A Novel System A Isoform Mediating Na+/Neutral Amino Acid Cotransport*

A cDNA clone encoding a plasma membrane alanine-preferring transporter (SAT2) has been isolated from glutamatergic neurons in culture and represents the second member of the system A family of neutral amino acid transporters. SAT2 displays a widespread distribution and is expressed in most tissues, including heart, adrenal gland, skeletal muscle, stomach, fat, brain, spinal cord, colon, and lung, with lower levels detected in spleen. No signal is detected in liver or testis. In the central nervous system, SAT2 is expressed in neurons. SAT2 is significantly up-regulated during differentiation of cerebellar granule cells and is absent from astrocytes in primary culture. The functional properties of SAT2, examined using transfected fibroblasts and in cRNA-injected voltage-clamped Xenopus oocytes, show that small aliphatic neutral amino acids are preferred substrates and that transport is voltage- and Na+-dependent (1:1 stoichiometry), pH-sensitive, and inhibited by α-(methylamino)isobutyric acid (MeAIB), a specific inhibitor of system A. Kinetic analyses of alanine and MeAIB uptake by SAT2 are saturable, with Michaelis constants (K m ) of 200–500 μm. In addition to its ubiquitous role as a substrate for oxidative metabolism and a major vehicle of nitrogen transport, SAT2 may provide alanine to function as the amino group donor to α-ketoglutarate to provide an alternative source for neurotransmitter synthesis in glutamatergic neurons.

System A is widely expressed in mammalian cells, where it mediates Na ϩ -coupled cellular uptake of small aliphatic amino acids, with alanine, serine, and glutamine being particularly good substrates. It is distinguished from other neutral amino acid transporters like systems L, ASC, and N by the fact that it recognizes N-methylamino acids such as ␣-(methylamino) isobutyric acid (MeAIB), 1 does not tolerate substitution of Li ϩ for Na ϩ , and exhibits trans-inhibition (1)(2)(3). An additional important feature of system A is that it is the major amino acid system subject to regulation by environmental conditions, proliferative stimuli, developmental changes, and hormonal signals (4 -6). Such regulation is thought to be directed at system A, because its substrates play key roles in the overall flux of amino acids between tissues and in delivering energy to amino acid metabolism.
In the intestine, system A is localized to basolateral membranes where it plays an important role in the generation of ␣-ketoglutarate, the preferred metabolic fuel of enterocytes over glucose (7). High intracellular alanine levels generated by system A in enterocytes provide the driving force for glutamine uptake through coupled exchange via system L. Glutamine is then converted to glutamate through phosphate-activated glutaminase (PAG), with production of NH 4 ϩ , which exits the basolateral membrane. Aminotransaminase converts glutamate and pyruvate to alanine and ␣-ketoglutarate. Alanine then exits via system L in exchange for additional glutamine. System A, therefore, plays a key role in regulating the availability of essential oxidative metabolites in enterocytes. Alanine is a particularly important nitrogen donor via alanine aminotransferase in peripheral tissues, e.g. the liver and skeletal muscle (8,9), and in the brain (10). In the liver, system A transports alanine to provide carbon atoms for gluconeogenesis and nitrogen for urea production (8,9). System A in liver is known to be up-regulated by glucagon and insulin, facilitating the conversion of amino acids to glucose and stimulating urea nitrogen production (11). In the heart and muscle, system A, together with the glutamine transport system N, likely plays an important role in the synthesis of oxidative fuel.
The different roles alanine plays in different tissues (8,9), the different factors (insulin, glucagon, amino acid deprivation, hyperosmotic stress, etc.) that affect the regulation of system A in different tissues (4 -6), the differences in substrate preferences (e.g. glutamine, proline, glycine) in different tissues (12)(13)(14), and the variants that emerge in transformed cells (15,16) strongly suggest that a multiplicity of discrete system A isoforms exist.
We have recently cloned and functionally identified a neuronal glutamine transporter (GlnT), the first member of the system A family of transporters (17). GlnT was cloned based upon its weak similarity to a family of plasma membrane transporters that are found in plants, yeast, and Caenorhabditis elegans amino acid/auxin permease (AAAP) and to the transporter that mediates uptake of inhibitory amino acid neurotransmitters (VIAAT/VGAT) into synaptic vesicles (18 -20). GlnT expression occurs predominantly in nervous tissue and is expressed on the plasma membrane of glutamatergic and some (GABA)ergic (e.g. cerebellar Purkinje cells) neurons. GlnT is a highly efficient glutamine transporter and may be a physiologically important gateway for the precursor of neurotransmitter glutamate via the glutamate/glutamine cycle. Examination of the transport properties of GlnT revealed that not all classical system A substrates (e.g. proline and glycine) could compete with [ 14 C]MeAIB for uptake. In addition, GlnT expression was not found in peripheral organs known to express system A. This indicated that additional members of system A exist and may be expected to express different substrate preferences and display unique tissue specificity.
