Amyotrophic Lateral Sclerosis-linked Glutamate Transporter Mutant Has Impaired Glutamate Clearance Capacity*

We have investigated the functional impact of a naturally occurring mutation of the human glutamate transporter GLT1 (EAAT2), which had been detected in a patient with sporadic amyotrophic lateral sclerosis. The mutation involves a substitution of the putative N-linked glycosylation site asparagine 206 by a serine residue (N206S) and results in reduced glycosylation of the transporter and decreased uptake activity. Electrophysiological analysis of N206S revealed a pronounced reduction in transport rate compared with wild-type, but there was no alteration in the apparent affinities for glutamate and sodium. In addition, no change in the sensitivity for the specific transport inhibitor dihydrokainate was observed. However, the decreased rate of transport was associated with a reduction of the N206S transporter in the plasma membrane. Under ionic conditions, which favor the reverse operation mode of the transporter, N206S exhibited an increased reverse transport capacity. Furthermore, if coexpressed in the same cell, N206S manifested a dominant negative effect on the wild-type GLT1 activity, whereas it did not affect wild-type EAAC1. These findings provide evidence for a role of the N-linked glycosylation in both cellular trafficking and transport function. The resulting alteration in glutamate clearance capacity likely contributes to excitotoxicity that participates in motor neuron degeneration in amyotrophic lateral sclerosis.

Dysfunction in glutamate homeostasis is implicated in a number of acute and chronic neurodegenerative diseases (26). An abnormality in the glutamate transport system has been reported in patients with amyotrophic lateral sclerosis (ALS). ALS is a progressive neurological disorder characterized by degeneration of upper and lower motor neurons. Although the primary pathogenic trigger is unknown, evidence is mounting to implicate a role for glutamate-mediated excitotoxicity in the disorder. A decrease in glutamate transporter activity was observed in synaptic preparations from motor and sensory cortex of sporadic ALS patients (27), which was subsequently ascribed to reduced levels of the glutamate transporter isoform GLT1, as assessed by immunodetection techniques (28). An aberrant RNA editing process was reported as a probable cause for loss of GLT1 in the sporadic form of ALS (29). However, the involvement of aberrant splicing processes for the GLT1 mRNA in ALS is still controversial (30 -35).
Our recent investigations indicated that oxidative alterations of human GLT1 caused by acquired reactivity of familial ALS-linked superoxide dismutase 1 mutants lead to impaired glutamate uptake (36). Given the accumulating evidence for an important role of GLT1 in both sporadic and familial ALS, the question arises of whether mutations in GLT1 could be a cause or a risk factor in ALS pathogenesis.
Using single-strand conformation polymorphism analysis of genomic DNA, we recently reported a mutation in the GLT1 gene associated with sporadic ALS, which substitutes an asparagine for a serine at position 206 (37). We speculated that this mutation might be of functional significance, because as-paragine 206 is a putative glycosylation site of GLT1 (12, 38 -40). The experiments illustrated in this report demonstrate the impact of this mutation (N206S) on the functional properties of the human glutamate transporter GLT1.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis-The cDNA encoding the human glutamate transporter GLT1 served as a template for site-directed mutagenesis performed by the unique site elimination procedure using the Transformer site-directed mutagenesis kit (CLONTECH, Palo Alto, CA). The procedure used two mutagenic primers: one primer (Mt-XhoI, 5Ј-CGGGCCCCCCCTGGAGGTCGACGGTA) served as a selection primer in the mutagenesis reactions to eliminate the unique XhoI site in the Bluescript II vector; a second mutated oligonucleotide (N206S, 5Ј-GACGAGGAGGCCAGCGCAACCAGCGCTGTTG) was designed to replace the codon for Asn 206 by Ser. The wild-type and mutated human GLT1 cDNAs were subcloned either into the pTNLII vector, for expression in Xenopus laevis oocytes, or into pcDNA3 (Invitrogen, Carlsbad, CA), for expression in COS7 cells. The presence of the mutated nucleotide sequence in the human GLT1 cDNA was verified by the dideoxy termination method, using the Sequenase kit (version 2.0; US Biochemicals, Cleveland, OH).
