Directed Mutations in the Poorly Defined Region of Porcine Liver Fructose-1,6-bisphosphatase Significantly Affect Catalysis and the Mechanism of AMP Inhibition*

Asn64, Asp68, Lys71, Lys72, and Asp74 of porcine liver fructose-1,6-bisphosphatase (FBPase) are conserved residues and part of a loop for which no electron density has been observed in crystal structures. Yet mutations of the above dramatically affect catalytic rates and/or AMP inhibition. The Asp74 → Ala and Asp74 → Asn mutant enzymes exhibited 50,000- and 2,000-fold reductions, respectively, in k catrelative to wild-type FBPase. The pH optimum for the catalytic activity of the Asp74 → Glu, Asp68 → Glu, Asn64 → Gln, and Asn64 → Ala mutant enzymes shifted from pH 7.0 (wild-type enzyme) to pH 8.5, whereas the Lys71 → Ala mutant and Lys71,72 → Met double mutant had optimum activity at pH 7.5. Mg2+cooperativity, K m for fructose 1,6-bisphosphate, and K i for fructose 2,6-bisphosphate were comparable for the mutant and wild-type enzymes. Nevertheless, for the Asp74 → Glu, Asp68 → Glu, Asn64→ Gln, and Asn64 → Ala mutants, the binding affinity for Mg2+ decreased by 40–125-fold relative to the wild-type enzyme. In addition, the Asp74 → Glu and Asn64 → Ala mutants exhibited no AMP cooperativity, and the kinetic mechanism of AMP inhibition with respect to Mg2+ was changed from competitive to noncompetitive. The double mutation Lys71,72 → Met increasedK i for AMP by 175-fold and increased Mg2+ affinity by 2-fold relative to wild-type FBPase. The results reported here strongly suggest that loop 51–72 is important for catalytic activity and the mechanism of allosteric inhibition of FBPase by AMP.

In mammals, FBPase is a homotetramer with a subunit M r of 37,000 (8). Each subunit of the tetramer (designated C1, C2, C3, and C4) has an allosteric AMP domain (residues 1-200) and a catalytic Fru-1,6-P 2 domain (residues 201-335), with the AMP-binding site being ϳ28 Å away from the substrate-binding site (9 -11). Structures of AMP complexes of the enzyme define the T-state, whereas structures in the presence substrate analogs and without AMP represent the R-state. The structural transition (R-to T-state) involves a 17°rotation of the C1-C2 dimer with respect to the C3-C4 dimer and a 1.9°r otation of the AMP domain relative to the Fru-1,6-P 2 domain within each subunit (10). The R-to T-state transition results in conformational changes at interfaces between subunits C1 and C2 and subunits C1 and C4 (as well as interfaces related to these by the symmetry of the tetramer). Metal-binding sites (up to three total) are at or near the active site at the interface between the two domains of the enzyme (12,13). Mg 2ϩ and AMP are mutually exclusive in their binding to FBPase (14,15). In fact, AMP inhibition is nonlinear and noncompetitive with respect to Fru-1,6-P 2 and nonlinear and competitive with respect to Mg 2ϩ . Yet crystallographic studies reveal similar metal coordination in the absence and presence AMP (either one Mg 2ϩ ion or two Zn 2ϩ or two Mn 2ϩ ions), although small perturbations in the active site are attributed to the 1.9°rotation of the AMP relative to the Fru-1,6-P 2 domain due to AMP binding (13).
