Reactivity and inhibitor potential of hydroxycitrate isomers with citrate synthase, citrate lyase, and ATP citrate lyase.

The four isomers of hydroxycitrate have been tested as substrates and inhibitors for citrate synthase, citrate lyase, and ATP citrate lyase. None of the isomers served as a substrate for citrate synthase and they were moderate to weak inhibitors of this reaction. Of the four isomers, only (pncit)-(2S)-2-hydroxycitrate did not serve as a substrate for citrate lyase while (pncit)-(4S)-4-hydroxycitrate was the only isomer which did not serve as a substrate for ATP citrate lyase. No consistent pattern of reactivity or inhibitor potency was seen with the different isomeric hydroxycitrates. It is proposed that more than one mode of binding is possible between the isomers and the three different active sites.

The four isomers of hydroxycitrate have been tested as substrates and inhibitors for citrate synthase, citrate lyase, and ATP citrate lyase. None of the isomers served as a substrate for citrate synthase and they were moderate to weak inhibitors of this reaction. Of the four isomers, only (pn,J-(2S)-2-hydroxycitrate did not serve as a substrate for citrate lyase while (pneit)-(4S)-4-hydroxycitrate was the only isomer which did not serve as a substrate for ATP citrate lyase. No consistent pattern of reactivity or inhibitor potency was seen with the different isomeric hydroxycitrates. It is proposed that more than one mode of binding is possible between the isomers and the three different active sites.
There are three citrate enzymes (1, 2) which catalyze the same bond-making and -breaking reaction which involves the equilibrium of citrate with oxalacetate and an acetyl moiety. The citrate re-synthase, analogous to enzyme 1, is excluded from this list. For enzymes 1, 2, and 3, when citrate synthesis occurs, the methyl carbon atom of acetate attacks the si-face of oxalacetate, and an inversion of configuration occurs as one hydrogen atom of the methyl group of acetate is replaced by oxalacetate (3-7). Citrate re-synthase has the opposite stereochemistry.
The equilibrium in the reaction catalyzed by citrate synthase favors citrate formation, while the equilibria for citrate lyase (a bacterial enzyme) and ATP citrate lyase (a cytosolic enzyme found in eukaryotic cells) favor the reverse reaction,  (22). The crystalline pig heart citrate synthase was purchased from Boehringer Mannheim and assaved bv the method-of Srere et al. (23). ATP-citrate lyase was purified fro& rat liver as described earlier (24) and the assay has also been described (25    by preincubation of the enzyme with Mgz+ or Zn*+ and various citrate analogs was greater with the K. aerogenes enzyme than the S. diacetilactis enzyme. Zn2+ caused a slower rate of apparent inactivation than did Mg*+ in the presence of all citrate analogs except (2S)-OHcit-@n,J.
The most potent inactivator, of all tricarboxylic acids examined with both enzymes, was (2S)-OHcit-(pn,,J. (4S)-OHcit-@n,,,) produced the slowest rate of inactivation. Citrate lyase contains an essential acetyl group, the loss of which may occur during the course of the reaction it normally catalyzes. This loss results in inactivation of the enzyme (10, 11). The nature of this inactivation with various citrate analogs was investigated.
[WAcetyl citrate lyase was incubated with citrate, tricarballylate, or the hydroxycitrate isomers and the mixture then placed over a Sephadex G-25 column (Fig. 4) isomers caused a complete deacetylation of the enzyme as evidenced by the absence of any radioactivity in the fractions (Peak A) containing the enzyme. Tricarballylate produced a partial deacetylation of the enzyme.
These data suggested that the inactivation produced by these tricarboxylic acids is due to deacetylation of the enzyme. This was confirmed by incubating the inactivated citrate lyase from K. aerogenes (produced by treatment with (2S)-OHcit-(pn,,J) with purified acetate:SH-(acyl carrier protein) enzyme ligase (AMP) in the presence of acetate and ATP. A recovery of 38% citrate lyase activity was obtained after a 40min incubation. On further incubation for a total of 6 h, 82% of the initial activity was recovered.
