pfk

  •  
  • rmPfk

    rabbit muscle Pfk

  •  
  • ScPfk

    Saccharomyces cerevisiae Pfk

  •  
  • TCEP

    Tris(2-carboxyethyl)phosphine

References

References
1
Sols
A.
Multimodulation of enzyme activity
Curr. Top. Cell Regul.
1981
, vol. 
19
 (pg. 
77
-
101
)
[PubMed]
2
Tejwani
G.A.
Morgan
M.
The role of phosphofructokinase in the Pasteur effect
Trends Biochem. Sci.
1978
, vol. 
3
 (pg. 
30
-
33
)
3
Schöneberg
T.
Kloos
M.
Brüser
A.
Kirchberger
J.
Sträter
N.
Structure and allosteric regulation of eukaryotic 6-phosphofructokinases
Biol. Chem.
2013
, vol. 
394
 (pg. 
977
-
993
)
[PubMed]
4
Reinhart
G.D.
Lardy
H.A.
Rat liver phosphofructokinase: kinetic activity under near-physylogical condition
Biochemistry
1980
, vol. 
19
 (pg. 
1477
-
1484
)
[PubMed]
5
Zancan
P.
Almeida
F.V.R.
Faber-Barata
J.
Dellias
J.M.
Sola-Penna
M.
Fructose-2,6-bisphosphate counteracts guanidinium chloride-, thermal-, and ATP-induced dissociation of skeletal muscle key glycolytic enzyme 6-phosphofructo-1-kinase: A structural mechanism for PFK allosteric regulation
Arch. Biochem. Biophys.
2007
, vol. 
467
 (pg. 
275
-
282
)
[PubMed]
6
Zancan
P.
Marinho-Carvalho
M.M.
Faber-Barata
J.
Dellias
J.M.M.
Sola-Penna
M.
ATP and fructose-2,6-bisphosphate regulate skeletal muscle 6-phosphofructo-1-kinase by altering its quaternary structure
IUBMB Life
2008
, vol. 
60
 (pg. 
526
-
533
)
[PubMed]
7
Usenik
A.
Legis
M.
Evolution of allosteric citrate binding sites on 6-phosphofructokinase
PLoS One
2010
, vol. 
5
 pg. 
e15447
 