Here, we report the cloning and characterization of SAT2, the second member of the system A family of neutral amino acid transporters. SAT2 displays a widespread distribution and exhibits broad specificity for neutral amino acids with a preference for alanine. SAT2 is also enriched on glutamatergic neurons in the brain, where it may be an important gateway for alanine uptake to serve, in part, as an amino group donor and alternative precursor of transmitter glutamate. Because alanine plays a key role as a vehicle of nitrogen transport and as an end product of nitrogen catabolism in many tissues, the identification SAT2 will facilitate molecular studies on its functional roles and regulation in various physiological and pathological states.

EXPERIMENTAL PROCEDURES
Primary Cultures-Cerebellar primary cultures were prepared from 8-day postnatal (P8) Harlan Sprague-Dawley rats as described (21) with minor modifications. Cerebella were removed from decapitated rat pups, minced, and digested with trypsin (0.25 mg/ml) for 15 min at 37°C in HEPES-buffered Lebowitz medium containing 0.3% bovine serum albumin. The cell suspension was then placed into Dulbecco's modified Eagle's medium (DMEM) with 5% fetal bovine serum (FBS) and centrifuged at 300 ϫ g for 10 min at 4°C. This treatment was followed by trituration with a Pasteur pipette in DNase I (50 g/ml) containing DMEM/5% FBS. The cell suspension was allowed to settle for 15 min at room temperature, and the supernatant was transferred to a centrifuge tube on ice. This trituration step was repeated two more times, and the combined supernatants were centrifuged as above. The cell pellet was resuspended in DMEM containing 10% fetal calf serum, 2 mM glutamine, and 100 g/ml penicillin/streptomycin and plated at a density of 15 or 3.75 ϫ 10 5 cells (neuronal and astrocytic cultures, respectively) in 65 mm-diameter plastic dishes coated with 10 g/ml polyornithine. The medium for neuronal cultures also contained 25 mM KCl (final concentration). After 18 h, 10 M cytosine arabinoside was added to the neuronal cultures to prevent the replication of non-neuronal cells. The astrocyte culture medium was changed at day 3 in vitro and every other day afterward. The neuronal culture medium was 50% refreshed at day 7.
Cloning of SAT2-An oligo(dT)-primed, size-selected cDNA library from cerebellar granule cells (10 days in culture) that was constructed in CDM7/amp, a modified T7 promoter bearing plasmid expression vector, was used (17). Successive subdivisions of the library were screened with a cDNA probe made from a 319-base pair fragment of expressed sequence tag (EST) 184296 amplified by polymerase chain reaction (PCR) using the following two primers, 5Ј-GGCAACTCATAT-TTCACTATGAAGAGGTAGC-3Ј and 5Ј-CCAGGTACTACTTCCTTTG-GAATGTCAGTA-3Ј from a human colon cDNA library, gel-purified, and 32 P-labeled using [ 32 P]dCTP (NEN Life Science Products) by random priming (Bio-Rad). This fragment corresponds to the conserved putative first membrane-spanning segment of the yeast AAAP homologues (20) and to region I of VIAA/VGAT (18,19). Southern blots of EcoRI restriction digests of plasmid prepared from overnight cultures were hybridized overnight at 45°C in a buffer containing 5ϫ SSC; 25% formamide; 5 ϫ Denhardt's solution; 50 mM NaPO 4 (pH 6.5); 0.1% SDS; 250 g/ml tRNA and the denatured radiolabeled cDNA probe (10 6 cpm/ ml). The filters were washed at 1ϫ SSC; 0.1% SDS at 60°C. Autoradiographs were analyzed using a BAS2000 phosphor-imaging system (Fuji Biomedical) after 12-h exposure. A 4.2-kilobase (kb) cDNA was obtained and subcloned into pUC18, and overlapping fragments were sequenced in both directions with the Thermo Sequenase Cycle sequencing kit (Amersham Pharmacia Biotech) according to the manufacturer's instructions.
Preparation of SAT2 Antibody-The coding region for the N-terminal hydrophilic portion of SAT2 (amino acids 1-65) was amplified by PCR using the primers, 5Ј-CGGGATCCATGAAGACCGAAATGGGAAGGT-C-3Ј and 5Ј-GGAATTCCTAGTCTGTTTCGTACTTCTTCTTCCCCAGA-T-3Ј, which include 5Ј-appended BamHI and EcoRI sites (underlined) and an internal stop (italics). This fragment was subcloned into the bacterial expression vector pGEX-KT, and the resulting plasmid was transfected into BL21 cells. Recombinant glutathione S-transferase fusion protein was isolated using a purification module (Amersham Pharmacia Biotech), and then the N terminus of SAT2 was released following thrombin cleavage according to the manufacturer's instructions. Polyclonal antibodies were produced using the resulting 6-kDa N terminus of rat SAT2 in rabbits by Macromolecular Resources at Colorado State University.
