Functional Regulation of γ-Aminobutyric Acid Transporters by Direct Tyrosine Phosphorylation*

Tyrosine phosphorylation regulates multiple cell signaling pathways and functionally modulates a number of ion channels and receptors. Neurotransmitter transporters, which act to clear transmitter from the synaptic cleft, are regulated by multiple second messenger pathways that exert their effects, at least in part, by causing a redistribution of the transporter protein to or from the cell surface. To test the hypothesis that tyrosine phosphorylation affects transporter function and to determine its mechanism of action, we examined the regulation of the rat brain γ-aminobutyric acid (GABA) transporter GAT1 expressed endogenously in hippocampal neurons and expressed heterologously in Chinese hamster ovary cells. Inhibitors of tyrosine kinases decreased GABA uptake; inhibitors of tyrosine phosphatases increased GABA uptake. The decrease in uptake seen with tyrosine kinase inhibitors was correlated with a decrease in tyrosine phosphorylation of GAT1 and resulted in a redistribution of the transporter from the cell surface to intracellular locations. A mutant GAT1 construct that was refractory to tyrosine phosphorylation could not be regulated by tyrosine kinase inhibitors. Activators of protein kinase C, which are known to cause a redistribution of GAT1 from the cell surface, were additive to the effects of tyrosine kinase inhibitors suggesting that multiple signaling pathways control transporter redistribution. Application of brain-derived neurotrophic factor, which activates receptor tyrosine kinases, up-regulated GAT1 function suggesting one potential trigger for the cellular regulation of GAT1 signaling by tyrosine phosphorylation. These data support the hypothesis that transporter expression and function is controlled by the interplay of multiple cell signaling cascades.

Tyrosine phosphorylation regulates multiple cell signaling pathways and functionally modulates a number of ion channels and receptors. Neurotransmitter transporters, which act to clear transmitter from the synaptic cleft, are regulated by multiple second messenger pathways that exert their effects, at least in part, by causing a redistribution of the transporter protein to or from the cell surface. To test the hypothesis that tyrosine phosphorylation affects transporter function and to determine its mechanism of action, we examined the regulation of the rat brain ␥-aminobutyric acid (GABA) transporter GAT1 expressed endogenously in hippocampal neurons and expressed heterologously in Chinese hamster ovary cells. Inhibitors of tyrosine kinases decreased GABA uptake; inhibitors of tyrosine phosphatases increased GABA uptake. The decrease in uptake seen with tyrosine kinase inhibitors was correlated with a decrease in tyrosine phosphorylation of GAT1 and resulted in a redistribution of the transporter from the cell surface to intracellular locations. A mutant GAT1 construct that was refractory to tyrosine phosphorylation could not be regulated by tyrosine kinase inhibitors. Activators of protein kinase C, which are known to cause a redistribution of GAT1 from the cell surface, were additive to the effects of tyrosine kinase inhibitors suggesting that multiple signaling pathways control transporter redistribution. Application of brain-derived neurotrophic factor, which activates receptor tyrosine kinases, up-regulated GAT1 function suggesting one potential trigger for the cellular regulation of GAT1 signaling by tyrosine phosphorylation. These data support the hypothesis that transporter expression and function is controlled by the interplay of multiple cell signaling cascades.
␥-Aminobutyric acid (GABA) 1 is the predominant inhibitory neurotransmitter in the central nervous system and plays a pivotal role in the balance between neuronal excitation and inhibition. The time course of neurotransmitter in the synaptic cleft is one determinant of synaptic transmission, and its re-moval from the cleft via specific transporters provides an efficient mechanism for termination of signaling. GABA transporters (GATs) are located on the plasma membrane of neurons and glia (1) and function to remove GABA through co-transport of ions down their electrochemical gradient (2). Demonstration of a physiological role for GABA transporters comes from experiments involving specific GABA uptake inhibitors; these inhibitors affect both GABA A and GABA B receptor-mediated synaptic signaling (3)(4)(5). GABA transporters also play a pathophysiological role in temporal lobe epilepsy (6) and are the targets of pharmacological interventions in epilepsy (7). An important property of GABA transporters, and neurotransmitter transporters in general, is their ability to be functionally regulated by a wide variety of signaling cascades (for review, see Refs. 8 and 9). Functional modulation occurs in part through second messengers such as kinases, phosphatases, arachidonic acid, and pH. These factors may act directly on the transporter protein (e.g. by phosphorylation; see Refs. 10 -13) or by regulating the interaction of the transporter with other synaptic proteins, such as syntaxin (14). A recurring theme is that the regulation occurs through changes in the number of functional surface transporters (15)(16)(17).
