Biosynthetic preparation of L-[13C]- and [15N]glutamate by Brevibacterium flavum.

The biosynthesis of isotopically labeled L-glutamic acid by the microorganism Brevibacterium flavum was studied with a variety of carbon-13-enriched precursors. The purpose of this study was twofold: to develop techniques for the efficient preparation of labeled L-glutamate with a variety of useful labeling patterns which can be used for other metabolic studies, and to better understand the metabolic events leading to label scrambling in these strains. B. flavum, which is used commercially for the production of monosodium glutamate, has the capability of utilizing glucose or acetate as a sole carbon source, an important criterion from the standpoint of developing labeling strategies. Unfortunately, singly labeled glucose precursors lead to excessive isotopic dilution which reduces their usefulness. Studies with [3-13C]pyruvate indicate that this problem can in principle be overcome by using labeled three-carbon precursors; however, conditions could not be found which would lead to an acceptable yield of isotopically labeled L-glutamate. In contrast, [1-13C]- or [2-13C]acetate provides relatively inexpensive, readily available precursors for the production of selectively labeled, highly enriched L-glutamate. The preparation of L-[15N]glutamate from [15N]ammonium sulfate was carried out and is a very effective labeling strategy. Analysis of the isotopic distribution in labeled glutamate provides details about the metabolic pathways in these interesting organisms.     

Microbial preparation of L-[15N]tyrosine and [15N]tyramine and their gas chromatographic-mass spectrometric analyses

The preparation of L-[15N]tyrosine and [15N]tyramine by microbial synthesis is described. Immobilized Erwinia herbicola cells were added to a reaction mixture containing phenol, pyruvic acid, and 15NH4Cl. The reaction was driven by excess nonlabeled pyruvate and phenol. Under these denaturing concentrations of phenol, immobilized cells were more effective than free ones. Gram quantities of L-[15N]tyrosine were obtained without label dilution. The conversion of this L-[15N]tyrosine into [15N]tyramine by Streptococcus faecalis was performed at maximal efficiency. Gas chromatographic-mass spectrometric studies and 1H and 15N NMR analyses of the labeled compounds are reported.     

A 13C-NMR study on the influxes into the tricarboxylic acid cycle of a renal epithelial cell line, LLC-PK1/Cl4: the metabolism of [2-13C]glycine, L-[3-13C]alanine and L-[3-13C]aspartic acid in renal epithelial cells

Perchloric acid extracts of LLC-PK1/Cl4 cells, a renal epithelial cell line, incubated with either [2-13C]glycine L-[3-13C]alanine, or D,L-[3-13C]aspartic acid were investigated by 13C-NMR spectroscopy. All amino acids, except labelled glycine, gave rise to glycolytic products and tricarboxylic acid cycle (TCA) intermediates. For the first time we also observed activity of gamma-glutamyltransferase activity and glutathione synthetase activity in LLC-PK1 cells, as is evident from enrichment of reduced glutathione. Time courses showed that only 6% of the labelled glycine was utilized in 30 min, whereas 31% of L-alanine and 60% of L-aspartic acid was utilized during the same period. 13C-NMR was also shown to be a useful tool for the determination of amino acid uptake in LLC-PK1 cells. These uptake experiments indicated that glycine, alanine and aspartic acid are transported into Cl4 cells via a sodium-dependent process. From the relative enrichment of the glutamate carbons, we calculated the activity of pyruvate dehydrogenase to be about 61% when labelled L-alanine was the only carbon source for LLC-PK1/Cl4 cells. Experiments with labelled D,L-aspartic, however, showed that about 40% of C-3-enriched oxaloacetate (arising from a de-amination of aspartic acid) reached the pyruvate pool.     

Microbial production of L-[15N]glutamic acid and its gas chromatography-mass spectrometry analysis

L-[15N]Glutamic acid was prepared in high yields via a fermentative process. Brevibacterium lactofermentum, growing on a medium containing 97% enriched 15NH4Cl as a sole isotopic precursor, excreted mostly L-[15N]glutamic acid. The L-[15N]glutamic acid was purified and identified. Gas chromatography-mass spectrometry analysis was performed to demonstrate its usefulness in clinical studies.     

