Transporters of the SLC6 (solute carrier 6) family play an important role in the removal of neurotransmitters in brain tissue and in amino acid transport in epithelial cells. In the present study, we demonstrate that mouse v7-3 (slc6a15) encodes a transporter for neutral amino acids. The transporter is functionally and sequence related to B0AT1 (slc6a19) and was hence named B0AT2. Leucine, isoleucine, valine, proline and methionine were recognized by the transporter, with values of K0.5 (half-saturation constant) ranging from 40 to 200 μM. Alanine, glutamine and phenylalanine were low-affinity substrates of the transporter, with K0.5 values in the millimolar range. Transport of neutral amino acids via B0AT2 was Na+-dependent, Cl−-independent and electrogenic. Superfusion of mouse B0AT2-expressing oocytes with amino acid substrates generated robust inward currents. Na+-activation kinetics of proline transport and uptake under voltage clamp suggested a 1:1 Na+/amino acid co-transport stoichiometry. Susbtrate and co-substrate influenced each other's K0.5 values, suggesting that they share the same binding site. A mouse B0AT2-like transport activity was detected in synaptosomes and cultured neurons. A potential role of B0AT2 in transporting neurotransmitter precursors and neuromodulators is proposed.
- amino acid transport
- neurotransmitter transporter family
- solute carrier 6 (SLC6) transporter family
- transport mechanism
Large neutral amino acids are frequently precursors for brain neurotransmitters. Aromatic amino acids are precursors for the neurotransmitters 5-hydroxytryptamine (serotonin), noradrenaline and dopamine. Branched-chain amino acids are important amino group donors for glutamate biosynthesis  and can serve as anaplerotic metabolites providing tricarboxylic acid cycle intermediates for glutamate and GABA (γ-aminobutyric acid) biosynthesis. It is often assumed that large neutral amino acids are transported into non-epithelial cells by the Na+-independent system L amino acid transporter. However, in contrast with this notion, Na+-dependent transport of large neutral amino acids has been described not only in epithelial cells, but also in a variety of brain cell preparations. Herrero et al.  reported Na+-dependent uptake of tryptophan into vesicles derived from synaptosomes. Similarly, Yudkoff et al.  demonstrated Na+-dependent uptake of leucine into synaptosomes. An Na+-dependent transporter for large neutral amino acids with a similar substrate specificity as system L was also detected on the abluminal side of the blood–brain barrier . These transporters are different from other well-known Na+-dependent amino acid transporters, such as system ASC or system A, none of which transport branched-chain or aromatic acids .
The SLC6 (solute carrier 6) family is one of the largest transporter families in the human genome and currently comprises 20 members . They are grouped into four subfamilies, namely the monoamine transporter branch, the GABA transporter branch, the amino acid transporter branch and the ‘orphan transporter branch’ . We have recently identified the Na+-dependent neutral amino acid transporter B0AT1 (SLC6A19) , which is most closely related to the orphan transporters, suggesting that the orphans may in fact be amino acid transporters. This notion was subsequently reinforced by the identification of the Na+- and Cl−-dependent IMINO transporter (SLC6A20), considered previously to be the orphan transporter XT3 or XTRP3 [8,9]. The orphan transporter v7-3 was initially cloned from rat brain , bovine brain  and human cerebellum , but no functional activity has been reported. The transporter failed to show active uptake of neurotransmitters, such as dopamine, noradrenaline, 5-hydroxytryptamine, GABA, glycine and glutamate, in two different expression systems . In the present paper, we show that slc6a15, previously called neurotransmitter transporter v7-3 or NTT7-3, is a high-affinity Na+-dependent transporter for large neutral amino acids. Because of its functional similarity to B0AT1, we suggest naming the transporter B0AT2.