The role of tyrosine phosphorylation in regulating GABA transporter function has yet to be elucidated, although there is reasonable evidence to suggest that this form of regulation of the transporter might occur: (i) protein phosphorylation of tyrosine residues acutely regulates neurotransmitter receptors (18) and ion channels (19 -22); (ii) the primary amino acid sequence of the rat brain GABA transporter GAT1 contains five putative intracellular tyrosine phosphorylation sites (23); and (iii) pharmacological manipulation of tyrosine kinases acutely regulates serotonin and dopamine transporters (24,25), although the mechanism underlying the modulation is not known. Long term regulation of serotonin transport by tyrosine kinases may be because of alterations in transporter mRNA levels (26). In the present report, we show that GAT1 is acutely regulated by direct tyrosine phosphorylation, that tyrosine phosphorylation acts to increase the expression of the transporter on the plasma membrane, and that a potential physiological trigger of this signaling cascade is receptor tyrosine kinases activated by brain-derived neurotrophic factor.

EXPERIMENTAL PROCEDURES
Cell Culture-Primary hippocampal cultures were prepared from postnatal day 0 -3 rats by mincing tissue in ␣ minimal essential medium supplemented with cysteine, glucose, and 100 units of Papain (Sigma or Worthington Biochemical). Tissue was incubated for 20 min at 37°C followed by gentle trituration, dilution, and plating onto poly-L-lysine-coated glass coverslips. To obtain pure neuronal cultures, mixed cultures were treated for 48 h with 10 M cytosine arabinoside (Sigma); treatment was initiated 24 h after plating. Cells were plated onto untreated 24-well plates and maintained in Earle's minimal essential medium supplemented with 10% fetal bovine serum. 1F9 cells (CHO cells stably expressing GAT1; 27) were maintained in ␣ minimal essential medium supplemented with 5% fetal bovine serum, L-gluta-mine, and penicillin-streptomycin. Transfections were carried out using LipofectAMINE (Life Technologies, Inc.) in Optimem I (Life Technologies, Inc.). The lipid-DNA mixture was incubated with the cells for 5 h; cells were then rinsed and refed with complete medium. Stable transformants were obtained by selection in 500 ng/ml G418 (Life Technologies, Inc.).
[ 3 H]GABA Uptake Assays-Pre-assay drug incubations (5-30 min) were performed in HBSS. Following preincubation, cells were rinsed three times in 1ϫ HBSS and allowed to equilibrate for 10 min in the final wash. Buffer was then exchanged with control HBSS or drugcontaining HBSS. GABA was added to initiate the assay. The final [ 3 H]GABA concentration of the assay solution was 40 nM; the total GABA concentration of the assay solution was 30 M. The assay was terminated by rapidly rinsing the cells three times with 1ϫ HBSS followed by solubilization in 300 l of 0.001-0.005% SDS at 37°C for 2 h. Aliquots were used for scintillation counting and to determine protein concentrations. Statistical analyses of the uptake data were performed using SPSS. Two-sample comparisons were made using t tests; multiple comparisons were made using one-way analysis of variances (ANOVAs) followed by Tukey's honestly significant difference post-hoc test.
Immunoprecipitations and Immunoblotting-Following preincubation with the appropriate drug, cells were lysed in buffer (9.1 mM dibasic sodium phosphate, 1.7 mM monobasic sodium phosphate, pH 7.4, 150 mM NaCl, 0.5% sodium deoxycholate, 0.1% SDS, 250 M phenylmethylsulfonyl fluoride, 1 g/ml aprotinin, 1.0 mM activated sodium orthovanadate, 5.0 mM sodium pyrophosphate) for 1 h at 4°C. Lysates were precleared with 10 l of protein G-agarose conjugate (Sigma) followed by immunoprecipitation of GAT1 using anti-GAT1 antibody 346J (28) and protein G-agarose. The product was washed in radioimmune precipitation buffer and run on a 10% acrylamide gel. Protein was transferred to a nitrocellulose membrane by electroblotting. Western blotting was performed using the antiphosphotyrosine antibody PY99 (Santa Cruz Biotechnology) and visualized using ECL reagents (Amersham Pharmacia Biotech).

