Calcium- and Syntaxin 1-mediated Trafficking of the Neuronal Glycine Transporter GLYT2*

Previously we demonstrated the existence of a physical and functional interaction between the glycine transporters and the SNARE protein syntaxin 1. In the present report the physiological role of the syntaxin 1-glycine transporter 2 (GLYT2) interaction has been investigated by using a brain-derived preparation. Previous studies, focused on syntaxin 1-transporter interactions using overexpression systems, led to the postu-lation that syntaxin is somehow implicated in protein trafficking. Since syntaxin 1 is involved in exocytosis of neurotransmitter and also interacts with GLYT2, we stimulated exocytosis in synaptosomes and examined its effect on surface-expression and transport activity of GLYT2. We found that, under conditions that stimulate vesicular glycine release, GLYT2 is rapidly trafficked first toward the plasma membrane and then internalized. When the same experiments were performed with synaptosomes inactivated for syntaxin 1 by a pretreat-ment with the neurotoxin Bont/C, GLYT2 was unable to reach the plasma membrane but still was able to leave it. These results indicate the existence of a SNARE-medi-ated regulatory mechanism that controls the surface-expression of GLYT2. Syntaxin 1 is involved in the arrival to the plasma membrane but not in the retrieval. Furthermore, by using immunogold labeling on purified preparations from synaptosomes,

Exocytosis takes place after an increase in the intracellular calcium concentration and requires the fusion of synaptic vesicles, containing neurotransmitter, with the plasma membrane (1,2). Three membrane proteins are known to be involved in this process; synaptobrevin (VAMP), 1 SNAP-25, and syntaxin 1 (3)(4)(5). SNAP-25 and syntaxin 1 form a complex on the plasma membrane and function as a receptor for VAMP, present on synaptic vesicles. When VAMP forms a heteromeric complex with syntaxin 1 and SNAP-25, synaptic vesicles come in close contact with the plasma membrane and fusion between this membrane and the synaptic vesicle membrane may occur (6).
Glycine is an inhibitory neurotransmitter mainly present in brainstem and spinal cord, where it plays a role in a variety of motor and sensory functions. Apart from its inhibitory effect on glycinergic neurons, glycine modulates the action of glutamate on the N-methyl-D-aspartate receptors. After the release of glycine, high affinity reuptake mechanisms play an important role in the removal of glycine from the extracellular space, thereby terminating the neurotransmission signal. The uptake of glycine is accomplished by two high affinity glycine transporters, GLYT1 and GLYT2. Both transporters have been cloned (7)(8)(9)(10)(11) and have been the subject of many studies (12). The fine tuning of reuptake mechanisms are thought to control the duration of the neurotransmission signal and therefore the overall synaptic action (13).
Protein trafficking plays a key role in the regulation of many aspects of neuronal function, and certainly many data point to the fact that neurotransmitter transporters are also regulated by protein redistribution mechanisms (14,15). Recently we reported the interaction between the glycine transporters and syntaxin 1 (16). This interaction was found to be similar to those reported for the ␥-aminobutyric acid transporter GAT1 (17), the N-and Q-type calcium channels (18), and the cystic fibrosis transmembrane regulator (CFTR) chloride channel (19). All these studies have in common that syntaxin 1 overexpression decreased the currents in the channels or the transport rate in the transporters. The emerging view from these studies is that syntaxin 1 somehow interferes in protein trafficking (16,20). The fact that syntaxin 1 is also involved in the fusion of synaptic vesicles with the plasma membrane during the release of neurotransmitter made us study the relation between exocytosis and GLYT2 trafficking, both in the presence and absence of functional syntaxin 1.
Here we show that a treatment that induces glycine release in synaptosomes also stimulates the fast arrival of GLYT2 to the plasma membrane, which is immediately followed by an internalization of the transporter. The trafficking of GLYT2 toward and from the plasma membrane might be supported by its localization in synaptic-like vesicles. We furthermore demonstrate that the arrival, but not the retrieval, of GLYT2 from the plasma membrane is mediated by syntaxin 1.