MATERIALS AND METHODS
cDNA cloning and plasmids
Total RNA was isolated from mouse brain using the Nucleospin RNA kit (Macherey-Nagel, Düren, Germany). To clone mB0AT2 (mouse B0AT2), cDNA was synthesized as described previously . The coding sequence was amplified using Pfu polymerase (Promega, Madison, WI, U.S.A.) during 40 PCR cycles of 95 °C, 30 s; 55 °C, 60 s; and 72 °C, 15 min using the sense primer, 5′-CCACCATGCCTAAGAATAGCA-3′, and the antisense primer, 5′-CCCTGCTTTCGTTCAGCTAC-3′. The sense primer corresponds to bases 435–450 of the mouse cDNA NM_175328 (GenBank® accession number) and, in addition, contains a Kozak consensus sequence (underlined) in front of the start codon (in boldface). The antisense primer corresponds to bases 2621–2640 of the same cDNA. The B0AT2 PCR product was purified by agarose gel electrophoresis, and the 2205 bp fragment was isolated using a gel elution kit (Qiagen, Clifton Hill, VIC, Australia). The isolated PCR fragment was initially cloned using the Zero Blunt TOPO PCR cloning kit (Invitrogen, Mulgrave, VIC, Australia); its sequence was confirmed (Biomolecular Resource Facility, Australian National University) and deposited under accession number AM085111 in the GenBank® database. For expression studies, mB0AT2 was excised with HindIII/XbaI and inserted into the same sites of the oocyte expression vector pGEM-He-Juel .
Oocytes and injections
Oocyte isolation and maintenance have been described in detail previously . For expression in oocytes, mB0AT2 in pGem-He-Juel was linearized with SalI and transcribed in vitro using the T7 mMessage mMachine Kit (Ambion, Austin, TX, U.S.A.). Oocytes were injected with 30 ng of cRNA encoding mB0AT2. Transport measurements were carried out after 3–10 days of expression. Rat MCT1 (monocarboxylate transporter 1) was used as described previously .
Neuron-rich primary cultures were derived from embryonic mouse brains as described by Brewer et al. . Embryos were removed from the uteri of pregnant mice on E17 (embryonic day 17). Complete brains were prepared and stored in Ca2+- and Mg2+-free HBSS (Hanks balanced salt solution; 137 mM NaCl, 5.4 mM KCl, 0.44 mM KH2PO4, 2.7 mM Na2HPO4 and 10 mM Hepes) supplemented with pyruvate (1 mM) and 10 mM Hepes, pH 7.4 (storage solution). Cells were dissociated by passage through a 211 μm nylon mesh and resuspended in the storage solution (0.2 ml per brain) by triturating with a 1 ml automatic pipette. Cells were diluted to twice the original volume with the storage solution supplemented with Ca2+ (1.26 mM) and Mg2+ (0.9 mM) and collected by centrifugation at 200 g for 5 min. The pellet of cells from one litter (10–13 animals) was resuspended in 20 ml of Neurobasal/B27 (Invitrogen, Mulgrave, VIC, Australia) supplemented with 0.5 mM glutamine, 0.025 mM glutamate, 20 units/ml penicillin and 0.02 mg/ml streptomycin; 3×105 cells were seeded on to poly(D-lysine)-coated 35-mm-diameter dishes. Cells were incubated at 37 °C in a humidified atmosphere of 19:1 air/CO2, in serum-free Neurobasal/B27 medium supplemented with 0.5 mM glutamine, 0.025 mM glutamate, 20 units/ml penicillin and 0.02 mg/ml streptomycin. Half of the medium was replaced by fresh medium without glutamate every fourth day. Transport experiments were carried out 9 days after seeding. To examine the purity of the cultures (Supplementary Figure 1 at http://www.BiochemJ.org/bj/393/bj3930421add.htm), cells were washed with PBS and then fixed for 10 min in 3.5% (w/v) paraformaldehyde in PBS. Paraformaldehyde was removed by washing dishes twice with PBS and once with 0.1% (w/v) glycine in PBS. As a preparation for staining, cells were permeabilized by treatment with 0.3% (v/v) Triton X-100 in PBS for 10 min. Antigens were detected by incubation with primary and secondary antibodies in 10% (v/v) goat serum in 0.1% (w/v) PBS for 2 h each. A monoclonal anti-GFAP (glial fibrillary acidic protein) antibody (diluted 1:5; Dianova, Hamburg, Germany) and a polyclonal rabbit anti-(neurofilament 200) antibody (diluted 1:10; Sigma, St. Louis, U.S.A.) were used as primary antibodies. Phycoerythrin-labelled sheep F(ab′)2 anti-(mouse IgG) (Amrad Biotech, Richmond, VIC, Australia) or FITC-labelled goat anti-(rabbit IgG) (Sigma) were used as secondary antibodies.
