Transporter-associated currents in the gamma-aminobutyric acid transporter GAT-1 are conditionally impaired by mutations of a conserved glycine residue.

To determine whether glycine residues play a role in the conformational changes during neurotransmitter transport, we have analyzed site-directed mutants of the gamma-aminobutyric acid (GABA) transporter GAT-1 in a domain containing three consecutive glycines conserved throughout the sodium- and chloride-dependent neurotransmitter transporter family. Only cysteine replacement of glycine 80 resulted in the complete loss of [(3)H]GABA uptake, but oocytes expressing this mutant exhibited the sodium-dependent transient currents thought to reflect a charge-moving conformational change. When sodium was removed and subsequently added back, the transients by G80C did not recover, as opposed to wild type, where recovery was almost complete. Remarkably, the transients by G80C could be restored after exposure of the oocytes to either GABA or a depolarizing pre-pulse. These treatments also resulted in a full recovery of the transients by the wild type. Whereas in wild type lithium leak currents are observed after prior sodium depletion, this was not the case for the glycine 80 mutants unless GABA was added or the oocytes were subjected to a depolarizing pre-pulse. Thus, glycine 80 appears essential for conformational transitions in GAT-1. When this residue is mutated, removal of sodium results in "freezing" the transporter in one conformation from which it can only exit by compensatory changes induced by GABA or depolarization. Our results can be explained by a model invoking two outward-facing states of the empty transporter and a defective transition between these states in the glycine 80 mutants.

During ion-coupled transport, conformational changes have to take place to expose ion and solute binding sites alternately to the two sides of the membrane. An important class of solute transporters is the sodium-coupled neurotransmitter transporters, which play a crucial role in synaptic transmission and are members of two distinct families (for reviews, see Ref. 1 and 2). Evidence for conformational changes during neurotransmitter transport has been obtained using ion-and substrate-dependent changes in accessibility of trypsin-sensitive sites (3,4) and engineered cysteines (5-7) as well as by changes in fluorescence of site-directed probes (8). Glycines can introduce flex-ibility in the protein structure, and recently a glycine residue has been implicated to act as a gating-hinge in potassium channels (9). Moreover evidence has been presented suggesting that glycine residues engineered into the proton-coupled lactose transporter confer conformational flexibility to it (10). Our aim was to evaluate the idea that one or more of the native glycine residues in neurotransmitter transporters may play a role in conformational transitions during transport. Therefore, we have examined the role of a stretch of six conserved residues of the GABA 1 transporter GAT-1, which contains three consecutive glycine residues and is located in the first extracellular loop and is highly conserved in the SLC6 transporter family.
GAT-1 catalyzes sodium/chloride/GABA co-transport with a stoichiometry of 2:1:1 (11)(12)(13)(14), but there is still some dispute on the precise role of chloride (15). GABA transport is electrogenic (11)(12)(13)(14)(15). Therefore, the process can be monitored not only by radioactive GABA uptake but also by sodium-dependent GABA-induced steady-state currents reflecting electrogenic GABA translocation (12)(13)(14)(15)(16). In addition to these steady-state currents, GAT-1 also mediates sodium-dependent transient currents in the absence of GABA. These transients are thought to reflect a charge-moving conformational change (13, 16 -19). Moreover, in the absence of both sodium and GABA, GAT-1 mediates cation-leak currents, which are usually monitored in lithium-containing media (16,17). These currents are suppressed by sodium in a non-competitive manner and probably carried by the sodium-free transporter (16,20). Here we describe that mutation of only one of the three external loop 1 glycines of GAT-1, glycine 80, to conserved residues causes the above conformational transitions to become impeded upon removal of sodium. However, this is a conditional loss because these transporter-associated currents can be "rescued" by GABA or by prolonged depolarization. Our results can be explained by a model invoking two distinct conformations of the outward-facing sodium-free transporter.

EXPERIMENTAL PROCEDURES
Generation and Sub-cloning of Mutants-Mutations were made by site-directed mutagenesis of the C74A-GAT-1 in the vector pBluescript SK (Ϫ) (Stratagene) according to the Kunkel method as described (21,22). Briefly, the parent DNA was used to transform Escherichia coli CJ236 (dut Ϫ , ung Ϫ ). From one of the transformants, single-stranded uracil-containing DNA was isolated upon growth in uridine-containing medium according to the standard protocol from Stratagene using helper phage R408. This yields the sense strand, and consequently mutagenic primers were designed to be antisense. Mutants were subcloned into a construct containing GAT-1 in the oocyte expression vector pOG1 (18) using the unique restriction enzymes ClaI and AvrII. The * This work was supported by The Israel Science Foundation Grant 488/03-16.1 and the Bernard Katz Minerva Center for Cellular Biophysics. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
[ 3 H]GABA Transport in HeLa Cells-HeLa cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 200 units/ml penicillin, 200 g/ml streptomycin, and 2 mM glutamine. HeLa cells plated on 24-well plates were infected with recombinant vaccinia/T7 virus vTF7-3 (23) and transfected with cDNA (pBluescript SK Ϫ with wild type or mutant transporter inserted downstream to the T7 promoter) using the transfection reagent DOTAP (N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl sulfate) as described (24). Uptake of [ 3 H]GABA into the cells was assayed 18 -20 h post-transfection. The wells were washed twice with a solution containing 150 mM choline chloride, 5 mM KP, pH 7.4, 0.5 mM MgSO 4 , and 0.3 mM CaCl 2 . Each well was then incubated with 0.4 Ci of [ 3 H]GABA (99 Ci/mmol) in a NaCl transport solution (150 mM NaCl, with KP i , MgSO 4 , and CaCl 2 as above). Transport reactions were carried out for 10 min at room temperature, and the assay was terminated by washing the cells twice with ice-cold NaCl transport solution. Cells were lysed with 1% SDS, and radioactivity was measured by liquid scintillation counting.