[PubMed]
8
Hesterberg
L.K.
Lee
J.C.
Sedimentation study of a catalytically active form of rabbit muscle phosphofructokinase at pH 8.55
Biochemistry
1980
, vol. 
19
 (pg. 
2030
-
2039
)
[PubMed]
9
Hesterberg
L.K.
Lee
C.
Ericksong
H.P.
Structural properties of an active form of rabbit muscle phosphofructokinase
J. Biol. Chem.
1981
, vol. 
256
 (pg. 
9724
-
9730
)
[PubMed]
10
Dunaway
G.A.
Kasten
T.P.
Sebo
T.
Trapp
R.
Analysis of the phosphofructokinase subunits and isoenzymes in human tissues
Biochem. J.
1988
, vol. 
251
 (pg. 
677
-
683
)
[PubMed]
11
Akkerman
J.
Gorter
G.
Sixma
J.
Staal
G.
Human platelet 6-phosphofructokinase purification, kinetic parameters and the influence of sulphate ions on enzyme activity
Biochim. Biophys. Acta
1974
, vol. 
370
 (pg. 
102
-
112
)
[PubMed]
12
Akkerman
J.
Gorter
G.
Over
J.
Sixma
J.
Staal
G.E.J.
Human platelet 6-phophofructokinase relation between inhibition by Mg*ATP2- and cooperativity towards fructose 6-phosphate and investigations on the formation of a ternary complex
Biochim. Biophys. Acta
1975
, vol. 
397
 (pg. 
395
-
404
)
[PubMed]
13
Brüser
A.
Kirchberger
J.
Kloos
M.
Sträter
N.
Schöneberg
T.
Functional linkage of adenine nucleotide binding sites in mammalian muscle 6-phosphofructokinase
J. Biol. Chem.
2012
, vol. 
287
 (pg. 
17546
-
17553
)
[PubMed]
14
Brüser
A.
Kirchberger
J.
Schöneberg
T.
Altered allosteric regulation of muscle 6-phosphofructokinase causes Tarui disease
Biochem. Biophys. Res. Commun.
2012
, vol. 
427
 (pg. 
133
-
137
)
[PubMed]
15
Sa
C.
Sa
V.
Arago
J.J.
Martínez-Costa
O.H.
Sánchez-Martínez
C.
Sánchez
V.
Aragón
J.J.
Chimeric phosphofructokinases involving exchange of the N- and C-terminal halves of mammalian isozymes: implications for ligand binding sites
FEBS Lett.
2007
, vol. 
581
 (pg. 
3033
-
3038
)
[PubMed]
16
Toscano
A.
Musumeci
O.
Tarui disease and distal glycogenoses: clinical and genetic update
Acta Myol.
2007
, vol. 
26
 (pg. 
105
-
107
)
[PubMed]
17
Layzer
R.B.
Rowland
L.P.
Ranney
H.M.
Muscle phosphofructokinase deficiency
Arch. Neurol.
1967
, vol. 
17
 (pg. 
512
-
523
)
[PubMed]
18
Smerc
A.
Sodja
E.
Legisa
M.
Posttranslational modification of 6-phosphofructo-1- kinase as an important feature of cancer metabolism
PLoS One
2011
, vol. 
6
 (pg. 
1
-
12
)
19
Yi
W.
Clark
P.M.
Mason
D.E.
Keenan
M.C.
Hill
C.
Goddard
W.A.
Peters
E.C.
Driggers
E.M.
Hsieh-Wilson
L.C.
Phosphofructokinase 1 glycosylation regulates cell growth and metabolism
Science
2012
, vol. 
337
 (pg. 
975
-
980
)
[PubMed]
20
Evans
P.R.
Farrants
G.W.
Hudson
P.J.
Britton
H.G.
Phosphofructokinase: structure and control [and discussion]
Philos. Trans. R. Soc. B Biol. Sci.
1981
, vol. 
293
 (pg. 
53
-
62
)
21
Schirmer
T.
Evans
P.R.
Structural basis of the allosteric behaviour of phosphofructokinase
