Identification of interactions that stabilize the transition state in Escherichia coli phosphofructo-1-kinase.

The activity of Escherichia coli phosphofructo-1-kinase depends upon the dissociation of a group with a pK of approximately 6.6. Mutation of the 2 active site residues most likely to be titrated in this range, Asp-127 and Asp-129, lowered activity but produced little change in the pH dependence, suggesting that these residues while important to activity were not responsible for the pH dependence of the enzyme. Alternatively, the pH dependence was thought to correspond to the negative charge on the phosphoryl group being transferred in the reaction. Mutation of Arg-72 to histidine, while lowering the activity substantially, shifted the pH optimum to approximately 6.2 with a secondary plateau on the alkaline limb that was eliminated in the double mutant R72H/R171H. Mutation of Arg-171 alone produced only a small decrease in maximal activity and little change in pH dependence from the wild type enzyme. The data support an associative mechanism with a transition state stabilized by the interaction between the negative charge on the phosphoryl group being transferred and the arginyl residue at position 72 on the enzyme.


Identification of Interactions That Stabilize the Transition State in
Escherichia coli Phosphofiucto-1-kinase" ( The activity of Escherichia coli phosphofructo-l-kinase depends upon the dissociation of a group with a pK of approximately 6.6. Mutation of the 2 active site residues most likely to be titrated in this range, Asp-127 and Asp-129, lowered activity but produced little change in the pH dependence, suggesting that these residues while important to activity were not responsible for the pH dependence of the enzyme. Alternatively, the pH dependence was thought to correspond to the negative charge on the phosphoryl group being transferred in the reaction. Mutation of Arg-72 to histidine, while lowering the activity substantially, shifted the pH optimum to approximately 6.2 with a secondary plateau on the alkaline limb that was eliminated in the double mutant R72W R171H. Mutation of kg-171 alone produced only a small decrease in maximal activity and little change in pH dependence from the wild type enzyme. The data support an associative mechanism with a transition state stabilized by the interaction between the negative charge on the phosphoryl group being transferred and the arginyl residue at position 72 on the enzyme. The ATP-dependent phosphofructo-1-kinase of Escherichia coli catalyzes the phosphorylation of Fru-6-P' to produce Fru-1,6-P,. The enzyme displays allosteric properties with cooperative binding of Fru-6-P that is influenced by the activators ADP or GDP or the inhibitor phosphoenolpyruvate. The steady state kinetics indicate a general adherence to the concerted allosteric mechanism (1) and a random mechanism that is nonequilibrating under some conditions (2, 3).
Deville-Bonne et al. (4) studied the pH dependence of the kinetic properties of the enzyme in the pH range of 6 to 9 and noted the catalytic rate constant to be controlled by the ionization of a critical group with a pK of approximately 6.6 in the presence of allosteric effectors. The critical group must be unprotonated for the enzyme to be active. The identity of this group has not been established but a number of functional group candidates can be inferred from crystallographic studies of the enzyme with products bound (5) and from site-directed mutagenesis studies (6,7). Deville-Bonne et al. (4) suggested that the pH dependence of kcat could be due to a carboxyl or histidine residue and that Asp-127 was a prime candidate for the critical residue on the basis of results of Hellinga and Evans (6) who showed that mutation of the Asp-127 residue to serine reduces activity by more than 5 orders of magnitude. Laine et al. (8) examined the pH dependence of both Asp-127 and Asp-* This work was supported by NIDDK Grant DK19912. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. P,, fructose 1,6-bisphosphate; MES, 4-morpholineethanesulfonic acid; The abbreviations used are: Fru-6-P, fructose 6-phosphate; Fru-1,B-PAGE, polyacrylamide gel electrophoresis; FPLC, fast protein liquid chromatography.
129 and concluded that the Asp-127 is not directly involved in the pH dependence of the enzyme and that a mutation of Asp-129 results in a shift of the pH response.
In the current study, we examined the pH dependence of a mutant involving both residues 127 and 129. We suggest that the pH dependente reflects the pK of a phosphoryl group of ATP that is interacting with a positively charged residue on the enzyme. Furthermore, single and double mutants involving basic residues reputed to be present in the active site of phosphofructo-1-kinase, Arg-72 and Arg-171, have been studied to identify the residue that is critical to charge stabilization in the transition state.
