Pyruvate Carboxylase from Chicken Liver

On the basis of initial velocity and product inhibition studies a nonclassical Ping Pong Bi Bi Uni Uni mechanism has been proposed for pyruvate carboxylase from chicken liver. The nonclassical feature of this mechanism is the proposal that each active site on the enzyme is composed of two separate and functionally distinct catalytic sites, i.e., a separate catalytic site exists for the reactants of each partial reaction. The two catalytic sites are presumably linked by the biotinyl residue which functions as the carboxyl carrier. The Bi Bi partial reaction, in which MgATP2-, MgADP-, HCOs-, and phosphate are the substrates, is proposed to utilize a rapid equilibrium random Bi Bi mechanism which includes the formation of two abortive complexes, E-HCOs-Pi and E-HCO,-MgADP-. A rate equation has been developed for the proposed mechanism by employing a combination of rapid equilibrium and steady state methodology. The proposed mechanism is directly analogous to the modsed Ping-Pong Bi Bi mechanism previously described for the biotin-enzyme methylmalonyl-CoA transcarboxylase (D. B. Northrop (1969) J. Biol. Chem., 244, 5808).

oxalacetate + MgADP + P; This enzyme is visualized as a tetrameric molecule when examined in the electron microscope after negative staining (1) and contains biotin and tightly bound manganese in equimolar ratio at a level approximating 4 moles per mole of enzyme (2,3). Besides the substrates listed in Reaction 1, optimal enzymatic ac-* This investigation was supported by Grants AM 11712 and AM 12245 from the National Institutes of Health and bv Atomic Energy Commission Contract AT-(U-1)-1242.
t Postdoctoral fellow of the National Institutes of Health.
1968 to 1970. § Present address, Department of Biochemistry, Temple University, Philadelphia, Pa. 19140. tivity requires the presence of Mg* in excess of that required for formation of the metal-nucleotide complex (4) and the presence of a univalent cation such as K+ or Tris+.l In addition, the enzyme exhibits an absolute requirement for activation by an acyl-CoA (e.g. acetyl-CoA) (5). A minimal reaction mechanism for pyruvate carboxylase from chicken liver has been proposed (Reactions 2 and 3) on the basis of isotope exchange studies and of the isolation of the carboxylated biotin-enzyme intermediate (6).
E-biotin + methylmalonyl-CoA e E-biotin-CO2 + propionyl-Coil (4) E-biotin-CO2 + pyruvate Fr? E-biotin + oxalacetate (5) It will be noted that pyruvate carboxylase and methylmalonyl-CoA transcarboxylase share a common partial reaction (Reactions 3 and 5). Although the initial velocity studies of the transcarboxylase reaction indicated conformity to the expected Ping Pong Bi Bi mechanism (8), anomalous results were obtained during studies of (a) the properties of the inhibition of the over-all reaction by products and dead end inhibitors (8) and (b) the specificity of inhibition of Reactions 4 and 5 (as measured by isotope exchange techniques) by various substrate analogs (9). These data were interpreted to suggest the operation of a modified Ping Pong Bi Bi mechanism in which separate sites exist on the enzyme for catalysis of the two partial reactions (Reactions 4 and 5). Northrop and Wood (9) suggested that the two separate sites are linked by a mobile biotinyl residue, which in this and other biotin-enzymes is mounted on a long flexible side chain (10).
('omprehensive initial velocity and product inhibition studies arc reported here for t.he acyl-Cob-dependent pyruvate carboxylase purified from chicken liver which are consistent with the proposal that catalysis by this enzyme is described by a nonclassical Ping Pong Bi Bi Uni Uni mechanism. Similar conclusions have been reached from recent studies of pyruvate carboxylase from rat liver by McClure et al. (11,12) although the evidence is somewhat different in nature.
Steady state kinetic studies of the acyl-Co&independent pyruvate carboxylase from Aspergillus niger (13) have also been reported.
i1 preliminary report of some of the data has appeared (14).
