The proton transfer step catalyzed by yeast pyruvate kinase.

The nature of the proton donor to the C-3 of the enolate of pyruvate, the intermediate in the reaction catalyzed by yeast pyruvate kinase, was investigated by site-directed mutagenesis and physical and kinetic analyses. Thr-298 is correctly located to function as the proton donor. T298S and T298A were constructed and purified. Both mutants are catalytically active with a decrease in k(cat) and k(cat)/K(m)(,PEP). Mn(2+)-activated T298S and T298A do not exhibit homotropic kinetic cooperativity with phosphoenolpyruvate (PEP) in the absence of fructose 1,6-bisphosphate, although PEP binding to enzyme-Mn(2+) is cooperative. The pH dependence of k(cat) for T298A indicates the loss of pK(a)(,2) = 6.4-6.9. Thr-298 affects the ionization (pK(a) approximately 6.5) responsible for modulation of k(cat). Fluorescence studies show altered dissociation constants of ligands to each enzyme complex upon Thr-298 mutations. The rates of the phosphoryl transfer and proton transfer steps in the pyruvate kinase-catalyzed reaction are altered; pyruvate enolization is affected to a greater extent. Proton inventory studies demonstrate solvent isotope effects on k(cat) and k(cat)/K(m)(,PEP). Fractionation factors are metal-dependent and significantly <1. The data suggest that a water molecule in a water channel is the direct proton donor to enolpyruvate and that Thr-298 affects a late step in catalysis.

Yeast pyruvate kinase (YPK) 1 (EC 2.7.1.4.0) is a key regulatory enzyme in glycolysis that catalyzes the phosphoryl transfer from phosphoenolpyruvate (PEP) to ADP to yield pyruvate and ATP. The reaction requires both monovalent and divalent cations, normally K ϩ and Mg 2ϩ or Mn 2ϩ . Fructose 1,6-bisphosphate (Fru-6-P 2 ) is the heterotropic activator of YPK. The net reaction catalyzed by YPK is the sum of at least two partial reactions. Phosphoryl transfer from PEP to M(II)ADP occurs by an apparent S n 2 mechanism with inversion of configuration at the phosphoryl group to yield the enolate of pyruvate and M(II)ATP (1). In the second partial reaction, a proton donor at the active site stereospecifically protonates the C-3 of enolpyruvate at the 2-si face of the double bond to form ketopyruvate (2)(3)(4). The enolate of pyruvate, the common species in both partial reactions, is a tightly bound intermediate in the net reaction catalyzed by PK (5). The two-step character of the net PK reaction is demonstrated by the ability of PK to catalyze the enolization of bound pyruvate without phosphoryl transfer (6 -8). Muscle PK requires a group such as inorganic phosphate, methyl phosphonate, or fluorophosphate as a cofactor (6,7). YPK requires ATP as a cofactor for this activity (8). PK will also catalyze the ketonization of enolpyruvate, generated in situ subsequent to the hydrolysis of PEP catalyzed by alkaline phosphatase (3).
Initial crystallographic data from cat muscle PK (9) indicated that Lys-269 (Lys-240 in the YPK numbering sequence) can serve as the putative proton donor because it was positioned close to the methyl carbon of the bound pyruvate. The ⑀-amino group of Lys-269 as the possible candidate for the acid/base catalyst in PK was supported by the measurement of multiple protons that are incorporated into pyruvate with rabbit muscle pyruvate kinase (10). Subsequent refined x-ray crystal structures of the yeast enzyme with bound phosphoglycolate (11) and of the rabbit muscle enzyme complexed with pyruvate (12) demonstrated that Lys-240 is at the 2-re face of the double bond of the enolate and is in apparent contact with the phosphoryl group of PEP. The recent studies of the K240M mutant of YPK support the role of this lysine in facilitating phosphoryl transfer but not enolpyruvate protonation (13). The location of the methyl group of pyruvate is oriented such that Thr-298 (Thr-327 in the muscle PK numbering sequence) is in the correct position to protonate the enolate intermediate at C-3 (12). Thr-298 of YPK relative to bound phosphoglycolate is in the same location (11). Rose et al. (14) have suggested that the proton donor in pyruvate kinase is a high pK a monoprotic acid that rapidly exchanges protons with solvent in the unliganded enzyme. The secondary alcohol of Thr-298 satisfies these requirements and that of the stereochemistry. On the other hand, pH rate profiles with rabbit muscle PK reveal an ionization with a pK a ϭ 8.3. This ionization has been interpreted as the pK a of the acid-base catalyst (15). The strict conservation of both Lys-240 and Thr-298 in all PKs that have been sequenced and their critical location within the active site of PK suggest that the side chains of both residues play important roles in catalysis. Studies of YPK by Bollenbach et al. (13) demonstrate that the pK a of 8.8 is lost on mutation of Lys-240 to methionine. Furthermore, the results of the kinetic characterization of K240M suggest that Lys-240 is not the direct proton donor to the enolpyruvate intermediate in YPK. In its protonated form, Lys-240 is important in stabilization of the pentavalent transition state of the phosphoryl group undergoing transfer.
The present study focuses on the role of Thr-298 in catalysis of the pyruvate kinase reaction. This question was addressed by altering Thr-298 by site-directed mutagenesis and by physical and kinetic analyses of the wild type and the resulting mutant enzymes. The results indicate that it is water at the active site that serves as the proton donor.
Site-directed Mutagenesis and Cell Growth-Mutagenesis reactions were performed in the pSelect TM -1 vector according to the Promega Altered Sites II manual. The mutagenic oligonucleotides used for constructing the specific point mutants are as follows: T298S, CCAAC-ATTTGGGAAGCACAGATAACTGG; T298A, CCAACATTTGGGCAGC-ACAGATAACTGG; and T298V, CCAACATTTGGACAGCACAGATAA-CTGG. The underlined sequences correspond to the mutated nucleotides. A 1402-bp XbaI/EcoRI fragment of the YPK gene was subcloned into the XbaI/EcoRI sites of the pSelect TM -1 vector (18). For each of the Thr-298 mutants, the 751-bp fragment of DNA containing the desired mutation in the YPK gene was excised from the pSelect vector by BglII/BstEII restriction digest, followed by isolation on a 1% agarose gel. The 751-bp fragments were cloned into the BglII/BstEII sites of the yeast shuttle vector, pPYK101, that contains the entire PK gene. In order to verify the presence of the desired mutations, the mutated genes were sequenced (DNA Sequence Facility, Iowa State University, Ames, IA). The mutant pPYK101 constructs were transformed into the pyruvate kinase-deficient yeast strain, pyk1-5 (19), using a lithium acetate procedure (20).
Yeast pyk1-5 containing the wild type pPYK101 was grown on rich media containing the following per liter: 10 g of yeast extract, 20 of g bactopeptone, and 2% glucose. Transformed pyk1-5 containing pPYK101 with the T298S and T298A mutations were grown on glucose minimal media containing the following per liter: 6.7 g of yeast nitrogen base, 10 of g succinic acid, 4 of g NaOH, 2% glucose, 0.5% casamino acids, 0.001% adenine, and 0.001% methionine. Cells were harvested, and the wild type YPK and T298S and T298A mutants were purified by the procedure published previously (16). Yeast pyk1-5 containing the T298V PYK101 was grown on glycerol-ethanol minimal media prepared the same as for T298S and T298A except containing 2% glycerol and 2% ethanol in place of 2% glucose.
Circular Dichroism Spectroscopy-CD spectra of apo wild type, T298S, and T298A YPK were acquired on an Aviv 62 DS circular dichroism spectrometer at 25 Ϯ 1°C. Protein solutions at concentrations of 0.04 mg/ml in potassium phosphate buffer (4 mM (pH 6.2)) were measured in a quartz cuvette with a path length of 0.4 cm. Each spectrum was recorded with a bandwidth of 0.5 nm and a scan rate of 1 nm/s. The scans were acquired from 280 to 200 nm. For each sample, 5 repetitive scans were obtained and averaged. A scan of the buffer is subtracted from each of these spectra.
