The rabbit kidney tubule simultaneously degrades and synthesizes glutamate. A 13C NMR study.

The rabbit kidney does not readily metabolize but synthesizes glutamine at high rates by pathways that remain poorly defined. Therefore, the metabolism of variously labeled [13C]- and [14C]glutamates has been studied in isolated rabbit kidney tubules with and without acetate. CO2, glutamine, and alanine were the main carbon and nitrogenous end products of glutamate metabolism but no ammonia accumulated. Absolute fluxes through enzymes involved in glutamate metabolism, including enzymes of four different cycles operating simultaneously, were assessed by combining mainly the 13C NMR data with a new model of glutamate metabolism. In contrast to a previous conclusion of Klahr et al. (Klahr, S., Schoolwerth, A. C., and Bourgoignie, J. J. (1972) Am. J. Physiol. 222, 813-820), glutamate metabolism was found to be initiated by glutamate dehydrogenase at high rates. Glutamate dehydrogenase also operated at high rates in the reverse direction; this, together with the operation of the glutamine synthetase reaction, masked the release of ammonia. Addition of acetate stimulated the operation of the “glutamate → α-ketoglutarate → glutamate” cycle and the accumulation of glucose but reduced both the net oxidative deamination of glutamate and glutamine synthesis. Acetate considerably increased flux through α-ketoglutarate dehydrogenase and citrate synthase at the expense of flux through phosphoenolpyruvate carboxykinase; acetate also caused a large decrease in flux through alanine aminotransferase, pyruvate dehydrogenase, and the “substrate cycle” involving oxaloacetate, phosphoenolpyruvate, and pyruvate.

Phosphate-activated glutaminase (hereafter referred to as glutaminase), the enzyme which initiates glutamine degradation and releases both glutamate and ammonia, is present in the rabbit kidney (1)(2)(3). However, its activity is low when compared with that in the kidney of other species (2,4). This is not surprising because the rabbit, like other herbivorous species, excretes an alkaline urine (2) that does not need to contain large amounts of ammonium ions. This also explains why rabbit kidney tubules do not readily use glutamine as substrate (5,6). By contrast, the rabbit kidney has a high capacity to trap ammonia and synthesize glutamine (1, 2, 4 -8) by pathways that are far from being fully defined. In his initial work on glutamine synthesis in mammalian tissues (1), Krebs demonstrated that addition of glutamate to the incubation medium significantly increased glutamine synthesis by rabbit kidney cortex slices; since no ammonia accumulated under this condition, he concluded that the ammonia released by the glutamate dehydrogenase reaction was immediately utilized for the synthesis of glutamine (1). Later, Klahr (9) and Klahr et al. (4) have also shown that glucose synthesis by rabbit kidney cortex slices is increased in the presence of glutamate. Since the latter authors observed glucose synthesis without concomitant accumulation of ammonium ions, they concluded, unlike Krebs (1), that glutamate metabolism in rabbit kidney slices is initiated by transamination rather than by oxidative deamination by glutamate dehydrogenase (4). More recently, Watford et al. (10) stated that rabbit kidney tubules do not readily use glutamate or glutamine as gluconeogenic substrate, possibly because of inhibition of glutamate dehydrogenase by ammonia (4). However, none of the above authors measured glutamate uptake in their study. Recent studies in our laboratory (5,6) have also shown that endogenous glutamate is converted at significant rates into glutamine in isolated rabbit kidney-cortex tubules which contain a high activity of glutamine synthetase (1,(7)(8)(9), the enzyme responsible for glutamine synthesis.
Thus, all the above observations suggest that both rabbit renal cortical slices and tubules have the capacity to remove glutamate, but the extent to which the rabbit renal cortex can utilize glutamate as substrate and the pathways involved in glutamate metabolism by this tissue remain uncertain. In an attempt to clarify this subject, we decided to study glutamate metabolism in rabbit kidney tubules; for this, we conducted a study involving metabolic balance of substrate, incorporation of label into metabolites of glutamate, and modeling of glutamate metabolism (44).
Our results clearly establish that rabbit kidney tubules readily utilize glutamate as substrate and convert it mainly into glutamine, alanine, and CO 2 and to a lesser extent into glucose and serine. Modeling of the data obtained by 13 C NMR spectroscopy indicates that, during glutamate utilization via mainly glutamate dehydrogenase, glutamine synthetase, and alanine aminotransferase, glutamate synthesis also occurs at high rates. In addition, our study demonstrates that addition of acetate reduces net removal of glutamate by inhibiting not only net flux through glutamate dehydrogenase in the oxidative deamination direction but also by inhibiting glutamine, alanine, and serine synthesis. Finally, the rate of the cycle involving pyruvate, oxaloacetate, and phosphoenolpyruvate that was high with glutamate as sole substrate is drastically reduced

Preparations of Kidney Tubules and Incubations
Kidney cortex tubules were prepared by collagenase treatment of renal cortex slices as described by Baverel et al. (11). Incubations other than those involving radioactive substrates were performed for 30 or 60 min at 37°C in a shaking water bath, in 25-ml stoppered Erlenmeyer flasks in an atmosphere of O 2 /CO 2 (19/1). The flasks contained 1 ml of the tubule suspension plus 3 ml of Krebs-Henseleit medium (12) either unsupplemented or supplemented with substrates, i.e. 5 mM (final concentration) L-[1-14 C]glutamate (10 3 Bq/flask), L- [1,[5][6][7][8][9][10][11][12][13][14] C]glutamate (2. 10 3 Bq/flask), L-[U-14 C]glutamate (5. 10 3 Bq/flask), L-[1,2-13 C]glutamate, L-[3-13 C]glutamate, L- [5-13 C]glutamate, in the absence and the presence of 10 mM (final concentration) acetate or 25 mM NaH 14 CO 3 (10 6 Bq/flask). These differently labeled glutamate were used in an attempt to define the fate of all the glutamate carbons which depends on the metabolic pathways involved (see Fig. 1). It should be mentioned that the C-1 of glutamate could be released as 14 CO 2 by the ␣-ketoglutarate dehydrogenase reaction. The C-2 and C-5 on one hand, and the C-3 and C-4 on the other hand, were expected to behave symmetrically beyond the stage of succinyl-CoA. Therefore, the [ 14 C]glutamates used allowed us to measure and/or calculate easily the contribution of each glutamate carbon to the release of CO 2 . Finally, 14 C-labeled bicarbonate was used to measure the pyruvate carboxylase-mediated incorporation of the bicarbonate carbon into the C-1 of alanine and of glutamate ϩ glutamine. In all experiments, each experimental condition was performed in quadruplicate. Incubation was stopped by adding perchloric acid (final concentration 2%, v/v) to each flask. Metabolite assays were conducted on the neutralized supernatant. In all experiments, zero time flasks, with and without substrates, were prepared by adding perchloric acid before the tubules. When radioactive glutamate or bicarbonate 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, measurement of bicarbonate fixation, and NMR spectroscopy measurements.