Inhibition Studies with Sulfhydryl Reagents-Before the transport measurements, the cells adhering to 24-well plates were washed twice with 1 ml of the transport medium containing 150 mM choline chloride instead of NaCl. Each well was then incubated at room temperature with 200 l of the preincubation medium containing the sulfhydryl reagent (the different compositions are indicated in the legend to Fig.  1). After 5 min the medium was aspirated, and the cells were washed twice with 1 ml of the transport solution. Subsequently they were assayed for [ 3 H]GABA transport at room temperature (22-26°C) unless indicated otherwise. The hydrophilic methanethiosulfonate reagents used during the preincubation were purchased from Toronto Research Chemicals, Inc.
cRNA Transcription, Injection, and Oocyte Preparation-Capped runoff cRNA transcripts were made from transporter constructs in pOG1 linearized with SacII using mMessage mMachine (Ambion Inc.). Oocytes were removed from anesthetized Xenopus laevis frogs and treated with collagenase (type 1A, Sigma C-9891) until most of the capillaries were absent and injected with 50 nl of undiluted cRNA (around 1 ng/nl) the same or the next day. Oocytes were maintained at 18°C in modified Barth's saline (88 mM NaCl, 1 mM KCl, 1 mM MgSO 4 , 2.4 mM NaHCO 3 , 1 mM CaCl 2 , 0.3 mM Ca(NO 3 ) 2 , and 10 mM HEPES, pH 7.5) with freshly added 2 mM sodium pyruvic acid and 0.5 mM theophylline and supplemented with 10,000 units/liter penicillin, 10 mg/liter streptomycin, and 50 mg/liter gentamycin.
Oocyte Electrophysiology-This was done as described (16). Oocytes were placed in the recording chamber, penetrated with two agarose (1%)-cushioned micropipettes (back-filled with 2 M KCl, resistance varied between 0.5 and 2 megaohms), and voltage-clamped using GeneClamp 500 (Axon Instruments, Inc.) and digitized using Digidata 1200A (Axon Instruments, Inc., Union City, CA), both controlled with the pClamp6 suite (Axon Instruments, Inc.). Currents were acquired with Clampex 6.03 and low pass-filtered at 10 kHz every 0.5 ms. Oocytes were stepped from Ϫ140 mV to ϩ60 mV in 25-mV increments using Ϫ25 mV as the holding potential. Each potential was held clamped for 500 ms. In the experiments depicted in Fig. 5, the voltage was stepped in 9-mV increments from Ϫ160 mV to ϩ11 mV in the case of C74A and from Ϫ100 mV to ϩ71 mV for G80C/C74A. In all experiments, after each voltage jump the oocytes were held at Ϫ25 mV for 500 ms before jumping to the next voltage. The membrane potential was measured relative to an extracellular Ag ϩ /AgCl electrode in the recording chamber. The recording solution (ND) contained 96 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , and 5 mM HEPES, pH 7.4. In substitution experiments sodium ions were replaced with equimolar choline or lithium. All electrophysiological experiments shown in the figures are in the absence of MTSET.

