A Family of Basic Amino Acid Transporters of the Vacuolar Membrane from Saccharomyces cerevisiae *

the Saccharomyces cerevisiae we identified genes involved in the transport into vacuoles of the basic amino acids histidine, lysine, and arginine. ATP-dependent uptake of histidine and lysine by isolated vacuolar membrane vesicles was impaired in YMR088c , a vacuolar basic amino acid transporter 1 ( VBA1 )-deleted strain, whereas uptake of tyrosine or calcium was little af-fected. This defect in histidine and lysine uptake was complemented fully by introducing the VBA1 gene and partially by a gene encoding Vba1p fused with green fluorescent protein, which was determined to localize exclusively to the vacuolar membrane. A defect in the uptake of histidine, lysine, or arginine was also observed in the vacuolar membrane vesicles of mutants YBR293w ( VBA2 ) and YCL069w ( VBA3 ). These three VBA genes are closely related phylogenetically and constitute a new family of basic amino acid transporters in the yeast vacuole. in Living Cells— Subcellular local- ization of Vba1p-GFP fusion protein in living S. cerevisiae cells was assessed using fluorescence microscopy (DeltaVision microscope; Ap- plied Precision) as described previously (15). To stain vacuolar mem-branes, FM 4–64 (Molecular Probes) was added to growing cultures to a final concentration of 5 (cid:1) M . The cells were further cultured for 20 min and harvested. After washing, the cells were resuspended in fresh YPD media for 30 min to allow the dye to stain the vacuole via endocytosis.

Vacuoles are the largest organelles in the yeast Saccharomyces cerevisiae, occupying ϳ25% of the cell volume. Like lysosomes, vacuoles function as a digestive compartment but also serve as a storage compartment in which the bulk of basic amino acids is localized (1,2). The concentration of arginine in the vacuoles of S. cerevisiae grown with arginine as the primary nitrogen source is about 20 times higher than in the cytoplasm (1,2). In contrast, vacuoles contain little glutamic acid, which is the most abundant amino acid in yeast (2). Knowledge about vacuolar compartmentalization of amino acids is prerequisite to understanding the regulation of nitrogen metabolism. The vacuolar membrane catalyzes the active transport of a variety of amino acids (3,4), a process that is driven by a proton electrochemical gradient generated via the action of the proton pumping vacuolar ATPase (5,6) and is likely mediated by a proton/amino acid antiporter.
About two decades ago, kinetics experiments of amino acid uptake by vacuolar membrane vesicles suggested seven independent transport systems for amino acids in the S. cerevisiae vacuole: arginine, lysine-arginine, histidine, phenylalaninetryptophan, tyrosine, asparagine-glutamine, and isoleucineleucine (4). However, for many years the genes for the proteins involved in transport of these amino acids into vacuoles remained unknown. However, a recent report identified some of the genes involved in this process (7), namely AVT1 for the uptake of glutamine, isoleucine, and tyrosine into vacuoles and both AVT3 and AVT4 for efflux of them. Moreover, AVT6 is likely responsible for aspartate and glutamate efflux from vacuoles (7).
Here, we have reported identification of genes encoding proteins involved in the uptake of basic amino acids into vacuoles. These genes constitute a new vacuolar transporter family in the major facilitator superfamily (MFS) 1 of S. cerevisiae.