Effect of Hydroxycitrates on ATP Citrate Lyase -(4S)-OHcit-@n,,J is a very potent linear competitive inhibitor of ATP citrate lyase from rat liver, as already reported (15). (4R l-OHcit-@n,,,), the other stereoisomer with a hydroxyl group substituted on carbon 4, was a less effective inhibitor, but a more potent inhibitor than the two isomers with hydroxyl groups substituted on carbon 2. These and other kinetic constants are summarized in Table I I1  II,  I  I  I  I  II  III  I  I  I  012  456  8  IO  20  012  456  8  IO  citrate synthase should be noted.
In the case of ATP citrate lyase, (4S)-OHcit-don,,,) did not serve as a substrate for this enzyme, while the other three isomers exhibited substrate activity. All four, however, caused release of phosphate from the phosphoenzyme produced with ATP. None of the four isomers were found to be a substrate for citrate synthase under three different conditions of our assay ((29) and Table IV). DISCUSSION No simple pattern of activity or inhibition of the three citrate enzymes by hydroxycitrates, as listed in Tables I and  IV, emerged from these studies. Therefore, we found it necessary to consider both the detailed mechanisms of each enzyme, as understood to date, and the various possible modes of binding of the hydroxycitrates to enzyme. In the hydroxycitrates there are four groups available for chelation to a metal ion, but, in an unstrained situation, only three of these four groups will take part in this chelation. The four groups under discussion are the central carboxyl group (on C(3)), the central hydroxyl group (on C(3)), the terminal hydroxyl group (on C(2) or C(4)), and the terminal carboxyl group (on C(1) or C(5)).
Englard and Siegel reported that n(-)-tartrate and mesotartrates are substrates of malate dehydrogenase (30), presumably giving rise to hydroxyoxalacetate and thus this compound is probably substrate for the malate dehydrogenase reaction. (4S)-OHcit- (pn,.,,) and (4R )-OHcit-(pn,,J give rise to oxalacetate and glyoxalate so that if reaction occurs their assay with malate dehydrogenase presents no problems.   Synthase -All four isomers of hydroxycitrate bind, although weakly, to citrate synthase, but none are cleaved by the action of this enzyme. The binding for (4R)-OHcit-@XL,,,) is similar to that of citrate, while the binding of the other three isomers, particularly the binding of (2R )-OHcit-@n,J, is much weaker. (4R )-OHcit-@n,J is an analog (OH replacing F) of the isomer of fluorocitrate that is a substrate of citrate synthase (13). The question then arises, why does no reaction occur with the hydroxycitrate when it can occur with the analogous fluorocitrate.
The conformations of (4R)-OHcit-@n,J and th e quivalent isomer of fluorocitrate are essen-e tially identical in the crystalline state (13), and therefore it is not surprising that both bind equivalently well. One possible explanation for the lack of substrate activity for the hydroxycitrates, in spite of binding, is that the presence of the second hydroxy group might favor y-lactone formation since this is a very stable form of the ion, but a form that cannot be made by fluorocitrate.
In fact the hydroxycitrates found in plants are generally isolated as y-lactones (17). Therefore the anhydride, postulated as an intermediate in the citrate synthase reaction (31), is less likely to form, i.e. lactone formation (5carboxyl to 2-hydroxyl group or l-car-boxy1 to 4-hydroxyl group) might take precedence over anhydride formation (5-carboxyl to 3-carboxyl). If this were the explanation it would lead us to predict that aminocitrates would also not be substrates of citrate synthase since they might also preferentially form -y-lactams. Oxalacetate is thought to bring citrate synthase into a good conformation for reaction (32). The two isomers of hydroxycitrate that are unsubstituted in the oxalacetate-derived (-2-@n,,J) end of citrate (i.e. (4S)- OHcit-@n,,,) and (4R)-OHcit-(pn.,,,)) bind best. Thus conformation changes in the oxalacetate-derived end of the parent citrate molecule are less well tolerated than those in the acetate-derived end.
Citrate Lyase -In the first part of the citrate lyase reaction, citrate, on binding, is exchanged with an acetyl group bound to a prosthetic group to give citryl enzyme and acetate (33)(34)(35). Cleavage of the citryl enzyme gives oxalacetate and regenerates acetyl enzyme. This cleavage probably involves initial ionization of the central hydroxy group of citrate (36)(37)(38), and this ionization is facilitated by metals. Indeed, divalent metal ions (magnesium) are necessary for the activity of this enzyme (and also for ATP citrate lyase). I OAC