HEPEST was obtained from Calbiochem, coenzyme A from P-L Biochemicals, acetic anhydride and pyruvic acid from Eastman, and 5,5'-dithiobis- (2-nitrobeneoic acid) from Aldrich. Other chemicals used in these studies were reagent grade.

Methods
The commercial preparations of malate dehydrogenase and lactate dehydrogenase were equilibrated with 0.01 M Tris-Cl, pH 7.8, on a Sephadex G-25 column, 20 x 1 cm. This procedure is necessary in order to remove SOt2-, which is a potent inhibitor of pyruvate carboxylase from chicken liver. 3 Acetyl-CoA was prepared from coenzyme A and acetic anhydride by the method of Simon and Shcmin (15). The thiolester content of acetyl-CoA preparations was determined with citrate synthase.
Pyruvate concentrations were determined with lactate dehydrogenase (17) and oxalacetate concentrations with malate dehydrogenase (1X). Protein was determined by the spectrophotometric method of Warburg and Christian (19). Concentrations of adenine nucleotide solutions were determined spectrophotometrically, assuming a molar absorptivity of 15,400 (20). The concentrations of free and complexed nucleotides were calculated by assuming that the apparent stability constants of MgATP" and MgADP were 70,000 and 4,000 31-1, respectively, at pH 7.8 (21). In assays of oxalacetate decarboxylation the free Mgz+ was maintained at 5 m&I and I\IgADP:ADP at 20. At these levels of iVg*+, phosphate concentrations were limited to 50 mM, or below, to avoid the formation of precipitates.
Pyruvate carboxylase was purified through Stage 5 and stored as described by Scrutton et al. (22), except that mosb prepara-.tions were further purified by chromatography on a DEAE-Sephadex A-50 co1umn.l Specific activities (determined as described in Reference 22) are reported as micromoles of product formed at 25" per min per mg of protein.
Pyruvate carboxylase preparations used in these studies varied in specific activity from .15 t,o 30 i.u. per mg.
The initial rate of CO2 fixation by pyruvate carboxylase was measured by estimating the rate of oxalacetate production in the presence of malate dehydrogenase and NADH (5). When oxalacetate was added as a product inhibitor the initial rate of the reaction was estimated from the increase in absorbance at 290 nm due to oxalacetate production. Where appropriate the observed rate was corrected for breakdown of added oxalacet,ate by measuring the change in absorbance in the absence of acetyl-CoA. When HC08-was the variable substrate all solutions were freshly prepared from water which had been boiled for 30 min and then cooled and stored under Nz. The actual concentration of HCO, in the cuvette was taken as the sum of the concentration of added HCO,-plus the measured endogenous HCOSconcentration.
The endogenous HCO;-concentration (always 20.4 mM) in the assay mix was calculated with the equation A = (0) K,/V, where K, and V are apparent constants estimated graphically with velocities at high concentrations of added HC03-and 0 is the velocity measured in the absence of added HCO,.
The initial rate of oxalacetate decarboxylation by pyruvate carboxylase was measured by estimating the rate of pyruvate production in the presence of lactate dehydrogenase and NADH (2). The rates were corrected for the decarboxylation of oxalacetate which was not due to catalysis of the over-all reaction by measuring the rate of pyruvate production in a system lacking acetyl-CoA.
When pyruvate was added as product inhibitor the initial rate of oxalacetate decarboxylation was determined by measuring ATP production in the presence of hexokinase, glucose, glucose 6-phosphate dehydrogenase, and NADP (2).
In all cases specific reaction conditions for each experiment are given in the appropriate figure legend. After equilibration of the assay systems to 25", the reactions were initiated by addition either of pyruvate carboxylase (CO2 fixation) or of pyruvate carboxylase + Mg2+ (oxalacetate decarboxylation).
