Cerebral glutamate was monitored in a superfused cerebral cortical preparation by 1H NMR spectroscopy using a semiselective spin-echo sequence N-acetyl aspartate (NAA) as an internal concentration reference. During controlled metabolic conditions, the cerebral 1H NMR-detected glutamate-to-NAA ratio was approximately 20-30% lower than expected from the ratio of neutralized perchloric acid extracts of the preparations. Inhibition of respiration in the presence of glucose did not change the 1H NMR glutamate-to-NAA ratio in brain slice preparation. In contrast, either complete depletion of ATP during cyanide poisoning together with 0 mM glucose, anoxia in the absence of glucose, or treatment with nigericin or with a protonophore, carbonyl cyanide-m-fluorophenylhydrazone, increased 1H NMR-detected glutamate/NAA in the cerebral preparations without a change in the relative and absolute concentration ratios determined from the tissue acid extracts. Spin-spin relaxation times of glutamate and NAA peaks in anoxic slices were 749 +/- 89 and 729 +/- 94 ms, respectively, and thus, the portion of glutamate that could not be detected by 1H NMR was quantified in absolute terms. It was calculated that an increase in the glutamate-to-NAA ratio from 0.55 +/- 0.02 to 0.67 +/- 0.02 during aglycemic anoxia corresponded to some 6 mmol/kg of tissue dry weight of glutamate from the total concentration of 28 mmol/kg dry weight. It is suggested that this 22% of total glutamate pool is present in a noncytoplasmic compartment during controlled metabolic state.