‡ ‡ To whom correspondence should be addressed: Faculty of Agriculture, Ehime University, 3-5-7 Tarumi, Matsuyama 790-8566, Japan. Tel./Fax: 81-089-946-9861; E-mail: ykaki@agr.ehime-u.ac.jp. was conducted by using primers for ⌬ymr088c::KanMX4, a primer pair to clone the pVBA1 as described below, and for ⌬ybr293w::KanMX4, a 5Ј primer, 5Ј-TTTGGTACCGAATAAGCAATTACAATACG-3Ј, and a 3Ј primer, 5Ј-CTGCCGCCTCAGATTCCC-3Ј. Cells were grown aerobically at 30°C in YPD medium; if necessary, medium was supplemented with 200 g/ml geneticin and 100 g/ml nourseothricin. Deletion of the corresponding gene was confirmed by chromosomal polymerase chain reaction (PCR). The gene YMR088c (VBA1) was amplified by PCR using a 5Ј primer with a KpnI site, TGGGGTACCACGAGGCTGGTCATGC, and a 3Ј primer with a SacII site, ATTACCGCGGGGGAAAGCCCTTTG. After digestion with KpnI and SacII, the PCR product was inserted into the same restriction sites of plasmid pRS316 (10), yielding the recombinant plasmid pVBA1. For construction of the Vba1p-green fluorescent protein (GFP) fusion protein, an XbaI site was created by site-directed mutagenesis using QuikChange (Stratagene) immediately before the stop codon of VBA1. A DNA fragment encoding a modified GFP (S65T) with XbaI sites on both sides was then ligated into the created XbaI site, yielding the recombinant plasmid pVBA1GFP. Plasmids pVBA1 or pVBA1GFP were introduced into ATCC 4006223 (⌬vba1 mutant) by the lithium acetate method of Ito et al. (11).
Transport Assay with Intact Cells and Vacuolar Membrane Vesicles-Amino acid uptake by intact cells was performed as follows. Cells were cultured in YPD medium, harvested during the logarithmic phase of growth, washed twice with water, and suspended in 10 mM MES-Tris (pH 6.4) containing 2 mM MgCl 2 , 25 mM KCl, and 10 mM glucose at a cell density of 5 ϫ 10 8 cells/ml. The amino acid uptake reaction (performed at room temperature) was initiated by the addition of a specific amino acid (2 mM final concentration in the medium). At specific time intervals, 0.5-ml aliquots of cell suspension were withdrawn and layered on 0.6 ml of corn oil/di-n-butyl phthalate (1:3) and centrifuged for 2 min at 10,000 ϫ g at room temperature. The amount of amino acid in the supernatant (water phase) was measured by the ninhydrin reaction (12). Purification of vacuolar membrane vesicles from various S. cerevisiae strains was performed as described previously (3,5). The uptake of [U-14 C]histidine (final 0.1 mM; 11.6 GBq/mmol), [U-14 C]lysine (0.1 mM; 11.8 GBq/mmol), [U-14 C]arginine (0.1 mM; 12.9 GBq/mmol), [U-14 C]tyrosine (0.1 mM; 16.9 GBq/mmol), and 45 CaCl 2 (0.5 mM; 18.5 GBq/mmol) by the membrane vesicles was performed as described previously (13) with minor modifications. Cellulose acetate membrane filters (0.45 m; ADVANTEC, Japan) were used, and the radioactivity was measured using a liquid scintillation counter with xylene scintillator. All radioactive materials were obtained from New England Nuclear Laboratories. Protein concentration was determined by the method of Lowry et al. (14) with bovine serum albumin as standard.
Fluorescence Microscopy of GFP in Living Cells-Subcellular localization of Vba1p-GFP fusion protein in living S. cerevisiae cells was assessed using fluorescence microscopy (DeltaVision microscope; Applied Precision) as described previously (15). To stain vacuolar membranes, FM 4 -64 (Molecular Probes) was added to growing cultures to a final concentration of 5 M. The cells were further cultured for 20 min and harvested. After washing, the cells were resuspended in fresh YPD media for 30 min to allow the dye to stain the vacuole via endocytosis.