The change in absorbance at the appropriate wave length (290 or 340 nm) was recorded continuously with Gilford spectrophotometers, model 2000 or 240, each of which was equipped with a Radiometer constant temperature apparatus maintained at 25". Full scale deflections in the range 0.1 to 1.0 absorbance unit were used. The chart speed was adjusted to give slopes of approximately 45". Data Analysis-The statistical analysis of the kinetic data n-as based on the procedures developed by Cleland (23). Reciprocal velocities were plotted graphically against the reciprocals of substrate concentrations.
When these plots were linear, the data were fitted to Equation 6. vs v=K, Least squares fits were performed on an IBM 1620 digital com-puter4 using FORTRAN computer programs obtained from W. W. Cleland of the University of Wisconsin.
These programs provide values for the constants in the fitted equation, standard errors of their estimates, and weighting factors for further analysis. The form of the over-all rate equation was determined from secondary plots in which slopes (K,/V) and intercepts (l/V) obtained from Equation 6 were plotted against tither the inhibitor concentration or the reciprocal of the changing fixed  Precision of the duplicate assays was within 3% with the coupled assay of COZ fixation and within 5 y0 for the measurements at 290 nm and for the coupled assays of oxalacetate decarboxylation.
Observed initial velocities which deviated markedly from the indicated pattern were discarded.
The lines drawn through the points in double reciprocal plots are calculated from a least square fit of the data (Equation 6) unless otherwise stated in the figure legend.
The slopes and intercepts of these lines then provided the data points in the secondary plots.
However, the lines shown in the secondary plots were calculated from a fit of the original experimental data to the appropriate over-all rate equation.

RlGXJLTS
AND DISCUSSION Dejinition of Enzyme species under Study-An acyl-CoA, e.g. acetyl-CoA, is an essential cofactor for pyruvate carboxylase from chicken lircr.
To simplify the interpretation of the present studies, all experiments were conducted at an essentially saturating concentration of acetyl-Co24 (50 PM).
shown in Fig. 1, this is more than twice the concentration of activator required to obtain the optimum maximum velocity with each of the substrates tested. Furthermore, the apparent K, values of the substrates arc independent of the acetyl-CoA concentration when the acetyl-CoX concentration exceeds 20 ~JM.~ Since a saturating acetyl-CoX concentration was employed, in the present studies the catalytic properties observed are those of the enzyme-modifier complex in which all kinetically significant modifier sites are occupied.
In addition to the requirement for activation by an acyl-CoA, expression of maximal catalytic activity by pyruvate carboxylase from chicken liver also requires the presence of a monovalent cation, e.g. K+, Tris+.l The studies reported here were performed in the presence of optimal concentrations of the cation activator.
Thus in assay systems buffered with Tris+ the concentration of l'ris cation was approximately 85 mM and small variations in the K+ concentration in this system did not effect the initial reaction rate. When K+-HEPES was employed as a buffer for the CO2 fixation assay system the K+ concentrat.ion was maintained at 100 to 125 rnM by addition of KCI. This conclusion is based on experiments showing that isotope exchange between ATP and 3'Pi is dependent on the presence of MgADP (6). Pyruvate and oxalacetate are involved in a separate partial reaction (Reaction 3) since the exchange of [14C]pyruvate with oxalacetate occurs in the absence of the other reaction components (6). Thus, the over-all reaction catalyzed by pyruvate carboxylase appears to be the sum of two partial reactions, one with a sequential Bi Bi kinetic mechanism and the other with a Uni Uni mechanism.
Cleland has defined procedures for predicting the initial velocity patterns which will be observed if the experimental dnta are consistent with the mechanism under consideration (25). For the kinetic mechanism proposed for pyruvate carboxylase on the basis of isotope exchange studies, i.e. Ping Pong Bi Bi Uni Uni, double reciprocal plots of the data should consist of a family of parallel lines when the varied pair of substrates is composed of one substrate from each partial reaction.