Isolation of Synaptosomes and Synaptic Vesicles-Synaptosomes were purified from rat (3 months old, male) brainstem and spinal cord as described (22), and resuspended in ice-cold HBM buffer (Hepesbuffered medium, composition in mM: NaCl 140, KCl 5, MgCl 2 1, Na 2 HPO 4 1.2, NaHCO 3 5, glucose 10, and HEPES-NaOH 20, pH 7.4). This buffer was used as a Ca 2ϩ -free medium. The preparation was then incubated for 90 min at 37°C with or without neurotoxin (80 nM Bont/C or 80 nM Bont/B) and immediately used in experiments.
Synaptic vesicles were isolated from rat (3 months old, male) brainstem and spinal cord as described (23). Instead of the last purification step (controlled pore glass chromatography), a flotation density gradient was performed (24). In brief, crude synaptosomes (P2) were prepared by differential centrifugation (10 min at 800 ϫ g, followed by 15 min at 9200 ϫ g on the first supernatant) and lysed by a hypoosmotic shock. After differential centrifugation (20 min at 25,000 ϫ g, supernatant was taken and centrifuged for 2 h at 165,000 ϫ g), the obtained pellet was resuspended in 40 mM sucrose. This preparation is referred to as crude synaptic vesicles (lysate of P2). This fraction was then further purified by a continuous (50 -800 mM) sucrose gradient. After centrifugation (3 h at 65,000 g), the 200 -400 mM sucrose band was taken and 2 M sucrose was added to give a final concentration of 1 M. This sample was then transferred to a tube and subjected to a flotation gradient by adding the following amounts of sucrose on top of the sample: 6 ml of 800 mM, 6 ml of 600 mM, 6 ml of 380 mM, and 6 ml of 250 mM. After centrifugation (3 h at 27,000 ϫ g), the band at the 600 -380 mM sucrose interface was taken, and the membranes isolated by centrifugation (2 h at 165,000 ϫ g). Finally, the pellet was resuspended in 1 ml of glycine buffer (300 mM glycine, 5 mM Hepes-NaOH, pH 7.3). During the purification, three samples were taken for Western blot analysis corresponding to crude synaptosomes (P2), crude synaptic vesicles (LP2), and purified synaptic vesicles (SV).

Release of [ 3 H]Glycine from
Synaptosomes-A suspension of synaptosomes (0.5 mg/ml of protein) in HBM buffer was preloaded with 150 nM radiolabeled glycine (PerkinElmer Life Sciences) for 15 min. After centrifugation (3 min, 10,000 ϫ g), the synaptosomes were washed 3 times with HBM. Release of [ 3 H]glycine was then evoked at 37°C by the addition of 1 mM 4-aminopyridine and 5 mM CaCl 2 (final concentrations). After several time intervals (0, 0.5, 1, 2, and 3 min), an aliquot (0.1 mg of protein) was taken and immediately transferred to ice. Medium and synaptosomes were then separated by centrifugation (3 min, 10,000 ϫ g), followed by washing the synaptosomes three times with ice-cold HBM buffer. Synaptosomes were finally resuspended in 200 l of 0.2 M NaOH. The [ 3 H]glycine present in the medium and in the resuspended synaptosomes was quantified by liquid scintillation counting.
Uptake Assays-Transport assays were performed at 37°C in HBM buffer as described (25) with some modifications. Synaptosomes (0.5 mg/ml of protein) were incubated with 150 nM [ 3 H]glycine in a glass tube under continuous agitation. Assays were performed in the presence of 1 mM sarcosine to block glycine transport by GLYT1. After 0.5, 1, 2, and 3 min of incubation, a sample (0.1 mg of protein) from the tube was taken and immediately transferred to ice. After centrifugation (3 min, 10,000 ϫ g), the synaptosomes were washed three times with ice-cold HBM. Finally the synaptosomes were lysed in 0.2 M NaOH and the [ 3 H]glycine present counted on a liquid scintillation counter. In kinetic parameter determinations, glycine concentrations ranging from 150 nM to 100 M were used in a 1-min transport assay. All assays were done in triplicate. Protein concentrations were determined by the method of Bradford (26).
