Modulation of the Neural Glutamate Transporter EAAC1 by the Addicsin-interacting Protein ARL6IP1*

Addicsin (Arl6ip5) is a murine homologue of rat glutamate transporter-associated protein 3-18 (GTRAP3-18), a putative negative modulator of Na+-dependent neural glutamate transporter-excitatory amino acid carrier 1 (EAAC1). Here we report that ADP-ribosylation factor-like 6 interacting protein 1 (Arl6ip1) is a novel addicsin-associated partner that indirectly promotes EAAC1-mediated glutamate transport activity in a protein kinase C activity-dependent manner. Like addicsin, Arl6ip1 is expressed in numerous tissues and proved likely to be co-localized with addicsin in certain neurons in the matured brain. Arl6ip1 was not translocated from the subcellular compartments under any of the test conditions and had no association with any molecules on the plasma membrane. Immunoprecipitation assay demonstrated that Arl6ip1 bound directly to addicsin and that the hydrophobic region located at amino acids 103–117 of addicsin was crucial to the formation of the Arl6ip1-addicsin heterodimer and addicsin homodimer. Glutamate transport assay revealed that increasing the expression of Arl6ip1 in C6BU-1 cells markedly enhanced Na+-dependent EAAC1-mediated glutamate transport activity in the presence of 100 nm phorbol 12-myristate 13-acetate. Under these conditions, kinetic analyses demonstrated that EAAC1 altered glutamate transport activity by increasing its glutamate affinity but not its maximal velocity. Meanwhile, increasing expression of addicsin Y110A/L112A mutant lacking binding ability for Arl6ip1 showed no enhancement of EAAC1-mediated glutamate transport activity, regardless of phorbol 12-myristate 13-acetate activation, suggesting that association between addicsin and Arl6ip1 causes altered EAAC1-mediated glutamate transport activity. Our findings suggest that Arl6ip1 is a novel addicsin-associated partner that promotes EAAC1-mediated glutamate transport activity by decreasing the number of addicsin molecules available for interaction with EAAC1.

Glutamate is the major excitatory neurotransmitter in excitatory synapses and the metabolic substrate for ␥-aminobutyric acid synthesis in the inhibitory neurons in the mammalian central nervous system. Extracellular glutamate levels are strictly regulated by Na ϩ -dependent high affinity excitatory amino acid transporters (EAATs), 4 which convey L-glutamate, L-aspartate, and D-aspartate from the extracellular space into the cells by means of the Na ϩ -K ϩ -coupled transport system (1). EAATs can be classified into five different homologues, designated EAAT1 (glutamate/aspartate transporter (GLAST)), EAAT2 (glutamate trasnporter-2 (GLT-1)), EAAT3 (excitatory amino acid carrier 1 (EAAC1)), EAAT4, or EAAT5 (2). In the CNS, GLAST and GLT-1 mainly localize to the astrocytes and play a major role in protecting neurons from glutamate-induced toxicity (3) as well as terminating glutamatergic transmissions (4,5). On the other hand, EAAC1 is widely expressed in the CNS and localizes in the somata and dendrites of certain neurons in the CNS (6,7). EAAT4 is expressed in cerebellar Purkinje cells and EAAT5 in the retina (8 -10).
The function of EAAC1 in the CNS has not been established. However, recent studies demonstrate that EAAC1 acts as a cysteine transporter and maintains neuronal glutathione metabolism (11), and that EAAC1 has a unique mitochondria-mediated anti-apoptotic function in injured motor neurons (12). Knock-out of EAAC1 in rodents leads to the development of epilepsy resulting from reduced synthesis of the neurotransmitter ␥-aminobutyric acid (13), dicarboxylic aminoaciduria, and significant motor impairment (14). Changes in glutamate transport activity of EAAC1 are also associated with long term potentiation and fear conditioning (15). These pieces of evidence suggest that EAAC1 may contribute physiologically to multiple neurotic functions as distinct from glutamate clearance in the CNS. Furthermore, a number of reports demonstrate that intracellular signaling molecules and accessory proteins regulate the expression and function of EAAC1 (16). For instance, protein kinase C (PKC) and phosphatidylinositol 3-kinase regulate the cell surface expression and intrinsic activity of EAAC1 (17).Furthermore,GTRAP3-18(glutamatetransporter-associated protein [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18] negatively modulates EAAC1-mediated glutamate transport by protein-protein interaction with EAAC1 (18).