Analytical Methods
Metabolite Assays-Glucose, glycogen, lactate, pyruvate, glutamate, glutamine, alanine, aspartate, citrate, ␣-ketoglutarate, fumarate, malate, glycerol, and glycerol 3-phosphate were determined by usual enzymatic methods, and the dry weight of tubules added to the flasks was determined as described previously (11,13). 13 C NMR Techniques-These were recorded at 100.6 MHz on a Bruker AM-400 WB spectrometer using a 10-mm broadband probe thermostated at 8 Ϯ 0.5°C. Magnet homogeneity was adjusted using the deuterium lock signal. Supernatants, obtained from four flasks for each experimental condition with either L-[1,2-13 C]glutamate, or L[3-13 C]glutamate, or L-[5-13 C]glutamate, with or without acetate as substrates, were pooled and lyophilized. Then, the freeze-dried material was redissolved in 3 ml of 2 H 2 O and centrifuged (5000 ϫ g, 4°C, 15 min). In order to obtain absolute quantitative results, special care was taken for data acquisition. Relaxation times, saturation, and nuclear Overhauser effects were minimized, and resolution was optimized. To reduce long relaxation times, 20 l of a 2 H 2 O solution containing 13 mM sodium EDTA and 11 mM gadolinium nitrate and [2-13 C]glycine as internal standard were added to each milliliter of sample. Under this condition, all the T1 relaxation times determined for the carbons of interest in our experiments and measured using the inversion-recovery method were less than 10 s. Acquisition parameters were as follows: spectral width, 25000 Hz; tilt angle, 90°; data size, 32K; 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 (14). We did not use the nuclear Overhauser enhancement during proton decoupling to avoid the use of corresponding correction factors. A 2-Hz line broadening was applied prior to Fourier transform. Peak areas were determined by measuring the integral heights. Chemical shifts were expressed as parts/million relative to tetramethylsilane. Assignments were made by comparing the chemical shifts obtained with those given in the literature (15,16).

Calculations
Net substrate utilization and product formation were calculated as the difference between the total flasks 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 micromoles of substances removed or produced/g dry wt/unit of time (30 or 60 min).
The rates of release of 14 CO 2 from the 14 C-labeled glutamate species used were calculated by dividing the radioactivity in 14 CO 2 by the specific radioactivity of each labeled carbon of the glutamate species of interest measured in each medium.
Fixation of 14 CO 2 from [ 14 C]bicarbonate was calculated by dividing the acid-stable radioactivity present in the incubation medium after incubation by the specific radioactivity of the total "bicarbonate ϩ CO 2 " pool measured in the zero time flasks. The formation of [1-14 C]glutamine ϩ [1-14 C]glutamate was measured by the method of Squires and Brosnan (17).
When a [ 13 C]glutamate species was the substrate, the transfer of the C-1 or C-2 or C-3 or C-5 of glutamate to a given position in a given metabolite was calculated by the following formula: where L m is the amount of 13 C measured in the corresponding NMR resonance, l m is the natural 13 C abundance (1.1%) multiplied by the amount of the metabolite assayed enzymatically, E s is the 13 C enrichment of the C-1 or C-2, or C-3 or C-5 of glutamate, and e s is the natural 13 C abundance. Table I, rabbit kidney tubules metabolized glutamate at high rates; glutamate utilization was linear with time over 60 min. Glutamine, which was the main nitrogenous product found, accumulated linearly with time; since gluta- mine synthesis is an ATP-consuming reaction, this means that our tubules were metabolically viable (their ATP content was 6.5 Ϯ 0.4, 7.4 Ϯ 0.6, and 7.5 Ϯ 0.5 mol/g dry wt after 0, 30, and 60 min of incubation, respectively; n ϭ 4; these mean values take into account the values found both in the absence and the presence of acetate because they were not significantly different). Significant amounts of alanine also accumulated at both incubation times, but alanine accumulation was not linear with time; nitrogen balance calculations reveal that alanine nitrogen accounted for 31.5 and 15.3% of the nitrogen removed as glutamate at 30 and 60 min, respectively. Small amounts of aspartate, glucose, and lactate were also found to accumulate but no significant accumulation of pyruvate, glycogen, glycerol 3-phosphate, glycerol, acetate, ketone bodies, or tricarboxylic acid cycle intermediates occurred. As found by Krebs (1) and Klahr (9) and Klahr et al. (4) in rabbit kidney slices, no ammonia accumulated; on the contrary, a small amount of the ammonia brought by the tubules at zero time was utilized as substrate during the incubation. Nitrogen balance calculations also indicate that the nitrogen found as glutamine, alanine, and aspartate exceeded the nitrogen removed as glutamate and ammonia; this means that the glutamine synthesis observed from endogenous sources in the absence of exogenous substrate (see Table I) still occurred, at least partially, in the presence of glutamate.

