The ‘glial’ glutamate transporter, EAAT2 (Glt-1) accounts for high affinity glutamate uptake into adult rodent nerve endings


  • Sachin K. Suchak,

    1. Biochemical Neuropharmacology Group, Centre for Neuroscience Research, GKT School of Biomedical Sciences, King's College London
    Search for more papers by this author
  • Nicoletta V. Baloyianni,

    1. Biochemical Neuropharmacology Group, Centre for Neuroscience Research, GKT School of Biomedical Sciences, King's College London
    Search for more papers by this author
  • Michael S. Perkinton,

    1. Biochemical Neuropharmacology Group, Centre for Neuroscience Research, GKT School of Biomedical Sciences, King's College London
    Search for more papers by this author
    • The current address of Michael S. Perkinton is the Department of Neuroscience, Institute of Psychiatry, King's College London.

  • Robert J. Williams,

    1. Biochemical Neuropharmacology Group, Centre for Neuroscience Research, GKT School of Biomedical Sciences, King's College London
    Search for more papers by this author
  • Brian S. Meldrum,

    1. Biochemical Neuropharmacology Group, Centre for Neuroscience Research, GKT School of Biomedical Sciences, King's College London
    Search for more papers by this author
  • Marcus Rattray

    1. Biochemical Neuropharmacology Group, Centre for Neuroscience Research, GKT School of Biomedical Sciences, King's College London
    Search for more papers by this author

Address correspondence and reprint requests to Marcus Rattray, Biochemical Neuropharmacology Group, Centre for Neuroscience Research, GKT School of Biomedical Sciences, King's College London, Hodgkin Building, Guy's Hospital, London SE1 1UL, United Kingdom. E-mail:


The excitatory amino acid transporters (EAAT) removes neurotransmitters glutamate and aspartate from the synaptic cleft. Most CNS glutamate uptake is mediated by EAAT2 into glia, though nerve terminals show evidence for uptake, through an unknown transporter. Reverse-transcriptase PCR identified the expression of EAAT1, EAAT2, EAAT3 and EAAT4 mRNAs in primary cultures of mouse cortical or striatal neurones. We have used synaptosomes and glial plasmalemmal vesicles (GPV) from adult mouse and rat CNS to identify the nerve terminal transporter. Western blotting showed detectable levels of the transporters EAAT1 (GLAST) and EAAT2 (Glt-1) in both synaptosomes and GPVs. Uptake of [3H]D-aspartate or [3H]L-glutamate into these preparations revealed sodium-dependent uptake in GPV and synaptosomes which was inhibited by a range of EAAT blockers: dihydrokainate, serine-o-sulfate, l-trans-2,4-pyrrolidine dicarboxylate (PDC) (+/–)-threo-3-methylglutamate and (2S,4R )-4-methylglutamate. The IC50 values found for these compounds suggested functional expression of the ‘glial, transporter, EAAT2 in nerve terminals. Additionally blockade of the majority EAAT2 uptake sites with 100 µm dihydrokainate, failed to unmask any functional non-EAAT2 uptake sites. The data presented in this study indicate that EAAT2 is the predominant nerve terminal glutamate transporter in the adult rodent CNS.

Abbreviations used

excitatory amino acid transporter


excitatory amino acid transporters 1 (GLAST)


excitatory amino acid transporter 2 (Glt-1)


excitatory amino acid transporter 3 (EAAC1)


excitatory amino acid transporter 4


glial plasmalemmal Vesicles






l-trans-2,4-pyrrolidine dicarboxylate








central nervous system


amyotrophic lateral sclerosis


Dulbecco's modified Eagle medium


maloney-monkey leukemia virus reverse transcriptase


ethylenediaminetetra-acetic acid




glial fibrillary acidic protein


sodium dodecyl sulfate


enhanced chemiluminence

Glutamate is the principal excitatory amino acid neurotransmitter within the mammalian central nervous system. Glutamate levels in the synaptic cleft are tightly regulated because elevated glutamate levels result in excitotoxic neuronal death, which involves activation of ionotropic glutamate receptors and excessive calcium influx into neurones (Choi et al. 1987). Excitotoxicity has been implicated in numerous neurodegenerative diseases including amyotrophic lateral sclerosis (ALS) (Fray et al. 1998; Sasaki et al. 2000) and epilepsy (Meldrum 1994), as well as in acute CNS insults such as stroke and head trauma (Meldrum 1995). The major mechanism of control of synaptic glutamate levels, are the plasma membrane high affinity, sodium-dependent excitatory amino acid transporter (EAAT) family. These proteins are thought to be responsible for the uptake of glutamate into glial cells and neurones (Danbolt 2001), with the bulk of glutamate uptake occurring in glial cells.