If both of the varied substrates are involved in the sequential Bi Bi partial reaction, the initial velocity pattern will be a family of lines intersecting to the left of the vertical axis.
The initial velocity pattern for each possible pair of substrates in both the direction of COs fixation and the direction of oxalacetate decarboxylation has been determined and the results are summarized in the first three columns of Table I  posed Bi Bi partial reaction (MgATP" and HC03 in the direction of CO2 fixation, MgADP and Pi in the direction of oxalacetate decarboxylation) the predicted pattern of intersecting lines is observed.
Additionally, the predicted pattern of parallel lines is observed for each experiment in which the varied pair of substrates is composed of one substrate from the proposed Bi Bi partial reaction and one substrate from the proposed Uni Uni partial reaction (pyruvate or oxalacetate). In these experiments velocities were measured in duplicate for a minimum of five different concentrations of one varied substrate and four dif-ferent. concentrations of the second varied substrate.
Preliminary analysis of the experimental data to determine whether the pattern was intersecting or parallel was performed as described under "Methods." Of the 12 patterns summarized by the data in Table I  The data for each experiment were fitted to the appropriate rate equation (cf. "Methods") to obtain the apparent K, and Ki for the various substrates.
These constants are summarized in Table I. According to Cleland, if the standard errors of the fitted constants are 10% or less of the fitted constants themselves, one can be quite confident that the proper rate equation was employed (23). For the data in Table I  These constants therefore reflect a good internal consistency between the rate equation employed and the experimental data tested.
The results of the iostope exchange experiments reported previously (6) and the initial velocity studies presented here are entirely consistent with the proposed Ping Pong Bi Bi Uni Uni mechanism depicted in Equations 2 and 3. Thus, the kinetically significant events in the reaction catalyzed by pyruvate carboxyl-6 Six different substrate pair combinations a*nd 16 different product-substrate combinations were examined during the course of these studies.
Only representative examples of the primary data are presented in this paper because presenting all of the primary data would produce an unnecessarily bulky paper. Also, in the opinion of the authors most of the primary data is of minimal interest to many readers. Therefore, except for data shown in the paper, the primary data from which the constants in Tables  I and III were (12) the kinetic behavior of a Ping Pong Bi Bi Uni Uni mechanism in the absence of products.
The initial velocity patterns shown in Table I are equally consistent with two fundamentally different types of Ping Pong Bi Bi Uni Uni mechanisms.
The distinction between the two types of mechanisms depends on the number of separate and independent catalytic sites which are assumed to be operative within each active site. Classical mechanisms assume that the active site functions as a single catalytic site, e.g. the glutamate-oxalacetate transaminase mechanism (29). An active site which functions as a single catalytic site in the pyruvate carboxylase reaction is represented schematically in Fig. 4A. On the other hand, nonclassical mechanisms assume that two separate and functionally independent catalytic sites are present within each active site; or more specifically, a separate catalytic site exists for the reactants of each partial reaction.
A schematic representation of an active site of this type is shown for pyruvate carboxylase in Fig. 4B. For the type of active site illustrated in Fig.  4B it is assumed that the biotinyl residue, which functions as a carboxyl carrier, can link the two separate catalytic sites and thus provide the element which unifies the active site.
The classical concept of the active site serves as a basis for the nomenclature and methodology developed by Cleland (7), and has been utilized successfully in describing the kinetic behavior of a large number of enzymes.
The nonclassical concept of the  4. A, a schematic illustration of an active site which functions as a single catalytic site. It is not assumed that the cr-keto acids and the nucleotides are bound by an identical set of amino acids, only that there is an "overlap" between the a-keto acid and nucleotide sites suflicient to prevent the binding of a-keto acids to the enzyme-nucleotide complex, and vice versa. B, a schematic illustration of an active site which is composed of separate catalytic sites for the reactants of each partial reaction.