Site-directed Mutagenesis-The preparation of the pRZ3 plasmid, which basically consists of the E. coli gene in pUC19, has been described previously (3). The pfi gene from pRZ3 was removed by digestion with BamHI and Hind111 and ligated into M13 for site-directed mutagenesis by the method of Kunkel(9) using the Mutagene M13 mutagenesis kit from Bio-Rad. Uracil-containing single-stranded DNA was generated in the dut-ung-mutant E. coli strain CJ236 and used as the template to synthesize the mutated strand with the synthetic primer. The double mutant R72H/R171H was produced by using two mutagenic primers and wild type template uracil-containing DNA. The triple mutant D127SID129NiR252Q was produced with two mutagenic primers, one of which had substitutions at both positions 127 and 129. These oligonucleotides, as well as the others used for mutagenesis and those used for DNA sequencing, were synthesized on Biosearch DNA Synthesizer and purified by polyacrylamide gel electrophoresis and C18 Sep-Pak cartridge. Table I lists the oligonucleotides employed for site-directed mutagenesis. The obtained double-stranded DNA was transformed into wild type E. coli strain TG-1 to select against the nonmutagenized strand. DNA sequencing by Sequenase protocol was used to screen the mutants and to verify the entire sequence of one strand of the mutant DNA.
Expression and Purification of the Enzymes-Wild type E. coli phosphofructo-1-kinase and the mutant phosphofructo-1-kinases were expressed in DF1020 which is a mutant E. coli strain with p@ genes deleted. The bacteria were grown in a medium consisting of 1% tryptone, 0.5% yeast extract, 0.5% NaCl, and 5 mg/ml ampicillin. Except where indicated, the enzymes were purified to homogeneity by the method of Kotlarz  Enzyme Assay-The initial velocity of phosphofructo-1-kinase activity was determined spectrophotometrically by measuring the decrease Active Site Residues of Phosphofructo-1-kinase of optical density at 340 nm, which represents indirectly the production of Fru-1,G-P. The reaction was carried out in the presence of 0.2 mM NADH, 1 mM dithiothreitol, 1 mM EDTA, 27 pg of aldolase, 5 pg of glycerophosphate dehydrogenase, and 1.12 pg of triose-phosphate isomerase in a 0.5-ml cuvette at 30 "C. The allosteric activator GDP was added in all experiments. The buffer system used for all experiments contained 0.1 &I MES, 0.051 M N-ethylmorpholine, and 0.051 M diethanolamine. The pH was adjusted to the desired value with HCI or NaOH. For substrate-dependent assays, one of the two substrates, ATP or Fru-6-P, was kept saturated and the other substrate was varied from 1/10 K,,, to 10 K,,. Magnesium ion concentration was kept 4 mM higher than the concentration of ATP under all conditions in order to ensure that virtually all of the ATP existed as the MgATP complex. Analysis of kinetic parameters was carried out using the GraFit graphical analysis program. Fru-6-PO,, was obtained using the Hill equation and the K, for ATP using the Michaelis-Menten equation.
Determination of Circular Dichroism Spectra-Spectra were determined using a Jasco 700 spectropolarimeter with a 1-mm path at room temperature and a time constant of 1 s at 50 ndmin. Native and mutant enzymes were dialyzed against 40 mM potassium phosphate at the indicated pH and diluted to a concentration of 0.18 mg/ml.
Other Methods-The concentration of protein was determined by Bradford's dye binding assay with bovine serum albumin as the standard (12). Gel electrophoresis of proteins was carried out using 1 2 8 polyacrylamide support according to the system of Laemmli (13).

RESULTS AND DISCUSSION
Purification of Mutants-With two exceptions, the wild type and mutant phosphofructo-1-kinases were purified to homogeneity by chromatography on Blue Sepharose. Mutant R72H bound weakly and could be eluted by salt, and R72WR171H did not bind to the Blue Sepharose column. FPLC using a Mono Q anion exchange column was used to purify these two mutants as described under "Experimental Procedures." For R72H, the elution buffer was adjusted to pH 8.1; for double mutant R72W R171H, either the initial ammonium sulfate precipitate or enzyme from the initial emuent of the Blue Sepharose column was chromatographed twice, using the same buffer, once a t pH 7.1 and the second a t pH 8.1. Fig. LA shows the elution profile from Mono Q of the double mutant R72WR171H a t pH 7.1, and the single major peak that was obtained after rechromatography a t pH 8.1 (Fig. 1B ). The SDS-PAGE gel shown in Fig. 2 describes the results from each step in the purification of mutant R72WR171H. The purity and molecular mass for all mutants were verified by SDS-PAGE.