Site-directed mutagenesis of the residues involved at subunit interfaces of porcine liver FBPase revealed significant changes in the kinetic mechanism of AMP inhibition and cooperativity (16 -19). However, a detailed mechanism of catalysis and allosteric regulation based on changes in interacting residues has been elusive. Perhaps published structures of FBPase do not reveal all of the essential structural elements for catalysis and regulation. Mutation of Arg 49 , a residue that precedes a disordered loop (residues 54 -71), causes dramatic changes in FB-Pase kinetics (17). Comparison of the amino acid sequences of mammalian FBPases reveals Gly 58 , Gly 61 , Asn 64 , Asp 68 , Asp 74 , Lys 71 , and Lys 72 as conserved residues within this disordered loop. Furthermore, loop 54 -71 is proteolytically sensitive in all known FBPases (20). To determine whether loop 54 -71 plays a major role in FBPase kinetics, mutant enzymes Asn 64 3 Gln, Asn 64 3 Ala, Asp 68 3 Glu, Lys 71 3 Ala, Lys 71,72 3 Met (double mutation), Asp 74 3 Ala, Asp 74 3 Asn, and Asp 74 3 Glu were expressed in Escherichia coli, purified to homogeneity, and evaluated by initial velocity kinetics. The above mutations have profound effects on both catalysis and regulation of FBPase.

EXPERIMENTAL PROCEDURE
Materials-Fru-1,6-P 2 , Fru-2,6-P 2 , NADP, AMP, ampicillin, tetracycline, and isopropyl-␤-D-thiogalactopyranoside were purchased from Sigma. DNA-modifying and restriction enzymes, T4 polynucleotide ki-* This work was supported by Research Grant NS 10546 from the National Institutes of Health and Grant MCB-9603595 from the National Science Foundation. This is Journal Paper J-17874 of the Iowa Agriculture and Home Economics Experiment Station (Ames, IA), Project 3191. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Expression and Purification of Wild-type and Mutant FBPases-The expression and purification of wild-type and mutant forms of FBPase were carried out as described previously (22) with slight modification. After centrifugation to remove cell debris, the supernatant was subjected to heat treatment, 30 -70% ammonium sulfate precipitation, Sephadex G-200 column chromatography, and CM-Sepharose column chromatography. Wild-type and mutant FBPases eluted as a single peak from the CM-Sepharose column by a NaCl gradient from 20 to 400 mM in 10 mM malonate buffer, pH 6.0. Protein purity was evaluated by 12% SDS-polyacrylamide gel electrophoresis according to Laemmli (23). Protein concentrations were determined as described by Bradford (24) using bovine serum albumin as the standard.
Kinetic Studies-Specific activity during purification was determined by the phosphoglucoisomerase and glucose-6-phosphate dehydrogenase coupled spectrometric assay at either pH 7.5 or 9.6 (1). All other kinetic experiments were done at either pH 7.5 or 8.5 and 25°C using a coupled spectrofluorometric assay (15). Initial rate data were analyzed using a computer program written in MINITAB language with an ␣ value of 2.0 (25). Cooperativity was evaluated using either the ENZFITTER program (26) or the MINITAB program.
Circular Dichroism Spectrometry-CD studies on the wild-type and mutant FBPases were carried out in a Jasco J710 CD spectrometer in a 1-mm cell at room temperature using a 0.2 mg/ml concentration of the enzyme. Spectra were collected from 200 to 260 nm in increments of 1.3 nm, and each spectrum was blank-corrected and smoothed using the software package provided with the instrument.

Purification of Wild-type and Mutant Forms of FBPase-
Purification of mutant and wild-type enzymes followed previous protocols (22). Mutant enzymes exhibited elution patterns similar to wild-type FBPase, except for Lys 71 3 Ala and Lys 71,72 3 Met, which eluted from the CM-Sepharose column at 100 mM NaCl. On the basis of SDS-polyacrylamide gel electrophoresis, all enzymes exhibit identical mobilities (M r ϳ 37, 500) with no evidence of proteolysis (27) and purity greater than 95% (data not shown). Activity ratios (pH 7.5:9.6) for the wild-type, Lys 71 3 Ala, and double mutant Lys 71,72 3 Met enzymes confirmed the absence of proteolysis.
Secondary Structure Analysis-The CD spectra of wild-type and mutant FBPases are essentially superimposable from 200 to 260 nm (data not shown), indicating the absence of global structural alteration changes caused by the mutations.