RESULTS
To initially determine whether tyrosine phosphorylation of GAT1 may play a role in the functional regulation of GABA transport, we treated neuronal cultures of dissociated rat hippocampus with drugs that affect tyrosine kinases and phosphatases. These cultures were maintained in normal serum levels until the assay. Fig. 1A shows that application of genistein, a nonspecific tyrosine kinase inhibitor, decreased transport of GABA in subsequent radiolabeled GABA uptake assays. This reduction in uptake did not occur in the presence of daidzein, an inactive genistein analogue. Furthermore, the genistein-mediated reduction in transport could be prevented by co-application of the tyrosine phosphatase inhibitor pervanadate. The action of these drugs on GABA transport is likely mediated through the rat brain GABA transporter GAT1 because SKF89976A, a high affinity antagonist of GAT1, blocks a majority of GABA uptake in these cultures. Wash-out experiments indicated that the genistein effect was reversible (Fig.  1B). GABA uptake returned to 90% of control levels 30 min following a 30-min treatment of 10 M genistein.
At the concentrations of genistein used (IC 50 ϳ1 M; Fig. 1C), there are multiple signaling pathways potentially affected that could mediate the reduction in GABA transport. Therefore, we also examined GABA transport in dissociated hippocampal neurons using the compound K252a. At low nanomolar concentrations, this drug is a selective inhibitor of tyrosine kinases (29). Concentration-response measurements revealed that K252a inhibited GABA transport with an IC 50 of ϳ3 nM.
The reversal of genistein-mediated GABA transport by the phosphatase inhibitor pervanadate raised the possibility that the amount of GABA transport is determined by a balance of tyrosine kinase and phosphatase activity. To determine whether GABA transport was already maximally activated in the control state, we incubated cultures with saturating concentrations of pervanadate alone (100 M) or in the presence of increasing concentrations of genistein. These results are shown in Fig. 1D. Pervanadate alone increased GABA transport; this increase was blocked by genistein. These data are consistent with the hypothesis that the GABA transporter in hippocampal neurons is regulated by a balance of tyrosine phosphatase and kinase activities. These functional changes in GABA transport are not likely because of nonspecific effects of these compounds (e.g. changes in membrane potential), because glutamate uptake measured in parallel neuronal cultures was not affected (data not shown).
We next tested the hypothesis that the functional effects of tyrosine kinase inhibition were correlated with a change in the phosphorylation state of the transporter directly. Dissociated rat hippocampal neurons were treated with or without genistein, immunoprecipitated with a GAT1-specific antibody, and immunoblotted with an antiphosphotyrosine antibody. The results of this experiment are shown in Fig. 2A. Based upon three experiments, genistein caused an average 54% reduction in the amount of GAT1-specific tyrosine phosphorylation (Fig. 2B). These data correlate well with the functional effects of genistein on GABA transport in hippocampal neurons (see Fig. 1A). The specific kinase inhibitor K252a had a similar effect to genistein in reducing GAT1 phosphorylation levels; this reduction in GAT1 phosphorylation could be prevented by co-application of pervanadate (Fig. 2C).
Given the data from hippocampal neurons suggesting that regulation of transport is mediated through GAT1, an expression system was sought that would mimic the endogenous phenomenology and would permit a detailed characterization of the mechanisms underlying the regulation. Therefore, experiments were repeated in 1F9 cells, a mammalian cell line stably expressing GAT1 (27). The genistein-induced reduction in V max is consistent with changes in the number of functional transporters. To test this hypothesis directly, GAT1 immunoreactivity following biotinylation of surface proteins was examined in 1F9 cells preincubated in the absence or presence of genistein. These data are shown in Fig. 3B. As shown in the representative immunoblot, preincubation of cells with genistein caused a decrease in the amount of GAT1 immunoreactivity in the biotinylated fraction, the fraction corresponding to the surface population of transporters. This decrease in surface immunoreactivity was correlated with a increase in intracellular GAT1 immunoreactivity. Two control experiments support these findings. First, the intracellular cytoskeletal protein actin was not labeled by the biotinylation reagent, suggesting that only surface proteins were being labeled; and second, immunoreactive bands were not seen in untransfected CHO cells immunoblotted with the GAT1 antibody (data not shown). The surface biotinylation experiments were repeated two more times. The percent reduction of surface GAT1 protein in the presence of genistein was 38, 52, and 39% in these three experiments.
