The individual subunits of the glutamate transporter EAAC1 homotrimer function independently of each other

Molecular Biology and Transient Expression

Wild-type EAAC1 cloned from rat retina was subcloned into pBK-CMV (Stratagene) as described previously (19) and was subjected to site-directed mutagenesis according to the QuikChange protocol (Stratagene, La Jolla, CA) as described by the supplier. Cys-less EAAC1 was produced by the same procedure by sequentially mutating the five cysteine residues in EAAC1_WT (C9, C158, C218, C255, C342) to serine. The primers for mutagenesis were obtained from the DNA core lab, Department of Biochemistry at the University of Miami School of Medicine. The complete coding sequences of mutated EAAC1 clones were subsequently sequenced. Wild-type and mutant EAAC1 constructs (named pCMV-EAAC1_WT or pCMV-EAAC1_mutantform) were used for transient transfection of sub-confluent human embryonic kidney cell (HEK293, ATCC number CGL 1573 or HEK293T/17, ATCC number CRL 11268) cultures using the calcium phosphate-mediated transfection method (20) as described previously (19, 21). For coexpression experiments cDNAs of the mutant and wild-type receptors were used for transfection in a 1:1 ratio, unless stated otherwise. Electrophysiological recordings were performed between days 1 to 3 post-transfection.

Membrane vesicle preparation and crosslinking

Glutamate transporter-expressing HEK293 cells were homogenized on ice in 10 mM HEPES-NaOH, pH 7.4, containing 100 mM NaCl, 1 mM Na₄-EDTA, Protease-Inhibitor Complete (Roche Applied Science, Mannheim, Germany) in a dilution according to the manufacturer’s instruction and centrifuged for 10 min at 400xg to remove nuclei, cell debris and undisrupted cells (P1 fraction). The supernatant was centrifuged for 30 min at 30,000×g at 4° C. The resulting pellet (P2 fraction/membrane vesicles) was processed according to the respective crosslinking protocols (next paragraphs). Crosslinking products were analyzed by SDS-PAGE and Western blotting. Controls were processed under identical conditions prior to and after the experiment, however, without adding enzymatic or chemical crosslinkers (see Fig 3, and Fig. 10 Supporting Information. A – F, lanes 1 and 6, respectively). The experimental procedures for chemical and enzymatic crosslinking can be found in the Supporting Information.

SDS-PAGE and Western-blot analysis

Crosslinking products were separated and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using a 6% separating gel (0.5 mm thick) and Western blot analysis. Immunoblots were prepared with affinity-purified antibodies directed against EAAC1 (BioTrend Chemikalien GmbH, Köln, Germany) in a dilution of 1:1000 as described previously (22). The antibody reaction was detected by chemiluminescence using Hyperfilm-ECL (Amersham ECL-Detection kit; Amersham Biosciences UK Limited, Little Chalfont, UK). In order to determine the apparent molecular weight (MW) of the crosslinking products, the Pre-stained Protein Marker, Broad Range (6,5–175 kDa) (New England Biolabs, Beverly, MA, USA), was used in parallel with “HiMark™” High Molecular Protein Standard (40–500 kDa) (Invitrogen GmbH, Karlsruhe, Germany) to determine molecular weights higher than 175 kDa. MW marker proteins on immunoblots were stained with Protogold (Plano GmbH, Wetzlar Germany). The MW of each crosslinking product was determined using the gel documentation software Total Lab (Nonlinear Dynamics, Newcastle upon Tyne, UK).

Co-immunoprecipitation

Transfected HEK293 cells were washed two times with phosphate-buffered saline, PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄ and 2 mM KH₂PO₄, pH 7.4). Cells were collected after adding cell lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, protease inhibitor cocktail 2, Sigma) by mechanical trituration. Cell membranes were further broken down by freezing and thawing three times and then passing through a 26 gauge needle 10 times. The resulting suspension was centrifuged at 1,000×g for 10 min to remove debris. The concentration of the supernatant (cell lysate) was adjusted to 1μg/μl. Protein A agarose beads (Upstate) were added to 500 μl cell lysate and incubated for 1 hour with gentle rocking at 4°C to pre-clear the cell lysate. After centrifugation and transfer of the supernatant to another tube, anti-GFP antibody (BioVision) was added and the solution was incubated with gentle rocking overnight at 4°C. Protein A agarose beads were added and incubated with gentle rocking for 2 hours at 4°C. The beads were washed three times with 500 μl of cell lysis buffer. Protein was eluted from the beads with 2 × Laemmli buffer and the sample was heated to 65°C for 10 minutes. The sample was loaded onto a 10% SDS-PAGE gel, analyzed by Western blotting, and protein bands were detected with ECL reagent.