Pyruvate Kinase Assay-Pyruvate kinase was assayed by a variation of the method of Teitz and Ochoa (21). Typical assay mixtures contained in 1 ml: 100 mM MES, (pH 6.2), 4% glycerol, 200 mM KCl, either 15 mM MgCl 2 or 4 mM MnCl 2 , 5 mM ADP, 5 mM PEP, 175 M NADH, 20 g of L-lactate dehydrogenase, and YPK (1-2 g of wild type, 2-5 g of T298S, and 10 g of T298A, respectively). When present, the concentration of Fru-6-P 2 was 1 mM. The decrease in absorbance at 340 nm due to NADH oxidation was measured as a function of time using a Gilford 240 or 250 spectrophotometer. The specific activity of YPK is expressed as mol of NADH oxidized/ml/min/mg protein. The concentration of YPK was determined by absorbance at 280 nm and by using the extinction coefficient ⑀ 280 ϭ 0.51 (mg/ml) Ϫ1 cm Ϫ1 .
Initial velocity data were fit to either the Michaelis-Menten equation (Equation 1) or the Hill equation (Equation 2), depending on which model best describes the experimental data.
pH Studies-The effect of pH on the maximal velocity of the reaction catalyzed by wild type YPK, T298S, and T298A, respectively, was studied in the pH range 4.5-9.1. The measurements were performed in the presence of 1 mM Fru-6-P 2 and with Mg 2ϩ (15 mM) or Mn 2ϩ (4 mM) as the divalent activator. The buffers used were acetate (pH 4.5-5.2), MES (pH 5.0 -6.8), HEPES (pH 6.5-7.5), and TAPS (pH 7.5-9.1). The buffers were all titrated to the desired pH by using either potassium hydroxide or hydrochloric acid. The pH of the kinetic assay mixtures, containing all of the assay components except the coupling enzyme (L-lactate dehydrogenase) and pyruvate kinase, was determined before the initiation of the reactions. The pH rate data were fit to either Equation 3 or Equation 4, which were previously derived (13), 2 or to Equation 5. Equation 3 describes the ionization of two groups in the enzyme-substrate complex (ES) that are required for catalytic activity (K A and K C ), and the ionization of a group in the ES complex (K B ) that modifies the rate of reaction. Equation 4 describes two ionizations, K A and K B , in the ES complex. Equation 5 corresponds to the situation where only the two groups K A and K C are required for catalytic activity. V max,app is the observed maximal rate of reaction at a given pH value. The maximal velocity in the plateau region, V maxЈ , can be related to the maximal velocity, V max , by a proportionality factor ␣, where V maxЈ ϭ ␣V max . The value of ␣ can be either greater than or less than 1, depending on whether the protons inhibit or activate the reaction rate, respectively.
Fluorescence Measurements-The dissociation constants of the ligands to various enzyme complexes were measured by steady-state fluorescence. Fluorescence quenching of the intrinsic single tryptophan residue of YPK upon ligand binding was monitored on an SLM model 8100 fluorimeter thermostated at 24 Ϯ 1°C. The excitation monochromator was a model MC400, and the emission monochromator was a model MC200. Fluorescence titrations were performed by monitoring the change in fluorescence intensity at 334 nm upon excitation at 295 nm. Titrations were performed by sequentially adding 1-10-l aliquots of a concentrated ligand solution to 900 l of a mixture containing 100 mM MES (pH 6.2), 5% glycerol, 200 mM KCl, 0.05-0.07 mg/ml wild type YPK or Thr-298 mutants, and other ligands as specified. The percent fluorescence quenching, Q, was calculated using the relationship: Q ϭ ((F 0 Ϫ F)/F 0 ) ϫ 100, where F and F 0 are the fluorescence intensities of YPK in the presence and absence of a ligand, respectively. F was corrected for dilution effects. Fluorescence titration data were fit to Equations 6 and 7, where [L] represents the variable ligand concentration; K D is the dissociation constant for ligand L to the respective enzyme complex; Q max is the percent maximal quenching; and n H is the Hill coefficient. Equations 6 and 7 are analogous to the Michaelis-Menten and the Hill equations, respectively.
Rate of Phosphoryl Transfer-The rate of phosphoryl transfer catalyzed by YPK was measured using the glycolate kinase activity, as described by Bollenbach et al. (13). The reaction mixture contained in 1 ml, 100 mM HEPES (pH 7.5), 4% glycerol, 200 mM KCl, 2 mM Fru-6-P 2 , 50 mM glycolate, 10 mM MnCl 2 or 15 mM MgCl 2 , 0.2-8 mM ATP, and YPK (100 -200 g of wild type and T298S, or 200 -300 g of T298A). The glycolate kinase reaction was followed as the decrease of ATP peaks over time using high pressure liquid chromatography separation. High pressure liquid chromatography analysis was performed on a Beckman 421 liquid chromatograph equipped with a Beckman 334 Gradient System and a Rainin 218TP54-C18 column (5 m, 4.6 ϫ 250 mm) with a pore size of 300 Å. Detection was at 260 nm. Integration and plotting of chromatograms were performed on a Spectra Physics SP4290 integrator. Initial velocities of phosphoryl transfer were calculated from measurements of ATP disappearance with time. The k cat for phosphoryl transfer catalyzed by YPK (v p ) was determined from initial velocities versus [ATP] and fitting the data to Equation 1. A nonenzymatic blank was run at each ATP concentration. There was no detectable decrease in ATP concentration over the course of the experiment in the absence of enzyme.
Rate of Pyruvate Enolization-The enolization of pyruvate was measured as the time-dependent exchange of tritium from 3-[ 3 H]pyruvate into water, as described previously (13). The activity of 3-[ 3 H]pyruvate was 75,170 dpm/mol pyruvate in the experiments with wild type YPK and T298S or 72,900 dpm/mol pyruvate in the experiments with T298A. The PK-catalyzed rate of pyruvate enolization, v T , is defined as mol of protons exchanged into water/min/mg protein.
Solvent Isotope Effect Studies-All buffers, divalent and monovalent cations, and substrates were exchanged in 99.9% D 2 O by dissolving solutes 3 times in D 2 O followed by lyophilization. In each case the final sample was re-dissolved in D 2 O to give the desired solute concentrations. The pD and pH values of the assay buffer were adjusted to 6.2 using KOD and KOH, respectively. The assay buffer was adjusted to pD 6.2 according to the relationship pD ϭ (pH) meter reading ϩ 0.4 (22). NADH was dissolved in the assay buffer, prepared as above, immediately prior to use. Assays (1 ml) were prepared and covered with parafilm prior to measurements. Assays were performed in duplicate. Solvent isotope effects (SIE) on k cat , D (k cat ), and on k cat /K m,PEP , D (k cat /K m,PEP ), were determined from fitting initial velocity versus [PEP]  represents the situation of a single proton in the transition state and one in the reactant state, respectively. Alternatively, one proton may have different contributions in the reactant state than in the transition state.
In these equations, T and R are the fractionation factors for the proton in the transition state and reactant state, respectively, and V n and V 0 represent the maximal velocities observed in a mixture of isotopic water of n mole fraction of D 2 O and in H 2 O, where n ϭ 0, respectively (23,24).

Cell Growth and Purification of T298S and T298A
T298S, T298A, and T298V were constructed and expressed using the same procedure as for wild type YPK. The pyruvate kinase-deficient Saccharomyces cerevisiae strain, pyk1-5, containing the pPYK101 plasmid with either the T298S or the T298A mutation was grown on media containing glucose as the sole carbon source. In both cases, 10 liters of glucose minimal media yielded ϳ100 g of cells. T298S and T298A were purified according to Mesecar and Nowak (16), with no modification from the protocol for wild type YPK. Yeast pyk1-5 containing the T298V pPYK101 was unable to grow on media containing glucose as the sole carbon source (glucose minimal media). These cells were successfully grown on glycerol-ethanol media. The inability of pyk1-5 containing the T298V pPYK101 to grow on media containing glucose is indicative of a catalytically inactive mutant or a mutant enzyme of very low activity (Ͻ5% relative to wild type YPK, the value for T298A). In the absence of an active PK gene, yeast do not have the ability to catabolize glucose to pyruvate. The T298S and T298A mutants were purified to greater than 95% homogeneity based on SDS-PAGE, with yields of 35 and 30 mg/liter of culture, respectively. The T298V mutant was not purified.

