Purification of and mechanism studies on citrate synthase. Use of biospecific adsorption-elution techniques.

Citrate synthases from animal tissues were found to bind to Sepharose-"ATP." A pure preparation of citrate synthase was obtained from a crude fraction of rat heart by the specific elution of the enzyme from the Sepharose-"ATP" with the dead end complex-forming substrates, oxalacetate and CoA. The proposed mechanisms of citrate synthase, obtained from steady state kinetics, were examined in light of the elution pattern of the enzyme obtained using combinations of substrates and substrate analogs.

but it can be specifically eluted with its dead end complexforming compounds, CoA and oxalacetate. Neither of the substrates of the citrate synthase reaction alone nor the other putative dead end complex formers, citrate and acetyl-CoA, could elute the enzyme from the column.
We also have used the results of these studies and the elution of citrate synthase with substrate analogs in an attempt to reconcile the different mechanisms proposed for the enzyme from steady state kinetic data. MATERIALS AND METHODS

RESULTS
Purification of Citmte Synthuse from Rat Heart: Homogenization-Frozen rat hearts (26 g), washed free of blood in cold 0.9% NaCl, were homogenized in 5 volumes of 20% alcohol and 0.4 M KCl. Homogenization was performed in a Waring Blendor for 10 ?&min periods at full speed, cooling the mixture between homogenization periods in a salt-ice bath. The homogenate was centrifuged at 20,000 x g for 40 min at 4', and the precipitate was discarded. The supernatant fluid was dialyzed overnight against 5 liters of cold 5 mM KPO, buffer, pH 7.4, with one change of buffer.
Ammonium Sulfate Fractionation-The dialyzed supernatant fluid was centrifuged, and the precipitate was discarded. The supematant fluid was brought to 50% saturation of ammonium sulfate with the solid salt (29.1 g/100 ml). The precipitate was removed by centrifugation, and the supernatant solution was brought to 75% saturation of ammonium sulfate (15.9 g/100 ml), stirred for 1 hour, and centrifuged again as described above. The precipitate was dissolved in a small quantity of 5 mM KPO, buffer, pH 7.4, and dialyzed against 2 liters of the same buffer at 4" for 24 hours with three changes of buffer.
Sepharose-"ATP" Column Chromatography-The supernatant solution (after centrifugation to remove a small amount of precipitate) was applied to a Sepharose-"ATP" column previously equilibrated with 5 mM KPO, buffer, pH 7.4. The column was washed with the same buffer until no more protein was eluted. Solutions containing 200 PM potassium citrate, 200 MM acetyl-CoA, 100~~ potassium citrate + 100 HIM acetyl-CoA, 200 PM oxalacetate, and 200 PM CoA, each in 5 mM KPO, buffer, pH 7.4, were successively applied to the column, but no enzyme was eluted. The enzyme was eluted with a mixture of 100 PM CoA and 100 pM oxalacetate in 5 mM KPO, buffer, pH 7.4. When no further enzyme activity appeared in the eluate, the column was washed with 200 mM KCl, which resulted in the elution of protein that contained no citrate synthase activity. The elution profile of the enzyme from the column is shown in Fig. 1. Fractions having the highest activities were pooled, and the protein was precipitated with (NH&SO,. The precipitate was collected by centrifugation and dissolved in 20 mM KPO, buffer, pH 7.4. The results of the purification procedure are summarized in Table I. Fig. 2, a and b, shows the disc gel electrophoresis patterns of the proteins in the original ammonium sulfate fraction applied to the column and in the final preparation. The eluate obtained with the CoA and oxalacetate wash from the Sepharose-"ATP" column shows a single protein band. Sodium dodecyl sulfate gel electrophoresis (Fig. 2c) of the enzyme also shows a single protein band. Comparison with standard proteins indicated a subunit M, of 40,000 to 50,000, in agreement with earlier results (3).
The column could be regenerated for use in another purification procedure by washing it with 200 mM KC1 followed by equilibration with 5 mM KPO, buffer, pH 7.4.
El&ion Behavior of Citrate Synthase-In additional experiments the order and concentration of eluants used was varied. The enzyme can be eluted immediately with a mixture of CoA and oxalacetate without prior washes with the individual substrates, that is to say the behavior of the enzyme on the Sepharose-"ATP" is not affected by previous noneluting conditions. If lower concentrations of either CoA or oxalacetate are used, then a slow elution of citrate synthase occurs yielding rather broad peaks of activity. Instead of complete elution in 2 or 3 tubes, 20% of the activity is found in a total of 10 or more tubes. CoA levels as high as 1 mM failed to displace the enzyme from the column. A mixture of acetyl-CoA (50 pM) and oxalacetate (50 PM) cannot elute the enzyme from the column.