Electrophysiology and Rapid Solution Exchange

Glutamate-induced EAAC1 currents were recorded with an Adams & List EPC7 amplifier under voltage-clamp conditions in the whole-cell current-recording configuration (23). The typical resistance of the recording electrode was 2–3 MΩ; the series resistance was 5–8 MΩ. Because the glutamate-induced currents were small (typically < 500 pA), series resistance (R_S) compensation had a negligible effect on the magnitude of the observed currents (< 4% error). Therefore, R_S was not compensated. The extracellular solution contained (in mM): 140 NaCl, 2 CaCl₂, 2 MgCl₂, and 30 4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES). Two different pipette solutions were used depending on whether mainly the non-coupled anion current (with thiocyanate) or the coupled transport current (with chloride) was investigated (19, 24). These solutions contained (in mM): 130 KSCN or KCl, 2 MgCl₂, 10 tetraethyl-ammonium chloride (TEACl), 10 ethylene glycol-bis-(β-aminoethyl ether)-N,N,N′,N′-tetracetic acid (EGTA), and 10 HEPES (pH 7.4/KOH). Thiocyanate was used because it enhances glutamate transporter associated currents and allows the detection of the EAAC1 anion-conducting mode (19, 25).

For the electrophysiological investigation of the Na⁺/glutamate homo-exchange mode (24) the pipette solution contained (in mM) 140 NaCl/NaSCN, 2 MgCl₂, 10 TEACl, 10 EGTA, 10 glutamate, and 10 HEPES (pH 7.4/NaOH). In this transport mode, the same concentrations of Na⁺ are used on both sides of the membrane, and concentrations of glutamate are used on the intra- and extracellular side which saturate their respective binding site. The affinity of EAAC1 for cytoplasmic glutamate is 280 μM (26). Therefore, we used 10 mM cytoplasmic glutamate in order to ensure saturation of the intracellular binding site. In contrast the affinity for extracellular glutamate is about 6 μM (19) and 100 – 200 μM extracellular glutamate is sufficient for saturation of this binding site. When permeating anions, such as SCN⁻, are present, establishment of homoexchange conditions leads to the permanent activation of an anion current (24, 27, 28). This permanent anion current was used in this work as a tool to study the behavior of mutant transporters in the homoexchange mode.

The currents were low pass filtered at 1–10 kHz (Krohn-Hite 3200), and digitized with a digitizer board (Axon, Digidata 1200) at a sampling rate of 10–50 kHz, which was controlled by software (Axon PClamp). All the experiments were performed at room temperature.

Rapid solution exchange was performed as described previously (19). Briefly, substrates were applied to the EAAC1-expressing cell by means of a quartz tube (opening diameter: 350 μm) positioned at a distance of ~0.5 mm to the cell. The linear flow rate of the solutions emerging from the opening of the tube was ~5–10 cm/s, resulting in typical rise times of the whole-cell current of 30–50 ms (10–90%).

Data Analysis

Non-linear regression fits of experimental data were performed with Origin (Microcal software, Northampton, MA) or Clampfit (pClamp8 software, Axon Instruments, Foster City, CA). Dose-response data were described with a Michaelis-Menten-type relationship. To describe the properties of mixed populations of wild-type and mutant transporters we used a sum of two Michaelis-Menten-type equations.

Functional characterization of EAAC1_R446Q

We first tested the specificity of EAAC1_R446Q expressed in HEK293 cells for neutral amino acids. When 500 μM alanine was applied to EAAC1_R446Q in the homoexchange mode (140 mM NaSCN and 10 mM alanine internal, see Fig. 1B for a graphical illustration of the transport modes used), it generated a large inwardly-directed anion current (Fig. 1C, left panel), caused by the outward movement of SCN⁻. In contrast, application of 500 μM glutamate to the same cell did not result in any detectable current (Fig. 1C, middle panel). Similar results were obtained for concentrations of glutamate up to 1 mM (Fig. 1D). The typical glutamate concentration used in the coexpression experiments was 50 μM. Thus, these results demonstrated that glutamate does not activate anion current in EAAC1_R446Q at the typical concentrations used.

