Assessing the Reversibility of the Anaplerotic Reactions of the Propionyl-CoA Pathway in Heart and Liver*

While a number of studies underline the importance of anaplerotic pathways for hepatic biosynthetic functions and cardiac contractile activity, much re-mains to be learned about the sites and regulation of anaplerosis in these tissues. As part of a study on the regulation of anaplerosis from propionyl-CoA precursors in rat livers and hearts, we investigated the degree of reversibility of the reactions of the propionyl-CoA pathway. Label was introduced into the pathway via NaH 13 CO 3 , [U- 13 C 3 ]propionate, or [U- 13 C 3 ]lactate (cid:1) [U- 13 C 3 ]pyruvate, under various concentrations of pro- pionate. The mass isotopomer distributions of propio-nyl-CoA, methylmalonyl-CoA, and succinyl-CoA revealed that, in intact livers and hearts, (i) the propionyl-CoA carboxylase reaction is slightly reversible only at low propionyl-CoA flux, (ii) the methylmalonyl-CoA racemase reaction keeps the methylmalonyl-CoA enantiomers in isotopic equilibrium under all conditions tested, and (iii) the methylmalonyl-CoA mutase reaction is reversible, but its reversibility

Adequate energy production via the citric acid cycle (CAC) 1 requires not only a constant supply of acetyl-CoA, but also a fairly constant pool of the catalytic intermediates which carry the acetyl groups as they are oxidized. Although the process of anaplerosis in micro-organisms was discovered in the 1960s (1), its importance for the homeostasis of mammalian cells (particularly cardiomyocytes) was recognized only in the 1980s (2, 3) (for a recent review, see Ref. 4). While the crucial role of anaplerosis for hepatic gluconeogenesis is self-evident, investigations in the heart suggested that stimulating anaplerosis from exogenous substrates could become part of the treatment of myocardial reperfusion injury and other cardiomyopathies (5)(6)(7)(8)(9)(10).
In a recent clinical study (11), the hypoglycemia as well as the mechanical performance of the heart and muscle of patients suffering from long chain fatty acid oxidation defects was greatly improved by replacing the fat component of their therapeutic diet, i.e. trioctanoin (a medium-even-chain triglyceride) by triheptanoin (a medium-odd-chain triglyceride). The only difference between the metabolisms of octanoate and heptanoate is the production of propionyl-CoA from the latter. In a follow-up investigation in pig hearts, propionate infusion was found to be a very effective anaplerotic substrate. Anaplerosis from 0.25 mM [U- 13 C 3 ]propionate amounted to 9% of the rate of the CAC (12,13). This led us to investigate the regulation of the propionyl-CoA pathway in livers and hearts of animals.
Early in vitro work with purified enzymes revealed different degrees of reversibility of the reactions catalyzed by propionyl-CoA carboxylase (21, 22), methylmalonyl-CoA racemase (23), and methylmalonyl-CoA mutase (24). In the presence of a mixture of purified propionyl-CoA carboxylase, methylmalonyl-CoA racemase, and methylmalonyl-CoA mutase, only a few percent of the radioactivity of [ 14 C]succinyl-CoA (prepared from [1-14 C]succinate) was recovered in CO 2 released by the reversal of propionyl-CoA carboxylase (16). From these studies on purified enzymes, one cannot draw conclusions on the degree of reversibility of the reactions of the propionyl-CoA pathway in intact organs.
To the best of our review of the literature, there have been only few attempts at testing the reversibility of the propionyl-CoA pathway in intact mammalian tissues, in the absence of biotin or cobalamine deficiency (25). A characterization of the reversibility and of the degree of thermodynamic equilibrium of the pathway would allow one to predict the efficiency of anaplerotic interventions with propionyl-CoA precursors. We took advantage of recent developments in mass isotopomer analysis 2 (26) to evaluate the reversibility of propionyl-CoA carboxylase and methylmalonyl-CoA mutase in intact rat livers and hearts.

