Logo of jcbfmLink to Publisher's site
J Cereb Blood Flow Metab. Jan 2011; 31(1): 384–387.
Published online Nov 10, 2010. doi:  10.1038/jcbfm.2010.199
PMCID: PMC3049474

Brain glutamine synthesis requires neuronal aspartate: A commentary


Inspired by the paper, ‘Brain glutamine synthesis requires neuronal-born aspartate as amino donor for glial glutamate formation' by Pardo et al, a modified model of oxidation–reduction, transamination, and mitochondrial carrier reactions involved in aspartate-dependent astrocytic glutamine synthesis and oxidation is proposed. The alternative model retains the need for cytosolic aspartate for transamination of α-ketoglutarate, but the ‘missing' aspartate molecule is generated within astrocytes during subsequent glutamate oxidation. Oxaloacetate formed during glutamate formation is used during glutamate degradation, and all transmitochondrial reactions, oxidations–reductions, and cytosolic and mitochondrial transaminations are stoichiometrically balanced. The model is consistent with experimental observations made by Pardo et al.

Keywords: aralar, astrocytes, glutamate oxidation, malate–aspartate shuttle

The recent paper ‘Brain glutamine synthesis requires neuronal-born aspartate as amino donor for glial glutamate formation' by Pardo et al (2010) is of considerable interest because synthesis of glutamate/glutamine is a major role for astrocytes. In the brain, in vivo pyruvate carboxylation catalyzed by astrocyte-specific pyruvate carboxylase accounts for one-third of astrocytic glucose metabolism (Hyder et al, 2006; Hertz et al, 2007); accordingly, approximately two-thirds of astrocytic glucose oxidation is associated with the production of glutamate/glutamine. This is because synthesis of a new molecule of the tricarboxylic acid (TCA) cycle constituent α-ketoglutarate, glutamate's precursor, requires carboxylation of one molecule of pyruvate to the TCA intermediate oxaloacetate (OAA in Figure 1, pathway 1) together with oxidative decarboxylation (by the ubiquitous pyruvate dehydrogenase) of another molecule of pyruvate to acetyl coenzyme A. Subsequently, acetyl coenzyme A condenses with OAA to citrate, which is converted by α-ketoglutarate to glutamate that is amidated to glutamine and transferred to neurons, where it forms transmitter glutamate. Total astrocytic oxidative metabolism (pyruvate dehydrogenase and pyruvate carboxylase (PC) mediated) accounts for 30% of that in the brain cortex (Hertz et al, 2007).

Figure 1
Proposed alternative model for cytosolic-mitochondrial trafficking associated with astrocytic production of glutamine (pathway 1), its transfer to glutamatergic neurons (without indication of any extracellular space, because there is no other function ...

Some observations in the paper by Pardo et al (2010) should be emphasized. Although astrocytes accounted for only 7% of protein expression of aralar, the glutamate/aspartate exchanger (AGC1) in the brain necessary for the operation of the malate–aspartate shuttle, this amount is not negligible, because astrocytes constitute ~20% of the total volume in the brain cortex (reviewed by Hertz (2008)). When expressed relative to cell volume, the astrocytic expression of aralar protein is therefore about one-third of that in neurons. The malate–aspartate shuttle regenerates NAD+ from NADH (nicotinamide adenine dinucleotide) formed during glycolysis by transferring reducing equivalents from the cytosol to the mitochondria by coupled oxidation–reduction and transamination reactions involving interconversion of malate, OAA, aspartate, α-ketoglutarate, and glutamate, which traverse the mitochondrial membrane and deliver NADH to the electron transport chain; the malate–aspartate shuttle requires two transmitochondrial transporters, AGC (AGC1 (aralar) in the brain) and a malate/α-ketoglutarate exchanger (OGC). In addition, among the six compounds studied by Pardo et al (namely aspartate, alanine, leucine, GABA, glutamate, and citrate), only aspartate promoted the synthesis of intracellular glutamate and medium glutamine in cultured astrocytes. It was suggested that in the brain in vivo, this aspartate originated from neurons, based on the observation that the astrocyte-specific substrate acetate could sustain the formation of aspartate in aralar−/− mice, but that this aspartate was unavailable for its own synthesis of glutamate, which seemed to require an exogenous source. However, I have severe reservations with regard to a neuronal source of aspartate required in astrocytes and transferred through the extracellular space. Aspartate is, like glutamate, an excitatory transmitter, which makes it unsuited to act specifically as a transcellular amino group donor. Moreover, the net transfer of aspartate from neurons to astrocytes would drain the neuronal amino-acid pool, violating the generally accepted conclusion that only astrocytes are capable of net synthesis of TCA cycle constituents and their amino-acid derivatives. Finally, in Pardo's Figure 4, cytosolic aspartate is both transferred to astrocytes and transaminated to OAA in the neurons themselves, creating stoichiometric problems. Nevertheless, one attractive feature of this figure is that one of the two molecules of pyruvate required for the synthesis of releasable glutamate can be formed without the involvement of AGC1; the process only requires transfer of α-ketoglutarate across the mitochondrial membrane through OGC, which shows similar expressions in both neurons and astrocytes (Berkich et al, 2007).