We have recently identified a novel mRNA encoding a protein designated "addicsin" that is highly up-regulated in amygdala nuclei in morphine-administered mice (19). Amino acid comparison reveals that addicsin is a murine homologue of GTRAP3-18, a putative modulator of EAAC1, and JWA, a human vitamin A-responsive factor (18,19). The protein known in various papers as addicsin, GTRAP3-18, or JWA has recently been renamed Arl6ip5. Addicsin/GTRAP3-18 is widely distributed in various tissues, including the brain (19,20). In the CNS, addicsin is predominantly expressed in the principal neurons of the mature brain such as the pyramidal cells and ␥-aminobutyric acid-ergic interneurons scattered in the hippocampal formation (21,22). Addicsin is thought to be localized in intracellular compartments and implicated in intracellular trafficking events, because it is a member of the prenylated Rab acceptor 1 domain family, member 3 (PRAF3) (23). In addition, JWA, a human homologue of addicsin, is responsive to environmental oxidative stress, such as hydrogen peroxide (24), and is up-regulated in the thalamus in schizophrenia (25). These data clearly demonstrate that addicsin may play crucial roles in basic physiological functions in vivo. However, at present, the molecular functions of addicsin remain largely unknown. To gain further insights into this issue, we particularly focused on the evidence that addicsin/GTRAP3-18 tends to form a multimeric complex (18,19). This has prompted us to attempt to identify addicsin-associated factors using a yeast two-hybrid screen. Here, we report that ADP-ribosylation factor-like 6 interacting protein 1 (Arl6ip1) is an addicsin-associated partner and positively modulates EAAC1mediated glutamate transport activity in a PKC-activitydependent manner.

EXPERIMENTAL PROCEDURES
Animals-Male ddY mice (6 weeks old; 25-30 g in body weight; Japan SLC, Inc., Shizuoka, Japan) were maintained in individual cages (12-hour light-dark cycle; 23-24°C) and used in the experiments. All animals received humane care in accordance with the National Institute of Advanced Industrial Science and Technology guidelines.
Yeast Two-hybrid Screening-Matchmaker two-hybrid system 3 (Clontech) was used for the yeast two-hybrid screen. A full length of addicsin cDNA was subcloned into a bait vector, pGBKT7, by fusion with the GAL4 DNA-binding domain. The oligo(dT)-primed cDNAs prepared from the amygdala in repeated morphine-administered mice were subcloned into a prey vector, pACT2, by fusion with the GAL4 activation domain. These constructs were co-transfected into the yeast AH109 strain using the lithium acetate method. After the transformants had grown on tryptophan-, leucine-, and histidinedeficient plates (SD-TLH) containing 20 mg/ml 5-bromo-4chloro-3-indolyl-␣-D-galactopyranoside (X-␣-gal) (Clontech) at 30°C, the resultant blue colonies were selected. The plasmids were isolated using Zymoprep (Zymo Research Corp., Orange, CA), and their cDNA sequences were then analyzed.
Western Blot Analysis-For distribution analysis, mouse brain S2 fractions were prepared as described previously (19). The S2 fractions (40 g) suspended in SDS sample buffer (62.5 mM Tris-HCl, pH 6.8, 20% glycerol, 2% SDS, 0.005% bromphenol blue, 5% 2-mercaptoethanol) were subjected to 12% SDS-PAGE and then transferred to a polyvinylidene fluoride membrane (Bio-Rad). For immunoprecipitation assay, whole cell lysates were prepared by dissolving in 100 l of SDS sample buffer. The whole cell lysates (10 l) were subjected to Tris-Tricine SDS-PAGE (16.5% T, 3% C) and then transferred to the polyvinylidene fluoride membrane. After the blots had been blocked with 10% dried milk in PBS containing 0.1% Tween 20 (PBS-T), they were incubated with specific antibodies for the target proteins. After incubation with HRP-conjugated secondary antibody, specific signals were detected by the enhanced chemiluminescence (ECL) system (BD Biosciences).
Immmunoprecipitation Assay-For in vitro immunoprecipitation assay, at 48 h after transfection using Lipofectamine 2000, NG108-15 or COS7 cells were dissolved in the RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1.0% Nonidet P-40, 0.1% sodium deoxycholate, 0.1% SDS, 1 mM EDTA) containing 0.1% protease inhibitor mixture. Cell extracts were incubated with 0.1 volume of anti-FLAG M2 affinity gel or antic-Myc affinity gel overnight at 4°C. After washing five times with TBS buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl), gels were suspended in SDS sample buffer and subjected to Western blotting. For in vivo immunoprecipitation assay, mouse whole brains were homogenized with a 10-fold volume of lysis buffer (20 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1.0% Triton X-100, 1 mM EDTA, and 0.1% protease inhibitor). After centrifugation at 100,000 ϫ g for 1 h, the supernatant (250 g/ml) was incubated with rabbit anti-Arl6ip1 serum or rabbit preimmune serum for overnight at 4°C. The supernatant was further incubated with protein A-Sepharose for 1-2 h at 4°C. After rinsing five times with the lysis buffer, the protein A-Sepharose was suspended in SDS sample buffer and then subjected to Western blotting.