As shown in
To test whether acetate, a substrate that circulates at high levels in the rabbit blood (18) and is readily metabolized by rabbit kidney tissue (19), can alter the synthesis of glutamine in rabbit kidney tubules by providing additional carbons, increasing concentrations of acetate (0.5-10 mM) were added to the incubation medium containing 5 mM glutamate. Fig. 2 shows that acetate addition induced a dose-dependent inhibition of both glutamate removal and glutamine accumulation; alanine accumulation was reduced only at the highest concentration of acetate used, whereas an increase in glucose accumulation was observed at all acetate concentrations employed.
In order to determine the precise pathways of glutamate metabolism (see Fig. 1) and the effect of 10 mM acetate (that had the greatest effect of the concentrations used) on these pathways, we performed a series of experiments in which we combined enzymatic, radioactive, and 13 C NMR spectroscopy measurements using specifically or uniformly 14 C-and/or 13 Clabeled glutamate in the absence and the presence of unlabeled acetate. Table II shows that 10 mM acetate, which was avidly removed by rabbit kidney tubules, caused effects that have already been partially commented above (see Fig. 2). This table provides the values (obtained by enzymatic methods) necessary to the 13 C (see the "Calculation" section) and flux calculations in which the glutamate removal is a key parameter. In addition, this table allows us to compare the formation of end products with calculated fluxes through the corresponding enzymes. Note also that the presence of glutamate stimulated the utilization of acetate.  (25 mM). With glutamate as sole substrate, it can be seen that a large proportion of the C-1 of this amino acid was released as CO 2 by the ␣-ketoglutarate dehydrogenase reaction; since the alanine and aspartate formed by the alanine and aspartate aminotransferase reactions can account for only 34% (64. 6: 190.6) of the ␣-ketoglutarate that has been decarboxylated, our data clearly indicate that most (66%) of the ␣-ketoglutarate synthesis from glutamate occurred thanks to the glutamate dehydrogenase reaction. The difference between the releases of 14 CO 2 from [1,5-14 C]glutamate and [1-14 C]glutamate gives the release of 14 CO 2 from [5-14 C]glutamate which is equal to a mean value of 153 mol/g dry wt. Since the releases as CO 2 of the C-2 and of the C-5 of glutamate, which occur beyond the stage of succinate and fumarate (two symmetrical molecules), are assumed to be equal, one can deduce that the release of 14 CO 2 from [2-14 C]glutamate is also equal to 153 mol/g dry wt. Thus, the C-1, C-2, and C-5 of glutamate accounted for about 80% (497.2: 635.2) of the total release of CO 2 from glutamate measured as the 14 CO 2 release from [U-14 C]glutamate (see Table III). Therefore, it appears that only a small proportion (about 10%) of the C-3 and of the C-4 of glutamate, which are assumed to behave symmetrically beyond the stage of succinyl-CoA, was released as CO 2 in rabbit kidney FIG

TABLE II
Effect of 10 mM acetate on the metabolism of 5 mM L-glutamate in rabbit kidney tubules Kidney tubules (57.9 Ϯ 2.5 mg dry wt per flask) were incubated for 60 min as described under "Experimental Procedures." Results (mol/g dry wt/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 without acetate: *, 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 Tables III-VI tubules. Another important observation made when glutamate was the only exogenous substrate is that a significant amount of 14 CO 2 was fixed from [ 14 C]bicarbonate and recovered in the form of the C-1 of glutamate ϩ glutamine and of the C-1 of alanine (Table III) This demonstrates that glutamate synthesis occurred even in the presence of a high concentration of glutamate which was removed in net amounts. Thus, our enzymatic measurements reflect only the net removal of glutamate, which is simultaneously removed and synthesized, and therefore underesti-mate the unidirectional removal of this substrate.
Since the C-1 of alanine was labeled in the presence of glutamate plus 14 C-bicarbonate, this reveals that part of the oxaloacetate synthesized by the pyruvate carboxylase reaction underwent equilibration with fumarate to give [1-14 C]malate and then [1-14 C]oxaloacetate. This [1-14 C]oxaloacetate was converted into [1-14 C]alanine by the combined action of phosphoenolpyruvate carboxykinase, pyruvate kinase, and alanine aminotransferase (see Fig. 1). Thus, the above observations demonstrate that the "substrate cycle" involving pyruvate, oxaloacetate, and phosphoenolpyruvate was operating during glutamate metabolism in rabbit kidney tubules.
Table III also reveals that, as far as the release of CO 2 from glutamate is concerned, only the release of 14 CO 2 from [1,5-14 C]glutamate was significantly inhibited by the addition of acetate. Since the release of 14 CO 2 from [1-14 C]glutamate did not change under this condition, this indicates that the oxidation of the C-5, and therefore of the C-2 of glutamate, was inhibited by acetate. Simple calculations reveal that acetate did not alter the oxidation of the C-3 and C-4 of glutamate. Surprisingly (at least at first sight), in the presence of glutamate plus acetate, the release of 14 CO 2 from [1-14 C]glutamate (Table III), which gives an estimate of the minimal removal of glutamate, was considerably higher than the net removal of glutamate measured enzymatically (see Table II). This reflects the fact that acetate considerably stimulated the synthesis of glutamate; further evidence for this is provided by the observation that acetate increased the accumulation of [1-14 C]glutamate ϩ [1-14 C]glutamine from glutamate plus [ 14 C]bicarbonate (Table III). Therefore, under this experimental condition (glutamate ϩ acetate), the unidirectional removal of glutamate was masked to a large extent by the simultaneous synthesis of glutamate, a fact which could not be demonstrated solely by enzymatic measurement of glutamate removal.