In Vivo Pig Experiments
Four anesthetized pigs were thoracotomized and fitted with a pumpcontrolled bypass between the femoral artery and a branch of the left anterior descending coronary artery (10). NaH 13 CO 3 was infused into the bypass at a rate calculated to achieve a 15% enrichment of HCO 3 Ϫ / CO 2 in the infused myocardial territory. The enrichment of HCO 3 Ϫ /CO 2 was measured in samples of venous blood from the infused territory just before taking punch biopsies of the myocardium at 10 and 60 min (two pigs at each time point).

Sample Preparation
Frozen tissue (0.2-1 g) is extracted with 4 volumes of ice-cold 6% perchloric acid using a Polytron homogenizer. After centrifugation, the acid extract is mixed with 3 volumes of 75 mM potassium phosphate buffer pH 4. The solution is transferred to a glass syringe from which it is slowly pushed through an oligonucleotide purification cartridge (ABI Masterpiece, Applied Biosystems) as described by Deutsch et al. (30) for the isolation of long chain acyl-CoAs from tissue extracts. The cartridge is then rinsed with 5 ml of water. The short chain acyl-CoAs are eluted from the cartridge with 1 ml of 50% acetonitrile in water. The acetonitrile/water fraction is evaporated in a Savant vacuum centrifuge at room temperature. The residue is dissolved in 1 ml of water which is divided into two aliquots for the assays of the mass isotopomer distributions of (i) propionyl-CoA via propionylsarcosine pentafluorobenzyl ester and (ii) methylmalonyl-CoA 3 and succinyl-CoA via the tert-butyldimethylsilyl derivatives of the corresponding acids (27,28). The measured mass isotopomer distributions were corrected for the contributions of naturally occurring heavy isotopes (31). In the experiments with NaH 13 CO 3 tracer, the enrichment of HCO 3 Ϫ /CO 2 in perfusate or blood was measured by gas chromatography-mass spectrometry (32).

Calculations
The fractional contribution of succinyl-CoA to methylmalonyl-CoA via the reversal of the methylmalonyl-CoA mutase reaction (FC Succ-CoA 3 MMA-CoA) can be calculated from the M2 enrichments of methylmalonyl-CoA and succinyl-CoA labeled from [U- 13  , which allows calculating the fractional contribution of methylmalonyl-CoA to propionyl-CoA via the reversal of the propionyl-CoA carboxylase reaction.

Data Presentation
Data are expressed as means of the parameters calculated from triplicate gas chromatography-mass spectrometry injections in the analysis of a given tissue or effluent perfusate sample for a given perfusion. We present data from about 50 organ perfusion experiments. For each of the experimental conditions chosen, we ran all perfusions with one parameter (time or concentration of one substrate) being allowed to vary.

RESULTS
We tested the reversibility of the reactions of the propionyl-CoA pathway in perfused rat livers and hearts using three types of isotopic tracers, i.e. NaH 13 CO 3 , [U-13 C 3 ]propionate, or [U-13 C 3 ]lactate ϩ [U-13 C 3 ]pyruvate, in the presence of increasing concentrations of propionate that are within physiological/ therapeutic range.
Perfusions with NaH 13 CO 3 -First, we followed the time course of the labeling of liver methylmalonyl-CoA and succinyl-CoA from bicarbonate. Fig. 2A shows that the labeling of methylmalonyl-CoA plateaued at about one-third that of bicarbonate. The labeling of succinyl-CoA was only about 1 ⁄10 of that of methylmalonyl-CoA. Similar data were obtained in perfused rat hearts, except that the labeling of methylmalonyl-CoA was only 1 ⁄10 that of bicarbonate (Fig. 2B).