\O
(2) (zs)-OHcit-(pncit) Citrate lyase contains an acetyl group bound to a sulfhydryl group of a CoA-like bound prosthetic group, so that a built-in thiol ester group (analogous to acetyl-CoA in citrate synthase) is present. Any mechanism which results in loss of the acetyl group from the enzymes leads to inactivation (10). Normally citrate binds and is cleaved, leaving an acetyl group bound to the enzyme and liberating oxalacetate. The acetyl enzyme is then available for cleavage of another citrate ion.
However, if the binding is nonproductive, the acetyl group is displaced on binding but is not replaced when the ligand leaves the active site. This nonproductive binding apparently occurs with (2S)-OHcit-(pn,,J, as described under "Results." The other three hydroxycitrates bind and react, and, as a result, leave acetate or glycolate bound to the enzyme. However, in these reactions some inactivation also occurs. The reaction is fastest with (2R)-OHcit-@~r,~J which gives an acetyl group on cleavage. (4S)-OHcit- (prz,,,) and (4R )-OHcit-@n,,,) undergo reaction at a rate which is about 9 to 23% of the rate with (2R )-OHcit-(pn,i,).
The product of this cleavage is presumably glycolate rather than acetate. However, the question arises, of the four isomers tested why does (2R)-OHcit- (pn,,,) undergo the fastest reaction in the presence of citrate lyase while (2S)-OHcit+n,,J (which also gives acetate on cleavage) undergoes no reaction at all. The answer probably lies in the fact that (ZSI-OHcit-@n,,,) can bind in more than one way to the enzyme and possibly binds predominantly in the nonproductive mode. Presumably, hydroxycitrates are bound in a "citrate-like" manner when they undergo cleavage as shown diagrammatically in Fig. 5. However, an alternate mode of binding, involving the terminal hydroxyl and carboxyl groups and the central carboxyl group, which is found in crystalline rubidium fluorocitrate (13), may be nonproductive and hence result in inhibition. We assume that the magnesium binding site involves the pro-R end of citrate (the oxalacetate-derived end) since chelation to the thioester portion of the acetyl-derived end is less likely. As a result we have included the (pn,,,)-1-carboxyl group but not the (pn,&i-carboxyl group in the metal chela- bind in this fluorocitrate-like manner and therefore competition with the citrate-like manner of binding does not occur.

ATP Citrate
Lyase-The ATP citrate lyase is believed to " " Y ,I " LIL, catalyze its reaction by way of the following steps. Initially,

Hydroxycitrates and Citrate Lyases
ATP interacts with the enzyme, in the presence of magnesium ions, to give phosphorylated enzyme (39) and ADP. Citrate is then believed to react with the phosphoenzyme to give an enzyme. citryl . phosphate complex (40). It is not clear whether magnesium is necessary at this second stage or not. Phosphate is released and the citryl enzyme then interacts with CoA, without the intervention of any metal ion, to give enzymebound citryl-CoA which is cleaved to give oxalacetate and acetyl-CoA (41).
When citrate reacts with phosphoenzyme, phosphate is liberated. The release of phosphate may not be the result of a direct displacement when the citrate binds since it is believed that an enzyme. citryl . phosphate complex can exist as an intermediate in the reaction.
The loss of phosphate is also caused by other tricarboxylic acids, including nL-isocitrate, tricarballylate, and cis-and trans-aconitate, but not by dicarboxylic acids (39). Since none of the four acids listed above contain central hydroxyl groups, the reaction (i.e. cleavage) cannot occur and therefore an inhibition occurs. The four hydroxycitrates bind well and all undergo a reaction except (4S)-OHcit-@~z,~J which is found to be an extremely potent binder (see Table  I). No reports of other analogs, e.g. isocitrate, binding as tightly, have been made. We suggest that, again, alternate modes of binding of (4S)-OHcit+n,J to the enzyme occur but that, in addition, interaction of the group not involved in binding to the normal binding sites can occur in some additional way with another active site group, perhaps the phosphate binding site.
Conclusion -We suggest that hydroxycitrates, with an extra hydroxyl group available for chelation, can bind in the same manner as citrate. However, alternate modes of binding are also possible, some of which compete with or replace the citrate-like binding with the result that no cleavage can occur, and inhibition occurs.
Acknowledgments -The skilled technical assistance of Carolyn Stewart is gratefully acknowledged. Drs. R. Guthrie and R. Kierstead of the Roche Chemical Division kindly provided the trisodium salts of the hydroxycitrate isomers.