Finally, Table III shows that, in the absence and the presence of acetate, 44 and 46%, respectively, of the alanine that accumulated was labeled on its C-1 in the presence of [ 14 C]bicarbonate. This suggests that a large fraction of the oxaloacetate giving rise to alanine equilibrated with fumarate. Fig. 3, A and B, shows the 13 C NMR spectra of perchloric extracts obtained after 60 min of incubation of rabbit kidney tubules in the presence of [3-13 C]glutamate either in the absence or the presence of acetate. As expected, these figures show that the main non-volatile carbon products of glutamate metabolism in the absence of acetate were glutamine and alanine and that acetate inhibited the accumulation of both alanine and glutamine. The fact that carbons of glutamate and glutamine other than that originally labeled in the added glutamate became labeled during incubation confirms that glutamate synthesis occurred during glutamate utilization as concluded above on the basis of enzymatic and radioactive measurements (see Tables II and III). Irrespective of the labeled glutamate species used as substrate (the spectra obtained with   (Table IV), the labeled C-2 of alanine was of the same order of magnitude as the labeled C-3 of alanine, and the labeled C-4 of glutamate and glutamine was virtually equal to the labeled C-5 of glutamate and glutamine; this is in agreement with the view that the C-3 and C-4 of the glutamate converted into ␣-ketoglutarate by either glutamate dehydrogenase or by alanine or aspartate or phosphoserine aminotransferase (see Fig. 1), and further metabolized in the tricarboxylic acid cycle, passed through the stage of succinate and fumarate, two symmetrical molecules. The passage through these symmetrical molecules is also consistent with the fate of the C-2 and C-5 of the glutamate converted into the ␣-ketoglutarate that was further metabolized in the tricarboxylic acid cycle because the label-ings of the C-1 (singlet) of glutamate, glutamine, and alanine in the presence of [5-13 C]glutamate as substrate were approximately the same as those obtained when [1,2-13 C]glutamate was the substrate (see Tables V and VI).
Since the C-2 and the C-3 of the oxaloacetate derived from the added [3-13 C]glutamate were converted at the same rates to the C-3 and C-2 (respectively) of the glutamate synthesized and therefore to the C-3 and C-2 of the glutamine derived from the latter glutamate, one may calculate that with glutamate as sole substrate, a mean value of 67 mol/g dry wt of glutamine arose directly from the added [3-13 C]glutamate, whereas 10.8 mol/g dry wt of [3-13 C]glutamine derived from synthesized [3-13 C]glutamate (see Table IV); in the presence of [3-13 C]glutamate plus acetate, the corresponding values are 34.6 and 11.3 mol/g dry wt, respectively. The latter values obtained for estimating the direct accumulation of glutamine from added [3-13 C]glutamate in the absence and the presence of acetate are in relatively good agreement with those obtained by measuring the accumulation of [5-13 C]glutamine from [5-13 C]glutamate (Table V) as well as by measuring the multiplets (spin-spin 13 C coupling of C-1 and C-2) resulting from the synthesis of [1,2-13 C]glutamine from [1,2-13 C]glutamate (Table VI).    (Table VI).
Table IV also shows that, with [3-13 C]glutamate as sole substrate, the C-2 and the C-3 of aspartate, lactate, alanine, and serine became labeled to the same extent. This reveals that the C-3 of glutamate, after having passed through the stage of succinate and fumarate, two symmetrical molecules, and then through the stage of malate and oxaloacetate, was metabolized by the action of phosphoenolpyruvate carboxykinase, pyruvate kinase, lactate dehydrogenase, aspartate, alanine, and phosphoserine aminotransferases. The metabolism of [3-13 C]gluta-mate also resulted in the labeling of the C-4 and C-5 of glutamate and glutamine (Table IV). This indicates that the labeled C-3 of glutamate led to the labeling of the C-2 and C-1 of acetyl-CoA that were converted to the C-4 and C-5 of citrate and then to the C-4 and C-5 of glutamate and glutamine. Thus, the pyruvate derived from glutamate underwent oxidative decarboxylation by the pyruvate dehydrogenase reaction. Addition of acetate did not alter the labeling of the C-2 and C-3 of aspartate but reduced the labeling of the C-2 and C-3 of lactate, alanine, and serine as well as the labeling of the C-4 and C-5 of glutamate and glutamine (Table IV). Since the sum of the labeled C-2 and C-3 of alanine (Table IV) was close to the accumulation of alanine found in Table II, virtually all the carbons of the alanine found were derived from glutamate.
Very little C-1 of glutamate and glutamine became labeled during the metabolism of [3-13 C]glutamate alone, but significant labeling of the C-1 of glutamate but not of glutamine occurred upon addition of acetate (Table IV). Such a labeling pattern results from prior labeling of the C-4 of oxaloacetate leading via the combined action of the tricarboxylic acid cycle and the glutamate dehydrogenase reaction to the labeling of the C-1 of glutamate.  [5-13 C]glutamate was further metabolized in the tricarboxylic acid cycle to lead to the accumulation of alanine and to the synthesis of glutamate which accumulated as glutamate and glutamine. Both in the absence and the presence of acetate, the labelings of the C-1 (singlet) of glutamate and glutamine and of alanine were almost identical when [5-13 C]glutamate or [1,2-13 C]glutamate was the substrate. Tables V and VI also show that, with [5-13 C]glutamate and [1,2-13 C]glutamate as substrate, both in the absence and the presence of acetate, the labeling of the C-1 (singlet) of glutamate, glutamine, and alanine was virtually half that of the C-2 of glutamate, glutamine, and alanine when [3-13 C]glutamate was the substrate (see Table IV).
