METHODS

Materials.

[2-¹³C]acetate sodium salt (99%) was purchased from Cambridge Isotope Laboratories (Andover, MA). AOA (hemihydrochloride), Dowex 50W resin (100–200 mesh, hydrogen form), and Dowex 1-X8 resin (100–200 mesh, chloride form) were purchased from Sigma Chemical (St. Louis, MO). All other reagents from commercially available sources were of the highest quality available. Male Sprague-Dawley rats, 250–300 g, were purchased from Sasco (Houston, TX).

Heart perfusions.

Rats were anesthetized in an ether atmosphere, and hearts were rapidly excised and placed in an ice-cold perfusion medium. The aorta was immediately cannulated, and hearts were perfused using standard Langendorff methods at a column height of 70 cmH₂O. Typical coronary flow rates were 15–18 ml/min. A modified Krebs-Henseleit (KH) buffer containing 119.2 mM NaCl, 4.7 mM KCl, 1.25 mM CaCl₂, 1.2 mM MgSO₄, and 25 mM NaHCO₃ was bubbled with 95% O₂-5% CO₂. The entire recirculation system was jacketed and maintained at 37°C. After an initial washout period of 10–15 min, hearts were perfused with 300 ml of recirculating KH buffer for 30 min before addition of 2 mM [2-¹³C]acetate. In the inhibitor experiments, AOA was present during this entire 30-min period to ensure equilibration of the inhibitor into all cellular compartments. After addition of [2-¹³C]acetate, 14 3-min proton-decoupled ¹³C spectra were collected during the approach to isotopic steady state (42 min). Hearts were then freeze-clamped, extracted with 3.6% cold perchloric acid, neutralized with KOH, freeze-dried, and dissolved in 0.6 ml of deuterated water (²H₂O) for NMR analysis and glutamate assays.

NMR.

Proton-decoupled ¹³C NMR spectra were obtained at 125.7 MHz on a GN-500 spectrometer. Intact heart spectra were collected in an 18-mm thin-walled NMR tube (Wilmad) with the heart bathed in perfusate, as previously described (15). The temperature of the heart was maintained at 37°C by control of the circulation perfusate temperature and by temperature control of the air surrounding the 18-mm NMR tube in the magnet using the GE VT accessory. A spherical bulb (∼100 μl) containing [3-¹³C]propionate positioned near the heart provided an external concentration and chemical shift standard. Typically, spectra were signal averaged over periods of 3 min using a 45° observe pulse, a sweep width of ±14,000 Hz, 16K data points, and a 1-s delay between pulses and Waltz bilevel ¹H decoupling.

Spectra of heart extracts were obtained in a 5-mm tube using a 45° observe pulse, 16K data points over ±14,000 Hz, and a 6-s delay between pulses. Broad-band proton decoupling was achieved using Waltz decoupling at two power levels, and the temperature was maintained at 25°C. All ¹³C spectra were zero filled to 32K points before Fourier transformation. The relative areas of the multiplet components in each glutamate resonance were determined by General Electric integration software and applied to a steady-state isotopomer analysis (14).¹H NMR spectra of heart extracts were obtained on the same spectrometer using a presaturation pulse sequence to eliminate residual water.

Myocardial O₂ consumption and tissue assays.

Myocardial O₂ consumption was calculated from the difference in O₂ content of perfusion medium in the supply line and coronary effluent collected from the pulmonary artery as described previously (18). Total tissue glutamate was measured fluorometrically in tissue extracts (2). Homogenized tissue samples were prepared from isolated, perfused hearts by Polytron homogenization. Glutamic-oxalacetic transaminase [GOT; aspartate aminotransferase (AST);l-aspartate:2-oxoglutarate aminotransferase; EC 2.6.1.1] activity in homogenized heart tissue samples was assayed spectrophotometrically using techniques outlined by Bergmeyer and Bernt (see Ref. 2).

Isolation of glutamate.