To date, 5 EAAT members have been cloned from a variety of mammalian sources. The predominant CNS glutamate transporter is EAAT2 (GLT1), accounting for over 90% of the total CNS glutamate uptake (Haugeto et al. 1996; Tanaka et al. 1997). EAAT2 is often described as a glial cell specific transporter protein. EAAT1 (GLAST) is also expressed predominantly by glia and is the major glutamate transporter in some non-neuronal cells, including Bergmann glia in the cerebellum and Müller cells in the retina (Danbolt 2001). Other glutamate transporters are mainly expressed by subsets of neurones: EAAT3 (EAAC1) is expressed in the cell bodies and dendrites of some neurones (Furuta et al. 1997); EAAT4 is expressed in cerebellar Purkinje neurones (Itoh et al. 1997); EAAT5 is found in both glia and neurones in the retina (Rauen 2000).

Glial glutamate uptake is assumed to account for the bulk of the glutamate uptake activity in the CNS. However, there is also strong evidence for the existence of a glutamate transporter in glutamatergic nerve endings. Numerous studies have demonstrated that neurones accumulate the glutamate transporter substrate, [3H]D-aspartate (Streit 1980; Gundersen et al. 1993; Gundersen et al. 1996). Preparations of nerve terminals, synaptosomes, mediate the uptake of radiolabelled l-glutamate and d-aspartate (Bridges et al. 1999; Divac et al. 1977; Danbolt 2001). Moreover, EAAT2 mRNA has been detected in glutamatergic neurones of the hippocampus (Akbar et al. 1998; Torp et al. 1994; Schmitt et al. 1996; Berger and Hediger 1998). Furthermore, under ischaemic conditions, estimates suggest that reverse transport via nerve terminal glutamate transporters contributes significantly to the release of glutamate (Rossi et al. 2000).

Despite this evidence, the molecular nature of the presynaptic neuronal glutamate transporter(s) has not yet been identified. Immunocytochemistry suggests low, if any, significant expression of EAAT2 in nerve terminals or on axons (Chaudhry et al. 1995; Lehre and Danbolt 1998). Transporters other than EAAT2 may be expressed on nerve terminals. In culture conditions, neurones express a range of glutamate transporters including EAAT2, EAAT1 and EAAT3 (Wang et al. 1998). In cell culture, Wang et al. (1998) suggested that EAAT3 is a major contributor to glutamate uptake in cortical neurones. However, the situation in mature neurones in vivo is not clear: even though synaptosomes possess glutamate uptake with a pharmacology similar to EAAT2, it quite possible that the uptake measured is due to glial contamination of the preparations (Vandenberg 1998).

In order to establish which excitatory amino acid transporter(s) are expressed and function in adult rodent neurones, this study has employed pharmacological and biochemical approaches using synaptosomes. The synaptosome preparation method used also provides glial plasmalemmal vesicles (GPV), which are small vesicles pinched off from glial membranes (Nakamura et al. 1993; Hirst et al. 1998). This approach allows a greater purification of nerve endings than other techniques and the direct comparison of preparations enriched in nerve endings and glia in the same experiments. Uptake of transportable substrates ([3H]L-glutamate and [3H]D-aspartate) into both synaptosomes and GPV were used to determine neuronal and glial excitatory amino acid uptake. Different EAAT blockers and western blotting were used to estimate which of the EAAT subtypes were responsible for uptake in glial and neuronal compartments.

Materials and methods


Unless otherwise stated, reagents were obtained from Merck (Dorset, UK) or Sigma Aldrich (Dorset, UK).


Apart from the cell culture, below, experiments were carried out either on male and female, 6–8 weeks old, Sprague–Dawley or Lister Hooded rats. Alternatively studies were performed on adult female Swiss mice (NIH, Harland UK). Animals were killed by CO2 asphyxiation and the brain and/or spinal cord was removed rapidly.

Cell culture

Primary neuronal cultures from the mouse cerebellum and cortex were prepared as previously described (el Etr et al. 1989; Perkinton et al. 2002). Briefly, cerebral cortex and striatum were dissected from E15/E16 Swiss mouse embryos (NIH, Harlan UK), and mechanically dissociated. Six-well poly-L-ornithine treated plates were prepared and the cells were seeded (106/mL) in a serum-free culture media comprised of F-12 nutrient and DMEM (Invitrogen, Paisley, Scotland, UK), supplemented with antibiotics and hormones as described (Perkinton et al. 1999). The cells were then cultured at 37°C in humidified 95% air and 5% CO2. Using light microscopy the cells were determined to be mainly neuronal, with no glial elements, after approximately 7 days in vitro and were thus ready for use. Cultures were analysed by immunocytochemistry revealing that less than 2% of the cells were GFAP positive.

Reverse transcriptase PCR

Total RNA was obtained from both mouse primary neuronal cultures and adult rat brain regions, using the RNAzol B method (Biogenesis Ltd, Dorset, UK) following the manufacturers protocol, with the addition of a second ethanol precipitation to improve purity of the sample. cDNA was produced by reverse transcription of 4 μg of RNA in a total reaction volume of 80 μL, containing 2.5 nmol/mL oligo-dT(15) primer (Promega, Southampton, UK); 1 mm dNTPs (Amersham Pharmacia Biotech, Little Chalfont, UK), 1 U/mL recombinant ribonuclease Inhibitor (RNasin, Promega), 2.5 U/μL M-MLV reverse transcriptase, 5 mm MgC12, 10 mm Tris-HCl (pH 9.0 at 25°C), 50 mm KCl, 0.1% Triton X-100. Reverse transcription reactions were carried out at 42°C for 1 h followed by 10 min at 95°C. cDNAs were stored at −20°C until use.

