The Anaplerotic Substrate Alanine Stimulates Acetate Incorporation into Glutamate and Glutamine in Rabbit Kidney Tubules

Although acetate, the main circulating volatile fatty acid in humans and animals, is metabolized at high rates by the renal tissue, little is known about the precise fate of its carbons and about the regulation of its renal metabolism. Therefore, we studied the metabolism of variously labeled [13C]acetate and [14C]acetate molecules and its regulation by alanine, which is also readily metabolized by the kidney, in isolated rabbit renal proximal tubules. With acetate as the sole substrate, 72% of the C-1 and 49% of the C-2 of acetate were released as CO2; with acetate plus alanine, the corresponding values were decreased to 49 and 25%. The only other important products formed from the acetate carbons were glutamine, and to a smaller extent, glutamate. By combining 13C NMR and radioactive and enzymatic measurements with a novel model of acetate metabolism, fluxes through the enzymes involved were calculated. Thanks to its anaplerotic effect, alanine caused a stimulation of acetate removal and a large increase in fluxes through pyruvate carboxylase, citrate synthase, and the enzymes involved in glutamate and glutamine synthesis but not in flux through α-ketoglutarate dehydrogenase. We conclude that the anaplerotic substrate alanine not only accelerates the disposal of acetate but also prevents the wasting of the latter compound as CO2.

Acetate is the main circulating volatile fatty acid in humans and other mammalian species. Its blood concentration is low (less than 0.2 mM) in fed and starved humans and starved herbivores but may reach the millimolar range in humans after alcohol consumption and in fed herbivorous species (1)(2)(3)(4)(5)(6)(7). The sources of blood acetate are on the one hand absorption of the acetate formed as a result of gastrointestinal bacterial fermentation, and on the other hand, the acetate formed and released by various tissues containing acetyl-CoA hydrolase activity (2,3,6,8).
On the basis of experiments performed in vivo with labeled acetate, it has been shown that the turnover of circulating acetate is rapid and that, depending on the species and nutritional state, the oxidation of this compound provides from 6 to 70% of the whole body energy expenditure (3,4,7,9). This means that acetate is removed and metabolized by peripheral tissues. Indeed, acetyl-CoA synthetase, the enzyme that initiates acetate degradation, has been demonstrated to be active in many tissues including the liver, kidney, heart, brain, adipose tissue, and skeletal muscle (3). It has been found that, besides the heart, the kidney contains a high activity of this enzyme (3).
In agreement with this observation, we have shown in a recent study that acetate is readily metabolized by suspensions of rabbit renal proximal tubules (10). In the same study (10), we have demonstrated that acetate significantly altered the metabolism of alanine, a major precursor of glutamine in these tubules. For this, we used 13 C-labeled alanine and unlabeled acetate in combination with enzymatic and 13 C NMR spectroscopy measurements to calculate metabolic fluxes related to alanine metabolism.
In an attempt to identify precisely the metabolic fate of acetate carbons and gain insight into the reciprocal effect of alanine on acetate metabolism, we have conducted concomitantly a study in which we incubated rabbit renal proximal tubules with 13 C-labeled acetates in the absence and the presence of unlabeled alanine. Thanks to the development of a novel model of acetate metabolism that is of general use (see "Appendix") and to the combination of enzymatic, radioactive, and 13 C NMR measurements, we were able to estimate fluxes through enzymes involved in acetate metabolism in rabbit renal proximal tubules. We showed that the addition of alanine, which increased the removal of acetate, also stimulated its metabolism through citrate synthase but not through ␣-ketoglutarate dehydrogenase. We also demonstrated that acetate carbons were converted to different extents not only into CO 2 but also into glutamate and glutamine, especially in the presence of the anaplerotic substrate alanine.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Preparation of Kidney Tubules and Incubation
Kidney cortex tubules were prepared by the treatment of renal cortex slices with collagenase as described by Baverel et al. (11). Incubations other than those involving radioactive substrates were performed for 60 min at 37°C in a shaking water bath in 25 ml stoppered Erlenmeyer flasks in an O 2 /CO 2 (19/1) atmosphere. The flasks contained 1 ml of the suspension plus 3 ml of Krebs-Henseleit medium (12), either supplemented or not with substrates, i.e. 5 mM (final concentration) [1-14 C]acetate or [2-14 C]acetate (10 3 Bq/flask), [1-13 C]acetate or [2][3][4][5][6][7][8][9][10][11][12][13] C]acetate in the absence and the presence of 5 mM L-alanine. These differently labeled acetates were used in an attempt to completely define the fate of the two acetate carbons. In all experiments, each experimental condition was performed in quadruplicate. Incubation was stopped by the addition of HClO 4 (final concentration 2% (v/v)) to each flask. In all experiments, zero-time flasks, with and without substrates, were prepared by the addition of HClO 4 before the tubules.