Glutamate was isolated from rat heart extracts using two short columns, one cationic (Dowex 50W hydrogen column) and one anionic (Dowex 1-X8 formate column) to separate the strongly acidic and neutral amino acids. A column containing 6 ml of Dowex 50W in a 10-ml disposable syringe was washed with 50 ml of 2 N HCl followed by washing with distilled water (final pH of eluant was between 5 and 6). A second column containing 3 ml of Dowex 1-X8 was assembled using 0.6-cm-diameter glass Pasteur pipette. This column was washed with 20 ml of 2 M formic acid followed by 20 ml of distilled water (final pH of eluant was near 7). Neutralized heart extract samples were dissolved in 2 ml of distilled water at pH 2.5–3.0 (adjusted by HCl) and applied to the Dowex 50W column. The column was washed with 25 ml of distilled water to remove carboxylic acids followed by 30 ml of 2 M NH₄OH to recover all amino acids. The amino acid fraction was freeze dried and redissolved into 2 ml of distilled water, and the pH was adjusted to 8 using KOH. This fraction was applied to the Dowex 1-X8 column and washed with 200 ml of distilled water to remove neutral amino acids and glycerol. Glutamate was eluted with 10 ml of 0.5 M formic acid, freeze dried to remove excess formic acid, and redissolved into 0.5 ml²H₂O for¹H NMR analysis.

Modeling.

A kinetic model similar to that of Chance et al. (3) was constructed. Single pools of citrate, α-KG, succinate, malate, oxalacetate, and glutamate were included in reactions that involved the Krebs TCA cycle, exchange between α-KG and glutamate, and anaplerosis. Differential equations describing the time-dependent changes in concentration of the individual ¹³C isotopomers in each metabolite pool were solved numerically. This kinetic model was used to fit the ¹³C fractional enrichment curves of glutamate C4 and C3 to optimal values ofV_x andV_TCA (whereV_x is transaminase flux, mitochondrial export flux, or some combination thereof and V_TCAis TCA cycle flux) and to simulate time-dependent changes in the glutamate spectrum (both C3 and C4 fractional enrichments and C4 multiplet areas) as a function ofV_x/V_TCAand total TCA cycle intermediate pool size.

Statistical analysis.

All results are reported as means ± SD. The Student’st-test was used to compare means;P < 0.05 was considered significant.

RESULTS

Rates of ¹³C enrichment in glutamate with or without AOA.

Temporal changes in proton-decoupled¹³C NMR spectra of isolated, perfused rat hearts in the absence and presence of 0.5 mM AOA are compared in Fig. 1. The average glutamate C3 and C4 resonance areas for five or six hearts in each group are shown plotted vs. time in Fig. 2. There are three obvious differences between these data: first,¹³C enrichment of glutamate C4 and C3 occurred more slowly in the heart perfused with the transaminase inhibitor; second, there was less glutamate detected by¹³C NMR at steady state in the hearts perfused with AOA (intensities of glutamate resonances in these plots were standardized relative to an external [3-¹³C]propionate reference at 11 ppm); and third, the difference in rates of¹³C incorporation into glutamate C4 and C3 was much less in the presence of inhibitor. Homogenized tissue from hearts perfused with 0.5 mM AOA had 60% less GOT activity than control hearts (in vitro assays), whereas coronary flow, heart rate, developed pressure, and O₂consumption (in vivo assays) were unaffected by the presence of the inhibitor (see Table 1).

In separate experiments, hearts were perfused with even higher concentrations of AOA to determine whether the kinetics and steady-state levels of¹³C-enriched glutamate could be further depressed. Heart function (as detected by rate-pressure product) was not significantly different in groups perfused with 0, 0.5, 1, or 2 mM AOA. The rates of glutamate enrichment with¹³C did appear to slow even further with increasing concentrations of AOA (spectra not shown), but less ¹³C-enriched glutamate was also detected by NMR at the higher inhibitor concentrations, and this precluded an accurate assessment of¹³C resonances areas.

¹³C NMR spectra of intact hearts at steady state.