Auto-induction medium for the production of [U-15N]- and [U-13C, U-15N]-labeled proteins for NMR screening and structure determination

Protocols have been developed and applied for the high-throughput production of [U-15N]- or [U-13C-, U-15N]-labeled proteins using the conditional methionine auxotroph Escherichia coli B834. The large-scale growth and expression uses a chemically defined auto-induction medium containing salts and trace metals, vitamins including vitamin B12, and glucose, glycerol, and lactose. The results from nine expression trials in 2-L of the auto-induction medium (500 mL in each of four polyethylene terephthalate beverage bottles) gave an average final optical density at 600 nm of approximately 5, an average wet cell mass yield of approximately 9.5 g L(-1), and an average yield of approximately 20 mg of labeled protein in the six instances in which proteolysis of the fusion protein was observed. Correlations between the cell mass recovered, the level of protein expression, and the relative amounts of glucose, glycerol, and lactose in the auto-induction medium were noted. Mass spectral analysis showed that the purified proteins contained both 15N and 13C at levels greater than 95%. 1H-15N heteronuclear single quantum correlation spectroscopy as well as 13C; 15N-edited spectroscopy showed that the purified [U-15N]- and [U-13C, U-15N]-labeled proteins were suitable for structure analysis.     

Microbial fermentative preparation of L-[15N2]lysine and its tracer: application to serum amino acid kinetic studies

The microorganism Brevibacterium flavum 21129 has been used to produce multigram batches of L-[15N2]lysine of high purity and isotopic enrichment by supplementation of the growth medium with (15NH4)2SO4 of 98.0 atom% excess. The doubly 15N-labeled lysine can be detected at dilutions 10 times greater than singly labeled lysine when isotope dilution curves are analyzed by gas chromatography-mass spectrometry. This enhanced sensitivity permits kinetic measurements of plasma free-lysine isotope content over a 300-fold dilution during 6 h following a single oral bolus of 5 mg/kg body wt. This inexpensive preparation method lends itself to the production of highly useful biochemical compounds for kinetic studies of human nutrition.     

Nitrogen-13 flux from L-[13N]glutamate in the isolated rabbit heart: effect of substrates and transaminase inhibition

Kinetic and biochemical parameters of nitrogen-13 flux from L-[13N]glutamate in myocardium were examined. Tissue radioactivity kinetics and chemical analyses were determined after bolus injection of L-[13N]glutamate into isolated arterially perfused interventricular septa under various metabolic states, which included addition of lactate, pyruvate, aminooxyacetate (a transaminase inhibitor), or a combination of aminooxyacetate and pyruvate to the standard perfusate containing insulin and glucose. Chemical analysis of tissue and effluent at 6 min allowed determination of the composition of the slow third kinetic component of the time-activity curves. 13N-labeled aspartate, alanine and glutamate accounted for more than 80% of the tissue nitrogen-13 under the experimental conditions used. Specific activities for these amino acids were constant, but not identical to each other, from 6 through 15 min after administration of L-[13N]glutamate. Little labeled ammonia (1.9%) and glutamine (4.7%) were produced, indicating limited accessibility of exogenous glutamate to catabolic mitochondrial glutamate dehydrogenase and glutamine synthetase, under control conditions. Lactate and pyruvate additions did not affect tissue amino acid specific activities. Aminooxyacetate suppressed formation of 13N-labeled alanine and aspartate and increased production of L-[13N]glutamine and [13N]ammonia. Formation of [13N]ammonia was, however, substantially decreased when aminooxyacetate was used in the presence of exogenous pyruvate. The data support a model for glutamate compartmentation in myocardium not affected by increasing the velocity of enzymatic reactions through increased substrate (i.e., lactate or pyruvate) concentrations but which can be altered by competitive inhibition of transaminases (via aminooxyacetate) making exogenous glutamate more available to other compartments.     

13N]Ammonia and L-[amide-13N]glutamine metabolism in glutaminase-sensitive and glutaminase-resistant murine tumors