cRNA Synthesis, Oocyte Injection, COS7 Transient Transfection, and Electrophysiological Studies-For oocyte injection, the linearized pTN-LII constructs were transcribed in vitro using the mMESSAGE mMA-CHINE transcription kit (Ambion, Austin, TX). Xenopus oocytes were enzymatically defolliculated with collagenase (2 mg/ml, 2 h at 18°C), and stage V-VI oocytes were injected with capped RNA (25 ng/50 nl or as indicated). Oocytes were used 2-4 days after injection. Two-electrode voltage clamp was used to measure the glutamate transporter-mediated currents at room temperature using the Clampator-1B device (Dagan, Minneapolis, MN) interfaced to a personal computer by a Digidata 1200 analog-to-digital controller. Data were acquired with the pCLAMP 8.0 software (Axon Instruments, Foster City, CA). Microelectrodes were filled with 3 M KCl and had a tip resistance of 0.5-2 M⍀. Voltage step commands were applied in 10-mV increments to generate current-voltage curves; the current output was low pass filtered at 1 kHz; and the signal was averaged three times before and after solution exchange. Recordings were made from oocytes bathed with a solution that contained 96 mM NaCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , 2 mM KCl, and 5 mM HEPES-Na, pH 7.4 (ND96). For reverse transport experiments, the solution contained 80 mM NaCl, 20 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , 4 mM BaCl 2 , and 5 mM HEPES-Na, pH 7.4 (ND80). Transportermediated efflux was determined by including the GLT1-nontransported inhibitor dihydrokainate (DHK, 1 mM) in the efflux buffer, and the carrier-mediated efflux current was estimated by subtracting the current evoked in ND80 to that in the presence of ND80 and DHK. For experiments in Fig. 6, oocytes were injected with a solution containing monosodium glutamate (MSG, 1 M, 23 nl). For expression studies, COS7 cells were transiently transfected with the pcDNA3 constructs using the GenePorter transfection kit (Gene Therapy Systems, San Diego, CA) following the manufacturer's recommendations. The vector alone was used for mock transfection.
Uptake Measurement-Uptake was measured 2-4 days after injection of cRNA in oocytes or 36 -48 h after transfection in COS7 cells cultured in Petri dishes (60 mm diameter) or six-well plates. ND96 buffer, containing 20 M D-aspartate ([ 3 H]D-aspartate, 10.5 Ci/mmol; isotopic dilution, 1:20,000), was used to measure the uptake in oocytes for 30 min. Transporter-mediated efflux of radiolabeled substrate was determined in oocytes previously injected with 23 nl of D-aspartate solution ([ 3 H]D-aspartate and 50 mM D-aspartate in 50 mM potassium phosphate buffer, pH 7.4; isotopic dilution, 1:50) and by exposing them for different times to the ND80 solution. After injection of the radioactive solution, oocytes were kept in ND96 containing 50 M DHK for 1 h. The efflux was started after rapid washing of oocytes in ND80 buffer. Incubation was carried out in 500 l of ND80 at room temperature.
Oocytes not expressing the transporters were used to assess the nonspecific leak of [ 3 H]D-aspartate. This leak, which on average represents ϳ60% of the total radioactivity released in the buffer at a given time, was subtracted out from the efflux data. For assessment of the oocyte radioactive content, oocytes were separately dissolved in 10% SDS, mixed with scintillation fluid, and counted for radioactivity. The efflux rate constants were obtained by fitting the data to the equation ln(1 Ϫ N t /N 0 ) ϭ Ϫk t , where t is the time, k is the first-order rate constant, N t is the radiolabeled substrate released in the medium, and N 0 is the radiolabeled substrate content of the oocyte at the start of the experiment. Uptake in transfected COS7 cells was measured for 30 min in the presence of 40 M D-aspartate (isotopic dilution, 1:2,500) and in buffer containing 140 mM NaCl, 2.5 mM KCl, 1.2 mM MgCl 2 , 1.2 mM CaCl 2 , 1.2 mM K 2 HPO 4 , 10 mM glucose, and 10 mM HEPES-Na, pH 7.4. Uptake was stopped by washing the cells three times with ice-cold choline buffer (no sodium). The uptake values were corrected by the protein content in each Petri dish. For the experiments with tunicamycin, this compound was added 24 h before measuring the uptake, at a concentration of 5 g/ml in 0.1% Me 2 SO. This concentration and application time of tunicamycin are compromises between avoiding cellular death and detachment from the dish and observing a deglycosylation pattern for GLT1.