Specific oligonucleotide primers were designed to amplify bases 653–1162 of rat EAAT1, GenBank accession number X63744 (rEAAT1M3: 5′-TCTTGGTTTCGCTGTCTGCCACG-3′; rEAAT1P6: 5′-TCCTCATTCATGCCGTCATCGTCC-3′); 1310–1636 of rat EAAT2 GenBank accession number X67857 (rEAAT2M9: 5′-ATGTCTTCGTGCATTCGGTGTTGGG-3′; rEAAT2P3: AGCCGTGGCACCATCTTCATA GC-3′); 1176–1550 of rat EAAT3 GenBank accession number U39555 (rEAAT3M4: CCCAAACGCATCACCCAGAAC G-3′; rEAAT3P7: 5′-TCGGCAACCCTT- CCAGTTACATTCC-3′), and 991–1665 of rat EAAT4 GenBank accession number U89608 (rEAAT4M8: 5′-GCTTGTGCCATGAGTGACTTATAG G-3′; rEAAT4P38: 5′-CGTGTCCTGAGGGATTTCTTC G-3′). While designed against the rat EAAT sequences, the primers were also able to amplify the mouse homologues of the rat sequences. Primers for the actin gene were used as a positive control and a method of checking for genomic DNA contamination, as previously described (Yip et al. 2001).

PCR was carried out in a reaction volume of 25 μL comprised of 2.5 μL of cDNA, 1.5 mm MgCl2 (Promega), 0.1 mm dNTP, Promega polymerase buffer (Promega), 0.1 mm of each primer, 0.625 U Taq DNA polymerase (Promega), 10 mm Tris-HCl (pH 9.0 at 25°C), 50 mm KCl, 0.1% Triton X-100. PCR amplification was performed for 27 cycles (30 s 95°C, 45 s 50°C, 30 s 72°C) followed by a 7-min extension at 72°C using a thermal cycler (PCR Cycler 2400, Applied Biosystems, Cheshire, UK). Amplified products were then subjected to electrophoresis in 1% agarose gels in TBE buffer, and visualised by ultraviolet illumination in the presence of ethidium bromide (0.1 mg/mL).

Amplified products were subcloned into the vector pCR2.1 (Invitrogen) following the manufacturers instructions. Plasmids were isolated using a kit (Qiagen Maxiprep, Qiagen Crawley, UK) and sequenced (Molecular Biology Unit, GKT School of Biomedical Sciences, King's College, London).

Glial plasmalemmal vesicle (GPV) and synaptosome preparation

Glial and nerve ending fractions were prepared as described by Nakamura et al. (1993) with minor modifications (Hirst et al. 1998). Whole brains without the cerebellum were weighed and homogenised in 0.32 m sucrose, 1 mm EDTA, pH 7.4, using a motor driven homogeniser. The preparation was centrifuged at 1000 g for 10 min (4°C) (Sorval Super T21). The resultant pellet was discarded, and the supernatant was gently transferred onto a four-step discontinuous Percoll gradient of 2%, 6%, 10% and 20% Percoll (Amersham Pharmacia Biotech), in SEDH solution (0.32 m sucrose, 1 mm EDTA, 0.25 mm dithiothreitol [DTT], 20 mm HEPES, pH 7.4). Tubes were centrifuged at 33 500 g for 5 min (4°C). The GPV fraction was collected from the interface of the 2% and 6% gradient, while the synaptosomes were collected from the interface of the 10% and 20% Percoll steps. Both fractions were washed twice by centrifugation, at 18 500 g for 5 min (4°C) and resuspended in SEDH. The GPV and synaptosome fractions were diluted to a final volume of 1.0 mL for GPVs and 1.5 mL for synaptosomes with ice-cold SEDH. The total preparation time was approximately 3½ hours.

Protein analyses

The protein content in GPVs and synaptosomes was determined according to the Bradford method (Bioquant reagent, Merck) using solutions of bovine serum albumin as standards (0.0–1.6 mg/mL). The absorbance readings were carried out using the Bio-kinetics reader (EL 312e-Bio-tek instruments).