Circular Dichroism-To determine whether or not global changes in conformation were caused by the mutations at the active site of the enzyme, the circular dichroism spectra of the mutants employed in this study were determined. Differences in spectra of all mutants were very small, generally less than 5% at any given point. Because many of the mutants were studied a t varying pH, the circular dichroism spectra were determined a t pH 6.5, 7.5, and 8.5. No significant differences were seen among the wild type and mutant enzymes a t any given pH, nor were significant differences promoted by the change in pH. The results for several of the mutants and wild type enzyme are shown in Fig. 3. The data suggest that there is no global conformational change in the mutant structures relative to wild type enzyme and that pH has no effect on the overall structure of either wild type or mutant enzymes. Kinetic Properties of Active Site Mutants-Previous studies from Evans' laboratory (7,14) have shown that mutations at Asp-127 and Asp-129 produce striking reductions in the catalytic activity of phosphofructo-1-kinase. Specifically, the D127S mutant showed a reduction in kc,, of 18,000 (7), and the D129S mutant was reduced by 160-fold (14). We have re-examined the effects of mutation of these two positions and confirm the basic observations of Evans' laboratory by generating three mutants: D127S, D127A, and the triple mutant D127S/D129N/R252Q. I t should be noted that D127A was produced as a desirable side product of the mutagenesis to produce D127S. D127A resulted from a change in only one position of the Asp codon, whereas the mutagenic primer contained two base substitutions. Several kinetic parameters of these mutants are described in Table  11. The D127S mutant gave results very similar to those of Hellinga and Evans (7) who suggested that Asp-127 acts as a general base interacting with the proton on the OH of C-1 of Fru-6-P. The alanine mutant ofhp-127 has more activity than the serine mutation. Perhaps the alanine mutation permits a water molecule to occupy the space (and partially fulfill the role of general base) of the carboxyl group of Asp that occupies that In that enzyme, mutagenesis ofAsp-151, a residue that is apparently homologous with Asp-127, resulted in a profound reduction in kc,, (15). In that case also, the Ala mutant had much greater activity than the Ser mutant. A comparison of the back reaction carried out by the two Asp-127 mutants is instructive. Hellinga and Evans (7) noted that the K,,, for the sugar bisphosphate is lower in the serine mutant as opposed to wild type enzyme. These results are confirmed by the data in Table I1 which indicate an apparent affinity of D127S for Fru-1,B-P, that is about 45 times higher than that of wild type, a value similar to that reported by Hellinga and Evans (7). This increase in apparent affinity of D127S is explained by the presumption that an Asp residue at this position does not favorably interact with the phosphoryl group on carbon 1 of the sugar phosphate whereas the Ser residue can form a favorable hydrogen bond. The D127A mutant, on the other hand, lacks the ability to form this bond and hence the apparent affinity of the Ala mutant for Fru-l,6-P2 is much lower than that of the Ser mutant.
With the triple mutant, in which both Asp-127 and Asp-129 are converted to neutral residues, a profound decrease in kc,, was observed, although the activity was somewhat higher than that seen with a single mutation at Asp-127. Note that in the mutants described in Table 11, the K,,, values for fructose-6-P are nearly identical with that of wild type phosphofructo-lkinase.