Western Blot Analysis-Oocytes were solubilized in ND96 and 1% Triton X-100 in presence of a mixture of protease inhibitors (Complete; Roche Molecular Biochemicals), whereas transfected COS7 cells were lyzed with phosphate-buffered saline (PBS), 1% SDS, and protease inhibitors. Proteins were resolved on SDS-polyacrylamide gel electrophoresis (10%) under reducing condition, blotted onto a polyvinylidene difluoride membrane (Amersham Pharmacia Biotech), and probed with a polyclonal antibody (Rb68518) raised against peptide 12-26 of the sequence of rat GLT1 (36). Protein bands were visualized by chemiluminescence using a horseradish peroxidase-conjugated secondary antibody.
Assessment of Cell Surface Expression by Biotinylation-The biotinylation reaction was performed following the protocol reported by Daniels and Amara (41). Briefly, COS7 cells were grown near confluence in six-well plates and transfected, and 36 h after transfection they were incubated with biotinylation buffer containing 1 mg/ml NHS-SSbiotin (Pierce). The cell extracts were assessed for protein content and eventually adjusted to the same value. An aliquot of the cell extracts was run on a gel to confirm equal protein concentration. Cell extracts were then incubated with UltraLink immobilized NeutrAvidin (Pierce) for immunoprecipitation of biotinylated proteins. The immunoprecipitates were separated on SDS-polyacrylamide gel electrophoresis and probed with anti-GLT1 antibody, and the biotinylated transporters were visualized by chemiluminescence.
Immunofluorescent Analysis-COS7 cells were plated on 35-mm Petri dishes and transfected with 3 g of plasmid/dish containing wildtype human GLT1, mutant N206S, or plasmid alone. Cells were fixed in 4% paraformaldehyde for 15 min, rinsed three times, and incubated 1 h with blocking solution (PBS containing 5% normal goat serum, 10% bovine serum albumin, and 0.2% Triton X-100). The anti-GLT1 antibody was incubated overnight (1:500 dilution) in PBS, 1% bovine serum albumin, and 0.1% Triton X-100 at 4°C. The secondary antibody (Cy3conjugated goat anti-rabbit IgG; Molecular Probes, Eugene, OR) was incubated for 1 h (1:1,000 dilution) in PBS, 1% bovine serum albumin, and 0.1% Triton X-100. Cells were washed twice with PBS and examined with a confocal microscope (Bio-Rad). Images obtained with Bio-Rad software were then formatted in Adobe Photoshop. Oocytes were embedded in Tissue-Tek and frozen, and 8-m sections were cut on a cryostat and mounted onto Superfrost Plus microscope slides (Fisher). Sections were dried for 30 min at room temperature and then fixed in PBS and 4% paraformaldehyde for 10 min. After rinses in PBS, sections were blocked in 2% normal donkey serum and 0.1% Triton X-100 in PBS for 1 h and incubated overnight at 4°C in the same solution to which the anti-GLT1 antiserum (1:500) was added. Next, the sections were rinsed, incubated for 1 h in Cy3-conjugated anti-rabbit serum (0.5%; Jackson ImmunoResearch) in PBS containing 1% normal donkey serum, rinsed again, mounted in Vectashield (Vector Laboratories), and analyzed for immunofluorescence. Stained oocytes were examined with a Leica DM microscope connected to a video camera and video-grabbing software.
Statistical Analysis-Data are expressed as mean Ϯ S.E. of at least three independent experiments.

N206S-human GLT1 Is a Less-glycosylated Transporter-
The amino acid asparagine at position 206 resides in the third large extracellular loop of the GLT1 carrier and is predicted to be glycosylated in vivo (12). The N206S mutation of GLT1 found in sporadic ALS was expected to suppress N-glycosylation (37). To verify the impact of this mutation on the glycosylation of the transporter, we transiently transfected COS7 cells with cDNA encoding either wild-type or N206S-human GLT1. Western blot analysis on homogenates prepared from transfected cells showed that the apparent molecular weight of N206S is reduced compared with that of wild type (Fig. 1, inset), suggesting a decrease in the extent of glycosylation. Our antibody, which was raised against the N-terminal region of the rat homologous to human GLT1 (amino acids 12-26; Ref. 36) recognized a band of ϳ70 kDa for wild type and 60 -65 kDa for N206S. The human GLT1 transporter has an additional putative glycosylation site at position 216 ( Fig. 2A). If treated with the N-glycosidase enzyme PNGase F (Fig. 2B), wild type and N206S originated a band of the same apparent molecular weight and with lower electrophoretic mobility, suggesting that the mutant GLT1 is indeed a partially glycosylated transporter and that the wild-type GLT1 has both the N206 and N216 sites glycosylated in vivo. A similar shift in electrophoretic mobility was observed by Danbolt and collaborators (40) when the purified rat GLT1 protein was treated with N-glycosydase.