The two catalytic sites are assumed to be physically distinct to the extent that there is no "overlap", i.e. occupancy of one site does not prevent occupancy of the other. Yet these sites must be sufficiently close to each other so that biotin can function as a carboxyl carrier between them. This scheme is adopted from Fig. 5

II.
If the Bi I3i partial reaction is assumed to have a rapid equilibrium random mechanism in the classical ping-pong reaction the predicted patterns are those shown in Column 4, Table II. With a nonclassical mechanism (Fig. 5B), the predicted patterns are those shown in Column 5. It is apparent that the inhibition patterns predicted for the nonclassical mechanism differ for each possible combination of J-sried substrate and product inhibition from those predicted for the two versions of the classical mechanism. The latter two show a fair degree of correspondence.
The classical and nonclassical mechanisms are most readily distinguished by examination of certain key pairs of product inhibitor and varied substrate which should lead to competitive inhibition patterns.
For the classical mechanisms ( Fig. 5A) a competitive interaction is predicted between pyruvate and MgADP-and also between oxalacetate and MgATP2 because in each case the varied subst,rat,e and the product inhibitor combine at the same site with the same form of the enzyme.
Thus the classical mechanism predicts competitive interactions only when the varied substrate and the product inhibitor are involved in different partial reactions.
On the other hand the nonclassical mechanism (Fig. 5B) predicts competitive interactions only between substrates and products which are in the same partial reaction, e.g. between pyruvate and oxalacetnte. The kinetic behavior predicted for the nonclassical mechanism is a direct consequence of the asrumption that a separate catalytic site exists for the reactants of each partial reaction.
The patterns observed with the various product inhibitors are shown in Table II If we examine the results of tests of the key pairs of varied substrate and product inhibitor mentioned above (Ta,ble II, Column 7) a-e find no instance in which the competitive inhibition pat.terns predicted by the classical mechanisms are observed.
Thus when MgXDP-is the product inhibitor and pyruvatc is the variable substrate, uncompetitive rather than competitive patterns are observed.
Similarly oxalacetate is a noncompetitive product inhibitor when MgATP2-is the varied substrate while MgATPz-is an uncompetitive inhibitor (Fig. 6) and II('O3 a noncompetitive inhibitor (Fig. 7) of osalacetate decarbosylation when oxalacetate is the varied substrate.
These result':: suggest strongly that the reaction catalyzed by pyruvate carbosylase from chicken liver is not described by a classical Ping Pong observed when the product inhibitor and variable substrate are involved in separate partial reactions.
Identification of the reaction mechanism as nonclassical is further suggested by the observation of competitive product inhibitions when the product inhibitor and the varied substrate are involved in the same partial reaction.
Thus MgATP2-is a competitive product inhibitor of osalacetate decarboxylation when either MgADP or phosphate is the varied substrate, and MgADP and phosphate are competitive inhibitors of CO2 fixation when MgATP* is the variable substrate (Fig. 8). As indicated in the previous section, MgATP", MgADP, and phosphate are all involved iu the Bi Bi partial reaction. Pyruvate and oxalacetate, which nre involved in the Uni Uni partial reaction, also show competitive interactions (Fig. 9, Table III The assignment of a specific nonclassical mechanism to pyruvate carboxylase from chicken liver requires identification of the sequential mechanism which characterizes the Bi Bi partial reaction. The competitive patterns observed when MgATP2interacts kinetically with MgADP-or phosphate argues that the binding of reactants at the Bi Bi site can be described by a rapid equilibrium random type of mechanism.6 However, a simple 6 These inhibition patterns do not distinguish between a rapid equilibrium random Bi Bi mechanism and a simple random Bi Bi mechanism but the former has been assumed here because the procedures which have been employed to derive the rate equation Each cuvette contained the following