When [3-13 C]glutamate was the substrate in the absence and the presence of acetate, the labeling (in mol/g dry wt) of the different carbons of glucose (␣ ϩ ␤ anomers) was as follows: C-1, 2.  Table II, it appears that acetate tended to stimulate the conversion of glutamate carbons into glucose carbons, although the 13 C resonances of the glucose carbons were small. Table VII contains the proportions of each metabolite converted into the next one(s). It should be emphasized here that these proportions which constitute the basis of our model do not give direct access to fluxes. Fluxes also take into account the formation of the substrate of the reaction of interest. Table VII shows that with glutamate as sole substrate, the proportion of oxaloacetate synthesized that was converted into phosphoenolpyruvate was much greater than that converted into citrate, whereas the contrary was true in the presence of acetate. Under the control condition without acetate, almost all the phosphoenolpyruvate synthesized was converted into pyruvate, a very small fraction being converted into 3-phosphoglycerate; addition of acetate increased the proportion of phosphoenolpyruvate converted into 3-phosphoglycerate at the expense of that converted into pyruvate. Most of the pyruvate synthesized was converted into oxaloacetate both in the absence and the presence of acetate; addition of acetate significantly changed the proportions of pyruvate converted into oxaloacetate, alanine, and lactate. Table VII also shows that the proportion of 3-phosphoglycerate synthesized from glutamate as sole substrate that was converted into 3-phosphohydroxypyruvate and then into serine was much greater than that converted into glucose; the reverse was observed upon addition    (Table VII); such a proportion, which may appear surprisingly high in view of the significant proportion of glutamate converted into glutamine, takes into account the operation of the "Glu 3 ␣KG 3 Glu" and of the "Glu 3 Gln 1 3 Glu" cycles but not the operation of the other cycles (see Fig. 1).
The following proportions presented inside specific braces, {}, also take into account the recycling through the "Glu 3 ␣KG 3 Glu" and the "Glu 3 Gln 3 Glu" cycles but not through the other cycles. The proportion { v Glu 3 Glu} of the added glutamate removed that accumulated as glutamate was increased in the presence of acetate. On the contrary, the proportion { Glu 3 Gln} of added glutamate removed that accumulated as glutamine was significantly reduced by the addition of acetate.
Note that 33 and 70% of the citrate-derived ␣-ketoglutarate { Cit ␣KG 3 Glu} was converted into glutamate in the absence and the presence of acetate, respectively. Note also that a high proportion { Cit ␣KG 3 OAA) of the citrate-derived ␣-ketoglutarate was converted into oxaloacetate. It can be seen in Table  VII that a significant proportion (OAA i ) of the oxaloacetate synthesized was inverted as a result of the equilibration with fumarate thanks to the reversible part of the tricarboxylic acid cycle catalyzed by malate dehydrogenase and fumarase. This means that, both in the absence and the presence of acetate, most of the oxaloacetate synthesized by the pyruvate carboxylase reaction underwent equilibration with fumarate. It can also be seen in Table VII that the proportion (TCA)) of oxaloacetate recycled at each turn of the tricarboxylic acid cycle that was relatively small with glutamate as sole substrate, dramatically increased in the presence of acetate. By contrast, the proportion (Pyr 1 ) OAA) of the oxaloacetate that was metabolized in the "OAA 3 PEP 1 3 Pyr 3 OAA" cycle, which was high in the absence of acetate, was considerably diminished in the presence of acetate. Both in the absence and the presence of acetate, a very small proportion (AcCoA 1 ) OAA) of the oxaloacetate synthesized was metabolized in the "OAA 3 PEP 3 Pyr 3 AcCoA 3 Cit 3 OAA" cycle. Table VII shows that the recycling ratio {␣KG ) Glu ϩ Gln} of ␣-ketoglutarate through the "Glu 3 ␣KG 3 Glu" and "Glu 3 Gln 3 Glu" cycles, which corresponds to the proportion of the citrate-derived ␣-ketoglutarate that passed through these 2 cycles, was increased in the presence of acetate. A significant proportion (AcCoA ϩ OAA 3 Cit) of glutamate-derived acetyl-CoA was condensed with oxaloacetate that did not originate from glutamate both in the absence and the presence of acetate (Table VII). Finally, a very large proportion (OAA ϩ AcCoA 3 Cit) of glutamate-derived oxaloacetate was condensed with acetyl-CoA that did not originate from glutamate, especially in the presence of acetate. Table VIII shows the absolute values of fluxes through enzymes involved in glutamate metabolism in the absence and the presence of acetate. It can be seen that high unidirectional fluxes occurred from glutamate to ␣-ketoglutarate and from ␣-ketoglutarate to glutamate at the level of glutamate dehydrogenase under both experimental conditions. Acetate considerably increased these two unidirectional fluxes, resulting in the diminution of the net difference (which was small) in favor of glutamate utilization by the latter enzyme. This effect caused by acetate was accompanied by an inhibition of flux through glutamine synthetase which is in agreement with the reduction of glutamine accumulation shown in Table II. Note that, both in the absence and the presence of acetate, flux through ␣-ketoglutarate dehydrogenase was severalfold greater than the sum of fluxes through glutamate dehydrogenase (net flux in the direction of oxidative deamination of glutamate), alanine, aspartate, and phosphoserine aminotransferases, which lead to ␣-ketoglutarate synthesis; this clearly indicates that most of the ␣-ketoglutarate oxidatively decarboxylated was synthesized by the tricarboxylic acid cycle. This is in agreement with the high flux found through citrate synthase. It is also clear from the data of Table VIII that, upon addition of acetate, the increased flux through ␣-ketoglutarate dehydrogenase can be entirely accounted for by the increase in flux through citrate synthase which occurred at the expense of the very large flux through phosphoenolpyruvate carboxykinase. Table VIII also shows that the acetate-induced diversion of oxaloacetate from phosphoenolpyruvate formation to citrate synthesis was logically accompanied by a large decrease in flux through pyruvate kinase and therefore in a decreased availability of pyruvate leading to a diminution of fluxes through alanine aminotransferase and pyruvate dehydrogenase. The increased glucose synthesis (see flux through glucose-6-phosphatase) leading to competition between the gluconeogenic pathway and pyruvate synthesis for utilizing the phosphoenolpyruvate available in decreased amounts in the presence of acetate may also explain in part the decreased availability of pyruvate and of 3-phosphohydroxypyruvate (see flux through 3-phosphoglycerate dehydrogenase in Table VIII).

