Some models of brain energy metabolism used to interpret in vivo (13)C nuclear magnetic resonance spectroscopic data assume that intramitochondrial alpha-ketoglutarate is in rapid isotopic equilibrium with total brain glutamate, most of which is cytosolic. If so, the kinetics of changes in (13)C-glutamate can be used to predict citric acid cycle flux. For this to be a valid assumption, the brain mitochondrial transporters of glutamate and alpha-ketoglutarate must operate under physiologic conditions at rates much faster than that of the citric acid cycle. To test the assumption, we incubated brain mitochondria under physiologic conditions, metabolizing both pyruvate and glutamate and measured rates of glutamate, aspartate, and alpha-ketoglutarate transport. Under the conditions employed (66% of maximal O(2) consumption), the rate of synthesis of intramitochondrial alpha-ketoglutarate was 142 nmol/min.mg and the combined initial rate of alpha-ketoglutarate plus glutamate efflux from the mitochondria was 95 nmol/min.mg. It thus seems that much of the alpha-ketoglutarate synthesized within the mitochondria proceeds around the citric acid cycle without equilibrating with cytosolic glutamate. Unless the two pools are in such rapid exchange that they maintain the same percent (13)C enrichment at all points, (13)C enrichment of glutamate alone cannot be used to determine tricarboxylic acid cycle flux. The alpha-ketoglutarate pool is far smaller than the glutamate pool and will therefore approach steady state faster than will glutamate at the metabolite transport rates measured.