Biotinylation Studies-At time point 0, 4-aminopyridine and CaCl 2 (final concentrations of 1 and 5 mM, respectively) were added to a suspension of synaptosomes (0.5 mg/ml in HBM). Immediately, at the indicated time points, aliquots (0.1 mg of protein) were manually taken with a pipette and transferred to a tube containing 1 ml of ice-cold HBM to stop all trafficking events. Subsequently, samples were centrifuged (3 min, 10,000 ϫ g), and the synaptosomes washed once with ice-cold HBM and then resuspended in 1 ml of HBM containing the cell membrane impermeable reagent sulfo-N-succinimidyl-6-biotinamidohexanoate (2 mg/ml, Pierce), and an incubation for 1 h at 4°C was performed. Then, the excess of reagent was quenched by adding lysine (100 mM final concentration) and the incubation was continued for another 10 min. Samples were centrifuged (5 min, 10,000 ϫ g), and the pellets were washed twice with HBM and resuspended in lysis buffer (150 mM NaCl, 50 mM HEPES-Tris, 5 mM EDTA, 1% Triton X-100, 0.25% deoxycholate, 0.5% SDS, 0.4 mM phenylmethylsulfonyl fluoride, pH 7.4). The biotinylated proteins were then incubated for 1 h with streptavidinagarose beads and precipitated by centrifugation (Sigma). The beads were washed three times with 1 ml of lysis buffer, and the proteins were eluted from the beads with Laemmli sample buffer and subjected to Western blotting (25). Bands were visualized with the ECL detection method (Amersham Pharmacia Biotech) and quantified on a model 300A densitometer in combination with ImageQuant software (Molecular Dynamics, Sunnyvale, CA) by using film exposures that were in the linear range. Standard errors in biotinylation studies were calculated after densitometry from at least four separate experiments. To test the biotinylation procedure, samples from different synaptosomal preparations were taken and subjected to biotinylation. Streptavidinprecipitated proteins were subsequently analyzed by Western blotting with antibodies against intracellular proteins (e.g. syntaxin 1 and synaptophysin). We never found immunoreactivity against intracellular proteins in biotinylated fractions, which indicated that only plasma membrane proteins were biotinylated using the above described protocol.
Immunogold Labeling and Electron Microscopy-Purified synaptic vesicles were attached to collodion carbon-coated electron microscopy grids, previously made hydrophilic by glow discharge. These grids were placed on drops of a synaptic vesicle suspension (0.7 mg/ml in glycine buffer; 300 mM glycine, 5 mM Hepes-NaOH, pH 7.3) and incubated for 5 min at room temperature. After removal of excess liquid, the absorbed material was fixed by incubation with 0.5% glutaraldehyde in TBS (30 mM Tris, 150 mM NaCl, pH 8.2) for 2 min. Grids were then incubated for 10 min in TBG (TBS containing 0.1% bovine serum albumin and 1% gelatin) to block nonspecific binding. Subsequently the grids were transferred to drops of anti-synaptophysin (1/10 dilution) or anti-GLYT2 (1/2 dilution) and incubated for 1 h. This step was omitted in control incubations. Then, after several washes with TBG, the grids were incubated for another 45 min with goat anti-mouse IgG conjugated to 10-nm colloidal gold or 10-nm gold-protein A conjugate for antisynaptophysin or anti-GLYT2, respectively. Both secondary antibodies (Biocell Research Laboratory, Cardiff, United Kingdom) were diluted 1/40 in TBG. Grids were then washed three times in TBS, followed by three washes in glycine buffer, and fixed with 2% glutaraldehyde in TBS for 2 min. Finally the grids were negatively stained for 40 s in a 2% aqueous solution of uranyl acetate. Samples were examined in a JEM-1010 electron microscope, and pictures were taken using a 792 Bioscan digital camera (Gatan Inc., Warrendale, PA).

RESULTS
The starting point of the present study was to delineate the physiological role of the interaction between syntaxin 1 and the glycine transporter GLYT2. Since syntaxin 1 plays a crucial role in the exocytosis of neurotransmitters and also interacts with several neurotransmitter transporters, we decided to stimulate exocytosis in a brain-derived preparation and study its effect on the glycine transporter GLYT2, both in the presence and the absence of functional syntaxin 1. The last is possible by inactivating syntaxin 1 in vivo with Bont/C, a neurotoxin that specifically cleaves syntaxin 1 (27). For these studies purified sealed synaptic terminals (synaptosomes) were used because these allow the treatment of Bont/C and additionally contain the complete functional machinery of exocytosis and reuptake of neurotransmitters.