Biophysical Characterization of T298S and T298A
The secondary structures of the apo T298S and T298A mutants were monitored by far UV-CD and compared with wild type YPK (data not shown). Both mutants showed a slight and consistent difference in the magnitude of the molar ellipticity relative to wild type enzyme. The general shape of the far UV spectra of the Thr-298 mutants and wild type YPK was the same over the full spectral range (200 -280 nm). These results indicate that neither mutation caused any significant changes in the secondary structure and that the wild type and mutant YPKs were folded into a similar if not identical structure.
The intrinsic tryptophan fluorescence emission spectra of apo T298S and apo T298A mutants were recorded between 310 and 400 nm and compared with wild type YPK. The single tryptophan, Trp-452, was selectively excited at 295 nm so that observed changes in the intrinsic fluorescence could be correlated with conformational changes at a specific localized region in YPK. The apo forms of T298S and T298A YPK have a tryptophan emission maximum at ϳ334 nm, similar to that for apo wild type YPK (16).

Steady-state Kinetics
Steady-state kinetic measurements were performed for the wild type YPK and for the mutant enzymes under identical conditions. An example of the steady-state kinetic rate profiles for T298S at variable PEP concentrations with Mg 2ϩ and with Mn 2ϩ as divalent activator in the absence or presence of Fru-6-P 2 is shown in Fig. 1. Similar profiles were obtained with T298A (data not shown). The kinetic responses of wild type and the Thr-298 mutants with either Mg 2ϩ or Mn 2ϩ as the activator and in the absence or presence of Fru-6-P 2 were fit to Equations 1 and 2 as appropriate. The best fits from the appropriate equations were used accordingly. A summary of the calculated steady-state parameters is presented in Table I.
The mutant enzymes showed differential allosteric responses, depending upon the divalent metal activator (Mg 2ϩ or Mn 2ϩ ). The T298S mutation showed alterations primarily in k cat . The k cat /K m,PEP values for wild type and T298S were similar. The divalent metal specificity with T298S was the same as with wild type. Based on k cat values, Mg 2ϩ Ͼ Mn 2ϩ . The T298S mutation had a minor effect on K m,PEP only in the presence of Mn 2ϩ . The Mn 2ϩ -activated T298S did not display kinetic cooperativity with PEP in the absence of Fru-6-P 2 , with n H ϭ 1 compared with n H ϭ 2.5 for wild type YPK. T298S activated by Mg 2ϩ showed cooperative behavior (n H ϭ 1.9) similar to that of wild type enzyme (n H ϭ 2.8).
The turnover rates for T298A decreased by a factor of 24 and in the presence and absence of Fru-6-P 2 (FBP). The velocity response in the presence of Fru-6-P 2 (q) is hyperbolic and is best fit to Equation 1. The velocity response in the absence of Fru-6-P 2 (OE) is sigmoidal and is best fit to Equation 2. B, the velocity response of Mn 2ϩ -activated (4 mM Mn 2ϩ ) T298S YPK is shown as a function of [PEP] in the presence (q) and absence (OE) of Fru-6-P 2 . The velocity response in the presence and absence of Fru-6-P 2 is hyperbolic, and both sets of data are best described by Equation 1. The parameters obtained from the fits are listed in Table I. 11 compared with wild type YPK when activated by Mg 2ϩ in the absence and in the presence of Fru-6-P 2 , respectively. Fru-6-P 2 increased the k cat of the reaction with T298A and decreased K m,PEP with Mg 2ϩ as the activator. The decrease in k cat for T298A relative to wild type YPK in the absence or presence of Fru-6-P 2 was ϳ18-fold with Mn 2ϩ as the activator. The divalent metal specificity with T298A based on k cat was the same as with wild type and with T298S. The steady-state interaction between PEP and YPK, as measured by K m and k cat /K m values, has been affected significantly by the T298A mutation. The thermodynamic interaction of PEP with the enzyme was unaffected (see below). In the absence of Fru-6-P 2 , the Mn 2ϩ -activated T298S and T298A enzymes did not exhibit positive cooperativity with PEP. T298A displayed slight negative cooperativity for PEP in the presence of Fru-6-P 2 (n H ϭ 0.7 compared with n H ϭ 1 in wild type). T298A activated by Mg 2ϩ had kinetic behavior similar to that observed for wild type enzyme and T298S YPK activated by Mg 2ϩ .
The K m,M 2ϩ for Mn 2ϩ and for Mg 2ϩ (Table I) is an apparent value based on the total metal concentration and was not corrected for the divalent metal bound by ADP in solution. The concentration of ADP is constant and saturating in all of the kinetic studies. Mutation of Thr-298 to serine had no effect on K m,Mg 2ϩ and resulted in a 3-fold decrease of K m,Mn 2ϩ compared with wild type YPK. Mutation of Thr-298 to alanine had no effect on the K m,M 2ϩ values.

Effects of pH on Catalysis by Wild Type YPK and Thr-298 Mutants
Any relevance of Thr-298 to an ionization that affects catalysis was addressed by measuring pH rate profiles. The effect of pH on the V max,app for the Mn 2ϩ -and the Mg 2ϩ -activated wild type, T298S, and T298A enzymes in the presence of Fru-6-P 2 was measured over the pH range of 4.5-9.1. The pH rate profiles for the enzymes activated by Fru-6-P 2 and with Mn 2ϩ and with Mg 2ϩ are shown in Fig. 2, A and B, respectively. A summary of the pK a values obtained from the fit of the data in Fig. 2 to Equations 3-5, as appropriate, is presented in Table  II. Interpretation of the pK a values obtained from kinetic studies should be done with care. They may not be true thermodynamic values reflecting the microscopic pK a of a specific ionizable group (25). The K m,PEP values measured for the Mn 2ϩactivated wild type, T298S, and T298A mutants in the presence of Fru-6-P 2 were invariant over this pH range allowing the calculation of pK a values for "E free ." The pH data for the Fru-6-P 2 -and M 2ϩ -activated wild type YPK in Fig. 2, A and B, were fit to Equation 3, and describe three ionizations for catalysis in the ES complex. A single deprotonation with a pK a of ϳ5.5 and a single protonation of a group with a pK a of ϳ8.5 were required for catalytic activity. Deprotonation of a group in the pH range of 6 -8 altered the rate of reaction (Table II) (13). Studies (13) with K240M YPK indicate that the pK a of 8.8 was lost upon mutation of Lys-240. Lys-240 plays a putative role in phosphoryl transfer in the mechanism of YPK catalysis.
The data for the pH effect on V max,app for the Mn 2ϩ -and Fru-6-P 2 -activated T298S ( Fig. 2A) were fit to Equation 3, and the resulting pK a values are very similar to the values obtained for wild type YPK (Table II). The data for the effect of pH on V max,app for the Mg 2ϩ -and Fru-6-P 2 -activated T298S (Fig. 2B) were fit to Equation 4. This equation describes only two ionizations, K A and K B , in the ES complex and gives a better statistical fit than Equation 3. The calculated pK a values are similar to the values for pK A and pK B obtained with the wild type YPK. The value for pK C appears to be greater than 9.1 and hence could not be measured because of experimental limitations.
The data for the pH effect on V max,app for T298A in the presence of Fru-6-P 2 and with Mn 2ϩ or Mg 2 (Fig. 2, A and B) were fit to Equation 5. This equation describes the model where only two ionizable groups are required for catalytic activity, K A and K C . The resulting pK a values are similar to the values for pK A and pK C obtained with the wild type YPK (Table II). The results indicate that the pK B ϭ 6.4 -6.9 was lost upon removal of the hydroxyl group at position 298.