Removal and Metabolic Fate of Glutamate in Rabbit Kidney Tubules
Our study clearly demonstrates that rabbit kidney tubules readily utilize glutamate as substrate; at equimolar concentration, the rate of glutamate removal observed in the present study is about twice that found for glucose in a recent study (6). This demonstration that glutamate is a substrate of rabbit kidney tubules is in agreement with the findings of Krebs (1) that glutamate is a precursor of glutamine in rabbit kidney slices and of Klahr (9) and Klahr et al. (4) who observed also in rabbit kidney slices that glutamate increases the production of glucose. Our results not only confirm that both glutamine and glucose are products of glutamate metabolism but also establish the relative importance of theses two end products. In addition, they establish the relative importance of alanine and CO 2 as carbon products of glutamate metabolism. From the results obtained, it is clear that glutamate is a potential energy provider to rabbit renal tubular cells.
The main fate of glutamate nitrogen in rabbit kidney tubules is glutamine. Both in the presence of glutamate as sole substrate and in the presence of glutamate ϩ acetate, the significant amount of glutamine formed, which cannot be entirely explained by endogenous formation (see Tables I and II), indicates that added glutamate must have contributed not only to the carbon skeleton but also to its amide nitrogen. This clearly demonstrates that some of the glutamate removed must have been deaminated via glutamate dehydrogenase to provide the ammonia needed for the synthesis of the glutamine amide group. This study also establishes that alanine and, to a much smaller extent, aspartate, and serine are nitrogenous products of glutamate metabolism in rabbit kidney tubules. In agreement with the findings of other authors (1,4,9), no ammonia accumulated, an observation explainable by the high rate of glutamine synthesis that masked the release of ammonia due to glutamate deamination.

Routes of Glutamate Metabolism in Rabbit Kidney
Tubules and Importance of the Information Obtained from the Combination of the 13

C NMR Data and the Model
The advantages of using 13 C NMR spectroscopy, also used by other authors for the study of intermediary metabolism of various organs in vitro and in vivo (8, 20 -36), together with a mathematical model of substrate metabolism (37), have been underlined in a previous study (6). As will be seen below and in the accompanying paper (44), further developments of our modeling approach allowed us to measure not only fluxes in three different cycles as in our previous study (6) but also unidirectional fluxes at the level of glutamate dehydrogenase allowing us therefore to measure enzymatic fluxes in a fourth cycle, i.e. the "Glu 3 ␣KG 3 Glu" cycle. Note that Chance and coworkers and Shulman and co-workers (23,31,33) were also able to assess the rate of the exchange between ␣-ketoglutarate and glutamate in the rat and human brain in vivo. Unfortunately, flux through glutamine synthetase was lower than the total glutamine found at the end of the incubation period. Therefore, we obtained no experimental evidence in the present study that the glutamine synthesized from glutamate was metabolized by glutaminase which is also active in rabbit kidney cortex (1-3); otherwise, we would have been able to measure simultaneously enzymatic fluxes through five different cycles operating during glutamate metabolism in rabbit kidney tubules as considered in our mathematical model of glutamate metabolism in the accompanying paper (44).
Note that the consistency of our analysis and measurements was satisfactory because there were no contradictions either in the qualitative or in the quantitative distributions of the different labeled glutamate carbons in the different products. For example the symmetrical behavior beyond the succinyl-coenzyme A step on the one hand of C-2 and C-5 and on the other hand of the C-3 and C-4 of glutamate was observed from differently 13 C-and 14 C-labeled glutamates. In addition the number of experimental results was greater than the number of parameters in the model; therefore, several ways of calculations giving virtually identical results were possible. Furthermore, the consistency of our model is also indicated by the fact that calculated and measured amounts of a few labeled products were also almost identical. Finally, virtually identical fluxes were obtained when they were calculated from the data obtained from differently labeled glutamates.
Role of Glutamate Dehydrogenase-In these experiments, the release of 14 CO 2 from [1-14 C]glutamate, which cannot be explained by the rates of alanine, aspartate, and serine synthesis (see Tables II and III), provides conclusive evidence that glutamate dehydrogenase plays an important role in initiating glutamate metabolism in rabbit kidney tubules. Our data therefore are in agreement with the view of Krebs (1) about the role of this enzyme, but they disagree with the conclusion of Klahr (9) and Klahr et al. (4) who did not notice that the release of ammonia by oxidative deamination of glutamate was masked by the high capacity of rabbit kidney-cortex slices for glutamine synthesis. Additional evidence for a high flux through glutamate dehydrogenase in the direction of oxidative deamination of glutamate is provided by the results drawn from the combination of the 13 C NMR data and the model of glutamate metabolism developed in this study (Table VIII).
Simultaneous Degradation and Synthesis of Glutamate-The fact that, in the presence of glutamate ϩ acetate, the release of 14 CO 2 from [1-14 C]glutamate, or half the release of 14 CO 2 from [1,5-14 C]glutamate, or one-fifth of the release of 14 CO 2 from [U-14 C]glutamate was higher than the removal of glutamate measured enzymatically (Tables II and III) strongly suggests that glutamate removal was accompanied by glutamate synthesis. That this was indeed the case not only in the presence of glutamate ϩ acetate but also in the presence of glutamate as sole substrate was unequivocally demonstrated by the conversion of the C-3 of added [3-13 C]glutamate to the C-1, C-2, C-4, and C-5 of glutamate (Table IV). As shown in Table VIII, flux through glutamate dehydrogenase in the direction of the reductive amination of ␣-ketoglutarate was high both in the absence and the presence of acetate; however, as expected since glutamate was removed in net amounts under both experimental conditions, flux in the oxidative deamination direction exceeded that in the reductive amination direction. The net flux through glutamate dehydrogenase, which was therefore in favor of glutamate deamination (Table VIII), gives the availability of ammonia for glutamine synthesis. It can be seen that, both in the absence and the presence of acetate, the ammonia derived from added glutamate and available for glutamine synthesis was not sufficient to explain all the glutamine found to accumulate (Table II); in agreement with the nitrogen balance calculations mentioned above, this means that the glutamine synthesis from endogenous substrates observed in the absence of exogenous substrates still occurred (at least partially) in the presence of glutamate alone and in the presence of glutamate ϩ acetate (Table II).