Radiolabelled neurotransmitter uptake by GPV and synaptosomes

The GPV and synaptosome aliquots were diluted to a final concentration of 1 mg protein per microlitre with ice-cold SEDH solution and were used immediately for uptake studies. Uptake of [3H]L-glutamate or [3H]D-aspartate was measured in uptake buffer (140 mm NaCl, 5 mm KCl, 1 mm CaCl2, 1.2 mm NaHPO4, 5 mm Glucose and 20 mm HEPES, pH 7.4). In some experiments, sodium concentration was lowered by replacing sodium chloride with choline chloride. Uptake was started by adding [3H]L-glutamate (L-[3H]Glutamic Acid, 20–60 Ci/mmol, Amersham Pharmacia Biotech) or [3H]D-aspartate (D-[2,3-3H]Aspartic Acid, 14.0 Ci/mmol, Amersham Pharmacia Biotech) to the reaction tubes containing either GPVs or synaptosomes, achieving a final concentration of 50 nm. The mixture was incubated for 5 min and the reaction terminated by rapid filtration with ice cold PBS onto glass fibre filter (Whatman GF/C) coated in 0.1% polyethyleneimine and the filters washed using a Skatron Cell Harvester (LIER, Norway.) Filters were dried overnight and counted on a liquid scintillation counter (WinSpectral 414) using Emulsifier Safe scintillation fluid (Packard, Pangbourne, UK).

The following EAAT inhibitors were used: Dihydrokainate (DHK) (Tocris Cookson, Bristol, UK); l-trans-2,4-Pyrrolidine Dicarboxylate (PDC) (Tocris Cookson); serine-O-sulfate (SOS); 2S,4R-4-methylglutamate (4MG, also known as SYM2081) (Tocris Cookson) and (+/–)-threo-3-methylglutamate (T-3MG) (Tocris Cookson).

Western blot analyses

Synaptosome and GPV protein samples were equalised and then mixed 4 : 1 with boiling buffer (50 mm Tris, pH 7.5, 2% SDS, 5% b-mercaptoethanol, 10% glycerol, 0.005% bromophenol blue) and heated to 95°C for 5 min. Western analysis methods were used as described by (Perkinton et al. 1999) with minor modifications. Samples were separated by polyacrylamide gel electrophoresis using 8%, acrylamide: bis acrylamide (29 : 1) containing 0.1% SDS. Proteins were transferred onto ECL Nitrocellulose membranes (Hybond C pure, Amersham Pharmacia Biotech) by semidry electroblotting for 90 min at 1.5 V/cm2) in blotting buffer (25 mm Tris, 0.13 m glycine, 20% methanol, 0.1% SDS). Blots were incubated in 4% skimmed milk powder in TBS (20 mm Tris-HCl, 0.5 m NaCl, pH 7.5) for 30 min, washed twice for 5 min in TBS containing 0.05% Tween-20 (TTBS).

Blots were then incubated overnight in 1% skimmed milk in TTBS containing primary antiserum: rabbit anti rat EAAT2 (Anti B12, 1 : 5000) or rabbit anti rat EAAT1 (Rb AntiA522, 1 : 5000). Both antisera were a kind gift of Dr N. Danbolt (Oslo, Norway). For controls to estimate the degree of glial and neuronal contamination of fractions, we used rabbit-anti cow Glial Fibrillary Acidic Protein antiserum (Anti-GFAP, 1, 2000, DAKO Laboratories), mouse-anti Neurofilament antiserum (RT97, 1 : 1000, developed by Dr J Wood and obtained from the Developmental Studies Hybridoma Bank, Iowa, USA) and rabbit anti rat EAAT3 (EAAC11, 1 : 1000, Autogen Bioclear, Calne, Wiltshire, UK).

After incubation with primary antibodies, membranes were washed three times in TTBS. Secondary antibodies were then applied (peroxidase-conjugated goat antirabbit IgG, 1 : 5000, Jackson Laboratories or peroxidase-conjugated goat anti-mouse IgG, 1 : 5000, Vector Laboratories, Peterborough, UK) and incubated for 1 h in peroxidase-conjugated goat anti-rabbit IgG (Jackson Laboratories/Stratech, Scientific, Bedforshire, UK) diluted 1 : 5000 in 1% skimmed milk in TTBS. Membranes were washed twice times in TTBS followed by a final wash in TBS, and proteins detected using the ECL plus detection system (Amersham Pharmacia Biotech), according to the manufacturers instructions. Membranes were apposed to films, which were scanned and the band densities obtained using Bio Image Intelligent Quantifier (B. I. Systems Corporation, Ann Arbor, Michigan, USA).


Neuronal expression of four rodent EAAT isoforms

Primer pairs for each EAAT subtype were used for RT-PCR to amplify EAAT mRNAs from various adult rat brain regions: cerebellum; cortex; hippocampus; hypothalamus; midbrain; pons; striatum and thalamus (Fig. 1a). Each region analysed was shown to express each of the four rat EAAT subtypes analysed. As a control, cDNA from each sample were used as templates for the amplification of actin mRNA. All cDNA tested (Fig. 1, Actin+) showed the presence of actin, while the RNA samples that had not been reverse-transcribed (Fig. 1, Actin−) showed no amplification at all, thus confirming the absence of genomic DNA contamination. Cloning and sequencing the PCR products confirmed the specificity of each PCR reaction (results not shown).

Figure 1.