Glycerol Gradient Centrifugation Analysis-Glycerol gradient centrifugation analysis was carried out as described previously (26). NG108-15 cells were co-transfected with pcDNA-Arl6ip1-FLAG and pcDNA-addicsin-myc using Lipofectamine 2000. At 24 h after transfection, cells were dissolved in RIPA buffer containing 0.1% protease inhibitor. 1 ml of cell lysate was layered onto 10 ml of a 10 -40% linear glycerol gradient containing 10 mM Tris-HCl, pH 8.0, 120 mM NaCl, 1 mM EDTA, and 1 mM 2-mercaptoethanol. After centrifugation for 24 h at 130,000 ϫ g in a Hitachi RPS40T rotor at 4°C, 0.5-ml fractions were collected from the bottom of each centrifugation tube. Samples were analyzed by Western blotting using anti-FLAG M2 monoclonal antibody and anti-c-Myc monoclonal antibody.
Glutamate Transport Assay-The glutamate transport assay was performed as described previously (27). Briefly, C6BU-1-pSw-Arl6ip1,C6BU-1-pSw-addicsin,orC6BU-1-pSw-addicsin-Y110A/L112A cells were cultured on the PRIMARIA 24-well plate (BD Biosciences). After exposure to vehicle (2.2 mM ethanol) or 10 nM Mifepristone for 24 h at 37°C, cells were further incubated in 10% FBS-DMEM supplemented with vehicle (1.3 mM DMSO) or 100 nM phorbol esters for 30 min at 37°C. Cells were rinsed three times with 500 l of sodium-or cholinecontaining buffer maintained at 37°C (5 mM Tris-HCl, 10 mM HEPES, 2.5 mM KCl, 1.2 mM CaCl 2 , 1.2 mM MgCl 2 , 1.2 mM K 2 HPO 4 , 10 mM glucose, 140 mM NaCl, or choline chloride) before incubation with 0.5 M L-[ 3 H]glutamate (2.5 Ci/ml) (GE Healthcare) and 30 M unlabeled glutamate for 5 min at 37°C in a final volume of 400 l/well. The glutamate uptake was terminated using triple washing with 500 l of ice-cold cholinecontaining buffer. After cells were solubilized in 400 l/well of 0.1 N NaOH, the radioactivity of 360 l/well lysates was analyzed using an LS6500 scintillation counter (Beckman Coulter Inc., Fullerton, CA). Moreover, a 20-l of aliquot was used to determine its protein concentration using a DC protein assay kit (Bio-Rad). Five independent glutamate uptake experiments were performed in quintuplicate. For the saturation analysis, L-glutamate concentrations ranged from 10 to 500 M. Kinetic parameters were determined using the Eadie-Hofstee equation. In the knockdown experiments, cells were grown in 12-well plates and transiently transfected with each double-stranded siRNA (60 pmol/well) for 24 h by using Lipofectamine 2000. After transfection, the cells were treated with 1.3 mM DMSO or 100 nM PMA for 30 min at 37°C and then subjected to glutamate transport assay. The sense sequence for the siRNA control was 5Ј-AGCAGGUGUUUCUUGCAAAdTdT-3Ј. The sense sequence for siRNA 1 was 5Ј-UUCUCCGAACGUGUCAC-GUdTdT-3Ј. The sense sequence for siRNA 2 was 5Ј-UUUG-CAAGAAACACCUGCUdGdG-3Ј. The siRNA 1 and siRNA 2 knocked down about 50 and 15%, respectively, of the total addicsin mRNA.
Biotinylation Assay-Biotinylation assay was performed as described previously (17). In brief, C6BU-1-pSw-Arl6ip1, C6BU-1-pSw-addicsin, and C6BU-1-pSw-addicsin Y110A/ L112A cells were cultured in 10% FBS-DMEM containing 100 g/ml Hyglomycin and 250 g/ml Zeocin on 35-mm culture dishes (BD Biosciences). Cells were washed three times with phosphate-buffered saline containing Ca 2ϩ and Mg 2ϩ (PBS-Ca 2ϩ -Mg 2ϩ ; 138 mM NaCl, 2.7 mM KCl, 1.5 mM KH 2 PO 4 , 9.6 mM Na 2 HPO 4 , 1 mM MgCl 2 , and 0.1 mM CaCl 2 , pH 7.3). The cells were then reacted with 0.5 mg/ml EZ-Link sulfo-NHSbiotin in PBS-Ca 2ϩ -Mg 2ϩ at 4°C for 20 min. Cells were incubated with 100 mM glycine in PBS-Ca 2ϩ -Mg 2ϩ at 4°C for 20 min to remove nonreacted biotin and dissolved in 250 l RIPA buffer containing 0.1% protease inhibitor. Cell extracts were centrifuged at 16,500 ϫ g for 30 min at 4°C to remove insoluble materials and a 125-l aliquot of total lysate fraction was obtained. The total lysate fraction was further incubated with equal volume of 50% slurry of avidin beads for 1 h and centrifuged at 16,500 ϫ g for 10 min at 4°C to recover the intracellular lysate fraction. The avidin beads were washed six times with 250 l of RIPA buffer and were subjected to 65 l of SDS sample buffer to elute the absorbed proteins while shaking for 30 min. All preserved samples were analyzed by Western blotting.