When radioactive acetate was present in the medium, incubation, deproteinization, collection, and measurement of the 14 CO 2 formed were performed as described by Baverel and Lund (13). After removal of the denaturated protein by centrifugation, the supernatant was neutralized with a mixture of 20% (w/v) KOH and 1% (v/v) H 3 PO 4 (8 M) for metabolite determination and NMR spectroscopy.

Analytical Methods
Metabolite Assays-Lactate, pyruvate, glucose, glutamate, glutamine, ammonia, alanine, citrate, ␣-ketoglutarate, fumarate, malate, acetate, acetoacetate, and 3-hydroxybutyrate as well as the dry weight of tubules added to the flasks were determined as described previously (11,13). Serine was measured by high pressure liquid chromatography with the use of the Pico-Tag method (14). 13 C NMR Techniques-Perchloric acid extracts were neutralized (11), freeze-dried, and reconstituted in D 2 O in the presence of [2 13 -C]glycine as internal standard. Data were recorded as described previously (15,16) at 100.6 MHz on a Bruker AM-400 WB spectrometer with a 10-mm broadband probe thermostatically controlled at 8 Ϯ 0.5°C. Acquisition parameters were as follows: spectral width, 25,000 Hz; tilt angle, 90°; data size, 32,000; repetition time, 30 s; number of scans, 2700; Proton decoupling was carried out during the data acquisition (0.65 s) using a standard WALTZ 16 pulse sequence for inverse-gated proton decoupling (17). Assignments were made by comparing the chemical shifts obtained with published values (18,19).

Calculations
Net substrate utilization and product formation were calculated as the difference between the total flask contents (tissue plus medium) at the start (zero-time flasks) and after the period of incubation. The metabolic rates, reported as means Ϯ S.E., are expressed in mol of substances removed or produced per flask per unit time (60 min).
The rates of release of 14 CO 2 from the 14 C-labeled acetate species used were calculated by dividing the radioactivity in 14 CO 2 by the specific radioactivity of the labeled carbon of the acetate species of interest measured in each medium. When [ 13 C]acetate species were the substrate, the transfer of the C-1 or C-2 of acetate to a given position in a given metabolite was calculated from (L m Ϫ l m )/(A s Ϫ a s ), in which L m is the amount of 13 C measured in the corresponding NMR resonance, l m is the natural abundance (1.1%) multiplied by the amount of metabolite assayed enzymatically, A s is the total 13 C abundance of the C-1 or C-2 of acetate, and a s is the natural 13 C abundance.

RESULTS
To determine the fate of acetate carbons and the metabolic pathways involved, experiments were performed in which rabbit renal proximal tubules were incubated with differently 14 C-and 13 C-labeled acetates with and without alanine. Substrate utilization and product formation were measured by combining enzymatic, radioactive, and 13 C NMR spectroscopy measurements.
Enzymatic Measurements of Substrate Utilization and Product Formation-Substrate removal and glutamate plus glutamine accumulation and labeling were approximately linear with time over a 60-min incubation period (n ϭ 2 in duplicate; results not shown). Table I shows that, when acetate was added as the sole exogenous substrate into the incubation medium, it was readily removed by rabbit renal proximal tubules. Al-though under this condition, no exogenous nitrogenous substrate was provided to the renal cells, they accumulated substantial amounts of glutamine, and to a lesser extent, of glutamate and serine. The fact that substantial amounts of glutamine were also synthesized in the absence of added substrate (Table I) means that, in the presence of acetate, these amino acids were formed, at least in part, from endogenous sources. Some acetoacetate was also formed in the presence of acetate as the sole exogenous substrate (Table I).
As shown in Table I, and in agreement with our recent results (10), alanine was also metabolized at high rates by the renal tubules when this amino acid was added as substrate together with acetate. Alanine addition caused a 32% stimulation of acetate removal and a great increase in glutamate and glutamine accumulation. In the presence of alanine, acetoacetate accumulation was also significantly reduced, and small amounts of pyruvate and lactate accumulated. Nitrogen balance calculations indicate that, in the presence of alanine, no significant room was left for glutamine and glutamate synthesis from endogenous substrates. Under none of the experimental conditions studied did we observe any substantial accumulation of glucose, ammonia, ␤-hydroxybutyrate, or intermediates of the tricarboxylic acid cycle.