(a) Specific primer pairs were used to amplify EAAT1, EAAT2, EAAT3, EAAT4 and actin mRNAs from reverse-transcribed RNA samples extracted from three adult rat brains. The following regions were analysed: Cerebellum (Cb); Cortex (Cx); Hippocampus (Hp); Hypothalamus (Hy); Midbrain (Mb); Pons (Po); Striatum (St) and Thalamus (Th) (n = 3). Sizes of the amplified products are shown to the right of each picture. As control, RNA not subjected to reverse transcription was used to amplify actin (actin -). The absence of a band reveals no genomic DNA contamination of samples. Lane 1: Hyperladder 1 (Bioline, London, UK). (b) Primers were used to amplify EAAT1, EAAT2, EAAT3 and EAAT4 from cDNA prepared from primary neuronal cultures from striatum (St) and cerebral cortex (Cx) of E15-E16 mice. Three independent cultures for each region were used. Lane 1: Lambda HinDIII marker (Promega).

These data show that EAAT1, EAAT2 and EAAT3 mRNAs were found at moderate to high levels in all CNS areas tested. For EAAT4, there was a very low level of expression in all brain regions apart from cerebellum. The data are consistent with consistent with those published (Berger and Hediger 1998; Danbolt et al. 1998; Dehnes et al. 1998; Massie et al. 2001).

The same primer pairs were used to amplify RNA extracted from three separate cultures of neurones from mouse cerebral cortex and three separate cultures of mouse striatal neurones. The data shows that all four EAAT subtypes are expressed in both striatal and cortical primary neuronal cultures (Fig. 1b).

EAAT1 and 2 protein are present in synaptosomes

Western blotting was performed on synaptosomes and GPV prepared from adult mouse whole brain minus cerebellum (n = 4). These blots reveal that the synaptosomes contain EAAT1 and EAAT2 proteins. The levels of EAAT1 and EAAT2 proteins in synaptosomes were significantly lower than the levels in GPVs (∼20% and ∼10%, respectively, p < 0.05) (Fig. 2). Similar results were found in synaptosomes and GPV derived from rat brain minus cerebellum (not shown).

Figure 2.

Western blotting was carried out on GPV and synaptosome fractions from mouse brain. For each protein studied, the figure shows a representative blot, with the size of the protein (KDa) indicated. Band densities were quantified and are shown as a bar chart, bars show means and SEM (n = 4). Densities in GPV were compared to densities in synaptosomes using a paired t-test. *p < 0.05; **p < 0.01; ***p < 0.001.

The glial marker, glial fibrillary acid protein (GFAP) was present in the synaptosomes, at about 70% of the levels found in GPV (not significant). Conversely, using the RT97 antibody to a neuronal marker, phosphorylated neurofilament, there were low levels of neurofilament and EAAT3 immunoreactivity present in the GPV (less than 10% of levels in the synaptosome fraction, p < 0.05). Similar results were found in synaptosomes and GPV derived from rat brain minus cerebellum (not shown).

These results suggested that it was not possible to achieve a complete separation of the glial and neuronal fractions. Nevertheless, these results show both synaptosomes and GPVs are enriched for nerve endings and glial fragments, respectively.

Both synaptosomes and GPV exhibit sodium dependent uptake

[3H]L-Glutamate or [3H]D-aspartate (50 nm) were used as substrates to measure glutamate transport. A range of concentrations of tritiated substrates were used for uptake, and this showed that uptake was consistent with high affinity, saturatable transport by EAATs (results not shown). A markedly reduced concentration of sodium ions (1.2 mm compared to 141.2 mm in controls) reduced uptake to less than 10% (p < 0.05, n = 4). This was true for synaptosomes and GPVs prepared from either mouse or rat whole brain minus cerebellum.

The uptake activity in the glial fraction (GPVs) was, as expected, greater than that found in the synaptosomal fraction. GPVs showed almost twice the amount of uptake compared to the synaptosomal fraction (p < 0.05, n = 4).

Pharmacological characteristics of EAA transport in rodent GPV and synaptosomes

The effect of a range of EAAT blockers on uptake of [3H]D-aspartate into rat synaptosomes and GPV were measured (Figs 3a and b). For each blocker, the IC50 values were calculated after fitting data to a sigmoidal dose–response curve (GraphPad, Prism) (Table 1). The data reveals the following rank order of potency PDC > SOS = 4MG > DHK > T-3MG in rat synaptosomes and PDC = SOS = 4MG > DHK > T-3MG in rat GPV. Analysis revealed that all data fitted to the one-site binding model suggesting that a single transporter predominates in both the synaptosome and GPV fractions (GraphPad, Prism). No significant differences in drug potencies between synaptosomes and GPVs were observed (Fig. 3a and b, and Table 1).

Figure 3.

Graphs show uptake of 50 nm [3H]D-aspartate (a and b) or 50 nm [3H]L-glutamate (c and d) into rat brain synaptosomes (a), rat brain GPV (b), mouse brain synaptosomes (c) or mouse brain GPV (d). Various EAAT blockers were used as indicated. Each experiment was carried out a minimum of 4 times, data points show mean ± SEM.