Mouse cortical synaptosomes were prepared essentially as described by Lopez-Perez . Briefly, the method employs differential centrifugation in a continuous sucrose gradient followed by a purification procedure based on a two-phase aqueous system of dextran 500 and PEG [poly(ethylene glycol)] 4000 . The cerebral cortex of four to seven adult mice was used for each preparation. The purified synaptosomes were suspended to a final protein concentration of 0.5 mg/ml in a buffer solution (pH 7.4) containing 0.32 M sorbitol, 0.1 mM potassium EDTA and 5 mM potassium phosphate. The suspension was frozen rapidly in liquid nitrogen and stored at −160 °C for up to 2 weeks. On the day of the experiment, the synaptosomal suspension was thawed at 37 °C and then maintained on ice for the duration of the experiment.
Flux experiments in oocytes were performed as described previously . ND96 (96 mM NaCl, 2 mM KCl, 1.8 mM CaCl2, 1 mM MgCl2 and 5 mM Hepes, titrated with NaOH to pH 7.4, unless otherwise indicated) was used as an incubation buffer. To achieve Na+-free conditions, NaCl was replaced by NMDG (N-methyl-D-glucamine)-Cl or LiCl. To determine the Na+-dependence of transport, NaCl-ND96 (pH 7.4) was mixed with NMDG-Cl-ND96 (pH 7.4, NaCl replaced by NMDG-Cl) in different proportions. Different pH values were adjusted by mixing Mes-buffered ND96 (5 mM Hepes replaced by 5 mM Mes) with Tris-buffered ND96 (5 mM Hepes replaced by 5 mM Tris base). Oocytes were depolarized by addition of 50 mM KCl. Uptake of [14C]proline increased linearly with time for up to 20 min. As a result, uptake was determined using an incubation period of 10–15 min. All oocyte experiments were performed at least three times.
Transport experiments with primary cultures of neurons were performed at 37 °C. Growth medium was aspirated, and cells were washed three times with 2 ml of modified HBSS (bicarbonate-free, Hepes-containing HBSS: 136.6 mM NaCl, 5.4 mM KCl, 2.7 mM Na2HPO4, 1 mM CaCl2, 0.5 mM MgCl2, 0.44 mM KH2PO4 and 0.41 mM MgSO4, containing 5 mM Hepes, pH 7.5). To initiate transport, 1.8 ml of modified HBSS containing [14C]proline and unlabelled proline at a final concentration of 50 μM and a specific radioactivity of 120 kBq/nmol was added to the cells. Unlabelled competitive inhibitors were added at a final concentration of 5 mM. After 5 min, transport was stopped by aspirating the transport buffer, followed by three washing cycles with 2 ml of ice-cold modified HBSS. To determine Na+-independent transport, NaCl was replaced by NMDG-Cl. Cells were lysed by addition of 1 ml of 0.1 M HCl. Of the resulting suspension, a 900 μl aliquot was mixed with 3 ml of scintillation cocktail (Ultima Gold; PerkinElmer Life and Analytical Sciences, Boston, MA, U.S.A.), and radioactivity was determined by liquid-scintillation counting. A portion of 100 μl was used for protein determination using the Bradford reagent (Sigma, Castle Hill, NSW, Australia).
To measure proline uptake in synaptosomes, 20 μl of synaptosomal suspension (8–10 μg of protein) was added to 180 μl of HBSS (1.26 mM CaCl2, 5.4 mM KCl, 0.4 mM KH2PO4, 0.5 mM MgCl2, 0.4 mM MgSO4, 136.6 mM NaCl or NMDG-Cl, 10 mM Hepes and 2.7 mM Na2HPO4, pH 7.5) at room temperature (23 °C), which generates a large inwardly directed Na+-gradient. To determine Na+-independent uptake, NaCl was replaced by NMDG-Cl. The synaptosomes were allowed to accumulate substrate for 5 min in the presence of 50 μM L-proline and 3.7 kBq of L-[14C]proline in the absence or presence of a 200-fold excess (10 mM) of competing substrates. The uptake was terminated by adding 2 ml of ice-cold buffer containing 5 mM unlabelled proline to the transport assay. Synaptosomes were collected on nitrocellulose filters (0.45 μm pore size) and accumulated radioactivity was measured using liquid-scintillation counting.