Radioactive Measurements of CO 2 Production from Acetate- Table II shows that, with acetate as the sole substrate, 72% of the C-1 and 49% of the C-2 of the acetate removed were released as CO 2 . This clearly indicates that, under this condition, the remainder of the C-1 and C-2 of the acetate removed was incorporated into the non-volatile carbon products found to accumulate, namely acetoacetate, glutamine, and glutamate.
The release of 14 CO 2 from [1-14 C]acetate did not change upon the addition of alanine; under this condition, it represented only 49% of the C-1 of the acetate removed. By contrast, the presence of alanine significantly diminished the production of 14 CO 2 from [2-14 C]acetate; under the latter condition, the production of CO 2 accounted for only 25% of the C-2 of the acetate removed. 13 C NMR Spectroscopy Measurements- Fig. 1, A and B, shows the 13 C NMR spectra of perchloric acid extracts obtained after 60 min of incubation of renal tubules with [2-13 C]acetate in the absence and the presence, respectively, of alanine. As all the C-1 and C-2 of the 14 C-labeled acetates removed could not be accounted for by the production of 14 CO 2 (Table II), it is not surprising that a significant amount of the C-2 of [2-13 C]acetate removed was recovered in glutamine and glutamate, especially in the presence of alanine. No substantial amount of lactate, acetoacetate, or serine was found to be labeled. Using these spectra and those obtained with [1-13 C]acetate as substrate without and with alanine (results not shown), we calculated the amount of labeled products after correction for the 13 C natural abundance (Tables III and IV).
With [2-13 C]acetate as substrate (Table III), the high labeling of the C-4 of glutamate and glutamine indicates that the C-2 of acetate gave the C-2 of acetyl-CoA, and then gave the C-4 of citrate, ␣-ketoglutarate, glutamate, and glutamine, via the successive operation of acetyl-CoA synthetase, citrate synthase, aconitase, isocitrate dehydrogenase, alanine, or (in the absence of alanine) other amino acid aminotransferases and glutamine synthetase. The fact that virtually equal amounts of C-2 and C-3 of glutamate plus glutamine were labeled is consistent with the previous observations made by other authors with acetate and other substrates in kidney and other tissues (20 -24). This is also in agreement with the view that the C-2 of acetate passed through succinate and fumarate, two symmetrical molecules, during the first tricarboxylic acid cycle turn, leading to the formation of either [2-13 C]oxalacetate or [3-13 C]oxalacetate and then to either [3-13 C]citrate or [2-13 C]citrate and ␣-ketoglutarate during the second tricarboxylic acid cycle turn, yielding finally glutamate and glutamine labeled on their C-2 and C-3 after the transaminase and glutamine synthetase reactions. Consistent with the increase in the synthesis of glutamine plus glutamate shown in Table I, the addition of alanine caused an increase in the incorporation of the C-2 of acetate into the C-2, C-3, and C-4 of glutamate and glutamine (Table III).
It should be noted that a fraction of the glutamate and glutamine molecules formed from [2-13 C]acetate were simultaneously labeled on their C-3 and C-4 as revealed by the doublets that indicate 13 C-13 C couplings between these two glutamine and glutamate carbons (Fig. 1). However, as the spectral resolution was not sufficient to quantify these doublets in a reliable manner, no attempt was made to quantify them to determine the relative proportions of labeled and unlabeled oxalacetate and acetyl-CoA molecules contributing to the synthesis of glutamate and glutamine. Table IV shows as expected that, with [1-13 C]acetate as substrate and both in the absence and the presence of alanine, the labeling of the C-5 of glutamate and glutamine was close to the labeling of the C-4 of these two amino acids observed when [2-13 C]acetate was the substrate (see Table III for comparison). The labeling of the C-1 of glutamate and glutamine is in agreement with the conversion of the C-1 of acetate into the C-4 of oxalacetate during the first tricarboxylic acid cycle turn and then the formation of the C-1 of citrate and ␣-ketoglutarate during the second tricarboxylic acid cycle turn before the transamination of ␣-ketoglutarate into glutamate followed by glutamine synthesis. The fact that the labeling of the C-1 of glutamate and glutamine when [1-13 C]acetate was the substrate was smaller than the labeling of either the C-2 or the C-3 of glutamate and glutamine when [2-13 C]acetate was the substrate is not surprising because, in the presence of [1-13 C]acetate, the label found in the C-1 and the C-4 of oxalacetate after the first tricarboxylic acid cycle turn was lost as CO 2 during the second turn via the isocitrate dehydrogenase and the ␣-ketoglutarate dehydrogenase step, respectively. As already seen with the incorporation of the C-2 of acetate into the C-2, C-3, and C-4 of glutamate and glutamine (Table III), incorporation of the C-1 of acetate into the C-1 and C-5 of glutamate and glutamine was increased in the presence of alanine.
