Probing the Origin of Acetyl-CoA and Oxaloacetate Entering the Citric Acid Cycle from the 13C Labeling of Citrate Released by Perfused Rat Hearts*

We present a strategy for simultaneous assessment of the relative contributions of anaplerotic pyruvate carboxylation, pyruvate decarboxylation, and fatty acid oxidation to citrate formation in the perfused rat heart. This requires perfusing with a mix of13C-substrates and determining the 13C labeling pattern of a single metabolite, citrate, by gas chromatography-mass spectrometry. The mass isotopomer distributions of the oxaloacetate and acetyl moieties of citrate allow calculation of the flux ratios: (pyruvate carboxylation)/(pyruvate decarboxylation), (pyruvate carboxylation)/(citrate synthesis), (pyruvate decarboxylation)/(citrate synthesis) (pyruvate carboxylation)/(fatty acid oxidation), and (pyruvate decarboxylation)/(fatty acid oxidation). Calculations, based on precursor-product relationship, are independent of pool size. The utility of our method was demonstrated for hearts perfused under normoxia with [U-13C3](lactate + pyruvate) and [1-13C]octanoate under steady-state conditions. Under these conditions, effluent and tissue citrate were similarly enriched in all 13C mass isotopomers. The use of effluent citrate instead of tissue citrate allows probing substrate fluxes through the various reactions non-invasively in the intact heart. The methodology should also be applicable to hearts perfused with other 13C-substrates, such as 1-13C-labeled long chain fatty acid, and under various conditions, provided that assumptions on which equations are developed are valid.

Tracing of pyruvate metabolism and of other reactions feeding into (anaplerosis) or out (cataplerosis) of the citric acid cycle (CAC) 1 with radioactive or stable isotopes is complicated by label recycling and exchange reactions between CAC intermediates and other metabolites such as aspartate and glutamate. Mathematical models of increasing complexity were developed for the study of the CAC in various organs or tissues (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12), including the heart (13)(14)(15)(16)(17). Solving for flux parameters in equations derived from these models requires measuring the incorporation of 14 C-or 13 C-labeled substrate(s) into various CAC metabolites (8,9) or the distribution of label between carbons of given molecules such as glutamate (1,(13)(14)(15)(17)(18)(19) or citrate (4,6). The use of 13 C-enriched labeled substrate(s) and measurements of 13 C labeling of CAC intermediates or related metabolites by nuclear magnetic resonance (NMR) or gas chromatography-mass spectrometry (GCMS) offers several advantages over classical 14 C methods. Also, these two techniques are complementary. One advantage of GCMS over NMR is its sensitivity. Thus, the 13 C labeling of the actual CAC intermediates can be determined.
Investigations on the cardioprotective effects of pyruvate have emphasized the concerted regulation of pyruvate decarboxylation and fatty acid oxidation (14 -15, 17, 20 -23). For this purpose, elegant 13 C NMR techniques were developed to quantitate from the 13 C labeling of carbons of glutamate, the relative contributions of pyruvate, fatty acid, and ketone body oxidation to acetyl-CoA formation in hearts perfused under various conditions (15, 17, 24 -25). By replenishing CAC intermediates, anaplerotic carboxylation of pyruvate could also be a key component for restoration of cardiac function following ischemia (26 -28). Substrate fluxes through pyruvate carboxylation were evaluated in hearts perfused with 14 C-labeled pyruvate (29 -30). However, in these studies, the partitioning of pyruvate between carboxylation and decarboxylation could not be directly quantitated. Such measurements could help clarify the relative importance of these two pathways in the cardioprotective effect of pyruvate (23, 26 -27, 29 -32).