Table 1.  Potency of EAAT uptake inhibitors acting on sodium dependent [3H]L-glutamate and [3H]D-aspartate uptake into rat and mouse synaptosomes and GPV
  1. ND = Not determined.

Mouse Synaptosomes1.58 ± 0.671.50 ± 0.3966.18 ± 6.89NDND
Mouse GPV0.87 ± 0.181.72 ± 1.1829.41 ± 7.73NDND
Rat Synaptosomes0.39 ± 0.3512.63 ± 3.3123.37 ± 3.9310.68 ± 1.72102.00 ± 1.35
Rat GPV6.70 ± 2.968.69 ± 2.2626.80 ± 6.699.26 ± 1.8381.21 ± 8.91

The effect of some EAAT blockers on [3H]L-glutamate uptake into synaptosomes and GPV prepared from mouse brain were also determined (Fig. 3c and d). Data was similar to that found in the rat (Table 1). The rank order of potency of EAAT blockers in both murine synaptosomes (Fig. 3c) and GPV (Fig. 3d) were as follows: PDC = SOS > DHK for synaptosomes (Table 1) and PDC = SOS > DHK in GPV.

EAAT-2 uptake blockers, DHK, T-3MG and 4MG, show IC50 values in both GPV and synaptosome preparations that are similar to those predicted from analysis of recombinant EAAT2 (Arriza et al. 1994; Vandenberg et al. 1997; Vandenberg 1998), this data would suggest that EAAT2 accounts for the main EAAT uptake activity in synaptosomes and GPV. However these data do not preclude the presence of other EAAT subtypes as each of the blockers will act upon all EAAT subtypes to a greater or lesser extent. T-3MG is also a blocker of EAAT4 (Eliasof et al. 2001), but as cerebellum was excluded from the tissues used to prepare synaptosomes and GPV, we do not expect a significant contribution of EAAT4 to the uptake measured here.

In order to determine whether transporters other than EAAT2 are involved in the uptake of [3H]D-aspartate by rat synaptosomes and GPV, 100 µm DHK was used to block the majority of EAAT2 uptake sites. This concentration was chosen to selectively block EAAT2 without affecting the activity of the other main EAAT subtypes in the synaptosomes and GPV, namely EAAT1 and EAAT3 (Aprico et al. 2001).

It was possible to estimate the amount of EAAT2 [3H]D-aspartate uptake sites that would be blocked (percentage inhibition) by this concentration of DHK using the following equation:


The concentration of [3H]D-aspartate [S] was 0.05 µm and the concentration of DHK, [I] was 100 µm, respectively. Published values for the Ki of DHK vary from 20 to 110 mm (Griffiths et al. 1989; Pines et al. 1992; Arriza et al. 1994; Bridges et al. 1994; Willis et al. 1996). The Km of d-aspartate varies from 2 µm to 13 µm (Arriza et al. 1994; Dowd et al. 1996; Shimamoto et al. 1998; Koch et al. 1999) depending on the expression system used to measure uptake. The mean theoretical percentage inhibition of EAAT2 [3H]D-aspartate uptake sites was calculated to be approximately 70%, range 47%−92%.

If transporters other than EAAT2 were active in the synaptosomal fraction, a 70% reduction in available EAAT2 [3H]D-aspartate uptake sites, caused by 100 µm DHK would be expected to unmask non-EAAT2 transporters and cause a large apparent decrease in the IC50 values for SOS. This is because SOS has a 6–10 fold higher affinity for EAAT1 over EAAT2 (Arriza et al. 1994; Vandenberg 1998) and a 7 fold higher affinity for EAAT3 over EAAT2 (Arriza et al. 1994).

100 µm DHK reduces total uptake in GPV and synaptosomes by 66% and 73%, respectively (Fig. 4a and b). In the presence of 100 mm DHK a slight, but not significant change in the IC50 for SOS can be observed: from 8.7 µm when used alone, to 19 µm when used in conjunction with 100 µm DHK (Fig. 4a). A similar situation is observed in GPV, where the IC50 is altered from 12.6 mm in the absence of 100 µm DHK to 20.7 in its presence (Fig. 4b, Table 2). Pre-treatment with 100 µm DHK caused a similar changes in the IC50 for PDC: 1.1 µm for PDC in synaptosomes, reducing to 6.0 µm in the presence of DHK; and 0.7 µm for PDC in GPV changing to 10.6 µm in combination with DHK. One-way anova tests revealed that these reductions were not significant. These data strongly suggest that transporters other than EAAT2 do not contribute significantly to [3H]D-aspartate uptake in rat synaptosomes.

Figure 4.

Graphs show uptake of 50 nm [3H]D-aspartate into rat brain synaptosomes (a) or GPV (b). The ability of the inhibitors PDC and SOS to block uptake were tested in the presence or absence of 100 mm DHK. IC50 values calculated after fitting data to a sigmoidal dose–response curve (GraphPad, Prism) are shown in Table 2. Each experiment was carried out 3 times, data points show mean ± SEM.