Calculations of Proportions- Table V shows the calculated proportions of metabolites converted into the next one(s). It should be mentioned here that, with the model used, these proportions allowed us to calculate enzymatic fluxes only when they were combined with the utilization of the substrate(s) of interest. As can be seen in Table V, some proportions could not be calculated when acetate was the sole substrate. This is because, under this condition, the 13 C resonances were not great enough to allow us to quantify the corresponding conversions. The fraction of pyruvate converted into oxalacetate,

Effect of 5 mM L-alanine on the metabolism of 5 mM acetate in rabbit kidney tubules
Kidney tubules (26.8 Ϯ 2.1 mg of dry weight/flask) were incubated for 60 min as described under "Experimental Procedures." Results (mol/h) for metabolite removal (Ϫ) or production are reported as means Ϯ S.E. for four experiments performed in quadruplicate. The paired Student's t test was used to measure the statistical difference against the control with acetate alone: *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001. The radioactivity and 13 C NMR data corresponding to these experiments are reported in Table 2 and Tables 3 and 4  4, glutamine C-4 (31.70); 5, glutamate C-4 (34.16); 6, glutamine C-2 (55.10); 7, glutamate C-2 (55.53). Numbers in parentheses indicate the chemical shifts in ppm referred to tetramethylsilane. An expanded view of the C 4 and C 3 resonances of glutamate and also glutamine including the coupling between C 4 and C 3 is shown.
which was close to unity in the absence of alanine, was significantly decreased in the presence of alanine, whereas a significant fraction of pyruvate was converted into acetyl-CoA only in the presence of alanine.
The decrease in the fraction of acetyl-CoA converted into acetoacetate caused by the addition of alanine was fully compensated by an increase in the acetyl-CoA converted into citrate. The addition of alanine also significantly increased the proportion of ␣-ketoglutarate converted into glutamate plus glutamine at the expense of the proportion of ␣-ketoglutarate converted into oxalacetate (Table V). Table V also shows that the proportions of the oxalacetate recycled at each turn of the three different cycles could be quantified only in the presence of alanine.
Enzymatic Fluxes- Table VI shows the absolute values of fluxes through enzymes involved in acetate metabolism both in the absence and the presence of alanine. Fluxes through acetyl-CoA synthetase were identical to the values of acetate removal reported in Table I. Fluxes through 3-ketothiolase were also identical to the acetoacetate accumulations presented in Table  I because there was no evidence of ␤-hydroxybutyrate accumulation. On the addition of alanine, flux through acetyl-CoA synthetase was stimulated, whereas that through 3-ketothiolase was inhibited. In the absence of alanine, the minimum flux through pyruvate carboxylase (1.35 mol/h; in Table VI, see {PC} Ϫ {PEPCK}) 1 logically matches the exit of ␣-ketoglutarate from the tricarboxylic acid cycle to form glutamate and glutamine (in Table VI, see Glx accumulated). It should be mentioned that the small 13 C labeling data obtained did not provide evidence for the existence of the OAA 3 P-enolpyruvate 3 Pyr A flux through pyruvate dehydrogenase was observed only in the presence of alanine; under this condition, the absence of labeling of the C-5 of glutamate and glutamine from [2-13 C]acetate indicates that pyruvate dehydrogenase metabolized to a significant extent only unlabeled pyruvate formed from the added unlabeled alanine. Similarly, flux through lactate dehydrogenase was derived from the accumulation of lactate shown in Table I because no 13 C-labeled lactate was found to accumulate.
