Studies on the Mechanism and Stereochemical Properties of the Oxalacetate Decarboxylase Activity of Pyruvate Kinase *

When cod fish muscle oxalacetate decarboxylase catalyzes the decarboxylation of oxalacetate in the presence of NaBH,, L-lactate results from the reduction of enzyme-bound pyruvate. However, D-lactate results when borohydride reduces the binary enzyme .pyruvate complex formed by adding pyruvate from solution, as reported by others. This observation suggests that there are alternate mechanisms for reduction that are either kinetically or sterically determined for the E. pyruvate forms produced in the two directions. In the process of investigating the mechanism of reduction, the cod fish muscle decarboxylase was discovered to be identical with pyruvate kinase. Decarboxylase activity appears to take place at a site which overlaps the phosphoenolpyruvate binding site on this enzyme, as discussed in the following paper. Crystalline rabbit muscle pyruvate kinase also contains significant decarboxylase activity indicating that the two reactions may be structurally related functions. In the presence of K+, orthophosphate, or ATP the rabbit muscle enzyme catalyzes the detritiation of enzyme-bound pyruvate formed during decarboxylation before release of pyruvate from the enzyme, in analogy with the detritiation of pyruvate formed from P[3-3H]enolpyruvate in the kinase reaction. This observation is consistent with the formation of an enolpyruvate intermediate common to the kinetic pathways of both reactions. Since the decarboxylase reaction is completely stereospecific, within the limits of detection, going with retention of configuration, the protonation of the enolpyruvate intermediate is completely determined by the enzyme as is the case with the enolpyruvate intermediate generated from P-enolpyruvate in the kinase reaction.

When cod fish muscle oxalacetate decarboxylase catalyzes the decarboxylation of oxalacetate in the presence of NaBH,, L-lactate results from the reduction of enzyme-bound pyruvate. However, D-lactate results when borohydride reduces the binary enzyme .pyruvate complex formed by adding pyruvate from solution, as reported by others. This observation suggests that there are alternate mechanisms for reduction that are either kinetically or sterically determined for the E. pyruvate forms produced in the two directions. In the process of investigating the mechanism of reduction, the cod fish muscle decarboxylase was discovered to be identical with pyruvate kinase. Decarboxylase activity appears to take place at a site which overlaps the phosphoenolpyruvate binding site on this enzyme, as discussed in the following paper. Crystalline rabbit muscle pyruvate kinase also contains significant decarboxylase activity indicating that the two reactions may be structurally related functions. In the presence of K+, orthophosphate, or ATP the rabbit muscle enzyme catalyzes the detritiation of enzyme-bound pyruvate formed during decarboxylation before release of pyruvate from the enzyme, in analogy with the detritiation of pyruvate formed from P- [3-3H]enolpyruvate in the kinase reaction. This observation is consistent with the formation of an enolpyruvate intermediate common to the kinetic pathways of both reactions. Since the decarboxylase reaction is completely stereospecific, within the limits of detection, going with retention of configuration, the protonation of the enolpyruvate intermediate is completely determined by the enzyme as is the case with the enolpyruvate intermediate generated from P-enolpyruvate in the kinase reaction.
While the investigation of oxalacetate-metabolizing enzymes is of great importance in understanding cellular metabolism, only recently has attention been given to enzymes that irreversibly decarboxylate oxalacetate in animal tissues.
Oxalacetate + H++pyruvate + CO, Schmitt et al. reported the purification from cod fish muscle of an oxalacetate decarboxylase which has an absolute divalent metal ion requirement and is insensitive to avidin (1). They also observed decarboxylase activity in rat skeletal muscle, although no attempt was made to isolate the enzyme responsible in this case. Kosicki and Westheimer were unable to show inactivation of the cod muscle enzyme with NaBH, either in the absence or presence of substrates (2). On the other hand, the requirement for a divalent metal both in the decarboxylation reaction and in the enolization of pyruvate (3) led to the proposal of metal ion acting to polarize the carbonyl group for activation at C-3. Further evidence for this was the requirement for divalent metal for a new reaction of the enzyme, the reduction of pyruvate by NaBH, which was stereoselective to form o-lactate (2). * This work was supported by United States Public Health Service Grants GM-20940, m-05539, and CA-06927 and by an appropriation from the Commonwealth of Pennsylvania.