The short-term metabolic fate of labeled nitrogen derived from [13N]ammonia or from L-[amide-13N]glutamine was determined in murine tumors known to be resistant (Ridgeway Osteogenic Sarcoma (ROS] or sensitive (Sarcoma-180 (S-180)) to glutaminase therapy. At 5 min after intraperitoneal injection of [13N]ammonia or of L-[amide-13N]glutamine, only about 0.7% of the label recovered in both tumors was in protein and nucleic acid. After [13N]ammonia administration, most of the label (over 80%) was in a metabolized form; a large portion of this metabolized label (50-57%) was in the urea fraction with a smaller amount in glutamine (37-42%). The major short-term fate of label derived from L-[amide-13N]glutamine was incorporation into components of the urea cycle with smaller amounts in the acidic metabolites and in acidic amino acids. No labeled urea was found during in vitro studies in which S-180 tumor slices were incubated with [13N]ammonia, suggesting that the [13N]urea formed in the tumor in the in vivo experiments was not due to de novo synthesis through carbamyl phosphate in the tumor. Both tumors exhibited very low glutamine synthetase activity. Following glutaminase treatment, glutamine synthetase and gamma-glutamyltransferase activities, while remaining low, increased in the resistant tumor but not in the sensitive tumor; this increase may be related to the insensitivity of the ROS tumor toward glutaminase treatment.     

Nitrogen-13 flux from L-[13N]glutamate in the isolated rabbit heart: effect of substrates and transminase inhibition

Kinetic and biochemical parameters of nitrogen-13 flux from L-[¹³N]-glutamate in myocardium were examined. Tissue radioactivity kinetics and chemical analyses were determined after bolus injection of L-[¹³N]glutamate into isolated arterially perfused interventricular septa under various metabolic states, which included addition of lactate, pyruvate, aminooxyacetate (a transminase inhibitor), or a combination of aminooxyacetate and pyruvate to the standard perfusate containing insulin and glucose. Chemical analysis of tissue and effluent at 6 min allowed determination of the composition of the slow third kind kinetic component of the time-activity curves. ¹³N-labeled aspartate, alanine and glutamate accounted for more than 80% of the tissue nitrogen-13 under the experimental conditions used. Specific activities for these amino acids were constant, but not identical to each other, from 6 through 15 min after administration of L-[¹³N]glutamate. Little labeled ammonia (1.9%) and glutamine (4.7%) were produced, indicating limited accessibility of exogenous glutamate to catabolic mitochondrial glutamate dehydrogenase and glutamine synthetase, under control conditions. Lactate and pyruvate additions did not affect tissue amino acid specific activities. Aminooxyacetate suppressed formation of ¹³N-labeled alanine and aspartate and increased production of L-[¹³N]glutamine and [¹³N]ammonia. Formation of [¹³N]ammonia was, however, substantially decreased when aminooxyacetate was used in the presence of exogenous pyruvate. The data support a model for glutamate compartmentation in myocardium not affected by increasing the velocity of enzymatic reactions through increased substrate (i.e., lactate or pyruvate) concentrations but which can be altered by competitive inhibition of transaminases (via aminooxyacetate) making exogenous glutamate more available to other compartments.     

Estimating protein turnover with a [15N,13C]leucine tracer: a study using simulated data.

We explore the use of [15N,13C]leucine tracer to estimate whole-body fractional rates of a fast-turning-over protein pool employing synthetic data. The kinetics of [15N,13C]leucine tracer are simplified compared with those of traditional leucine tracers and benefit from irreversible transamination to [13C]alpha-ketoisocaproaic acid (KIC) resulting in a simplified model structure. A three-compartment model of [15N,13C]leucine kinetics was proposed and evaluated using data generated by a Reference Model (based on a model by Cobelli et al.). The results suggest that fractional turnover rates of a fast-turning-over protein pool can be estimated with a low but acceptable precision during a six-hour constant intravenous infusion of [15N,13C]leucine with frequent sampling of plasma tracer-to-tracee ratio (TTR) of [15N,13C]leucine. We conclude that [15N,13C]leucine may be useful for the measurement of protein kinetics and its full potential should be explored in clinical studies with compartmental data analysis.     

Simultaneous determination of [2-15N]- and [5-15N]glutamine with gas chromatography-mass spectroscopy: applications to nitrogen metabolic studies

A gas chromatographic-mass spectrometric method for analysis of L-[2-15N]- and L-[5-15N]glutamine is described. The method is based on direct acylation of glutamine with trifluoroacetic anhydride and the formation of the N,N-bis-trifluoroacetyl-L-glutamine derivative. This simple and sensitive method is capable of detecting approximately 0.5 atom% excess 15N in as little as 10 microliter of plasma with a mean coefficient of variance of 11.6%. The method was applied to determine the appearance of 15N enrichment in plasma amino-N and amide-N of glutamine in a healthy adult volunteer during a constant infusion of 15NH4Cl. A plateau level of 3.7 and 2.6 atom% excess was observed in amide-N and amino-N, respectively, at 1 and 2 h after 15NH4Cl infusion was started.     