Table VI also shows that the addition of alanine caused a doubling of flux through citrate synthase but no statistically significant increase in flux through ␣-ketoglutarate dehydrogenase because the increased synthesis of ␣-ketoglutarate was accompanied by an augmented conversion of ␣-ketoglutarate into glutamate and glutamine. This is in agreement with the increased labeling of glutamate and glutamine carbons from the 13 C-labeled acetates ( Fig. 1 and Tables III and IV). The increased accumulation of glutamate plus glutamine and the stimulation of flux through glutamine synthetase are also in agreement with the results presented not only in Tables I, III, and IV but also in Fig. 1.

DISCUSSION
Thanks to enzymatic, radioactive, and 13 C NMR measurements in combination with an original model of acetate metabolism that is of general use, this study not only establishes the fate of acetate in rabbit renal proximal tubules in the absence and the presence of alanine but also provides a precise quantification of fluxes through the enzymes related to acetate metabolism.
Fate of Acetate-Confirming previous results (10,16), acetate was avidly metabolized by rabbit kidney tubules. Carbon balance calculations, comparing on the one hand the removal of acetate (Table I) and on the other hand the sum of the 14 CO 2 formation (Table II) and the 13 C incorporated into glutamate plus glutamine (Tables III and IV), indicate that most, if not all, of the acetate metabolized could be accounted for by the products measured. The small discrepancies observed may be due to small experimental errors in the determination of substrate removal or product formation. In this respect, it is also probable that small amounts of acetoacetate were labeled with carbon 13 during the incubations but were lost during the sample preparation for 13 C NMR measurements because of the well established instability of this ketone body.
With acetate as the sole exogenous substrate, most (72%) of the C-1 but only about half (49%) of the C-2 of acetate was recovered as CO 2 . This clearly means that a substantial fraction of acetate was metabolized beyond the first tricarboxylic acid cycle turn because the acetyl-CoA carbons incorporated into citrate cannot be released as CO 2 during the first tricarboxylic acid cycle turn. This observation also means that part of the acetate carbons removed was incorporated into non-volatile compounds. Our 13 C NMR measurements revealed that these compounds were glutamine and, to a lesser extent, glutamate. This finding implies that an amount of oxalacetate equivalent to the accumulation of labeled glutamate and glutamine was formed from endogenous sources to replenish the tricarboxylic acid cycle. The diversion of ␣-ketoglutarate carbons from the tricarboxylic acid cycle to convert them into non-volatile compounds also explains why the two carbons of acetate were not converted into CO 2 at the same rates. Indeed, at the end of the first tricarboxylic acid cycle turn, the C-1 of acetate was recovered either as the C-1 or the C-4 of oxalacetate, whereas the C-2 of acetate was recovered either as the C-2 or the C-3 of oxalacetate. Because the C-1 and C-4, but not the C-2 and C-3, of oxalacetate could be released as CO 2 during the next tricarboxylic acid cycle turn, the probability to recover the C-1 of acetate into CO 2 , and not into glutamate and glutamine, was necessarily greater. In other words, for a given accumulation of glutamate and glutamine, the greater the recycling in the tricarboxylic acid cycle, the greater the ratio of the release of CO 2 from the C-2 of acetate to that from the C-1 of acetate.
It is important to mention that, although the addition of alanine stimulated acetate utilization, the release as CO 2 of the C-1 of acetate did not increase, and that of the C-2 significantly diminished. This was due to an increased incorporation of both acetate carbons into glutamate and glutamine as revealed by the 13 C NMR data (Tables III and IV).
Note that in the presence of alanine, only 49% of the C-1 and 25% of the C-2 of acetate were released as CO 2 (Tables I and II). Thus, depending on the intensity of the diversion of ␣-ketoglutarate carbon from the tricarboxylic acid cycle, which is necessarily matched by an equivalent anaplerotic activity, the proportions of the acetate carbons released as CO 2 may greatly vary. This observation is of general importance for all in vivo studies in which acetate oxidation is studied (3,7,25) or in which acetate is used as a labeled precursor for measuring metabolic fluxes (26 -28). It should be emphasized that the measurement of the differential yield in labeled CO 2 from acetate may seem at first sight confirmatory of a concept established for a long time. In fact, it is a key measurement allowing us to validate our calculations of the proportion of ␣-ketoglutarate converted into oxalacetate both in the absence and in the presence of alanine. Indeed, the alanine-induced decrease in this proportion (Table V), which reflects the recycling in the tricarboxylic acid cycle, is in excellent agreement with the finding reported in Table II that alanine reduced the CO 2 from the C-2 of acetate but not that from the C-1 of acetate.
