Biochemical properties of mutant and wild-type fructose-1,6-bisphosphatases are consistent with the coupling of intra- and intersubunit conformational changes in the T- and R-state transition.

The significance of interactions between AMP domains in recombinant porcine fructose-1,6-bisphosphatase (FBPase) is explored by site-directed mutagenesis and kinetic characterization of homogeneous preparations of mutant enzymes. Mutations of Lys42, Ile190, and Gly191 do not perturb the circular dichroism spectra, but have significant effects on ligand binding and mechanisms of cooperativity. The Km for fructose 1,6-bisphosphate and the Ki for the competitive inhibitor, fructose 2,6-bisphosphate, decreased by as much as 4- and 8-fold, respectively, in the Q32L, K42E, K42T, I190T, and G191A mutants relative to the wild-type enzyme. Q32L, unlike the other four mutants, exhibited a 1.7-fold increase in Kcat. Mg2+ binding is sigmoidal for the five mutants as well as for the wild-type enzyme, but the Mg2+ affinities were decreased (3-22-fold) in mutant FBPases. With the exception of Q32L (8-fold increase), the 50% inhibiting concentrations of AMP for K42E, K42T, I190T, and G191A were increased over 2,000-fold (>10 mM) relative to the wild-type enzyme. Most importantly, a loss of AMP cooperativity was found with K42E, K42T, I190T, and G191A. In addition, the mechanism of AMP inhibition with respect to Mg2+ was changed from competitive to noncompetitive for K42T, I190T, and G191A FBPases. Structural modeling and kinetic studies suggest that Lys42, Ile190, and Gly191 are located at the pivot point of intersubunit conformational changes that energetically couple the Mg2+-binding site to the AMP domain of FBPase.

FBPase is a tetrameric enzyme consisting of identical subunits of molecular weight 37,000 (6). The activity of FBPase, as a function of Mg 2ϩ concentration, is sigmoidal at neutral pH with a Hill coefficient of approximately 2 (7). X-ray diffraction studies showed that there are two divalent metal-binding sites per monomer; however, in the case of Mg 2ϩ , only one metal ion binds per subunit (8). AMP inhibition of FBPase is cooperative with a Hill coefficient of 2 (9 -11). The structure of porcine kidney FBPase, however, reveals one AMP-binding site per subunit (12). The structural basis for the observed cooperativity in AMP and Mg 2ϩ interactions is still unclear. AMP and Mg 2ϩ are mutually exclusive in their binding to FBPase (13)(14)(15)(16); AMP inhibition with respective to Mg 2ϩ is competitive at neutral and alkaline pH (13,17). Crystallographic studies have shown that AMP interactions perturb metal-binding sites (18,19).
FBPase can exist in at least two forms: (i) the R-state, in which the enzyme is complexed with substrate or product with or without metal ions and (ii) the T-state, in which FBPase is complexed with AMP in the presence or absence of metal ions (12,20). The structural transition (R-to T-state) involves a 17°r otation of the C1-C2 dimer with respect to the C3-C4 dimer and a 1.9°rotation of the AMP domain (residues 1-200) relative to the FBP domain (residues 200 -335) (18). The N-terminal residues, helices H1, H2, and H3, and ␤-turn T4, which comprise part of the C1-C4 and symmetry-related C2-C3 interfaces, putatively are involved in the propagation of structural changes throughout the tetramer during the R-to T-state transition (18). Mutation of residues in helices H1 and H3 changed the affinity of AMP or Mg 2ϩ (21,22). Here, we present the effects of mutation of Gln 32 and Lys 42 (helix H2) and Ile 190 and Gly 191 (␤-turn T4). The mutations cause a dramatic alteration in Mg 2ϩ and AMP affinities. In addition, for some of the mutants, the mechanism of AMP inhibition relative to Mg 2ϩ changes from competitive to noncompetitive.