Northern Analysis-Poly(A) ϩ RNA was purified from different rat tissues by guanidine isothiocyanate extraction and ultracentrifugation through a cesium trifluoroacetic acid cushion followed by a single round of oligo(dT)-cellulose chromatography (22). 5 g of RNA (A 260/280 Ͼ 1.9) was electrophoresed through formaldehyde-agarose gels and stained with ethidium bromide to assure that equivalent amounts of RNA were loaded into each lane. The RNA was then electroblotted onto nylon membranes and hybridized with a oligonucleotide probe (48-mer) against 3Ј non-coding sequences of SAT2 (bases 1926 -1973) radiolabeled using terminal deoxynucleotidyltransferase (Life Technologies, Inc.) and [ 32 P]dATP (NEN Life Science Products) or random-primed full-length SAT2 cDNA in buffer containing 5ϫ SSC, 50% formamide, 5ϫ Denhardt's solution, 50 M NaPO 4 (pH 6.5), 0.1% SDS, 250 g/ml tRNA for 18 h at 45°C. The filters were washed in 1ϫ SSC, 0.1% SDS at 60°C (oligo probe) or 0.1ϫ SSC, 0.1% SDS at 55°C (cDNA probe) and exposed to Kodak X-Omat film with an intensifying screen at Ϫ70°C.
In Situ Hybridization-Sagittal fresh-frozen rat brain sections (15 m) were fixed on slides in 4% formaldehyde in phosphate-buffered saline (PBS) for 15 min at room temperature followed by three 5-min washes in PBS. The slides were then immersed in 0.1 M triethanolamine, pH 8, and acetic anhydride (0.25% final) was added under fast stirring. After 10-min incubation, the sections were rinsed in diethylpyrocarbonate-treated water, dehydrated, and delipidated as follows: 70% EtOH for 1 min, 96% EtOH for 1 min, twice 100% EtOH for 2 min, twice chloroform for 5 min, 100% EtOH for 2 min, 96% EtOH for 1 min, and then air-dried.
The SAT2 cRNA hybridization probe was prepared as follows: a KpnI-HindIII 3Ј-non-coding fragment of SAT2 (bases 2583-2913) was subcloned into pBluescript KS (Stratagene) and linearized with KpnI and HindIII to generate riboprobes in antisense (T7 polymerase-catalyzed transcription) and sense (T3 polymerase-catalyzed transcription) orientation, respectively. In vitro transcription was performed using 250 ng of linearized plasmid, 62 Ci of [ 35 S]dUTP (NEN Life Science Products), and the Riboprobe Combination System T3/T7 (Promega) according to the manufacturer's instructions.
For hybridization, the sections were covered with hybridization mix containing 80,000 dpm/l 35 S-labeled cRNA probe in hybridization buffer (50% formamide, 20 mM Tris-HCl, pH 7.4, 1 mM EDTA, 315 mM NaCl, 10% dextran sulfate, 1ϫ Denhardt's, 0.1 mg/ml salmon sperm DNA, 0.25 mg/ml yeast tRNA, 0.25 mg/ml yeast total RNA, 10 mM dithiothreitol (DTT), 0.1% sodium thiosulfate, 0.1% sodium dodecyl sulfate) and then coverslipped. After hybridization for 18 h in a humid chamber placed in a 56°C convection oven, coverslips were removed in 1ϫ SSC/1 mM DTT. Sections were washed twice for 15 min in 1ϫ SSC/1 mM DTT; treated with 20 g/ml RNase A and 1 unit/ml RNase T in 0.5 M NaCl, 10 mM Tris-HCl, pH 8, 1 mM EDTA for 30 min at 37°C; and then washed successively in 1, 0.5, and 0.2ϫ SSC/1 mM DTT for 15 min each at room temperature and twice in 0.2 ϫ SSC/1 mM DTT for 20 min at 60°C. The slides were briefly dipped in water then in 70% ethanol, air-dried, and exposed to Kodak Biomax x-ray film for 96 h.
Membrane Preparations and Western Analysis-For primary cultures, cells were rinsed with PBS, harvested by scraping in PBS, centrifuged, and resuspended in PBS containing 1 mM phenylmethylsulfonyl fluoride, 5 mg/l pepstatin, 5 mg/l leupeptin, and 5 mg/l aprotinin, sonicated, and centrifuged at 2000 ϫ g for 5 min. The resulting supernatants were kept frozen at Ϫ80°C until further use. 5 g of SDSsolubilized cell extracts were size-fractionated under reducing conditions on 8% acrylamide gels and electrophoretically transferred to an nitrocellulose membrane using standard protocols. SAT2, synaptophysin (p38), and glial fibrillary acidic protein were detected using the respective primary antibodies, horseradish peroxidase-conjugated secondary antibodies (Sigma) followed by exposure to film.