RESULTS
Cysteine Scanning Mutagenesis-The amino acid residues of the consensus sequence NGGGAF (residues 77-82 in GAT-1), located in the first extracellular loop and conserved throughout the sodium-and chloride-dependent neurotransmitter transporter family, were replaced one at a time by cysteine. This was done in the background of the C74A mutant, because cysteine 74 is the only cysteine in wild type GAT-1 sensitive to modification by impermeant methanethiosulfonate reagents. Using the vaccinia/T7 recombinant virus-based expression system, we have shown previously that the sodium-and chloride-dependent [ 3 H]GABA transport activity of HeLa cells expressing C74A is almost as high as those expressing wild type (6). Significant transport activity was exhibited by four of the six cysteine mutants: N77C/C74A, G78C/C74A, G79C/C74A, and A81C/C74A (Fig. 1A). No sodium-and chloride-dependent [ 3 H]GABA uptake could be detected in the case of G80C/C74A and F82C/C74A (Fig. 1A). Surface-biotinylation experiments showed that this was not due to inefficient targeting of G80C and F82C to the plasma membrane (data not shown). The four active cysteine mutants were all functionally affected by a FIG. 1. Transport activity of cysteine substitution mutants and sensitivity to MTSET. Cysteine mutants in the C74A background were transiently expressed in HeLa cells as described under "Experimental Procedures." 18 h after transfection, sodium-dependent [ 3 H]GABA transport was measured at room temperature for 10 min (A). The cells expressing GAT-1-C74A or the indicated mutants were treated with 200 M MTSET in the choline chloride-containing medium followed by the uptake assay (B). The cells were preincubated for 5 min in the presence of MTSET (200 M MTSET for C74A, N77C, and A81C; 50 M MTSET for G78C and G79C) in media containing choline chloride (ChCl), NaCl, NaCl plus 1 mM GABA, and sodium gluconate (Nagluc) followed by the uptake assay (C). The results are expressed as percent of GAT-1-C74A activity (A) or as the percent of activity after preincubation with MTSET relative to that of cells preincubated similarly without MTSET (B and C) and are given as the means Ϯ S.E. of at least three separate experiments performed in triplicate. The mean values of the cysteine mutants were compared with C74A (B) using a one way analysis of variance with a post hoc Dunnett multiple comparison test with p Ͻ 0.01 taken as significant (*). The mean values for the percent of control activity under the various preincubation conditions for each mutant were compared similarly to the values obtained after preincubation in choline chloride medium. 5-min preincubation with a 0.2 mM concentration of impermeant positively charged sulfhydryl reagent MTSET; in the case of G78C/C74A, G79C/C74A, and A81C/C74A, [ 3 H]GABA uptake was inhibited, whereas that by N77C/C74A was stimulated by the reagent (Fig. 1B). Similar results with the four mutants were obtained upon preincubation with a 1 mM concentration of the negatively charged (2-sulfonatoethyl)methanethiosulfonate (data not shown). Even though the mutants are in the background of C74A, one possibility that explains the effects by MTSET is that one or more of the mutants might expose a previously inaccessible endogenous cysteine residue. However, the fact that introduction of an alanine or serine residue at these positions did not render the transporters sensitive to MTSET (Fig. 1B) strongly suggests that the introduced cysteines themselves are modified by the sulfhydryl reagent.
To assess if the reactivity of N77C, G78C, G79C, and A81C is influenced by the conformation of the transporter, the effect of the three substrates on the sensitivity of the mutants to MTSET was determined. In the case of N77C/C74A the stimulation by MTSET in sodium-containing medium was less than that in its absence. In the presence of sodium, GABA did not have much of an effect, but removal of chloride resulted in a diminished stimulation (Fig. 1C). In G78C/C74A there was not much of an effect of the composition of the external medium on sensitivity to the sulfhydryl reagent, but both in G79C/C74A and A81C/C74A there was a significant protection by sodium, and this protection was diminished either in the presence of GABA or when chloride was removed (Fig. 1C).
Effect of GABA on Sodium Transients-Because the lack of [ 3 H]GABA transport of G80C/C74A and F82C/C74A is not due to defective targeting to the plasma membrane, it is possible that one or both of the mutants have a defect in one of the steps of the transport cycle but still may be able to execute some of the other steps. This possibility was addressed by analyzing the ability of the mutant transporters upon expression in X. laevis oocytes to carry out reactions linked to the transport cycle. One of those is the capacitative sodium-dependent transient current that is thought to represent a charge-moving conformational change subsequent to sodium binding to GAT-1 (Refs. 13, 16, 17, and 19 and Fig. 2). When the membrane potential of oocytes expressing wild type GAT-1 was jumped to a range of voltages from Ϫ140 to ϩ60 mV (for 500 ms at each voltage) from a holding potential of Ϫ25 mV, transient currents were observed in sodium-containing medium ( Fig. 2A) but not in sodium-free medium (Fig. 2B). These sodium-dependent transients were inward when the jumps were to potentials more negative than the holding potential and outward when the jumps were to potentials more positive than Ϫ25 mV. When the voltage was returned to the holding potential, transients of identical magnitude but opposite sign were observed. These transients are many orders of magnitude slower than expected for the rate of diffusion-limited sodium binding. The transients are rather a readout of sodium binding to the transporters and are thought to represent sodium-dependent charge movements by the transporter through the membrane electric field during the "on" steps and back during the "off" steps (13,16,17,19). It was also shown that when external sodium concentrations were progressively lowered, the contribution of the inward charge movements during jumps from the same holding potential became more and more pronounced, as if at the holding potential less and less sodium was bound to the transporter. Thus at lower extracellular sodium, more negative membrane potentials are required for the transporters to bind sodium. As expected, when after prior sodium removal the oocytes expressing wild type GAT-1 were perfused again with this cation, the transient currents returned (Fig. 2C). The addition of nontransportable solutes, such as L-aspartate, did not have an effect on the transients (Fig. 2D). As observed previously (13,16,17,19), the addition of GABA resulted in an almost complete disappearance of the transient current, which now became a maintained steady-state current (Fig. 2E). This is because, rather than sodium-dependent charge movements into and out of the membrane electric field, now sodium and GABA are transported across the membrane in an electrogenic manner. After GABA was removed, the transients were observed again (Fig. 2F).