We interpreted data of Fig. 2 as suggesting that the enrichment of methylmalonyl-CoA was diluted via the reversal of the methylmalonyl-CoA mutase reaction. We reasoned that the reversal of the mutase reaction would be decreased by increasing the flow of unlabeled carbon through the propionyl-CoA pathway. We perfused livers and hearts with buffer containing Perfusions with [U- 13 C 3 ]Propionate-In livers and hearts perfused with increasing concentrations of [U-13 C 3 ]propionate, the M3 enrichment of methylmalonyl-CoA increased with the [U-13 C 3 ]propionate concentration (Fig. 4, A and B). The M3 enrichment of propionyl-CoA was very close to that of methylmalonyl-CoA (not shown). A trace of M2 enrichment of propionyl-CoA was detected in the presence of 0.1 mM [U-13 C 3 ]propionate, but its value could not be determined with precision. In addition, small amounts of M2 and M1 isotopomers of methylmalonyl-CoA were formed (Fig. 4, A and B). The mass isotopomer distribution of succinyl-CoA (Fig. 4, C and D) showed substantial proportions of M1 to M3 isotopomers.
Perfusions with [U- 13 Fig. 5, A and B). In addition, a very small amount of M2 propionyl-CoA was detected (2.8% in liver, 3.9% in heart). When the concentration of unlabeled propionate was increased, (i) the proportions of M1 to M3 isotopomers of methylmalonyl-CoA decreased to almost zero (Fig. 5, A and B), and (ii) no label was detected in propionyl-CoA.
The same livers and hearts contained substantial proportions of M1, M2, and M3 isotopomers of succinyl-CoA (Fig. 5, C  and D). In the hearts, the distribution of the succinyl-CoA isotopomers was not affected much by increasing the concentration of unlabeled propionate (Fig. 5D). In contrast, in livers perfused under identical conditions, the proportions of the M1 to M3 isotopomers decreased with the concentration of unlabeled propionate (Fig. 5C).
Reversibility of the Methylmalonyl-CoA Mutase Reaction-The M2 enrichments of methylmalonyl-CoA and succinyl-CoA labeled from [U-13 C 3 ]propionate (Fig. 4) were introduced into Equation 1 to compute the fractional contribution of succinyl-CoA to methylmalonyl-CoA, an index of the reversibility of the methylmalonyl-CoA mutase reaction. Fig. 6 (solid symbols) shows this index as a function of the propionate concentration. To compute this index for 0 mM propionate (Fig. 6, open symbols), we used the M2 enrichments of methylmalonyl-CoA and succinyl-CoA labeled from [U-13 C 3 ]lactate ϩ [U-13 C 3 ]pyruvate (left ends of the four panels of Fig. 5). Fig. 6 shows that the contribution of succinyl-CoA to methylmalonyl-CoA is very high (Ͼ90%) in liver and heart in the absence of extracellular propionate. As the concentration of propionate increases, this contribution decreases to low levels (about 10%).
To estimate the level of reversibility of the methylmalonyl-CoA mutase reaction in vivo, we infused NaH 13 CO 3 in a branch of the coronary artery of four anesthetized pigs. In the infused myocardial area, the M1 labeling ratios [methylmalonyl-CoA]/ [bicarbonate] and [succinyl-CoA]/[bicarbonate] ranged from 36 to 53% and from 1.0 to 2.3%, respectively. These relative per-centages are similar to those observed in rat hearts perfused in the absence of propionate (Fig. 2B). The arterial concentration of propionate in the pigs was about 0.03 mM. DISCUSSION The present study was designed to characterize the reactions of the propionyl-CoA pathway in liver and heart, emphasizing the reversibility of the reactions and their impact on the labeling pattern of the CoA intermediates. For this purpose, we perfused livers and hearts with NaH 13 CO 3 , [U-13 C 3 ] propionate, or [U-13 C 3 ]lactate ϩ [U-13 C 3 ]pyruvate at various propionate concentrations.
