Mutations in the Hinge of a Dynamic Loop Broadly Influence Functional Properties of Fructose-1,6-bisphosphatase*

Loop 52–72 of porcine fructose-1,6-bisphosphatase may play a central role in the mechanism of catalysis and allosteric inhibition by AMP. The loop pivots between different conformational states about a hinge located at residues 50 and 51. The insertion of proline separately at positions 50 and 51 reduces k cat by up to 3-fold, with no effect on the K m for fructose 1,6-bisphosphate. TheK a for Mg2+ in the Lys50→ Pro mutant increases ∼15-fold, whereas that for the Ala51 → Pro mutant is unchanged. Although these mutants retain wild-type binding affinity for AMP and the fluorescent AMP analog 2′(3′)-O-(trinitrophenyl)adenosine 5′-monophosphate, the K i for AMP increases 8000- and 280-fold in the position 50 and 51 mutants, respectively. In fact, the mutation Lys50 → Pro changes the mechanism of AMP inhibition with respect to Mg2+ from competitive to noncompetitive and abolishes K+ activation. The K i for fructose 2,6-bisphosphate increases ∼20- and 30-fold in the Lys50 → Pro and Ala51 → Pro mutants, respectively. Fluorescence from a tryptophan introduced by the mutation of Tyr57 suggests an altered conformational state for Loop 52–72 due to the proline at position 50. Evidently, the Pro50 mutant binds AMP with high affinity at the allosteric site, but the mechanism of allosteric regulation of catalysis has been disabled.

Loop 52-72 could play an important role in the allosteric mechanism of AMP inhibition, existing in at least two conformational states (9,10). In the absence of AMP, the loop interacts with the active site (engaged loop conformation), positioning the side chain of an essential catalytic residue, Asp 74 (9,11). The binding of AMP to its allosteric pocket putatively displaces the loop from the active site and stabilizes the disengaged loop conformation (9,10). Subunits with disengaged loops evidently have reduced affinity for essential metal ions and, as a consequence, may be impaired with respect to catalysis. Mutations of conserved residues in Loop 52-72 have significant effects on catalysis and allosteric inhibition of catalysis by AMP (11).
The metabolic flux through FBPase is sensitive to prevailing levels of AMP and Fru-2,6-P 2 . AMP inhibits FBPase allosterically with a Hill coefficient of 2 (12,13), whereas Fru-2,6-P 2 inhibits FBPase by direct ligation of the active site with a Hill coefficient of unity (1)(2)(3). Inhibition of FBPase by AMP is nonlinear and noncompetitive with respect to Fru-1,6-P 2 , but nonlinear and competitive with respect to Mg 2ϩ (6,14). In contrast, inhibition of FBPase by Fru-2,6-P 2 is linear and competitive with respect to Fru-1,6-P 2 and noncompetitive with respect to Mg 2ϩ . Due to the equilibrium reaction catalyzed by adenylate kinase, concentrations of AMP in vivo are nearly constant under normal physiological conditions, whereas the concentration of Fru-2,6-P 2 varies in response to extracellular regulators such as glucagon and epinephrine. Hence, Fru-2,6-P 2 is the probable dynamic regulator of FBPase in vivo, but AMP still plays a significant role. Fru-2,6-P 2 lowers the apparent inhibition constant for AMP at least 10-fold (15) by decreasing the k off of AMP (16). Levels of AMP association with the allosteric sites of FBPase will then necessarily change in response to Fru-2,6-P 2 levels, even if the concentration of AMP in vivo is constant. AMP/Fru-2,6-P 2 synergism may arise from the ability of each ligand to stabilize a common thermodynamic state of FBPase (disengaged Loop 52-72) through different mechanisms (9).