Catalytic Rates of FBPase Mutants-Initial rate studies were done at saturating concentrations of Fru-1,6-P 2 or Mg 2ϩ that do not cause substrate inhibition. Kinetic parameters are in Table I. The k cat values for the Asp 74 3 Ala and Asp 74 3 Asn mutants decreased 50,000 and 2,000 times, respectively, whereas a shift in the pH of optimum activity from pH 7.5 to 8.5 was exhibited by the Asp 74 3 Glu mutant with only a 20-fold reduction in k cat . A similar alteration in the pH of optimum activity occurred with mutants Asn 64 3 Gln, Asn 64 3 Ala, and Asp 68 3 Glu. On the other hand, replacements Lys 71 3 Ala and Lys 71,72 3 Met did not alter the pH of optimum activity. In Table I, the kinetic parameters for the Asp 74 3 Glu, Asn 64 3 Gln, Asn 64 3 Ala, and wild-type FBPases were measured at pH 8.5, and for mutants Asp 68 3 Glu, Lys 71 3 Ala, and Lys 71,72 3 Met at pH 7.5. A slight change in Fru-1,6-P 2 affinity was observed in all mutants relative to wild-type FBPase.
Mg 2ϩ Activation-The activity of wild-type FBPase as a function of Mg 2ϩ concentration is sigmoidal at neutral pH with a Hill coefficient of ϳ2, but is hyperbolic at pH 9.6 (3, 4). The Hill coefficient for Mg 2ϩ and the K a for Mg 2ϩ of the wild-type and mutant forms of FBPase were determined here at a saturating Fru-1,6-P 2 concentration (30 M) at pH 7.5 or 8.5. Hill coefficients at pH 8.5 dropped from 2.2 for the wild-type enzyme to ϳ1.  Fru-2,6-P 2 Inhibition-Fru-2,6-P 2 is a competitive inhibitor of Fru-1,6-P 2 and competes with the substrate for the active site of FBPase (7,29). K i values for Fru-2,6-P 2 decreased by Ͻ4-fold for all mutant enzymes compared with the wild-type enzyme. On the other hand, the Lys 71,72 3 Met mutant exhibited a 7-fold increase in K i for Fru-2,6-P 2 .
Kinetics of AMP Inhibition-AMP is an allosteric regulator of FBPase (30). The action of AMP inhibition is nonlinear and noncompetitive with respect to Fru-1,6-P 2 (31), but nonlinear and competitive relative to Mg 2ϩ for wild-type FBPase at either neutral or alkaline pH (15). AMP binding to FBPase is cooperative with a Hill coefficient of ϳ2 (4). The wild-type and Lys 71 3 Ala, Lys 71,72 3 Met, Asp 68 3 Glu, and Asn 64 3 Gln FBPases exhibited competitive inhibition patterns for AMP relative to Mg 2ϩ , whereas the Asp 74 3 Glu and Asn 64 3 Ala mutants showed noncompetitive inhibition. Fig. 1 shows double-reciprocal plots of 1/velocity versus 1/[Mg 2ϩ ] 2 at various fixed concentrations of AMP for the Lys 71,72 3 Met mutant of FBPase. The data of Fig. 1 are consistent with a steady-state random mechanism represented by Equation 1, where v, V m , A, B, I, K a , K b , K ia , K i , K ii , K iii , and K iv represent initial velocity; maximal velocity; the concentrations of Mg 2ϩ , Fru-1,6-P 2 , and AMP; the Michaelis constants for Mg 2ϩ and Fru-1,6-P 2 ; the dissociation constant for Mg 2ϩ ; and the dissociation constants for AMP from the enzyme⅐AMP, enzyme⅐ AMP⅐AMP, enzyme⅐Fru-1,6-P 2 ⅐AMP, and enzyme⅐Fru-1,6-P 2 ⅐ AMP⅐AMP complexes, respectively. n represents the Hill coefficient for AMP. When n ϭ 2, the binding of AMP to FBPase is cooperative; on the other hand, there is no cooperativity when n ϭ 1. In the case of the wild-type, Lys 71 3 Ala, Asp 68 3 Glu, and Asn 64 3 Gln FBPases, the kinetic data (similar to those shown in Fig. 1) fit best to Equation 1 when n ϭ 2. The "goodness of fit" was 4% when n ϭ 2 as opposed to 11% when n ϭ 1.