$ Recipient of National Institutes of Health Fellowship AM-54188.
The present study grew out of an attempt to determine if the cod muscle decarboxylase reaction occurred with retention of configuration as found for other decarboxylases. The subsequent discovery that the decarboxylase was, in fact, identical with pyruvate kinase and that both the decarboxylase and kinase reactions take place at overlapping sites on the enzyme, as described in the following paper (4), prompted the investigation of the mechanistic and stereochemical relationship between these two activities. aqueous solution, having a pH of 4 to 6, was adjusted to pH 7, diluted to 10 ml with distilled water, and run onto a Dowex l-Clcolumn.
HCl (2 mM) was used to elute lactic acid. Lactate and Malate Assays-Total lactic acid was determined by the calorimetric procedure of Barker (8). D-and L-lactate were determined by the method of Holzer and Soling (9) from the increase in optical density at 366 nm (6 = 9,l x lo3 at pH 9.5) due to reduction of acetylpyridine NAD + (0.  of P-[3-aH]enolpyruvate, 12.4% of the radioactivity was found in the first peak. The second peak shown in Fig. 1  first peak was all L-lactate, a portion of it was oxidized to pyruvate in the presence of the acetylpyridine analog of DPN+ and L-specific lactate dehydrogenase and chromatographed on Dowex l-Cl-. As shown in Fig. 1 (bottom plate) essentially all of the counts were eluted with 40 mM HCl where pyruvate is found. Any unoxidized n-lactate should have appeared in the 2 mM HCl region. No contaminating n-specific lactate dehydrogenase activity was found in the L-specific lactate dehydrogenase used in this experiment.
In an important control experiment, [2-'Y: lpyruvate was reduced by NaBH, in the absence of enzyme. As shown in Fig.  2 (top plate) a single peak was eluted with 2 mM HCl. This was not a substrate for L-lactate dehydrogenase (Fig. 2). When a portion of this peak was acidified, treated with methanol, and rechromatographed it migrated as free lactate acid and assayed as racemic lactate (Fig. 2, bottom plate).
To be certain that L-lactate production reflected the action of the decarboxylase and not reduction of pyruvate on P-enolpyruvate carboxylase, this experiment was repeated by infusing NaBH, (0.60 pmol/min) and oxalacetate 0.0376 hmol/min) in separate streams into a solution containing 0.94 unit of the decarboxylase from cod muscle. When a total of 1.11 pmol of oxalacetate had been infused, the reaction mixture was acidified and processed in the same way as described above. The first peak eluted from the Dowex l-Cl-column was pure L-lactate (-0.10 pmol) and the second was the borate ester of racemic lactate (-0.28 rmol). The results of this experiment are in accord with those found previously where oxalacetate was enzymically generated in the reaction mixture instead of being added directly and indicate that optically pure L-lactate is produced by the decarboxylase.
Borohydride Trapping of Enzyme .Pyruvate Complex Formed by Addition of Pyruvate from Solution-Based on optical rotation Kosicki and Westheimer (2) concluded that an excess of n-lactate was formed when free pyruvate was reduced in the presence of the cod decarboxylase. Their experiment was repeated using the methods described here for isolating lactic acid. As shown in Fig. 3, two radioactive peaks were eluted from the Dowex l-Cl-column, one of which migrates as unesterified lactic acid (A)  lactic acid (B). Using L specific and D specific lactate dehydrogenase, peak A was found to be greater than 90% n-lactate while peak B analyzed as the hydrolyzable borate ester of racemic lactate. The yield of n-lactate from both peaks was 72 + 4% which compares well with the 25% excess n-lactate originally reported by Kosicki and Westheimer (2). In retrospect it is evident that in their isolation of lactic acid these authors used conditions that would have led to the hydrolysis of any borate esters present and hence could not have discriminated between lactate resulting from the reduction of pyruvate on the enzyme surface and that occurring free in solution.