Table 2.  Potency of SOS and PDC acting of [3H]D-aspartate uptake into synaptosomes and GPV in the presence and absence of 100 mm DHK
+ DHK20.7 ± 7.2919.0 ± 2.11+ DHK20.7 ± 7.29
− DHK12.6 ± 3.278.7 ± 2.56− DHK12.6 ± 3.27


This study has shown that neurones can express EAAT1, EAAT2, EAAT3 and EAAT4, as shown by RT-PCR of primary cultures of mouse neurones. We have confirmed the expression of EAAT2 and EAAT1 in cultured neurones by immunocytochemistry (results not shown). We have demonstrated that nerve terminals derived from adult mouse or rat brains express protein for the two major glutamate transporters, EAAT2 and EAAT1. Furthermore, functional studies show that synaptosomes are capable of supporting glutamate and d-aspartate uptake in a sodium dependent manner with a pharmacological profile that most closely resembles that of EAAT2. Uptake into synaptosomes shows similar pharmacological properties to uptake mediated by glial plasmalemmal vesicles obtained from the same preparations. Blockade of the majority of EAAT2 uptake sites with DHK fails to reveal any increase in the sensitivity of the remaining uptake to the drug SOS, which is more effective at the main non-EAAT2 transporters (specifically EAAT1 and EAAT3). Therefore these data suggest that EAAT2 is the principal uptake site in nerve terminals.

EAAT2 mediates glutamate uptake into adult nerve endings

The predominant CNS glutamate transporter is EAAT2, accounting for over 90% of the total CNS glutamate uptake (Haugeto et al. 1996; Tanaka et al. 1997). EAAT2 is often referred to as the glial glutamate transporter, implying an exclusive expression with the glial population. Whilst it is clear that nerve terminals take up glutamate (and d-aspartate) via sodium dependent high affinity uptake mechanisms (Fonnum 1984), the exact subtype or subtypes responsible for this uptake are not yet known. This current study has provided evidence that EAAT2 is the principal glutamate uptake site on nerve terminals. This conclusion is based on the presence of EAAT2 protein in nerve terminals (the synaptosome preparation) and the pharmacological characterisation of [3H]D-aspartate or [3H]L-glutamate into synaptosomal and GPV preparations.

Our study has shown evidence for functional EAAT2 within the nerve terminals, yet to date there is no clear anatomical evidence to confirm this finding. Chaudhry et al. (1995) used immunocytochemistry and electron microscopy to examine EAAT2 expression. They found that the vast majority of the EAAT2 protein was detected in the glial cells and that around 10% of the EAAT2 was found on neuronal processes. Despite this the authors could not conclude that there was neuronal EAAT2 expression. The Western blotting data (Fig. 1) agrees with the Chaudhry study by showing that synaptosomal fraction contained about 10% of the EAAT2 levels compared to the GPV. Thus the present data are compatible with the anatomical studies.

A potential problem with the current approach is that synaptosomes are not purely nerve endings and do contain glial elements that may contribute to the uptake observed in the preparation. Contamination of synaptosomes by glial elements is a recognized problem with this technique (Daniels and Vickroy 1998), and data shown here. This study has utilized a method (Hirst et al. 1998) that is designed to partially purify the glial elements: GPV. The major advantage of this is in the present study that it allows synaptosomes and GPV to be directly compared in the same experiments. Thus even if pure populations of glial elements and nerve terminals are not obtained, any major differences in the glutamate transporter profiles will result in observable differences in the pharmacology of the two preparations. We have attempted ways of further purification of synaptosomes, including GFAP immunoaffinity purification and treatment with the glial toxin fluorocitrate. However, both of these approaches were unsuccessful as they caused non-specific deterioration of the synaptosomal fractions. Direct methods of visualising [3H]D-aspartate in synaptosomes would help to resolve the relative levels of neuronal uptake compared to glial uptake.

We have demonstrated that uptake of [3H]D-aspartate and [3H]L-glutamate in to both synaptosomes and GPV can be inhibited by dihydrokainate (DHK) (+/–)-threo-3-methylglutamate (T-3MG) and 2S, 4R-4-methylglutamate (4MG, also referred to as SYM2081). Each of these drugs is known to be highly potent for the EAAT2 transporter, and with the exception of T-3MG (Eliasof et al. 2001), little or no activity at other EAAT subtypes (for a full review of the pharmacology see Bridges et al. 1999). These data seem to indicate that EAAT2 predominates within the synaptosomes and GPV.