Although no transient currents were observed with the F82C/C74A mutant, the G80C/C74A transporters did exhibit sodium-dependent transients, but in contrast to the wild type, in 96 mM external sodium they were only observed when the voltage was jumped to positive potentials from a holding potential of Ϫ25mV ( Figs. 2A and 3A) as if at the holding potential all the mutant transporters were in the sodium-bound state. Indeed lowering the sodium concentration resulted in the transients becoming more and more symmetrical and eventually inward (data not shown, but see Fig. 4). These transients reflect capacitative charge movements because upon return to Ϫ25 mV, transients of similar size but opposite sign were observed (Fig. 3A). In the absence of sodium (choline replacement) no transients by G80C were observed (Fig. 3B). Remarkably, when after extensive perfusion with the sodium-free medium, the G80C/C74A oocytes were perfused again with sodium Ringer for as long as 10 min, almost no transients were detected (Fig. 3C), and this remained the same even after the perfusion with sodium was continued for an hour (data not shown). Similar results were obtained when sodium was re- moved by perfusion with N-methyl-D-glucamine (data not shown). Interestingly, exposure to 1 mM GABA caused a marked and immediate recovery of the transients (Fig. 3E). Half-maximal recovery of the transients was obtained at around 100 M GABA. This effect of GABA was specific because 1 mM L-aspartate was not able to bring about any recovery of the transients (Fig. 3D). Subsequent removal of GABA caused a further increase of the transients (Fig. 3F). The recovery of the transients upon exposure to GABA was completely dependent on the simultaneous presence of sodium; after exposure to GABA in lithium-or choline-containing media followed by washout of the GABA, the sodium-dependent transients did not recover (data not shown). On the other hand the recovery of the transients by GABA was also observed in the absence of chloride (gluconate substitution, data not shown). Observations similar to those on G80C/C74A were also made when glycine 80 was replaced by alanine and, albeit to a much lesser extent, when the replacement was by serine or threonine. When gly-cine 80 was replaced by proline or aspartate, no sodium-dependent transients were observed regardless of the presence of GABA (data not shown). This was not due to a targeting defect, at least not in the G80P/C74A mutant, as evidenced by surfacebiotinylation experiments (data not shown). The phenomenon of poor recovery of the transients after removal of sodium was neither observed with G78C/C74A nor with G79C/C74A (data not shown).
It is important to see the full extent of the transients to quantify the phenomenon. However, even at ϩ60 mV the full extent of the transients was not yet observed since their size was still increasing as a function of the voltage (Fig. 3). It is difficult to determine the full extent of the transients of G80C/ C74A at 96 mM sodium because at the holding potential all the transporters are in the sodium-bound form, and jumping to potentials more positive than ϩ100 mV is required to obtain the maximal charge movement (data not shown). Such very positive potentials were not well tolerated by most batches of oocytes. At reduced sodium concentrations, less positive potentials should be required to cause dissociation of sodium from the transporters. Therefore, the effect of GABA was determined in a solution containing only 10 mM sodium. Before perfusion with the 10 mM sodium-containing solution, the oocytes were perfused first with a sodium-free solution. The results are given as non-subtracted records to show more clearly the opposite effect of GABA on the transient currents by C74A (Fig. 4, A and B) and the mutants (Fig. 4, C-F). In 10 mM sodium, transients in the G80C/C74A and G80A/C74A mutants were almost undetectable (Fig. 4, C and E), but exposure to 1 mM GABA resulted in the induction of the transients (Fig. 4, D and F). As predicted, the transients by G80C/C74A at 10 mM sodium were symmetrical, in contrast to the outward transients measured at 96 mM sodium (Fig. 2). The size of the transients saturated in the voltage range used (Fig. 4, D and  F). Again, the addition of 1 mM L-aspartate did not cause the induction of any transients (data not shown). In the absence of GABA reintroduction of sodium at 10 mM to oocytes that expressed the parent construct C74A resulted in the appearance of transients (Fig. 4A). Because of the reduced sodium concentration, most of the charge movement was observed when the voltage was jumped to negative potentials (compare with Fig. 2 where 96 mM sodium is used). The addition of GABA (Fig. 4B, see also Fig. 2E) resulted in an almost total disappearance of the transient current, which now became a maintained steadystate current because of coupled electrogenic sodium and GABA translocation (13). These GABA-induced currents are apparent when comparing the steady-state currents at negative potentials in Fig. 4B with those in Fig. 4A. For instance, in this experiment the GABA-induced current at Ϫ140 mV was Ϫ279.5 nA. In the case of other mutants at position 80, also in the presence of 10 mM sodium small transients could be induced in the serine and threonine replacement mutants by GABA but not in those where glycine 80 was replaced with proline or aspartate (data not shown). The induction of the transient currents in G80A/C74A and G80C/C74A was dependent on the GABA concentration. In the case of G80A/C74A, the charge movements were maximal at 1 mM. At 10 mM GABA a small but significant decrease of the Q max was observed, 87 Ϯ 3% that at 1 mM (n ϭ 6), apparently due to the emergence of GABA-induced steady-state currents (see also Fig. 6). At 0.1 mM GABA, the Q max was 40 Ϯ 4% that at 1 mM GABA (n ϭ 6). In the case of G80C/C74A, the Q max at 0.1 mM GABA was only 9 Ϯ 4% that at 1 mM, whereas at 10 mM its size increased to 216 Ϯ 25% that at 1 mM GABA (n ϭ 5). This indicates that the G80C/C74A mutant has an ϳ10fold lower apparent affinity for GABA than G80A/C74A.