The data to this point demonstrate that inhibition of tyrosine kinase activity decreases GAT1 function and that this decrease is correlated with (i) a reduction in the amount of GAT1 tyrosine phosphorylation, (ii) a decrease in the transport capacity, and (iii) a loss of GAT1 immunoreactivity from the cell surface. We reasoned that if direct tyrosine phosphorylation of GAT1 was necessary for the effects of genistein, then removal of potential tyrosine phosphorylation sites on GAT1 should prevent its modulation by tyrosine kinases and phosphatases. Therefore, we created a GAT1 mutant construct lacking all five putative intracellular consensus tyrosine phosphorylation sites (23). This mutant was transiently expressed in CHO cells, and its biochemistry and function was compared with wild-type GAT1 also transiently expressed in these cells. Fig. 4A shows that immunoprecipitation of the wild-type GAT1 with a GAT1 antibody followed by immunolabeling with the phosphotyrosine antibody reveals an immunoreactive band of the appropriate molecular weight. However, this band is absent in cells transfected with the GAT1-5YA mutant. These results are not because of low expression levels of the mutant transporter because comparable expression levels (based upon uptake assays, see Fig. 4B) were chosen nor are the results due to an inability of the GAT1 antibody to recognize the mutant, because immunoblotting of total cell lysates with the GAT1 antibody revealed an immunoreactive band of the appropriate molecular weight (data not shown). Fig. 4B shows functional data indicating that neither genistein nor K252A could modulate GABA uptake in cells transfected with the GAT1-5YA mutant. Taken together, these data suggest that direct tyrosine phosphorylation of GAT1 is necessary for the functional effects of compounds that affect tyrosine kinase and phosphatase pathways. GAT1 is regulated by PKC (for review, see Ref. 9); activators of PKC cause a reduction in GABA transport that correlates with a decrease in the amount of GAT1 immunoreactivity on the cells surface. Thus, PKC activation and tyrosine kinase inhibition have comparable effects. This raises the possibility that the net internalization of GAT1 by both second messenger signaling pathways proceeds through a common redistribution mechanism. If this hypothesis is true, then the effects of compounds that act on each system, at saturating concentrations, should not be additive; however, if the effects are additive, it suggests that multiple independent mechanisms are acting to control the distribution and function of the transporter. To test this hypothesis, we incubated hippocampal neurons in compounds that inhibit components of tyrosine signaling cascade, in compounds that activate PKC, or both. The results of this experiment are shown in Fig. 5A. When applied alone at saturating concentrations, both the tyrosine kinase inhibitor K252A and the PKC-activating phorbol ester phorbol 12-myristate 13-acetate significantly reduced GABA transport. When applied together, these compounds significantly decreased GABA transport to a level below that seen with either compound alone. In addition, phorbol 12-myristate 13-acetate inhibited the increase in GABA transport mediated by the tyrosine phosphatase inhibitor pervanadate.
If the PKC-dependent and tyrosine phosphorylationdependent effects on GABA transport occur through separate mechanisms, then inhibition of one pathway should not effect inhibition by the other pathway. To test this hypothesis, we examined the effect of K252a on GABA uptake in the presence of the PKC antagonist bisindolylmaleimide II. These data are shown in Fig. 5B. Bisindolylmaleimide II reversed phorbol 12-myristate 13-acetate-induced inhibition of GABA uptake but had no effect on the inhibition mediated by K252a. Taken together, these data suggest that PKC-mediated changes in GABA transport and tyrosine kinase/phosphatasemediated changes in GABA transport occur through distinct mechanisms.