In the present paper, we present a strategy for the direct and simultaneous assessment of the relative contributions of anaplerotic pyruvate carboxylation, pyruvate decarboxylation, and fatty acid oxidation to citrate formation in the intact perfused rat heart. This is accomplished using a combination of 13 C-substrates and requires determination of the 13 C mass isotopomer distributions (MIDs) of the acetyl and oxaloacetate (OAA) moieties of a single metabolite, effluent citrate, by GCMS. Part of this work was presented in abstract form (33)(34)(35). THEORY

Principle of Mass Isotopomer Analysis
The relative contributions of pathways feeding OAA and acetyl-CoA for citrate synthesis were determined in hearts perfused with a mix of 13 C-substrates under steady-state conditions. [U- 13 C 3 ]Lactate and [U- 13 C 3 ]pyruvate were supplied at physiological concentrations and in a ratio to clamp the redox state. Also a source of acetyl-CoA other than pyruvate, was provided, namely a 1-13 C-labeled fatty acid, octanoate. Calculation of the various substrate flux ratios requires GCMS determination of the MIDs of (i) the acetyl and OAA moieties of citrate and of (ii) pyruvate.

Calculation of Relative Substrate Fluxes
The following notations are used in developing the equations: Metabolites: AC, acetyl-CoA; CIT, citrate; OAA, oxaloacetate; OCT, octanoate; PYRi and PYRe, intracellular and extracellular pyruvate; AC CIT , the acetyl moiety of citrate, equivalent to carbons 4 and 5 of citrate (C-4 ϩ 5); OAA CIT , the OAA moiety of citrate equivalent to C-1 ϩ 2 ϩ 3 ϩ 6. Isotopomer specifications: OAA Mi , mass isotopomer of a given metabolite, for example OAA, labeled with i atoms of 13 C; MF, mol fraction in a given mass isotopomer (Mi) of a metabolite, calculated as MF (M i ) ϭ A Mi /(⌺A Mi ) where A represents the peak area of each fragmentogram, determined by computer integration and corrected for naturally occurring heavy isotopes (6 -7, 18, 36); MPE, molar percent enrichment in a given mass isotopomer of a metabolite, equivalent to MF ϫ 100. Flux rates: FC PYRi3 OAA , fractional contribution (FC) of one metabolite to the total flux of the other, e.g. intracellular pyruvate to OAA. For a given reaction, the sum of the different FCs equals 1.
The principle of the calculation of relative input fluxes from measured MID of given metabolites was described in previous publications for liver and heart perfusions (6 -7, 37-38). Briefly, we first assume, for simplicity, that carboxylation of M3 pyruvate is the only pathway forming M3 OAA. Then, we consider the possible formation of M3 OAA through the metabolism in the CAC of some citrate isotopomers. Conditions where PYRe M3 is the MF in M3 of the incoming pyruvate tracer. The contribution of unlabeled pyruvate arising from glucose or glycogen through glycolysis is (1 Ϫ FC PYRe3 PYRi ). Since M3 OAA is the only source of M3 citrate labeled in its OAA moiety (OAA CIT ), the balance of M3 isotopomers of OAA moiety of citrate and adjacent metabolites yields Equation 3, where FC OAA3 CIT is the fractional contribution of OAA to citrate via citrate synthase, and OAA M3 CIT is the MF in M3 of the OAA moiety of citrate. Multiplying Equation 1 by Equation 3 yields the fractional contribution of pyruvate to citrate via the carboxylation of pyruvate (FC PYRi3 CIT(OAA) ). The latter expression is equivalent to the flux ratio (pyruvate carboxylation)/ (citrate synthesis), also named factor "y" (2-3, 6, 39 -41), when flux through citrate synthase is set equal to 1 (see Equation 4).
Similar reasoning yields the flux ratios (pyruvate decarboxylation)/(citrate synthesis) or factor "x" (2-3, 6, 39) and (octanoate oxidation)/(citrate synthesis) using Equations 5 and 6, respectively, where (i) FC PYRi3 AC and FC OCT3 AC are the fractional contributions of pyruvate decarboxylation and of octanoate oxidation to acetyl-CoA, respectively; (ii) FC AC3 CIT is the fractional contribution of acetyl-CoA to citrate via citrate synthase; and (iii) OCTi M1 is the MF in M1 of intracellular octanoate, which we assumed equal to that of extracellular octanoate, since octanoate is not a physiological fatty acid. The factor 4 in Equation 6 takes into account that octanoate is oxidized to 4 acetyl-CoA; only one of which is labeled. The sum of Equations 5 and 6 (FC PYRi3 CIT ϩ FC OCT3 CIT ) reflects the relative combined contributions of pyruvate decarboxylation and octanoate oxidation to the acetyl moiety of citrate; the contribution of other substrates (FC OS3 AC(CIT) ), such as endogenous fatty acids and/or leucine, is given by Equation 7.