Nature
1990
, vol. 
343
 (pg. 
140
-
145
)
[PubMed]
22
Martinez-Oyanedel
J.
McNae
I.W.
Nowicki
M.W.
Keillor
J.W.
Michels
P.A.M.
Fothergill-Gilmore
L.A.
Walkinshaw
M.D.
The first crystal structure of phosphofructokinase from a eukaryote: Trypanosoma brucei
J. Mol. Biol.
2007
, vol. 
366
 (pg. 
1185
-
1198
)
[PubMed]
23
McNae
I.W.
Martinez-Oyanedel
J.
Keillor
J.W.
Michels
P.A.M.
Fothergill-Gilmore
L.A.
Walkinshaw
M.D.
The crystal structure of ATP-bound phosphofructokinase from Trypanosoma brucei reveals conformational transitions different from those of other phosphofructokinases
J. Mol. Biol.
2009
, vol. 
385
 (pg. 
1519
-
1533
)
[PubMed]
24
Parmeggiani
A.
Luft
J.H.
Love
D.S.
Krebs
E.G.
Crystallization and properties of rabbit skeletal muscle phosphofructokinase
J. Biol. Chem.
1966
, vol. 
241
 (pg. 
4625
-
4637
)
[PubMed]
25
Sträter
N.
Marek
S.
Kuettner
E.
Kloos
M.
Molecular architecture and structural basis of allosteric regulation of eukaryotic phosphofructokinases
FASEB J.
2011
, vol. 
25
 (pg. 
89
-
98
)
[PubMed]
26
Banaszak
K.
Mechin
I.
Obmolova
G.
Oldham
M.
Chang
S.H.
Ruiz
T.
Radermacher
M.
Kopperschlager
G.
Rypniewski
W.
The crystal structures of eukaryotic phosphofructokinases from baker's yeast and rabbit skeletal muscle
J. Mol. Biol.
2011
, vol. 
407
 (pg. 
284
-
297
)
[PubMed]
27
Kloos
M.
Brüser
A.
Kirchberger
J.
Schöneberg
T.
Sträter
N.
Crystallization and preliminary crystallographic analysis of human muscle phosphofructokinase, the main regulator of glycolysis
Acta Crystallogr. F Struct. Biol. Commun.
2014
, vol. 
70
 (pg. 
578
-
582
)
[PubMed]
28
Kabsch
W.
XDS
Acta Crystallogr. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
125
-
132
)
[PubMed]
29
Evans
P.
Scaling and assessment of data quality
Acta Crystallogr. D Biol. Crystallogr.
2006
, vol. 
62
 (pg. 
72
-
82
)
[PubMed]
30
Stein
N.
CHAINSAW: a program for mutating pdb files used as templates in molecular replacement
J. Appl. Crystallogr.
2008
, vol. 
41
 (pg. 
641
-
643
)
31
McCoy
A.J.
Grosse-Kunstleve
R.W.
Adams
P.D.
Winn
M.D.
Storoni
L.C.
Read
R.J.
Phaser crystallographic software
J. Appl. Crystallogr.
2007
, vol. 
40
 (pg. 
658
-
674
)
[PubMed]
32
Terwilliger
T.C.
Grosse-Kunstleve
R.W.
Afonine
P. V
Moriarty
N.W.
Zwart
P.H.
Hung
L.W.
Read
R.J.
Adams
P.D.
Iterative model building, structure refinement and density modification with the PHENIX AutoBuild wizard
Acta Crystallogr. D. Biol. Crystallogr.
2008
, vol. 
64
 (pg. 
61
-
69
)
[PubMed]
33
Emsley
P.
Cowtan
K.
Coot: model-building tools for molecular graphics
Acta Crystallogr. D. Biol. Crystallogr.
2004
, vol. 
60
 (pg. 
2126
-
2132
)
[PubMed]
34
Bricogne
G.
Blanc
E.
Brandl
M.
Flensburg
C.
Keller
P.
Paciorek
W.
Roversi
P.
Smart
O.
Vonrhein
C.
Womack
T.
2009
 