If either Asp-127 or Asp-129 represents the anionic residue involved in the rate-limiting step of catalysis or in stabilization of the transition state of the wild type enzyme, then the pH dependence of the mutants should be altered. The pH dependence of the mutants involving these residues should no longer were used in each measurement in the wild type and two mutants, respectively. The solid curues describe the theoretical titration of a single group with the pK indicated on the figure. describe the dissociation constant of a group in the range of 6 to 7 but should be shifted t o that characteristic of one or more different new functional residues, unless, of course, the new critical residue(s) of the rate-limiting step has the same titration characteristics as the wild type enzyme. Fig. 4 describes the pH dependence of wild type phosphofructo-1-kinase and those of mutations at Asp-127, Asp-129, and Arg-252. Note that while maximal activity has been altered, the mutants display pH dependences that are not strikingly different from that of wild type. A shift of about 0.4 pH unit in seen in the pK of the mutants relative to that of wild-type enzyme. The data on D127S are in agreement with that of Laine et al. (8) who also described a pK of 7 for the pH dependence of kcat for this mu- tant. An examination of the active site of the enzyme as derived from x-ray crystallography (5) does not immediately suggest any other residue with a pK in the range of 6 to 7. The only other negatively charged residue near the area is Asp-103, whose mutation to Ala by Berger and Evans (14) led to a modest reduction in kcat compared to the effects of mutation on Asp-127 and Asp-129, suggesting that Asp-103 does not play a major role in transition state stabilization. On the other hand, the secondary pK of the terminal phosphoryl residue of the substrate ATP would be expected to be about 6.5. This leads t o the suggestion that the pH dependence of the enzyme could be a

Kinetic parameters of arginine mutants
Assays were performed at the indicated pH and in the presence of 1 mM GDP. For the determination of the K , for Fru-6-P, 1 mM ATP was used with wild type and R72H, and 3 mM was used with R171H and the double mutant. For the K,,, for ATP, 1 mM Fru-6-P was used with wild type and R72H and 3 m M was used with R171H and the double mutant, mutants, respectively, without substantially influencing the K, for either substrate. We reasoned that if either of these 2 basic residues was involved in transition state stabilization by neutralizing the negative charge on the terminal phosphate ofATP, then a mutation should influence the pH-dependent behavior of kcat. A mutation of the Arg residues to His was thought to provide the positive charge to neutralize the negative charge of the phosphate while providing a vastly different dependence upon pH. There are problems of geometry between His and Arg residues, although these difficulties are lessened by the fact that electrostatic interactions operate over greater distances than other types of interactions. The kinetic parameters of the R72H, R171H, and the double mutant, R72HiR171H, are given in Table 111. Mutations in either of the basic residues produced very small changes in K,,, for both ATP and Fru-6-P. Furthermore, pH had a relatively small effect on apparent substrate affinity for either substrate. The largest differences were less than 2.5-fold, as seen in the affinity of the wild type enzyme for Fru-6-P. On the other hand, mutation of Arg-72 led to a decrease in kc, of more than 100-fold, while mutation in Arg-171 led to a relatively modest decrease in activity. Note that the maximal activity of wild type enzyme is greater at the higher pH value (8.5), whereas the activity of the Arg-72 mutant is greater at pH 6.5. A more complete description of the effect of pH on the activity of the Argmutants is provided by Fig. 5, the data ofwhich should be compared with the wild type pH dependence described in Fig.  4. Mutation of Arg-72 to histidine led to a change in the pH dependence of the reaction by shifting the maximum to a lower pH. The pH maximum was shifted to approximately pH 6.2, suggesting that the activity is for the most part controlled by ionization of the terminal phosphate leading t o the ascending limb of the curve and deionization of histidine leading to the descending limb on the alkaline side of the histidine dissociation. The pH dependence of the activity of R72H plateaus at a value near one-half of the maximal activity in the region of pH 8, indicating the influence of a second positively charged group, possibly Arg-171. The double mutant in which both Arg-72 and Arg-171 have both been changed to histidine displays a n even greater decrease ink,, on the alkaline side, eliminating the plateau seen in the single mutant, presumably as a result of the elimination of the secondary influence of Arg-171. The kinetic data obtained with the above described mutants confirm the important role of Asp-127 in the mechanism of phosphofi-ucto-1-kinase. However, the rate-controlling step appears to involve the chemical step transferring the phosphoryl group to the acceptor sugar phosphate. In nonenzymatic reactions involving phosphate transfer, a distinction is made between associative and dissociative mechanisms, but the distinctions are somewhat blurred in enzymatic reactions where donating and accepting molecules are both in place at the time of the chemical event (17). The transfer of the metaphosphate ion generated in the dissociative mechanism would be impeded by a nearby cationic group. An associative mechanism would receive assistance from a positive charge (18). The results with mutant R72H indicate the importance of charge neutralization in the transition state and hence favor the associative mechanism. The reaction rate is highest when the negative charge on transferring phosphate is neutralized by the positive charge at position 72. In wild type enzyme, position 72 is occupied by Arg which will be positively charged below pH 12, and the negative charge on the phosphoryl group will be titrated at about pH 6.5. In the R72H mutant, the histidine will lose a positive charge in the same pH range that the phosphoryl group will gain a negative charge, leading to a pH optimum between the pK values of the two groups. This generates the pH-dependent curve shown in Fig. 4 which confirms the importance of the interaction between the negative charge on the terminal phosphoryl group of ATP and the positive charge on the arginine residue at position 72 of phosphofructo-1-kinase.