Transient Expression and Transport Assay in Mammalian Cells-
Monkey kidney fibroblasts (CV-1 cells) were plated on collagen-coated 12-well dishes (2 ϫ 10 5 cells/well) in DMEM containing 10% FBS, penicillin (100 units/ml), streptomycin (100 mg/ml), and glutamine (4 mM). Cells were rinsed with DMEM without serum, antibiotics, or glutamine, infected with recombinant T7 vaccinia virus at 10 plaqueforming units/cell for 30 min, and then transfected with plasmid containing SAT2 cDNA in CDM7/amp (1 g/ml) by lipofection in the same medium (23). After  C]MeAIB. Cells were placed in a 37°C incubator for 2.5 min, and uptake was terminated with a 2.5-ml wash in 4°C buffer on ice. For Na ϩ -free media, NaCl was replaced with equimolar choline Cl or LiCl. For Cl Ϫ -free media, NaCl was replaced with equimolar NaNO 2 or sodium gluconate. Cells were then solubilized in 1 ml of 1% SDS, and radioactivity was measured by scintillation counting in 5 ml of EcoScint (National Diagnostics). Transport measurements were performed in duplicate and repeated at least three times using independent infection/transfections. All experimental conditions with SAT2-transfected cells had corresponding mock-transfected cells in adjacent wells.
Functional Characterization of SAT2 in Xenopus Oocytes-Oocytes were isolated from Xenopus laevis (under 2-aminoethylbenzoate anesthesia), treated with collagenase A (Roche Molecular Biochemicals) and stored at 18°C in modified Barths' medium (24). SAT2 cDNA was subcloned into the pTLNii vector under SP6, and SAT2 cRNA was synthesized in vitro using the Ambion mMESSAGE mMACHINE 228 kit. Oocytes were injected with 50 ng of SAT2 cRNA and incubated 2 days before performing radiotracer or voltage-clamp experiments at 22°C. Standard Na ϩ uptake medium comprised 100 mM NaCl, 2 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES (pH 7.5 with Tris base). For Na ϩ -free or low Na ϩ media, NaCl was replaced by equimolar choline chloride. For pH sensitivity experiments, Na ϩ media were buffered at pH 6.5-8.6 using 0 -5 mM MES, 0 -5 mM HEPES, and 0 -5 mM Tris base.
A two-microelectrode voltage clamp was used to measure currents in control (water-injected) oocytes and oocytes injected with SAT2 cRNA. Microelectrodes (resistance, 0.5-3 M⍀) were filled with 3 M KCl. Oocytes were clamped at a holding potential (V h ) of Ϫ50 mV and step-changes in membrane potential (V m ) applied (from ϩ50 to Ϫ150 mV in 20-mV increments), each for a duration of 100 ms, before and after the addition of substrate; current was low-pass-filtered at 500 Hz and digitized at 5 kHz. Test solutions were washed out with substrate-free medium (100 mM choline chloride) at pH 7.5 for several minutes. Steady-state data (obtained by averaging the points over the final 16.7 ms at each V m step) were fitted to Eq. 11.3 of Mackenzie (24), for which I is the evoked current (i.e. the difference in steady-state current measured in the presence and absence of substrate), I max the derived current maximum, S the concentration of substrate S (Na ϩ or amino acid), K 0.5 S the substrate concentration at which current was half-maximal, and n H the Hill coefficient for S. Additionally, continuous current monitoring was performed at V h ϭ Ϫ70 mV, for which the current was low-pass-filtered at 20 Hz (sampling at 1 Hz).
Na ϩ /L-alanine coupling stoichiometry was determined by direct comparison of net inward charge with [ 3 H]alanine accumulation in individual oocytes under voltage clamp (24,25). Oocytes were clamped at V h ϭ Ϫ70 mV and superfused with standard Na ϩ medium (pH 7.5) plus 500 M L-[ 3 H]alanine (NEN Life Science Products; final specific activity, 1-2 MBq⅐mmol Ϫ1 ) for 5 min before washing out with Na ϩ medium. The alanine-evoked current was integrated with time to obtain the alaninedependent charge (Q Ala ) and converted to a molar equivalent using the Faraday conversion. Oocytes were solubilized with 10% SDS, and the 3 H content was measured by liquid scintillation counting.