Metabolism of alpha-D-[1,2-13C]glucose pentaacetate and alpha-D-glucose penta[2-13C]acetate in rat hepatocytes

Hepatocytes from fed rats were incubated for 120 min in the presence of alpha-D-[1,2-13C]glucose pentaacetate (1.7 mM), both D-[1,2-13C]glucose (1.7 mM) and acetate (8.5 mM), alpha-D-glucose penta[2-13C]acetate (1.7 mM), or D-[1,2-13C]glucose (8.3 mM). The amounts of 13C-enriched L-lactate and D-glucose and those of acetate and beta-hydroxybutyrate recovered in the incubation medium were comparable under the first two experimental conditions. The vast majority of D-glucose isotopomers consisted of alpha- and beta-D[1,2-13C]glucose. The less abundant single-labeled isotopomers of D-glucose were equally labeled on each C atom. The output of 13C-labeled L-lactate, mainly L-[2-13C]lactate and L-[3-13C]lactate, was 1 order of magnitude lower than that found in hepatocytes exposed to 8.3 mM D-[1,2-13C]glucose, in which case the total production of the single-labeled species of D-glucose was also increased and that of the C3- or C4-labeled hexose was lower than that of the other 13C-labeled isotopomers. In cells exposed to alpha-D-glucose penta[2-13C]acetate, the large majority of 13C atoms was recovered as [2-13C]acetate and, to a much lesser extent, beta-hydroxybutyrate labeled in position 2 and/or 4. Nevertheless, L-[2-13C]lactate, L-[3-13C]lactate, and single-labeled D-glucose isotopomers were also produced in amounts higher or comparable to those found in cells exposed to alpha-D-[1,2-13C]glucose pentaacetate. However, a modest preferential labelling of the C6-C5-C4 moiety of D-glucose, relative to its C1-C2-C3 moiety, and a lesser isotopic enrichment of the C3 (or C4), relative to that of C1 (or C6) and C2 (or C5), were now observed. These findings indicate that, despite extensive hydrolysis of alpha-D-glucose pentaacetate (1.7 mM) in the hepatocytes, the catabolism of its D-glucose moiety is not more efficient than that of unesterified D-glucose, tested at the same molar concentration (1.7 mM) in the presence of the same molar concentration of unesterified acetate (8.5 mM), and much lower than that found at a physiological concentration of the hexose (8.3 mM). The present results also argue against any significant back-and-forth interconversion of D-glucose 6-phosphate and triose phosphates, under conditions in which sizeable amounts of D-glucose are formed de novo from 13C-enriched Krebs cycle intermediates generated from either D-[1,2-13C]glucose or [2-13C]acetate.     

Incorporation of Label from 13C-, 2H-, and 15N-Labeled Methionine Molecules during the Biosynthesis of 2[prime]-Deoxymugineic Acid in Roots of Wheat

The biosynthetic pathway of 2[prime]-deoxymugineic acid, a key phytosiderophore, was investigated by feeding 13C-, 2H-, and 15N-labeled methionine, the first precursor, to the roots of hydroponically cultured wheat (Triticum aestivum L. cv Minori). The incorporation of label from each methionine species was observed during their conversion to 2[prime]-deoxymugineic acid, using 2H-, 15N-, and 13C-nuclear magnetic resonance (NMR). L-[1-13C]Methionine (99% 13C) was efficiently incorporated, resulting in 13C enrichment of the three carboxyl groups of 2[prime]-deoxymugineic acid. Use of D,L-[15N]methionine (95% 15N) resulted in 15N enrichment of 2[prime]-deoxymugineic acid at the azetidine ring nitrogen and the secondary amino nitrogen. When D,L-[2,3,3,-2H3-S-methyl-2H3]methionine (98.2% 2H) was fed to the roots, 2H-NMR results indicated that only six deuterium atoms were incorporated, and that the deuterium atom from the C-2 position of each methionine was almost completely lost. [2,2,3,3-2H4]1-Aminocyclopropane-1-carboxylic acid (98% 2H) was not incorporated into 2[prime]-deoxymugineic acid. These data and our previous findings demonstrated that only the deuterium atom from the C-2 position of L-methionine was lost, and that other atoms were completely incorporated when three molecules of methionine were converted to 2[prime]-deoxymugineic acid. These observations are consistent with the conversion of L-methionine to azetidine-2-carboxylic acid, suggesting that L-methionine is first converted to azetidine-2-carboxylic acid during biosynthesis leading to 2[prime]-deoxymugineic acid. Based on these results, a hypothetical pathway from L-methionine to 2[prime]-deoxymugineic acid was postulated.     