Ligand Binding to Wild Type and Thr-298 Mutants of YPK
The effect of the mutations at Thr-298 of YPK on the binding of Mn 2ϩ , PEP, and Fru-6-P 2 to various YPK complexes was quantitated by measuring the quenching of the intrinsic tryptophan fluorescence as a function of ligand concentration. The steady-state fluorescence data were fit to Equations 6 or 7. The dissociation constants, K D , maximal quenching, Q max , and Hill coefficients, n H , for ligand interactions with the enzyme complexes were obtained from appropriate fits to the data and are reported in Table III. The fluorescence response of apo wild type YPK and apo Thr-298 mutants to Mn 2ϩ follows simple saturation behavior that were fit to Equation 6. The K D,Mn 2ϩ and Q max values do not  Thr-298 mutants in the presence of saturating Fru-6-P 2 could not be measured by steady-state fluorescence because no additional change in the intensity of the fluorescence emission spectra of the YPK⅐Fru-6-P 2 or YPK⅐PEP⅐Fru-6-P 2 complexes was observed upon the addition of Mn 2ϩ . Binding of Mg 2ϩ to YPK complexes of wild type YPK and the Thr-298 mutants could not be monitored by steady-state fluorescence. Titration of Mg 2ϩ to apo YPK and YPK⅐PEP complexes did not result in measurable quenching of the fluorescence emission spectra of these complexes. These data also indicate that the binding of the two activating cations, Mg 2ϩ and Mn 2ϩ , induce different conformational responses of YPK.
Titration of the apo forms of the enzymes with PEP resulted in significant quenching of the tryptophan fluorescence (Q max ϳ20%). PEP binds to apo wild type YPK and apo Thr-298 mutants in a hyperbolic manner (data not shown) and with similar K D values. Fig. 3, A and B, shows data for the fluorescence response of T298S and T298A to PEP in the presence of 15 mM Mn 2ϩ . The response of both Thr-298 mutants is sigmoidal, and the data were fit to Equation 7. With Mn 2ϩ as the activating cation and in the absence of Fru-6-P 2 , neither Thr-298 mutant showed kinetic cooperativity with PEP as the variable substrate ( Table I). The K D,PEP was increased 2-fold with T298S⅐Mn 2ϩ and 70-fold with T298A⅐Mn 2ϩ , relative to wild type YPK⅐Mn 2ϩ . 3 The interaction of PEP with the YPK⅐Mg 2ϩ complexes of wild type YPK and the two Thr-298 mutants is described by a simple hyperbola (Equation 6). K D,PEP to the enzyme⅐Mg 2ϩ complex was not significantly affected upon mutation of Thr-298 to Ser or Ala. PEP binding to the YPK⅐Mg 2ϩ complex is weaker than binding to the YPK⅐Mn 2ϩ complex by 280-fold with wild type YPK, 150-fold with T298S, and 3-fold with T298A. The binding of PEP to the YPK⅐Fru-6-P 2 complexes could not be monitored by steady-state fluorescence for the same reason that the binding of Mn 2ϩ to these complexes could not be measured; quenching of YPK by Fru-6-P 2 precludes further quenching.
Binding of Fru-6-P 2 to the apo forms of wild type YPK and Thr-298 mutants resulted in a large quenching of the intrinsic  Table II.   TABLE II pH effects on k cat for wild type, T298S, and T298A YPK The effect of pH on k cat was measured as described under "Experimental Procedures," with PEP as a variable substrate when Mn 2ϩ was the divalent activator, and at saturating concentrations of PEP when Mg 2ϩ was the divalent activator. In each experiment, Fru-6-P 2 was present at saturating concentrations. tryptophan fluorescence (Q max ϳ50%). The interaction of Fru-6-P 2 with the apo forms of the enzymes was cooperative, with Hill coefficients of 2.4 -2.8. Fru-6-P 2 binding to apo T298S was 10-fold tighter than binding to wild type YPK, and Fru-6-P 2 binding to apo T298A was 2-fold tighter. The positive cooperativity in binding of Fru-6-P 2 to wild type YPK and to the two Thr-298 mutants in the presence of saturating PEP remained unchanged (n H ϭ 2.3 to 2.7). The K D,Fru-6-P 2 decreased 1.5-fold with wild type-YPK⅐PEP and increased 4-fold with T298S⅐PEP and T298A⅐PEP compared with Fru-6-P 2 binding to the apo forms of these enzymes. The interaction of Fru-6-P 2 with the YPK⅐M 2ϩ (Mn 2ϩ or Mg 2ϩ ) complex of wild type and Thr-298 mutants was enhanced compared with binding to apo YPK or YPK⅐PEP complexes of these enzymes, and the positive cooperativity remained unchanged (n H ϭ 2.0 -3.1). Saturation of wild type YPK and Thr-298 mutants with both Mn 2ϩ and PEP reduced but did not abolish the positive cooperativity in Fru-6-P 2 binding (n H ϭ 1.2 to 1.5). Fru-6-P 2 binds significantly tighter to the ternary YPK⅐Mn 2ϩ ⅐PEP complex than to either of the respective binary complexes of wild type and Thr-298 mutants. Hill coefficients of 1.3-1.4 were observed for the interaction of Fru-6-P 2 with the YPK⅐Mg 2ϩ ⅐PEP complex of wild type and T298S but was significantly greater with T298A (n H ϭ 2.7).
In summary, the mutation of threonine 298 to serine or to alanine altered binding of Mn 2ϩ to the YPK⅐PEP complex, binding of PEP to the YPK⅐Mn 2ϩ complex, and binding of Fru-6-P 2 to all of the respective enzyme complexes (Table III).

Fluorescence Spectra of YPK Complexes
The conformational response of wild type YPK and of the two Thr-298 mutants to ligand binding was analyzed by fluorescence spectroscopy. The single tryptophan, Trp-452, was excited at 295 nm. Fig. 4, A and B, shows the results of emission scans of T298S and T298A, respectively. Each enzyme was sequentially titrated with Mg 2ϩ , PEP, and Fru-6-P 2 . The apo form of the two mutants in the presence of buffer and KCl had an emission maximum at ϳ334 nm. Upon saturation of T298S with Mg 2ϩ , the emission maximum was shifted to the red by ϳ4 nm, and the maximal fluorescence quenching was 7%. The addition of kinetically saturating amounts of Mg 2ϩ to T298A did not cause a change in the emission spectrum of Trp-452. Mg 2ϩ binding to wild type YPK did not cause significant changes in the fluorescence emission spectrum (17). Saturation of the YPK⅐Mg 2ϩ complex with PEP caused 25% fluorescence quenching in T298S and 16% quenching and a 4 nm red shift in T298A. In wild type YPK, addition of PEP to the YPK⅐Mg 2ϩ complex resulted in 23% quenching and a 2 nm red shift. In both Thr-298 YPK mutants, addition of Fru-6-P 2 to the YPK⅐Mg 2ϩ ⅐PEP complex caused a large red shift (ϳ12 nm in T298S and 10 nm in T298A), a total fluorescence quenching of 37%, and a broadening of the spectra. The overall fluorescence quenching of 50% observed for the wild type YPK⅐  3. Interaction of PEP with YPK⅐Mn 2؉ . The binding of PEP to YPK⅐Mn 2ϩ was measured by steady-state fluorescence quenching at 334 nm. A, the interaction of PEP with T298S⅐Mn 2ϩ . The data are best fit to Equation 7 (-), which describes a sigmoidal response. For comparison purposes the fit of the data to Equation 6 (---), which describes a rectangular hyperbola, is also shown. B, the interaction of PEP with T298A⅐Mn 2ϩ . The data are best fit to Equation 7. The best fit parameters are listed in Table III. Mg 2ϩ ⅐PEP⅐Fru-6-P 2 complex was significantly different from the value of 37%, as measured for the two Thr-298 mutants. This suggests that the environment surrounding Trp-452 in the fully ligated complex with Mg 2ϩ is different in the two Thr-298 mutants from that of wild type YPK. It is clear from Fig. 4, A and B, that binding of Mg 2ϩ to YPK and the subsequent titration of PEP resulted in different responses in the two Thr-298 mutants. Fig. 4, C and D, shows the results of fluorescence emission of T298S and T298A sequentially titrated with Mn 2ϩ , PEP, and Fru-6-P 2 , respectively. Upon saturation of YPK with Mn 2ϩ , a fluorescence quench of 11 and 7% was obtained with T298S and T298A, respectively. A slight red shift in the fluorescence emission spectrum was observed for the T298A⅐Mn 2ϩ complex. In wild type YPK, Mn 2ϩ binding causes a 2 nm blue shift and a fluorescence quenching of 9% (16). Saturation of the YPK⅐Mn 2ϩ complex with PEP caused 32% fluorescence quenching and a 2 nm blue shift in T298S and 36% quenching in T298A. Addition of PEP to the wild type YPK⅐Mn 2ϩ complex resulted in a red shift of 2 nm and an increase in the fluorescence quenching of 41%. In both Thr-298 YPK mutants, addition of Fru-6-P 2 to the YPK⅐Mn 2ϩ ⅐PEP complex caused a large red shift (ϳ13 nm in T298S and 15 nm in T298A), a total fluorescence quenching of 38%, and a broadening of the spectra. Saturation of the wild type YPK⅐Mn 2ϩ ⅐PEP complex with Fru-6-P 2 elicited a red shift of ϳ10 nm and an overall fluorescence quenching of 52%.
