Substrate mediated regulation of phosphate-activated glutaminase in nervous tissue

Glutamine uptake by System A transporters maintains neurotransmitter GABA synthesis and inhibitory synaptic transmission

GABA synthesis is necessary to maintain synaptic vesicle filling, and key proteins in its biosynthetic pathways may play a role in regulating inhibitory synaptic stability and strength. GABAergic neurons require a source of precursor glutamate, possibly from glutamine, although it is controversial whether glutamine contributes to the synaptic pool of GABA. Here we report that inhibition of System A glutamine transporters with alpha-(methyl-amino) isobutyric acid rapidly reduced the amplitude of inhibitory post-synaptic currents and miniature inhibitory post-synaptic currents (mIPSCs) recorded in rat hippocampal area cornu ammonis 1 (CA1) pyramidal neurons, indicating that synaptic vesicle content of GABA was reduced. After inhibiting astrocytic glutamine synthesis by either blocking glutamate transporters or the glutamine synthetic enzyme, the effect of alpha-(methyl-amino) isobutyric acid on mIPSC amplitudes was abolished. Exogenous glutamine did not affect mIPSC amplitudes, suggesting that the neuronal transporters are normally saturated. Our findings demonstrate that a constitutive supply of glutamine is provided by astrocytes to inhibitory neurons to maintain vesicle filling. Therefore, glutamine transporters, like those for glutamate, are potential regulators of inhibitory synaptic strength. However, in contrast to glutamate, extracellular glutamine levels are normally high. Therefore, we propose a supportive role for glutamine, even under resting conditions, to maintain GABA vesicle filling.

Relationships between glutamine, glutamate, and GABA in nerve endings under Pb-toxicity conditions

Glutamine (Gln), glutamate (Glu) and gamma-amino butyric acid (GABA) are essential amino acids for brain metabolism and function. Astrocytic-derived glutamine is the precursor of the two most important neurotransmitters: glutamate, an excitatory neurotransmitter, and GABA, an inhibitory neurotransmitter. In addition to their roles in neurotransmission these neurotransmitters act as alternative metabolic substrates that enable metabolic coupling between astrocytes and neurons. The relationships between Gln, Glu and GABA were studied under lead (Pb) toxicity conditions using synaptosomal fractions obtained from adult rat brains to investigate the cause of Pb neurotoxicity-induced seizures. We have found that diminished transport of [(14)C]GABA occurs after Pb treatment. Both uptake and depolarization-evoked release decrease by 40% and 30%, respectively, relative to controls. Lower expression of glutamate decarboxylase (GAD), the GABA synthesizing enzyme, is also observed. In contrast to impaired synaptosomal GABA function, the GABA transporter GAT-1 protein is overexpressed (possibly as a compensative mechanism). Furthermore, similar decreases in synaptosomal uptake of radioactive glutamine and glutamate are observed. However, the K(+)-evoked release of Glu increases by 20% over control values and the quantity of neuronal EAAC1 transporter for glutamate reaches remarkably higher levels after Pb treatment. In addition, Pb induces decreased activity of phosphate-activated glutaminase (PAG), which plays a role in glutamate metabolism. Most noteworthy is that the overexpression and reversed action of the EAAC1 transporter may be the cause of the elevated extracellular glutamate levels. In addition to the impairment of synaptosomal processes of glutamatergic and GABAergic transport, the results indicate perturbed relationships between Gln, Glu and GABA that may be the cause of altered neuronal-astrocytic interactions under conditions of Pb neurotoxicity.

Effects of protein kinase C modulation on NMDA receptor mediated regulation of neurotransmitter enzyme and c-fos protein in cultured neurons

The role of protein kinase C (PKC) in N-methyl-D-aspartate (NMDA) receptor-mediated biochemical differentiation and c-fos protein expression was investigated in cultured cerebellar granule neurons. The biochemical differentiation of glutamatergic granule cells was studied in terms of the specific activity of phosphate-activated glutaminase, an enzyme treatment in the synthesis of the putative neurotransmitter pool of glutamate. When the partially depolarized cells were treated with NMDA for the last 1 to 3 days (between 2 and 5 days in vitro), it elevated the specific activity of glutaminase. In contrast, NMDA had little effect on the activity of aspartate aminotransferase or of lactate dehydrogenase. Treatment of 10-day old granule neurons with NMDA also resulted in a marked increase in the immunocytochemically measured expression of c-fos protein. The increases in both the activity of glutaminase and the steady state level of c-fos protein were specific to the activation of NMDA receptors, as they were completely blocked by D,L-2-amino-5-phosphonovaleric acid. The specific stimulation of NMDA receptors in PKC-depleted granule neurons or in the presence of reasonably specific PKC inhibitors also produced significant elevation in the activity of glutaminase and the expression of c-fos protein. These increases were similar in magnitude to those observed in the granule neurons of the respective control groups. Our findings demonstrate that PKC is not directly involved in the NMDA receptor-mediated signal transduction processes associated with biochemical differentiation and c-fos induction in cerebellar granule neurons.