Our observations on the dependence of the transient currents on the sodium concentration and voltage indicate that the glycine mutants have an increased apparent sodium affinity. In the voltage range used there was no saturation of the charge movements (representing sodium dissociation) of G80C/C74A in response to depolarization at 96 mM sodium (Fig. 2), and the same was true for sodium binding to C74A in response to hyperpolarization at 10 mM sodium (Fig. 4A). To directly compare the Q/V relationship of G80C/C74A with that of C74A we employed an intermediate sodium concentration (20 mM , Fig.  5). The curve of the mutant was right-shifted by around 80 mV relative to that of the control; V 1 ⁄2 for C74A was Ϫ86.6 Ϯ 2.3 mV (n ϭ 10), Ϫ8.4 Ϯ 2.6 mV for G80C/C74A (n ϭ 4), and Ϫ11.7 Ϯ 1.7 mV for G80A/C74A (n ϭ 3). Importantly, the Q/V relationship of the G80C/C74A mutant after the removal of sodium and its subsequent re-addition was similar to that before sodium removal (Fig. 5A). The same was true after addition and subsequent removal of L-aspartate and GABA (Fig. 5A). Similar observations were made with G80A/C74A (data not shown). The charge movements in 20 mM sodium were compared under four consecutive conditions, 1) before sodium removal, 2) after sodium removal by perfusion with sodium-free solution, 3) after prior perfusion with 20 mM sodium and 1 mM L-aspartate, and 4) after prior perfusion with 20 mM sodium and 1 mM GABA. For each oocyte the size of the charge movements from each condition was normalized to that of condition 1, and the data are presented in Fig. 5, B and C, represent the average normalized values for each condition of all oocytes tested. The data indicate that GABA primarily influenced the number of the G80C/C74A transporters capable to carry out the charge-moving conformational change (Fig. 5B). Interestingly, at 20 mM sodium, also the recovery of the transients by C74A after sodium removal was not entirely complete, namely 76 Ϯ 5% of the Q max before sodium removal (n ϭ 10) (Fig. 5C). Remarkably, after exposure to GABA and its subsequent removal, the transients recovered to their original size, 105 Ϯ 4% (n ϭ 10) of the Q max before sodium removal (Fig. 5C). At 96 mM sodium the phenomenon was less pronounced, and the corresponding values were 91 Ϯ 2 and 106 Ϯ 4%, respectively (n ϭ 7). The data collected in this figure allow us to estimate the expression levels of the G80C/C74A mutant and the wild type. Before sodium removal the Q max for the mutant in 20 mM sodium was 5.74 Ϯ 1.2 nC (n ϭ 4). However, GABA stimulated the charge movements by around 3.5-fold (Fig. 5B) so that under optimal conditions Q max of the mutant was around 20 nC, which is compared with a Q max of 24.00 Ϯ 0.25 nC (n ϭ 10) for C74A. Steady-state GABA Currents-In the 96 mM sodium-containing medium, small steady state currents induced by 1 mM GABA were observed with G80A/C74A (Fig. 6A). At Ϫ140 mV these were Ϫ80.6 Ϯ 5.7 nA (n ϭ 5) for the mutant and Ϫ502 Ϯ 38 nA (n ϭ 5) for C74A. In the case of the G80A/C74A mutant, when GABA-induced steady-state currents were measured after sodium was removed and subsequently was added back, they were found to be 102.3 Ϯ 3.7% (n ϭ 5) at Ϫ140 mV of the corresponding currents monitored before sodium removal. Uptake of [ 3 H]GABA under voltage clamp at Ϫ 80 mV showed charge/flux ratios of 1.62 Ϯ 0.09 for C74A (n ϭ 5) and 1.56 Ϯ 0.13 for G80A/C74A (n ϭ 4). These values are in between the proposed charge/flux ratios of 1 (11)(12)(13)(14) and 2 (15) for GABA transport, but it is obvious that the mutation does not cause a change in the stoichiometry of coupled GABA transport. The apparent affinity for GABA in G80A/C74A was markedly reduced relative to C74A (Fig. 6A). In the case of G80C/C74A, inward currents were almost undetectable at 1 mM GABA, but at 10 mM very small inward currents were observed (data not shown). Again, this indicates that the apparent GABA affinity of G80C/C74A is substantially lower than that of G80A/C74A. In the case of G80S/C74A and G80T/C74A no GABA-induced currents could be observed even at 10 mM. The sodium concentration dependence of the GABA-induced currents by G80A/ C74A was shifted toward much lower concentrations than in C74A. At 2 mM sodium the currents induced by 10 mM GABA in G80A/C74A were 50.3 Ϯ 4.7% that of those at 96 mM sodium at Ϫ140 mV, whereas in C74A the corresponding value was 7.2 Ϯ 3.0% (n ϭ 3). The GABA-induced current by C74A increased at more negative voltages, possibly due in part to an increased binding of sodium to the transporter. Consistent with this idea is that this voltage dependence in the mutant, which has a much higher apparent affinity for sodium, is much more shallow (Fig. 6B). The percentage of the size of the GABA-induced currents at ϩ60 mV relative to those at Ϫ140 mV was 56.1 Ϯ 4.5% for G80A/C74A (n ϭ 5) and 5.4 Ϯ 1.6% for C74A (n ϭ 5), respectively. The size of the currents at Ϫ140 mV was Ϫ145 Ϯ 12 nA for G80A/C74A and Ϫ490 Ϯ 36 nA for C74A (n ϭ 5).