Cloning and Structural
Features of SAT2-Our cloning strategy began with a search of the EST data base for distant mammalian homologues of the vesicular GABA/glycine transporter. A human colon EST was found that displayed moderate sequence identity (25-45%) within the first putative transmembrane domain (TMD) to a large family of putative amino acid permeases in the AAAP family of plants, yeast, and C. elegans and in mammals. This EST was homologous to, but differed significantly from, the degenerate oligonucleotide probe used to clone GlnT (17). We screened a rat cDNA library prepared from glutamatergic neuronal cultures with PCR-amplified sequences from this human EST that we obtained from a human colon cDNA library. A cDNA clone with an open reading frame of 1514 bp predicting a highly hydrophobic protein of 504 amino acids and a molecular mass of 55,554 daltons was obtained (Fig. 1). Hydropathy analysis suggests 11 hydrophobic membrane-spanning segments (TMD). The absence of a signal sequence for membrane insertion in the N terminus suggests that it is retained in the cytoplasm leaving the short C terminus located extracellularly. The granule cell cDNA clone has two potential sites for N-linked glycosylation between putative TMDs 5 and 6. Consensus sequences for phosphorylation by protein kinase C exist on the predicted cytoplasmic domain between putative TMDs 6 and 7. Interestingly, SAT2 displays significant homology (approximately 50 -55% identity) to both rat GlnT and to SN1/NAT, a system N transporter (26,27).
Distribution of SAT2-By Northern analysis, the mRNA (ϳ4.8 kb) was found to be expressed in most tissues examined, including heart, adrenal gland, skeletal muscle, stomach, fat, brain, spinal cord, colon, and lung, with lower levels detected in spleen, and was absent from testis (Fig. 2). Surprisingly, SAT2 mRNA was present at very low levels in liver, if at all. Equivalent levels of RNA loaded in each well and the lane assignments were confirmed by ethidium staining and by probing with GlnT, a neuron-enriched system A isoform, and with a recently identified homologous isoform specifically enriched in liver (data not shown). In addition, an independent Northern blot was reprobed with SAT2, and no expression in liver was observed. To avoid possible cross-hybridization with homologous family members of system A, Northern blots were probed with a 3Ј-non-coding oligonucleotide as well as with a fulllength cDNA probe washed at high stringency with identical results. Interestingly, the areas containing the most SAT2 mRNA were regions containing neurons (brain and spinal cord) and neuroendocrine cells (adrenal) as well as the intestine and kidney. In situ hybridization of rat brain sections revealed strong labeling in the pyriform cortex, hippocampus, and cerebellar granule layer, regions of the brain that contain neurons that use glutamate as their transmitter (Fig. 3A). Specificity of labeling was confirmed using the sense SAT2 probe (Fig. 3B). In contrast to GlnT, SAT2 mRNA was not observed in cerebellar Purkinje neurons. In situ hybridization of rat spinal cord sections with SAT2 revealed strong labeling of cholinergic motor neurons (data not shown). We also examined the expression of SAT2 in primary cultures of rat cerebellar granule cells, which are a good model for glutamatergic neurons. High levels of expression are observed in neuronal cultures, whereas parallel astrocyte cultures are devoid of immunoreactivity (Fig.  3C). In the neuronal cultures, specific immunoreactivity increases concomitantly with the morphological differentiation of the cells (Fig. 3D). The stimulus-coupled release of glutamate also develops gradually and concomitantly with neuronal differentiation in these cultures (21).
Functional Identification of SAT2 in Mammalian Cells-To determine whether SAT2 was a system A transporter like GlnT (17) or a system N transporter like SN1/NAT (26,27), the cDNA was transiently expressed in fibroblasts using the vaccinia virus/bacteriophage T7 hybrid system and the cultures were incubated with [ 14 C]MeAIB (Fig. 4). The time course of SAT2-specific uptake of [ 14 C]MeAIB (10 M) became saturated by 5 min and remained stable for 20 min. The total uptake of [ 14 C]MeAIB was approximately 5-fold greater in the transfected cells (29 pmol/well) than in the mock-transfected cells (6 pmol/well) at 2.5 min of incubation. Kinetic analysis of MeAIB uptake by SAT2 at pH 7.4 was saturable, with a Michaelis constant (K m ) of 530 Ϯ 53 M; n ϭ 9 (Fig. 4A). The transporter encoded by SAT2 has an absolute requirement for extracellular Na ϩ ions, because choline and lithium cannot effectively substitute for it in transfected cells (Fig. 4B). Replacement of chloride with nitrate or acetate resulted in uptake changes of less than 30%, suggesting that anions may not be thermodynamically coupled to [ 14 C]MeAIB uptake. A characteristic feature of SAT2 is its pH sensitivity. [ 14 C]MeAIB uptake mediated by SAT2 is strongly pH-sensitive with uptake at pH 8.2 being approximately six times greater than that seen at pH 6.6 ( Fig.  4C). To further define the characteristics of SAT2 we tested the ability of various amino acids (1.5 mM) to compete with [ 14 C]MeAIB (10 M) for uptake (Fig. 4D). Alanine is the preferred substrate for SAT2. Serine, proline, and methionine are also effective [ 14 C]MeAIB uptake competitors. Asparagine, glutamine, glycine, and histidine are less effective inhibitors, whereas arginine, ␤-alanine, glutamic acid, isoleucine, leucine, lysine, phenylalanine, threonine, tryptophan, and valine are ineffective inhibitors. MeAIB inhibition of [ 14 C]MeAIB uptake was approximately 80% at 1.5 mM concentration. A direct comparison of alanine and glutamine transport by SAT2 was also determined (Fig. 4E). Kinetic analysis indicate that the affinity of alanine for SAT2 was similar to MeAIB (K m ϭ 529 Ϯ 50 M; V max ϭ 656 Ϯ 35 nmol/min/well; n ϭ 4) but significantly greater than for glutamine (K m ϭ 1.65 Ϯ 0.27 mM; V max ϭ 389 Ϯ 40 nmol/min/well; n ϭ 5).