Mutations were confirmed by NruI/XhoI digestion and by fluorescent dideoxy chain-termination sequencing at the Nucleic Acid Facility at Iowa State University. The sequencing primer, 5Ј-GTCATTGGAGAG-GACATC-3Ј, was used to confirm the mutation of Q32L, K42E, K42T, and 5Ј-TCTGAGAAGGACGCACTG-3Ј was used to confirm the mutation of I190T, and G191A. The mutagenesis plasmid was finally transformed into Escherichia coli DF 657, a strain deficient in the FBPase gene.
Protein concentration was assayed as described by Bradford (26) with bovine serum albumin (from Sigma) as the standard. The protein purity was determined by using 12% SDS-polyacrylamide gel electrophoresis according to Laemmli (27).
Kinetic Studies-Specific activity during purification was determined by the phosphoglucoisomerase and glucose-6-phosphate dehydrogenase coupled spectrophotometric assay at pH 7.5 and 9.6 (13, 28). All other kinetic experiments were done at pH 7.5 (Hepes buffer) and 24°C by using a coupled spectrofluorometric assay (13). Initial-rate data were analyzed by using a computer program written in MINITAB with an ␣ value of 2.0 (13,29). Cooperativity was evaluated by using either the ENZFITTER program (30) or the MINITAB program.
Circular Dichroism Spectrometry-CD studies on the wild-type and mutant forms of FBPase were carried out in a Jasco J710 CD spectrometer in a 1-mm cell at room temperature. Spectra were collected from 200 to 260 nm in 1.3-nm increments, and each spectrum was blankcorrected and smoothed by using the software package provided with the instrument.

Purification of the Wild-type and Mutant Forms of FBPase-
Purification of the mutants was similar to that of the wild-type enzyme, except that K42E and K42T were eluted from the CM-Sephadex column by 50 mM malonate buffer (pH 6.0). The enzyme purity of the wild-type and the five mutant forms of porcine liver FBPase was evaluated by SDS-polyacrylamide gel electrophoresis (data not shown). The six enzymes exhibit identical mobilities and are greater than 96% homogeneous with use of electrophoresis as a criterion. The pH 7.5/pH 9.5 activity ratios shown in Table I, along with the SDS-polyacrylamide gel electrophoresis results, suggest that the enzymes had not un-dergone proteolysis (28).
Secondary Structure Analysis-CD spectrometry was used to analyze the secondary structure of the mutant FBPases and the wild-type enzyme. The CD spectra of the six enzymes are superimposable from 200 to 260 nm (data not shown). These results indicated that no major conformational changes occurred in the mutant FBPases with use of CD as a criterion of secondary structure of the proteins.
Kinetic Properties of FBPase Mutants-To evaluate the effects of mutations on residues at the C1-C4 and C2-C3 interface of FBPase, initial-rate kinetic studies were undertaken with the wild-type and mutant forms of the enzyme at a concentration of Fru-1,6-P 2 or Mg 2ϩ that does not cause substrate inhibition. The kinetic parameters are summarized in Table I. A 1.7-fold increase in K cat compared with that of wild-type enzyme was found for Q32L, suggesting that the catalytic efficiency is slightly enhanced in this mutant. Small decreases (2-fold) in Fru-1,6-P 2 K m were found for Q32L, K42T, K42E, and I190T, and a 4-fold decrease was noted for G191A relative to wild-type FBPase.
Mg 2ϩ Activation-The Hill coefficient for Mg 2ϩ and the K a for Mg 2ϩ of the wild-type and mutant forms of FBPase were determined at saturating Fru-1,6-P 2 concentrations (10 -15 M). 3-, 9-, 15-, 17-, and 22-fold increases in K a values for Mg 2ϩ were found for Q32L, K42E, K42T, I190T, and G191A, respectively, relative to the wild-type enzyme ( Table I). The five mutants as well as the wild-type enzyme showed sigmoidal kinetics for Mg 2ϩ . Different K a values for Mg 2ϩ relative to that of the wild-type enzyme were observed for other mutants of the subunit interface of FBPase (21,22).