Immunohistochemical Analysis-For distribution analysis of Arl6ip1 and addicsin in vivo, fresh frozen coronal sections 16 m thick were prepared from mouse adult brain (6-week-old male ddY mouse, Bregma ϩ4.2, Ϫ1.8 or Ϫ5.3 mm). The sections were stained using the Vectastain TM Elite ABC kit (Vector, Burlingame, CA) according to the manufacturer's protocol. In brief, the sections were fixed with acetone/methanol solution (w/w, 1:1) for 15 min and blocked with 1.5% normal goat serum for 20 min. They were then exposed to the rabbit antimouse Arl6ip1 IgG polyclonal antibody or rabbit anti-addicsin IgG overnight at 4°C. The sections were then incubated with biotinylated goat anti-rabbit IgG secondary antibody for 30 min and streptavidin-conjugated HRP complex for 30 min. Signals were detected using the 3,3Ј-diaminobenzidine solution. Images were acquired using a BX21 light microscope (Olympus, Tokyo, Japan). For immunostaining of C6BU-1-pSw-Arl6ip1 and C6BU-1-pSw-addicsin cells, cells were fixed with 4% paraformaldehyde in PBS for 10 min and permeabilized with 0.25% Triton X-100 in PBS containing 1% normal goat serum for 5 min. After blocking with 1% normal goat serum for 1 h, the cells were reacted with anti-V5 monoclonal antibody overnight at 4°C and then exposed to Alexa Fluor 488 goat anti-mouse IgG secondary for 1 h. After the cells had been coverslipped in Gel/Mount (Biomeda, Foster City, CA), images were acquired using a Fluoview FV1000 confocal laser-scanning microscope (Olympus, Tokyo, Japan).

Identification of Arl6ip1 as an Addicsin-associated Factor-
To identify potential factors associating with mouse addicsin, a yeast two-hybrid screen was carried out using a full length of addicsin cDNA as bait. From a prey cDNA library prepared from the amygdala of repeatedly morphine-administered mice, 49 positive clones were obtained. Of these positive clones, two clones clearly displayed ␣-galactosidase activity. These two clones yielded an identical cDNA sequence encoding mouse Arl6ip1 corresponding to its 1-59 amino acids (GenBank TM accession number AF223953) (data not shown). The amino acid sequence analysis revealed that addicsin had no homology with Arl6ip1. However, the hydrophobic profile of addicsin was almost the same as that of Arl6ip1 (data not shown). Reconfirmation testing using a full length of mouse Arl6ip1 as prey or bait revealed specific interactions with addicsin (data not shown). Furthermore, to test this interaction with addicsin and Arl6ip1 in vitro, we prepared crude cell lysates from NG108-15 cells in which the FLAG-tagged Arl6ip1 (Arl6ip1-FLAG) and Myc-tagged addicsin (addicsin-myc) were expressed. Immunoprecipitation assay using these cell lysates demonstrated that Arl6ip1-FLAG specifically interacted with addicsin-myc in the cell extracts prepared from co-expression cells (Fig. 1A, lanes 4  and 8), but not from Arl6ip1-FLAG (lanes 2 and 6) or addicsinmyc single-expression cells (lanes 3 and 7). No proteins were immunoprecipitated with either anti-FLAG or anti-c-Myc antibodies in sham NG108-15 cell extracts (Fig. 1A, lanes 1 and  5), suggesting that there is no specific binding to immunobeads and no cross-reaction of antibodies. Glycerol gradient centrifugation analysis revealed that the elution profile of Arl6ip1-FLAG was very similar to that of addicsin-myc (Fig. 1B). The elution peaks of both proteins were observed at fraction 18 with a deduced molecular mass of 24 kDa (Fig. 1B). In addition, the elution peak of addicsin homodimer was detected in the fractions corresponding to average molecular mass of 44 kDa (Fig.  1B, lower panel). These observations support the proposition that addicsin can form both addicsin-Arl6ip1 heterodimer and addicsin-addicsin homodimer in vitro. Furthermore, to test their interactions in vivo, an antiserum was generated using synthetic peptides corresponding to amino acids 185-199 of mouse Arl6ip1. The resultant antiserum specifically recognized an ϳ24-kDa single band in crude mouse whole brain lysates, consistent with the calculated molecular weight of Arl6ip1 (Fig.  1C, left panel, I), whereas the preimmune serum completely failed to react to the single band (Fig. 1C, left panel, P). The in vivo immunoprecipitation assay using mouse matured whole brain crude lysates demonstrated that addicsin was immunoprecipitated with anti-Arl6ip1 antibody but not preimmune antibody (Fig. 1C, right panel). These results suggest that addicsin interacts with Arl6ip1 in the mouse matured CNS.