Regardless of the activating divalent cation, the total fluorescence quenching in the fully ligated YPK⅐M 2ϩ ⅐PEP⅐Fru-6-P 2 complex was ϳ38% with the two Thr-298 mutants and 52% with wild type YPK. These results indicate that the environments surrounding Trp-452 in the fully ligated complex are similar in the two Thr-298 mutants regardless of the divalent cation but are different from the environment around Trp-452 in wild type YPK. It is evident from Fig. 4 that the two Thr-298 mutants show differences in the responses to M 2ϩ binding and subsequent PEP titration. Small conformational changes introduced at the catalytic site upon mutation of Thr-298 to Ser or to Ala are reflected in differences in quenching of the maximal fluorescence emission of Trp-452.

Partial Reactions
The effects of the Thr-298 mutations on the two partial reactions catalyzed by YPK were quantitatively assessed to determine the relative influence of Thr-298 on each catalytic step.
Phosphoryl Transfer-The effects of mutation of Thr-298 to Ser or Ala on the phosphoryl transfer half-reaction were determined by measuring the glycolate kinase activity with wild type and the two Thr-298 mutants. Glycolate kinase activity is a secondary kinase reaction where PK catalyzes the ATP-dependent phosphorylation of glycolate (15,26). Because glycolate lacks the C-3 vinyl group, glycolate kinase reflects phosphoryl transfer in the absence of proton transfer. The rates of the glycolate kinase reaction were measured in the presence of Fru-6-P 2 and with Mn 2ϩ or with Mg 2ϩ as the cation activator, at pH 7.5. The initial velocity response to ATP concentration with wild type YPK and the two Thr-298 mutants followed Michaelis-Menten kinetics. The rate constant for phosphoryl transfer, v p , by wild type YPK was 60.5 and 8.7 min Ϫ1 in the presence of Mn 2ϩ and Mg 2ϩ , respectively. With T298S, the rate of the glycolate kinase reaction in the presence of Mn 2ϩ was 80% of the activity measured with wild type and in the presence of Mg 2ϩ had the same glycolate kinase activity as wild type YPK. With T298A, the rate of the glycolate kinase reaction was ϳ20% of the rate measured with wild type YPK in the presence of either Mn 2ϩ or Mg 2ϩ as activators (Table IV).
The apparent turnover number of the glycolate kinase reaction divided by K m,ATP with each of the metal ions varied by less than a factor of 2 among wild type YPK, T298S, and T298A. These results indicate that the two mutations at position 298 have a minimal effect on the interaction of the enzyme with the metal-nucleotide substrate.
tored via the enzyme-catalyzed exchange of methyl protons of 3-[ 3 H]pyruvate into solvent (pyruvate enolization) (6). The rate of pyruvate enolization by T298S and T298A was measured and compared with the enolization rate for wild type YPK. The normalized rates are net rates corrected for the rate of spontaneous detritiation of pyruvate in solution. Mutation of Thr-298 to Ser or Ala resulted in a significant decrease of the rate of pyruvate detritiation relative to wild type. The observed effects are also metal-dependent. The rates of detritiation for Fru-6-P 2 -activated Thr-298 mutants in the presence of Mn 2ϩ are ϳ10% of the rate measured with wild type YPK under similar conditions. The rates of exchange for Fru-6-P 2 -activated T298S and T298A in the presence of Mg 2ϩ are 60 and 40% of the detritiation activity measured with wild type YPK, respectively. Both mutants gave similar values for v T . With wild type YPK, Mn 2ϩ is a 6-fold better activator of pyruvate enolization than is Mg 2ϩ in the presence of Fru-6-P 2 . With the two Thr-298 mutants, Mn 2ϩ and Mg 2ϩ appear to activate enolization of pyruvate to the same extent (Table V). There was no detectable detritiation of pyruvate by Mg 2ϩ -activated wild type YPK and T298S in the absence of Fru-6-P 2 . The results of these experiments indicate that the loss of the functional group at position 298 (T298A) in YPK does not eliminate proton transfer. Mutation of Thr-298 to Ser or Ala affected both the phosphoryl transfer and proton transfer steps in the PK-catalyzed reaction. Both mutations of Thr-298 affected pyruvate enolization to a greater extent than they affected phosphoryl transfer.

Solvent Isotope Effects
Solvent isotope effects reflect the importance of solvent-exchangeable protons in catalysis. If Thr-298 is directly involved in proton transfer in the PK-catalyzed reaction, then mutation to serine or alanine should decrease the reaction rate of proton transfer relative to rates of other steps in the catalytic process. Removal of the functional group at this position (T298A) should elicit a strong change in the SIE on k cat if enolate protonation by bulk water occurs. Solvent isotope effects were measured for the Fru-6-P 2 -activated wild type YPK and for the two Thr-298 mutants in the presence of either Mn 2ϩ or Mg 2ϩ as divalent activator and with PEP as the variable substrate. This allowed the measurements for solvent isotope effects on k cat , [ D k cat ], and on the k cat /K m,PEP , [ D (k cat /K m,PEP )]. The results indicate that with wild type YPK, the values for D k cat are metal-dependent. The isotope effect on k cat was metal-independent with the Thr-298 mutants (Table VI).
The isotope effect on k cat /K m,PEP describes the solvent effect on all catalytic steps up to and including the first irreversible step of the reaction. The first irreversible step is presumed to be phosphoryl transfer from PEP to ADP. The values for D (k cat / K m,PEP ) are listed in Table VI. The isotope effect on k cat /K m,PEP was similar for the Mn 2ϩ -and Mg 2ϩ -activated wild type YPK, respectively. The values for D (k cat /K m,PEP ) for T298S were similar to those for wild type YPK. The isotope effect on k cat /K m,PEP was eliminated for T298A regardless of the metal activator.