Fru-2,6-P 2 Inhibition- Table I shows that K i for Fru-2,6-P 2 was decreased 3-fold for K42E and K42T and 8-fold for I190T and G191A relative to the wild-type enzyme. Fru-2,6-P 2 competes with Fru-1,6-P 2 for the active site of wild-type FBPase (31,32) and the six mutants of the present study. The mutations at interface residues located in the AMP domain resulted in changes in Fru-1,6-P 2 and Fru-2,6-P 2 affinity, suggesting that these residues may be involved in communication between the Fru-1,6-P 2 and AMP domains.
Kinetics of AMP Inhibition-AMP is an allosteric inhibitor of FBPase, exhibiting sigmoidal binding with a Hill coefficient of 2.0 (9 -11). The pattern of AMP inhibition is nonlinear, non- competitive relative to Fru-1,6-P 2 , but nonlinear, competitive relative to Mg 2ϩ for wild-type FBPase at either neutral or alkaline pH (13). The expected competitive inhibition pattern of AMP relative to Mg 2ϩ was found in wild-type, Q32L and K42E FBPases; however, the kinetics were changed to noncompetitive inhibition in the case of mutants K42T, I190T, and G191A. Figs. 1 and 2 show double-reciprocal plots of 1/velocity versus 1/(Mg 2ϩ ) 2 at various fixed concentrations of AMP for the Q32L and K42E mutants of FBPase, respectively. The data in Fig. 1 gave excellent fits to Equation 1, which is consistent with a steady-state random mechanism, when n ϭ 2. The Goodness of Fit was 6% when n ϭ 2 and 16% when n ϭ 1. The data in Fig.  2, however, fit better to Equation 1 when n ϭ 1. The Goodness of Fit was 5% when n ϭ 1 and 17% when n ϭ 2. The form of Equation 1 is as follows, 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 concentration of Mg 2ϩ, Fru-1,6-P 2 , AMP, the Michaelis constants for Mg 2ϩ and Fru-1,6-P 2 , the dissociation constants for Mg 2ϩ and the dissociation constants for AMP from the enzyme-AMP, the enzyme-AMP-AMP, the enzyme-Fru-1,6-P 2 -AMP, and the enzyme-Fru-1,6-P 2 -AMP-AMP complexes, respectively. n represents the Hill coefficient for AMP with FBPase. When n ϭ 2, the binding of AMP to FBPase exhibits cooperativity, on the other hand, there is no cooperativity when n ϭ 1. In other words, mutation of Gln 32 to Leu results in no alteration in AMP cooperativity for FBPase; however, the K42E mutation causes loss of AMP cooperativity. Fig. 3 illustrates double-reciprocal plots of 1/velocity versus 1/(Mg 2ϩ ) 2 at various fixed concentrations of AMP for the K42T mutants. Unexpectedly, a family of lines was obtained intersecting to the left of 1/v axis. A similar result was found with the I190T and G191A mutants (data not shown). These findings indicate that the inhibitor, AMP, exhibited noncompetitive inhibition relative to Mg 2ϩ with these mutants. The data fit best to Equation 2, the form of which is as follows, where v, V m , A, I, K a , K i , and K ii represent initial velocity, maximal velocity, the concentration of Mg 2ϩ , the concentration of AMP, the Michaelis constants for Mg 2ϩ , the dissociation constants for AMP from the enzyme-AMP and the enzyme-Mg 2ϩ -AMP complexes, respectively. n represents the Hill coefficient for AMP with FBPase. When n ϭ 2, the binding of AMP to FBPase exhibits cooperativity, on the other hand, there is no cooperativity when n ϭ 1. The data for K42T, I190T, and G191A all fit best to Equation 2 when n ϭ 1. The Goodness of Fit is 4%, 5%, and 5% when n ϭ 1 in K42T, I190T, and G191A, respectively. Goodness of Fit values are all more than 12% in all instances when n ϭ 2. The proposed scheme from which Equation 2 was derived is as follows.
All five mutants investigated, as well as wild-type FBPase, exhibited noncompetitive inhibition by AMP relative to Fru-1,6-P 2 (data not shown). The K i value (I 0.5 ) for AMP for Q32L was increased 8-fold; however, mutations at Lys 42 , Ile 190 , and Gly 191 caused dramatic effects on AMP affinity (2,000 -4,000fold decreases). These findings indicate that Lys 42 , Ile 190 , and Gly 191 play important roles in AMP affinity, cooperativity, and the mechanism of AMP binding.