The singular result requiring further exploration is that BH,stereospecifically reduces the re face of enzyme-bound pyruvate formed from the decarboxylation of oxalacetate while the opposite (or si) face is stereospecifically reduced in the enzyme .pyruvate complex formed by direct addition of pyruvate to the enzyme from solution. Assuming that the two stereospecific reduction reactions are catalyzed by the same enzyme species in the apparently homogeneous enzyme preparation, three general hypotheses may potentially explain the above results. First, pyruvate generated on the enzyme from decarboxylation may be at a site distinct from the site to which pyruvate, added from the product side, binds. Differences between these topographically distinct complexes result in the formation of different enantiomers of lactate after BH,reduction. Second, alternate enzyme-pyruvate forms at the same site may be reduced by BH,-depending on which direction in the kinetic sequence pyruvate is generated at the active site (Equation 3). Differences in the conformation of these complexes would expose different faces of the carbonyl bond of pyruvate to reduction by BH,-. A third hypothesis assumes that BH,-binds to a site on the enzyme which is identical with or overlaps with the presumed p-carboxyl binding site for oxalacetate. Borohydride binds to this site Effect of BF,-on Decarboxylase Activity-As a test of the presence of a common binding site for BH,-and oxalacetate on the decarboxylase, the effect of NaBF,, a presumed analog of NaBH,, on the K, of oxalacetate in the decarboxylase assay was examined and found to be a linear competitive inhibitor with a K, of 18 mM (Fig. 4). As a comparison, 100 mM F-or NO,-gave less than 20% inhibition when oxalacetate was at its K, concentration (0.25 mM). This corresponds to a K, of >200 mh4.
Evidence That Decarboxylase and Pyruuate Kinase Activity Are due to Single Active Site-The failure of NaBF, to inhibit the enolization of pyruvate would also result if the enolization were due to another unrelated enzyme present in the decarboxylase preparation.
Since rabbit muscle pyruvate kinase is known to catalyze this reaction (13), P-enolpyruvate-ADP phosphoryltransfer activity in the cod muscle enzy*me was assayed for by coupling to lactate dehydrogenase. Unexpectedly, the kinase activity exceeded by 230-fold the oxalacetate decarboxylase rate in this apparently homogeneous preparation of decarboxylase. The dilemma of the inconsistent reduction stereochemistries seemed, briefly, to be explained by the fraction procedures, rates of inactivation of the two activities, and modifier effects is presented in the following paper (4). In addition, Tsai and Mildvan' have established that the MnZ+ activator constants for both the decarboxylase and pyruvate kinase activities are similar for the cod muscle enzyme preparation.
Therefore, it became of particular interest, because of the opposite actions of NaBH,, to ask whether the kinase and decarboxylase activities occur at the same site on the enzyme.
The conclusion that both activities occur at the same or overlapping sites is supported by the findings of the following paper that (a) P-enolpyruvate is a strong linear competitive inhibitor with respect to oxalacetate, K, = 3.25 CM, in the cod muscle decarboxylase reaction, (b) 4-ethyloxalacetate, previously shown to be a strong inhibitor of cod muscle decarboxylase, 1 has approximately the same K, value, w 11 PM, as a linear enzyme. Evidence supporting this conclusion based on protein * C. S. Tsai and A. S. Mildvan, unpublished results.
competitive inhibitor against oxalacetate and P-enolpyruvate in the decarboxylase and pyruvate kinase reactions, respectively, and (c) both P-enolpyruvate and oxalacetate exhibit cooperative binding kinetics in their respective reactions and are almost equally responsive to fructose 1,6-bisphosphate, an allosteric activator which shifts the Hill constant toward one (4). Finally, as evidence that decarboxylase and kinase may be structurally linked activities, crystalline rabbit muscle pyruvate kinase was observed to have significant decarboxylase activity (0.5 unit/mg) that was completely inhibited by P-enolpyruvate.