Proton Inventory
Proton inventory studies were performed for the Mn 2ϩ -and for the Mg 2ϩ -activated wild type YPK and for the two Thr-298 mutants in the presence of Fru-6-P 2 . This study was undertaken in an effort to dissect the individual contributions to the observed net SIE in wild type YPK and for the Thr-298 mutants. The proton inventories with Mn 2ϩ -and Fru-6-P 2 -activated wild type YPK and Thr-298 mutants were all linear, and the data were fit to Equation 8 (Fig. 5A). A linear proton inventory is indicative of a solvent-sensitive step involving a single proton in the transition state that contributes to the SIE. The calculated fractionation factors, T , were 0.36, 0.47, and 0.64 for wild type YPK, T298S, and T298A, respectively (Table  VI). With Mg 2ϩ as the activator, wild type YPK gave a linear proton inventory effect. The T298S and T298A mutants show a downward curvature in the proton inventory plots (Fig. 5B). The data in Fig. 5B were fit to Equation 8 for wild type YPK and to Equation 9 for the two Thr-298 mutants. Equation 9 describes contributions of a proton in the transition state and a proton in the reactant state to the observed isotope effect. This is the simplest model that gives the best fit to the data obtained with the Mg 2ϩ -activated Thr-298 mutants. The values obtained for the fractionation factors are listed in Table VI. The fractionation factors are significantly less than 1 with wild type YPK and the two Thr-298 mutants. Such low fractionation factors are distinct and suggest that the proton responsible for the observed overall solvent isotope effect in each case may be derived from a metal-bound water. The nonlinear proton inventory data for T298S and T298A can alternatively be fit by assuming two protons are involved in the transition state. If this model is used, equally good fits are obtained. For T298S the fit to this model gives T 1 ϭ 0.21 and T 2 ϭ 1.64. With T298A, the fit for such a model gives T 1 ϭ 0.43 and T 2 ϭ 1.36. In this model, one fractionation factor is still quite low, whereas the value of T for the second proton is Ͼ1 indicating perhaps a hindered proton that binds tighter than proton binding in bulk water.
Table VI summarizes the theoretical values for the solvent isotope effect on k cat obtained from the ratio of the experimentally measured value for k cat in H 2 O (n ϭ 0), (k cat ) 0 , and the fitted value for k cat in D 2 O (n ϭ 1), (k cat ) 1 , to Equation 8 or 9.

DISCUSSION
Threonine 298 is located at the active site of pyruvate kinase based on the recent x-ray crystal structures of both the yeast (11) and the muscle (12) enzymes. The orientation of Thr-298 relative to the 2-si face of PEP, determined by modeling PEP relative to phosphoglycolate or pyruvate, suggests a putative role in the protonation of the enolate of pyruvate at the C-3 position. Thr-298 was mutated to serine and to alanine in YPK in an attempt to clarify its role in catalysis. If Thr-298 is the proton donor in yeast pyruvate kinase, substitution with serine  (13). Reaction mixtures consisted of 100 mM TAPS (pH 8.0), 4% glycerol, 200 mM KCl, 100 mM 3-[ 3 H]pyruvate (75,170 dpm/mol pyruvate in the experiments with wild type YPK and T298S or 72,900 dpm/mol pyruvate in the experiments with T298A), 2 mM ATP, and 100 -300 g of YPK in 100 mM HEPES (pH 7.5). The divalent metal was either 10 mM MnCl 2 or 15 mM MgCl 2 , as indicated. The concentration of Fru-6-P 2 , when present, was 1 mM.
Complex is expected to result in minor alterations in net catalytic activity and a reaction where the enzyme still retains the proton donating ability. Mutation of Thr-298 to alanine would be expected to abolish completely the proton donating ability and hence to eliminate the net catalytic activity of the enzyme. It is possible that phosphoryl transfer might remain "normal," but the enolate of pyruvate is released from the enzyme as the second product. The enolate would then undergo protonation in solution.
The T298S mutant of YPK is catalytically active with minor alterations in the kinetic constants as expected. The effect of T298S on the k cat , K m,PEP and k cat /K m,PEP are minor. This conservative mutation results in small alterations in steadystate kinetic responses except with Mn 2ϩ as the activator. In the presence of Mn 2ϩ but in the absence of Fru-6-P 2 , this mutant does not demonstrate cooperative kinetics with PEP.
The T298A mutant results in a catalytically active enzyme that was superficially not anticipated. The value for k cat is decreased by about an order of magnitude relative to wild type. The k cat value for the Mg 2ϩ -activated enzyme is doubled in the presence of Fru-6-P 2 compared with the value in its absence (Table I). An increase in the total Mg 2ϩ concentration to 22 mM in the kinetic assay with T298A in the absence of Fru-6-P 2 did not change the value for k cat . The presence of Fru-6-P 2 induces a significant decrease in the K m,PEP and an increase in k cat for the Mg 2ϩ -activated T298A resulting in a 20-fold increase in catalytic efficiency. The apparent K m for Mg 2ϩ with T298A has been measured to be ϳ10.4 mM in the absence of Fru-6-P 2 (data not shown). The modification of Thr-298 to alanine at the active site of YPK must cause an alteration at the catalytic site such that in the absence of Fru-6-P 2 , a less active conformation of the enzyme is induced. This altered active conformation at the catalytic site in T298A is less competent to accommodate Mg 2ϩ and PEP binding than that with wild type or T298S YPK. This conclusion is reinforced by the fluorescence emission spectra of T298S and T298A sequentially titrated with Mg 2ϩ , PEP, and Fru-6-P 2 (Fig. 4, A and B). The T298A mutation decreases the k cat /K m,PEP significantly relative to the value for wild type YPK.
The elimination of the alcohol function at the active site affects the steady-state interaction between PEP and free enzyme. Mn 2ϩ -activated T298S and T298A do not exhibit homotropic kinetic cooperativity with PEP as the variable substrate in the absence of Fru-6-P 2 . PEP binding to the YPK⅐Mn 2ϩ complex is cooperative, however (n H of 1.5 and 2.0 for T298S⅐Mn 2ϩ and T298A⅐Mn 2ϩ , respectively). The metal activators elicit different kinetic behavior for both Thr-298 mutants. Mutation of the active site Thr-298 to serine or alanine results primarily in k cat effects indicating alterations in the catalytic process. The steady-state kinetic data with T298S and T298A support an important but not a critical function of Thr-298 in YPK catalysis. The turnover rate with T298A, which is still significant, suggests that Thr-298 is not the direct proton donor in the YPK-catalyzed reaction.
Small conformational alterations introduced at the active site of yeast PK by mutation of Thr-298 to serine or alanine trigger changes at the allosteric site that is located more than 40 Å away (11). These changes are also translated in an alteration of the allosteric response of the mutant enzymes. The altered allosteric response with the T298S and T298A YPK mutants is not surprising in the light of the data reported by Rigden et al. (27) on the allosterically regulated PK from the FIG. 5. Proton inventory plots for wild type, T298S, and T298A YPK. Initial velocities were measured at pL 6.2 (L ϭ H, D) and at saturating concentrations of the substrate PEP (5 mM) and of all ligands (4 mM MnCl 2 or 15 mM MgCl 2 , 5 mM ADP, 1 mM Fru-6-P 2 , and 200 mM KCl). A, proton inventories for Mn 2ϩ -and Fru-6-P 2 -activated wild type (WT, q), T298S (f), and T298A (OE) YPK. The lines represent the best fit of the data to Equation 8. B, proton inventories for Mg 2ϩand Fru-6-P 2 -activated wild type (q), T298S (f), and T298A (OE) YPK. The lines represent the best fit of the data to Equation 8 for wild type YPK and to Equation 9 for T298S and T298A YPK. The best fit parameters from the fits are given in Table VI. protozoan parasite Leishmania mexicana (allosteric activator fructose 2,6-biphosphate). Comparison of C ␣ root mean square differences among the crystal structures of the Leishmania, Saccharomyces, E. coli, and rabbit muscle enzymes show that the Leishmania and Saccharomyces enzymes are most similar (27). Superposition of the catalytic domain of PK from Leishmania (T-state) and Saccharomyces (R-state) reveals that the polypeptide backbone between residues 296 and 310 (298 -312 in YPK numbering) differs substantially between the T-and R-states. The conformational location of the two residues Thr-296 and Arg-310 (Thr-298 and Arg-312 in YPK numbering) appears to play an important role in the T-state to R-state allosteric transition. Previous studies (16,28) have shown that binding of Mn 2ϩ or Mg 2ϩ causes a conformational change that favors the communication between the PEP and Fru-6-P 2 sites. A plausible theory that emanates from the study of Rigden et al. (27) is that the binding of PEP to PK elicits a change in conformation of Thr-298 (YPK numbering). A transmission of this change through the main chain of the protein to Arg-312 is required to effect an allosteric transition from the active site. Two separate signals may be necessary to determine an allosteric transition from the catalytic site of PK, one from the enzyme-bound divalent activator and the second from PEP via Thr-298. The interaction between the M 2ϩ activator and PEP with YPK is highly coupled (16,28). The role of Thr-298 in the reaction catalyzed by YPK seems to be complex and involves both catalytic and regulatory functions.