We found that the rank order of potency for the transporter blockers in rat synaptosomes is PDC ≥ SOS = 4MG > DHK > T3MG while it is PDC > SOS > DHK in mice synaptosomes. There is an apparent discrepancy with the published potency (Ki values) for the same compounds acting on recombinant EAAT2 expressed by Xenopus oocytes, which were: 4MG ≥ PDC > DHK > T3MG > SOS, with Ki values of 3 mm (Robinson et al. 1993), 2–7 mm (Arriza et al. 1994; Vandenberg et al. 1997), 9.2 mm (Arriza et al. 1994), 18 mm (Vandenberg et al. 1997) and 240 mm (Vandenberg 1998), respectively. Xenopus oocytes may not provide an ideal comparison to the present data, as the transporters expressed in this system may differ in their sensitivity to EAAT blockers in comparison to native transporters in rodent tissues. As such a better comparator may be studies in which synaptosomes were used to determine the IC50 of the various EAAT blockers. From these studies, the following IC50 values were obtained: SOS (2–55 mm) > T3MG(99 mm) ≥ DHK (80–170 mm) (Garlin et al. 1995; Eliasof et al. 2001, respectively). The IC50 values found in this study are similar to, but do not exactly match, those expected from those published data in synaptosomes. We found IC50 values for DHK to be between 23 and 66 mm in comparison to the higher values typically found by other researchers. For example (Robinson et al. 1991) found that in synaptosomes the IC50 value for DHK varied between 80 and 170 mm suggesting that there is some variability in the results obtained with this compound.

Blockade of EAAT2 with DHK fails to reveal other EAATs are active in synaptosomes

As yet no subtype selective inhibitors have been developed and even compounds such as DHK and T-3MG, which are potent blockers of EAAT2 mediated uptake, have affinity for other members of the EAAT family. This has the effect of complicating the pharmacological profile of the uptake in both synaptosomes and GPV. In an attempt to side step this issue this study has utilised DHK and SOS together. We reasoned that even if low levels of functional non-EAAT2 transporters existed (i.e. EAAT1 and EAAT3) within the two preparations they would be revealed as an increase in sensitivity to the drug SOS when the majority of the EAAT2 uptake sites are blocked with 100 mm DHK. Given that SOS has a higher affinity for EAAT1 and EAAT3 over EAAT2 by 6 and 10 times, respectively, we would expect to see a large reduction in the IC50 of SOS when comparing its effects in the presence and absence of 100 mm DHK. As expected the application of 100 mm DHK resulted in a reduction in total uptake of [3H]D-aspartate. However, no significant reductions in the IC50 values of SOS (with and without DHK) were observed in synaptosomes or GPV, which would strongly suggest that EAAT2 transporter are the only active transporters in synaptosomes. We note that there are relatively high levels of EAAT3 and some EAAT1 and protein detected in the synaptosomal fraction, but conclude that these transporters do not contribute significantly to synaptosomal [3H]D-aspartate under the conditions used here.

Using RT-PCR we have shown that mRNA for EAAT1, EAAT2, EAAT3 and EAAT4 are expressed in neurones, cultured from embryonic mouse striatum and cerebral cortex. This data confirms and extends other studies, which have shown that EAAT 1 and 2 protein and mRNA are expressed in cultured neurones (Brooks-Kayal et al. 1998; Mennerick et al. 1998; Wang et al. 1998). The present data raises the possibility that adult neurones may express a wider variety of EAAT subtypes than previously thought. However when considering these data two important points must be considered. Firstly presence of mRNA does not indicate, nor does it guarantee the presence of the corresponding protein; secondly neuronal cultures possess several embryonic phenotypic features and as such are not exact models of adult neurones. For example, EAAT2 protein and mRNA are expressed by embryonic neurones in vivo (Furuta et al. 1997), and as such these data obtained from primary neuronal cultures may be a developmental phenomenon.

Evidence for localisation of EAAT2 on nerve terminals

Our findings point to EAAT2 being responsible for the uptake observed in nerve terminals. Yet this raises the question that if the pharmacological profile of synaptosomal uptake resembles that of EAAT2 so closely and synaptosomal uptake is represents such a large percentage of glial uptake, why has no EAAT2 protein been detected in adult nerve terminals to date?

In adult brain, some neurones have been shown to express mRNA for EAAT2 (Torp et al. 1994; Schmitt et al. 1996; A kbar et al. 1998; Berger and Hediger 1998), however, as mentioned above, there are no studies which have convincingly shown the presence of this protein. The resolution of current techniques may not be sufficient to detect very low levels of EAAT2 protein in neurones. As stated previously the Chaudhury study (1995) used high-resolution electron microscopy to examine EAAT2 distribution. It was determined that the EAAT2 that they observed on neurones may have been due to chance, however, they could not rule out the possibility that low levels were indeed present.

Additionally it may be possible that neurones are expressing a variant of EAAT2. Recent work by two groups has demonstrated that both mRNA and protein for this variant form are expressed within neurones in the CNS (Chen et al. 2002; Schmitt et al. 2002). These two studies have demonstrated that EAAT2 variants are abundantly expressed in neuronal populations. It is therefore possible that the uptake activity in synaptosomes here is due to these isoforms. At the present time, it is not possible to determine whether the uptake measured here is due to EAAT2 or the variant of EAAT2. Both transporters have high affinity for DHK, and seem to have very similar pharmacological properties. In addition, the antiserum used here would be predicted to recognise both isoforms on western blots.


SKS was a recipient of a Medical Research Council PhD studentship.