Amino-acid-induced currents were analysed by two-electrode voltage clamp recording. The recordings were performed with 1×LU and 10×MGU headstages connected to a Geneclamp 500B electronic amplifier (Axon Instruments, Union City, CA, U.S.A.). The output signal was amplified 10-fold and filtered at 50 Hz. The analogue signal was converted by a Digidata 1322A (Axon Instruments), and data were sampled using pCLAMP software (Axon Instruments). Oocytes were chosen that had a membrane potential more negative than −30 mV. Once a stable membrane potential was reached under current-clamp conditions, the amplifier was switched to voltage-clamp mode, holding the oocytes at −50 mV. ND96 was used as control solution in all recordings. A complete change of the bath to a new solution was accomplished in approx. 10 s. During steady-state measurements, data were sampled at 3 Hz. To study voltage-dependence of the transporter, the holding potential was increased stepwise from −150 mV to +50 mV in increments of 10 mV; the step length was 500 ms and the sampling rate was increased accordingly. To determine the stoichiometry of B0AT2, the flow-through of ND96 was stopped and 20 μl of 10 mM [14C]proline was added to the bath, giving a final concentration of 1 mM. After 15 min, the flow-through was restarted to remove the radioactive substrate. After the inward current ceased, the oocyte was removed and placed in a scintillation vial to determine its radioactivity. The amount of accumulated radioactivity was compared with that of the integrated inward current.
RT (reverse transcription)–PCR
A 1-kb actin cDNA fragment was amplified as a control during 30 cycles using the following primers: 5′-GCTCACCATGGATGATGATATCGC-3′ and 5′-GGAGGAGCAATGATCTTGATCTTC-3′.
Calculations, statistics and computer analysis
Each datapoint or column in the Figures, Table 2 and Table 3 represents the activity (mean±S.D.) for seven to ten mB0AT2-expressing oocytes. In flux experiments, the activity (mean±S.D.) of seven to ten non-injected oocytes was subtracted from this activity. Flux studies in cells and synaptosomes were performed in triplicate. Kinetic constants were derived by non-linear curve-fitting using Origin7.0 software (OriginLab Corporation, Northampton, MA, U.S.A.). To determine K0.5 (half-saturation constant) and Imax (maximum current), the Michaelis–Menten equation: or the Hill equation: were used, where h is the Hill coefficient, I is actual current and [S] is substrate concentration. Sequence alignment was calculated using programs of GCG and PHYLIP packages supplied by ANGIS (Australian National Genomic Information Service). Sequence alignment was performed using ClustalW .
Cloning of mB0AT2 (v7-3)
The SLC6 family comprises transporters for monoamines, GABA, taurine, creatine and amino acids in addition to a large group of orphan transporters. We showed recently that two orphan transporters, namely SLC6A19 (B0AT1) and SLC6A20 (IMINO), are amino acid transporters [7,8]. Subsequently, we wanted to test the hypothesis that other members of the orphan transporter branch are also amino acid transporters. To this end, we amplified the slc6a15 sequence from mouse brain and deposited its sequence under accession number AM085111 in the GenBank® database. The transcript encodes a membrane protein of 729 amino acids in length. Alignment with the other mouse members of the SLC6 family reveals that mB0AT2 is more closely related to NTT4 (slc6a17) than it is to mB0AT1. mB0AT2 shares 65% identical amino acids with mNTT4 but shows only 35% identity with mB0AT1 (slc6a19), XT2 (slc6a18) and IMINO (slc6a20), with all other SLC6 family members having between 21% and 35% identity with mB0AT2.