Conditions Where M3 Pyruvate Is Not the Sole Source of M3 OAA-In developing Equations 1-6, we assumed that carboxylation of M3 pyruvate was the only reaction forming M3 OAA molecules. Such assumptions are likely to prevail in hearts perfused with a low enrichment of [U-13 C 3 ]lactate and [U-13 C 3 ]pyruvate. However, in hearts perfused simultaneously with a mix of 13 C-substrates, the probability of condensation between labeled acetyl-CoA and labeled OAA is increased. Some citrate molecules labeled in both their acetyl and OAA moieties could be metabolized to M3 OAA in the CAC. Therefore, to assess the validity of our initial assumption, we evaluated the proportion of the M3 OAA molecules formed through the recycling of some citrate isotopomers in the CAC and subsequently corrected the measured MF M3 of the OAA moiety of citrate (OAA M3 CIT ). This was done using Equation 8 and requires estimating the (i) MPE of citrate isotopomer precursor of M3 OAA (OAA M3 PR ) and (ii) the 13 C dilution in the CAC (DF). Values for (i) MPE OAA M3 PR and (ii) DF, obtained using Equations 9 and 10 (see below), are approximations. More precise estimation of these values would require extensive modeling of the metabolism of our mix of 13 C-substrates in the CAC, where (i) OAA M3 PR is the enrichment of citrate isotopomer precursor of M3 OAA and (ii) DF is the correction factor for the 13  Upon metabolism in the CAC, citrate isotopomers B and C form M3 OAA, whereas isotopomers A and D form M2 OAA. For simplicity, it is assumed that citrate isotopomers A to D are similarly enriched, although this is an approximation of the actual situation (see "Discussion" for details). Assessing the validity of this assumption requires more extensive modeling of the metabolism of our mix of 13 C-substrates in the CAC. Then, the enrichment of citrate isotopomers B and C is estimated CIT and OAA M1 CIT are experimentally determined enrichments of the acetyl and OAA moieties of citrate, and the number 0.5 takes into account that only half of citrate isotopomers formed through the condensation of M2 acetyl-CoA and M1 OAA are metabolized to M3 OAA. This constitutes the first term of Equation 9. Note that for clarity, the superscript CIT was omitted. A similar reasoning was applied for each term of this equation.
The 13 C dilution in the CAC due to entry of unlabeled molecules was estimated from the total enrichment in tissue citrate (CIT ⌺Mi ) and succinate (SUC ⌺Mi ) using Equation 10, where the factor f takes into account that a fraction of all citrate molecules enriched with one 13 C in any one carbon of its OAA moiety are converted to unlabeled succinate. Note that the magnitude of f depends on the nature of the 13 C-substrate.