Determination of the Arl6ip1-binding Region on Addicsin-To determine the region of addicsin responsible for binding to Arl6ip1, we carried out immunoprecipitation assays using several addicsin truncations. Arl6ip1 was able to associate with a full length of addicsin (wild type), a deletion mutant lacking the C-terminal region at amino acids 145-188 (d1), a truncation of the N-terminal domain at amino acids 1-102 (d2), and a mutant deleting the region containing the C-terminal PKC phosphorylation motif at amino acids 136 -144 (d3) ( Fig. 2A  and B). However, Arl6ip1 failed to interact with a mutant lacking a portion of the second hydrophobic region at amino acids 103-117 of addicsin (d4) ( Fig. 2A and B). As in the results for Arl6ip1, addicsin was able to interact with the wild type, d1, d2, and d3 mutant but not the d4 truncation (Fig. 2B). On the other hand, EAAC1 interacted with all the mutants (Fig. 2B). These results thus suggest that the hydrophobic region at amino acids 103-117 of addicsin is crucial to the homo-and hetero-multimerization of addicsin (Fig. 2C).
Distribution of Arl6ip1 mRNA and Protein in Vivo-To investigate the tissue distribution of Arl6ip1 transcript and protein using RT-PCR, we performed Western blotting and immunohistochemical analysis. RT-PCR analysis showed that Arl6ip1 mRNA was found in both the neuronal and non-neuronal tissues tested (Fig. 3A). Moreover, Arl6ip1 mRNA was co-expressed with addicsin mRNA in all the tissues examined (Fig. 3A). Western blot analysis demonstrated that Arl6ip1 was broadly expressed in various brain regions consistent with the ubiquitous localization of addicsin in matured CNS (Fig. 3B). Furthermore, immunohistochemical analysis revealed that Arl6ip1 was widely localized in matured brain and expressed in neuron-like cells (Fig. 3C, panels a, c, and e). For instance, in the olfactory bulb, Arl6ip1 immunoreactivity (IR) was detected in the mitral cell layer, glomerular layer, and granular layer (Fig.  3C, panel a). In the hippocampal formation, Arl6ip1 IR was observed in the granular layer of the dentate gyrus and stratum pyramidal of the CA1-3 fields (Fig. 3C, panel c). Arl6ip1 IR was also detected in the molecular layer, Purkinje cell layer, and granular cell layers of the cerebellum (Fig. 3C, panel e). Consistent with these results, addicsin IR also showed the same localization pattern as Arl6ip1 IR in these brain regions tested (Fig.  3C, panels b, d, and f), strongly suggesting that Arl6ip1 is colocalized with addicsin in the matured CNS.
