Role of the divalent metal cation in the pyruvate oxidase reaction.

Purified pyruvate oxidase requires a divalent metal cation for enzymatic activity. The function of the divalent metal cation was studied for unactivated, dodecyl sulfate-activated, and phosphatidylglycerol-activated oxidase. Assays performed in the presence of Mg2+, CA2+, Zn2+, Mn2+, Ba2+, Ni2+, Co2+, Cu2+, and Cr3+ in each of four different buffers, phosphate, 1,4-piperazinediethanesulfonic acid, imidazole, and citrate, indicate that any of these metal cations will fulfill the pyruvate oxidase requirement. Extensive steady state kinetics data were obtained with both Mg2+ and Mn2+. All the data are consistent with the proposition that the only role of the metal is to bind to the cofactor thiamin pyrophosphate (TPP) and that it is the Me2+-TPP complex which is the true cofactor. Values of the Mg2+ and Mn2+ dissociation constants with TPP were determined by EPR spectroscopy and these data were used to calculate the Michaelis constant for the Me2+-TPP complexes. The results show that the Michaelis constants for the Me2+-TPP complexes are independent of the metal cation in the complex. Fluorescence quenching experiments show that the Michaelis constant is equal to the dissociation constant of the Mn2+-TPP complex with the enzyme. It was also shown that Mn2+ will only bind to the enzyme in the presence of TPP and that one Mn2+ binds per subunit. Steady state kinetics experiments with Mn2+ were more complicated than those obtained with Mg2+ because of the formation of an abortive Mn2+-pyruvate complex. Both EPR and steady state kinetics data indicated complex formation with a dissociation constant of about 70 mM.

11 Supported by National Institutes of Health Grant GM 7768. that there are basically 3 kinetic forms of the oxidase: unactivated, monomeric amphiphile-activated, and aggregated amphiphile-activated. The two activated forms of the enzyme are very similar; the only difference is that enzyme activated by aggregated amphiphiles displays cooperative kinetics with respect to the TPP cofactor, while enzyme activated by monomer amphiphiles does not (6).
In this paper, attention is drawn to the role of the divalent metal cation in the pyruvate oxidase reaction. The requirement for a divalent metal cation was recognized early in the purification of the enzyme (1). Historically, Mg2+ has been the metal cation of choice in TPP'-catalyzed reactions. All previous work with pyruvate oxidase has been conducted in the presence of a saturating concentration of Mg2+ (10 mM). Like many TPP-requiring enzymes (9-ll), however, the oxidase will permit other divalent metal cations to substitute for M e . As part of the goal of determining the mechanism of amphiphile activation of pyruvate oxidase, the function of the divalent metal cation was studied for all three kinetic forms of the enzyme using steady state kinetics, fluorescence, and EPR spectroscopy.
EXPERIMENTAL PROCEDURES Materials-All metal cations investigated were obtained as their chloride salts. Gadolinium'" and europium"+ were purchased from Apache Chemicals, Inc. Terbium'" was purchased from ICN Pharmaceuticals, Inc. All other metal cations were reagent grade. Magnesium'+ was obtained from Mallinckrodt Chemical Works. Nickel", cobalt", and manganese"+ were obtained from J. T. Baker Chemical Co. Copper" and zinc"+ were obtained from Fisher. Barium" and chromium2+ were obtained from Allied Chemical Co. Finally, cal-cium2+ was obtained from Matheson, Coleman, and Bell.
Thiamin pyrophosphate, sodium pyruvate, Pipes, and phosphatidylglycerol were purchased from Sigma. Sodium dodecyl sulfate was obtained from Eastman. Sodium ferricyanide was obtained from K & K Laboratories, Inc. All other chemicals were reagent grade.
Pyruvate oxidase was purified according to previously published procedures (12). Enzyme assays were conducted as previously described (6).