Heart Perfusions-Fed male Sprague-Dawley rats (Charles River, Quebec) weighing 120 -220 g were anesthetized by intraperitoneal injection of sodium pentobarbital (65 mg/kg). After opening the chest and insertion of a cannula into the aorta, hearts were excised and transferred to a Langendorff set up as described previously (38,42). Briefly, hearts (wet weight 1.1-1.3 g) were perfused for up to 60 min retrogradely through the aorta at a constant pressure of 80 mmHg with a non-recirculating Krebs-Ringer bicarbonate buffer containing 1.3 mM calcium, 11 mM glucose, 1 or 0.5 mM lactate, 0.2 or 0.05 mM pyruvate, 0.2 mM octanoate, with or without 0.5 mM glutamate or 0.1 mM acetate. The buffer was gassed with 95% O 2 :5% CO 2 (pH 7.4) at 38°C. The following parameters were continuously monitored through instruments linked to a microcomputer: (i) coronary flow, using an electromagnetic flow probe (model FM501, Carolina Medical Electronics, King, NC) installed above the aortic cannula; (ii) temperature using a thermocouple (Yellow Springs Instrument, Yellow Springs, OH) attached to the surface of the heart; and (iii) contractile function using a latex balloon filled with fluid inserted into the left ventricular cavity and connected to a pressure transducer (Digi-Med Heart Performance Analyzer, Micro-Med, Louisville, KY). Hearts that did not show an increase in coronary flow on release of a 20 -25-s period of in-flow occlusion were discarded. Under our conditions, hearts were beating spontaneously at 280 Ϯ 19 beats/min and maintained a contractile function (dP/dT) of 2099 Ϯ 152 mmHg/s and a coronary flow rate of 9.3 Ϯ 0.6 ml/min.
After a 15-20-min equilibration period, one or more unlabeled substrate(s) were replaced by the corresponding labeled substrates either acetate. Starting 10 min before the labeling period, samples of effluent perfusate (20 ml) were collected every 5 min and processed as follows: (i) 7 ml was immediately made 10 mM hydroxylamine-hydrochloride and sonicated for 1 min to convert ␣-ketoglutarate (␣KG) to its oxime derivative (43); (ii) 10 ml was made 1% sulfosalicylic acid; and (iii) 1 ml was left untreated. Samples were stored at Ϫ20°C until analysis. After perfusing for 40 min with 13 Csubstrate(s), hearts were freeze-clamped and stored in liquid nitrogen.
GCMS Methodology-The MID of tissue and effluent perfusate citrate and other CAC metabolites (␣KG, OAA, succinate, fumarate, malate) was determined by published GCMS methods (6,18,(37)(38)43). For subcellular fractionation, tissue was homogenized in Tris buffer 50 mM with sucrose 250 mM (pH 7.4) (15% w/v), followed by sequential 10 min centrifugation at 4°C, at 1,000 ϫ g, and at 10,000 ϫ g (44). Supernatant and pellet were extracted with ethyl acetate and analyzed for citrate. For determination of the MID of the OAA moiety of effluent citrate, effluent perfusate samples treated with sulfosalicylic acid were adjusted to pH 7.6 and reacted with an excess of NaBH 4 (10 mol) to convert OAA to malate. After 30 min at room temperature, the solution was acidified with concentrated HCl to destroy residual NaBH 4 , and the pH was readjusted to 7.6 before the addition of 750 l of 500 mM triethanolamine buffer (pH 7.4) containing 100 mM MgSO 4 , 50 mM EDTA, 50 mol hydroxylamine-hydrochloride, and 2.5 units of citrate lyase. After 5 min at 37°C, samples were sonicated (for complete conversion of OAA to its oxime derivative (43)). After addition of 350 l of saturated sulfosalicylic acid and centrifugation at 3,000 ϫ g for 15 min, the oxime was extracted from the supernatant with 3 ϫ 18 ml of diethyl ether. From the measured MIDs of effluent citrate and of its OAA moiety, we calculated the MIDs of the acetyl moiety of effluent citrate (see below for details).
GCMS Conditions-All metabolites were analyzed as their t-butyldimethylsilyl derivatives on a Hewlett-Packard 5890 Series I or II gas chromatograph coupled to a 5988A mass spectrometer or 5972A mass selective detector. Both instruments were equipped with a HP 5 capillary column (30 or 50 m, 0.2-mm inner diameter, 0.33-m film thickness). The mass spectrometer was operated in the electron impact mode (70 eV) after automatic and manual calibrations (to optimize sensitivity in the high mass range). Split ratio injection was about 10/1, carrier gas helium was at 0.6 -0. 8 Table II; Conditions A and B). Therefore, ⌺f i equals ((0.79 ϫ 1) ϩ (0.21 ϫ 0.5)) or 0.895. Alternatively, the relative contributions of these 13 C-substrates can be estimated from the M1 and M2 enrichments of the acetyl moiety of citrate, 88.0 Ϯ 0.3 and 12.0 Ϯ 0.7%, respectively (Table IIIC). temperature programs, elution times, and ions monitored for the analysis of the MID of CAC metabolites were previously reported (6,18,(37)(38)43). Areas under each fragmentogram were determined by computer integration and corrected for naturally occurring heavy isotopes (6 -7, 18, 36).