Regulation of neurotransmitter enzyme by quisqualate subtype glutamate receptors in cultured cerebellar and hippocampal neurons

The possible involvement of ionotropic and metabotropic quisqualate (QA) receptors in neuronal plasticity was studied in cultured glutamatergic cerebellar or hippocampal cells in terms of the specific activity of phosphate-activated glutaminase, an enzyme important in the synthesis of the putative neurotransmitter pool of glutamate. When cerebellar or hippocampal neurons were treated with QA, it elevated the specific activity of glutaminase in a dose-dependent manner. The half-maximal effect was obtained at about 0.1 microM, the maximum increase was at about 1 microM, but levels higher than 10 microM QA produced progressive reduction in glutaminase activity. In contrast, QA had little effects on the activities of lactate dehydrogenase and aspartate aminotransferase and the amount of protein, indicating that the increase in glutaminase was relatively specific. The QA-mediated increase in glutaminase was mimicked by the ionotropic QA receptor agonist alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA; EC50, about 0.5 microM), but not by the metabotropic QA receptor agonist trans-(+-)-1-amino-cyclopentyl-1,3,dicarboxylate (t-ACPD; up to 0.5 mM). The specific ionotropic QA receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) inhibited QA- and AMPA-mediated increases in glutaminase activity in a dose-dependent manner, whereas other glutamate receptor antagonists, D,L-2-amino-5-phosphonovalerate, gamma-D-glutamyl aminomethyl sulphonic acid and gamma-D-glutamyl diethyl ester were ineffective. The elevation of neurotransmitter enzyme was Ca(2+)-dependent.(ABSTRACT TRUNCATED AT 250 WORDS)

GABA synthesis in brain slices is dependent on glutamine produced in astrocytes

The rate of gamma-aminobutyric acid (GABA) synthesis in rat-brain slices was determined by inhibiting GABA transaminase with 20-microM gabaculine and measuring the increase of GABA. Added 500-microM glutamine increased the rate of GABA synthesis by 50%, indicating that glutamate decarboxylase is not saturated in brain slices. The stimulation of GABA synthesis with added glutamine in brain slices was much less than that reported for synaptosomes. The lower stimulation in slices was attributable to astrocytic glutamine production, as the rate of GABA synthesis decreased by 44% when glutamine production was inhibited with methionine sulfoximine. Added glutamine restored the rate to the maximal value observed in brain slices. The rate of GABA synthesis was decreased by 65% in slices pretreated with an inhibitor of glutaminase, and added glutamine did not reverse this effect. These results suggest that glutamine produced by astrocytes is a quantitatively important precursor of GABA synthesis in cortical slices.

Cell-type specific effects of N-methyl-D-aspartate on biochemical differentiation of subcortical neurons in culture

The possible involvement of N-methyl-D-aspartate (NMDA) receptors in the biochemical differentiation of cultured neurons derived from the medial frontal part of the forebrain containing the septum-diagonal band region was studied in terms of the activities of enzymes important in the synthesis of neurotransmitter compounds. The activity of choline acetyltransferase (ChAT) was used as a marker for cholinergic neurons, glutamate decarboxylase (GAD) for GABAergic neurons and phosphate-activated glutaminase (GLNase) and aspartate aminotransferase (ASP-AT) for glutamatergic neurons, while lactate dehydrogenase (LDH) was included as an ubiquitous enzyme. The exposure of cultures to a depolarizing concentration of K+ (40 mM) for the last 3 days (i.e. between 2 and 5 days in vitro) significantly enhanced the expression of ChAT, GAD and GLNase activities, but high K+ caused little alteration in the activities of ASP-AT and LDH. On the other hand, treatment with NMDA markedly elevated the specific activities of GAD and GLNase only, and the compound had no significant effects on the activities of ChAT, ASP-AT and LDH enzymes. The enhancements of the specific activities of GAD and GLNase were completely blocked by the NMDA receptor antagonist, 2-amino-5-phosphonovaleric acid, and by the NMDA receptor-linked Ca2+ ion channel blocker, MK-801. On the basis of the present findings it is concluded that, (a) contrary to an earlier proposal, ASP-AT does not appear to be a good marker for the glutamatergic neurons, (b) the failure of the subcortical cholinergic neurons to respond by an increase in ChAT activity to NMDA may indicate that these nerve cells lack NMDA subtype excitatory amino acid receptors, and (c) as the septal GABAergic input in the hippocampus is involved in the modulation of long-term potentiation, the presence of NMDA receptors on these neurons would now suggest that NMDA receptors are linked to both the initiation and the modulation of hippocampal plasticity in the mammalian brain.