Bont/C Inhibits Glycine Release from Synaptosomes-Synaptosomes were prepared from rat spinal cord and brainstem because it has been reported that GLYT2, the model transporter in our studies, is abundant in neurons from these areas (28). Synaptosomes were preloaded with [ 3 H]glycine, and calcium-dependent exocytosis was provoked by plasma membrane depolarization using 4-aminopyridine in the presence of extracellular Ca 2ϩ . 4-Aminopyridine blocks certain voltagegated K ϩ channels, causing spontaneous repetitive firing in synaptosomes and a sustained rise in synaptosomal free Ca 2ϩ concentration (29,30). Furthermore, it was suggested that 4-aminopyridine closely mimicked physiological stimulation of calcium channels, resulting in a calcium influx (31). This treatment provoked a release of [ 3 H]glycine from synaptosomes ( Fig. 1), whereas no release was measured when 4-aminopyridine or extracellular calcium were added individually (data not shown). A 30-min preincubation of the synaptosomes with 1.0 M bafilomycin, an inhibitor of the V-ATPase that prevents the entrance of neurotransmitter into synaptic vesicles (32), strongly reduced the release of [ 3 H]glycine (Fig. 1). This demonstrates that the release of [ 3 H]glycine from the synaptosomes is mainly due to exocytosis of [ 3 H]glycine-filled synaptic vesicles. Accordingly, preincubation of synaptosomes with Bont/C resulted in an inhibition of [ 3 H]glycine release, and a partial (approximately 55%) proteolysis of syntaxin 1 (Fig. 1, inset). It has been reported that Bont/C causes a 60% maximal breakdown of syntaxin 1 in synaptosomes, and that this is sufficient to completely block vesicular neurotransmitter release (27).
Syntaxin 1 Inactivation Leads to a Reduction in GLYT2 on the Plasma Membrane-This study was performed on the glycine transporter GLYT2, which, in contrast to isoform 1, is mainly neuronal (28). Fig. 2A shows the time course of [ 3 H]glycine transport in synaptosomes at a concentration of 150 nM [ 3 H]glycine. The uptake of [ 3 H]glycine was completely abolished in the absence of chloride ions in the uptake assay ( Fig. 2A), demonstrating that low affinity glycine transport systems, which do not depend on Cl Ϫ ions (33), do not contribute to glycine transport. Additionally, all transport assays were carried out in the presence of sarcosine, a specific inhibitor of GLYT1. Therefore, glycine transport in our experiments is mainly due to the activity of GLYT2. A 90-min pre-treatment of synaptosomes with 80 nM Bont/C reduced glycine transport by approximately 50% (Fig. 2A). A pre-treatment with Bont/B, a neurotoxin that blocks neurotransmitter release by proteolytic cleavage of VAMP (34), reduced glycine uptake in a similar way as Bont/C did ( Fig. 2A). Kinetic constants measurements showed that Bont/C treatment mainly affected the V max of GLYT2 (K m ϭ 26.7 Ϯ 1.3 M and V max ϭ 4.39 Ϯ 0.18 mol/mg of protein/min for control synaptosomes, and K m ϭ 20.1 Ϯ 0.4 M and V max ϭ 2.27 Ϯ 0.14 mol/mg of protein/min for Bont/ C-treated synaptosomes). Changes in V max are usually indicative for changes in the number of plasma resident transporters; therefore, surface biotinylation experiments were performed to study this issue in more detail. These studies revealed a reduction in GLYT2-surface immunoreactivity after Bont/C (55% of control) or Bont/B (65% of control) treatment (Fig. 2B), which explains the observed inhibition in glycine uptake after neurotoxin treatment. This is in agreement with that found for the ␥-aminobutyric acid transporter GAT1 in hippocampal cul-tures, where GAT1 partially disappeared from the plasma membrane after Bont/C incubation (35). Our results show that two SNARE proteins (syntaxin 1 and VAMP) somehow control the number of GLYT2 proteins on the plasma membrane. To determine why GLYT2 disappears partly from the plasma membrane, we hypothesized that neurotoxin treatment might interfere in GLYT2 trafficking and impede the transporter to reach the plasma membrane. To study this issue, we analyzed in more detail surface expression of GLYT2 in Bont/C-treated and non-treated synaptosomes.