MID of the Acetyl Moiety (C-4 ϩ 5) of Citrate-The MID of the acetyl moiety of citrate (AC M i CIT ) was calculated from the measured MIDs of effluent citrate (CIT Mi ) and of its OAA moiety (OAA M i CIT ), obtained after cleavage with citrate lyase (see Scheme 1). The enrichment of each mass isotopomer of citrate is a function of those of its acetyl and OAA moieties (Equations 11-17). Note that for clarity, the superscript CIT was omitted from the factors "AC" and "OAA" in Equations 11-17.
In Equations 11-17, the MID data are expressed as MF. Note that the sum of the MF of all mass isotopomers of a given substrate equals 1. These equations were solved for the M1 and M2 enrichments of the acetyl moiety of citrate (i) using a spreadsheet program (Lotus, release 3.1) and (ii) by simplification. First, using our program, MIDs of citrate were calculated from measured MIDs of its OAA moiety and different theoretical MID values of its acetyl moiety. Solutions for the M1 and M2 enrichments of the acetyl moiety of citrate were considered optimal when the difference between calculated and measured MIDs of citrate was 10% or less. Second, for simplification of Equations 11 ). Therefore, we concluded that Equations 18 and 19 provide a simple and reliable mean to calculate the M1 and M2 enrichments of the acetyl moiety of citrate from the measured MIDs of citrate and of its OAA moiety. In the present paper, we report only values calculated using these equations.

Expression of Results and Statistical Analysis-The
MIDs of the various metabolites, corrected for natural abundance, are expressed as molar percent enrichment (MPE). For calculations of the MID of acetyl-CoA and of relative flux parameters (FC values), MID data are expressed as mol fraction (MF). Data are presented as means Ϯ S.E. Statistical significance at p Ͻ 0.05 was assessed using the indicated test.

RESULTS AND DISCUSSION
Origin of Effluent Citrate-Rat hearts perfused with 13 Csubstrates under normoxic conditions constantly released small amounts of 13 C-labeled citrate, the MID of which can be analyzed with precision by GCMS. The average coefficient of variations (% CV ϭ S.D./mean ϫ 100) for all measured enrichments was 6 Ϯ 1%. To identify the origin of this effluent citrate, the 13 C labeling of citrate isolated from the effluent perfusate collected after 35-40 min of heart perfusion with [U-13 C 3 ]lactate and [U-13 C 3 ]pyruvate was compared with that of citrate isolated from different centrifugal preparations of heart homogenates enriched in either mitochondria or cytosol. As shown in Fig. 1, citrate isolated from all sources was similarly enriched in all mass isotopomers, indicating a similar origin. Citrate could be formed (i) from OAA and acetyl-CoA via citrate synthase and/or (ii) from ␣KG and bicarbonate through the reversal of the aconitase and isocitrate dehydrogenase reactions. To determine the relative contributions of these pathways, hearts were perfused with unlabeled lactate, pyruvate, SCHEME 1. Determination of the MID of the acetyl moiety of citrate. This scheme summarizes the experimental procedure followed to determine the M1 and M2 enrichments of the acetyl moiety of citrate from the measured MID of effluent citrate and of its OAA moiety, obtained after cleavage with citrate lyase. octanoate, and 0.5 mM [U-13 C 5 ]glutamate following a protocol developed for perfused livers (37). From the measured M5 enrichment of tissue citrate and ␣KG, shown in Table I, one calculates the fractional contribution of ␣KG to citrate through the reversal of the aconitase and isocitrate dehydrogenase reactions (FC ␣KG3 CIT ) to be 5.0 Ϯ 1.0%. The remaining 95% of citrate molecules comes from OAA. This low rate of reversal of the aconitase and isocitrate dehydrogenase reactions in the heart contrasts with that in perfused livers where as much as 45% of all citrate molecules was formed from ␣KG (37).