Lithium Leak Currents-The lithium-leak current is another transporter-associated current that in wild type does not require GABA (16,20). In G80A/C74A and G80C/C74A, these currents were almost undetectable in the absence of GABA, Ϫ23 Ϯ 9 nA and Ϫ20 Ϯ 5 nA respectively, at Ϫ140 mV (n ϭ 3). In contrast to C74A, where GABA did not have an effect on the lithium-leak currents (Fig. 7A), in the G80A/C74A and G80C/ C74A mutants, robust currents were only observed when 10 mM GABA was present, Ϫ1921 Ϯ 403 nA and Ϫ417 Ϯ 54 nA at Ϫ140 mV, respectively (Fig. 7, B and C). The concentration of GABA where a half-maximal effect was observed was 2.5 Ϯ 0.5 and 4.0 Ϯ 0.5 mM for G80A/C74A and G80C/C74A, respectively (n ϭ 4). This effect by GABA was specific since it was not observed with 10 mM L-aspartate (Fig. 7, B and C). The non-transportable GABA analogue SKF 100330A also was not able to induce any currents (data not shown). Similar to the lithiumleak currents in wild type (16,20), these GABA-dependent currents in the mutants were potently inhibited by sodium, 72.4 Ϯ 1.6 and 63.6 Ϯ 0.8% inhibition by 1 mM sodium for G80A/C74A and G80C/C74A, respectively, at Ϫ140 mV (n ϭ 3). Measurements of charge/flux ratios at Ϫ80 mV showed that no [ 3 H]GABA uptake took place in lithium-containing media (data not shown).
Even though in G80C/C74A sensitivity of [ 3 H]GABA transport to MTSET could not be determined because of the lack of activity (Fig. 1A), we could assess the impact of the sulfhydryl reagent on the GABA-dependent lithium-leak currents and sodium-transient currents. Neither of these currents was sensitive to 1 mM MTSET regardless of if the preincubation was performed in choline or sodium containing media with or without 1 mM GABA (data not shown).

Effect of Depolarization on Transient and Leak Currents-
The data presented thus far indicate that GABA is able to correct defective conformational changes in the glycine 80 mutants. Because the binding order of sodium and GABA could be random, it is possible that GABA exerts its effect by bypassing the affected step in the transport cycle. We considered the possibility that the membrane potential may perhaps directly influence this affected step. During our voltage-jump protocols the oocytes were held at varying membrane potentials for 500 ms. After perfusion with a choline-containing medium, oocytes expressing G80A/C74A and G80C/C74A were exposed to a medium containing 10 mM sodium. As documented in Fig. 4, no transients were observed under these conditions unless GABA was added. Before monitoring the transient currents at the holding potential of Ϫ25 mV, the oocytes were held clamped for 2 min at various potentials beginning at Ϫ85 mV. At positive potentials, the transients recovered in the absence of GABA (Fig. 8). The size of the recovery was time-dependent, reaching a maximum at 2 min, and this time dependence was similar at ϩ35 and ϩ65 mV (data not shown). The size of the transients by C74A after 96 mM sodium was added back to the oocytes and was increased by the depolarization pre-pulse by 35 Ϯ 7% (n ϭ 3); moreover, when the oocytes were held in the lithium-containing media for 2 min at ϩ65 mV, leak currents could subsequently be monitored in the absence of GABA; Ϫ549 Ϯ 66 nA for G80A/C74A (n ϭ 3) and Ϫ374 Ϯ 76 nA for G80C/C74A (n ϭ 4), both at Ϫ140mV. When, after the depolarizing pre-pulse the leak currents were monitored at several time points after the return to the holding potential (Ϫ25 mV), a progressive decay was observed with a half-time of ϳ45 s (data not shown). The same was true for the transients but only when they were monitored at low sodium concentrations; at 10 mM sodium the half-time of decay was around 60 s, but at 96 mM sodium no measurable decay was observed even after keeping the oocytes at the holding potential for 4 min (data not shown). As will be FIG. 6. GABA induced currents. A, currents induced by the indicated concentrations of GABA, obtained by subtraction of the currents in the absence of GABA, were measured in ND medium with oocytes expressing GAT-1-C74A, G80A/ C74A, and G80C/C74A. The currents at Ϫ140 mV are normalized to those induced by 1000 M GABA with GAT-1-C74A. B, I-V relationships of the currents induced 10 mM GABA in ND. Currents are normalized to those by G80A/C74A at Ϫ140 mV. The data means Ϯ S.E. are averages from five oocytes from three different batches.