Temperature Sensitivity of Wild-type and Mutant Forms of FBPases-To evaluate the influence of mutations on the C1-C4 (C2-C3) subunit interface, the thermosensitivity of wild-type and mutant FBPases were studied. The wild-type enzyme re-tains its activity after incubation at 57°for 10 min (Fig. 4); however, the relative activity was decreased 70% at 67°after a 10-min incubation. At 47°, K42E and K42T exhibited a decreased activity of approximately of 50%, and I190T and G191A activities were decreased 73 and 84%, respectively. The four mutants were all inactivated at 57°. Compared with wild-type enzyme, K42E, K42T, I190T, and G191A exhibited increased thermosensitivity. Subtle changes in the enzyme structure 17°i n the C1-C4/C2-C3 interface may cause the observed thermosensitivity in the FBPase mutants. The CD spectra of mutants and wild-type FBPases are superimposable, ruling out significant changes in secondary structure. The Q32L mutation shows a slight change in activity at 57°but retains 70% of its activity when incubated at 67°, which is about 40% higher than that of wild-type enzyme. The elevated thermostability of Q32L may be a consequence of enhanced hydrophobic interactions in the mutant. DISCUSSION The allosteric transformation of FBPase is initiated putatively by structural perturbations in the AMP domain upon the binding of AMP (18). Conformational changes in the AMP domain lead to a 1.9°rotation of this domain relative to the FBP domain and to a 17°twist about a 2-fold axis of the tetramer (Fig. 5). The 1.9°rotation of the AMP domain relative to the FBP domain affects the Mg 2ϩ -binding site and may be the basis for allosteric regulation of activity. The model above has pedagogic value, but does not provide a framework sufficiently precise for the understanding of biochemical properties of mutant FBPases reported here and elsewhere (21,22). Sev-eral issues are difficult to address within the framework of the simplified description of the allosteric transition. First, can the T-state of FBPase exist in equilibrium with the R-state and in the absence of AMP? Second, how is it possible for AMP to bind near the C1-C4/C2-C3 interface and drive the change in conformational relationships of the AMP and FBP domains? AMP could more effectively leverage a conformational change by binding directly to the interface between the AMP and FBP domains. Finally, is the 1.9°rotation, mentioned above, coupled to the 17°rotation? Allosteric signals may pass directly from the AMP to the FBP domain within a subunit (18), or conversely, allosteric signals might also pass indirectly from the AMP to the FBP domain, by way of conformational changes in the C1-C4/C2-C3 interface.
The more detailed description of the allosteric transition, given below, is motivated by observations reported by Zhang et al. (18). The axis of 17°rotation of the AMP domains does not coincide with a molecular 2-fold axis of the tetramer (18). Instead, that rotation axis is parallel to the molecular twofold axis, but passes through approximately NZ of Lys 42 (18). Thus, the intersubunit conformational change is due to 17°rotations about two axes displaced from each other by approximately 20 Å (Fig. 6). The 17°rotation about axes so displaced from each other will produce domain reorientations not only in the C1-C4/C2-C3 interface but also in the C1-C2 and C3-C4 interfaces. Reorientations are observed in the C1-C2 and C3-C4 interfaces between AMP domains (18). The interactions between the FBP , and G191A (ϩ) were incubated 10 min at 30, 37, 47, 57, and 67°C, respectively. The enzyme activity was assayed at 25°C by spectrophotometry (as described under "Experimental Procedures") immediately after the incubation, which is expressed as a percentage of the relative activity at 30°C. The protein concentration was 0.12 mg in each assay. Each reaction was either in duplicate or triplicate.
domains of the C1-C2 and C3-C4 interfaces, however, are unchanged. As demonstrated in Fig. 6, the combination of reoriented AMP domains and static FBP domains in the C1-C2 and C3-C4 interfaces requires the reorientation of AMP domains relative to FBP domains within a subunit. The direction of rotation of the AMP domain relative to the FBP domain reproduces the observed 1.9°rotation. Thus the small reorientation of the AMP domain relative to the FBP domain within a sub-unit is coupled to the 17°rotations about offset axes. Rather than directly effecting the relative orientation of the AMP and FBP domains, AMP may exert its influence indirectly by acting on the C1-C4/C2-C3 interface.