The pH dependence of k cat for wild type YPK is described by the same three pK a values regardless of the activating cation (Table II). Bollenbach et al. (13) showed that the catalytically important group with a pK a of 8.8 has been lost on mutation of Lys-240 to methionine. The mutation of Thr-298 to serine had little effect on the pH dependence of k cat when Mn 2ϩ was the divalent activator. In the pH rate profile for the Mg 2ϩ -activated T298S, pK C cannot be determined. This results from a shift of the basic pK a for this group. Hence a pK C Ͼ9.1 cannot be measured because of experimental limitations. The pH dependence of k cat for Mn 2ϩ -and Mg 2ϩ -activated T298A indicates that the second ionization with a pK a of 6.4 (Mn 2ϩ ⅐YPK) and of 6.9 (Mg 2ϩ ⅐YPK), respectively, has been lost on mutation. It is this ionization that is responsible for modulation of k cat in the YPK-catalyzed reaction. Although it may be tempting to attribute pK B ϭ 6.4 -6.9 to Thr-298, it is likely that Thr-298 is important for an ionization with pK a 6.5. Cleland (25) has cautioned against the temptation to interpret such pK a values as intrinsic thermodynamic values.
Glycolate kinase activity and proton exchange rates were measured to determine independently how each of the two specific chemical steps in the reaction may be affected by the alteration of Thr-298. The phosphoryl transfer step in pyruvate kinase can be measured independently by the "glycolate kinase" activity of PK, one of several PK-catalyzed secondary kinase reactions (15,26,29). The second half-reaction catalyzed by pyruvate kinase, the proton transfer to enolpyruvate, can be measured by the PK-catalyzed deprotonation (enolization) of pyruvate in the presence of ATP. This reaction occurs in the absence of net phosphoryl transfer (6). The rates of the glycolate kinase reaction measured with wild type YPK and the Thr-298 mutants in the presence of Fru-6-P 2 are metal-dependent. Mn 2ϩ is a better activator of glycolate kinase than is Mg 2ϩ . Mutation of Thr-298 to serine does not affect the glycolate kinase activity. The mutation of Thr-298 to alanine results in a 5-fold decrease of the glycolate kinase activity compared with wild type YPK. Phosphoryl transfer precedes proton transfer in the PK-catalyzed reaction, and the two reactions were suggested to be decoupled from each other in the wild type enzyme (6). The relative rates of the glycolate kinase reaction with the Thr-298 mutants compared with wild type YPK do not correlate with the relative k cat values for the net reaction measured by steady-state kinetics. This suggests that the active site Thr-298 does not directly affect the phosphoryl transfer step of the PK-catalyzed reaction. There is evidence 4 that the mutations of Thr-298 elicit conformational changes of bound PEP at the catalytic site where phosphoryl transfer does occur, although these changes do not extend into the nucleotide-binding site. The kinetic studies of glycolate kinase activity indicate that the interaction of ATP with the enzyme is unaffected by mutation of Thr-298.
The rates of detritiation of pyruvate measured with wild type YPK and the Thr-298 mutants are also metal-dependent. Mn 2ϩ is the preferred activator compared with Mg 2ϩ . Mutation of Thr-298 to serine or alanine results in a decrease in the rate of pyruvate enolization relative to wild type enzyme ( Table V). The T298A mutant catalyzes pyruvate enolization at 13-42% relative to wild type YPK. Loss of the functional group at position 298 in yeast pyruvate kinase decreases but does not eliminate the enolization reaction. Mutation of Thr-298 to serine or alanine alters the rates of both the phosphoryl transfer and proton transfer steps in the PK-catalyzed reaction. In neither reaction is the effect proportional to the effect on the overall reaction, nor are the effects parallel for the two partial reactions. Both mutations of Thr-298 affect pyruvate enolization to a greater extent than they affect phosphoryl transfer. These results strongly suggest that Thr-298 is not the direct proton donor in the PK-catalyzed reaction but may play a role in the proton transfer step. The results summarized in Table V can be explained by a role for the divalent activator both in stabilizing the enolate intermediate and in fostering the process of proton transfer. Based on the above results, the two steps in the net catalytic reaction, phosphoryl transfer and pyruvate enolization, appear to be coupled. Such a conclusion was also reached from the studies of the K240M mutant of YPK (13).
The single tryptophan residue (Trp-452) of yeast pyruvate kinase provides a unique probe for monitoring local conformational changes in the vicinity of the Fru-6-P 2 site. The intrinsic fluorescence of Trp-452 is quenched upon ligand binding, allowing for the measurement of thermodynamic dissociation constants of ligands from various YPK complexes (16,28). Binding of Mn 2ϩ and PEP to apo YPK and of PEP to the YPK⅐Mg 2ϩ complex has not been affected by mutation of Thr-298 at the active site to serine or alanine. The binding of Mn 2ϩ to the YPK⅐PEP complex is 3-and 20-fold weaker in T298S and T298A, respectively, relative to wild type YPK. The K D of PEP from the YPK⅐Mn 2ϩ complex is increased by 2-and 70-fold in T298S and T298A, respectively, compared with wild type YPK. It is possible that the thermodynamically "preferred" ordered pathway occurring in wild type YPK, where either PEP or Mn 2ϩ binds first to form the YPK⅐Mn 2ϩ ⅐PEP complex before ADP binds (16), is changed to a more ordered pathway in T298A, with PEP binding first followed by Mn 2ϩ . These results suggest that the mutations of Thr-298 affect the conformational response of YPK to substrate (and activator) binding. PEP binding to the YPK⅐Mn 2ϩ complex of both Thr-298 mutants is cooperative. Although kinetic cooperativity with PEP is no longer observed with both mutants in the presence of Mn 2ϩ and in the absence of Fru-6-P 2 , cooperative binding of PEP to the YPK⅐Mn 2ϩ complex occurs. Similar behavior was observed with the slow substrate Br-PEP for Mn 2ϩ -activated wild type YPK in the absence of Fru-6-P 2 (30). In the latter case, it was proposed that the cooperative binding of Br-PEP to wild type YPK is masked by a late slow kinetic step and therefore cooperativity in kinetics is not observed. The possibility of kinetic factors in the cooperative response of T298S was investigated by measuring the velocity response to variable [PEP] in a temperature range from 8 to 40°C. 4 The apparent Hill coefficients for PEP do not change significantly over the temperature range studied. This suggests that if there is a kinetic step (or steps) that attenuates the cooperative response of PEP by the Mn 2ϩ -activated enzyme in the absence of Fru-6-P 2 , then this step is not temperature-sensitive. It is possible that in the presence of Mn 2ϩ but in the absence of Fru-6-P 2 , the two Thr-298 mutants already acquire the active conformation that is induced in wild type YPK by Fru-6-P 2 binding. The divalent cations Mg 2ϩ and Mn 2ϩ elicit different conformational effects on the enzyme (Table III). Binding of the allosteric activator Fru-6-P 2 to all the enzyme complexes studied has been altered upon mutation of the active site Thr-298. The effects are metal-dependent, reinforcing the fact that the coupling between the Fru-6-P 2 and PEP sites in yeast PK is modulated through the enzyme-bound metal (16,28). Again, it is evident that small conformational changes introduced at the active site of YPK by mutation of Thr-298 to serine or alanine can induce long range effects at the Fru-6-P 2 -binding site situated more than 40 Å away.