Substrate specificity of mB0AT2 (v7-3)
Expression of mB0AT2 in Xenopus laevis oocytes resulted in a significant increase of neutral-amino acid uptake activity compared with control oocytes. Significant uptake above non-injected oocytes was observed for proline, leucine, isoleucine, alanine and phenylalanine (Figure 1A). Because there is negligible endogenous proline uptake in oocytes, we determined the substrate specificity of B0AT2 by challenging the uptake of 50 μM [14C]proline with a 100-fold excess of unlabelled L-amino acids (Figure 1B). Proline uptake was strongly (>90%) inhibited by methionine, leucine, isoleucine, valine, proline, alanine and the amino acid analogue AIB (aminoisobutyric acid). Partial inhibition (40–90%) was exerted by phenylalanine, serine, threonine, glutamine, asparagine, histidine, hydroxyproline and by the amino acid analogues BCH (2-aminobicyclo[2,2,1]heptane-2-carboxylic acid) and nipecotic acid. No significant inhibition was observed on addition of glycine, tyrosine, tryptophan, cysteine, arginine, lysine, aspartate or glutamate. A number of amino-acid-related compounds such as MeAIB [(N-methyl)AIB], β-alanine, GABA, betaine and [leucine]enkephalin (100 μM) also did not inhibit [14C]proline uptake. In summary, it appears that large aliphatic neutral amino acids plus proline were the preferred substrates of mB0AT2. The transporter was stereospecific, as indicated by the lack of inhibition of L-proline transport by D-proline (Figure 1B). The basic characteristics of mB0AT2 as determined using flux experiments were confirmed by electrophysiological recordings of substrate-induced transporter currents (Figure 1C and Table 2). When tested at a concentration of 1 mM, currents were induced in the following order: Pro>Leu=Met=Ala=Val=AIB>Ile>Thr>Asn=Ser>Phe>Gln, confirming the order of transport activity as determined with labelled amino acids. All other amino acids tested did not elicit significant currents (Table 2). mB0AT2 is functionally related and sequence-related to mB0AT1  but has a more restricted substrate specificity (Table 2). Neither mB0AT1 nor mB0AT2 accepted neurotransmitters or other substrates of SLC6 family members, such as taurine, creatine or betaine (Table 2). Two groups of mB0AT2 substrates were identified. The high-affinity substrates (methionine, isoleucine, leucine and proline) were transported with apparent K0.5 values of 40–200 μM (Table 3), whereas the low-affinity substrates (alanine, phenylalanine, glutamine and pipecolic acid) were transported with millimolar values of K0.5 (Table 3).
Transport mechanism of mB0AT2 (v7-3)
Uptake of [14C]proline was Na+-dependent; replacement of NaCl by NMDG-Cl completely abolished the transport activity, and replacement by LiCl reduced the transport activity by 80% (Figure 2A). Similar to the properties of B0AT1 , replacement of Cl− with gluconate did not reduce the transport activity of B0AT2. Addition of 50 mM KCl to the transport buffer, a manipulation that reduces the membrane potential of oocytes from −40±7 mV to −16±3 mV, reduced proline uptake by 43% (Figure 2A). Addition of 100 mM sucrose did not change proline uptake significantly, suggesting that the inhibitory effect of KCl was not caused by the increased osmolarity (Figure 2A). This result confirmed that transport of amino acids via mB0AT2 was indeed electrogenic, suggesting that the inward currents were coupled to amino acid transport and were not generated by a transport-associated ion conductance . A coupling of substrate transport to Na+-translocation was supported further by the voltage-dependence of proline-induced currents, which remained inwardly directed at all holding potentials tested (Figure 2B). To determine the number of charges translocated together with each substrate molecule, we determined [14C]proline uptake under voltage-clamp conditions. A plot of the integrated current against accumulated [14C]proline indicated a stoichiometry of 0.65±0.07 charges/proline (Figure 2C). Transport of proline via mB0AT2 was also dependent on extracellular pH, increasing with alkalinity (Figure 3A). The strong pH-dependence may point to a proton antiport mechanism. To test this possibility, we modulated the intracellular pH by coexpression of MCT1 . Co-expression of MCT1 and B0AT2 did not change the expression levels of the two transporters significantly (Figure 3B). Addition of 20 mM lactate to the oocytes, a manipulation that decreases the intracellular pH by approx. 0.7 pH unit , did not alter [14C]proline uptake by B0AT2, suggesting that the pH-dependence of B0AT2 was caused by an exofacial pH-modifier site and not by a participation of protons in the transport mechanism.