For Loop 52-72 to move between its various states, significant conformational change must occur in hinge elements at residues 50 and 51 and residues 71 and 72. The combined mutation of lysines 71 and 72 to methionine dramatically increases the K i for AMP, presumably due to the stabilization of the R-state over the T-state conformer of FBPase (11). If the loop-mediated mechanism of AMP inhibition is valid, mutations at the second hinge element must qualitatively cause a similar increase in the K i for AMP, yet mutations of Lys 50 to methionine (17) and glutamine and alanine (18) cause virtually no change in the K i for AMP. What is clear from crystal structures, however, is a large difference in main chain angles (, ) at position 50 in the R-and T-states of FBPase (7,9,10). Here we explore the consequence of mutations Lys 50 3 Pro and Ala 51 3 Pro, which directly influence the backbone conformation of residues 50 and 51. The mutations have little effect on the binding affinity of a fluorescent analog of AMP, but increase the kinetic K i value for that analog by several orders of magnitude. The kinetic K m value for Fru-1,6-P 2 is unchanged by the mutations, but the K a for Mg 2ϩ rises 15-fold in the case of the Pro 50 mutant and is 40-fold less active under standard conditions of assay. The data suggest that Loop 52-72 is an essential element in the allosteric mechanism of FBPase.
Mutagenesis of Wild-type FBPase-Mutations were accomplished by specific base changes in double-stranded plasmid using the Transformer TM site-directed mutagenesis kit (CLONTECH). The mutagenic primers for the Lys 50 3 Pro, Ala 51 3 Pro, and Tyr 57 3 Trp mutants were 5Ј-GTCCGCCCGGCGGGCATC-3Ј, 5Ј-CCGCAAGCCGGGCATCGC-3Ј, and 5Ј-CGCACCTCTGGGGAATTG3Ј, respectively (codons for mutation are in boldface and underlined). The selection primer 5Ј-CAGC-CTCGCCTCGAGAACGCCA-3Ј (the digestion site is in boldface and underlined) changed an original NruI site on the plasmid into a XhoI site. The mutation and integrity of the gene were confirmed by sequencing the entire gene. The Iowa State University sequencing facility provided DNA sequences using the fluorescent dye dideoxy terminator method.
Expression and Purification of Wild-type and Mutant FBPases-Protein expression and purification were performed as described previously (11), but with minor alterations. After cell breakage, FBPase was enriched and concentrated by 70% ammonium sulfate fractionation. The precipitate was taken up in 20 mM Tris (pH 7.5) and desalted on a Sephadex G-100 column using 20 mM Tris (pH 7.5). The active fractions were loaded directly onto a Cibacron blue column and eluted with a 0.5-1 M NaCl gradient. FBPase eluted at ϳ800 mM NaCl. The eluent was dialyzed against 20 mM Tris-HCl (pH 8.3), loaded onto a DEAE-Sephadex column, and then developed with a 0 -300 mM NaCl gradient. FBPase eluted at a salt concentration of 150 mM, and its purity was confirmed by SDS-polyacrylamide gel electrophoresis (19). Protein concentration was determined using the Bradford assay (20) with bovine serum albumin as a standard.
Circular Dichroism Spectroscopy-CD studies on wild-type and mutant FBPases were done at room temperature on a Jasco J710 CD spectrometer in a 1-cm cell using a protein concentration of 0.35 mg/ml. Spectra were collected from 200 to 260 nm in increments of 1.3 nm. Each spectrum was blank-corrected and smoothed using the software package provided with the instrument.
Kinetic Experiments-Assays for the determination of specific activity, k cat , and activity ratios at pH 7.5 and 9.5 employed the coupling enzymes phosphoglucose isomerase and glucose-6-phosphate dehydrogenase (1). The reduction of NADP ϩ to NADPH was monitored directly at 340 nm. All other assays employed the same coupling enzymes, but monitored NADPH evolution by its fluorescence emission at 470 nm (excitation wavelength of 340 nm). Initial rate data were analyzed using either programs written in the MINITAB language using an ␣ value of 2.0 (14) or ENZFITTER (21). The kinetic data were fit to a variety of models. Only models and parameters for the best fits are reported below.