Efficient syntheses of [3-15N]uracil and [3-15N]thymine.

Regiospecific syntheses of [3-15N]uracil and [3-15N]thymine are described using [15N]ammonium sulfate as a source of labeled nitrogen. The overall yields are excellent, and the reactions are amenable to production of multigram quantities of labeled material.     

Characterization of peptidoglycan stem lengths by solid-state 13C and 15N NMR

Lyophilized whole cells of Aerococcus viridans (Gaffkya homari) grown on a synthetic medium containing D-[2-13C, 15N]Ala, or containing both L-[1-13C]Lys and D-[15N]Ala, have been examined by double cross-polarization magic-angle spinning 13C and 15N nuclear magnetic resonance. Results from the double-labeled alanine experiment confirm the absence of metabolic scrambling of alanine by A. viridans. Results from the combined single-label experiment can be used to count directly the number of adjacent L-Lys and D-Ala units in peptide chains of cell-wall peptidoglycan. This count leads to the conclusion that there are no terminal D-Ala or D-Ala-D-Ala units in uncross-linked chains of the peptidoglycan of A. viridans.     

Growth of Paracoccus denitrificans on [2,3-13C]succinate and [1,4-13C]succinate. II. Isoleucine biosynthesis

Paracoccus denitrificans was grown on either [2,3-13C]succinate or [1,4-13C]succinate, and extracts were analysed by using gas chromatography-mass spectrometry. The distribution of label in isoleucine indicated that the 2-ketobutyrate required for isoleucine biosynthesis was mainly produced from pyruvate by 2-keto-acid chain elongation (i.e. the 'pyruvate elongation pathway'). Approximately 10% of isoleucine was produced by a second pathway involving propionyl CoA. Threonine and glutamate were not utilized by P. denitrificans as a source of 2-ketobutyrate production for isoleucine biosynthesis under the growth conditions used.     

Efficient methods for the synthesis of [2-15N]guanosine and 2'-deoxy[2-15N]guanosine derivatives

The nucleophilic addition-elimination reaction of 2',3',5'-tri-O-acetyl-2-fluoro-O6-[2-(4-nitrophenyl)ethyl]inosine (8) with [15N]benzylamine in the presence of triethylamine afforded the N2-benzyl[2-15N]guanosine derivative (13) in a high yield, which was further converted into the N2-benzoyl[2-15N] guanosine derivative by treatment with ruthenium trichloride and tetrabutylammonium periodate. A similar sequence of reactions of 2',3',5'-tri-O-acetyl-2-fluoro-06-[2-(methylthio)ethyl]inosine (9) and the 6-chloro-2-fluoro-9-(beta-D-ribofuranosyl)-9H-purine derivative (11), which were respectively prepared from guanosine, with potassium [15N]phthalimide afforded the N2-phthaloyl [2-15N]guanosine derivative (15; 62%) and 9-(2,3,5-tri-O-acetyl-beta-D-ribofuranosyl)-6-chloro-2-[15N]phthalimido-9H-purine (17; 64%), respectively. Compounds 15 and 17 were then efficiently converted into 2',3',5'-tri-O-acetyl [2-15N]guanosine. The corresponding 2'-deoxy derivatives (16 and 18) were also synthesized through similar procedures.     

Complex Glutamate Labeling from [U-13C]glucose or [U-13C]lactate in Co-cultures of Cerebellar Neurons and Astrocytes