The rapid labeling of methylmalonyl-CoA and succinyl-CoA from 40% enriched bicarbonate reflected a rapid turnover of the intermediates (Fig. 2, A and B). The enrichment of succinyl-CoA was low because of dilution of label in the CAC. However, the enrichment of methylmalonyl-CoA, formed by propionyl-CoA carboxylation, was much lower than that of bicarbonate in both livers and hearts, suggesting that the enrichment of methylmalonyl-CoA was diluted via the reversal of the methylmalonyl-CoA mutase reaction. As the concentration of unlabeled propionate in the perfusate was increased, the labeling of methylmalonyl-CoA increased until it equaled that of extracellular bicarbonate (Fig. 3, A and B). This confirmed that, in the

FIG. 4. Mass isotopomer distributions of methylmalonyl-CoA (A and B) and succinyl-CoA (C and D) in rat livers (A and C) and hearts (B and D) perfused for 20 min with increasing concentrations of [U-13 C 3 ]propionate. MPE, molar percent enrichment.
absence of extracellular propionate, the labeling of methylmalonyl-CoA from bicarbonate was diluted by the reversal of the methylmalonyl-CoA mutase reaction. Also, the identical enrichment of methylmalonyl-CoA 2 and bicarbonate at high propionate concentration demonstrated, albeit indirectly, the isotopic equilibration of the methylmalonyl-CoA enantiomers.
We then set out to explore the reversibility of methylmalonyl-CoA mutase by channeling label either via propionyl-CoA (using [U-13 C 3 ]propionate) or via the CAC and succinyl-CoA (using [U-13 C 3 ]lactate ϩ [U-13 C 3 ]pyruvate). The use of uniformly labeled substrates and mass isotopomer distribution analysis provides a greater wealth of information than singly labeled substrates, especially when label passes through the CAC (33,34). In livers and hearts perfused with increasing concentrations of [U-13 C 3 ]propionate, the mass isotopomer distribution of methylmalonyl-CoA was characterized mostly by the M3 isotopomer, the abundance of which was greater than 90% (Fig. 4, A and B). This also reflects the isotopic equilibration of the methylmalonyl-CoA enantiomers. In addition, we observed small but clearly identifiable amounts of M2 and M1 isotopomers of methylmalonyl-CoA, which could not be derived directly from M3 propionate. The origin of the M2 and M1 isotopomers of methylmalonyl-CoA becomes evident when one examines the mass isotopomer distributions of succinyl-CoA, which include M3, M2, and M1 isotopomers (Fig. 4, C and D). In the heart, the proportions of the three enriched mass isotopomers of succinyl-CoA are similar. The markedly different isotopomer enrichment patterns of liver and heart succinyl-CoA at the various propionate concentrations reflects these organs' differential capacity for propionate metabolism as well as for anaplerosis. In the livers, the proportions of the three enriched mass isotopomers of succinyl-CoA are different (M3 Ͼ M2 Ͼ M1), presumably because of additional isotopic exchanges in the pyruvate 3 oxaloacetate 3 phosphoenolpyruvate 3 pyruvate cycle. In both livers and hearts perfused with [U-13 C 3 ]propionate, the M2 and M1 mass isotopomers of succinyl-CoA are clearly the precursors of the M2 and M1 isotopomers of methylmalonyl-CoA. This provides direct evidence for the reversibility of the methylmalonyl-CoA mutase reaction.