Positive Modulation of EAAC1-mediated Glutamate Transport Activity by Arl6ip1-To investigate the effect of Arl6ip1 on EAAC1-mediated glutamate transport activity, we established stable C6BU-1 cell lines, designated C6BU-1-pSw-addicsin and C6BU-1-pSw-Arl6ip1, respectively. In these cell lines we were able to induce the V5-tagged addicsin (addicsin-V5) or Each fraction was analyzed by Western blotting using both anti-FLAG M2 antibody and anti-c-Myc antibody. In the lower panel, the open triangle indicates addicsin homodimer. C, in vivo immunoprecipitation assay using mouse whole brain lysates. In the left panel, to confirm the specificity of the rabbit anti-Arl6ip1 serum, mouse whole brain lysates were incubated with this serum (50 g) (I) or the rabbit preimmune serum (50 g) (P) overnight at 4°C. The 24-kDa band of Arl6ip1 was specifically observed. In the right panel, in vivo immunoprecipitation assay was carried out using mouse whole brain lysates. Western blotting using anti-addicsin antibody demonstrated that this Arl6ip1 antibody immunoprecipitated with addicsin in vivo.  V5-tagged Arl6ip1 (Arl6ip1-V5) in a closely mifepristone-dependent manner (Fig. 4, A and B). Next, to evaluate the effect of increasing expression of Arl6ip1-FLAG or addicsin-myc on EAAC1-mediated glutamate transport activity, we measured the EAAC1-mediated glutamate transport ability of C6BU-1-pSw-addicsin and C6BU-1-pSw-Arl6ip1 under various conditions. Compared with the control, the EAAC1-mediated glutamate transport activity was unchanged by up-regulation of addicsin-V5 or Arl6ip1-V5 (Fig. 4, A, lanes 1 and 2, and B, lanes  1 and 2). However, when these cells were stimulated with 100 nM PMA, the glutamate uptake activity in C6BU-1-pSw-addicsin fell steeply with increasing expression of addicsin-V5 (Fig.   3 and 4). By contrast, the glutamate transport activity in C6BU-1-pSw-Arl6ip1 significantly increased with up-regulation of Arl6ip1-V5 (Fig. 4B, lanes 3 and 4). When these cells were exposed to 100 nM 4␣-phorbol, a nonstimulating analogue of PMA, glutamate uptake activity in both cells remained at basal levels under all the conditions tested (Fig. 4, A, lanes 5 and  6, and B, lanes 5 and 6). Furthermore, cell viability assay demonstrated that neither cell line displayed any cytotoxicity because of up-regulation of addicsin-V5 or Arl6ip1-V5 (data not shown). Thus, these data strongly support the proposition that Arl6ip1 acts as a positive modulator of EAAC1-mediated glutamate transport in a PKC activity-dependent manner. Furthermore, to investigate the effect of decreased addicsin expression on glutamate uptake activity, we performed a knockdown experiment by transient transfection of double-stranded siRNAs into C6BU-1-pSw-Arl6ip1. Cells treated with addicsinspecific siRNAs had about a 200% increase in glutamate transport activity compared with cells treated with control siRNA under 100 nM PMA stimulation (Fig. 4C). Under DMSO treatment, cells transfected with siRNA 1 showed an increasing trend in basal glutamate transport activity (Fig. 4C). These data strongly support the results in Fig. 4B. Next, to evaluate the biochemical nature of altered EAAC1-mediated glutamate uptake, the kinetics of C6BU-1-pSw-Arl6ip1 was analyzed. When Arl6ip1-V5 was conditionally expressed, the PMAtreated cells showed an increase in affinity without any shift in maximal velocity (K m ϭ 647 M; V max ϭ 1.5 ϫ 10 3 pmol/mg/ min) as compared with DMSO-treated cells (K m ϭ 824 M; V max ϭ 1.5 ϫ 10 3 pmol/mg/min) (Fig. 4D). These data indicate that Arl6ip1 alters glutamate transport activity by increasing the catalytic efficiency of EAAC1.
Analysis of Subcellular Localization of Arl6ip1 in C6BU-1 Cells-To examine the molecular mechanism of the altered glutamate transport activity in C6BU-1-pSw-Arl6ip1, the subcellular localization and protein-protein interaction between EAAC1 and Arl6ip1 was analyzed. Western blot analysis revealed that Arl6ip1-V5 levels were unchanged by PMA treatment (Fig. 5A). Furthermore, immunohistochemical analysis demonstrated that Arl6ip1-V5 was localized in cytoplasmic structures and showed no changes in localization pattern as a result of 100 nM PMA treatment (Fig. 5B, panels a and b). Cell surface biotinyl assay indicated that Arl6ip1 had no association with any molecules on the cell surfaces under any of the conditions tested (Fig. 5C). Moreover, immunoprecipitation assay indicated that Arl6ip1 failed to interact with EAAC1, whereas addicsin could associate with EAAC1 (data not shown). These results suggest that Arl6ip1 is localized in intracellular compartments and has no interaction with EAAC1.