Figure 3Bcompares glutamate signal intensities in three different intact hearts after perfusion to steady state with [2-¹³C]acetate and 0.5, 1, or 2 mM AOA. Although the amount of¹³C-enriched glutamate in these three hearts appeared to be lower at the high levels of AOA, slight variations in heart size could have contributed to these differences. Thus we also compared glutamate intensities from a single heart, before and after exposure to AOA. The stacked plot shown in Fig.3A compares the glutamate resonance intensities in a heart first perfused to steady state with [2-¹³C]acetate and then every 3 min after addition of 0.5 mM AOA (with [2-¹³C]acetate still present). The temporal intensity changes detected in all three glutamate resonances after addition of AOA showed quite clearly that less glutamate was detected by ¹³C NMR when the inhibitor was present.

¹³C NMR spectra of extracts of tissue at isotopic steady state.

The observations described above from intact, perfused hearts suggested that either 1) less [2-¹³C]acetate entered the TCA cycle as [2-¹³C]acetyl-CoA when the transaminase inhibitor was present,2) addition of inhibitor resulted in a decrease in total tissue glutamate, or3) glutamate was sequestered in some cellular compartment in intact hearts perfused with AOA that rendered it at least partially invisible by¹³C NMR. High-resolution¹³C spectra of extracts of hearts perfused to steady state with or without AOA were not significantly different (not shown). A steady-state isotopomer analysis (14) of those spectra indicated that the fraction of acetyl-CoA derived from [2-¹³C]acetate (F_C2) and relative anaplerosis (y) was identical in the presence and absence of inhibitor (Table 2). Total tissue glutamate (as measured enzymatically and from¹³C NMR spectra of extracts) was also independent of inhibitor. ¹H NMR spectra of isolated, purified glutamate from these tissues provided an independent measure of the ¹³C fractional enrichment glutamate C4. Those values, reported in the last column of Table 2, were significantly less than F_C2 as measured by¹³C NMR. This indicated that there was a small pool of glutamate in these hearts (∼20% of total glutamate) that was not in exchange with TCA cycle intermediates and that the size of this nonexchanging pool was not affected by the presence of AOA (Table 2).

Time-dependent evolution of glutamate multiplets.

¹³C NMR spectra such as those shown in Fig. 1 had sufficiently good signal-to-noise to allow deconvolution of the singlet (S) and doublet (D34) components of glutamate C4. These data, shown in Fig. 4, indicate there was a dramatic difference in the temporal evolution of C4S and C4D34 (and other glutamate resonance multiplets as well; not shown) in hearts perfused with or without AOA. Although C4S was high and C4D34 low in the first 3-min spectrum of hearts in the absence of AOA, C4D34 was already higher than C4S in the first 3-min spectrum of hearts exposed to AOA. This was confirmed in extracts of hearts freeze-clamped at 5 min after addition of [2-¹³C]acetate with and without AOA. The glutamate C4 resonance from two representative spectra are shown in Fig. 5. The average C4D34/C4S from four such experiments was 0.64 ± 0.11 in hearts without AOA and 1.30 ± 0.12 in hearts with AOA. Because O₂ consumption and hence TCA cycle flux were identical in the two groups of hearts, the larger C4D34/C4S and the faster enrichment of glutamate C3 (relative to C4, see Fig. 2) in the AOA group suggested that the size of the TCA cycle intermediate pools in exchange with glutamate was smaller in hearts exposed to AOA than in control hearts. This was confirmed by¹H spectroscopy of these same samples (Fig. 5). The larger ¹³C satellite wings around the glutamate H4 resonance centered at 2.34 ppm verified that a larger fraction of total tissue glutamate had exchanged with α-KG in hearts without AOA. The average C4F at 5 min was 0.65 ± 0.05 in hearts without AOA vs. 0.22 ± 0.01 in hearts with AOA (n = 4 for each group). This confirmed that the kinetics of exchange among TCA cycle intermediate pools were different in the two experiments even though glutamate reached the same level of ¹³C fractional enrichment and isotopomeric composition after a period of 40–50 min in both groups of hearts (compare F_C2values for 2 groups at steady state; Table 2).