components in a total volume of 1.0 ml (in micromoles) : K+ HEPES, 100; K+ pyruvate, 3; KHCOa, 16; free Mg2+, 4; phosphate and MgATP2+, as indicated; malate dehydrogenase (20 rg) and pyruvate carboxylase (specific activity = 4.2 pg). Other conditions as listed for Fig. 2. Inset, secondary plot of slopes versus phosphate concentration. Each cuvette contained the following components in a total volume of 1.0 ml (in micromoles) : K+ HEPES, pH 7.8, 200; MgATP+, 2; free Mg2+, 3; NaHC03, 10; acetyl-CoA, 0.05; and pyruvate carboxylase (specific activity = 18.5, 60 pg). Velocity is change in absorbance per min at 290 nm. Inset, secondary plot of slopes versus the concentrations of oxalacetate. rapid equilibrium random Bi Bi mechanism is not completely adequate, since such a mechanism requires that HC03-and MgiTPz-show the same kinetic behavior (cf. Table II, Column 5).-However, HC03 shows noncompetitive interactions with (cf. next section) require the assumption that the binding events be in a state of rapid equilibrium. It might be noted that Gulbinsky and Cleland (personal communication) have shown that most random mechanisms will appear to be of the rapid equilibrium type when analyzed by steady state kinetic methods. 2; free Mg 2+, 5; bicarbonate and phosphate, as indicated; and lactate dehydrogenase (17 pg) and pyruvate carboxylase (specific activity = 22, 43 pg). Other conditions were as listed for Fig. 2. Inset, secondary plot of slopes and intercepts versus HC03 concentration.

MgADP
and phosphate (Fig. 10, Table II, Column 7), not the competitive interactions exhibited by MgATP*. The kinetic behavior of HCOa-therefore suggests that abortive enzyme-reactant complexes can form in which bicarbonate and a product of the Bi Bi partial reaction are simultaneously bound to the enzyme.
If the formation of two abortive complexes, namely, E-HC03-Pi and E-HC03-MgADP, is combined with a rapid equilibrium random Bi Bi mechanism (Fig. llA), the rate equation derived for the over-all reaction (see next section) predicts the product inhibition patterns listed in Column 5 of Table  II. The inclusion of the above abortive complexes causes changes in those patterns which are underlined in Table II and the predicted patterns are now identical with the experimental data, except for a single case. (An uncompetitive pattern is predicted when HC03 is a product inhibitor of oxalacetate decarboxylation and oxalacetate is the varied substrate, but a noncompetitive pattern is observed. This discrepancy will be discussed in a later section.) The two proposed abortive complexes are those which might be expected to form at a site which binds MgATP" and HC03-in the productive sequence. Analogous abortive complexes are commonly encountered in kinetic studies of other enzymes which have random mechanisms (3% The major features of the reaction mechanism proposed for pyruvate carboxylase from chicken liver are therefore the following: (a) the over-all reaction is the sum of two partial reactions and can be described as Ping Pong Bi Bi Uni Uni; (b) separate catalytic sites exist for the reactants of each partial reaction; (c) the two catalytic sites are connected by a biotinyl residue which isomerizes back and forth, carrying the carboxyl group from one site to the other; (~2) the mechanism of the Bi Bi partial reaction is rapid equilibrium random Bi Bi with the additional existence of two kinetically significant abortive complexes (EmHCOS--Pi and E-HC03-MgADP). The procedure (33) (see text).

Derivation of Rate Equation for
Proposed Two-Site Mechanism-the reactants involved in the complex, e.g. FAB. A separate set The procedure employed by Northrop (8) was used to derive a of fractional concentration factors is required for each catalytic rate equation for the mechanism proposed for pyruvate carbox-site, since the two sites behave independently.
The scheme in ylase. The derivation of rate equations for such complex re- Fig. 11A must be analyzed in order to obtain the necessary action mechanisms has been greatly facilitated by the develop-fractional concentration factors for the Bi Bi half-reaction, FAB ment of a procedure which combines features of the steady state and FpQ. The expressions obtained for these factors are shown and rapid equilibrium methods (32). hpplication of this method in Equations 13 and 14.