Global Phasing Limited, Cambridge
35
Terwilliger
T.C.
Grosse-Kunstleve
R.W.
Afonine
P. V
Moriarty
N.W.
Adams
P.D.
Read
R.J.
Zwart
P.H.
Hung
L.-W.
Iterative-build OMIT maps: map improvement by iterative model building and refinement without model bias
Acta Crystallogr. D. Biol. Crystallogr.
2008
, vol. 
64
 (pg. 
515
-
524
)
[PubMed]
36
Adams
P.D.
Afonine
P. V
Bunkóczi
G.
Chen
V.B.
Davis
I.W.
Echols
N.
Headd
J.J.
Hung
L.-W.
Kapral
G.J.
Grosse-Kunstleve
R.W.
, et al. 
PHENIX: a comprehensive Python-based system for macromolecular structure solution
Acta Crystallogr. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
213
-
221
)
[PubMed]
37
Chandra
N.
Acharya
K.R.
Moody
P.C.E.
Analysis and characterization of data from twinned crystals
Acta Crystallogr. D Biol. Crystallogr.
1999
, vol. 
55
 (pg. 
1750
-
1758
)
[PubMed]
38
Murshudov
G.N.
Vagin
A.
Dodson
E.J.
Refinement of macromolecular structures by the maximum-likelihood method
Acta Crystallogr. D. Biol. Crystallogr.
1997
, vol. 
53
 (pg. 
240
-
255
)
[PubMed]
39
Chen
V.B.
Arendall
W.B.
Headd
J.J.
Keedy
D.A.
Immormino
R.M.
Kapral
G.J.
Murray
L.W.
Richardson
J.S.
Richardson
D.C.
MolProbity: all-atom structure validation for macromolecular crystallography
Acta Crystallogr. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
12
-
21
)
[PubMed]
40
Reinhart
G.D.
Lardy
H.A.
Rat liver phosphofructokinase: use of fluorescence polarization to study aggregation at low protein concentrationt
Biochemistry
1980
, vol. 
19
 (pg. 
1484
-
1490
)
[PubMed]
41
Reinhart
G.D.
Lardy
H.A.
Rat liver phosphofructokinase: kinetic and physiological ramifications of the aggregation behavior
Biochemistry
1980
, vol. 
19
 (pg. 
1491
-
1495
)
[PubMed]
42
Lawrence
F.
Trujillo
J.
Quaternary structure of pig liver phosphofructokinase
J. Biol. Chem.
1980
, vol. 
255
 (pg. 
10537
-
10541
)
[PubMed]
43
Ling
K.H.
Marcus
F.
Lardy
H.A.
Purification and some properties of rabbit skeletal muscle phosphofructokinase
J. Biol. Chem.
1965
, vol. 
240
 (pg. 
1893
-
1899
)
[PubMed]
44
Paetkau
V.
Lardy
H.A.
Phosphofructokinase: correlation of physical and enzymatic properties
J. Biol. Chem.
1967
, vol. 
242
 (pg. 
2035
-
2042
)
[PubMed]
45
Mansour
T.E.
Wakid
N.W.
Sprouse
H.M.
Purification, crystallization and properties of activated sheep heart phosphofructokinase
Biochem. Biophys. Res. Commun.
1965
, vol. 
19
 (pg. 
721
-
727
)
[PubMed]
46
Mansour
T.E.
Wakid
N.W.
Sprouse
H.M.M.
Studies on heart phosphofructokinase
J. Biol. Chem.
1966
, vol. 
241
 (pg. 
1512
-
1521
)
[PubMed]
47
Li
Y.
Rivera
D.
Ru
W.
Gunasekera
D.
Kemp
R.G.
Identification of allosteric sites in rabbit phosphofructo-1-kinase
Biochemistry
1999
, vol. 
38
 (pg. 
16407
-
16412
)
[PubMed]
48
Kemp
R.G.
Gunasekera
D.
Evolution of the allosteric ligand sites of mammalian phosphofructo-1-kinase
Biochemistry
2002
, vol. 
41
 (pg. 
9426
-
9430
)
[PubMed]
49
Martinez-Costa
O.H.
Hermida
C.
Sanchez-Martinez
C.
Santamaria
B.
Aragon
J.J.
Identification of C-terminal motifs responsible for transmission of inhibition by ATP of mammalian phosphofructokinase, and their contribution to other allosteric effects
Biochem. J.
2004
, vol. 
377
 (pg. 
77
-
84
)
[PubMed]
50
Uyeda
K.
Phosphofructokinase
Adv. Enzym. Relat. Areas Mol. Biol.
1979
, vol. 
48
 (pg. 
193
-
244
)
51
Poorman
R.A.
Randolph
A.
Kemp
R.G.
Heinrikson
R.L.
Evolution of phosphofructokinase–gene duplication and creation of new effector sites
Nature
1984
, vol. 
309
 (pg. 
467
-
469
)
[PubMed]
52
Heinisch
J.
Ritzel
R.G.
von Borstel
R.C.
Aguilera
A.
Rodicio
R.
Zimmermann
F.K.
The phosphofructokinase genes of yeast evolved from two duplication events