Transport Properties of the Glycosylation Mutant-The impact of N206S on the transport activity was first assessed by measuring [ 3 H]D-aspartate uptake in transiently transfected COS7 cells. The expression of the wild type resulted in a 300fold increase in uptake over mock-transfected cells. Expression of N206S resulted in an ϳ50% reduction in uptake compared with the wild type ( Fig. 1). Treatment of COS7 expressing the transporters with tunicamycin, a drug that interferes with the glycosylation pathways of proteins, caused a reduction in glutamate uptake (Fig. 2C) for both wild type and N206S. The reduction in activity was paralleled by a defect in glycosylation (Fig. 2D).
We also expressed mutant and wild-type human GLT1 in Xenopus oocytes. As seen in COS7 cells, N206S expressed in oocytes exhibited a lower apparent molecular weight compared with the wild type (Fig. 3A), and the uptake of glutamate was ϳ40% less than in wild type (Fig. 3B). Wild type and N206S showed the same apparent affinity for L-glutamate when determined by measuring glutamate-evoked uptake currents for a wide range of membrane potentials (Ϫ50 mV shown in Fig.  3C). However, there was a significant reduction in maximal transport velocity (Fig. 3C). At an external Na ϩ concentration of 96 mM, the apparent affinity for glutamate was 17 Ϯ 4 M for wild type and 18 Ϯ 2 M for N206S. When fitting the Na ϩ dependence of the inward currents (V m ϭ Ϫ50 mV; [Glu] ϭ 100 M) to the Hill equation, the apparent affinity for Na ϩ was 31 Ϯ 4 mM (n ϭ 3; n Hill ϭ 1.6 Ϯ 0.2) for the wild type and 28 Ϯ 5 mM (n ϭ 3; n Hill ϭ 1.5 Ϯ 0.2) for N206S (Fig. 3D). Thus, the strict Na ϩ dependence of human GLT1 is not affected by the N206S mutation. To study the possible effects of N206S on the voltage dependence of the forward uptake current, we subtracted the currents recorded in the absence from those measured in the presence of a saturating concentration (200 M) of L-glutamate and in response to a series of command voltage pulses. The wild-type-and mutant-mediated glutamate uptake currents exhibited no noticeable difference in the voltage dependence (Fig. 3E), and the plots could be well superimposed when normalized for the maximal current (data not shown). The lack of an outward current at depolarized membrane potentials (ՆϪ20 mV), despite the presence of Cl Ϫ in the external solution, is consistent with a negligible contribution of a thermodynamically uncoupled anion conductance to the current-voltage curves generated by both the wild type and N206S mutant. The N206S mutation seemed to have no impact on the affinity for the competitive nontransported inhibitor DHK (38). DHK blocked the glutamate-evoked transport current without inducing a current per se (5 M to 1 mM DHK; data not shown), and Hanes analysis revealed no difference in the K i for DHK in wild type and N206S (Fig. 3F). Overall, these results suggest that the N206S mutation in human GLT1 affects maximal forward uptake capacity but not the apparent affinity of the carrier for glutamate, Na ϩ , or transport inhibitor.