Functional Characterization of SAT2 in Xenopus Oocytes-The functional characteristics of SAT2 were also explored by applying radiotracer and voltage-clamp techniques in Xenopus oocytes expressing SAT2. SAT2 mediates saturable, Na ϩ -dependent amino acid transport in a rheogenic manner. In the presence of Na ϩ , although L-alanine evoked currents of ϽϪ3 nA in control oocytes (see also Ref. 28), L-alanine evoked reversible inward currents of up to Ϫ1000 nA in oocytes expressing SAT2 (Fig. 5A). Smaller currents were obtained in response to L-glutamine, and the current evoked by MeAIB was only about 25% of the L-alanine-evoked current. SAT2 stimulated  Eleven putative transmembrane domains, potential sites for N-linked glycosylation (three-pronged branches) and phosphorylation by protein kinase C (P inside the circle) are indicated. White balls (with black letters) represent conserved residues between SAT1 and SAT2. Black balls (with white letters) represent amino acids unique to SAT2. The cytoplasm is shown above and the extracellular space below; Ϫ, represents acidic residues; ϩ, basic residues. L-[ 3 H]alanine (500 M) uptake 25-fold (61 Ϯ 16 pmol⅐min Ϫ1 , mean Ϯ S.E.) over control oocytes (2.4 Ϯ 0.5 pmol⅐min Ϫ1 ) and [ 14 C]MeAIB uptake (200 M) 5-fold to 0.9 Ϯ 0.07 pmol⅐min Ϫ1 (n ϭ 6 -12 oocytes in each group). MeAIB evokes a tiny outward current in native oocytes (28) and accelerates endogenous ala-nine transport (29), so that a phenomenon specific to the oocyte expression system may, in part, account for why MeAIB fluxes and currents were significantly smaller than those for L-alanine despite the similarity of the K m values for these substrates observed in transfected CV-1 cells. The substrate selectivity of SAT expressed in oocytes (data not shown) reflected the pattern observed using the T7 vaccinia system, and SAT2 was stereospecific for the L-isomer, at least in the case of alanine.
In oocytes expressing SAT2, the L-alanine-evoked currents (at saturating amino acid) showed a curvilinear dependence on membrane potential (V m ) (Fig. 5B). The current/voltage relationship was roughly linear between Ϫ150 and Ϫ30 mV, and no reversal of the currents was observed up to ϩ50 mV. The L-alanine-evoked currents followed Michaelis-Menten saturation kinetics. At Ϫ70 mV, K 0.5 Ala (L-alanine concentration at which currents were half-maximal) was around 200 M (Fig.  5C) with a Hill coefficient (n H ) for alanine of Ϫ1. K 0. 5 Ala was largely independent of V m , except at depolarized potentials (more positive than Ϫ10 mV), rising to 1 mM at ϩ30 mV.
The current-voltage relationships for subsaturating (200 M) L-alanine were determined as a function of extracellular Na ϩ concentration (Fig. 5F). At any given V m , the currents were larger at higher [Na ϩ ] o , indicating that L-alanine transport mediated by SAT2 is driven by the Na ϩ electrochemical gradient. The saturation kinetics for Na ϩ were determined as a function of V m at 0.2 mM L-alanine. At Ϫ70 mV, K 0.5 Ala was 20 mM and the Hill coefficient (n H ) for Na ϩ was Ϫ1 (Fig. 5G). n H for Na ϩ did not vary with V m , however, K 0.5 Ala exhibited significant voltage dependence, rising from Ͻ10 mM at hyperpolarized V m to almost 50 mM at ϩ50 mV (Fig. 5H). These data suggest that only one Na ϩ binds to SAT2 (or that there is no cooperativity between the binding of multiple Na ϩ ) and that Na ϩ binding is voltage-sensitive.