Effect of K+, P,, and ATPon Enolization-The enolization of pyruvate by pyruvate kinase is known to require K' and ATP (or P,) as well as MgZ+ (or Mn*+) (13). Thus the enolization of pyruvate formed during oxalacetate decarboxylation (Table I) may be expected if the same catalytic site is involved since KPO, was present at 10 mM. Using rabbit muscle pyruvate kinase it is observed that neither K+ nor ATP (or P,) is required for decarboxylase activity, nor do they significantly influence the oxalacetate to pyruvate rate. On the other hand it could be shown (Table II) that both K+ and ATP were required for the loss of tritium from [3-SH]oxalacetate to occur. Exchange was previously seen (Table I) with K+ and P, and the cod enzyme. It follows that the same factors required for detritiation of pyruvate cause detritiation of the pyruvate generated on the enzyme from oxalacetate. However, since neither K+ nor P, (or ATP) is required for the formation of pyruvate from oxalacetate, it would seem that their role is not a mechanistic one but a kinetic one, Robinson and Rose have suggested that the effect of high K+ in causing excessive detritiation of P-[3-3H]enolpyruvate during the P-enolpyruvate-ADP phosphotransferase reaction is to slow the departure of the enzyme-bound pyruvate (11). A similar explanation would satisfy the observations with [3-SH]oxalacetate.
Here ATP (or Pi) is also needed whereas with P-[3-aH Jenolpyruvate the ATP is generated on the enzyme. The absence of effects of K+ and ATP on decarboxylation rate suggests that neither the ketonization nor product release step is rate-determining for decarboxylation. Although the similar effect of K+ in causing the excessive enolizations from pyruvate derived from either oxalacetate or  Table I more than 60% of the tritium of oxalacetate was lost in decarboxylation by the cod enzyme. Since detritiation at the prochiral carbon of oxalacetate, if it occurred, would probably be stereospecific, it is unlikely that the difference would be due to enolization of oxalacetate. The finding that in the presence of NaBH, no tritium is lost (Fig. 1) is further evidence that the detritiation occurs at the pyruvate level only.
Stereochemistry of Decarborylation-Oxalacetate, stereospecifically labeled with deuterium and tritium at C-3, was decarboxylated with the rabbit muscle enzyme after rigorously excluding K+, NH,+, and Pi from the reaction mixture. Under these conditions it was not necessary to trap the pyruvate produced. Fig. 6 shows the scheme for establishing whether decarboxylation involves retention or inversion of configuration using (3R) [3-*H,sH]oxalacetate, generated from (3R) [3-*H,SH]malate, as a substrate for the decarboxylase. The pyruvate product was isolated and converted to malate in a coupled system using pyruvate carboxylase and malate dehydrogenase. The carboxylase reaction is known to go by retention of configuration and exhibits an intramolecular isotope effect where k,,/k, = 3.3 (7). The L-malate product was then treated with fumarase which labilizes only the pro-R hydrogen isotope during the formation of fumarate. If the decarboxylation reaction involves retention of configuration and (3R) [3-*H,*H]oxalacetate is the substrate, the fumarase step will result in a majority of the tritium being lost to solvent. On the other hand, if inversion is involved, most of the tritium will be retained in fumarate.
The results of Table III indicate retention of configuration. Using (3R) [3-2H,sH]malate as a source of the substrate, 84% of the counts of the recovered malate were lost to solvent in the fumarase step, close to the expected discrimination on the basis of the reported intramolecular isotope effect in the pyruvate carboxylase step (7). Retention of configuration is confirmed using (3s) [3-2H,3H]malate as the starting material since most of the label is found in the pro-S position of malate in the fumarase step. The poorer discrimination observed with this enantiomer likely reflects the greater decrease in specific activity going from the starting malate to the malate species used in the fumarase reaction. Contaminating NH,+ and Pi in the incubation mixture during the decarboxylation reaction could account for this decrease.