E P R Experiments-The EPK experiments were conducted at room temperature on a Varian E-9 EPR spectrometer. The EPR spectrum of 1 mM MnClr was recorded in 1) distilled water; 2) 0.05 M Pipes buffer, pH 6.0; 3 ) Pipes buffer plus 0.25,0.5,0.75, I.0,2.0, and 4.0 mM TPP; 4) Pipes buffer plus 4 mM T P P and 5, 10 subunits. All spectra were recorded at 9.5 GHz at room temperature, All fluorescence measurements were made with a Perkin-Elmer MPF-44A fluorimeter equipped with a circulating constant temperature bath. The temperature of the sample is measured using a YSI ' The abbreviations used are: TPP, thiamin pyrophosphate; Pipes, 1,4-piperazinediethanesulfonic acid. 9605 Ions in the Pyruvate Oxidase Reaction model 42SC telethermometer. To avoid large inner filter corrections due to TPP absorption, all fluorescence titrations were done with excitation a t 300 nm. Fluorescence intensity was followed a t 333 nm using an excitation band pass of 6 nm and an emission band pass of 10 nm. The Raman scattering intensity was negligible.
At the end of a titration, the fluorescence intensity values were corrected for small volume changes (always less than 5%j and for T P P inner filter effects. Since thiamin does not bind to pyruvate oxidase as judged from steady state kinetics, a thiamin quenching curve can be used as an empirical inner filter correction for the TPP titration data.

RESULTS
Survey of Metal Cations-Eight divalent cations and one trivalent cation were surveyed for their ability to fulfill the metal cation requirement shown by the oxidase. The results are shown in Table I. In the appropriate buffer, all of the metals could be observed to support pyruvate oxidase activity. In many cases, attempts to reach saturating concentrations of the metal cation led to precipitation and loss of activity. In those cases where the concentration of the metal cation could be increased to apparent saturation without precipitation occurring, the maximum velocity of the oxidase showed no apparent fluctuation. Thus, neither the metal ion nor the buffer appeared to influence the maximum velocity of the enzyme.  designates that no pyruvate oxidase activity was observed. + designates that pyruvate oxidase activity was present, but precipitation prevented full activity from being seen.
Repeated efforts to support pyruvate oxidase activity with the trivalent Tb"+, E d + , or Gd"+ ions were unsuccessful. All three ions at very low concentrations (1 PM or less) caused precipitation of the enzyme.
Kinetic Studies with Mg'?+"Initial velocity experiments were conducted on all three kinetic forms of the enzyme with Mg'+ in 0.1 M phosphate buffer. Lineweaver-Burk plots of velocity" versus TPP" a t fixed levels of Mg'+ and the alternate plots of velocity-' versus (Mg"')" at fixed levels of TPP are presented for unactivated, dodecyl sulfate-activated, and phosphatidylglycerol-activated pyruvate oxidase in Figs. 1, 2, and 3, respectively. The concentrations of each component were chosen such that the same data points could be used for each pair of plots. In all cases, the plots of velocity" versus the varied substrate-' at different levels of the changing fixed substrate yielded a family of lines that intersected at a common point on the velocity" axis. Secondary plots of the slope versus the changing fixed substrate" were linear and passed through the origin.
Similar experiments with Mg'+ and pyruvate or ferricyanide yielded the data summarized in Table 11. All these plots gave families of lines that intersected at a common point to the left of the velocity" axis. Secondary plots in all cases were linear and did not pass through the origin (data not shown).
Kinetic Studies with Mn"+-Initial velocity experiments similar to those just described were performed with Mn'+ in 0.05 M Pipes buffer. Lineweaver-Burk plots of velocity" uersus TPP" at fixed levels of Mn2+ and the alternate plots of velocity" versus (Mn'+)" at fixed levels of TPP were constructed for unactivated, dodecyl sulfate-activated, and phosphatidylglycerol-activated pyruvate oxidase (data not shown). The concentrations of each component were again chosen such that the same data points could be used for each pair of plots. As was the case when Mg2+ was the cation, all plots of velocity" uersus the varied substrate" at different levels of the changing fixed substrate yielded a family of lines that intersected at a common point on the velocity" axis. Secondary plots of the slope versus the changing fixed substrate" were linear and passed through the origin.
Similar experiments with Mn2+ and ferricyanide yielded the data also summarized in Table 11. All these plots gave families of lines that intersected a t a common point to the left of the In the presence of phosphatidylglycerol, the binding of TPP to the oxidase is cooperative. The data are plotted as a function of TPP" " to take into account this cooperativity. did not pass through the origin (data not shown).