The fact that glutamate carbons other than those labeled in the added glutamates became labeled during incubation (Tables IV-VI) implies that the ␣-ketoglutarate derived from added glutamate as a result of glutamate deamination or transamination was further metabolized in the tricarboxylic acid cycle to yield newly synthesized glutamate. Thus, at any incubation time point, the glutamate added at the start of incubation was mixed with small amounts of variously labeled glutamates synthesized by rabbit kidney tubules through various pathways. Thanks to the NMR data we were able to identify these pathways and to quantify them by combining the NMR data and the model developed. An interesting observation made from the data of Table IV (Table VI) found on the other hand, both with glutamate alone or with glutamate ϩ acetate.
Note that the two reactions of the glutamate Ϫ ␣-ketoglutarate interconversion could occur within one or different populations of mitochondria or within different cells. In this respect, it should also be emphasized here that no evidence was found for different pools of glutamate as will be discussed in the next section.
Synthesis of Glutamine-Although we recently presented some evidence that, in the presence of glucose plus NH 4 Cl, glutamine accumulation resulting from the action of glutamine synthetase was accompanied by glutamine degradation by glutaminase (6), we obtained in the present study no evidence that glutaminase operated; therefore, we were unable to quantify flux through the latter enzyme which is present along the entire length of the rabbit proximal tubule (3).
In this study, we were able to distinguish between the direct (from added glutamate) and indirect (from synthesized glutamate) synthesis of glutamine (see Tables IV-VI and "Results" ). This raises the question of whether or not glutamine was synthesized from a single or from two different glutamate pools. In this respect, the observations made from Tables IV-VI that (i) acetate increased the accumulation of [2-13 C]-and [3-13 C]glutamate from [3-13 C]glutamate but not the indirect accumulation of [2-13 C]-and [3-13 C]glutamine from the latter glutamates, and (ii) acetate caused a decrease in the direct accumulation of glutamine (Tables IV-VI) but not in the indirect accumulation of this amino acid should be examined with great caution. At first sight, the latter observations might be interpreted as indicative of the existence of two different glutamate pools from which two different glutamine pools would be synthesized. In fact, the absence of a (2-fold) decrease in the indirect accumulation of [2-13 C]-and [3-13 C]glutamine in the presence of acetate can simply be explained by a large (2-fold) increase in the net synthesis of glutamate as reflected by the accumulation of [2-13 C]glutamate ϩ [2-13 C]glutamine (see Table IV). Thus, there is no evidence in the present study for the existence of two different glutamate or glutamine pools in rabbit kidney tubules.
Note that it is conceivable that the glutaminase reaction also operated especially in the presence of glutamate as sole substrate when the glutamine concentration reached a level higher (Table II) than that at which the glutaminase reaction has been demonstrated to function in rabbit kidney tubules (5); however, the extent to which this reaction operated, if at all, remains uncertain because of the presence of high concentrations of glutamate, a well established end product inhibitor of renal glutaminase (38). It should also be mentioned here that any operation of the glutaminase reaction in the present study would imply that our value of flux through glutamine synthetase is underestimated.
Other Pathways of Glutamate Metabolism-An interesting observation is that the "substrate cycle" involving the operation of phosphoenolpyruvate carboxykinase, pyruvate kinase, and pyruvate carboxylase, which has recently been shown to operate during glucose metabolism in rabbit kidney tubules (6), was also found in the present study to operate at high rates during glutamate metabolism (Tables VII and VIII). Note that the operation of pyruvate carboxylase, which resulted in the labeling of the C-4 of oxaloacetate and then of the C-1 of glutamate and glutamine in the presence of glutamate ϩ [ 14 C]bicarbonate (Table III), cannot explain the labeling of the C-1 of alanine, glutamate, and glutamine as a result of the refixation of the 13 CO 2 released from the differently 13 C-labeled glutamates used as substrates; as a matter of fact, simple calculations taking into account the total "CO 2 ϩ bicarbonate" pool of our flasks reveal that the 13 CO 2 released was considerably diluted by unlabeled CO 2 and could account for only negligible labelings of the C-1 of the three latter amino acids.
Of the pyruvate synthesized by pyruvate kinase, a significant fraction was metabolized by pyruvate dehydrogenase; this reveals the existence of an additional cycle involving the successive operation of the latter enzyme, the tricarboxylic acid cycle from citrate to oxaloacetate, phosphoenolpyruvate carboxykinase, and pyruvate kinase. Thus, glutamate metabolism in rabbit kidney tubules involves four different cycles; in addition to the latter cycle, the "Glu 3 ␣KG 3 Glu", the tricarboxylic acid and the "OAA 3 PEP 3 Pyr 3 OAA" cycles are all operative at their own rate which was determined by the enzyme operating at the lowest rate.
The fact that, with glutamate as sole substrate, flux through alanine aminotransferase (Table VIII) was found slightly higher than alanine accumulation (Table II) can be simply explained by the fact that some of the alanine brought by the tubules at the start of incubation was metabolized and replaced by the alanine synthesized from glutamate; this suggests that bidirectional fluxes also occurred at the level of alanine aminotransferase, a fact that is not surprising for a near-equilibrium enzyme.
That some glucose was synthesized from glutamate in our experiments is in agreement with the existence of key gluconeogenic enzymes in the rabbit proximal tubule (39). The small amount of [ 13 C]serine found suggests that, like in the rat kidney (40), 3-phosphoglycerate dehydrogenase, phosphoserine aminotransferase, and phosphoserine phosphatase are also functional in rabbit kidney tubules. It should be mentioned here that, given the very small amounts of glucose synthesized from glutamate in this study, no account was taken of the possible utilization of glucose previously demonstrated in rabbit kidney tubules (6,41).