Steady-state Fluorescence Measurements-Fluorescence data were collected using an SLM 8100C fluorometer (Spectronic Instruments). The single tryptophan in the Trp 57 mutant was excited selectively using radiation of 295-nm wavelength. Fluorescence emission spectra were recorded in steps of 1 nm from 310 to 400 nm using a slit width of 2 mm for both excitation and emission. Each spectrum was an average of five scans. Fluorescence from TNP-AMP employed excitation and emission wavelengths of 400 and 535 nm, respectively. Conditions under which specific spectra were recorded are provided below and in the tables and figure legends. Each point is the average of 15 1-s acquisitions. Enzyme concentrations ranged from 0.3 to 0.5 M. All spectra were corrected for dilution and inner filter effects (absorbance of light by AMP and TNP-AMP) using Equation 1 (22), where F c is the corrected fluorescence, F is the fluorescence intensity experimentally measured, B is the background, V i is the volume of the sample for a specific titration point, V 0 is the initial volume of the sample, A ex is the absorbance measured at the wavelength of excitation (295 nm for AMP and 410 nm for TNP-AMP), and A em is the absorbance measured at the wavelength of emission (535 nm for TNP-AMP). As a control, ligands by themselves caused no change in fluorescence emission from a solution of tryptophan (100 M). TNP-AMP titration data were analyzed by nonlinear least-squares fits using Equation 2, where ⌬F is the change in fluorescence caused by the addition of ligand (L), F 0 is the fluorescence in the absence of ligand, ⌬F max is the theoretical maximum change in fluorescence at saturating ligand concentration, K d is the dissociation constant, and n is the Hill coefficient. Structural Modeling of FBPase Mutants-The starting model for the engaged conformation of Loop 52-72 is the crystal structure of the Zn 2ϩ ⅐product complex of FBPase (PDB accession code 1CNQ), in which the tetramer assumes the R-state conformation (9). The starting model for the disengaged conformation of Loop 52-72 is a Zn 2ϩ ⅐product⅐AMP complex of FBPase, in which the tetramer has the T-state conformation (10). The latter structure is similar to other T-state structures of FB-Pase (7,23), differing primarily in the ligands present at the active site. Proline residues were placed individually at positions 50 and 51 using the software XTALVIEW and a Silicon Graphics workstation. The initial models were passed through three protocols: 1) 100 cycles of conjugate gradient minimization; 2) 100 cycles of conjugate gradient minimization, followed by simulated annealing (starting temperature, 500 K; final temperature, 300 K) (24) and then another 100 cycles of conjugate gradient minimization; and 3) the same procedure as Protocol 2, but using a starting temperature of 1000 K. During simulated annealing and energy minimization, the C ␣ atoms (except for residues 50 -72) were restrained to their initial crystallographic positions by a harmonic pseudo-potential with a force constant of 50 kcal/mol. The justification for restraining all C ␣ atoms outside of the loop is the 10-year history of FBPase crystal structures, which collectively reveal only small changes in the protein conformation outside of Loop 52-72. Fig. 1. The main chain angle of Lys 50 differs significantly in the T-and R-states, neither value being compatible with that of proline, where must be close to Ϫ60°. Insertion of proline then results in an initial model with poor stereochemistry at position 50. Energy minimization alone and energy minimization with simulated annealing at a starting temperature of 500 K (Protocols 1 and 2 in "Experimental Procedures") leave the angle of Pro 50 some 30°a way from its preferred value in both the T-and R-states. Further improvement in the stereochemistry of Pro 50 requires a starting temperature of 1000 K in simulated annealing (Protocol 3), the consequence of which is a significant drift of the Loop 52-72 away from either its T-or R-state conformation. Evidently, proline at position 50 is incompatible with the T-state disengaged and R-state engaged loop conformations.