Stimulation of the N-methyl-D-aspartate receptor promotes the biochemical differentiation of cerebellar granule neurons and not astrocytes

Cerebellar granule cells are believed to be glutamatergic, but, as they receive excitatory amino acidergic input from mossy fibers, they also possess N-methyl-D-aspartate (NMDA) receptors. The possible involvement of these NMDA receptors in the biochemical differentiation of cultured granule neurons was studied in terms of the specific activity of phosphate-activated glutaminase, an enzyme important in the synthesis of the putative neurotransmitter pool of glutamate. When the partially depolarized cells were treated with NMDA for the last 3 days (i.e. between 2 and 5 days in vitro), it elevated specific activity of glutaminase in the dose- and time-dependent manners. The half-maximal effect was obtained at about 10 microM NMDA, whereas the maximum concentration, which produced about a 2.7-fold increase in 5-day-old cultures, was about 50 microM NMDA. This increase in glutaminase was completely blocked by the NMDA receptor antagonist, 2-amino-5-phosphonovaleric acid, and by the NMDA receptor-linked Ca2+ ion channel blockers, MK 801 and Mg2+. The effect of NMDA was not related to the survival of the granule cells, as the experiments were carried out before the dependence on high K+ for the survival of granule cells develops in culture, and during the period of investigation none of the compounds used compromised the survival of these cells. The enhancement of glutaminase activity was due to an induction in enzyme protein, since it was completely blocked by cycloheximide and actinomycin D. In contrast to granule neurons, the treatment with NMDA had no significant effect on the activity of glutaminase and glutamine synthetase in cultured cerebellar astroglial cells. Our present results on glutaminase enzyme would indicate that an increase in the cellular concentration of free Ca2+ mediated through the NMDA induced increase in Ca2+ conductance, leads to long term changes in differentiating cerebellar granule neurons, and it is possible that this kind of physiological stimulation of granule cells is normally provided in vivo by the presynaptic glutamatergic mossy fibers.

Effect of potassium depolarization on phosphate-activated glutaminase activity in primary cultures of cerebellar granule neurons and astroglial cells during development

The cerebellar granule cells are believed to be glutamatergic neurons. During the normal development of granule cells grown in a chemically defined medium, the specific activity of phosphate-activated glutaminase increased from 60 at 3 days to 150 (nmol/h/mg protein) at 15 days in vitro. Treatment with 25 mM K+ for the last 2 days elevated glutaminase activity in an age-dependent manner: about 100% at 3 and 6 days, 75% at 10 days, and 40% at 15 days in vitro. The enhancement of glutaminase in granule cells was dose-dependent. The half-maximal effect was obtained at about 20 mM K+, whereas the maximum concentration, which produced about a 2.5-fold increase in 3-day-old cultures was about 40 mM K+. The voltage-sensitive Na+ channel inhibitor tetrodotoxin had no effect on the depolarization-induced activity in granule cells. However, the increase in glutaminase by 25 mM K+ was significantly blocked by both organic (nifedipine) and inorganic (Ni2+ and Mg2+) calcium antagonists, indicating that elevation in activity may be mediated through transmembrane Ca2+ entry into granule cells. In contrast to neurons, in cultured cerebellar astrocytes, the activity of glutaminase slightly decreased during development, and treatment with 25 mM K+ had no significant effect on this enzyme activity. The present findings, together with previous observations, would indicate that depolarization with K+, which is believed to mimic in vivo presynaptic stimulation, could be one of the mechanisms that selectively controls the development and function of neurons, when measured in terms of the activity of the enzymes involved in the synthesis of cell-specific neurotransmitters.