From the above data, we concluded that hearts perfused under normoxia with [U-13 C 3 ]lactate, [U-13 C 3 ]pyruvate, and/or [1-13 C]octanoate released a small amount of 13 C-labeled citrate, the MID of which reflected that of tissue citrate formed predominantly through citrate synthase. Citrate efflux, which was shown to increase with citrate concentration in isolated rat heart mitochondria (45), may be favored by the inclusion in our perfusion buffer of several citrate precursors (glucose, lactate, pyruvate, octanoate, see "Discussion" in Comte et al. (61)). The rate of citrate efflux (ϳ10 -50 nmol/min) is compatible with the reported low activity of the mitochondrial tricarboxylate transporter of rat heart (45)(46). 13 C Labeling of the Acetyl and OAA Moieties of Citrate-The 13 C labeling of the acetyl and OAA moieties of effluent citrate was examined in hearts perfused with 11 mM glucose, 1 mM lactate, 0.2 mM pyruvate, and 0.2 mM octanoate with or without 0.1 mM acetate. In any one perfusion, one or more unlabeled substrate was replaced by the corresponding 13 (Fig. 2). Near isotopic steady state was attained more rapidly for M3 and M2 (10 -15 min) than for M1 (20 -30 min) isotopomers. Table II compares the MIDs of effluent citrate and of its OAA moiety when hearts were perfused for more than 30 min with one or more 13 C-substrates. These MID data were then introduced into Equations 18 and 19 to calculate the MID of the acetyl moiety of citrate, shown in Table III. Note that a good precision was obtained for these calculated values when measured 13 C enrichments were above 5%. For example, the mean M2 enrichment of the acetyl moiety of citrate (ϮS.D.) and CV (%) for 3-5 injections was 9.1 Ϯ 0.5 and 5%, respectively, for an effluent sample collected from hearts perfused with [U-13 C 2 ]acetate but was 0.75 Ϯ 0.31% and 42%, respectively, for a sample collected from hearts perfused with [U-13 C 3 ](lactate ϩ pyruvate).
Despite variations, the labeling patterns observed in Tables  II and III are compatible with the known metabolism of the 13 C-substrates. This is apparent for hearts perfused with one 13 (Tables II and III, A, B, and D), but not with a mix of 13 C-substrates (Tables IIC and IIIC). For example, [1-13 C]octanoate is oxidized to [1-13 C]acetyl-CoA which is converted to [5-13 C]citrate. Upon further metabolism, [5-13 C]citrate is converted to [ 13 C]OAA labeled on any one carbon. Accordingly, when hearts were perfused with [1-13 C]octanoate as sole tracer, citrate as well as its acetyl and OAA moieties, were predominantly enriched in M1 isotopomers (Table IIA, 1st and 2nd lines, and Table IIIA). The small enrichment in M2 citrate (Table IIA, 1st line) results mostly from the condensation of M1 OAA with M1 acetyl-CoA. However, the presence of some M2 (Ͻ1%) in the acetyl moiety of citrate (Table IIIA) cannot be explained by known pathways of [1-13 C]octanoate metabolism. Therefore, the M2 was ascribed to background noise and was subtracted from the M2 enrichment of the acetyl moiety of citrate calculated for heart perfusions with all three 13 C-substrates (see below for further discussion on the precision of the calculated MPE of the acetyl moiety of citrate).