Similar experiments conducted with Mn2+ and pyruvate resulted in data quite different from that obtained in the presence of Mg'+. Lineweaver-Burk plots of velocity" versus pyruvate" a t fixed levels of Mn" and the alternate plots of velocity" versus (Mn2+)" a t fixed levels of pyruvate were constructed for unactivated (Fig. 4), dodecyl sulfate-activated, and phosphatidylglycerol-activated pyruvate oxidase. For all three kinetic forms of the oxidase, the plots of velocity" versus pyruvate" a t fixed levels of Mn'+ were nonlinear and concave up. Thus, at low levels of Mn2+, pyruvate was apparently a very strong substrate inhibitor. The plots of velocity" versus (Mn")" at fixed levels of pyruvate yielded families of lines that were linear, but that did not pass through a common   (14). The data are plotted in the form of a double reciprocal plot (Fig. 5B). The dissociation constant obtained is 0.49 PM for the Mn'+-TPP complex to pyruvate oxidase. T h s close agreement between the dissociation constant and the Michaelis constant for TPP is consistent with earlier studies performed in the presence of Mg2+ (14).
Fluorescence quenching of pyruvate oxidase was also used to determine the stoichiometry of Mn2+ binding to the enzyme. It has already been demonstrated that TPP does not bind to the enzyme in the absence of a divalent cation (14). EPR experiments clearly demonstrate that Mn2+ will not bind to pyruvate oxidase in the absence of the cofactor TPP. All experiments were performed in the presence of 0.1 M Pipes buffer, pH 5.7. The results (Table 111) indicate that pyruvate oxidase has a substantial effect on the signal intensity from Mn" only in the presence of TPP. Little if any interaction between the metal and the enzyme is observed in the absence of the cofactor.
When the survey experiments were being conducted, it was observed that the colored cations (Ni", Co')+, CUI+, and C?+)  all exhibited subtle changes in their visible spectra when mixed with pyruvate. This observation indicated the possibility of an interaction between the metal cations and pyruvate. EPR experiments identical with those described above were performed with pyruvate instead of TPP, and indeed, at high concentrations of pyruvate, a Mn2+-pyruvate complex was detected. Titration with various concentrations of pyruvate yielded a dissociation constant of 71 mM for the Mn'+-pyruvate complex. Addition of Mg'+ up to 100 mM did not displace Mn2+ from the Mn'+-pyruvate complex. Therefore, under the conditions of the steady state assay, it can be assumed that no complex between Mg" and pyruvate exists. The data derived from EPR are compared to the steady state Michaelis constants in Table IV.

DISCUSSION
Pyruvate oxidase displays a lack of specificity for the metal cation required for enzymatic activity. The group of metal cations that will replace Mg2+ in the assay differ sufficiently in their properties to suggest that perhaps the metal cation is not directly involved in the catalytic process. Its function, then, could perhaps be that of binding to the TPP cofactor.
The resultink metal-cofactor complex would then be the actual cofactor. This rather passive role for the metal cation is quite common in enzymatic reactions in which nucleotide dior triphosphates participate. Alternatively, the metal cation could bind directly to the enzyme and promote subsequent TPP binding or act as an essential activator of the enzyme.
The experiments presented in this paper were performed to distinguish between these alternatives. All the data are consistent with the proposition that the true cofactor is a metal-TPP complex and indicate no other role for the metal. The steady state kinetics data are described using the equations derived under "Appendix," based on the model of the metal-TPP complex serving as the cofactor. The family of curves ( e g . Figs. 1-3) are consistent with Equations 4 and 8 (see "Appendix"). Various other schemes (Segal (15) lists seven such schemes) all lead to initial velocity equations that are more complex than Equation 4 and do not fit the data.
The only complication arises when steady state kinetics are performed with variable concentrations of Mn2+ and pyruvate. At low MnZ+ concentrations, pyruvate acts as a substrate inhibitor (Fig. 4) and effectively removes free Mn2+ from the solution. EPR demonstrates that, in fact, Mn2+ can form a complex with pyruvate with a dissociation constant of 71 mM. The apparent inhibitory effect of pyruvate can thus be explained and the kinetics data are consistent with Equation 8 (see "Appendix"). Equation 8 predicts the observed nonlinear behavior for the plots of velocity" versus pyruvate" at changing fixed levels of Mn2+. Furthermore, Equation 8 predicts Obtained with a saturating value of TPP by varying the concentration of Mn2+.