Modeling of FBPase Mutants-A structural overview of the region of interest is provided in
Proline at position 51, on the other hand, is less disruptive. The angle of residue 51 is near Ϫ60°in both the R-and T-states of FBPase. Backbone amide 51 hydrogen bonds with backbone carbonyls 47 and 48 in the R-state and with backbone carbonyl 46 in the T-state. Proline substitution at position 51 clearly disrupts the hydrogen bonds in both the R-and T-states. Moreover, the C ␦ atom of proline makes close contacts of 1.9 and 1.8 Å with backbone carbonyls 48 (R-state) and 46 (T-state), respectively. Either energy minimization alone or energy minimization with simulated annealing (initial temperature, 500 K) relaxes the contact in the R-state (the loop re-mains engaged with the active site) by a 90°rotation of the plane of the polypeptide linkage between residues 47 and 48. The close contact of the T-state remains, however, at 2.4 and 2.8 Å after modeling Protocols 1 and 2, respectively. Contacts from within the subunit and from the neighboring subunit constrain the T-state conformation of residues 52-57 (Fig. 1). Furthermore, as backbone carbonyl 46 is an integral part of helix H2, it cannot relieve its close contact with Pro 51 by a rotation of its polypeptide linkage. Hence, proline at position 51 probably increases the internal energy of the T-state, destabilizing it relative to the R-state of FBPase.
Expression and Purification of Wild-type and Mutant FB-Pases-Mutant and wild-type FBPases behaved identically throughout purification, including gel exclusion chromatography. The wild-type and mutant enzymes were at least 95% pure with no evidence for proteolysis on the basis of SDS-polyacrylamide gel electrophoresis.
Secondary Structure Analysis-The CD spectra of wild-type FBPase in the T-and R-states diverge minimally, but reproducibly in the vicinity of 210 nm (17). The CD spectra of the Ala 51 3 Pro, Lys 50 3 Pro/Tyr 57 3 Trp, and wild-type FBPases superimpose in the presence of P i (5 mM), Fru-6-P (5 mM), and saturating Mg 2ϩ (5 mM for the wild-type and Ala 51 3 Pro enzymes and 35 mM for Lys 50 3 Pro/Tyr 57 3 Trp). The addition of AMP (200 M) to the wild-type enzyme produced small changes in the CD spectrum near 210 nm, as noted above, but caused no change in the CD spectra of the Lys 50 3 Pro/Tyr 57 3 Trp and Ala 51 3 Pro enzymes. (Concentrations of AMP in excess of 200 M degraded CD spectra due to the elevated absorbance of radiation.) The CD spectra of the Lys 50 3 Pro mutant also differ from the corresponding wild-type spectra (Fig. 2). The CD spectrum of the Lys 50 3 Pro mutant did not change in the presence of P i (5 mM), Fru-6-P (5 mM), and saturating Mg 2ϩ (50 mM) in the presence or absence of AMP (200 M).
Catalytic Rates, Michaelis Constants for Mg 2ϩ and Fru-1,6-P 2 , and K ϩ Activation-Initial rate kinetics employ maximal substrate concentrations, sufficient to saturate the active site, but low enough to avoid substrate inhibition. The ratios of catalytic rate constants at pH 7.5 and 9.6 are comparable for wild-type and Ala 51 3 Pro FBPases (Table I), both consistent with reported activity ratios for an FBPase free of proteolysis. The activity ratios for the Lys 50 3 Pro and Lys 50 3 Pro/Tyr 57 3 Trp mutants are low, however; but as discussed below, the low value does not stem from limited proteolysis of that mutant. All enzymes have comparable K m values for Fru-1,6-P 2 (Table I); but beyond this, the quantitative similarities end. Under standard conditions of assay for the wild-type enzyme (5 mM Mg 2ϩ and 20 M Fru-1,6-P 2 , pH 7.5), the mutation of Lys 50 to proline resulted in a 40-fold loss of specific activity. However, the decline in specific activity was due primarily to a 15-fold increase in the K a for Mg 2ϩ for the Pro 50 mutant. The k cat value for the Pro 50 mutant (saturated with Mg 2ϩ ) is one-third of that for the wild-type enzyme. The Hill coefficients for Mg 2ϩ for the Ala 51 3 Pro, Lys 50 3 Pro/Tyr 57 3 Trp, and wild-type enzymes are similar, but are reduced significantly in the Lys 50 3 Pro mutant. Potassium ion activation is retained in the Ala 51 3 Pro and Lys 50 3 Pro/Tyr 57 3 Trp mutants at levels comparable to that of the wild-type enzyme, but is abolished altogether in the Pro 50 mutant.