The strong coupling of intra-and intersubunit transitions, proposed above, necessarily requires that FBPase be a twostate system. Conceivably, a dynamic equilibrium exists between the R-and T-states with AMP influencing the equilibrium in favor of the T-state. Such a hypothesis is consistent with the observed binding of AMP near the C1-C4/C2-C3 interface, where it can directly stabilize the T-state conformation of that interface. Furthermore, given the pairwise juxtaposition of AMP-binding sites across the C1-C4/C2-C3 interface, one can intuitively understand the observed Hill coefficient of 2 for AMP cooperativity.
Within the framework of the above model mutations will fall within three categories: (i) those which influence the R-and T-state equilibrium, (ii) those which affect the putative coupling of intra-and intersubunit conformational changes, and (iii) those which directly affect the relationship of the AMP and FBP domains within a subunit. Mutations presented here fall into the first two categories.
Mutations of Lys 42 , Ile 190 , and Gly 191 lie near the 17°rotation axes. Here, Lys 42 and a symmetry related mate participate in a complex network of hydrogen bonds, which is conserved in both the T-and R-states (Fig. 7). NZ of Lys 42 interacts with backbone carbonyls 190 and 191 and the carboxyl group of Glu 192 . Glu 192 makes an additional hydrogen bond with symmetry-related Thr 39 . All of the mutants near the rotation axes disturb intersubunit hydrogen bonds. The addition of two methyl groups between loops T4 in G191A causes steric repulsion and presumably the weakening of the hydrogen bond network around Lys 42 . I190T introduces an oxygen atom, which can make a hydrogen bond with either the carboxylate of Asp 187 or Glu 192 . This interaction may influence the hydrogen bonding networks near the rotation axes. The mutation of Lys 42 to Thr disrupts hydrogen bonds, whereas the Glu 42 mutant introduces an opposite electrostatic charge. All mutants of this area change the AMP and Mg 2ϩ affinities by similar amounts. All mutants lack AMP cooperativity and for three of the four mutants AMP inhibition with respect to Mg 2ϩ is noncompetitive. AMP inhibition with respect to Mg 2ϩ is competitive for wild-type FBPase. These significant effects caused by mutations near the 17°rotation axes and approximately 25 Å away from both the AMP and Mg 2ϩ -binding sites are consistent with the putative coupling of inter-and intrasubunit confor- mational changes.
Gln 32 in the C1-C4/C2-C3 interface is far from the 17°rotation axes (peripheral interface, Fig. 8). Mutations of the peripheral interface as a group affect either AMP or Mg 2ϩ affinity. The kinetic parameters (Table I) of the Q32L mutant are consistent with other mutants (22) at the peripheral interface. The Q32L mutant is to some extent an antipode of the M18I mutant described earlier (22). Indeed, the substitution of Met 18 with Ile enhances hydrophobic interactions in the T-state (Fig. 8). The reduced Mg 2ϩ affinity exhibited by the M18I mutant could result from a shift in equilibrium toward the T-state. Gln 32 participates in hydrogen bonds between helix H1 and helix H2 of a symmetry related subunit when the enzyme exists in the T-state, but is not involved in intersubunit hydrogen bonds in the R-state (Fig. 8). In addition, its environment in the T-state seems to be more hydrophilic. Substitution of hydrophilic glutamine with hydrophobic leucine should destabilize the T-state relative to the R-state. As a result, changes in kinetic param-eters are similar to mutants with a destabilized T-state (Table  I). All mutants of the peripheral areas of the C1-C4/C2-C3 interface may influence the equilibrium between R-and Tstates and thereby alter the observed kinetic properties of the system.  (8,19) (Protein Data Bank entries 1fbc and 1fpe) by using MOLSCRIPT (33).