The Na ϩ /L-alanine coupling stoichiometry was directly determined by comparing the alanine-dependent charge Q Ala (converted to a molar equivalent assuming monovalence) with  4. Functional identification of SAT2 as a system A transporter. A, kinetic analysis of uptake of MeAIB by SAT2 in transiently transfected CV-1 cells. Saturation isotherm of initial velocity (2.5 min) of MeAIB (0.1-3.2 mM) uptake by SAT2. Inset, Lineweaver-Burk analysis of SAT2 initial uptake velocity. B, MeAIB uptake by SAT2 is Na ϩ -dependent and Li ϩ -intolerant. Transport of [ 14 C]MeAIB (10 M) was measured at 2.5 min in NaCl-containing buffer (control, 100%), Na ϩ -deprived buffer (replaced by Li ϩ or choline), or Cl Ϫ -deprived buffer (replaced by gluconate or NO 2 ). Error bars represent the standard error of the mean (n ϭ 3-9). C, MeAIB uptake by SAT2 is strongly pH-sensitive.  (Fig. 5I). Q Ala correlated with 3 H accumulation in a linear fashion (with a slope of 1.04), justifying our use of the amino acid-evoked current as a direct index of amino acid transport. The Na ϩ /L-alanine coupling coefficient (n) was 1.19 Ϯ 0.09, and the mean Q Ala was not significantly different from the 3 H accumulation in a paired t test, indicating that Na ϩ -coupled, L-alanine transport in SAT2 proceeds with 1:1 stoichiometry. Unlike the Na ϩ /glucose cotransporter SGLT1 (24), there was no appreciable slippage (uncoupled current or uniport) of Na ϩ , because the shift in inward current as a result of switching from 0 to 100 mM Na ϩ in oocytes expressing SAT2 (Fig. 5A) was not significantly different from the current observed in control oocytes. DISCUSSION We have cloned and functionally identified SAT2, the second member of the system A family of neutral amino acid transporters. SAT2 shares 55% identity with GlnT (17), which we renamed SAT1. In addition, SAT2 shares 52% identity with SN1/NAT, a recently identified system N transporter expressed predominantly in liver but also found in brain and kidney (26,27). The functional properties of the transport activity of SAT2 observed in transfected mammalian cells indicate that it is alanine-preferring, capable of transporting MeAIB and small aliphatic amino acids, and intolerant of Li ϩ substitution for Na ϩ . These characteristics are discriminating features of system A and system N in tissue-cultured cells. Additional properties of SAT2 revealed in this study include voltage dependence (1:1 Na ϩ stoichiometry) and pH sensitivity consistent with previously described properties of system A (1-6). Although both systems A and N exhibit Na ϩ dependence and are pHsensitive, SN1/NAT has been identified as a proton antiporter and appears to behave as a glutamine-gated Na ϩ /H ϩ exchanger (26). Thus, changes in intra-or extracellular pH will affect the direction of flux through SN1/NAT. Changes in extracellular pH significantly increase the efficiency of both SAT1 Ala ϭ 0.19 Ϯ 0.03 mM, I max Ala ϭ Ϫ284 Ϯ 11 nA, and n H for L-alanine was 0.9 Ϯ 0.1 (r 2 ϭ 0.99). The Hill coefficient (n H ) for L-alanine did not differ significantly at other V m (data not shown). The voltage dependence of K 0.5 Ala (D) and I max Ala (E) are also presented. F, I/V relationships for 0.2 mM L-alanine as a function of Na ϩ concentration. G and H, Na ϩ saturation kinetics were determined from the currents evoked by subsaturating (0.2 mM) L-alanine at 4, 5, 10, 20, 30, 50, 60, 80, and 100 mM NaCl. At V m ϭ Ϫ70 mV (G), the derived kinetic parameters were K 0.5 Na ϭ 20.2 Ϯ 1.5 mM, I max Na ϭ Ϫ96 Ϯ 3 nA, and n H for L-alanine was 1.1 Ϯ 0.1 (r 2 ϭ 1.0). The Hill coefficient for Na ϩ did not significantly vary with V m (data not shown). The relationship of K 0.5 Na to V m is shown in H. I, Na ϩ /L-alanine coupling coefficient determined by L-[ 3 H]alanine uptake under voltage clamp. Oocytes were clamped at Ϫ70 mV, and 500 M L-[ 3 H]alanine was applied for 5 min. The alanine-dependent charge (Q Ala ) was compared with concomitant tracer accumulation in the same oocytes expressing SAT (filled circles). Charge and tracer accumulation in control oocytes (mean Ϯ S.E., bidirectional, n ϭ 3 oocytes; open circle) was first subtracted. Linear regression of SAT data (r 2 ϭ 0.43) gave a slope of 1.04. The ratio of Q Ala : 3 H accumulation was 1.19 Ϯ 0.09 (mean Ϯ S.E., n ϭ 6 oocytes), and there was no significant difference between Q Ala and 3 H accumulation (inset) according to Student's paired t test (t ϭ 1.9). and SAT2 function by affecting the V max of transport. However, the direct measurement of the 1:1 stoichiometry of Na ϩ :alanine transport by SAT2 observed in voltage-clamped Xenopus oocytes argues against protons being involved in the thermodynamic coupling of system A-mediated transport.