Kinetics of AMP Inhibition-Concentrations of AMP needed for 50% inhibition increased 280-, 8000-, and 400-fold relative to that of the wild-type enzyme for the Ala 51 3 Pro, Lys 50 3 Pro, and Lys 50 3 Pro/Tyr 57 3 Trp mutants, respectively ( Fig.  3 and Table I). Additionally, the Lys 50 3 Pro/Tyr 57 3 Trp double mutant displays biphasic behavior toward AMP. The data of Fig. 3 for the wild-type, Ala 51 3 Pro, and Lys 50 3 Pro enzymes were fit to Equation 3, where v is the observed velocity at a specific concentration of AMP, V 0 is the fitted velocity in the absence of AMP, I is the concentration of AMP, and IC 50 is the concentration of AMP that causes 50% inhibition. The exponent of 2 represents the Hill coefficient for AMP cooperativity, determined independently as described below. The data of Fig. 3  where v, V 0 , and I are as defined above for Equation 3 and IC 50-high and IC 50-low represent concentrations of AMP that cause 50% relative inhibition due to the ligation of high and low affinity sites, respectively. Ligation of the high affinity sites, as shown below, is cooperative with a Hill coefficient of 2, whereas the cooperativity with respect to the ligation of low affinity sites is an adjustable parameter (n) in Equation 4. The fitted value for n is 1.4. AMP inhibition of wild-type FBPase is nonlinear and noncompetitive with respect to Fru-1,6-P 2 and nonlinear and competitive with respect to Mg 2ϩ . The Ala 51 3 Pro mutant retains the nonlinear, competitive relationship between AMP and Mg 2ϩ (goodness-of-fit, 5.5%) (Fig. 4), consistent with Equation 5, where v, V m , A, I, K a , K i , and n represent initial velocity, maximal velocity, Mg 2ϩ concentration, AMP concentration, the Michaelis constant for Mg 2ϩ , the dissociation constant for AMP from the enzyme⅐AMP complex, and the Hill coefficient for AMP, respectively. The binding of AMP is cooperative at n ϭ 2, whereas cooperativity is absent when n ϭ 1. Data from the wild-type and Ala 51 3 Pro enzymes are consistent with n ϭ 2 (AMP cooperativity). When assayed at low AMP concentrations, Lys 50 3 Pro/Tyr 57 3 Trp retains the competitive mech-  anism with n ϭ 2. In contrast, AMP inhibition of the Lys 50 3 Pro mutant fits best to a mechanism of noncompetitive inhibition with respect to Mg 2ϩ (goodness-of-fit, 3.1%) (Fig. 4), represented by Equation 6, where the terms of Equation 6 are defined as for Equation 5 and K ii represents the dissociation constant for AMP from the enzyme⅐Mg 2ϩ ⅐AMP complex. The data are consistent with n ϭ 2, reflecting AMP cooperativity. In wild-type, Ala 51 3 Pro, and Lys 50 3 Pro/Tyr 57 3 Trp FBPases, the mechanism of AMP inhibition with respect to Fru-1,6-P 2 is nonlinear and noncompetitive (Equation 7), where v, V m , I, and n are as defined above, K i and K ii represent the dissociation constants for AMP from the enzyme⅐AMP and the enzyme⅐Fru-1,6-P 2 ⅐AMP complexes, respectively, B is the concentration of Fru-1,6-P 2 , and K b is the Michaelis constant for Fru-1,6-P 2 . The high value for IC 50 associated with AMP inhibition of the Lys 50 3 Pro mutant (Table I) precludes the determination of a kinetic mechanism.