The ability of the tyrosine signaling cascade to alter GAT1 phosphorylation and function raises the question as to the physiological triggers capable of regulating this effect. One potential trigger of this pathway is the neurotrophin family of signaling molecules, which act through tyrosine kinase-containing receptors to modulate second messenger cascades. In hippocampus, receptors for brain-derived neurotrophic factor (BDNF) are found on both pyramidal cells and interneurons (30) and thus are a likely choice to mediate GABA transporter inhibition. To test this hypothesis, we treated hippocampal neurons with BDNF and examined the resulting effects on GABA transport. These results are shown in Fig. 6. In cells cultured in normal serum, BDNF failed to cause an increase in transport in subsequent GABA uptake assays; however, in cultures deprived of serum for 24 h, BDNF caused an increase in transport in subsequent GABA uptake assays. In addition, basal GABA transport levels were reduced following serum deprivation. Application of BDNF caused an increase in transport comparable to that seen in normal serum neurons treated with BDNF. The BDNF-mediated changes in GABA transport occurred in the presence of the PKC inhibitor bisindolylmaleimide II (data not shown), as expected if BDNF is exerting its effects through tyrosine kinase pathways and if the mechanism is distinct from the PKC-mediated effects on GABA transport (as shown in Fig. 5). These data suggest that BDNF can regulate transporter function and that it may also be responsible for contributing to basal transport levels. To verify that BDNF was causing the expected increase in GAT1 tyrosine phosphorylation, we examined serum-deprived neuronal cultures that were untreated, treated with BDNF alone, or treated with BDNF and K252a as in the functional experiments. Cultures were pharmacologically treated, immunoprecipitated with a GAT1-specific antibody, and immunoblotted with an antiphosphotyrosine antibody. The immunoblot shows that levels of GAT1 tyrosine phosphorylation are increased in the presence of BDNF; the increase in GAT1 phos- phorylation is prevented by co-application of K252a. DISCUSSION Homologous recombination experiments in which different neurotransmitter transporter genes have been manipulated demonstrate that the maintenance of appropriate levels of transmitter in the synaptic cleft by transporters is crucial for normal brain function (31,32). These results raise the likelihood that cells need to rapidly regulate transporter function to control these extracellular transmitter levels. In the present paper, we demonstrate that tyrosine kinases and phosphatases regulate the activity of a GABA transporter endogenously expressed in hippocampal neurons and heterologously expressed in mammalian cells. Correlated changes occur among the levels of tyrosine phosphorylation of the transporter, transporter expression on the cell surface, and GABA uptake. Removal of five putative intracellular tyrosine residues from the rat brain GABA transporter GAT1 eliminates tyrosine phosphorylation of the transporter and prevents modulation by tyrosine kinases and phosphatases, suggesting a direct action of this signaling cascade on the transporter. Preliminary data suggest that multiple intracellular tyrosine residues are responsible for the phosphorylation activity of GAT1. 2 Recent evidence for many different transporter systems suggests that multiple triggers (e.g. G protein-coupled receptors (33) and direct transmitter action on the transporter (34 -36)) and multiple mechanisms (e.g. protein-protein interactions (14) and PKC phosphorylation (12)) exist for the control of transporter function (for review, see Ref. 9). For the GABA transporter GAT1, PKC effects changes in function (37)(38)(39) and transporter redistribution (16,28); in hippocampal neurons, there is a decrease in surface transporter levels (33). Some of the PKC-mediated action occurs through altering the interaction of GAT1 with the plasma membrane protein syntaxin 1A (14). These PKC-mediated effects can be triggered by the stimulation of multiple G protein-coupled receptors (33). Other signal transduction mechanisms, such as GABA acting directly on the transporter (35) and now direct tyrosine phosphorylation, have been shown to increase surface GABA transporter expression. Thus, surface transporter expression and function will be the sum total of multiple signaling pathways acting separately and in concert to induce transporter redistribution.
Tyrosine kinase-mediated regulation of GAT1 is induced by BDNF, thus identifying a potential physiological trigger for GABA transporter regulation by protein-tyrosine kinases and phosphatases. BDNF activates tyrosine kinase-containing receptors to induce both acute (e.g. via increases in tyrosine kinase activity) and chronic (e.g. via increases in transcription) cellular changes. Because the effects on GAT1 function and redistribution reported here occur on a time scale of minutes, it is unlikely that the BDNF-mediated effects are related to its action on transcription. However, this does not rule out long term changes in GAT1 function and expression as a result of transcriptional regulation by BDNF. The ability of nanomolar amounts of the tyrosine kinase inhibitor K252a to regulate GAT1 function suggests that receptor tyrosine kinases, rather than soluble tyrosine kinases, are sufficient for GAT1 modulation (29).
Although pathways that involve tyrosine kinase activity have been shown to regulate redistribution of receptors (40), channels (41), and now transporters, the mechanism by which direct tyrosine phosphorylation induces GAT1 subcellular redistribution is not known. It is hypothesized that direct phosphorylation of the serotonin transporter by PKC serves as a tag that identifies transporters to be internalized (12). This might occur because the tag is indicative of a transporter in an appropriate conformational state for internalization. The evidence that serotonin transporter substrates prevent PKC-dependent phosphorylation and increase surface transporter expression supports a role for conformational changes in the transporter as being important for redistribution (36). Conformational changes in GAT1 are suggested to underlie the different rates of internalization that occur in the presence of transporter agonists and antagonists (34). Thus, it may be that tyrosine phosphorylation of GAT1 induces conformational states that, like transporter agonists, confer a relative slowing of transporter internalization.