N206S Results in Impaired Plasma Membrane Expression-Reduction in maximal transport activity may be attributable to reductions in the catalytic activity, biosynthesis, or plasma membrane expression of the transporter. Indirect immunofluorescence, in combination with confocal laser-scanning microscopy of transfected COS7 cells, showed diffuse membrane labeling for wild-type human GLT1 (Fig. 4A), a significant decrease in plasma membrane and an increase in cytoplasmic  2. Extent of N-linked glycosylation of human GLT1. A, schematic diagram of human GLT1 shows the two putative glycosylation sites in the large extracellular loop between the third and fourth transmembrane domains. The GLT1 moiety from the sixth transmembrane domain to its end remains to be fully understood, and it is left undetermined. A novel topology for GLT1 was recently proposed (24). However, the segment of GLT1 comprising the glycosylation sites is represented as in our diagram. B, transfected COS7 cells were homogenized and treated with PNGase F (3 l; specific activity, 1,800,000 U/mg; New England Biolabs) for 3 h at 37°C and resolved on 10% SDS-polyacrylamide gel electrophoresis. C, COS7 were grown to near confluence in six-well plates, transfected, and treated with tunicamycin (5 g/ml, 0.1% Me 2 SO, 24 h) before measuring [ 3 H]D-aspartate uptake (30 min, room temperature). Cells were collected by scraping the dish with 1 ml of PBS plus 1% SDS. The lysate was used to determine the protein content, count of radioactivity, and gel electrophoresis. It should be noted that tunicamycin caused cell detachment from the wells. The groups treated with tunicamycin exhibited ϳ60% less protein content. Uptake values were normalized by protein content and expressed as percentage of wild-type GLT1 treated with 0.1% Me 2 SO. D, Western blot analysis of transfected COS7 treated with tunicamycin. The arrow indicates the fully deglycosylated transporter. hGLT, human GLT; WT, wild type. staining for N206S (Fig. 4A). No staining was observed for mock-transfected cells (data not shown). Immunofluorescence analysis was also performed in Xenopus oocytes. Injection of cRNA for wild type caused robust expression of the transporter in the plasma membrane with virtually no intracellular staining (Fig. 4A). Conversely, the N206S mutant resulted in a marked increase in the cytoplasmic punctate immunoreactivity, in parallel with a decrease in the plasma membrane expression. Water-injected oocytes exhibited negative labeling for both plasma membrane and cytoplasmic compartments (data not shown).
In addition, diminished plasma membrane expression of N206S was confirmed by selective labeling of the transporters at the cell surface by a biotinylation reaction. Compared with wild type, the decrease of the biotinylated N206S mutant at the plasma membrane of transfected COS7 cells is evident for both the oligomeric and monomeric forms of the transporter (Fig. 4B).
These findings strongly suggest that N206S is either targeted inefficiently to the cell surface or removed from the plasma membrane at a higher rate than wild type.
Glutamate Efflux in the N206S Mutant-Glutamate transporters are capable of running in both forward and reverse modes, and the overall direction is determined by the transmembrane electrochemical gradients of the transporter substrates and coupling ions. In GLT1-expressing oocytes, reverse transport of glutamate can be triggered by depolarized membrane potentials (V m Ն Ϫ20 mV) and measured as DHK-sensitive currents in the presence of high extracellular potassium and low extracellular sodium and the absence of external glutamate. Xenopus oocytes have a free pool of glutamate in the cytoplasm (ϳ10 mM), making this amino acid available as an intracellular substrate of the transporter (42). Under the above-mentioned conditions, wild-type human GLT1 generated an outwardly directed current, which could be blocked by extracellular application of DHK (Fig. 5, A and B, insets; Ϫ20 mV; Ref. 36). This DHK-sensitive reverse transport current at ϩ60 mV was 34 Ϯ 4% (n ϭ 12) of the forward uptake current recorded at Ϫ60 mV in Na ϩ buffer (ND96). In contrast, for N206S, the DHK-sensitive reverse transport current was 69 Ϯ 8% of the forward uptake current (Fig. 5, compare A and B). Moreover, in N206S, the reverse transport current at 0 mV represented ϳ80% of the total transporter-mediated currents (forward and reverse transport currents), whereas in wild-type hGLT1, the reverse transport current was ϳ 35% of the total transport currents (Fig. 5C). This suggests a marked propensity of the N206S mutant to extrude glutamate out of the cell. The same ratio was observed at Ϫ20 mV, whereas at ϩ20 mV

FIG. 3. Properties of N206S expressed in oocytes.