Fluxes through Enzymes of Acetate Metabolism-The values of fluxes presented in Table VI, which are mainly derived from 13 C NMR data (see the equations of the model under "Appendix"), are consistent with the enzymatic and radioactive measurements shown in Tables I and II. As expected in the presence of acetate as the sole exogenous substrate, most of the acetyl-CoA formed from acetate was metabolized by the citrate synthase reaction, and most of the ␣-ketoglutarate derived from citrate was metabolized through ␣-ketoglutarate dehydrogenase. Because no exogenous nitrogenous substrate was present in the incubation medium in the absence of alanine, the formation of glutamate occurred by transamination of ␣-ketoglutarate with endogenous amino acids derived from proteolysis. Some of the latter glutamate was probably oxidatively deaminated by glutamate dehydrogenase, which is very active in rabbit kidney tubules (29), to provide the ammonia needed for the synthesis of the glutamine accumulated. Note that the minimum flux through the anaplerotic enzyme, pyruvate carboxylase, corresponds to the net conversion of ␣-ketoglutarate into glutamate (Table VI) and that in the presence of acetate as the sole exogenous substrate, the pyruvate metabolized by the pyruvate carboxylase reaction was formed exclusively from endogenous substrates.
The fact that alanine addition, which caused an increased synthesis of oxalacetate by pyruvate carboxylase, stimulated the flux through acetyl-CoA synthetase by 32% strongly suggests that the availability of oxalacetate was limiting for acetate utilization. The diminution of flux through 3-ketothiolase probably resulted from a diversion of the acetyl-CoA synthesized from acetoacetate formation to citrate formation.
The very large stimulation of the flux through citrate synthase caused by alanine addition resulted not only from an increased availability of oxalacetate derived from alanine but also from an increased formation of acetyl-CoA also derived from alanine thanks to the pyruvate dehydrogenase reaction, which functioned at a high rate despite the presence of acetate. It is of interest to observe that the stimulation of the flux through citrate synthase was not accompanied by a statistically significant increase in flux through ␣-ketoglutarate dehydrogenase but rather by a considerable stimulation of the conversion of ␣-ketoglutarate into glutamate and glutamine. The increased synthesis of glutamate can be explained by the increased transamination of ␣-ketoglutarate with alanine, leading to stoichiometric increases in flux through both glutamate dehydrogenase and glutamine synthetase (Table VI).
It should be emphasized that the appearance of tracer carbons from acetate in glutamine represented an unidirectional conversion of acetate carbons into glutamine carbons thanks to the glutamine synthetase reaction and not a substantial exchange of carbons as a result of the concomitant action of glutamine synthetase and glutaminase, two enzymes that function unidirectionally. This view is supported by the following observations. (i) The amount of glutamine present at the start of incubation (in zero-time flasks) was 0.12 Ϯ 0.01 mol.
(ii) Nitrogen balance calculations indicate that glutamine was progressively synthesized from endogenous sources during the incubation period in the absence of exogenous substrate and in the presence of acetate as the sole exogenous substrate but not in the presence of acetate plus alanine (see "Results"). In the presence of acetate as the sole exogenous substrate, the concentration of glutamine at the end of the incubation period was 0.4 mM. Given that the production of glutamine was linear with time, one can calculate from results obtained previously in this laboratory (30) that, in our rabbit kidney tubules, flux through glutaminase, whose activity is low (31), was negligible when compared with that through glutamine synthetase, whose activity is high (31,32). It is also interesting to underline that neither at zero time nor at the end of the incubation period did we observe any glucose accumulation; therefore, no exchange of labeled acetate carbons could occur with glucose.
It is important to mention that the values of the enzymatic fluxes obtained in the present study with 13 C-labeled acetates plus unlabeled alanine are very close to those found with 13 Calanines plus unlabeled acetate (10). In our opinion, this represents a strong support not only for the validity of our results obtained with differently labeled substrates but also for the validity of our methodological approach. This approach combines experimental results derived mainly from 13 C NMR data with new mathematical models of the metabolic pathways involved when acetate and alanine are concomitant substrates of the rabbit kidney tubules. It is worthwhile to emphasize that such an approach is directly applicable to any cell that has the capacity to metabolize both acetate and alanine and contains significant activities of the enzymes involved in our study.