Solvent isotope effects on k cat and k cat /K m,PEP for the wild type and the two mutants of YPK were measured to investigate the influence of the solvent on proton transfer and net catalysis. 5 These experiments were further analyzed by proton inventory studies in an effort to determine the number of protons "in flight" in the isotope-sensitive step(s) with wild type YPK and the two Thr-298 mutants. A linear proton inventory response (V max versus the mole fraction of D 2 O in the solvent of D 2 O ϩ H 2 O) can be fit by the linear form of the Gross-Butler equation (Equation 8) that models a single proton in the transition state involved in the isotope-sensitive step in the reaction. Nonlinear responses can be more complex. They can be attributed to a proton in the transition state and in the reactant (Equation 9), multiple protons in the transition state, effects of commitment factors on the intrinsic isotope effects, and other complications. Some of these issues and their treatment have been addressed (23,24,33). Because we have limited knowledge about details of the solvent-sensitive steps of wild type YPK and even less information regarding the effects of Mn 2ϩ versus Mg 2ϩ and the effects of the Thr-298 mutants, our interpretation of these results are conservative.
It is feasible that upon alteration of Thr-298, enolpyruvate may be released into solution as the product. The SIE for protonation of free enolpyruvate in solution at pD 6.4 is 6.0 (3). Hence a maximum SIE on k cat of 6.0 would be expected if the acid/base catalyst is lost by YPK and if protonation of enolpyruvate in solution occurs and is rate-limiting in the catalytic process. The measured values for D k cat of 1.4 (with Mn 2ϩ ) and 1.3 (with Mg 2ϩ ) for T298A rule out this possibility.
The observation of D k cat values of Ͼ1 for wild type, T298S, and T298A indicates that the isotope-sensitive step is expressed in this kinetic parameter with each of the enzymes. These values are metal-dependent for wild type YPK and metal-independent for the two mutants. Values for D (k cat / K m,PEP ) Ͼ1 are also observed. The k cat /K m,PEP term reflects the interaction of PEP to the enzyme up to and including the first irreversible step. This is most likely phosphoryl transfer. The linear response to the proton inventory studies that were obtained with the Mn 2ϩ -activated wild type YPK and both Thr-298 mutants indicates a single solvent-related proton in the transition state. The fractionation factor () for each enzyme is significantly less than one suggesting that 1 H and not 2 H accumulates at the site of exchange. Values Ͻ1 for the fractionation factors have been identified with cysteine residues ( ϭ 0.40 -0.46) or metal-bound water (e.g. ϭ 0.69 (24)). Most other functional groups including alcohols have a ϭ 1.0. There is no cysteine residue in the active site of YPK, nor is there any evidence that such a residue plays a role in PK catalysis. Hence, our conclusion is that the proton in question is derived from metal-bound water in the active site of PK. The magnitude of the fraction factor increases in the order wild type Ͻ T298S Ͻ T298A suggesting that in T298A, the proton in transit does not interact with the donor water in the putative channel as selectively as in wild type YPK. In the case of T298A, the binding of hydrogen at the exchangeable site begins to approach the binding as in bulk water ( approaches unity). The effects observed are cation-dependent. The fractionation factors for the Mn 2ϩ -activated enzymes are larger than for the Mg 2ϩ -activated enzymes.
The proton inventory with Mg 2ϩ as divalent activator is linear with wild type YPK, whereas with T298S and T298A, the proton inventories are dome-shaped. The downward curvature of the proton inventories can be due to more complex phenomena. One model to explain the results obtained with the Mg 2ϩactivated Thr-298 mutants is that the net solvent isotope effect observed with the mutants arises from opposing contributions from a proton in the transition state ( T ) (as in the case with Mn 2ϩ ) and a proton in the reactant state site ( R ). The fractionation factors are significantly less than 1 with wild type YPK and both Thr-298 mutants, with either Mg 2ϩ or Mn 2ϩ as divalent activator (Table VI).
The proton inventory data for the two mutants activated by Mg 2ϩ can also be fit with a model of multiple protons in the transition state. A fit assuming two protons in the transition state gives one fractionation factor significantly lower than 1, which is consistent with this proton coming from a metal-bound water. The second proton has a fractionation factor greater than 1 and could reflect a tightly bound water where its effects are manifest in these mutants. This proton might either be that shared by Glu-334 and Ser-332 or a water bound to Glu-334 which is at the end of the channel. Our treatment of the solvent isotope data indicates that water plays a key role in catalysis by YPK. The low value for the fractionation factor suggests the catalytically important water is bound to the metal ion at the catalytic site.
If Thr-298 is part of a water network at the active site that is directly or remotely hydrogen-bonded to a metal-bound water, then the pK a of 6.4 -6.9 that is lost in T298A may be related to the ionization of water in the proposed channel. One of the water molecules in the channel is liganded to the enzymebound metal. The remote effect of the metal-bound water on enolpyruvate protonation is consistent with the weak correlation of the rate of pyruvate enolization with the pK a of metalbound water. From the semilog plot of the rate constant of pyruvate enolization by muscle PK with the pK a of metal-bound water, the apparent Hammett constant is 0.24 Ϯ 0.06 (34). Although a pK a of 6.4 -6.9 is quite low for the ionization of metalbound water (the pK a of Mg 2ϩ -bound water is ϳ12.8 (35)), it is important to realize that the pK a values obtained from kinetic studies do not necessarily reflect microscopic pK a values but rather group pK a values. A pK a of 6.8 has been measured for the Zn 2ϩ -bound water in carbonic anhydrase II (36), although the microscopic pK a of Zn 2ϩ -bound water is ϳ8.7 (25).
An alternative explanation for the lost pK a of 6.4 -6.9 comes from the studies of Larsen et al. (37). The crystal structure of the bis(Mg 2ϩ )-ATP-oxalate complex of rabbit muscle pyruvate kinase at 2.1 Å resolution reveals two water molecules "connecting" the hydroxyl groups of Ser-361 and Thr-327 to the carboxyl(ate) of Glu-363 (these are residues Ser-332, Thr-298, and Glu-334 in YPK). The side chains of Ser-332 and Thr-298 project into the region above the 2-si face of the proposed enolate of pyruvate. The side chain of Glu-334 is buried in the active site and is not in proximity of any cationic charge. Hence, Glu-334 might function as an acid/base catalyst in YPK through a relay of bound water and the hydroxyl groups of Ser-332 and Thr-298. It is possible that the pK a of 6.4 -6.9 could correspond to the carboxylate of Glu-334, perturbed by its environment.
A proton transfer network has been shown to operate at the active site of camphor cytochrome P450 monooxygenase (P450 cam ) (38,39). This channel involves the conserved residues Thr-252, Asp-251, and Glu-366 and three water molecules all of which can act as a proton-relay system from the solvent toward the dioxygen-ferryl complex (38,39). One water of this channel appears to serve as the proton donor for the hemebound O 2 . In P450 cam the hydroxyl group of Thr-252 was proposed to stabilize both the heme-bound O 2 and the active site water through hydrogen bonding (39).
In carbonic anhydrase II, ionization of a Zn 2ϩ -bound water is required in the last step of the mechanism to generate the nucleophilic Zn 2ϩ -bound hydroxide ion. The rate determining proton transfer is achieved by diffusion of a net proton across a hydrogen-bonded solvent network between Zn 2ϩ -bound solvent and His-64 (36). Hydrogen bonding of the resulting Zn 2ϩ -bound hydroxide with Thr-199 and Glu-106 orients the nucleophile with the optimal geometry required for nucleophilic attack at substrate CO 2 (39).
In conclusion, the identity of the general acid/base in the PKcatalyzed reaction is still unclear. It is unlikely that Thr-298 is the direct proton donor to the enolpyruvate intermediate based on the data presented from this study. Thr-298 plays a role in a late step in the catalytic mechanism involving proton transfer. Our data suggest that the direct proton donor to enolpyruvate in YPK may be a water molecule at the catalytic site that is part of a water channel. One water in this channel is coordinated to the enzyme-bound metal. Thr-298 is suggested as being the amino acid that interacts with the terminal water molecule of the proton circuit. The Thr-298 affects the pK a of the water in the channel and therefore its reactivity. The proposal of a water molecule as the ultimate proton donor and as part of a water channel is supported by the x-ray structure of the bis(Mg 2ϩ )-ATP-oxalate complex of rabbit muscle PK at 2.1 Å resolution, where specific water molecules are indicated (37).