In the derivation four species arc aesumed to contribute to the steady state distribution of the enzyme (Fig. 11B). Of these species m and m represent. enzyme with free and carbosyl-:ited biotin, respectively, at t.he site for the Bi Bi partial reaction, aand m represent t'he enzyme with the free and carbosylated biotin at the site for the ITni I-ni partial reaction.
Each of these species represents the sum of free enzyme plus, all enzyme-rcactant compleses which have the appropriate form of biotin at th? indicated binding site.
For the ol-keto acid catalytic site the relevant fractional concentration factors are obtained by analysis of the following scheme : This complete rate equation can be modified for initial velocity studies of COZ fixation by setting those terms containing the concentrations of the products P, Q, and R to zero. The reciprocal form of the resulting equation is identical with Equation 12. Hence Equation 18 predicts the observed initial velocity patterns. In order to test whether Equation 18 predicts the product inhibition patterns which are observed (cf . Table III), those terms containing the two products which are not present in the experiment are set equal to zero. Such an analysis has been conducted for each of the reactants as product inhibitor and the predicted patterns coincide with the experimental data, except for the case in which HCOj-is the product inhibitor and oxalacetate is the varied substrate (see Table II, Columns 5 and 6). A rationale for the observed inconsistency is presented below.
On the basis of the results of these studies and of the studies with methylmalonyl-Coh transcarboxylase by Northrop (8), prediction of the product inhibition patterns expected for a twosite mechanism appears to conform to the following principles. (a) Competitive patterns are observed only when the product inhibitor and the varied substrate are members of the same partial reaction.
(b) Noncompetitive and uncompetitive patterns are observed when the product inhibitor and the varied substrate are members of different partial reactions; the pattern is uncom-7 &o-inhibition constants (7)  The data presented in Table III permit a crude evaluation of the latter two of these relationships. The ratio VI: V2 was taken as 10 (2), and values for Kii,,Kiib,and Ki,(0.28,2.4,and 0.08 mM,respectively) were estimated from slope data for product inhibition experiments in which oxalacetate was the product inhibitor.
There is a reasonable agreement between K,, estimated from these data (Kep = 9 and 12) and the K,,, measured for the reaction at equilibrium (K,,, = 10.2 at pH 8.0), and total Mg2+ = 4.5 mM (34). However the close agreement between the values may be fortuitous since none of the kinetic constants used in evaluating the Haldane relationships have been extrapolated to infinite concentration of the fixed substrate or substrates.
pctitive if the product, inhibitor is a mrmbrr of a bireactant (01 higher reactancy) partial reaction, the pattern is IloncompetitiT-e if the product inhibitor is a member of a unireactant partial reaction.
(c) Substrate analogs which are true dead end inhibitors show competitive or uncompetitive inhibition; competitivcx inhibition versus the reactants to lvhich they are structurall\-rclated and uncompetitive inhibition uer~us the reactants of the other partial reaction (cf. References 8 and 11).
Concluding Discussion-The proposal of a two-site mt~chanism for pyruvate carbosylase from chicken liver is supportrd by the present kinetic studies and also by previous studies on the l)roperties of the interaction of substrates and inhibitors with this enzyme.
The two sites involved in catalysis of the partial reactions of pyruvate carboxylase appear to act independently of earh other. For instance rate constants for interconversion steps at one sit'e do not appear to be influenced by the presence or absence of a reactant at the other site. Such interaction would give rise to an apparent activation or inhibition which would be expressed as nonlinearity in the double reciprocal plots. No such deviations from linearity were observed in the studies reported here. The proposed independence in function of the two catalytic sit'es is further supported by previous observations. First, the properties of the interaction of the bound manganese with the substrates of the Uni Uni partial reaction (Equation 3) are unaffected by the presence of the substrates and cofactors of the Bi Bi partial reaction (Equation 2), which fail to show any significant interaction with the bound metal (35, 36). Second, oxalate acts as a specific inhibitor of the exchage of [14C]pyruvate with osalacetate (Reaction 3) and has no significant effect on the rate of exchange of [32P]phosphate with ATP (Reaction 2) (35).