Physiological Importance-As shown in previous works (10,16) and in this study, the rabbit proximal tubule has the capacity in vitro to avidly metabolize not only acetate, the most concentrated volatile fatty acid in the rabbit blood (6), but also alanine, an important precursor of glutamine (10,(33)(34)(35). It is very likely that in vivo, the rabbit kidney also metabolizes acetate and then, besides providing energy for the renal transport of mineral and organic solutes, contributes to the addition of bicarbonate to the urine, which is alkaline in this herbivorous species (36). In terms of acid base balance equilibrium, it is of interest to note that the alanine-induced stimulation of the removal of acetate, an anion, was balanced by an increased accumulation of glutamate and lactate, two other anions (Table  I). Thus, the addition of alanine did not increase the production of bicarbonate as a result of an increased acetate metabolism and, therefore, did not alter the acid base balance equilibrium. Concomitantly, alanine greatly stimulated the synthesis of glutamine, which incorporated a significant fraction of the acetate carbons used. Thus, the stimulation of acetate metabolism by alanine, if it occurs in vivo, might appear as a salvage mechanism that does not increase the loss of acetate carbons as CO 2 but rather incorporates them in the form of glutamine, an amino acid reabsorbed by the kidney and made available to other tissues.
Finally, this study also illustrates the utmost importance of anaplerosis for the renal metabolism of compounds which, like acetate, provide only two carbon units to intermediary metabolism and are potentially incorporated into non-volatile compounds such as glutamate and glutamine. The accumulation of glutamate and glutamine reflects, at least in part, a physiological cataplerosis that constantly depletes the concentration of citric acid cycle intermediates. Therefore, as for liver and cardiac cells (37,38), for the renal cells to continue generating energy, a constant anaplerosis was necessary in the present study. This anaplerosis was made possible by metabolizing alanine carbons through pyruvate carboxylase. APPENDIX Scheme 1 shows that acetate metabolism consists of a multicycle made of five different cycles operating simultaneously: (i) the citric acid cycle, (ii) the OAA 3 P-enolpyruvate 3 Pyr 3 OAA cycle, (iii) the OAA 3 P-enolpyruvate 3 Pyr 3 AcCoA 3 OAA cycle, (iv) the Glu 3 ␣KG 3 Glu cycle, and (v) the Glu 3 Gln 3 Glu cycle. This scheme also shows the proportions of metabolite conversions. Scheme 2, derived from Scheme 1, indicates the fate during one turn of each metabolic cycle of the individual carbons 1 and 2 of the acetate molecule. One can deduce the fate over a theoretically infinite number of multicycle turns. Note that a complete turn of the metabolic cycle is considered to have occurred as soon as the metabolite of interest of the cycle has been resynthesized once.
From the fate of these individual carbons, the amounts of acetate utilized, and the different end products accumulated (total and labeled species), our model allows the calculation of parameters necessary to obtain the amount of the different intermediate metabolites formed. Then, fluxes can be calculated since a flux through a given enzyme is taken as the formation of one product of the reaction catalyzed by this enzyme.
Notations-Let us call [C y Met] CzAc the amount of the metabolite (Met) formed and labeled on its carbon y (where 1 Յ y Յ 5) arising from acetate labeled on its carbon z, where z is equal to 1 or 2 because we used [1-13 C] Ac(ϩAla) when the calculation procedures are the same. This principle is also applied to accumulated metabolites and enzymatic fluxes. The accumulated amount of a metabolite is indicated by a ∧ sign added on the left side of its name inside the square brackets. An enzymatic flux is represented by the abbreviated name of the enzyme placed inside braces.
Calculations  Table I.
This model takes into account the ␣-ketoglutarate recycling through glutamate and glutamine; the corresponding proportion is noted (Љ). Let us call g, h, and z the proportions of any metabolite resynthesized after each complete turn of the citric acid cycle, the OAA 3 P-enolpyruvate 3 Pyr 3 OAA cycle, and the OAA 3 P-enolpyruvate 3 Pyr 3 AcCoA 3 OAA cycle, respectively. Oxalacetate, the common intermediate to these three cycles, is either converted into citrate or phosphoenolpyruvate with the proportion a and b, respectively. Also, because at each bifurcation, the sum of all proportions adds up to one, the result is as shown in Eq. where sЉ is the proportion of citrate-derived ␣-ketoglutarate that has been converted into oxalacetate after partial recycling through glutamate and glutamine. Therefore, g takes into account the recycling of ␣-ketoglutarate through glutamate and glutamine.