Syntaxin 1 Is Necessary in the Arrival of GLYT2 to the Plasma Membrane-As syntaxin 1 is both involved in exocytosis and in the regulation of GLYT2, we provoked plasma membrane depolarization in synaptosomes by addition of 4-aminopyridine and extracellular calcium, and studied its effect on GLYT2 trafficking by biotinylation. 4-Aminopyridine and CaCl 2 were added to a suspension of synaptosomes, and, at different time points, samples were taken and placed on ice. Subsequently, the synaptosomes in these samples were subjected to biotinylation and the amount of GLYT2 present on the plasma membrane revealed by Western blotting with GLYT2 antibodies. Fig. 3 shows that GLYT2 disappeared gradually from the plasma membrane during conditions of sustained exocytosis. This effect reached a maximum after 15 min and was not seen when 4-aminopyridine or CaCl 2 were individually added (data not shown). Then, after 15 min of incubation, the remaining extracellular calcium was removed by washing with EGTA, and the synaptosomes were incubated for another 15 min in the absence of 4-aminopyridine and Ca 2ϩ . This treatment resulted in a re-incorporation of GLYT2 into the plasma membrane of control synaptosomes, although not completely to the level at the beginning of the experiment. Even after longer incubation times, the starting level could not be reached (Fig.  3). This may be explained by the fact that EGTA only removes extracellular calcium and does not affect intracellular levels.
Anyhow, this experiment clearly shows that plasma membrane depolarization in the presence of extracellular calcium results in a removal of GLYT2 from the plasma membrane on a minute time scale, and that removal of the extracellular calcium partly reverses this effect. When this experiment was performed with synaptosomes pretreated with Bont/C, GLYT2 was still able to leave the plasma membrane after depolarization, but lost its capacity to return after EGTA washing (Fig. 3). This last result demonstrates that syntaxin 1 is necessary in the arrival and the incorporation of GLYT2 into the plasma membrane, and is not involved in the removal of the transporter from the plasma membrane. This may also explain why GLYT2 is less abundant on the plasma membrane after Bont/C incubations (Fig. 2B). During the 90-min preincubation with Bont/C, GLYT2 proteins are able to leave the plasma membrane, but are subsequently unable to return, resulting in an overall removal of GLYT2 from the plasma membrane after the preincubation.
Calcium Influx Results in an Immediate Incorporation of GLYT2 into the Plasma Membrane-To test if calcium also has a fast effect on GLYT2 trafficking, we examined, in an experiment similar to that described above, what happened to GLYT2 in the first few seconds after plasma membrane depolarization and calcium influx. When 4-aminopyridine and CaCl 2 were added to synaptosomes at 37°C, a slight decrease in the amount of GLYT2 present on the plasma membrane was im-mediately observed (data not shown). However, when the same experiment was performed at 25°C, an immediate up-regulation was seen directly after stimulation (Fig. 4). This effect was most prominent after the first second of incubation with 4-aminopyridine and CaCl 2 , and was not detected when 4-aminopyridine and CaCl 2 were individually added (data not shown). The biotinylation method chosen to study this fast arrival of GLYT2 to the plasma membrane allows to take samples as soon as 1 s, but it is not possible to study this phenomenon during the first second. This is probably the reason why this effect was not observed at 37°C because, at this temperature, the process is most likely much faster and over when the first sample was taken. Nonetheless, with this method, we were able to detect an immediate increase in the number of GLYT2 on the plasma membrane by a factor of approximately 2.4 (2.4 Ϯ 0.6; n ϭ 4). This increase is immediately followed by a retrieval of GLYT2 from the plasma membrane, as was already demonstrated in Fig. 3. When the experiment was performed at 25°C with synaptosomes previously treated with Bont/C, this rapid increase was not observed, again demonstrating that syntaxin 1 is necessary for GLYT2 to arrive to the plasma membrane (Fig. 4).