The independence in function of the two catalytic sites implies a spatial separation between these sites. In this case the required intersite linkage may be provided by the biotinyl residue which has an essential role in catalysis at both sites (6). If this residue is attached in peptide linkage to the e-NH2 group of a lysine residue in apopyruvate carboxylase, as has been shown for other biotin carboxylases (I 0)) movement between catalytic sites as much as 28 A apart would be possible if the flexible arm is fully extended.
Such movement may be facilitated by the change in net charge on the biotinyl residue which accompanies carboxylation.
The simplifying assumption that the carboxylation of the biotinyl residues has no significant effect on the interaction of substrates and products with pyruvate carboxylase is less well documented.
Oxalate acts as an uncompetitive inhibitor of ('02 fixation when pyruvate is the varied substrate, but as a competitive inhibitor of oxalacetate decarboxylation when oxalacrtate is the varied substrate (35). These inhibition patterns, whirh indicate preferential binding of oxalate to the noncarboxylated form of the enzyme, show that this inhibitor does discriminate between the two forms of the enzyme.
Similar conclusions have previously been deduced from the properties of the oxalate inhibition of exchange of [Wlpyruvate whith oxalacetate catalyzed by methylmalonyl-CoA transcarboxylase (9). However, in the over-all reaction catalyzed by this latter enzyme a comples pattern of inhibit.ion is observed when pyruvate is the varied substrate (9). The properties of this inhibition have been interpreted as indicating that oxalate has the capacit.y to trap the biotinyl residue at the oc-keto acid site (9) predicts that HCOs-will be an uncompetitive inhibitor of oxalacetate decarboxylation when oxalacetate is the varied substrate, but instead a noncompetitive pattern is observed (Fig. 7, Table II), This apparent inconsistency can be explained if the rate of the reaction in the direction of oxalacetate decarboxylation is not limited solely by the interconversion of the central complex but also in part by release of MgATl' from the enzyme.
Under such circumstances HCO, could react with E-hIgATP and biotin (cf. Fig. 5B) to regenerate the carboxy biotin form. This will result in an inhibition of the reaction which cannot be overcome by increasing the concentration of oxalacetate, hence yielding a noncompetitive pattern.
A similar %MgATP complex has been reported for the rat liver enzyme by McClure et al. (12).
Direct evidence that an active site may be separated into two different catalytic sites has been obtained in structural studies of the related biotin enzyme acetyl-CoA carboxylase from Escherichia coli. This enzyme may be separated into three classes of subunits ail of which are required for catalysis of the over-all reaction (37). The biotinyl residues of this protein are carried on a small subunit, the biotin carrier protein (37). Another subunit (E,) catalyzes carboxylation of free biotin in the presence of ATP + HCO,-and must therefore carry a binding site for this cofactor (37). A binding site for biotin also exists on the third subunit (&) (38) which is required for transcarboxylation for l'-N-carboxybiotinyl enzyme to acetyl-CoA (39). These observations provide definitive proof that each active site of acetyl-CoA carboxylase from E. coli is composed of two separate and distinct catalytic sites. A similar separation of an enzyme into subunits responsible for catalysis of different portions of the over-all reaction has been shown for pyruvate dehydrogenase complexes obtained from both mammalian and microbial sources by Reed (cf. Reference 40). However a thorough steady state kinetic analysis has not been reported for either acetyl-CoA carboxylase from E. coli or pyruvate dehydrogenase.
Thus at the present time we do not have both structural and kinetic evidence for a two-site-type over-all reaction mechanism for any given enzyme.