Effect of Acetate on Glutamate Metabolism
Our data indicate that the rabbit kidney, like the kidney of other species (42,43), has a high capacity to utilize acetate as substrate (Table II). Furthermore, they show that acetate has a number of important regulatory effects on the metabolism of glutamate in rabbit kidney tubules (Table VIII). The acetateinduced stimulation of flux through glutamate dehydrogenase in the oxidative deamination direction is a priori surprising because one would rather expect an inhibition of this unidirectional flux as a result of the expected shift of the mitochondrial redox potential toward a more reduced state secondary to acetate oxidation; one partial explanation may be an increased availability of glutamate for its oxidative deamination as a result of the inhibition of glutamate utilization by the glutamine synthesis and glutamate transamination pathways. The stimulation by acetate of the unidirectional flux through glutamate dehydrogenase in the reductive amination direction may have resulted from (i) an increased supply of ␣-ketoglutarate secondary to the stimulation of flux through citrate synthase caused by acetate and (ii) the more reduced state of the mitochondrial NADH/NAD ϩ couple mentioned above.
The inhibition of flux through glutamine synthetase caused by acetate cannot be attributed to a diminution of glutamate availability which on the contrary increased. Similarly, the provision of ATP (which can be calculated from the data of Table VIII), which is needed for the glutamine synthetase reaction (1), was not decreased but rather was more than doubled in the presence of acetate. This inhibition can be explained by a significant reduction of the release of ammonia, a substrate of the glutamine synthetase reaction, by glutamate dehydrogenase (Table VIII).
It is noteworthy that, despite the increased provision of oxaloacetate via the tricarboxylic acid cycle secondary to the acetate-induced stimulation of flux through ␣-ketoglutarate dehydrogenase, there was no increase in aspartate accumulation and a large reduction of flux through phosphoenolpyruvate carboxykinase (Table VIII). This can be explained by the increased availability of acetyl-CoA due to acetate utilization which directed oxaloacetate to the citrate synthase reaction at the expense of the aspartate aminotransferase and the phosphoenolpyruvate carboxykinase reactions. The inhibition of phosphoenolpyruvate synthesis caused by acetate led to a reduction of flux through pyruvate kinase and therefore of the synthesis of pyruvate which explains the observed diminution of fluxes through both pyruvate carboxylase and pyruvate dehydrogenase as well as the fall of alanine accumulation. It should be mentioned that, in the presence of acetate, the small fraction of the phosphoenolpyruvate synthesized that was further metabolized in the gluconeogenic pathway was directed to the synthesis of glucose at the expense of the synthesis of 3-phosphohydroxypyruvate (see fluxes through 3-phosphoglycerate dehydrogenase and glucose-6-phosphatase, Table VIII). This effect may result from an increased flux through glyceraldehyde-3-phosphate dehydrogenase due to an increased availability of NADH in the cytosol as a result of acetate oxidation. This is suggested by the unchanged lactate accumulation despite a reduced availability of pyruvate in the presence of acetate. Another possible explanation for the increased accumulation of glucose in the presence of acetate is that the latter compound inhibited the reutilization of the glucose synthesized since it is known that glucose is utilized by rabbit kidney tubules (6,41) and that acetate inhibits glucose utilization in these tubules. 2 Unfortunately, the results obtained in the present study do not allow us to test the latter hypothesis because the glucose synthesis and the 13 C resonances of glucose were too small to draw reliable conclusions on the possible existence of glucose reutilization.

Physiological Significance
It is clear that in the rabbit kidney, like in the kidney of other herbivorous animals, which excretes an alkaline urine (2) and contains a high glutamine synthetase activity, glutamine synthesis represents an important mechanism whereby the ammonia synthesized by the glutamate dehydrogenase and possi-bly glutaminase reactions can be trapped. It is therefore of physiological, biochemical, and nutritional interest to characterize not only the metabolism of glutamate, a substrate of glutamine synthetase and a regulator of glutaminase, but also the regulation of glutamate metabolism by important circulating substrates such acetate, which represents most of the volatile fatty acids in the rabbit blood (18).
It should be emphasized here that the glutamate concentration used in the present study is much higher than the blood glutamate concentration found in the rabbit (0.2 mM). 2 We used such a high glutamate concentration and a large amount of kidney tubules to increase the rate of glutamate removal and the rate of 13 C-labeled product formation so that it was possible to measure substrate removal and product formation in a reliable manner to compensate for the poor sensitivity of 13 C NMR spectroscopy. This led us also to use a 10 mM acetate concentration so that this substrate, which is avidly metabolized by rabbit kidney tubules (Table II) and is present at millimolar concentrations in the rabbit blood (18), was available in sufficient amounts until the end of the 60-min incubation period.
It should be underlined that the effects of acetate observed in the present study are probably of physiological relevance; indeed, the metabolism of sodium acetate leads to the production of bicarbonate which should be eliminated in the urine, thus leading to urine alkalinization and therefore to a decreased need for renal ammonia production and excretion. In the present study, the decreased flux through glutamate dehydrogenase (Table VIII), which releases ammonia, caused by the presence of acetate fits well with a physiological adaptation to the metabolic alkalosis caused by acetate oxidation. In this respect, one would also expect an increased flux through glutamine synthetase and an increased glutamine synthesis which, on the contrary, were found to decrease in this study (Table II and VIII). The absence of this expected increase in glutamine synthesis is probably related to the decreased availability of ammonia mentioned above. Whether acetate, which leads to an increased flux through the tricarboxylic acid and the "Glu 3 ␣KG 3 Glu" cycles and to a decreased flux through the "OAA 3 PEP 3 Pyr 3 OAA" and the "OAA 3 PEP 3 Pyr 3 AcCoA