The alanine-induced current was dependent on the alanine concentration (Fig. 1D), saturating with an apparent K_m of 20.1 ± 1.0 μM (n = 3). On average, the alanine-induced anion current at saturating alanine concentrations was −398 ± 155 pA (from 12 cells). This compared well to the −475 ± 120 pA (from 20 cells) obtained for glutamate activation of EAAC1_WT anion current under homoexchange conditions, showing that the R446Q amino acid exchange did not result in a dramatically changed expression level of the mutant as compared to the wild-type transporter. In addition to alanine, we also tested glutamine as a substrate for EAAC1_R446Q. As shown in Fig. 1D, glutamine generated maximum anion currents indistinguishable from the alanine-induced currents, but activated the transporter with a 4-times higher apparent affinity of 5.1 ± 1.0 μM (n = 3). Finally, we tested whether alanine and glutamine are substrates for EAAC1_WT. At the concentration range used for the coexpression experiments (up to 500 μM), no activation of anion current was observed for either of these two neutral amino acids, indicating that they were not recognized as substrates by EAAC1_WT.

Properties of mixed EAAC1_WT-EAAC1_R446Q

Next, we characterized the functional properties of a mixed population of wild-type and R446Q-mutant transporters. Initially, these experiments were performed in the homoexchange mode because EAAC1_R446Q does not catalyze steady-state forward transport (Fig. 1C, right panel). Application of 50 μM glutamate to HEK293 cells expressing a mixed population of these transporters resulted in generation of inwardly-directed anion current (Fig. 2A, left panel). A slightly smaller anion current (I_Gln/I_Glu = 0.78 ± 0.13, n = 5) was induced by 50 μM glutamine (Fig. 2A, right panel). When both substrates were applied at the same time the anion current was 1.80 ± 0.21 times the current in the presence of only glutamate (Fig. 2A, middle panel), suggesting that the currents induced by the two substrates are purely additive, pointing to independent operation of the subunits of the trimer. The statistical analysis of the results from 5 cells is shown in Fig. 2B.

To test for statistical coassembly of EAAC1_WT and EAAC1_R446Q, we expressed the two proteins in varying amounts by altering the cDNA ratio used for the transfection from 1:4 to 4:1. As shown in Fig. 2C, the ratio of the currents induced by glutamate and alanine was in agreement with that expected for pure statistical coassembly of wild-type and mutant transporter.

Although is has been previously shown that EAAT2 forms homotrimers (14, 17), the subunit stoichiometry of EAAC1 has so far not been determined by using biochemical methods. Therefore, we used Western blotting to test whether EAAC1 also assembled as a trimer. We used a crosslinking approach based on three different crosslinking reagents, two chemical (thiol-specific (BMDB) and amino-specific (BS³)) and one enzymatic crosslinking approach (transglutaminase reaction). Crosslinking with BMDB resulted in the formation of two prominent bands after Western blotting (Fig. 3A), corresponding to the monomer (69 kDa) and the trimer (200 kDa (*b in Fig. 3A). In addition, some lower intensity bands were observed that indicated the formation of dimers and of a higher molecular weight species, possibly a tetramer. Multimer formation was not observed when a cysteine-less variant of EAAC1 was treated with BMDB (Fig. 3B). A detailed description of the crosslinking data, including crosslinking with BS³ and transglutaminase, can be found in the Supporting Information. Together, the crosslinking data suggest that EAAC1 forms homotrimers.

In order to directly test whether EAAC1_R446Q can co-assemble with wild-type EAAC1, we performed co-immunoprecipitation (co-IP) experiments. Wild-type transporter and EAAC1_R446Q were co-expressed with an EAAC1_WT-YFP fusion protein. Anti-GFP antibody was used to perform the co-IP. Proteins were separated by 10% SDS-PAGE and analyzed by Western blot with anti-EAAC1 as the primary antibody. As shown in Fig. 4, EAAC1_R446Q protein was co-immunoprecipitated from cells that co-express EAAC1_R446Q and EAAC1_WT-YFP. No EAAC1_R446Q protein was co-immunoprecipitated from cells that expressed EAAC1_R446Q alone which excluded the possibility that anti-GFP antibody can interact with EAAC1_R446Q. As a control, co-immunoprecipitation of wild-type transporter and EAAC1_WT-YFP was also performed with similar results (data not shown). These results clearly show that EAAC1_R446Q can co-assemble with wild-type EAAC1.