An activation analysis of proline uptake as a function of the Na+ concentration showed a hyperbolic dependence, suggesting that only one Na+ ion is co-transported with proline (Figure 4A). Hill coefficients varied between 0.6±0.2 and 1.0±0.2, but were never >1. To characterize the interaction between Na+ and proline further, we determined kinetic constants at different substrate and co-substrate concentrations. When measured at a substrate concentration of 100 μM, half-maximal transport velocity was reached at a Na+ concentration of 16±4 mM (Figure 4A). However, the K0.5 of Na+ decreased significantly to 4.5±0.7 and 2.8±0.4 mM when increasing the proline concentration to 0.5 mM and 5 mM respectively (Figure 4A). The Na+ concentration also affected the K0.5 of proline. At 100 mM NaCl, proline was transported with a K0.5 of 195±10 μM. This value increased to 510±40 μM and 740±90 μM when the NaCl concentration was decreased to 30 mM and 3 mM respectively (Figure 4B). These results suggest that substrate and co-substrate influence each other's binding. The kinetic parameters of substrate and co-substrate also depended on the holding potential. When changing the holding potential from −80 mV to +20 mV, the K0.5 value of proline changed from approx. 0.2 mM to 0.8 mM (Figure 5A). The calculated Imax increased linearly over this range of the electrical driving force (Figure 5B). The K0.5 value (determined at a proline concentration of 0.1 mM) of Na+ changed from 8.7±0.4 mM to 19±2 mM and > 50 mM when the holding potential was changed from −100 mV to −50 mV and −10 mV respectively (Figure 5C).
Tissue distribution of mB0AT2 (v7-3)
RT–PCR experiments showed significant expression of mB0AT2 only in brain, lung and kidney (Figure 6). Expression of mB0AT2 was abundant in all three major brain regions, namely the cortex, the cerebellum and the brain stem. The RT–PCR data were confirmed by analysis of the EST (expressed sequence tag) database. Of 74 ESTs corresponding to NM_175328 shown in the Unigene database, 45 are found in the brain and 22 are found in the eye. Between one and three ESTs were detected in mammary gland, kidney, skin and testis. Transcripts of mB0AT2 were detected throughout development, starting with the pre-implantation embryo.
Proline uptake in cultured neurons and synaptosomes
The transport activity of mB0AT2 demonstrates its expression at the oocyte cell surface. To determine whether mB0AT2 is also expressed in the plasma membrane of neural cells, we characterized proline transport in cultured neurons and synaptosomes. Uptake of [14C]proline (50 μM) in cultured neurons was entirely Na+-dependent. Marginal proline uptake was observed when NaCl was replaced by NMDG-Cl (Figure 7A). In the presence of Na+, proline uptake was strongly inhibited by a 100-fold excess of the B0AT2 substrates leucine, alanine, AIB and proline itself. Other amino acids or related compounds, which are not substrates of B0AT2, such as MeAIB, GABA, arginine or glutamate, inhibited proline transport only slightly. Using RT–PCR, six proline transporters were detected in cultured neurons, namely the system A isoforms SNAT1 (system N/A transporter 1) and SNAT2 , the proton amino acid transporters PAT1 (proton amino acid transporter 1) and PAT2 , the neuronal proline-specific transporter PROT (proline transporter)  and B0AT2 (Figure 7B). The cultures contained less than 3% astrocytes (Supplementary Figure 1 at http://www.BiochemJ.org/bj/393/bj3930421add.htm), arguing against a contribution of other cell types to proline uptake.
A [leucine]enkephalin-sensitive proline transporter has been described in synaptosomes as having similar properties to the proline-specific transporter PROT (slc6a7) . Since B0AT2 is resistant to [leucine]enkephalin inhibition (Figure 1), we wondered whether a B0AT2-like activity could also be detected in synaptosomes. Similar to neurons, proline uptake in synaptosomes was entirely Na+-dependent. In contrast with cultured neurons, however, proline uptake was strongly inhibited by MeAIB and, to a similar extent, by [leucine]enkephalin (Figure 8A). Combining the two inhibitors did not increase inhibition, suggesting that MeAIB and [leucine]enkephalin inhibit the same transporter. To study the MeAIB-resistant (and [leucine]-enkephalin-resistant) proline uptake (approx. 40% of total proline uptake), we subsequently characterized proline uptake in synaptosomes in the presence of 10 mM MeAIB. The MeAIB-resistant fraction of proline uptake was strongly inhibited by leucine, alanine, methionine and proline itself. Other amino acids which are not substrates of B0AT2, such as arginine, GABA or glutamate, did not inhibit MeAIB-resistant proline uptake (Figure 8B). In summary, it seems that a transport activity matching the properties of B0AT2 can be detected in both cultured neurons and cortical synaptosomes.