Steady-state Fluorescence Measurements-TNP-AMP inhibited the wild-type enzyme competitively with respect to Mg 2ϩ and inhibited FBPase synergistically with Fru-2,6-P 2 . Fluorescence emission from TNP-AMP increased significantly in the presence of FBPase. In addition, AMP reduced the fluorescence from bound TNP-AMP, presumably through competition for the allosteric effector site (data not shown). TNP-AMP bound to wild-type, Lys 50 3 Pro, and Ala 51 3 Pro FBPases with high    6 for B using the parameters of Table I. A.U., arbitrary units. affinity, comparable to AMP. On the basis of fluorescence titration data (Table II), however, the Hill coefficient for TNP-AMP is near unity in all cases.
FBPase has no tryptophan, so the mutation of any single residue to tryptophan introduces a unique fluorophore. Fluorescence emission from tryptophan at position 57 is sensitive to the conformational state of Loop 52-72. The indole of Trp 57 is exposed to solvent in the T-state disengaged loop conformation, whereas it resides in a hydrophobic pocket in the R-state engaged loop conformation. The Tyr 57 3 Trp mutant has been thoroughly studied by x-ray crystallography and initial velocity kinetics (25). Fluorescence emission spectra from the Pro 50 / Trp 57 double mutant differ significantly from those of the Trp 57 single mutant (Fig. 6). Fluorescence emission from the single Trp 57 mutant is maximum in the presence of products/metals, conditions that promote the R-state engaged loop conformation (9,10). The addition of AMP reduced fluorescence emission to a level comparable to that observed from the Trp 57 mutant in the absence of products and metals (apo form of FBPase). In contrast, the apo form of the Pro 50 /Trp 57 double mutant has the highest fluorescence emission. The addition of either product/ metals or product/metals/AMP caused only a decrease in fluorescence emission. DISCUSSION Low activity ratios (ϳ0.3) are a property of wild-type FBPase after limited proteolysis by papain or subtilisin (26). The maximum effect on the activity ratio due to proteolysis occurs when one subunit per tetramer is cut. Thus, for the wild-type enzyme, an activity ratio of 1, as observed for the Lys 50 3 Pro mutant, requires an equal mixture of intact and once-proteolyzed tetramers. Proteolysis of one of eight subunits then could account for a diminished activity ratio in any FBPase. In SDSpolyacrylamide gel electrophoresis of the Lys 50 3 Pro mutant, however, no fragments typical of proteolysis were evident. Furthermore, loss of AMP inhibition due to proteolysis of wild-type FBPase varied proportionately with the extent of subunit cleavage. The 8000-fold increase in K i for AMP far exceeds the effect on AMP inhibition due to proteolysis, even in a completely proteolyzed system. Hence, the kinetic properties of the Lys 50 3 Pro mutant are not the consequence of proteolysis, but rather the influence of a proline at position 50.
Allosteric inhibition of wild-type FBPase by TNP-AMP is essentially identical to that caused by AMP, the only difference being the absence of cooperativity in the kinetics and in the binding of the fluorescent analog. In fact, TNP-AMP is indistinguishable from formycin 5Ј-monophosphate in its inhibition and binding to wild-type FBPase. Formycin 5Ј-monophosphate also exhibits a Hill coefficient of unity (27), and crystal structures of complexes of formycin 5Ј-monophosphate and AMP with human FBPase are identical (28). The difference in Hill coefficient may stem from subtle differences in conformation, evident perhaps only in partially ligated states of FBPase. Nonetheless, the fluorescence data clearly demonstrate tight binding of TNP-AMP and the displacement of that analog from FBPase by the addition of AMP. Hence, wild-type and Pro 50 mutant FBPases are indistinguishable in their binding of TNP-AMP and its displacement by AMP.
On the basis of kinetics, however, Pro 50 mutant FBPase is altogether insensitive to AMP. Inhibition of the Pro 50 mutant by AMP requires an 8000-fold increase in ligand concentration and exhibits a different kinetic mechanism (noncompetitive versus competitive with respect to Mg 2ϩ ). The apparent change in kinetic mechanism may arise from the unmasking of a secondary mechanism, due to the complete loss of allosteric inhibition by AMP. That secondary mechanism may be the direct coordination of AMP to the metal-ligated active site of Pro 50 mutant FBPase. Hence, if AMP binds tightly to the Pro 50 mutant, as evidenced by the fluorescence data, but does not inhibit, then proline at position 50 must interrupt the transmission of the allosteric signal.