So far, we have characterized the glutamate and alanine translocation properties of EAAC1_WT-EAAC1_R446Q by locking the transporters in the homoexchange mode. Since the translocation of the K⁺-bound transporter form is impaired in EAAC1_R446Q, we next measured steady-state transport currents in the mixed population of transporters, in order to test whether cooperativity of the subunits is required for the relocation reaction. As shown in Fig. 5A (left panel), application of 50 μM glutamate to the transporters in the forward transport mode resulted in the activation of large anion currents (−290 ± 70 pA, n = 4, KSCN intracellular solution), even in the total absence of alanine. This result suggested that occupation of the alanine binding site of EAAC1_R446Q was not required for forward glutamate transport in the wild-type transporter. As expected, application of 200 μM alanine in the forward transport mode to the same cells generated only 13% of the glutamate-induced inward anion currents (−39 ± 11 pA, n = 4, Fig. 5A, middle panel). Application of both substrates at the same time resulted in an anion current that was purely additive, as shown in the right panel of Fig. 5A.

The magnitude of the anion current carried by SCN⁻ outflow is a good measure for the forward transport activity of wild-type and mutant glutamate transporters, since it is believed that anion conductance is associated with few specific states within the transport cycle, such as the Na⁺, H⁺, and glutamate-bound state (see for example (19, 24, 28, 29)). Furthermore, it was shown that the anion current is absent when steady-state transport is inhibited in the wild-type transporter, or by specific mutants such as E373Q (21) and R446Q (this work). However, to directly test the effect of subunit interaction on glutamate transport, we measured transport currents in the absence of a driving force for anions across the membrane (symmetrical Cl⁻ concentrations on both sides of the membrane, 0 mV, Fig. 5C), in order to confirm the results obtained from the anion currents in the forward transport mode. As expected, 50 μM glutamate elicited inward transport currents in mixed EAAC1_WT-EAAC1_R446Q transporters (−48 ± 2 pA, n = 4, KCl intracellular solution), even in the absence of alanine or glutamine (Fig. 5C, left panel). In contrast, application of 50 μM glutamine did not induce any measurable currents (−1 ± 1 pA, n = 4, Fig. 5C, middle panel). The same result was found throughout the tested range of transmembrane potentials from −90 to +60 mV (Fig. 5D). The absence of glutamine-induced currents was not due to the absence of EAAC1_R446Q transporters, since application of glutamine to the same batch of cells in the presence of 140 mM extracellular SCN⁻ resulted in the activation of outwardly-directed anion currents (+220 ± 34 pA, n = 4, 140 mM NaCl, 10 mM alanine containing intracellular solution).

To further test for crosstalk between the individual subunits, we determined whether the substrate affinity of one subunit depended on the occupancy of a neighboring subunit in the trimer. In trimers that contained only wild-type transporter, the K_m for glutamate in the homoexchange mode was 13 μM. For trimers that contained both EAAC1_WT and EAAC1_R446Q transporters, the K_m for glutamate was 8.9 ± 1.0 μM (n = 3) in the absence of alanine. In the presence of 500 μM alanine the K_m for glutamate did not change within experimental error (8.0 ± 1.1 μM, n = 3, Fig. 5B). The same experiments were repeated in the forward transport mode. In this mode, the substrate binding sites of EAAC1_R446Q should be mainly exposed to the intracellular side in the presence of alanine. We found a K_m for glutamate of 5.4 ± 0.6 μM (n = 3, Fig. 5B) in the absence of alanine and 5.6 ± 1.1 μM (n = 3, Fig. 5B) in the presence of alanine. Together, these results suggest that the substrate affinity of one subunit in the trimer is independent of the state of the neighboring subunits.

Finally, we tested specific competitive inhibitors of either the wild-type transporter or EAAC1_R446Q for their effects in the mixed trimer. The glutamate transporter inhibitor DL-threo-β-benzyloxyaspartate (TBOA) is a potent inhibitor of EAAC1 with a K_i in the sub-micromolar range (19, 30), but it did not block alanine-induced currents in EAAC1_R446Q at concentrations up to 200 μM. When 100 μM TBOA was applied to trimers containing both wild-type transporter and EAAC1_R446Q an outward current was induced (+44 ± 4 pA, n = 3, Fig. 6A, left panel). This outward current was caused by inhibition of the sustained leak anion conductance of EAAC1 in the presence of intracellular SCN⁻ (19, 31). However, TBOA did not inhibit the inward anion current generated by 500 μM alanine (Fig. 6A, left panel). In addition, we used benzylserine, which is a competitive inhibitor of the neutral amino acid transporter ASCT2 (32). 1 mM benzylserine did not induce any currents in EAAC1, but it generated outward current in EAAC1_R446Q, due to the inhibition of the sustained anion leak (determined by the voltage dependence of the benzylserine-induced current, data not shown), which is conserved in this mutant transporter. The apparent affinity of EAAC1_R446Q for benzylserine was determined to 160 ± 10 μM (n = 4, data not shown). In the mixed EAAC1_WT-EAAC1_R446Q trimer, application of 1 mM benzylserine also resulted in outward current (+49 ± 5 pA, n = 3, Fig. 6B, left panel). Similar to TBOA, however, benzylserine did not significantly inhibit glutamate-elicited inward current (Fig. 6B, left panel). When both inhibitors where applied at the same time, an outward current of 84 ± 9 pA (n = 3) was observed, which is 91% of the current expected if both inhibitors inhibited their respective subunits independently. The results are statistically summarized in Fig. 6C. Together, these results indicated that inhibition of one subunit in the trimer did not affect the function of the neighboring subunits.