In the present study we identified the orphan neurotransmitter transporter v7-3 as a Na+-dependent transporter for large neutral amino acids. Because of its functional similarity to B0AT1 , we named it B0AT2. We suggest that the orphan transporter branch of the SLC6 family should be renamed into the amino acid transporter branch (II), discriminating it from the amino acid transporter branch (I), which comprises two glycine transporters, the neuronal proline transporter PROT and the general amino acid transporter ATB0,+ . The Na+-dependent amino acid transporters of the SLC6 family furthermore belong to a much larger family of Na+-dependent nutrient amino acid transporters (NAT family) , with members found in bacteria , insects [26,28,29] and a variety of other lower and higher eukaryotes . It appears that this family initially evolved to serve in amino acid uptake from nutrient sources in unicellular and multicellular organisms, whereas neurotransmitter transport is likely to be a more recent development.
Our results indicate that B0AT2 mediates a 1:1 Na+/amino acid co-transport as supported by the following evidence: (i) the transporter is Na+-dependent and Cl−-independent; (ii) the Na+-activation curves are hyperbolic; (iii) the transporter is electrogenic, translocating approx. one charge per substrate molecule; and (iv) protons are unlikely to participate in the transport mechanism and the pH-dependence is most probably caused by an exofacial modifier site. These properties match those of B0AT1 , suggesting that both transporters have the same transport mechanism. Increasing the electrical driving force increased the apparent affinity of B0AT2 for substrate and co-substrate. A possible explanation for this observation is that the negative membrane potential causes an increase of the local Na+ concentration at its binding site (ion-well effect). An increase of the local Na+ concentration also would decrease the proline K0.5, thereby explaining the effect of the membrane potential on the apparent proline affinity. The mutual influence of substrate and co-substrate on each other's K0.5 values suggests a possible interaction between substrate and co-substrate at the binding site. This view is supported by the recently published crystal structure of the prokaryotic SLC6-related transporter LeuTAa from the bacterium Aquifex aeolicus , which has 26% amino acid residues similar to, or identical with, those in mB0AT2 (Figure 9, and Supplementary Figure 2 at http://www.BiochemJ.org/bj/393/bj3930421add.htm). The occurrence of highly conserved residues throughout the sequence allowed us to develop a topological model of mB0AT2, which aligns well with the structure of the bacterial transporter (Figure 9). In particular, residues in helix 1 and helix 6, which form the substrate- and Na+-binding sites of the bacterial transporter, are highly conserved. The loops and termini are significantly longer in the mammalian transporter and are not conserved. In the LeuTAa structure, the carboxy group of the substrate leucine contributes to the co-ordination of the Na+ ion. This overlapping binding site elegantly explains the mutual influence of substrate and co-substrate on each other's K0.5 values. Na+ will provide an improved binding site for leucine and vice versa. Notably, all but one residue implicated in substrate- and Na+-binding in the bacterial transporter are fully conserved in B0AT2. Also conserved are several residues that form the substrate binding pocket (helix 8) (accommodating leucine in both transporters) and the four residues that are proposed to form the cytosolic and extracellular gates.