DISCUSSION

Dynamic NMR measurements of ¹³C enrichment in glutamate C3 and C4 offer considerable potential for estimating TCA cycle flux in vivo; yet the assumptions involved in such measurements and their validity have generated considerable scientific debate. The first concern is whether the rate of¹³C appearance in glutamate is limited to any significant extent by an exchange [α-KG↔glutamate (Glu)] or transport (Mito↔Cyto) process. In an early kinetic study using¹³C fractional enrichment data derived from spectra of tissue extracts, Chance et al. (3) estimated that exchange between α-KG and glutamate was about threefold faster than TCA cycle flux in hearts perfused with acetate. They concluded that the rate of exchange between cytosolic and mitochondrial metabolite pools was fast relative to other fluxes and that the size of each metabolite pool in exchange in the cycle was equal to that measured enzymatically in tissue extracts. A few years later, the first in vivo NMR demonstration of ¹³C kinetics to measure TCA cycle flux was elegantly demonstrated by Fitzpatrick et al. (5). They fitted the¹³C enrichment curves of glutamate C4 and C3 in rat brain obtained by¹H observe-¹³C edited spectroscopy to a kinetic model that also assumed that the entire cerebral glutamate pool was in rapid exchange with α-KG. This gave a reasonable 1.4 μmol ⋅ g⁻¹ ⋅ min⁻¹value for TCA cycle flux, in agreement with values determined using other techniques. This kinetic model was later extensively validated (16, 17) and applied to ¹³C NMR data collected from human brain during infusion of [1-¹³C]glucose (6). It was concluded that exchange between α-KG and glutamate (V_x) was 72 times faster thanV_TCA in human brain (16).

More recent ¹³C NMR kinetic investigations on isolated, perfused hearts have reported that the rates at which ¹³C appears in glutamate C3 and C4 may be influenced by processes other than TCA cycle flux (4, 28). Chatham et al. (4) followed¹³C incorporation into glutamate C3 and C4 in isolated rat hearts perfused with [2-¹³C]acetate and fitted those data using a similar kinetic model as described earlier by Chance et al. (3). However, they found that α-KG↔Glu exchange was 18% that of TCA cycle flux, a result quite different from the earlier Chance et al. report, in which the data were derived from spectra of tissue extracts. Somewhat later, Yu et al. (28) applied a mathematically simplified kinetic model to¹³C fractional enrichment data from intact rabbit hearts perfused with [2-¹³C]acetate and reported that two rate-limiting processes, TCA cycle flux (i.e.,V_TCA) and α-KG↔Glu exchange (F₁), contribute nearly equally to the glutamate enrichment kinetics. A comparison of their kinetically determined F₁ flux variable with estimates of total cytosolic transaminase flux (GOT_Cyto) from in vitro assays led them to conclude that flux through GOT_Cyto was much too high to contribute to the NMR observables, and thus F₁ probably reflects some transport process, perhaps transport of α-KG from the mitochondria to the cytosol. No attempt was made to estimate mitochondrial transaminase flux (GOT_Mito) in that study probably due to the uncertainties in the distribution of α-KG, glutamate, aspartate, and oxalacetate among the cytosolic and mitochondrial compartments. Finally, Weiss et al. (25), using a kinetic model similar to that used by Chance et al. (3) and Chatham et al. (4), reported that α-KG↔Glu exchange via transamination was nearly equal to TCA cycle flux in isovolumic, paced rat hearts perfused with [2-¹³C]acetate. In each of these studies, total tissue glutamate, as measured enzymatically (3, 28) or as detected by¹³C NMR (4, 25), was used to fit the kinetic data. Thus four independent¹³C kinetic modeling studies of hearts perfused with [2-¹³C]acetate have yielded somewhat disparate results. Chance et al. (3) concluded that flux through the transaminases was a factor of 3 larger than TCA cycle flux, Chatham et al. (4) concluded that transaminase flux (as part of malate-aspartate shuttle) was smaller than TCA cycle flux, and Weiss et al. (25) concluded that transaminase flux and TCA cycle flux were comparable, whereas Yu et al. (28), on the basis of direct experimental evidence for high GOT_Cyto flux, suggested that flux of α-KG from the mitochondria to the cytosol (but not transaminase flux) was comparable to TCA cycle flux.