GLYT2 Is Present in Synaptic-like Vesicles-The active trafficking of GLYT2 toward and from the plasma membrane prompted us to investigate if GLYT2 is being transported by small vesicles. Therefore, crude synaptosomes were lysed and the synaptic vesicle population purified. During the purification, several samples were taken and analyzed by Western blotting. Fig. 5A demonstrates the enrichment of the synaptic vesicle marker synaptophysin in the purified fraction. On the other hand, no plasma membrane contamination was present in the purified synaptic vesicle preparation since no Na ϩ /K ϩ ATPase could be detected after the purification. Immunoblots using an endosomal marker (transferrin receptor) showed that fraction SV was almost free of endosomal membranes. Syntaxin 1 was found to recycle with synaptic vesicles (36), which explains why this SNARE protein was present in the purified synaptic vesicle fraction. Interestingly, we detected GLYT2 immunoreactivity in the purified fraction (Fig. 5A), although GLYT2 immunoreactivity was not enriched, suggesting that FIG. 3. Syntaxin 1 is necessary for the arrival to but not for the removal of GLYT2 from the plasma membrane. At time 0, 1 mM 4-aminopyridine and 5 mM CaCl 2 (4AP/Ca 2ϩ ) were added to control synaptosomes. At the indicated time points, aliquots were taken and immediately chilled on ice. After 15 min of incubation at 37°C, the remaining synaptosomes were washed with 2 mM EGTA to remove all extracellular calcium, and, after another 15 and 30 min of incubation, the final samples were collected. All samples were subsequently subjected to biotinylation. The isolated biotinylated proteins, which represent the plasma membrane fraction, were then revealed by Western blotting using anti-GLYT2. The results show that calcium stimulates the removal of GLYT2 from the plasma membrane on a minute time scale, and after washing the transporter returns partly to its origin (control). The same experiment was then performed with Bont/Ctreated synaptosomes. The results of this experiment are similar compared with control synaptosomes, except that, after removal of the extracellular calcium, GLYT2 is unable to return to the plasma membrane. The graph shows the densitometry analysis of the data obtained from the Western blots, and every bar is the average of four independent experiments (mean Ϯ S.E.). Asterisks indicate that data are significantly different from time 0 data as determined by Student's t test (p Ͻ 0.01) .   FIG. 4. Fast effect of calcium on GLYT2. Membrane depolarization in the presence of extracellular calcium stimulates the incorporation of GLYT2 into the plasma membrane, which is directly followed by its internalization. This effect was observed by biotinylation studies performed as described in Fig. 3 at the indicated time points and at 25°C (control), and was abolished by treating the synaptosomes with Bont/C. In the graph, the up-regulation of GLYT2 is demonstrated after densitometry of the bands obtained in the Western blots and shows the average of four independent experiments (mean Ϯ S.E.). Asterisk indicates that data are significantly different from time 0 data as determined by Student's t test (p Ͻ 0.05). most of GLYT2 is present on the plasma membrane or on other intracellular compartments. Immunogold detection using anti-GLYT2 confirmed the presence of GLYT2 in small (40 -60-nm) synaptic-like vesicles (Fig. 5B). Approximately 5-10% of the vesicles were labeled with anti-GLYT2, whereas synaptophysin-immunogold labeling on the same vesicle preparation demonstrated that this synaptic vesicle marker was present in most (80 -90%) of the vesicles (Fig. 5C). These experiments show for the first time GLYT2 immunoreactivity in small synaptic-like vesicles, although the precise nature of these GLYT2containing vesicles has yet to be determined. DISCUSSION In the present study, we have shown that treatment of synaptosomes with an agent that promotes plasma membrane depolarization in the presence of extracellular calcium stimulates the arrival of the neuronal glycine transporter GLYT2 to the plasma membrane, and that this effect is immediately followed by an internalization of the transporter. The arrival, but not the removal, of GLYT2 from the plasma membrane is mediated by syntaxin 1 and probably other SNARE proteins. Moreover, the present data demonstrate that GLYT2 is present in synaptic-like vesicles, which might represent GLYT2 that is being trafficked. These results reveal the close correlation between release and uptake of neurotransmitter.