Probing the propionyl-CoA pathway with NaH 13 CO 3 or [U-13 C 3 ]propionate could not yield information on the possible reversibility of the propionyl-CoA carboxylase reaction. This is why we labeled the pathway via the CAC and succinyl-CoA using [U-13 C 3 ]lactate ϩ [U- 13 C 3 ]pyruvate in the presence of increasing concentrations of unlabeled propionate. With 0 mM propionate, the mass isotopomer distribution of heart succinyl-CoA showed similar proportions of the M1 to M3 isotopomers (left side of Fig. 5D). This again reflects and is compatible with the relatively low entry of unlabeled carbon through anaplerosis in the heart compared with liver. The proportions of the M1 to M3 isotopomers of succinyl-CoA are not markedly affected by increasing concentrations of unlabeled propionate. This probably results from the small ratio of (anaplerotic flux of propionyl-CoA)/(CAC), i.e. less than 10%, observed under similar conditions (12,13). In the same hearts, the methylmalonyl-CoA pool also included M1 to M3 isotopomers in a proportion that decreased as the concentration of unlabeled propionate increased in the perfusate. Comparison between the mass isotopomer patterns of heart succinyl-CoA and methylmalonyl-CoA supports the conclusion that methylmalonyl-CoA is derived from succinyl-CoA. This also demonstrates that the methylmalonyl-CoA mutase reaction is reversible in the heart and that this reversibility is decreased by increasing the flux through the propionyl-CoA pathway. We detected a low M2 labeling of propionyl-CoA from [U-13 C 3 ]lactate ϩ [U-13 C 3 ]pyruvate (2.8% in liver, 3.9% in heart), but only in the absence of exogenous propionate. When the M2 enrichments of propionyl-CoA and methylmalonyl-CoA were introduced in Equation 2, the fractional contributions of methylmalonyl-CoA to propionyl-CoA were 6 and 19% in liver and heart, respectively. When unlabeled propionate was added to the perfusate, no enrichment of propionyl-CoA from [U-13 C 3 ]lactate ϩ [U-13 C 3 ]pyruvate was detectable. Therefore, under our conditions, there was a slight reversibility of the propionyl-CoA carboxylase reaction, but only in the absence of exogenous propionate.
The data obtained in livers perfused with [U-13 C 3 ]lactate ϩ [U-13 C 3 ]pyruvate and increasing concentrations of unlabeled propionate (Fig. 5, A and C) were similar to the perfused heart experiments (Fig. 5, B and D), with one exception. In livers, the proportions of the M1 to M3 isotopomers of succinyl-CoA markedly decreased as the concentration of unlabeled propionate in the perfusate was increased. Again this reflects the greater capacity of the liver for anaplerosis and for the metabolism of propionate to oxaloacetate and pyruvate.
A quantitative index of the reversibility of methylmalonyl-CoA mutase, i.e. the fractional contribution of succinyl-CoA to the methylmalonyl-CoA pool was calculated from the M2 enrichments of methylmalonyl-CoA and succinyl-CoA (Fig. 6). In isolated livers and hearts perfused without propionate, more than 90% of methylmalonyl-CoA was derived from the reversal of methylmalonyl-CoA mutase. This high fractional contribution does not represent a net supply of methylmalonyl-CoA, but a rapid interconversion of the two CoA esters at no energy cost. Thus, in the absence of exogenous propionate, the net supply of methylmalonyl-CoA (less than 10%) results from the small flow of propionyl-CoA presumably derived from the catabolism of endogenous aminoacids. However, the rapid interconversion of methylmalonyl-CoA and succinyl-CoA has a major effect on the former's labeling pattern. This is because the labeling pattern of succinyl-CoA is strongly influenced by the CAC and related pathways. This explains the major dilution of the enrichment of methylmalonyl-CoA labeled from NaH 13 CO 3 (Fig. 2, A and B, and left sides of Fig. 3, A and B). As the concentration of propionate is increased in the perfusate, the reversibility of methylmalonyl-CoA mutase decreases (Fig. 6), and the labeling of methylmalonyl-CoA from NaH 13 CO 3 becomes very close to that of the latter (Fig. 3, A and B).
To what extent is the methylmalonyl-CoA mutase reaction reversible in vivo? In non-ruminant mammals, the arterial concentration of propionate is very low (5 M in dogs (35), 30 M in pigs (13)). From Fig. 6, one would expect marked reversibility of the reaction in the heart. This conclusion is supported by our data obtained in the pig heart perfused in situ with 15% 13 C-labeled bicarbonate. In the liver of non-ruminant mammals, the portal vein propionate concentration is about 0.2 mM (35), which should somewhat slow down the reversibility of the mutase reaction. In both heart and liver, the reversibility of the mutase should be markedly decreased after ingestion of a diet rich in triheptanoin, a precursor of propionyl-CoA used to boost anaplerosis in patients with long chain fatty acid oxidation defects (11). Last, in ruminant mammals with high propionate concentrations in the portal and arterial blood, there is probably little reversibility of the mutase.