El&ion with Oxalacetate Analogs-If the oxalacetate is replaced with a-ketoglutarate in the elution mixture with CoA, no enzyme is eluted. However, a slow elution is obtained with 200 WM (S)-malate and 100 fiM CoA, and a rapid elution with 200 HIM (R)-malate and 100 PM CoA (Fig. 3). Similarly (data not shown) when the elutions were carried out separately with  results in a much more rapid elution of the enzyme than does the latter. Acetyl-CoA in combination with either (R)-or (S)-malate does not elute the enzyme from the column.
Other Citrate Synthases-Neither the crude nor the pure preparation of Escherichia coli citrate synthase binds to Sepharose-"ATP." This is expected, since E. coli citrate synthase is only poorly inhibited by ATP (12). On the other hand, citrate synthases from human heart and yeast were bound by the column and could be completely purified by this method, whereas the rat liver enzyme behaved identically and could be partially purified.
Pig heart citrate synthase also binds to Sepharose-"ATP" and can be eluted slowly with the single substrates CoA or acetyl-CoA or rapidly with the combination of CoA and oxalacetate. Other Immobilized Nucleotides-Sepharose-"AMP" prepared in a manner similar to that described for Sepharose-"ATP" binds citrate synthase, but with this adsorbent the enzyme can be eluted with a single substrate. In addition, we have prepared Sepharose-AMP and Sepharose-ATP according to the method of Mosbach et al. (1) (the AMP and ATP derivatives were kindly supplied by Dr. Mosbach). In this case the nucleotide is linked to the column through its Ns position, and a hexyl side chain is linked to the Sepharose. The Sepharose-ATP prepared in this way behaved precisely like the Sepharose-"ATP" in the purification of human heart citrate synthase. The Sepharose-AMP (prepared according to Mosbath et al. (1)) binds citrate synthase, but elution can be accomplished with single substrates, as it is on the Sepharose-"AMP" column. DISCUSSION There are a number of satisfactory methods for the purification of citrate synthases from a variety of sources (3, 13). None of these methods matches the ease, rapidity, and high recovery of the procedure described in this paper. Pure enzyme from several eukaryotic cells was obtained using Sepharose-"ATP" as an adsorbent and the combination of oxalacetate and CoA as a specific eluant mixture. Although many proteins bind to the Sepharose-"ATP" column, only citrate synthase forms a complex with both CoA and oxalacetate, so that it is the only protein eluted from the column by this mixture. This confirms the usefulness of the technique of elution by "dead end" complex formation (14, 15), which has been used so successfully on lactate dehydrogenase.
The binding of a protein to an immobilized ligand is undoubtedly a more complex process than simple ligandprotein interaction (16). After the initial binding step there is probably a secondary interaction between the hydrophobic spacer arm and the protein, and perhaps even a tertiary interaction between the matrix and the bound protein. It is difficult to determine the extent of the latter two interactions. If, in a control experiment, unmodified Sepharose or Sepharose with the spacer arm is used, the interactions with the protein cannot be the same as those that occur when the immobilized ligand is present. When the protein is bound through the ligand, then regions of the protein are brought in close proximity with the spacer arm and matrix so that the secondary and tertiary interactions may occur. In the absence of the initial interaction between active site on the protein and ligand, the probability of secondary and tertiary bindings is small. If secondary and tertiary interactions occur to an important extent, then the following discussion would have to be modified.
We can consider two different explanations for the observed elution patterns. First, it is possible that the enzyme is bound to the ligand column not only at its adenine nucleotide (acetyl-CoA or CoA) site but also at its oxalacetate site. It would be necessary therefore to have both the substrates present to elute the enzyme from the column. On the other hand, the enzyme may be bound to ATP through just one binding site (the CoA site), and the binding of oxalacetate to its own site on the enzyme may cause a conformation change which enhances the binding of CoA to the enzyme without a concomitant increase in ATP binding.