Calculation of Relative Fluxes-The MIDs of the acetyl and   1.5% (n ϭ 3), indicating that tissue pyruvate arose predominantly from exogenous pyruvate and/or lactate (from Equation 2). Tissue pyruvate enrichment in M1 and M2 isotopomers was negligible (not significant; mean tested against the null hypothesis), suggesting low pyruvate recycling through pyruvate 3 OAA 3 malate 3 pyruvate, or pyruvate 3 acetyl-CoA 3 CAC 3 malate 3 pyruvate. However, because of evidence supporting compartmentation of myocardial pyruvate (30, 50 -51), one cannot exclude recycling of a very small mitochondrial pool of pyruvate which would not be detected by GCMS assay of pyruvate in whole heart homogenates.
Methodological Considerations: Advantages and Limitations-In this paper, we show how one can use the 13 C labeling pattern of citrate released by hearts perfused with 13 C-substrates to probe the origin of the acetyl-CoA and OAA moieties of citrate. This reflects the relative contributions of pyruvate carboxylation, pyruvate decarboxylation, and fatty acid oxidation. One could also analyze tissue citrate after purification through ion-exchange chromatography (53). However, the use of effluent instead of tissue citrate allows probing substrate fluxes non-invasively through the various reactions in hearts perfused under different conditions. For our perfusion experiments, we chose 13 C-substrates and concentrations to allow competition at the acetyl-CoA and OAA branch point of pyruvate metabolism. Also, anaplerotic pyruvate carboxylation was favored by the inclusion of (i) physiological concentrations of lactate and pyruvate and of (ii) a medium chain fatty acid (26). In hearts, unlike long chain fatty acids, medium chain fatty acids are not esterified but completely oxidized in mitochondria (54). For future studies, one may consider replacing octanoate with another more physiological fatty acid. By measuring the M3 enrichment of tissue pyruvate, we evaluated the carboxylation and decarboxylation of pyruvate from all sources, exogenous and endogenous. Alternatively, one could assay the M3 enrichment of tissue alanine, the labeling of which reflects that of pyruvate molecules committed to the CAC in hearts perfused with [3-13 C]pyruvate (55). Whether the MID of effluent alanine also reflects that of tissue pyruvate remains to be examined. Note that the contribution of intracellular pyruvate to the acetyl moiety of citrate does not reflect its contribution to the total acetyl-CoA pool; part of the acetyl-CoA could be converted to acetylcarnitine and channeled to the cytosol (56 -57).
The reliability of the various flux ratios calculated from our MID data depends (i) on the precision of the measured and calculated enrichment data and (ii) on the validity of assumptions on which Equations 1-6 were developed. Low precision on calculated enrichment values of the acetyl moiety of citrate below 5% results in part from the high natural background of the M1 and M2 ions for the t-butyldimethylsilyl of citrate and OAA (for e.g. M2/M0 ϭ 0.183 and 0.163, respectively). Precision is further decreased when these low enrichments occur at a C labeling of effluent citrate and of its OAA moiety isolated from hearts perfused with one or more 13 C-substrate(s) Hearts were perfused under normoxia with non-recirculating buffer containing 11 mM glucose and 0.2 mM octanoate. Other additions were lactate and pyruvate at 0.5 and 0.05 mM (A and D) or 1 and 0.2 mM (B and C), respectively, and acetate at 0.1 mM (A and D). After 15-20 min equilibration, one or more unlabeled substrate(s) were replaced by the corresponding labeled substrate(s), as indicated. The perfusion was continued for another 40 min. MID data, corrected for natural abundance in heavy isotopes, are shown as MPE and are means Ϯ S.E. for n different perfusions. ND, not determined. 13 (Table IIIB). To decrease this variability, two strategies could be considered. First, one could calculate the enrichments of the acetyl moiety of citrate by (i) introducing into Equations 11-17 the uncorrected MID data of citrate and of its OAA moiety, and (ii) including their natural abundance in a fitting routine (58). This one-step calculation requires an overdetermined system, i.e. that the number of citrate isotopomers measured exceeds the number of isotopomers of its acetyl moiety to be calculated. This should be advantageous because it compensates for errors or variations in the correction factors for natural abundance of all measured isotopomers. Alternatively, one could assay directly the MID of the acetyl moiety of citrate after cleavage with ATP-citrate lyase and conversion of acetyl-CoA to acetylglycine (59). The latter has a low natural abundance background at the M1 and M2 ions (M1/M0 ϭ 0.0504; M2/M0 ϭ 0.00715). The second approach requires, however, purification of ATP-citrate lyase from liver. Direct analysis of the 13 C labeling of the acetyl moiety of citrate either as (i) acetate released by cleavage of citrate with citrate lyase or (ii) as acetyl-CoA released by cleavage of citrate with ATP-citrate lyase, as described previously (53), was not feasible because of contamination with unlabeled acetate or a too low concentration of citrate in the effluent perfusate (ϳ1 M), respectively (not shown).