Regulation of GLYT2 by Calcium-Recently the proline transporter was found to be localized in synaptic vesicles (37), and the noradrenaline transporter in secretory granules of PC12 cells (38). Both neurotransmitter transporters belong to the same family as the glycine transporters and are thought to be responsible for the clearance of, respectively, proline and noradrenaline from the extracellular space. Here data are presented that indicate GLYT2 is also present in synaptic-like vesicles. At present, it is not known whether transporters and neurotransmitters share the same vesicles or use specialized individual vesicles. Nonetheless, a coordination in the exocytosis of such vesicles should exist making a connection between release of neurotransmitter and its reuptake by transporters. The observation that the number of GLYT2 proteins on the plasma membrane rapidly augments after membrane depolarization of nerve terminals now strengthens this hypothesis of an exocytotic-like transportation of neurotransmitter transporters from an intracellular compartment to the plasma membrane. Jayanthi et al. (39) described a fast and slow effect of calcium on the substrate-induced currents mediated by the proline transporter. These experiments, however, were done in transfected HEK cells, a cell line not known to have a specialized exocytotic machinery. Additionally, previous work on the serotonin transporter expressed in JAR cells also showed a long term inhibition of calcium on the transport rate. This inhibition reached 85% after 1 h of incubation with the calmodulin antagonist CGS 9343B and was found to affect mainly the V max of the serotonin transporter (40). A decrease in V max is usually indicative for a reduction in the number of transporters on the plasma membrane. Taking all these data together, it seems that the long term effect of calcium on the neurotransmitter transporters might be modulated by protein kinases. This idea is strengthened by the fact that activation of protein kinase C resulted in an inhibition of glycine transport in glioblastoma cells (41); moreover, other neurotransmitter transporters that belong to the same family as GLYT2, like the serotonin and dopamine transporters and GAT1, were found to be internalized upon protein kinase C mediation (42)(43)(44).
Fusion of synaptic vesicles with the plasma membrane, followed by the compensatory endocytosis, induces the redistribution of certain proteins present on the plasma membrane. This was recently found to be the case for several SNARE FIG. 5. Immunogold labeling on a synaptic vesicle preparation demonstrates the presence of GLYT2 in this fraction. Crude synaptosomes were purified from rat spinal cord and brainstem by differential centrifugation (fraction P2) and subsequently lysed by an osmotic shock. Then large membranes were removed by centrifugation (20 min at 25,000 ϫ g) and the supernatant was subjected to another centrifugation step (2 h at 165,000 ϫ g) to concentrate the remaining membranes. The pellet was resuspended (fraction LP2) and layered on top of a continuous (50 -800 mM) sucrose gradient. After centrifugation (3 h at 165,000 ϫ g), the band (200 -400 mM sucrose) corresponding to synaptic vesicles was removed and used in a flotation density gradient as described under "Experimental Procedures." The band (380 -600 mM sucrose) containing the synaptic vesicles was subsequently isolated and called SV. The amounts of synaptophysin (synaptic vesicle marker), Na ϩ /K ϩ ATPase (plasma membrane marker), transferrin receptor (endosomal marker), syntaxin 1 (localized in plasma membrane and synaptic vesicles), and GLYT2 in the samples taken during purification (P2, LP2, and SV) were determined by Western blotting (10 g of protein/lane, A). The purified synaptic vesicles were then immunogold labeled with anti-GLYT2 (B) or anti-synaptophysin (C), negatively stained using uranyl acetate, and examined by electron microscopy. Only small vesicular membrane structures were seen in fraction SV. In most vesicles (80 -90%), anti-synaptophysin label was detected, whereas, in approximately 5-10% of the vesicles, GLYT2 was labeled. In controls, where the first antibody (anti-GLYT2) was omitted in the incubations, no immunoreactivity was observed. Bar, 100 nm.
proteins, e.g. syntaxin 1 and SNAP-25 (45). These SNAREs, which are mainly present on the plasma membrane, are internalized with vesicle proteins during endocytosis coupled to exocytosis and are subsequently re-inserted into the plasma membrane although at different rates (45). In addition, the dopamine transporter was recently found to be internalized upon protein kinase C activation and targeted to the endosomal/lysosomal pathway (43,46). This dynamin-dependent mechanism might also be used by GLYT2 during coupled exocytosis/endocytosis. Future work, however, should clarify the exact mechanism by which GLYT2 and possibly other neurotransmitter transporters are being trafficked to and from the plasma membrane upon calcium activation.