A, wild-type and N206S human GLT1 cRNAs were injected into Xenopus oocytes. Three days after injection, oocytes were homogenized in ND96 containing Triton X-100 (1%), resolved on SDS-polyacrylamide gel electrophoresis, blotted, and probed with an anti-rat GLT1 antibody (4 g/ml). B, glutamate uptake was measured for 30 min in oocytes expressing the glutamate transporters. Data are normalized to the uptake value of oocytes expressing wild-type human GLT1 and are the average of at least four experiments (n ϭ 10 for each group). C, Michaelis-Menten kinetics for glutamate-evoked uptake current recorded at Ϫ60 mV (n ϭ 4). Data are normalized to the uptake current evoked by 1 mM glutamate. D, sodium dependence of I gly . A representative concentrationresponse curve for extracellular sodium is shown. The uptake current was recorded at Ϫ50 mV and evoked by application of 100 M glutamate. Equimolar choline ϩ was used to compensate for the lack of Na ϩ ions. E, representative current-voltage relationship of oocytes expressing wild type and N206S in response to 100 M glutamate. the outward current tended to be the predominant current both for wild type and N206S.
This behavior of N206S was confirmed by radiotracer measurements. We injected an equal amount of radiolabeled Daspartate into oocytes expressing wild type and N206S to normalize the intracellular content of radioactivity. Incubation of the oocytes for 15 or 30 min in efflux buffer (ND80) caused a release of radioactivity that was statistically greater for N206S than for wild type (Fig. 5D), despite lower expression of N206S in the plasma membrane. The rate constant of the efflux process was 0.52 ϫ 10 Ϫ3 /min for N206S and 0.34 ϫ 10 Ϫ3 /min for wild-type human GLT1. On average, the ratio of the N206S and wild-type efflux rate constants from three independent experiments was 1.81 Ϯ 0.26. The difference in the efflux rate constant was even more pronounced if the expression of the wild type was adjusted down to the approximate level of expression of N206S, assessed as the extent of glutamate-evoked uptake current. At 30 min, the efflux rate constant of D-aspartate for N206S was 2.24-fold that of wild type (N206S-cRNA, 25 ng/50 nl, versus wild-type cRNA, 10 ng/50 nl; data not shown).
Another approach we followed to evaluate the efflux capacity of the N206S implied manipulation of the intracellular concentrations of Na ϩ and glutamate. A similar approach was used for glycine transporters 1b and 2a expressed in oocytes (43). Oocytes expressing wild-type or N206S-human GLT1 were injected with a solution containing MSG (23 nl, 1 M), thereby increasing the intracellular concentrations of Na ϩ and glutamate to ϳ40 mM (assuming an internal oocyte volume of 0.5 l). Fig. 6A shows a representative recording from an oocyte expressing wild-type human GLT1. Intracellular injection of MSG evoked an outward current that was partly blocked by a saturating concentration (1 mM) of DHK (Fig. 6A, #2). This concentration of DHK fully inhibited the forward uptake current evoked by application of 100 M glutamate in both wildtype and N206S-human GLT1 (Fig. 3) without inducing a current per se (recording not shown). The expected alteration of the ionic gradients for Na ϩ and glutamate was confirmed by the decreased extent of the forward uptake current evoked by application of extracellular glutamate in normal sodium solution (ND96; Fig. 6A, #1). We therefore attribute the DHKsensitive component of the outward current to the current evoked by the reverse operation of the transporter. The remaining outward current was 29.8 Ϯ 8 nA for wild type (n ϭ 8) and and wild-type human GLT1 (filled symbols) were injected with 23 nl of radiolabeled plus cold D-aspartate. Radioactivity injected into each oocyte was ϳ10,000 cpm. A nonspecific leak of D-aspartate was measured in oocytes that did not express the transporter. Data are from a representative experiment (n ϭ 10 in each group) and represent the value obtained by subtracting the nonspecific leak of D-aspartate and the DHK-insensitive component (measured in the presence of 1 mM DHK). The nonspecific leak represents ϳ 60% of the total release of D-aspartate, and the DHK-insensitive component of the remaining release was not statistically different in wild type (42 Ϯ 2.3%) and N206S (39 Ϯ 4%). hGLT, human GLT.

FIG. 6. Cytoplasmic injection of MSG evoked reverse transport currents in wild-type and N206S-human GLT1 oocytes.