As for the assumptions of Equations 1-6, they are as follows: (i) the 13 C labeling of effluent citrate reflects that of tissue citrate; (ii) citrate is formed only through the citrate synthase reaction; (iii) the M1 enrichment of the acetyl moiety of citrate results exclusively from the oxidation of fatty acid, in the present study, [1-13 C]octanoate, and (iv) the carboxylation of [U-13 C 3 ]pyruvate is the only reaction generating M3 OAA. Under our conditions, the first three assumptions were valid. First, there were no significant differences in the measured MIDs of effluent and tissue citrate (Fig. 1). Second, the rate of citrate formation by the reversal of the aconitase and isocitrate dehydrogenase reactions was low (Table I). Third, perfusing hearts with [1-13 C]octanoate, but not with [U-13 C 3 ](lactate ϩ pyruvate) or [U-13 C 2 ]acetate, resulted in significant M1 enrichment of the acetyl moiety of effluent citrate (Table IIIA and B). Note that the second and third assumptions might not be valid in other perfusion conditions or other organs or cell systems such as liver (6 -7, 37) or ␤-cells (60) where there is substantial pyruvate recycling through the reactions: pyruvate 3 OAA 3 malate 3 pyruvate, or pyruvate 3 OAA 3 phosphoenolpyruvate 3 pyruvate. With respect to the fourth assumption, its validity was evaluated using Equations 8 -10. Under our conditions, the MF of citrate isotopomers, precursors of M3 OAA (OAA M3 PR ), was 0.0116 (Equation 9). 3 The dilution factor was 1.13 Ϯ 0.04 (n ϭ 8, Equation 10), which is similar to that reported for the perfused rat heart (12). Thus, more than 82% of M3 OAA molecules was formed through carboxylation of [U-13 C 3 ]pyruvate. Accordingly, correcting the measured M3 enrichment of the OAA moiety of citrate for the contribution of citrate isotopomers (Equation 8) did not significantly modify the flux ratios (pyruvate carboxylation)/(citrate synthesis) (0.063 Ϯ 0.009 versus 0.077 Ϯ 0.011; paired t test). The formation of M3 OAA, from citrate isotopomers recycling through the CAC, should be minimized by lowering the enrichment of the incoming 13 C-substrates. Under our conditions, however, [U-13 C 3 ]pyruvate and [U-13 C 3 ]lactate ought to be supplied at a high enrichment to label significantly the acetyl and OAA moiety of citrate; the enrichment of [1-13 C]octanoate could be lowered to 25 or 50%.
In conclusion, we developed a strategy to assess directly and simultaneously the relative contributions of pyruvate carboxylation, pyruvate decarboxylation, and fatty acid oxidation to citrate formation in perfused rat heart. This requires perfusing hearts with a mix of 13 C-substrates and determining the 13 C labeling pattern of citrate. The use of effluent citrate instead of tissue citrate allows probing substrate fluxes non-invasively through the various reactions in hearts perfused under various conditions. The utility of this method was demonstrated for hearts perfused under normoxia with [U-13 C 3 ](lactate ϩ pyruvate) and [1-13 C]octanoate. The methodology should also be applicable to hearts perfused with other 13 C-substrates, such as [1-13 C]labeled long chain fatty acid, and under various conditions, provided that assumptions on which equations are developed are valid.