Let Y be the amount of Ac utilization. Schemes 1 and 2 indicate that citrate formed from Ac-derived AcCoA is equal to Y⅐(1 Ϫ u)⅐sЉ, corresponding to the beginning of the first multicycle turn for AcCoA carbons. SCHEME 1. Metabolic cycles operating during acetate metabolism. This scheme shows five cycles functioning simultaneously. At each turn of this multicycle, each carbon atom undergoes successive shifts of its position inside metabolite molecules. In the citric acid, the OAA 3 P-enolpyruvate 3 Pyr 3 OAA, and OAA 3 P-enolpyruvate 3 Pyr 3 AcCoA 3 OAA cycles, the proportion of metabolite resynthesized at each turn is g, h, and z, respectively. k 1 and k 2 are the recycling proportions in the Glu 3 ␣KG 3 Glu and Glu 3 Gln 3 Glu cycles, respectively. The proportions g and z take into account the recycling of ␣-ketoglutarate through these two latter cycles. Scheme 1 also shows that oxalacetate is an important metabolite common to three cycles. The oxalacetate that condensed to acetyl-CoA to yield citrate was partially resynthesized in the multicycle; it originated mainly from endogenous precursors in the presence of acetate as the sole added substrate or from alanine in the presence of acetate plus alanine.
An equal amount of OAA is necessary to form these citrate molecules, but simultaneously, a fraction (1 Ϫ a)/a (Scheme 1) of this amount of OAA is utilized through another pathway. Thus, even at the beginning of the multicycle, it is obvious that not all of the OAA participating in the metabolism can be seen through acetate metabolism.
Note that in the condition with alanine plus acetate, the total OAA formed can be obtained from the fate of alanine carbons (10). In a classical representation, if OAA was resynthesized only in the citric acid cycle with a proportion g at each turn, the total amount of OAA formed over a theoretically infinite number of turns would be given by the OAA, noted [OAA] 0 , formed at the beginning of the first cycle turn multiplied by Eq. 7. In this study, the total amount of OAA formed is expressed as a function of [ v AcCoA] 0 , the AcCoA utilized at the beginning of the first multicycle turn, as expressed in Eq. 8 (see Schemes 1 and 2).
It should be stressed that, especially in the presence of added alanine, a significant part of OAA does not require AcCoA to be formed but is derived from endogenous or exogenous OAA precursors such as alanine. Therefore, to explain the total OAA formation, it was necessary to replace g, the tricarboxylic acid cycle recycling factor, by F, which is an equivalent recycling factor. Note that the higher the ratio of AcCoA utilized to OAA formed, the lower the ratio F/g.
Note also that, during this metabolism, a part of the AcCoA can be formed from OAA and thus, gives rise to successive turns of the OAA 3 P-enolpyruvate 3 Pyr 3 AcCoA 3 OAA cycle, the other part being obtained from the added acetate.
From Schemes 1 and 2, one can demonstrate the following equations (Eqs. 9 -15). Knowing Y, the amount of acetate utilized, Eq. 17 allows us to SCHEME 2. Metabolic fate of the C-1, and C-2 of acetate in rabbit kidney tubules. This scheme, derived from Scheme 1, shows the metabolic fate of acetate labeled on its C-1 or C-2, which is represented by 1,2 Ac. Metabolites formed are represented by ␣,␤ Met. Met corresponds to any acetate-derived metabolite, and ␣ and ␤ indicate the labeled carbon of these metabolites when the labeled carbon of the acetate added as substrate was C-1 or C-2, respectively. Unlabeled carbons of acetate-derived metabolites are represented by a minus sign. The amount (in mol/h) of labeled acetate utilized is represented by Y. Lowercase letters (or numbers) indicate the proportion of metabolite converted at each enzymatic step. The relative amount of the substrate (labeled acetate) converted into any labeled intermediate or end product is obtained by multiplying the successive proportions found in the pathway from the substrate to the intermediate or end product of interest. The amount of intermediate formed or end product accumulated during the incubation period (1 h) is obtained by multiplying the corresponding relative amount by the amount [Y] of labeled acetate utilized. In this scheme, the proportion (1 Ϫ s)Ј of ␣KG converted into glutamate takes into account not only the direct formation of Glu from ␣KG but also its indirect formation through Gln. The proportions (Ј) and (Љ) are equal to 1/(1 Ϫ k 1 Ϫ k 2 ) and (1 Ϫ k 1 )/(1 Ϫ k 1 Ϫ k 2 ), respectively.