Inhibition of proline transport by branched-chain amino acids is a unique property of B0AT2. Other neuronal proline transporters such as PROT , PAT1/2  and the system A isoforms SNAT1 and SNAT2  do not accept large bulky neutral amino acids as substrates. Inhibitors of these transporters, such as MeAIB (SNAT1/2), GABA (PAT1 and PAT2) and [leucine]-enkephalin (PROT), on the other hand, do not inhibit B0AT2, therefore allowing discrimination of the different transport activities in cultured neurons and synaptosomes. In cultured neurons, we found little inhibition of proline uptake by GABA and MeAIB, excluding a significant contribution to proline uptake by SNAT1, SNAT2, PAT1 and PAT2. Although we found mRNAs of all four transporters, they appeared to be located in intracellular membranes and hence their transport activity was inaccessible. This notion is supported by previous studies showing that surface expression of SNAT1 and SNAT2 is suppressed by amino acids in the medium  and that PAT1 is mainly localized in intracellular membranes in neurons . We found very little mRNA for PAT2, hence neurons may not express significant amounts of the protein. In contrast with the limited inhibition by MeAIB and GABA, a 100-fold excess of the B0AT2 substrates leucine, methionine and AIB inhibited [14C]proline uptake by approx. 75%, which is the same extent of inhibition as exerted by unlabelled proline itself. Moreover, it excludes a significant contribution of the proline-specific transporter PROT, confirming that PROT is also mainly localized in intracellular membranes . The properties of B0AT2 are remarkably similar to the Na+-dependent proline transport activity described in rat brain slices , which is inhibited by leucine, norleucine and norvaline. The situation appears to be different in synaptosomes. We confirmed that a significant fraction of proline uptake in synaptosomes is inhibited by [leucine]enkephalin and MeAIB . Both compounds appear to inhibit the same transporter because their effect was not additive. MeAIB is commonly used to delineate system A. However, MeAIB also inhibits most proline transporters, such as PAT1 and IMINO, and is likely to inhibit PROT as well. The [leucine]enkephalin-resistant proline uptake (approx. 40% of total uptake), in contrast, matched the properties of B0AT2. As a result, it appears that synaptosomal proline uptake has two components, one attributable to PROT and the second to B0AT2. It is worth noting that a [leucine]enkephalin-resistant fraction of synaptosomal proline uptake was noted previously  and was characterized as Cl−-independent, in line with the properties of B0AT2.
The distribution of the neurotransmitter transporter v7-3 has been studied in some detail in rat brain by in situ hybridization . Expression of the transporter was detected in neurons, but the distribution pattern was not compatible with any specific neurotransmitter, as seen by the presence of hybridization signals in dopaminergic neurons (substantia nigra), serotonergic neurons (raphe nuclei), noradrenergic neurons (locus coeruleus), glutamatergic neurons (hippocampus, olfactory bulb) and cholinergic neurons (motor neurons). This seemingly diffuse distribution can be explained by a role of B0AT2 in the provision of neurotransmitter precursors. The nitrogen of leucine, for example, is efficiently transferred to α-oxoglutarate, thereby forming glutamate . Isoleucine, methionine and valine can be metabolized to form succinyl-CoA and hence can act as precursors for glutamate biosynthesis. In motor neurons, methionine provides the methyl groups for acetylcholine biosynthesis [38,39]. Methionine is also a precursor of homocysteic acid, a glutamate analogue, which has been suggested as an excitatory amino acid in the brain . The physiological role of proline in the brain is still ill-defined. Proline can be converted into glutamate, involving proline dehydrogenase and glutamate–semialdehyde dehydrogenase, but these enzymes appear to be associated with astrocytes , whereas PROT and B0AT2 are expressed in neurons. The role of B0AT2 in aminergic neurons remains to be defined.
In summary, it is likely that the major function of B0AT2 is the transport of neurotransmitter precursors into neurons. The physiological role of neuronal proline accumulation remains to be determined.
This study was supported by grants from the Australian Research Council (ARC Discovery Project Grants DP 0208961 and DP0559104) and the National Health and Medical Research Council (Project Grant 224229) to S. B. L. K. B. carried out this work as a visiting student to the Australian National University.
The nucleotide sequence data reported will appear in DDBJ, EMBL, GenBank® and GSDB Nucleotide Sequence Databases under the accession number AM085111.
Abbreviations: (Me)AIB, (N-methyl)aminoisobutyric acid; BCH, 2-aminobicyclo[2,2,1]heptane-2-carboxylic acid; mB0AT2, mouse B0AT2; EST, expressed sequence tag; GABA, γ-aminobutyric acid; HBSS, Hanks balanced salt solution; MCT1, monocarboxylate transporter 1; NMDG, N-methyl-D-glucamine; PAT1, proton amino acid transporter 1; PROT, proline transporter; RT, reverse transcription; SLC6, solute carrier 6; SNAT1, system N/A transporter 1
- The Biochemical Society, London