Figure 1 illustrates an alternative model for aspartate-dependent glutamate/glutamine synthesis with four critical properties: (1) it satisfies the experimental observations made by Pardo et al without requiring extracellular trafficking of aspartate; (2) it explains why the provision of aspartate would stimulate astrocytic glutamate/glutamine synthesis; (3) it shows that astrocytes can form aspartate from acetate in aralar−/− animals, but are unable to transfer this aspartate to the astrocytic cytosol in the absence of aralar; and (4) it fulfills the important criterion that all transport processes across the mitochondrial membranes are stoichiometrically balanced. The figure shows that α-ketoglutarate, which was synthesized in the astrocytic TCA cycle—as in the Pardo model—is carried to the cytosol by OGC in exchange with malate, and transaminated to glutamate, and that a need is created for cytosolic aspartate (pathway 1). The generated glutamate is amidated with NH4+ to glutamine, which is transferred to neurons, where it is released as the transmitter glutamate (pathway 2). To simplify matters, the conversion of glutamine to transmitter glutamate in neurons by phosphate-activated glutaminase is indicated as entirely occurring in the cytosol, but it makes no difference for astrocytic reactions whether this is the case, or glutamate has to enter the mitochondria as described by Palaiologos et al (1989).

After termination of increased glutamatergic activity, excess glutamate is degraded, predominantly by oxidative metabolism of glutamate in astrocytes (pathway 3). Pronounced oxidative deamination of glutamate by glutamate dehydrogenase has been shown in cultured astrocytes (see Hertz et al (2007)), and astrocytes express a mitochondrial glutamate/hydroxyl carrier (Berkich et al, 2007). Nevertheless, the predominant, but possibly not the only, reaction in the brain in vivo seems to be the formation of mitochondrial α-ketoglutarate by transamination (Zaganas et al, 2009), coupled with the conversion of OAA to aspartate, and following glutamate/aspartate exchange by AGC1 activity as shown in pathway 3. The ‘excess' molecule of mitochondrial OAA created in pathway 1 now provides the OAA for the transamination as indicated by pathway 4, and the cytosolic aspartate generated in pathway 3 can provide the aspartate required in pathway 1. The formation of cytosolic aspartate would be absent in aralar−/− animals, and, in both wild-type and aralar-knockout cells, glutamate/glutamine synthesis can be predicted to be stimulated by addition of extracellular aspartate, because the aspartate may be normally delivered to pathway 1 with considerable delay, representing the time period between glutamate synthesis and glutamate oxidation, thus creating an apparent dependence on exogenous aspartate.

After formation in the astrocytic mitochondria, α-ketoglutarate is oxidized in the TCA cycle by conversion to malate that exits the TCA cycle and the mitochondria (pathway 3), followed by conversion to pyruvate in the astrocytic cytosol and reentry of pyruvate into the TCA cycle. If malate stays in the cycle, α-ketoglutarate will, after introduction of another molecule acetyl coenzyme A (and NADH formation) serve as the source for another molecule of glutamate (pathway 3, Figure 1). The relatively low activity of aralar (or that of another redox shuttle) may suffice for transfer of reducing equivalents for these reactions and for generation of the second pyruvate molecule required for glutamate synthesis (left part of Figure 1), because they do not need to occur very rapidly. However, as shown in Table 1, aralar activity in astrocytes corresponding to 7% of that in the brain is approximately twice of what is required for glutamate production and degradation and would be sufficient for the entire in vivo oxidation of glucose in astrocytes.