A kinetic model (Jeffrey, unpublished results) similar to that of Chance et al. (3) was used to fit the¹³C fractional enrichment data of Fig. 2 to obtain best values ofV_x (flux through an undefined process involving interchange of α-KG and glutamate carbons) and V_TCAfor both groups of hearts (AOA). The solid lines drawn through the data of Fig. 2 represent the best fit of the data to a model that fixed F_C2 at 0.93,y at 0.05, total tissue glutamate at 22 μmol/g dry wt, and the remaining TCA cycle intermediates at values reported by Chance et al. (3) for acetate-perfused hearts. The fit of the data in the absence of inhibitor gaveV_x = 7.5 μmol ⋅ min⁻¹ ⋅ g dry wt⁻¹ andV_TCA = 7.5 μmol ⋅ min⁻¹ ⋅ g dry wt⁻¹, quite similar to the values reported by Yu et al. (28) for acetate-perfused rabbit hearts. Although our fitted value ofV_TCA was somewhat lower than that predicted by our O₂ consumption measurements (expected value was 19/2 = 9.5 μmol ⋅ min⁻¹ ⋅ g dry wt⁻¹), the agreement is reasonable considering the assumptions involving the sizes of the all TCA cycle intermediate pools. Importantly, our data in the absence of inhibitor indicate thatV_x ≈V_TCA, in agreement with Weiss et al. (V_x/V_TCA= 0.86; Ref. 25) and Yu et al. (V_x/V_TCA= 0.92; Ref. 28) but clearly different from the conclusions reached by Chance et al. (V_x/V_TCA= 2.8; Ref. 3) and Chatham et al. (V_x/V_TCA= 0.18; Ref. 4).

Effects of AOA on pre-steady-state ¹³C enrichment of glutamate C3 and C4.

We have shown in this study that an inhibition of total cellular transaminase activity by 60% (as measured in vitro) has a dramatic effect on the rates of ¹³C appearance in glutamate C3 and C4. Inasmuch as flux through GOT_Cyto has been estimated at 20 times greater than TCA cycle flux in isolated rabbit hearts perfused with acetate (28), how can AOA at these levels have such a dramatic effect on the rates of ¹³C enrichment at glutamate C3 and C4? If GOT_Cyto flux is indeed much greater than TCA cycle flux in the rat heart, then one must consider a possible role of GOT_Mito in this process. Using the kinetic parameters for GOT_Mito and metabolite concentrations reported by Yu et al. (28) and making the assumption that 20% of total tissue aspartate and α-KG is mitochondrial and 10% of total tissue water is mitochondrial, one can arrive at an estimate of 90 μmol ⋅ min⁻¹ ⋅ g dry wt⁻¹ for GOT_Mito flux. Clearly, this estimate may suffer from the likely oversimplifying assumption that the kinetic parameters of GOT_Mito in situ are the same as those measured in vitro (28). Nevertheless, the calculation suggests that neither GOT_Cyto nor GOT_Mito may limit the rate of¹³C appearance in glutamate C3 and C4 in the absence of inhibitor. If one further assumes that AOA does nothing more than inhibit both GOT_Cyto and GOT_Mito by 60%, then we should not have detected the dramatic changes in the rate of¹³C appearance in glutamate C3 and C4 shown in Fig. 2. This suggests that AOA must be affecting other metabolic processes in addition to partial inhibition of the transaminases. This conclusion was supported in our attempts to fit the¹³C fractional enrichment data for hearts perfused with 0.5 mM AOA (Fig.2A) using the same kinetic model as described above. The solid lines through the inhibitor data of Fig. 2show the best fit using the same assumptions as stated above plusV_TCA fixed at 7.5 μmol ⋅ min⁻¹ ⋅ g dry wt⁻¹. AlthoughV_x tended to decrease in all calculations (curves shown are for a value of 5.8 μmol ⋅ min⁻¹ ⋅ g dry wt⁻¹), the poor agreement between experimental and calculated data indicates that this kinetic model is too simplistic. Again, the near coincidence of the C4 and C3 enrichment curves indicates that much smaller pool(s) of intermediates were in rapid exchange with the TCA cycle reactions in this group of hearts (see modeling results below).