Topographical distribution of neurochemical changes in Alzheimer's disease

Biochemical indices of cortical nerve cells affected in Alzheimer's disease have been proposed (excitatory dicarboxylic amino acid, EDAA, sodium-dependent carrier; phosphate-activated glutaminase activity; serotonin type 2 recognition site; somatostatin-like immunoreactivity). These and the content of EDAAs and two related amino acids, and choline acetyltransferase (ChAT) activity have been measured in up to 13 areas of cerebral cortex and the cerebellar cortex from 16 patients with Alzheimer's disease and 17 controls. Reduction of the index of the serotonin recognition site, somatostatin content and another biochemical index of interneurones coincide and indicate a rather unexpected focal loss of such neurones from the parietal lobe. No unequivocal measure of the integrity of pyramidal neurones could be established as the content of no amino acid was reduced, the index of the EDAA carrier showed evidence of change in few brain regions and glutaminase activity was subject to unexplained variability. ChAT activity alone closely paralleled a previous report of the distribution of morphological degeneration. The results are discussed in relation to therapy and positron emission tomography.

Evidence of glutamatergic denervation and possible abnormal metabolism in Alzheimer's disease

Excitatory dicarboxylic amino acids previously have been ascribed several functions in the brain. Here their total concentration and proposed neurochemical markers of neurotransmitter function have been measured in brain from patients with Alzheimer's disease (AD) and controls. Specimens were obtained antemortem (biopsy) approximately 3 years after emergence of symptoms and promptly (less than 3 h) postmortem some 10 years after onset. Early in the disease a slight elevation in aspartic acid concentration of cerebral cortex was observed in the patients with AD. A reduction in glutamic acid concentration of a similar magnitude was found. It is argued that this, together with a decrease in CSF glutamine content and lack of change in the phosphate-activated brain glutaminase activity of tissue, reflects an early metabolic abnormality. Later in the disease evidence of glutamatergic neurone loss is provided by the finding that in many regions of the cerebral cortex the Na+-dependent uptake of D-[3H]aspartic acid was almost always lowest in AD subjects compared with control when assessed by a method designed to minimise artifacts and epiphenomena. Release of endogenous neurotransmitters from human brain tissue postmortem did not appear to have the characteristics of that from human tissue antemortem and rat brain.

Role of thyroid hormone and nerve growth factor in the development of choline acetyltransferase and other cell-specific marker enzymes in the basal forebrain of the rat

The effects of treatment with L-thyroxine (subcutaneously 0.3 microgram/g body weight daily from birth, i.e., day 1) and 2.5S nerve growth factor (NGF; intraventricularly 2 micrograms on 1, 3, 5, 7, and 9 postnatal days), separately and together, were studied on the biochemical development of different cell types in the basal forebrain of 10-day-old rats. The development of cholinergic, gamma-aminobutyric acid-ergic (GABAergic), and glutamatergic neurons was monitored respectively in terms of choline acetyltransferase (ChAT), glutamate decarboxylase (GAD), and glutaminase activities, whereas glutamine synthetase (GS) and 2',3'-cyclic nucleotide-3'-phosphohydrolase (CNPase) activities were used to judge the maturation of astroglial and oligodendroglial cells. Treatment with either thyroid hormone or NGF from birth significantly increased the expression of ChAT activity in the basal forebrain of neonatal rats. When both agents were administered to the same animal, in agreement with our earlier in vitro findings, the stimulation in ChAT activity was much greater than the sum of the individual effects. In hypothyroid rats, significant effects of NGF at the low doses used were not detectable, although the increase of ChAT activity induced by thyroxine was potentiated by NGF in these animals. Under the present experimental conditions neither thyroxine nor NGF treatment had an appreciable effect on the activities of glutaminase, GS, and lactate dehydrogenase. However, the administration of thyroxine markedly increased CNPase activity in normal rats, whereas in hypothyroid rats the effect on both CNPase and GAD was also significant. Similar elevations in CNPase and GAD activities were not observed after NGF treatment, suggesting that the effect of NGF was specific to the cholinergic cells.(ABSTRACT TRUNCATED AT 250 WORDS)

Acute and long-term effects of chlorpromazine on glutamine synthetase and glutaminase in rat brain