A and B, recordings from representative oocytes show activation of outward currents after intracellular injection of MSG, as indicated by the arrows (23 nl, 1 M). Oocytes were held at Ϫ20 mV, and 50 M glutamate was applied to evoke the forward uptake current (filled bars). A, glutamate was applied in ND96 either before or after cytoplasmic injection on the steady level of the outward current. A decrease in the forward uptake current was noticed when glutamate was applied on top of the outward current because of the reduction of ionic gradients caused by the intracellular injection of MSG. The outward current was partly abolished by application of DHK (1 mM), and this component of the current was regarded as specific for human GLT1. B, the DHK-sensitive component of the outward current was compared in oocytes expressing similar levels of functional glutamate transporters. Oocytes were injected with 10 ng/50 nl of cRNA for wild type and 25 ng/50 nl of cRNA for N206S and clamped at Ϫ20 mV. C, Averaged forward and reverse transport currents recorded from wild-type (n ϭ 8) and N206S-human GLT1 (n ϭ 6) at Ϫ20 mV as shown in B. hGLT, human GLT; WT, wild type. 24.9 Ϯ 8 nA (n ϭ 6) for N206S and was statistically not different from the one measured in control oocytes (data not shown). It was, therefore, not further investigated.
The DHK-sensitive outward current was compared between wild type-and N206S-expressing oocytes. ND80 solution (20 mM K ϩ and 80 mM Na ϩ ; see "Experimental Procedures") was perfused after intracellular injection of MSG (Fig. 6B). For appropriate comparison, the expression level of wild-type human GLT1 was scaled down to match the forward glutamate uptake current of N206S. Under these conditions, the outward current blocked by DHK (reverse transport current) was much higher (Ͼ2-fold) for the N206S transporter compared with wild type (Fig. 6B). Moreover, the reverse transport current was 46.7 Ϯ 3.2% (n ϭ 8) of the forward uptake current in wild type and 285 Ϯ 20% (n ϭ 6) of the forward uptake current in N206S (Fig. 6C).
N206S causes selective down-regulation of wild-type human GLT1 activity-Our analysis on mutations in the human GLT1 gene in patients with ALS indicated that N206S is present in one allele of chromosome 11 (37). Because both normal and mutated genes coexist in vivo, we investigated whether N206S had any influence on the activity of wild-type human GLT1. A subsaturating amount of cRNAs (10 ng each) encoding wildtype or N206S-human GLT1 was injected in combination or alone into oocytes, and the resulting glutamate uptake-evoked current was measured at different membrane potentials. The neuronal glutamate transporter human EAAC1 was coinjected together with N206S or wild-type as a control to evaluate the selectivity of the effect of the N206S mutant. Although wildtype human GLT1 manifested greater activity then N206S, when coexpressed with N206S, the latter dominantly downregulated the wild-type to the level of activity determined by the mutant alone (Fig. 7). In contrast, coexpression of N206S with wild-type EAAC1 had no statistically significant effect on the uptake current mediated by the neuronal subtype. DISCUSSION We have analyzed the functional characteristics and potential pathophysiological impact of a naturally occurring mutation of the human glutamate transporter GLT1. The mutation involves the substitution of one of the two putative glycosylation sites, asparagine 206 to serine, and was identified in a patient with sporadic ALS (37). On the basis of Western blot analysis of homogenates prepared from transiently transfected COS7 cells and Xenopus oocytes, our data provide evidence for a defect in glycosylation in the GLT1 mutant N206S and document a 5-10-kDa reduction in apparent molecular mass. Furthermore, our analysis of the impact of the N206S mutation on the functional properties of the GLT1 carrier indicates that this mutation does not affect the affinities for glutamate, Na ϩ , and the nontransported inhibitor DHK (Fig. 3). These findings are consistent with previous observations reported for GLAST, the other known glial glutamate transporter isoform (44), indicating that the N-linked glycosylation is not involved in substrate recognition and Na ϩ coupling. However, we found substantial functional differences between wild-type and mutant transporter in terms of maximal transport capacity and the propensity to reverse the mode of operation. Oocytes and COS7 cells expressing N206S exhibit an ϳ50% reduction in V max . This reduced transport rate parallels an increase in immunoreactivity for GLT1 in the cytoplasmic compartment and a concurrent decrease in immunoreactivity on the plasma membrane of the cells.