where v, V M , A, I, K a , K i , and K ii are defined as above, and n represents the Hill coefficient for AMP. The data for Asp 74 3 Glu and Asn 64 3 Ala all fit best to Equation 2 when n ϭ 1. The goodness of fit is 5 and 6% when n ϭ 1 for Asp 74 3 Glu and Asn 64 3 Ala, respectively, and Ͼ18% when n ϭ 2. relative to the wild-type enzyme, respectively. AMP inhibition relative to Fru-1,6-P 2 for Asp 74 3 Glu, Asn 64 3 Gln, and Asn 64 3 Ala gave a family of parallel lines (Fig. 3), indicating uncompetitive inhibition. Equation 3 with n ϭ 1 best accounts for the data, where v, V M , A, I, n, K a , and K i represent initial velocity, maximal velocity, the concentration of Fru-1,6-P 2 , the concentration of AMP, the Hill coefficient, the Michaelis constants for Fru-1,6-P 2 , and the dissociation constant for AMP from the enzyme⅐AMP complex, respectively. The goodness of fit was Ͻ6%. Thus, cooperativity for AMP inhibition is lost in addition to the change of the kinetic mechanism.

DISCUSSION
Amino acid sequences of all known mammalian FBPases are well conserved, particularly in segment 52-72, where 10 amino acids are highly conserved, suggesting an important functional or structural role. In all published crystal structures of FBPase (9 -13), however, little or no electron density is present for residues 52-72, which is indicative of conformational disorder or proteolytic damage. The latter possibility is unlikely as microsequencing detected no proteolysis (10). Reduced k cat values for the Asp 74 3 Ala, Asp 74 3 Asn, and Asp 74 3 Glu FB-Pases by 50,000-, 2,000-, and 20-fold, respectively, along with the shift in the pH of optimum activity from neutral to alkaline for the latter mutant, demonstrate the importance of position 74 in catalysis. Given the sensitive nature of position 74, conformational changes in the adjacent segment 52-72 could le- verage significant alterations in the catalytic properties of FB-Pase. Mutations Asn 64 3 Ala, Asn 64 3 Gln, and Asp 68 3 Glu, for instance, change the pH of optimum activity from neutral to alkaline and decrease k cat values from 3-to 4-fold relative to that of the wild-type enzyme, and the double mutation Lys 71,72 3 Met elevates the K i for AMP nearly 175-fold. The activity profiles of the Asp 74 3 Glu, Asn 64 3 Ala, Asn 64 3 Gln, and Asp 68 3 Glu mutants are very similar to those obtained with the proteolyzed enzyme in which residues 1-64 and 1-25 are missing (32)(33)(34).
The analysis above is strengthened by the crystal structure 2 (2.3-Å resolution, R-factor ϭ 0.165, R-free ϭ 0.24) of recombinant wild-type porcine liver FBPase with the products Fru-6-P, P i , and Zn 2ϩ , which reveals strong electron density for loop 52-72 and a significant network of hydrogen bonds involving Asp 68 , Asn 64 , and other residues of the active site. In the proposed catalytic mechanism of hydrolysis of Fru-1,6-P 2 (12,13), a water molecule, bound to a metal cation, is activated for a nucleophilic attack of the 1-phosphate. The side chains of Asp 74 and Glu 98 , relative to the attacking water and to the metal ion at the "catalytic" site (called metal site 2 in crystal structures of FBPase) (Fig. 4), are in position to act as general base catalysts. The distance between OD2 of Asp 74 and OE2 of Glu 98 and the cation at site 2 is ϳ3.3 Å, but neither side chain is in the inner coordination sphere of that cation. Replacement of Glu 98 by glutamine resulted in a 1,600-fold reduction in k cat relative to wild-type FBPase (35). The Asp 74 3 Asn mutation caused a decrease in catalytic rate of the same order of magnitude, whereas the Asp 74 3 Ala mutation resulted in almost the total loss of catalytic activity. Thus, Asp 74 and Glu 98 may act in concert to remove a proton from the attacking water molecule.