The effect of administration of chlorpromazine on the activity of glutamine synthetase and glutaminase and the content of glutamate and gamma-aminobutyric acid (GABA) in different regions of rat brain was studied in an investigation of the possible role of these amino acids in the lowering of the seizure threshold following prolonged administration of chlorpromazine. Chlorpromazine was administered at a dose of 20 mg/kg of body weight s.c. For the acute study, the animals were killed 20 min after a single injection. For the long-term study, the animals were treated every day with the same dose for 21 days and were killed 20 min after the last injection. The results showed an increase in glutamate level in each brain region investigated following long-term administration, but only in the cerebral cortex after a single dose. GABA levels showed an increase in the brainstem only in acute experiments. Glutamine synthetase activity was increased in all three regions after a single dose and only in cerebral cortex after long-term administration. Glutaminase activity showed a decrease in cerebral cortex only after long-term administration of the drug. These results suggest the possible occurrence of a state of increased excitability in the brain as a result of long-term administration of chlorpromazine, thus contributing to the known complication of seizures.

Effect of thyroid deficiency on the regional development of glutaminase, a glutamatergic neuron marker, in the rat brain

The effect of thyroid deficiency on the activity of phosphate-activated glutaminase (the marker for glutamatergic neurons) was studied in different parts of the rat brain at ages 5, 10, 15 and 25 days, and at day 130 following 102 days of rehabilitation. The brain regions investigated were the cerebral cortex, basal forebrain, hippocampus and cerebellum. During normal development, the activity of glutaminase increased relatively earlier in the cerebral cortex and hippocampus than in the cerebellum, while the absolute value reached a much higher level in the hippocampus than in other brain regions. In the basal forebrain, the developmental pattern of glutaminase was bimodal, and the rise in enzyme activity after 15 days coincided with the decrease in the cerebral cortex. These regional developmental changes in glutaminase activity correlated well with known information on the formation of glutamatergic cells and pathways in the brain. Neonatal thyroid deficiency had little effect on the developmental patterns of enzyme activity, the exception being a transient decrease in 10-day-old hypothyroid hippocampus. The present results, together with previous findings, indicate that the effect of thyroid hormone on neural maturation is cell-type specific and the glutamatergic neurons are not the main targets of thyroid hormone action.

Calcium stimulation of glutamine hydrolysis in synaptosomes from rat brain

Calcium stimulates the hydrolysis of glutamine in synaptosomes prepared from rat brain both by the sucrose- (12) and the Ficoll/sucrose-gradient techniques (13). The calcium activation is phosphate-dependent and maximal effect is obtained at a calcium concentration of 0.5-1.0 mM. It is reduced by increasing the numbers of synaptosomes in the incubation mixture, and abolished by the product inhibitors of glutaminase, glutamate and ammonia, but unaffected by the uncoupler 2,4-dinitrophenol which inhibits the mitochondrial proton pump. Moreover, since the hydrolysis of glutamine is mediated by glutaminase (EC 3.5.1.2), and calcium does not activate the purified enzyme, an indirect phosphate-dependent effect of calcium on glutaminase is most likely. Calcium activates preferentially the N-ethylmaleimide insensitive fraction of glutaminase. The calcium activation is not dependent on synaptosomal membranes as it is found in synaptosomes subject to previous freezing. It is also found in isolated synaptosomal mitochondria and is thus a property of nerve endings. The calcium activation of glutaminase is unaffected by potassium in depolarizing concentrations, and may not be directly involved in the neurotransmission processes, but possibly in replenishing depleted stores of transmitter glutamate.

Regulation of glutaminase by exogenous glutamate, ammonia and 2-oxoglutarate in synaptosomal enriched preparation from rat brain

Phosphate activated glutaminase in synaptosomal enriched preparation from rat brain is very sensitive to inhibition by low concentration of glutamate, ammonia and 2-oxoglutarate when added to the incubation medium at pH 7.6. By increasing the concentration of either of these compounds up to 0.5 mM a pronounced initial inhibition is followed by little or no further effect when the concentration is increased beyond this level. By lowering the pH of the reaction mixture to 7.0, the inhibition by glutamate is almost abolished and that of ammonia reduced. Glutamate inhibits mainly the N-ethylmaleimide-sensitive fraction of glutaminase which previously is suggested to be localized to the outer phase of the mitochondrial inner membrane whereas ammonia inhibits both the N-ethylmaleimide-sensitive and -insensitive fraction. Evidence has been produced to show that the inhibition by 2-oxoglutarate is caused by glutamate formation by aminotransferase reactions. Since 2-oxoglutarate is produced by the tricarboxylic acid cycle, the operation of this cycle may regulate the glutaminase reaction by controlling glutamate formation via the aminotransferase reactions.