It has been proposed that the enzymes of pyruvate metabolism are related in evolution as suggested by similar stereochemistries in all reactions involving P-enolpyruvate (2-si face addition) and in proton replacement on pyruvate (retention). Retention has also been observed with malic enzyme (malate dehydrogenase decarboxylating) and with pyruvate carboxylase.
Chiral Pyruuate-The establishment of retention of configuration for the decarboxylation reaction of rabbit muscle pyruvate kinase provides a direct method for producing chirally labeled [SH,PH]pyruvate, that is more convenient, using only commercially available enzymes, than that previously described (7). For example, deuterium can be introduced into the pro-R position of malate using the fumarase equilibrium in I),0 (6). The malate can be isolated on Dowex Cland used as a starting material in the malate+oxalacetate-pyruvate interconversion (in 3H,0) described in the first part of the legend to Table III of this paper. The resulting pyruvate will be of the S configuration.
Pyruvate of the R configuration would be formed if tritium were introduced by fumarase and deuterium in the decarboxylase step. Precautions must be taken to exclude K+ and orthophosphate from the pyruvate kinase incubation to prevent randomization of the label on the pyruvate. The resulting pyruvate should be of very high chiral purity.

CONCLUDING REMARKS
The oxalacetate decarboxylase activity of pyruvate kinase exhibits stereospecificity as would be expected of an enzyme-catalyzed process (Table III). Thus, each step in the reaction, which includes C-C bond cleavage and subsequent protonation of the enolpyruvate intermediate, is completely determined by the enzyme. From earlier studies on the stereochemistry of the pyruvate kinase reaction (7), proton addition to enolpyruvate derived from P-enolpyruvate was shown to come from the 2-si face. If it were certain that the enolpyruvate formed from oxalacetate had the same orientation on the enzyme as that from P-enolpyruvate, it could be concluded that both CObdeparture and proton addition occur from the 2-si face as well. However, this may be an unsafe assumption. First, K+ and ATP or P, are not required for, nor do they significantly effect, the decarboxylase reaction while these compounds are required in the catalyzed enolization of pyruvate (13). Second, the loss of tritium to solvent from [3-3H]oxalacetate during decarboxylation is much greater than the loss to solvent from p-[3-3H]enolpyruvate during the kinase reaction. The ratio of rate constants for enolization and pyruvateoff steps (k, and k, of Equation 2) should be the same if the enzyme-pyruvate complex formed from oxalacetate and Penolpyruvate were the same. Finally, the alternate borohydride reduction stereochemistries may be a consequence of an additional E.pyruvate intermediate in the decarboxylase reaction not normally produced in the kinase reaction (Equation 3).
Attempts to catch enzyme-bound pyruvate generated from P-enolpyruvate with NaBH, has not yet been successful. Earlier results of Robinson and Rose (11) ( Table III) showed that in the P-enolpyruvate-pyruvate reaction most of the enzyme was present in the enzyme-pyruvate form. The greater accessibility to BH,-of the oxalacetate-derived enzyme pyruvate complex may be related to its greater exchange rate with water, referred to already.
All of the differences cited might result from an interaction between the active site and oxalacetate not like that interaction found in the enzyme-bound intermediates of the pyruvate kinase fraction. This interaction may be reflected in the low K, (0.2 mM) of oxalacetate and K, of 4-ethyloxalacetate  for the cod muscle enzyme compared to the K,,, of -3 mM for pyruvate inhibition of decarboxylase activity reported by Schmitt et al. (1). It might be that a strong interaction with the C-4 carboxyl group puts the enzyme into a conformation that allows different interactions with the medium and active site to occur throughout the catalytic process.