Kinetic data indicate the disruption of Loop 52-72 as the root cause for the loss of allosteric properties in Pro 50 mutant FB-Pase. The K a for Mg 2ϩ is elevated 15-fold in the Pro 50 mutant, and K ϩ activation is lost. The kinetic mechanism of AMP inhibition with respect to Mg 2ϩ is competitive in wild-type FBPase. Hence, in wild-type FBPase, AMP elevates the apparent dissociation constant of metals. A greatly elevated K a for Mg 2ϩ , accompanied by the complete loss of AMP inhibition, is consistent with an FBPase unable to achieve its high affinity state for metal cations. On the basis of recent crystal structures (10), the R-state engaged conformation of Loop 52-72 represents the high affinity state for metal cations and the catalytically productive state of FBPase. Pro 50 mutant FBPase then cannot achieve a engaged loop conformation.
Molecular modeling, CD spectroscopy, and fluorescence emission from Trp 57 strengthen the argument above. The mutation of Lys 50 to proline destabilizes the R-state engaged loop and T-state disengaged loop conformations in slow cooling/ energy minimization protocols. The CD spectrum of Pro 50 mutant FBPase supports an altered state, conformationally unresponsive to the binding of ligands. The indole group of Trp 57 in the Pro 50 /Trp 57 double mutant does not enter its hydrophobic pocket, as evidenced by the low fluorescence emission in the presence of saturating products/metals. The several lines of evidence provided here suggest that Loop 52-72 in Pro 50 mutant FBPase can achieve neither its T-state disengaged nor its R-state engaged loop conformation. The complete loss of allosteric inhibition in Pro 50 mutant FBPase suggests furthermore, that Loop 52-72 is an essential element in the allosteric regulation of FBPase.
The effects of the Ala 51 3 Pro mutation are less dramatic, in harmony with the results of modeling and CD spectroscopy. The Pro 51 mutant retains K ϩ activation and wild-type affinity for divalent cations, but exhibits a significant increase in the K i for AMP. The mutation evidently destabilizes the T-state disengaged conformation of Loop 52-72 relative to the R-state engaged conformation, again consistent with the results of modeling. In this respect, the Pro 51 mutant and the Met 71/72 double mutant (11) have similar effects on FBPase function. In both instances, lessened AMP inhibition probably arises from a perturbation in the equilibrium between engaged and disengaged loop conformations in favor of the engaged state.
The K i for Fru-2,6-P 2 increases 20-and 30-fold in the Lys 50 3 Pro and Ala 51 3 Pro mutants, respectively. The change in binding affinity for Fru-2,6-P 2 in these mutants must be indirect and is most likely through a perturbation of conformation states accessible to Loop 52-72. The increase in K i for Fru-2,6-P 2 due to loop mutations does not come with an increase in the K m for Fru-1,6-P 2 , as has been observed in other mutations that increase the K i for Fru-2,6-P 2 (29). The observation above is consistent with findings that suggest that Fru-2,6-P 2 and Fru-1,6-P 2 , although binding at the same site, evoke different conformational responses from FBPase. For instance, at high concentrations of Fru-2,6-P 2 , the Hill coefficient for AMP inhibition changes from 2 to 1, and the saturation curve for Fru-1,6-P 2 changes from hyperbolic to sigmoidal (3,30). At low concentrations of Fru-2,6-P 2 , the Hill coefficient for Mg 2ϩ changes from 2 to 1 (31). Finally, Fru-2,6-P 2 protects FBPase from loss of AMP inhibition and catalytic activity due to modification by acetylimidazole, whereas Fru-1,6-P 2 protects only against loss of activity (31). Although a complete understanding of the above phenomena will come with further study, mutations of the loop underscore a fundamental difference in FBPase in its Fru-2,6-P 2 -and Fru-1,6-P 2 -ligated states.