shown under "Discussion," these observations fit very nicely with the model to be presented therein. DISCUSSION In this study we have shown that mutation of glycine 80 of GAT-1, a residue conserved throughout the SLC6 transporter family, results in a remarkable phenotype. Unlike the wild type, the sodium-dependent transient currents of G80C (Figs. 3  and 4) and G80A (Fig. 4) do not recover when, after sodium removal, this cation is added back. However, the addition of GABA corrects this defect (Figs. 3E and 4, D and F), and the ability to carry out these transients persists after removal of GABA (Fig. 3F). Sodium removal is also a prerequisite for the measurement of the lithium leak current because sodium is a potent inhibitor of this current (16,20). Indeed, in the G80C and G80A mutants the lithium leak currents are almost undetectable, but also here they emerge in the presence of GABA (Fig. 7). Thus, it appears that after removal of sodium, the mutant transporters are "frozen" in a state from which they cannot exit to undergo the conformational changes involved in the transient and leak currents. This idea is further underscored by the observation that not only GABA but also a prolonged depolarization (Fig. 8) can restore both activities, as if either of these treatments can release the mutant transporters from their frozen state.
The frozen state appears to be outward-facing, since extracellular sodium and GABA can bind to it and cause the mutant transporters to regain the ability to mediate the transients. However, there exists yet another outward-facing state of the mutant transporters that is reached after the depolarizing steps in the voltage jump protocol (Fig. 3A, no prior sodium removal). Jumping to positive potentials causes the dissociation of sodium from the transporter and the outward charge movement. When after 500 ms the potential is stepped back to the holding potential, sodium rebinds, as evidenced by the ensuing opposite charge movement (Fig. 3A). To distinguish between these two outward-facing states, the frozen state and the one able to bind sodium and undergo the charge movement, they will be referred to as T and T*, respectively. The T* state is believed to represent a genuine intermediate of the transport cycle (13,16,17,19). Importantly, T* also represents the leak mode of the transporter, but when sodium binds to it and is occluded, the leak pathway is inhibited (16,20). Regarding the T state we know that addition of sodium and GABA to G80A/ C74A, after prior sodium removal, results in steady-state transport currents of the same magnitude as those recorded before sodium removal. Thus, there must be a sodium-and GABA-dependent transition between T and the translocation complex even though the transition from T to T* is attenuated (Fig. 3). These features are incorporated into a model of which we present two variants, which both can explain the observations on the glycine 80 mutants (Fig. 9, A and B). These variants have the same basic features, and the only difference between them is that in Fig. 9A, the T state is reached by the transporter when the empty inward-facing transporter reori-  8. Effect of the pre-pulse potential on the sodium-dependent transients. Oocytes expressing G80A/C74A and G80C/C74A were re-perfused with the 10 mM sodium containing medium after prior removal of sodium by perfusion with choline chloride medium. After each depolarization pre-pulse (manually adjusted) for 2 min at the indicated potential (starting with the most negative voltage, Ϫ85 mV), the membrane potential was readjusted to Ϫ25 mV, and sodium-dependent transients were recorded as under "Experimental Procedure." The data are normalized to the Q max (integrated using Clampfit) at ϩ65 mV and are the averages Ϯ S.E. from 3 or 4 oocytes. ents and exposes its binding sites to the outside. In Fig. 9B, the T* state is reached upon that reorientation.
In the variant depicted in Fig. 9A, the outward-facing transporter has to undergo a conformational change to reach the T* conformation. Sodium binding to T* and its subsequent occlusion yields T* Na . Binding of GABA to T* Na enables the formation of the translocation complex T* Na,G and coupled transport (Ref. 25 and Fig. 9). If we assume that due to the mutation of glycine 80 the transition from T to T* is impeded much more than the reverse transition (dashed arrow in Fig. 9A), all our observations on the glycine mutants can be easily explained. Thus, when sodium is removed, the transporter goes from T* Na to T*, and in the continued absence of sodium T* it converts to T. In the wild type, T is readily able to return to T*, but in the mutant this step is dramatically slowed down (Fig. 9A). Because the T* state represents the leak mode and also the state to which sodium binds and gets occluded, the lithium leak currents (Fig. 7, B and C) as well as the sodium-dependent transients (Figs. 3C and 4, C and E) are impeded in the mutants after sodium removal. However, in the presence of sodium, GABA can bind to T (K 0.5 around 100 M), yielding the translocation complex T* NaG . In this state sodium is occluded, and outward transients can be observed due to the fact that the rate of forward transport is low, and therefore, relatively little of this form is dissipated in contrast to the wild type, where the transients are suppressed by GABA because of rapid sodiumcoupled translocation (Refs. 13, 16, and 18 and Fig. 4, A and B). Upon washout of GABA, GABA dissociates from T* NaG to yield T* Na , and as a result the transients are even increased (Fig.  3F). When the mutant transporters are in the T state, GABA can bind to them even in the absence of sodium, albeit with a much lower apparent affinity (Fig. 7, B and C). This results, analogous to the T to T* Na,G transition (Fig. 9), in the formation of T* G state, which also is a leak mode.