Table 1
Aralar activity compared with its requirement

Synthesis of glutamate and its subsequent oxidation do not necessarily occur in the same astrocyte, let alone the same mitochondrion. Gap junction-mediated trafficking of NAD(P)H and perhaps OAA and aspartate among astrocytes may minimize problems related to substrate supply and local redox balance, as might fluctuations in the production of lactate from pyruvate and transastrocytic lactate transport (Gandhi et al, 2009a, 2009b). However, mitochondrial OAA cannot be transferred in a similar manner, and some mitochondria may gain and others may lose TCA cycle intermediates during glutamate formation and degradation.

Although the transmitochondrial pathway for glutamine conversion to glutamate in neurons suggested by Palaiologos et al (1989) is not required in the model, this pathway might explain the dissociation between labeling of glutamate and glutamine from the astrocyte-specific substrate acetate in the intact brain of aralar−/− mice reported by Pardo et al. As shown in pathway 1 of Figure 1, the astrocytic synthesis of glutamate is not directly dependent on AGC1, although it will with time be impaired in aralar−/− mice because of the absence of aspartate return (pathway 4). However, if newly synthesized glutamate in the neurons initially enters the mitochondria, neuronal glutamine hydrolysis to glutamate and NH4+ would immediately become AGC1 dependent and compromised in aralar−/− mice. The released NH4+ might be required for glutamine synthesis in astrocytes, which accordingly would be unable to incorporate label from acetate into glutamine.


The alternative model retains the need for cytosolic aspartate, established by Pardo et al, but is stoichiometrically balanced and involves no transcellular aspartate transport. Overall, 7% aralar activity in astrocytes is sufficient for the model to operate.


The author declares no conflict of interest.


This study was supported by Grant no. 30711120572 from the National Natural Science Foundation of China.


  • Berkich DA, Ola MS, Cole J, Sweatt AJ, Hutson SM, LaNoue KF. Mitochondrial transport proteins of the brain. J Neurosci Res. 2007;85:3367–3377. [PubMed]
  • Beauvoit B, Chance B. Time-resolved spectroscopy of mitochondria, cells and tissue under normal and pathological conditions. Mol Cell Biochem. 1998;184:445–455. [PubMed]
  • Gandhi GK, Cruz NF, Ball KK, Theus SA, Dienel GA. Selective astrocytic gap junctional trafficking of molecules involved in the glycolytic pathway: impact on cellular brain imaging. J Neurochem. 2009a;110:857–869. [PMC free article] [PubMed]
  • Gandhi GK, Cruz NF, Ball KK, Dienel GA. Astrocytes are poised for lactate trafficking and release from activated brain and for supply of glucose to neurons. J Neurochem. 2009b;111:522–536. [PMC free article] [PubMed]
  • Hertz L. Bioenergetics of cerebral ischemia: a cellular perspective. Neuropharmacology. 2008;55:289–309. [PubMed]
  • Hertz L, Peng L, Dienel GA. Energy metabolism in astrocytes: high rate of oxidative metabolism and spatiotemporal dependence on glycolysis/glycogenolysis. J Cereb Blood Flow Metab. 2007;27:219–249. [PubMed]
  • Hyder F, Patel AB, Gjedde A, Rothman DL, Behar KL, Shulman RG. Neuronal-glial glucose oxidation and glutamatergic-GABAergic function. J Cereb Blood Flow Metab. 2006;26:865–877. [PubMed]
  • Palaiologos G, Hertz L, Schousboe A. Role of aspartate aminotransferase and mitochondrial dicarboxylate transport for release of endogenously and exogenously supplied neurotransmitter in glutamatergic neurons. Neurochem Res. 1989;14:359–366. [PubMed]
  • Pardo B, Rodrigues TB, Contreras L, Garzón M, Llorente-Folch I, Kobayashi K, Saheki T, Cerdan S, Satrústegui J. 2010. Brain glutamine synthesis requires neuronal-born aspartate as amino donor for glial glutamate formation J Cereb Blood Flow Metab(e-pub ahead of print; 25 August 2010) [PMC free article] [PubMed]
  • Zaganas I, Kanavouras K, Mastorodemos V, Latsoudis H, Spanaki C, Plaitakis A. The human GLUD2 glutamate dehydrogenase: localization and functional aspects. Neurochem Int. 2009;55:52–63. [PubMed]

Articles from Journal of Cerebral Blood Flow & Metabolism are provided here courtesy of Nature Publishing Group
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem Compound links
  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...