Weiss et al. (25) also examined the effects of AOA (at a 5-fold lower concentration) on the ¹³C enrichment curves of glutamate but did not observe coalescence of the C4 and C3 enrichment curves or a change in the total amount of glutamate detected by ¹³C NMR. Interestingly, they reported that 0.1 mM AOA inhibited 90–95% of the total aspartate aminotransferase activity in heart tissue (as assayed in vitro); yet a fit of their¹³C NMR dynamic data to their kinetic model indicated that transaminase flux was reduced by only 50% in vivo (25). Although Weiss et al. ascribed this to differences between in vitro and in vivo enzyme kinetics, the rather substantial differences between their results and ours indicate that the same level of inhibition may not have been achieved in the two experiments. The data presented here indicate that the empirical method advocated by Weiss et al. (24), which uses the Δt₅₀ for C4 vs. C3 as an index of TCA cycle flux, cannot always be used with confidence, even if a change in¹³C NMR-detected metabolite levels is considered. For example, in our experiments, Δt₅₀ was very small in the presence of 0.5 mM AOA, and yet O₂ consumption (hence TCA cycle flux) was unchanged.

Evolution of glutamate multiplets with time.

We also demonstrated for the first time that the evolution of glutamate¹³C isotopomers, as reported by time-dependent evolution of C4S and C4D34, provides additional kinetic evidence about exchanging pools that is not easily detected in¹³C fractional enrichment measurements. To illustrate this point, we have used our kinetic model to simulate the changes expected in the shape of the¹³C enrichment curves of glutamate C4 and C3 as α-KG↔Glu exchange (V_x) is reduced at a fixed V_TCA. Figure 6 shows the resulting simulations at two intermediate pool sizes (differing by a factor of 10) and for values ofV_x/V_TCA= 2 and 0.1, withV_TCA held constant. As originally predicted Mason et al. (17) and demonstrated here experimentally, the simulation shows that the intensities of glutamate C3 and C4 should approach one another asV_x ≪V_TCA. This is quite intuitive. By slowing communication (exchange) between a relatively small pool of TCA cycle intermediates and the relatively large glutamate pool, we have effectively increased the rate at which¹³C reaches the C3 position relative to C4 without increasingV_TCA. A comparison of Fig. 6, left andright, indicates that the shapes of the C3 and C4 kinetic curve are relatively insensitive to the size of the TCA cycle intermediate pool undergoing rapid¹³C turnover, both forV_x/V_TCA= 2 (simulating our control hearts) and forV_x/V_TCA= 0.1 (simulating hearts inhibited by AOA). This same conclusion was also reached by Chatham et al. (4) and Yu et al. (28) for control hearts. Because the curves shown in Fig.6A for both sets of conditions did not change as the total TCA cycle intermediate pool size was decreased by a factor of 10, we must conclude that it is not possible to judge the size of the exchanging pools based upon C4 and C3 fractional enrichment data alone. However, this situation is quite different for the glutamate multiplets. Figure 6, bottom, illustrates that the glutamate C4 multiplet, C4D34, is quite sensitive to bothV_x/V_TCAand intermediate pool sizes. In particular, C4D34 reaches its maximum much more rapidly whenV_x/V_TCA and the combined size of the TCA cycle intermediate pools are both small.

The experimental data of Fig. 4 show that C4D34 is small early during isotopic enrichment of glutamate in the absence of AOA and the shape of its temporal evolution is similar to the solid curves shown in Fig. 6,bottom, forV_x/V_TCA= 2. Experiments performed in the presence of inhibitor, however, show that C4D34 is significantly larger not only during early glutamate turnover but throughout its approach to isotopic steady state (Fig. 4). This could only be simulated by allowing the total size of the TCA intermediate pools in exchange within the cycle to decrease significantly (dotted line in Fig. 6B,bottom). Thus AOA appears to not only decrease V_xbut also the size of intermediate pools undergoing¹³C turnover in the cycle. Although the simulation did not reproduce some of the finer experimental details (in particular, experimental C4D34 and C4S values were relatively constant during early turnover and then gradually changed between 20 and 40 min), it does illustrate that C4D34 can quickly rise above C4S when TCA cycle intermediate pools are small and exchange into the much larger glutamate pool is slow compared with TCA cycle flux. Thus the standard assumption that the size of the intermediate pools in exchange in the cycle is equal to the total metabolite pools measured in extracts may not always be valid and any kinetic model based on that assumption should be used with the present observations in mind.