Properties of phosphate activated glutaminase in astrocytes cultured from mouse brain

Astrocytes in primary cultures contain a relatively high activity of phosphate activated glutaminase, although it is significantly lower than that of synaptosomal enriched preparations. The relatively high glutaminase activity in the astrocytes appears not to be caused by substrate induction, since a 10-fold variation in the glutamine concentration of the culture medium does not affect the activity. Of the reaction products, only glutamate inhibits astrocytic glutaminase whereas that of synaptosomal enriched preparations is inhibited by both glutamate and ammonia. Similar to the synaptosomal enzyme, glutaminase in astrocytes is inhibited about 50% by N-ethylmaleimide, indicating N-ethylmaleimide-sensitive and -insensitive compartments of the enzyme. Calcium activates glutaminase in astrocytes as in synaptosomes, by promoting phosphate activation. Except for the lower activity and the lack of effect of ammonia, the properties of the astroglial glutaminase has been found to be no different from that of the synaptosomal one. The relatively unrestrained astroglial glutaminase may, however, argue against the concept of a glutamine cycle operating in a stoichiometric manner.

The activities in different neural cell types of certain enzymes associated with the metabolic compartmentation glutamate

The cellular distribution of certain enzymes associated with the metabolic compartmentation of glutamate was estimated in ultrastructurally preserved and metabolically competent perikarya fractions that were enriched in astrocytes, granule cells and Purkinje cells and derived from 8-day-old rat cerebellum, and in monolayer cultures (14 days in vitro) composed principally of interneurones or astrocytes. The neuronal activities of glutamine synthetase and glutamate dehydrogenase were respectively about 4- to 8-fold and 2- to 5-fold lower than in astrocytes, depending upon the class of neurone and the type of preparation used for comparison. By contrast glutaminase activity was about 3- to 12-fold higher in neuronal than in astroglial preparations. Estimations of the specific activity of succinate dehydrogenase differed less between cell types, indicating that the differences in glutamate dehydrogenase and glutaminase were not simply related to variations in the concentration of mitochondria relative to the other cellular constituents. The findings presented provide direct evidence in support of our model assigning the 'small' glutamate compartment, where most of the labelled glutamine is synthesized, to glial cells, and the 'large' compartment to neurones, and also underline the metabolic interaction between these two cell types in the brain.

Evidence for compartmentalization of glutamate in rat brain synaptosomes using the glutamate sensitivity of phosphate-activated glutaminase as a functional test

The glutamate content of rat brain synaptosomes was measured by high performance liquid chromatography to be 39 micromol/g protein. If uncompartmentalized this glutamate (4 mM) would inhibit phosphate-activated glutaminase considerably. Since the action of any endogenous effector on the enzyme is assumed to be negligible following disruption of the synaptosomes, due to dilution with the incubation medium, the inhibition by glumate and activation by phosphate were compared in intact and disrupted synaptosomes. The inhibition by endogenous glutamate in intact synaptosomes was found to correspond to less than that of 0.5 mM of added glutamate to disrupted synaptosomes, indicating that the major fraction of synaptosomal glutamate is compartmentalized.

Evidence for compartmentation of synaptosomal phosphate-activated glutaminase

Phosphate-activated glutaminase (EC 3.5.1.2) in synaptosomal preparations is inhibited 40-60% by the sulphydryl group reagent N-ethylmaleimide (NEM), forming the basis for distinction between NEM-sensitive and NEM-insensitive glutaminases. The NEM effect cannot be explained by differential effects on distinct glutaminases because other glutaminases have not been detected, and the synaptosomal glutaminase activity can be fully accounted for by the activity of phosphate-activated glutaminase. By fractionation of mitochondria isolated from synaptosomal preparations, which are preincubated with and without NEM, both NEM-sensitive and NEM-insensitive glutaminases are found to be localized to the inner mitochondrial membrane. Variations in pH (7.0-7.6) and the phosphate concentration (5-10 mM) affect chiefly NEM-sensitive glutaminase, demonstrating that this glutaminase may be subject to regulation by compounds in the cytosol having restricted permeability to the inner mitochondrial membrane. Since p-hydroxymercuribenzoate, which is known to be impermeable to the inner mitochondrial membrane, inhibits glutaminase similarly to NEM, phosphate-activated glutaminase is assumed to be compartmentalized within the inner mitochondrial membrane. Thus, NEM-sensitive glutaminase is localized to the outer face and NEM-insensitive glutaminase to the inner region of this membrane and probably also to the matrix region.