A defect in plasma membrane targeting was also reported for other glycosylation-deficient transporters (45,46). The cytoplasmic retention of N206S may be caused by a slower rate of insertion in the plasma membrane from the Golgi apparatus. Alternatively, there may be increased instability of the newly synthesized protein, caused by aberrant folding of the glycosylation-deficient mutant, resulting in its accelerated retrieval from the plasma membrane and degradation. These alterations in membrane targeting or turnover of GLT1, and the resulting reduction in glutamate clearing capacity from the synaptic cleft, may significantly contribute to the increased susceptibility of neurons, in particular motor neurons, to excitotoxic insults. Evidence supporting a crucial role of GLT1 in preventing glutamate excitotoxicity comes from studies of GLT1-knockout mice, which are considerably more vulnerable to excitotoxic insults than control mice and exhibit exacerbation of brain injury (47).
The sporadic ALS patient with the N206S mutation was heterozygous for this mutation (37). Therefore, we became interested in exploring how the simultaneous presence of wildtype and mutant GLT1 affects cellular glutamate uptake capacity. We show that N206S has a substantial dominant negative impact on wild-type activity. This may significantly impair the capacity of GLT1 to mop up glutamate at synapses. However, how the mutated transporter prevents the wild-type from being fully active is not yet known. It is reasonable to speculate that the efficient expression of the transporter in the plasma membrane requires its assembly as a fully glycosylated oligomer in the endoplasmic reticulum. Evidence in this direction came from Eskandari and colleagues (48), who showed that one member of the glutamate transporter family, the neuronal glutamate transporter EAAC1, assembles in the plasma membrane as a pentamer made of identical subunits. Therefore, it is possible that oligomers formed by incompletely glycosylated monomers may be recognized as aberrant and may not be targeted efficiently to the plasma membrane or, if correctly targeted, may be actively removed from the membrane and degraded. Data from the literature suggest that, in vivo, GLT1 monomers may self-associate to form homo-oligomers (49). Dominant-negative interactions between transporter monomers produced by aberrant splicing of the RNA have already been reported to account for a loss of glutamate uptake mediated by GLT1 in sporadic ALS (29). The finding that N206S does not exert this dominant negative effect on another glutamate transporter isoform, such as the neuronal subtype EAAC1, supports the concept that, whereas the differ-FIG. 7. The N206S mutant exerts a dominant negative effect on wild-type human GLT1 activity. 10 ng of cRNA was injected for each transporter. Current-voltage curves were obtained by bath application of 100 M glutamate. Data are normalized either to the uptake-evoked current generated by wild-type human GLT1 (WT; A) or by human EAAC1 (B) at Ϫ100 mV and are averaged from three different experiments (n ϭ 12). ent glutamate transporters are made up by homo-oligomers, subunits of two different transporter subtypes do not assemble together.
Glutamate transporters are capable of working in both forward and reverse directions, depending on the intracellular and extracellular ionic concentrations. The normal electrochemical gradients of Na ϩ , H ϩ , and K ϩ ensure glutamate uptake against a large concentration gradient (forward transport). In pathological situations such as during brain ischemia or even during intense and repetitive neuronal firing, there is depolarization of the cell membranes caused by elevation of extracellular K ϩ concentration in the synaptic cleft and rundown of electrochemical gradients. This induces the glutamate transporters to reverse their direction, allowing glutamate to exit the cells (reverse transport; Refs. 42,50,51). An interesting feature of N206S is its marked propensity to reverse transport operation relative to the forward uptake (Figs. 5 and 6). This capability is less pronounced for the wild type, both in terms of DHK-sensitive outward current and efflux of radiolabeled transporter substrate. How can a mutation in one of the glycosylation sites facilitate the reverse operation of the transporter? Although we do not have an explanation for this yet, we anticipate that the N206S mutation affects a specific step in the glutamate transport cycle, which leads to an increase in reverse transport. Unfortunately, the resolution of radiotracerbased flux techniques in conjunction with the relatively low levels of expression of N206S and the relatively slow solution exchange protocol typically used in the two-electrode voltage clamp approach limit the possibilities for dissecting these features of the transporter.
It is well established that both sporadic and familial ALS are associated with impairment of the glutamate transport system mediated by the glial glutamate transporter GLT1 (27-29, 36, 52, 53). The combined evidence that a missense mutation of human GLT1 is present in a patient with sporadic ALS, and that such a mutation affects both the capacity of a cell to clear glutamate and the ability to release glutamate, supports the concept that this mutation contributes to the excitotoxicity that occurs in ALS. In addition, its analysis has provided new insights into the role of N-glycosylation on glutamate transporter function and trafficking.