On the basis of these results, we arrived at the hypothesis that Arl6ip1 might regulate the amount of addicsin homodimer by formation of Arl6ip1-addicsin heterodimer on the cytoplasmic structure (Fig. 7). To test this idea, we searched for an addicsin mutant that lacked interaction with Arl6ip1 but not with addicsin itself. Finally, we noticed two amino acids at positions 110 and 112 of addicsin. These amino acids are located in the addicsin and Arl6ip1 association region (Fig. 2C) and are completely conserved among species from Drosophila to humans (Fig. 6A). Immunoprecipitation assay demonstrated that a double-mu-tated form of addicsin, designated addicsin Y110A/L112A (addicsin YL), had markedly less binding activity to Arl6ip1 (39.5 Ϯ 5.5% of wild-type addicsin), although it had the same binding activity to addicsin itself (113.9 Ϯ 9.6% of wild-type addicsin) (Fig. 6B). Furthermore, to investigate the effect of addicsin YL on EAAC1-mediated glutamate transport activity, we established the C6BU-1-pSw-addicsin YL cell line that exclusively up-regulated the V5-tagged addicsin YL (addicsin YL-V5) in a mifepristone-dependent manner (Fig. 6D, inset). Cell-surface biotinylation assay indicated that wild-type addicsin had a strong association with some molecules on the cell surface (Fig. 6C). However, in this cell line, addicsin YL had markedly less association with any molecules on the cell surfaces under all of the conditions tested (Fig. 6C), suggesting that addicsin YL had lost its binding ability to EAAC1. Interestingly, when cells were stimulated with 100 nM PMA, the glutamate uptake activity in C6BU-1-pSw-addicsin YL remained unchanged with increasing expression of addicsin YL-V5 (Fig. 6D,  lanes 3 and 4). By contrast, glutamate uptake activity was sustained at basal levels under all the conditions tested when the To examine the intracellular localization of Arl6ip1, immunohistochemistry was carried out in C6BU-1-pSw-Arl6ip1 cells using anti-V5 monoclonal antibody. Arl6ip1 appeared to be localized at the intracellular compartment. Scale bars correspond to 25 m. C, cell surface biotinyl assay in C6BU-1-pSw-Arl6ip1 cells. To investigate whether Arl6ip1 was associated with any cell surface molecules, cell surface biotinyl assay was carried out. No Arl6ip1 was associated with any molecules located on the plasma membrane. cells were stimulated with DMSO (Fig. 6D, lanes 1 and 2) or 100 nM 4␣-phorbol (Fig. 6D, lanes 5 and 6). These data suggest that the interaction of Arl6ip1 and addicsin may be an important element in the expression of enhanced EAAC1-mediated glutamate transport activity in a PKC activity-dependent manner.

DISCUSSION
Arl6ip1 was first identified as an ADP-ribosylation factorlike 6 (ARL-6)-association factor by yeast two-hybrid screening (28). Arl6ip1 was also isolated as a down-regulatory factor during myeloid differentiation by differential display (29). However, the biological functions of Arl6ip1 have not been fully elucidated. Amino acid analysis demonstrates that Arl6ip1 has two putative casein kinase II phosphorylation motifs (amino acids 18 -21 and 128 -131) and protein kinase C phosphorylation motifs (amino acids 94 -96, 115-117, and 128 -130) (data not shown). Arl6ip1 also contains an N-glycosylation motif (amino acid 6 -9), a prenyl group-binding motif (amino acids 72-75), and an endoplasmic reticulum retention signal (amino acids 200 -203). Consistent with these data, Arl6ip1 appeared to be localized in subcellular compartments, including the endoplasmic reticulum in C6BU-1 cells, according to our immunohistochemical analysis (Fig. 5B). Furthermore, ARL-6, an Arl6ip1-associated partner, belongs to the Ras superfamily of small GTP-binding proteins and is essential for membraneassociated intracellular trafficking processes (30). In agreement with this information, Arl6ip1 is structurally similar to addicsin (data not shown), which is thought to be localized in intracellular compartments that are implicated in intracellular trafficking events. Thus, this evidence strongly suggests that Arl6ip1, like addicsin, may be implicated in trafficking events, including protein transport, membrane trafficking, and cell signaling.