Propionyl-CoA carboxylase is generally referred to as catalyzing a reversible reaction (20). Kaziro et al. (15) commented that, based on the equilibrium constant of 5.7 and the ⌬F of Ϫ1028 calories/mol, the reaction is readily reversible. Our data are at variance with this conclusion for two reasons. First, in perfusions with [U-13 C 3 ]propionate, propionyl-CoA was only M3 labeled, while methylmalonyl-CoA was labeled in M3, M2, and M1 (Fig. 4, A and B). Second, in perfusions with [U-13 C 3 ]lactate ϩ [U-13 C 3 ]pyruvate and 0 mM propionate, the labeling of propionyl-CoA was very low (less than 4% M2, not shown). When unlabeled propionate was added to the perfusate, no label was detectable in propionyl-CoA. In contrast, the labeling of methylmalonyl-CoA from [U-13 C 3 ]lactate ϩ [U-13 C 3 ]pyruvate was substantial in M3, M2, and M1 (Fig. 5, A  and B).
We propose that the very low reversibility of the propionyl-CoA carboxylase reaction in intact livers and hearts results from the fact that the mass action ratio of this reaction is far from equilibrium. The equilibrium constant of the propionyl-CoA carboxylase reaction allows one to calculate that, at thermodynamic equilibrium, the mass action ratio we would expect the mass action ratio [(R,S)-methylmalonyl-CoA]/[propionyl-CoA] to be about 20 if the combined propionyl-CoA carboxylase ϩ methylmalonyl-CoA racemase reaction were at thermodynamic equilibrium. In fact, we found that the ratio [(R,S)-methylmalonyl-CoA]/[propionyl-CoA] is about 0.6 in rat livers perfused without or with 1 mM propionate (27). We can, therefore, conclude that the propionyl-CoA carboxylase reaction is quite far from thermodynamic equilibrium in the intact liver.
This rationale can be extended to the combined equilibrium constant of the reactions catalyzed by propionyl-CoA carboxylase, methylmalonyl-CoA racemase, and methylmalonyl-CoA mutase (KЈ ϭ 18.6 (24)). Then, the [succinyl-CoA]/ [propionyl-CoA] ratio should be 177 at thermodynamic equilibrium. In fact, we found that the ratio is (i) 3.7 and 0.14 in rat livers perfused without or with 1 mM propionate, respectively, and (ii) 18 and 0.3 in hearts perfused without or with 1 mM propionate, respectively. The measured [succinyl-CoA]/[propionyl-CoA] ratio is one to three orders of magnitude smaller than its thermodynamic equilibrium value (12,27). The dis-equilibrium of the propionyl-CoA to succinyl-CoA reaction sequence reflects the drawing of succinyl-CoA derived from propionyl-CoA into the CAC. This explains the efficiency of anaplerosis from a low concentration (0.25 mM) of [U-13 C 3 ]propionate in pig heart (9% of the flux through the CAC (13)).
In conclusion, our data demonstrate that in intact normal rat livers and hearts, (i) the methylmalonyl-CoA mutase reaction is partly reversible and that this reversibility is modulated by the propionyl-CoA flux, (ii) the methylmalonyl-CoA enantiomers are maintained in isotopic equilibrium via the reversible methylmalonyl-CoA racemase reaction, and (iii) the reversibility of the propionyl-CoA carboxylase reaction is minor. The dis-equilibrium of the propionyl-CoA to succinyl-CoA reaction sequence explains the effectiveness of anaplerosis from low concentrations of propionyl-CoA precursors such as propionate or heptanoate (11,13).