Physiological Role of Syntaxin 1 in Glycine Reuptake-The co-localization of GLYT2 and syntaxin 1 on the neuronal plasma membrane and the functional and physical interaction between these two membrane proteins suggested a physiological role for this interaction (16). Overexpression of GLYT2 and syntaxin 1 in COS cells resulted in an internalization of the transporter (16). However, the data presented here obtained from a brain-derived preparation demonstrate that syntaxin 1 does not stimulate internalization of the transporter but is involved in its arrival to the plasma membrane. These seemingly different results might be explained by a disruption in the equilibrium of the SNARE system caused by overexpression of syntaxin 1, which may vary the rate of exocytosis, protein trafficking, or even compete with native SNARE proteins (20). Deken et al. (35) proposed both a trafficking and a functional inhibitory role for syntaxin 1 in GAT1 regulation, and both processes might be separated in place and time. It could well be that this also holds for GLYT2, although future experiments may unravel the details of this possible dual role for syntaxin 1. To conform to the results presented previously for GAT1, where syntaxin 1 was proposed to be a positive regulator for GAT1 expression (35), we now propose that syntaxin 1 is involved in the arrival and insertion of GLYT2 into the plasma membrane. These results demonstrated a syntaxin 1 involvement both in the fast arrival of GLYT2 after plasma membrane depolarization, as well as in calcium-independent arrival of GLYT2 to the plasma membrane. The last became evident from the results presented in Fig. 3. After endocytosis of GLYT2 and subsequent EGTA washing, GLYT2 was unable to return to the plasma membrane in Bont/C-treated synaptosomes, in contrast to non-treated synaptosomes.
We have shown here that calcium controls the number of GLYT2 proteins on the plasma membrane and that at least two SNARE proteins (syntaxin 1 and VAMP) are involved in GLYT2 trafficking. The results, therefore, suggest that GLYT2 trafficking is accomplished by a process where GLYT2-containing vesicles fuse with the plasma membrane. Previous studies revealed a direct physical contact between syntaxin 1 and GLYT2 (16), similar to the one found for GAT1 (35). Moreover, these interactions reduced transport rates in both transporters, suggesting a regulatory inhibitory function for syntaxin 1. From the results presented here, however, it does not become clear if GLYT2 trafficking and its incorporation into the plasma membrane requires a direct interaction between the transporter and syntaxin 1. However, it seems reasonable to think that fusion of GLYT2-containing vesicles with the plasma membrane, like synaptic vesicle fusion, requires only SNARE proteins, and therefore the syntaxin 1-GLYT2 interaction might only play a role in the inhibitory function of syntaxin 1.
In conclusion, calcium seems to have two effects on the arrival and the retrieval of the glycine transporter GLYT2 toward and from the plasma membrane. First, it causes a fast syntaxin 1-dependent increase in the concentration of GLYT2 on the plasma membrane. This up-regulation is directly followed by a removal of GLYT2 from the plasma membrane under sustained exocytotic conditions. At the moment, we can only speculate about the physiological role of these processes. It could well be that the nerve terminal regulates in this way the duration of the neurotransmission signal, both in the short term and the long term. For the short term regulation of GLYT2, it seems reasonable to think that during fast neurotransmission more GLYT2 might be transiently needed to clear the neurotransmitter from the synaptic cleft in an efficient way, keeping the neurotransmitter signal short. On the other hand, the long term down-regulation of GLYT2 might result in a longer presence of glycine in the synaptic cleft in situations of sustained exocytosis. Within this context, it has been proposed that the redistribution of membrane proteins may play an important role in synaptic plasticity (15,37). Without any doubt, many data now point to the fact that the fine tuning of neurotransmitter transporters is crucial in neurotransmission; as shown in this and other reports, this mechanism should be controlled tightly in space and time.