NMR visibility of glutamate.

This study has shown that the NMR visibility of glutamate can be altered by a common metabolic inhibitor, without changing total tissue glutamate, heart rate, developed pressure, or O₂ consumption. Safer et al. (21) have also reported that 0.4 mM AOA does not markedly affect O₂ consumption by hearts but does alter the cytosolic-mitochondrial distribution of several metabolites associated with the malate-aspartate shuttle. It has been shown by Safer et al. in heart (21) and Kauppinen et al. in brain (7) that AOA at these levels increases the ratio of cytosolic NADH/NAD⁺, which in turn increases cytosolic malate and decreases cytosolic aspartate. This would in turn induce aspartate efflux from mitochondria and glutamate influx into mitochondria (21), perhaps leading to a net redistribution of glutamate toward the mitochondrial compartment at equilibrium. Our observation that the ¹³C-enriched glutamate signal decreases by ∼30% in hearts perfused with AOA is entirely consistent with these prior metabolic observations. This is important because kinetic models generally assume that all glutamate in exchange with the TCA cycle is NMR visible and that it equals total tissue glutamate as measured in extracts. In most cases, this appears to be a reasonable assumption (4, 16, 17, 24, 28), but our study demonstrates that there may be metabolic situations in which glutamate can become partially NMR invisible, perhaps by redistributing among subcellular compartments.

Summary.

Dynamic ¹³C NMR studies of hearts perfused with [2-¹³C]acetate with or without the transaminase inhibitor, AOA, have shown that this inhibitor has multiple effects on metabolism in the heart. We have shown that as little as 0.5 mM AOA inhibits total transaminase activity by 60% and effectively slows communication between the mitochondrial and cytosolic spaces of the myocardium without altering TCA cycle flux. One unanticipated finding was that AOA also altered the amount of glutamate detected by NMR. The NMR data (both extract spectra and in vivo spectra) indicate that a small exchanging pool of metabolites turns over quite rapidly in the presence of inhibitor. This small, presumably mitochondrial, pool then exchanges with mitochondrial and cytosolic glutamate (Glu_Mito and Glu_Cyto, respectively) at rates determined by residual transaminase activity and transport processes. This was reflected in an unusually high C3/C4 and unusally large C4D34 early during cycle turnover. The sizes of Glu_Mito and Glu_Cyto cannot be firmly established by our experimental data nor can we determine whether multiple subcompartmental pools exist within the cytosolic and mitochondrial spaces. Glu_Mito has been reported to be as high as 30% (26) and as low as 10% (8) of total tissue glutamate in normoxic hearts; so if our NMR observations in the presence of AOA indeed reflect a redistribution of ∼20% of Glu_Cyto to Glu_Mito, most of the total tissue glutamate would still remain in the cytosol. Given that the spin-lattice relaxation times and nuclear Overhauser enhancements of¹³C-enriched glutamate in hearts perfused with [2-¹³C]acetate are nearly identical to those measured in aqueous saline at the same temperature (22), we assume that all glutamate detected by¹³C NMR in vivo is Glu_Cyto and that Glu_Mito could be either invisible or less visible due to slower diffusion of molecules in the highly viscous mitochondrial matrix (23). Furthermore, our observation that ∼20% of total cellular glutamate is totally sequestered and never becomes enriched with ¹³C in acetate-perfused hearts (in agreement with previous reports; Refs. 11,12) illustrates that multiple glutamate pools can indeed exist in hearts, so that dividing total glutamate into further “kinetic” subcellular compartments may not be totally unreasonable.