Both SAT1 and SAT2 are capable of transporting MeAIB, yet they exhibit differences with respect to their substrate selectivity and their distribution in various tissues. Although both recognize alanine as a substrate, SAT1 is a more efficient glutamine transporter than SAT2. The efficiency (V max /K m ) of alanine transport by SAT2 is approximately four times greater than that of glutamine. SAT2 effectively recognizes proline compared with SAT1. SAT1 is expressed predominantly in brain and spinal cord and is absent from peripheral tissues like heart, muscle, kidney, and lung where SAT2 mRNA is present. The expression of SAT2 in the intestine is consistent with its role in regulating the availability of oxidative metabolites. Preliminary in situ hybridization data suggest the mRNA signal in kidney resides in the medulla rather than in cortex (data not shown), but the precise localization of SAT2 in kidney has not yet been determined. It is possible that it transports amino acids as osmolytes into renal cells in response to hyperosmotic stress (30). SAT2 mRNA is not observed in liver, a tissue where both insulin and glucagon stimulate system A activity to facilitate the conversion of amino acids to glucose and to stimulate urea nitrogen production. A related article on this subject reports on the sequence ATA2 that is identical to SAT2, with a Northern blot showing dramatic expression in the liver (31). This discrepancy in tissue distribution may be related to the higher stringency conditions used for Northern analysis in the present study. The existence of additional distinct system A isoforms has been suggested in the literature (15,16), and the unique role system A plays in liver is consistent with the concept that distinct isoforms exist to allow appropriate regulation, for example in response to a high protein diet or glucose deficiency. Alternatively, SAT2 expression in hepatocytes might be tightly regulated and intimately dependent upon the metabolic/functional disposition of the animal.
SAT2 mRNA is enriched in glutamatergic neurons in the brain, is selectively expressed in neuronal-rich cultures and not in astrocyte-rich cultures, and is up-regulated during neuronal differentiation. These results suggest that SAT2 may also play an important role in glutamatergic function in addition to SAT1. Although glutamine is the preferred precursor for neurotransmitter glutamate, it cannot be the only source because the amount of glutamate released exceeds the amount of glutamine that enters glutamatergic neurons (32). In addition, PAG is present at much lower levels in some glutamatergic pathways than in others (33), further suggesting that alternative sources for transmitter glutamate must exist. It is well recognized that glutamatergic neurons depend upon glia for maintenance of their neurotransmitter pool (34 -42). After release from neurons during neurotransmission, glutamate is rapidly cleared from the synaptic region by glutamate transporters that are expressed predominantly on astrocytes (43,44). There, glutamate is partly converted to glutamine by the astrocyte-specific enzyme glutamine synthetase (45). Recently, a system N transporter (SN1/NAT) that acts as a Na ϩ /H ϩgated glutamine effluxer was identified and is restricted to glial cells in the rat brain (26,27). SAT1, a glutamine-preferring system A transporter, is expressed on glutamatergic neurons and is absent from astrocytes (17). The identification of systems A and N in the brain supports the notion of the glutamate/glutamine cycle in the recycling of transmitter glutamate. Thus, glutamine may exit astrocytes via SN1/NAT and enter glutamate neurons via SAT1 to be made available to PAG, a mitochondrial enzyme (46 -49), for the synthesis of the neurotransmitter pool. A second major metabolic fate of glutamate once accumulated by astrocytes is a reaction with pyruvate to form alanine via the alanine aminotransferase pathway (50). ␣-Ketoglutarate together with alanine (an amino group donor) have been shown to be efficiently converted to transmitter glutamate via an aminooxyacetic acid-inhibitable transamination reaction and can maintain the release of glutamate in cerebellar granule cells and in brain slices (51,52). The major synthesis of the tricarboxylic acid intermediates such as ␣-ketoglutarate also occurs in astrocytes, because only they possess pyruvate carboxylase and are able to perform carbon dioxide fixation (53)(54)(55). Thus, neuronal ␣-ketoglutarate and alanine, in addition to glutamine, are supplied by astrocytes (56 -58) and are actively accumulated into neurons (10, 14, 50, 59 -62). The astrocyte-to-neuron metabolic shuttle via SN1/NAT, SAT1, and SAT2 may enable not only replenishment of neurotransmitter glutamate (and GABA) but also may provide neurons, in general, with the capacity for oxidative metabolism and ATP generation.