The C6BU-1 glioma cell is a suitable model for analyzing the molecular mechanisms of EAAC1-mediated glutamate transport, because only EAAC1 is endogenously expressed as a glutamate transporter (31). The glutamate transport activity of EAAC1 can be explained by either changing the number of transporters expressed on the plasma membrane or by changing the catalytic efficiency of the transporter located on the plasma membrane. In C6BU-1-pSw-Arl6ip1 cells, PMA stimulation elicited the acceleration of glutamate transport activity with an increase in affinity but not in maximal velocity once cells expressed Arl6ip1-V5 (Fig. 4D). These data indicate that Arl6ip1 alters the catalytic efficiency of EAAC1 located on the plasma membrane. In C6BU-1-pSw-Arl6ip1 cells, Arl6ip1 could not directly interact with EAAC1 on the plasma membrane and stayed in the subcellular compartment as a result of PMA treatment (Fig. 5, A-C), suggesting that Arl6ip1 indirectly modulated EAAC1-mediated glutamate transport activity. Previous reports have revealed that GTRAP3-18, a rat homologue of addicsin, inhibited EAAC1-mediated glutamate transport activity by direct interaction with EAAC1. Our immunoprecipitation assay demonstrated that the addicsin-interacting region is identical to the Arl6ip1-binding region in addicsin (Fig. 2, A-C). Moreover, the addicsin YL mutant, which had markedly less interaction ability with Arl6ip1, failed to promote EAAC1-mediated glutamate transport activity (Fig. 6,  B-D). These data therefore support the idea that activated FIGURE 6. Effect of binding ability of addicsin to Arl6ip1 on EAAC1-mediated glutamate transport activity. A, comparison of amino acid sequences in the second hydrophobic region of addicsin from different species. Accession numbers are human AB052638, mouse D87211, rat AF240182, zebrafish AB052637, and Drosophila AB052636. The asterisk indicates an identical amino acid from Drosophila to humans. B, binding activity of the addicsinY110A/L112A (addicsin YL) mutant to the wild type of Arl6ip1 or addicsin. To evaluate the binding ability of addicsin YL, immunoprecipitation (IP) assay was carried out using COS7 cell extracts co-expressing both addicsin YL and Arl6ip1 (upper right panel) or coexpressing both addicsin YL and addicsin (bottom right panel). The binding ability of addicsin YL to Arl6ip1 (upper left panel) or addicsin itself (bottom left panel) was quantified by Western blotting. The expression level of actin was used as an internal control. Data are means Ϯ S.E., n ϭ 3; *, p Ͻ 0.001, N.S., no significance. C, cell surface biotinyl assay in C6BU-1-pSw-addicsin and C6BU-1-pSw-addicsin YL cells.
To investigate whether the addicsin YL mutant could be associated with any cell surface molecules, cell surface biotinyl assay was carried out. Addicsin YL had markedly less binding activity to molecules located on the plasma membrane compared with wild type (WT) of addicsin. D, glutamate transport activity in C6BU-1-pSw-addicsin YL cells. Glutamate transport assay was performed on C6BU-1-pSw-addicsin cells under the indicated conditions. Addicsin YL had no influence on the EAAC1-mediated glutamate transport activity under any conditions, even when cells were exposed to 100 nM PMA. Data are means Ϯ S.E., n ϭ 5; N.S., no significance.
PKC might increase in addicsin-Arl6ip1 interaction, thereby decreasing the number of addicsin molecules available for interaction with EAAC1. The resulting decrease in the addicsin-EAAC1 complex causes a change in the catalytic efficiency of EAAC1 (Fig. 7).
The glutamate transport activity changes in the presence of PMA treatment in our experiments (Fig. 4, A and B), suggesting that PKC activation is essential for the dynamics of EAAC1-mediated glutamate transport. At present, this precise molecular mechanism remains unclear. However, EAAC1, addicsin, and Arl6ip1 themselves have putative PKC phosphorylation motifs (19,29,32). Thus, these molecules may be phosphorylated by PKC and modulate the function of EAAC1. Interestingly, the accumulated evidence reveals that PKC␣ regulates the redistribution of EAAC1 from the intracellular compartment to the plasma membrane and that PKC⑀ controls EAAC1-mediated glutamate transport by a trafficking-independent mechanism in C6BU-1 cells (33). Thus, in C6BU-1-pSw-Arl6ip1 cells, PKC⑀ may be involved in the enhancement of glutamate transport induced by increasing expression of Arl6ip1. However, further studies will be needed to elucidate the PKC-dependent modulation of EAAC1.
In summary, our biochemical and histological analyses show that Arl6ip1 can specifically interact with addicsin both in vitro and in vivo and bind a small portion of the hydrophobic region of addicsin, which is located at its amino acids 103-117. Furthermore, Arl6ip1 positively modulated EAAC1-mediated glutamate transport by increasing its affinity for glutamate, without any shift in maximal velocity, in a PKC activity-dependent manner. Arl6ip1 was localized in subcellular compartments and had no interacting ability to EAAC1 on cell surfaces in C6BU-1 cells. It is thus possible that Arl6ip1 is an addicsin-associated partner in vivo, and that Arl6ip1 promotes the catalytic efficiency of EAAC1 by reducing the interaction of addicsin with EAAC1 in a PKC activity-dependent manner.  FIGURE 7. Schematic presentation of the deduced regulatory mechanism of EAAC1-mediated glutamate transport. Addicsin can interact with either Arl6ip1 or addicsin itself. If addicsin makes a homodimer, addicsin can interact with EAAC1 and then negatively modulate EAAC1-mediated glutamate transport in a PKC activity-dependent manner. On the other hand, if addicsin makes a heterodimer with Arl6ip1, the number of addicsin homodimers available for interaction with EAAC1 decreases because Arl6ip1 is localized in cytoplasmic structures such as the endoplasmic reticulum. The resulting decrease in the addicsin-EAAC1 complex appears to cause changes in the catalytic efficiency of EAAC1.