Glutamate metabolism was studied in co-cultures of mouse cerebellar neurons (predominantly glutamatergic) and astrocytes. One set of cultures was superfused (90 min) in the presence of either [U-¹³C]glucose (2.5 mM) and lactate (1 mM) or [U-¹³C]lactate (1 mM) and glucose (2.5 mM). Other sets of cultures were incubated in medium containing [U-¹³C]lactate (1 mM) and glucose (2.5 mM) for 4 h. Regardless of the experimental conditions cell extracts were analyzed using mass spectrometry and nuclear magnetic resonance spectroscopy. ¹³C labeling of glutamate was much higher than that of glutamine under all experimental conditions indicating that acetyl-CoA from both lactate and glucose was preferentially metabolized in the neurons. Aspartate labeling was similar to that of glutamate, especially when [U-¹³C]glucose was the substrate. Labeling of glutamate, aspartate and glutamine was lower in the cells incubated with [U-¹³C]lactate. The first part of the pyruvate recycling pathway, pyruvate formation, was detected in singlet and doublet labeling of alanine under all experimental conditions. However, full recycling, detectable in singlet labeling of glutamate in the C-4 position was only quantifiable in the superfused cells both from [U-¹³C]glucose and [U-¹³C]lactate. Lactate and alanine were mostly uniformly labeled and labeling of alanine was the same regardless of the labeled substrate present and higher than that of lactate when superfused in the presence of [U-¹³C]glucose. These results show that metabolism of pyruvate, the precursor for lactate, alanine and acetyl-CoA is highly compartmentalized. Special issue dedicated to John P. Blass.     

Intestinal renal metabolism of L-citrulline and L-arginine following enteral or parenteral infusion of L-alanyl-L-[2,15N]glutamine or L-[2,15N]glutamine in mice

Previously, we observed increased plasma arginine (ARG) concentrations after glutamine (GLN)-enriched diets, in combination with clinical benefits. GLN delivers nitrogen for ARG synthesis, and the present study was designed to quantify the interorgan relationship of exogenous L-GLN or GLN dipeptide, by enteral or parenteral route, contributing to intestinal citrulline (CIT) and renal de novo ARG synthesis in mice. To study this, we used a multicatheterized mouse model with Swiss mice (n = 43) in the postabsorptive state. Stable isotopes were infused into the jugular vein or into the duodenum {per group either free L-[2,(15)N]GLN or dipeptide L-ALA-L-[2,(15)N]GLN, all with L-[ureido-(13)C-(2)H(2)]CIT and L-[guanidino-(15)N(2)-(2)H(2)]ARG} to establish renal and intestinal ARG and CIT metabolism. Blood flow was measured using (14)C-para-aminohippuric acid. Net intestinal CIT release, renal uptake of CIT, and net renal ARG efflux was found, as assessed by arteriovenous flux measurements. Quantitatively, more de novo L-[2,(15)N]CIT was produced when free L-[2,(15)N]GLN was given than when L-ALA-L-[2,(15)N]GLN was given, whereas renal de novo L-[2,(15)N]ARG was similar in all groups. In conclusion, the intestinal-renal axis is hereby proven in mice in that L-[2,(15)N]GLN or dipeptide were both converted into de novo renal L-[2,(15)N]ARG; however, not all was derived from intestinal L-[2,(15)N]CIT production. In this model, the feeding route and form of GLN did not influence de novo renal ARG production derived from GLN.     

Metabolism of [1-(13)C)glucose and [2-(13)C]acetate in the hypoxic rat brain.

The effects of hypoxia on the metabolism of the central nervous system were investigated in rats submitted to a low oxygen atmosphere (8% O(2); 92% N(2)). [1-(13)C]glucose and [2-(13)C]acetate were used as substrates, this latter being preferentially metabolized by glial cells. After 1-h substrate infusion, the incorporation of 13C in brain metabolites was determined by NMR spectroscopy. Under hypoxia, an important hyperglycemia was noted. As a consequence, when using labeled glucose, the specific enrichment of brain glucose C1 was lower (48.2+/-5.1%) than under normoxia (66.9+/-2.5%). However, relative to this specific enrichment, the (13)C incorporation in amino acids was increased under hypoxia. This suggested primarily a decreased exchange between blood and brain lactate. The glutamate C2/C4 enrichment ratio was higher under hypoxia (0.62+/-0.01) than normoxia (0.51+/-0.06), indicating a lower glutamate turnover relative to the neuronal TCA cycle activity. The glutamine C2/C4 enrichment ratio was also higher under hypoxia (0.87+/-0.07 instead of 0.65+/-0.11), indicating a new balance in the contributions of different carbon sources at the acetyl-CoA level. When using [2-(13)C]acetate as substrate, no difference in glutamine enrichment appeared under hypoxia, whereas a significant decrease in glutamate, aspartate, alanine and lactate enrichments was noted. This indicated a lower trafficking between astrocytes and neurons and a reduced tricarboxylic acid cycle intermediate recycling of pyruvate.