Why do "native" G80A-or G80C-expressing oocytes exhibit the transient currents? These oocytes have been in a sodiumcontaining medium for 4 -5 days since the moment of injection of the cRNA encoding for the mutant transporters. The rate of T to T* transition is slow in the mutants but not zero, and due to the high apparent sodium affinity any mutant transporters that reach the T* state will be immediately trapped in the T* Na state. If the membrane potential is jumped to positive potentials, sodium dissociates from the transporter, again yielding T*. However because the jump is only for 500 ms, this it too short for the T* to T transition to occur, and when the potential is jumped back to the holding potential all the transporters revert from T* to T* Na . However, after a prolonged absence of sodium, the transporters do undergo the T* to T transition, and in this case the transients cannot recover without GABA or a long depolarization.
The model predicts that when in the absence of sodium the mutant transporters are brought in the T* state by the depolarizing pre-pulse, after return to Ϫ25 mV the transporters will revert from T* to T in a time-dependent process. This is expected to result in a decay of the lithium leak current, and this has indeed been observed. When the depolarizing pre-pulse is given in the presence of saturating sodium, the transporters will become trapped in the T* Na state and remain there because of their high apparent affinity for sodium. Indeed, the transients do not decay when the oocytes are held at Ϫ25 mV in the standard sodium-containing medium for up to 4 min. In the presence of low sodium, part of the transporters are in T* and part in T* Na , and the decay of the transients is predicted to increase as sodium is lowered, exactly as we have observed.
According to the alternative variant (Fig. 9B), reorientation of the inward-facing binding sites of the transporter to the extracellular side yields the T* state. The explanation of all phenomena is exactly the same as above, as long as also here the transition from T to T* is impaired in the mutants while the reverse transition is not. In this variant entering the T state might be considered a side-reaction rather than part of the main translocation path. Nevertheless, sodium and GABA can bind to T to yield the translocation complex T* Na,G. Therefore, the T* to T transition would rather represent an alternative transport pathway, analogous to a random binding order of substrates in an enzymatic reaction. There is one phenomenon that is more readily explained by the model in Fig. 9B. When sodium is removed and subsequently added back, residual transients are observed in the mutant (Fig. 3C), and these residual transients persist even when sodium removal continues for up to 1 h. According to Fig. 9B the transporters transit from T* to T under these conditions, but it is possible that some move to the inward-facing conformation instead. When sodium is replenished, those transporters emerging from the inwardfacing conformation to T* may bind the cation (resulting in T* Na ), which can explain the residual transients observed in Fig. 3C.
It is important to note that the unusual behavior of the glycine mutant occurs to some extent also in the wild type, FIG. 9. Conformational transitions of the outward-facing GAT-1 transporter. Shown is an abbreviated diagram of the transport cycle (clockwise for influx) and the two outward-facing conformations of the transporter T and T* as described in "Discussion." The only difference between A and B is that the unloaded transporter reorients to the outside to yield T (A) or T* (B). We assume that in the glycine 80 mutants, the transition from T to T* (dashed arrow) is impaired, and after removal of external sodium from oocytes expressing G80C and G80A the mutant transporters are stuck in state T and can only slowly return to T*. The T to T* transition by the mutants is speeded up by depolarization. GABA binding to the mutants enables the transition from T to T* G in the absence of sodium (not shown) and from T to the translocation complex T* NaG in the presence of sodium. T* and T* G mediate the lithium leak currents. Transitions from T* to T* Na and from T to T* NaG are accompanied by a charge-moving conformational change. After coupled transport, GABA and the co-ions are released to the inside (not shown), giving rise to the inward-facing unloaded transporter (not shown), which reorients to T (A) or T* (B). For the sake of simplicity we have omitted the second sodium ion and the chloride ion, which also participate in GABA translocation. especially when monitored in low sodium (Fig. 5C). This makes sense in terms of the model, regardless of the variant considered. Low sodium promotes the transition from T* Na to T*, and therefore in low sodium there is an increased probability that the T state is reached. In low sodium the effect of GABA on transients, which are mediated by the wild type, immediately after GABA removal is less than in the mutants since in the wild type the transition from T to T* is not impaired. The observation that after sodium was added back to oocytes expressing wild type a depolarization pre-pulse resulted in increased transients also suggests that the mutations at glycine 80 enable us to observe more clearly a phenomenon that already exists in the wild type.
It is interesting that proline cannot substitute for glycine 80. Even though proline introduces kinks in ␣-helices, it seems that it is rather the flexibility afforded by glycine that is important. Glycine residues introduced into lactose permease can confer conformational flexibility to this transporter (10), and glycine has been shown to act as a gating-hinge in potassium channels (9). Glycine 80 is the third of three conserved glycines throughout the SLC6 transporter family. Remarkably, only it behaves as if it is required for conformational changes, indicating the uniqueness of this position. It seems likely that also in other transporter families some of the conserved glycine residues may fulfill a role similar to that of glycine 80 in GAT-1.