The results of experiments on the elution of rat heart and pig heart citrate synthase from Sepharose-"ATP" are summarized in Table II. It is difficult to interpret all of these results with a single hypothetical mechanism. The difference between the pig heart and rat heart enzyme, i.e. the ability of the pig heart enzyme to be eluted with single substrates, may be due to a difference in the dissociation constants for the substrates and the two enzymes. Yet the kinetic behavior and kinetic constants for the two enyzmes as measured in the steady state are It is true, however, that oxalacetate facilitated the ability of CoA to elute the pig heart enzyme making the difference between the two enzymes less marked. This result -is not in agreement with Johansson and Pettersson's finding that the affinity of pig heart citrate synthase for acetyl-CoA and propionyl-CoA, but not for CoA, increases about 20-fold on the binding of oxalacetate to the enzyme. Such a cosubstrate-induced binding of acyl-CoA was attributed to binding of the acyl moiety of the CoA derivative (17). It should also be remembered that human heart and yeast citrate synthases behave identically with the enzyme from rat tissue.
The inability of the mixture of acetyl-CoA and oxalacetate to elute the enzyme may be due to the rapid enzymic conversion of these substrates to citrate and CoA, but no such reasoning can be used to explain the ineffectiveness of mixture of acetyl-CoA and either (R)-or (S)-malate.
It was surprising that the acetyl-CoA and CoA which have dissociation constants of 106 and 30 PM with the rat heart enzyme, respectively, could not elute the enzyme even at a concentration of 200 PM and 1 mM respectively. One might assume therefore that either the effective concentration of ATP available to the nucleotide site on the enzyme is quite high, or the "off' constant for Sepharose-"ATP" enzyme is small. In terms of a possible conformation change, our results would indicate that the change in conformation of citrate synthase induced by (R)-malate for the fit of CoA is better than that induced by (S)-malate.
This agrees with the results of Weidman et al. (18) who reported that (R)-malate (but not (S)-malate) could form a ternary complex with citrate synthase and a spin label analog of acetyl-CoA.
The other putative dead end complex formed with acetyl-CoA and citrate does not effect elution of the enzyme. This may be due to the fact that citrate and acetyl-CoA have overlapping binding domains on the enzyme surface, and thus no change in the binding of acetyl-CoA is possible. An ordered mechanism for the citrate synthase reaction has been considered a number of times, since it is known that oxalacetate causes an apparent conformation change in the pig heart enzyme (19; 20) and that (S)-malate, an analog for oxalacetate, but not (R)-malate, induces an acetyl-CoA-enolase activity in the enzyme (21). We have shown also that oxalacetate and (S)-malate (but not (R)-malate) can protect citrate synthase against urea denaturation (20), again in contrast to the results of Weidman et al. (18) (see above) and our elution data.
The steady state kinetics of citrate synthase has been studied thoroughly only for the rat tissue and pig heart enzymes. Our kinetic analysis of the rat enzyme as well as partial kinetic analyses on other citrate synthases indicated that the mechanism may be a random one in both the forward and reverse direction, with two "dead end" complexes (E*AcCoA*Cit and E=CoA.OAA) formed in a kinetically significant manner (13). The kinetic data for both the pig heart and rat citrate synthases showed that oxalacetate did not affect the K, of acetyl-CoA, and that acetyl-CoA did not affect the K, of oxalacetate.
Such data indicate that K, for the substrates should equal their Kdlss. Whereas the K, for oxalacetate (5 PM The data presented here are consistent with the idea that oxalacetate does effect the binding of the acetyl-CoA. However, this situation is analogous to the situation for yeast hexokinase, which Cleland describes as a random mechanism involving several rate-limiting ternary complex conformation changes during its course (24).
The observation that E. coli citrate synthase does not bind to the Sepharose-"ATP" column, compared to pig heart, rat heart, rat liver, human heart, and yeast citrate synthases, all of which bind to the column, reflect the known difference between the poor ability of ATP to inhibit the E. coli enzyme compared to its ability to inhibit the eukaryotic enzymes. Further, the fact that the eukaryotic citrate synthases can be eluted from Sepharose-AMP columns with a single substrate is probably related to the observation that AMP is a poor inhibitor of citrate synthase. Since identical results were obtained with Sepharose-"ATP" where the linkage is through the ribose portion of the ATP molecule, and with Sepharose-ATP (according to Mosbach et al. (1)) where the linkage is through the N-6 group, it seems unlikely that strong interaction occurs through those parts of the ATP molecule and the enzyme.
Although our present results do not yield an unequivocal mechanism, it appears that affinity chromatography can serve both as a useful tool for the purification of the enzymes and as an interesting method for the study of the reaction mechanisms (25). In the present case the results for the rat citrate synthase are consistent with a mechanism in which the binding of oxalacetate increases the ability of the enzyme to bind acetyl-CoA.