Asn 64 and Asp 68 are also in the active site (Fig. 5), with the former hydrogen bonding with Glu 97 and Glu 98 and the latter with Arg 276 . Evidently, Asn 64 maintains Glu 97 and Glu 98 in proper orientations for metal binding and the activation of the attacking water molecule. The above is consistent with the reduction in the Hill coefficient for Mg 2ϩ and the 50-fold decrease in affinity for Mg 2ϩ , exhibited by the Asn 64 3 Ala and Asn 64 3 Gln mutants relative to wild-type FBPase. Mutation of Arg 276 to methionine reduces activity by 2,000-fold, abolishes Mg 2ϩ cooperativity, and alters the kinetic mechanism of FBPase (36). Evidently, the Asp 68 -Arg 276 hydrogen bond is a determinant for many of the kinetic properties of FBPase; its loss could destabilize loop 52-72 and impair the catalytic function of Asp 74 .
AMP induces a modest structural change in the helix (H2) immediately preceding loop 52-72, and in all AMP-bound complexes of FBPase, loop 52-72 is disordered (13,28). The new crystal structure has an ordered loop in the absence of AMP. Hence, AMP could inhibit FBPase by the disruption of interactions between loop 52-72 and the active site. As Asn 64 and Asp 68 probably stabilize metal binding to the active site, the observed competition between AMP and Mg 2ϩ in kinetics and NMR investigations may stem from the displacement of loop 52-72 from the active site due to the AMP-induced perturbation of helix H2.
The 175-and 10-fold increases in K i of AMP inhibition with respect to Fru-1,6-P 2 for the Lys 71,72 3 Met and Lys 71 3 Ala enzymes, respectively, could originate from the destabilization of loop 52-72 in its disordered, AMP-induced conformation. AMP may be less effective in the displacement of loop 52-72 of the double mutant (in which positions 71 and 72 are methionine instead of lysine) because of the unfavorable thermodynamics of exposing two methionyl side chains to the solvent. Indeed, the double mutant apparently exhibits preference toward the active R-state, relative to the less active or inactive T-state. Its k cat /K m increases 3-fold, and its K a for Mg 2ϩ decreases 2-fold, whereas its K i for AMP and Fru-2,6-P 2 increases 175-and 7-fold, respectively. In fact, the double mutant may represent the kinetic properties of FBPase locked into the R-state.
Mutations that bring about a change in the kinetic mechanism of FBPase from competitive to noncompetitive (AMP inhibition relative to Mg 2ϩ ) are possible in the context of a steady-state random mechanism (16,17) and now can be understood in terms of a conformational mechanism. To a first approximation, loop 52-72 is the instrument by which AMP exerts its allosteric effect on the active site. In the wild-type system with a fully functional loop, AMP and Mg 2ϩ (presumably at site 2) are probably antagonists with respect to the conformation that they stabilize for loop 52-72. Hence, in wildtype FBPase, loop 52-72 strongly couples the AMP-and Mg 2ϩbinding sites. In some mutants of loop 52-72, however, the strong coupling of the AMP and the cation site is diminished; AMP binds and perturbs the metal at site 2, thereby causing inhibition, but is no longer completely effective in displacing the cation from a catalytically productive association with the active site.
A rational basis for an uncompetitive mechanism is more challenging, however. Four separate mutations (Asp 74 3 Glu, Asn 64 3 Ala, Asn 64 3 Gln, and Arg 49 3 Cys) change AMP inhibition with respect to Fru-1,6-P 2 from noncompetitive to uncompetitive. The uncompetitive mechanism implies that AMP binds as an inhibitor only when Fru-1,6-P 2 is productively bound at the active site. Hence, these mutations may stabilize a conformation of loop 52-72 that does not favor the binding of AMP (as an inhibitor), but permits the association of Fru-1,6-P 2 with the active site. Productively bound Fru-1,6-P 2 could then induce a conformational change in loop 52-72 that re-establishes the coupling between the AMP and the substrate-binding site. A more detailed explanation of the above phenomenon, however, must await crystallographic and ligand binding studies of mutant FBPases.