Properties of mixed EAAC1_WT-EAAC1_H295K oligomers

So far, the data obtained from the coexpression experiments with the wild-type and R446Q transporter did not strictly exclude the possibility that glutamate and the cotransported cations are transported through a single pore in the middle of the trimeric subunit assembly. To test this possibility, we coexpressed the wild-type transporter with EAAC1_H295K. We had shown previously that EAAC1_H295K has a dramatically reduced affinity for glutamate (33). Therefore, we expected that coexpression of EAAC1_WT and EAAC1_H295K would produce transporters with an affinity intermediate of the wild-type and the mutant transporter, if both mutant and wild-type subunits would line a central permeation pathway for glutamate. In contrast, the experimental dose response curve for glutamate shown in Fig. 7A could not be described by a Michaelis-Menten-like relationship with intermediate glutamate affinity (dotted line in Fig. 7A), but was rather consistent with two independent populations of transporters with K_m values for glutamate of 7.4 ± 0.9 μM and 730 ± 80 μM (n = 4), respectively. These K_m values fit very well with the K_m of the pure wild-type and H295K transporters of 13 μM and 600 ± 60 μM (n = 5), respectively, as illustrated by the dashed lines in Fig. 7A. The contribution of the EAAC1_H295K current to the total current of the mixed population was 43 ± 4 %, in agreement with previous findings, showing that EAAC1_H295K catalyzes less anion current than wild-type EAAC1 (33).

In addition to having a dramatically different K_m for glutamate, EAAC1_H295K has the effect of slowing the activation of the glutamate-induced anion current by about a factor of 100 (33), as shown in Fig. 7B. Whereas application of 1 mM glutamate to EAAC1_WT resulted in a single-exponential anion current activation with a time constant of 45 ± 5 ms (Fig. 7B, left panel), which reflects the time resolution of the rapid solution exchange system, activation of the anion current in EAAC1_H295K occurred with a time constant of 790 ± 5 ms (n = 4, Fig. 7B, right panel). In contrast to the single-exponential rising behavior of the pure transporters, anion currents of the mixed population of wild-type and H295K transporters were activated with a double-exponential time course (Fig. 7B, middle panel). The time constants for the two phases of the current rise were 33 ± 3 ms and 1090 ± 40 ms, respectively. These results suggested that the kinetic properties of the mixed wild-type and H295K-mutant transporters can be described by a simple sum of the kinetic properties of the individual transport proteins.

Properties of mixed EAAC1_WT-EAAC1_E373Q

The glutamate residue E373 is conserved in the mammalian glutamate transporter family (34), but not in ASC transporters (6, 21). We had previously proposed that E373 is involved in proton cotransport by EAAC1, since the E373Q mutation abolishes the intrinsic pH dependence of glutamate transport (21). Here, we coexpressed EAAC1_WT with EAAC1_E373Q and determined the functional properties of the mixed transporters as a function of extracellular pH. At pH 7.3 EAAC1_WT and EAAC1_E373Q have very similar apparent affinities for glutamate (19), 13 μM and 10 μM, respectively. Consistently, the dose response relationship of the mixed transporter population, shown in Fig. 8 (open circles), was well described by a Michaelis-Menten-like relationship with an apparent K_m of 12.6 ± 1.2 μM (n = 3). In contrast, the dose-response relationship was more complex at an extracellular pH of 10.0 (Fig. 8, closed circles). Assuming an additive response of EAAC1_WT and EAAC1_E373Q to glutamate, the dose response curve could be described by a sum of two Michaelis-Menten-like curves, one yielding an apparent K_m of 11.3 ± 0.3 μM, the other component yielding a K_m of 420 ± 71 μM. These results are consistent with the existence of two independent transporter populations, one being the wild-type transporter with a shift of the K_m to 610 μM at pH 10.0 (right dotted line in Fig. 8 (26)), the other being EAAC1_E373Q with its virtually pH-independent K_m that remains at 17 μM at pH 10 (21).