Gene
1989
, vol. 
78
 (pg. 
309
-
321
)
[PubMed]
53
Shirakihara
Y.
Evans
P.R.
Crystal structure of the complex of phosphofructokinase from Escherichia coli with its reaction products
J. Mol. Biol.
1988
, vol. 
204
 (pg. 
973
-
994
)
[PubMed]
54
Lau
F.T.
Fersht
A.R.
Hellinga
H.W.
Evans
P.R.
Site-directed mutagenesis in the effector site of Escherichia coli phosphofructokinase
Biochemistry
1987
, vol. 
26
 (pg. 
4143
-
4148
)
[PubMed]
55
Chang
S.H.
Kemp
R.G.
Role of Ser530, Arg292, and His662 in the allosteric behavior of rabbit muscle phosphofructokinase
Biochem. Biophys. Res. Commun.
2002
, vol. 
290
 (pg. 
670
-
675
)
[PubMed]
56
Sánchez-Martínez
C.
Estévez
A.M.
Aragón
J.J.
Sa
C.
Este
A.M.
Arago
J.J.
Phosphofructokinase C isozyme from ascites tumor cells: cloning, expression, and properties
Biochem. Biophys. Res. Commun.
2000
, vol. 
271
 (pg. 
635
-
640
)
[PubMed]
57
Tornheim
K.
Lowenstein
J.M.
Control of phosphofructokinase from rat skeletal muscle. Effects of fructose diphosphate, AMP, ATP, and citrate
J. Biol. Chem.
1976
, vol. 
251
 (pg. 
7322
-
7328
)
[PubMed]
58
Kitajima
S.
Uyeda
K.
A binding study of the interaction of beta-D-fructose 2,6-bisphosphate with phosphofructokinase and fructose-1,6-bisphosphate
J. Biol. Chem.
1983
, vol. 
258
 (pg. 
7352
-
7357
)
[PubMed]
59
Tornheim
K.
Activation of muscle phosphofructokinase by fructose 2,6-bisphosphate and fructose 1,6-bisphosphate is differently affected by other regulatory metabolites
J. Biol. Chem.
1985
, vol. 
260
 (pg. 
7985
-
7989
)
[PubMed]
60
Sträter
N.
Marek
S.
Kuettner
E.B.
Kloos
M.
Keim
A.
Brüser
A.
Kirchberger
J.
Schöneberg
T.
Molecular architecture and structural basis of allosteric regulation of eukaryotic phosphofructokinases
FASEB J.
2011
, vol. 
25
 (pg. 
89
-
98
)
[PubMed]
61
Frieden
C.
Gilbert
H.R.
Bock
P.E.
Phosphofructokinase III. Correlation of the regulatory kinetic and molecular properties of the rabbit muscle enzyme
J. Biol. Chem.
1976
, vol. 
251
 (pg. 
5644
-
5647
)
[PubMed]
62
Arechaga
I.
Martínez-Costa
O.H.
Ferreras
C.
Carrascosa
J.L.
Aragón
J.J.
Martinez-Costa
O.H.
Aragon
J.J.
Electron microscopy analysis of mammalian phosphofructokinase reveals an unusual 3-dimensional structure with significant implications for enzyme function
FASEB J.
2010
, vol. 
24
 (pg. 
4960
-
4968
)
[PubMed]
63
Luther
M.A.
Gilbert
H.F.
Lee
J.C.
Self-association of rabbit muscle phosphofructokinase: role of subunit interaction in regulation of enzymatic activity
Biochemistry
1983
, vol. 
22
 (pg. 
5494
-
500
)
[PubMed]
64
Biophys
M.A.B.
York
N.
Biochem
R.A.E.J.
Luther
M.A.
Cai
G.Z.
Lee
J.C.
Thermodynamics of dimer and tetramer formations in rabbit muscle phosphofructokinase
Biochemistry
1986
, vol. 
25
 (pg. 
7931
-
7937
)
[PubMed]
65
Marinho-Carvalho
M.M.
Costa-Mattos
P.V.
Spitz
G.A.
Zancan
P.
Sola-Penna
M.
Calmodulin upregulates skeletal muscle 6-phosphofructo-1-kinase reversing the inhibitory effects of allosteric modulators
Biochim. Biophys. Acta
2009
, vol. 
1794
 (pg. 
1175
-
1180
)
[PubMed]
66
Mayr
G.W.
Interaction of calmodulin with muscle phosphofructokinase. Interplay with metabolic effectors of the enzyme under physiological conditions
Eur. J. Biochem.
1984
, vol. 
143
 (pg. 
521
-
529
)
[PubMed]
67
Mayr
G.W.
Heilmeyer
L.M.
Phosphofructokinase is a calmodulin binding protein
FEBS Lett.
1983
, vol. 
159
 (pg. 
51
-
57
)
[PubMed]
68
Buschmeier
B.
Meyer
H.E.
Mayr
G.W.
Characterization of the calmodulin-binding sites of muscle phosphofructokinase and comparison with known calmodulin-binding domains
J. Biol. Chem.
1987
, vol. 
262
 (pg. 
9454
-
9462
)
[PubMed]