Obtained by analysis using Equation 4 under Appendix. Details are given in the text. the Pyruvate Oxidase Reaction that plots of velocity" versus manganese" at changing fixed levels of pyruvate will consist of a family of straight lines that do not have a common intersection point. Equation 8 also predicts that secondary plots of the slope versus pyruvate" will be nonlinear and that secondary plots of the intercept versus pyruvate" will be linear. This is in fact what is observed and, thus, the proposed scheme explains the unusual nature of the data. The K O (Mn"+-pyruvate dissociation constant) values calculated from unactivated, dodecyl sulfate-activated, and phosphatidylglycerol-activated pyruvate oxidase are listed in Table IV, along with the value obtained from the EPR titration. These calculations are based on Equation 8 (see "Appendix"). The close agreement in these values indicates that the proposed explanation for the Mn'+ versus pyruvate kinetics is probably valid. The fact that no Mg"-pyruvate complex could be detected by EPR spectroscopy is consistent with the kinetic behavior observed between Mg'+ and pyruvate.
The EPR experiments also provided values for the dissociation constants of M$+-TPP and Mn"-TPP. Consideration of these values and the various concentrations of Mg", Mn'+, and TPP used in the steady state experiments indicates that the concentration of the metal-TPP complex is usually well below that of the free metal or free TPP concer.tration.
The values for the metal-TPP dissociation constants also provide the information necessary to calculate the actual Michaelis constants for the metal-TPP ccmplexes. It is evident from Equation 4 (see "Appendix") that the slopes of the secondary plots in Figs. 1-3 are (K,,KM~L+-TPP)/V,,,. Since V,,;,, is known from the primary plot and the K,! values are known from the EPR experiments, KM,L+-TPP can be calculated. The results of these calculations are also shown in Table  IV. These calculations are based on data collected using Pipes buffer since phosphate buffer can bind Mg2+ (16). The calculated K , for the Mn"-TPP complex with unactivated pyruvate oxidase, 1.04 p~, is in excellent agreement with the measured value, 0.88 p~, obtained in the presence of a large excess of Mns+ ( Fig. 5A and Table IV), and the dissociation constant, 0.49 p~ ( Table 4). It is interesting to note from Table IV that the Michaelis constants for the metal-TPP complexes are apparently independent of the nature of the metal cation involved in the complex.
In summary, this work has demonstrated that the steady state kinetics of pyruvate oxidase with respect to the divalent cation can be easily understood in terms of a simple model in which the cofactor which binds to the enzyme is actually the metal-TPP complex. The major complicating feature in the kinetics analysis is the demonstrated complex formation between Mn2+ and pyruvate. These conclusions have been supported by binding data using fluorescence and EPR spectroscopy in which it is demonstrated that only the metal-TPP complex binds strongly to the enzyme. The dissociation constant (0.49 p~) and Michaelis constant (0.88 pM j for the Mn"-'TPP complex are in excellent agreement.

Relevant Steady State Kinetics Equations
The observed kinetic behavior is easily explained by the proposal that only the Me'+-TPP complex binds to the enzyme and functions as a cofactor. The initial velocity, u, can be expressed simply as: will become more complex but will remain symmetrical with respect to the total concentrations of Me2+ and TPP. EPR evidence indicates that Mn2+ can form a complex with pyruvate. Assuming 1) that pyruvate and Mn2+-TPP combine with the oxidase in a random fashion and 2) that a Mn"pyruvate complex that forms does not bind to the enzyme. The initial velocity is given by: yields a curve whose slope is given by: Taking the first derivative with respect to [PI yields: Setting this equal to zero, and solving for KO: The value of K,, is obtained from the intercept secondary plot, and, thus, the value of KO can be obtained if the value of [PI is known where the change in slope of the slope secondary plot equals zero, i.e. at the minimum of the concave-up secondary plot (Fig. 4 ) . Values obtained using this calculation are included in Table IV.