RESULTS
Amino Acid Uptake by S. cerevisiae Mutants-Based on the complete genome sequence of S. cerevisiae (16,17), a computeraided analysis suggested the presence of the MFS comprising permeases that typically contain 12 transmembrane-spanning domains. This MFS is distributed in both prokaryotes and eukaryotes and includes uniporters, symporters, and antiporters. All MFS proteins can be further clustered into many permease families; the multidrug permease homologues, which confer multidrug resistance, constitute such a family (16,17). It is now accepted that the multidrug permease family is further divided into two subfamilies (17). The multidrug permeases most likely are drug/proton antiporters and are therefore expected to be dependent on a proton electrochemical gradient across the membrane for activity. Because amino acid transport by yeast vacuoles is dependent on the proton electrochemical gradient across the vacuolar membrane (3,4) and thus is likely mediated by an amino acid/proton antiporter, we expected yeast vacuolar amino acid transporters to be members of this multidrug permease family consisting of 28 proteins with 12 or more membrane-spanning regions (16). We purchased most of these deletion strains from ATCC and investigated their features. Fig. 1 shows histidine uptake by intact S. cerevisiae cells. Histidine uptake was abrogated in mutant YPH499 cells lacking vacuolar H ϩ -ATPase (⌬vma1 RH104 cells; Fig. 1A). The vma1 mutation also impaired the uptake of other amino acids such as arginine, lysine, isoleucine, tyrosine, and phenylalanine (data not shown), each a substrate of vacuolar transport systems depending on the proton electrochemical gradient (3,4). These results suggest that amino acid uptake by intact cells involves active transport into vacuoles. We therefore examined amino acid uptake by intact cells carrying putative multidrug permease mutants to identify a candidate gene involved in amino acid transport into vacuoles. Of the mutants tested, two (⌬ymr088c or ⌬ybr293w) showed minimal histidine uptake compared with the parent BY4741 cells, and the⌬ycl069w mutant showed transient uptake (Fig. 1B). The tyrosine and isoleucine uptake activities in these three mutants was equivalent to that in BY4741 (data not shown), excluding the possibly that these mutations promoted nonspecific defects in amino acid uptake. Thus, histidine uptake into vacuoles may be mediated by one or more of these genes.
Uptake by Vacuolar Membrane Vesicles of Mutant VBA1 (⌬ymr088c)- Fig. 2 shows amino acid and calcium uptake by vacuolar membrane vesicles of strains BY4741 (parent) and ⌬ymr088c. Active transport of calcium by S. cerevisiae vacuoles is primarily mediated by a calcium/proton antiporter (18). ATPdependent uptake of histidine, arginine, lysine, tyrosine, and 45 Ca was clearly observed in BY4741 cells (Fig. 2, open circles). All these activities were completely inhibited by the protonophore carbonylcyanide m-chlorophenylhydrazone (data not shown). Consistent with the result for intact ⌬ymr088c mutant cells (Fig. 1B), ATP-dependent histidine uptake was nominal in vacuolar membrane vesicles ( Fig. 2A, closed circles). The effect of this mutation on arginine uptake by vesicles was insignificant (Fig. 2B), but lysine uptake was severely impaired (Fig.  2C, closed circles). Tyrosine uptake (Fig. 2D), as described above for intact cells, was normal as was calcium uptake (Fig.  2E). Defects in both histidine and lysine uptake by the vesicles of the ⌬ymr088c mutant were clearly recovered by introducing the YMR088c gene on pRS316 (Fig. 2, A and C, open triangles), but not by the vector alone (open squares). These results sug-gest that the product of the YMR088c gene (VBA1, vacuolar basic amino acid transporter) is required for vacuolar uptake of at least histidine and lysine.

Localization of Vba1p-GFP on the Vacuolar Membrane-
Complementation for a defect in histidine or lysine uptake by ⌬vba1 mutant vesicles was also attempted with plasmid pVBA1GFP, encoding GFP fused to the carboxyl terminus of Vba1p (see "Experimental Procedures"). Expression of pVBA1GFP recovered the defect, albeit incompletely (Fig. 2, A and C, closed triangles). Thus, the Vba1p-GFP fusion protein functions as an amino acid transporter for the vacuole. Vba1p-GFP localized exclusively to the vacuolar membrane in log-phase ⌬vba1 mutant cells (Fig. 3), and GFP fluorescence coincided with that of FM 4 -64, which selectively stains the vacuolar membrane (15,19). These results indicate that Vba1p is a vacuolar membrane protein that, at a minimum, is involved in the transport of basic amino acids into S. cerevisiae vacuoles.
Uptake Activities of VBA2 (⌬ybr293w) and VBA3 (⌬ycl069w) Mutant Vacuolar Membrane Vesicles-VBA2 (YBR293w) and VBA3 (YCL069w) mutants (Fig. 4) express candidate genes for histidine transport into vacuoles (Fig. 1B), so we measured uptake activities of vacuolar membrane vesicles from these cells. ATP-dependent uptake of histidine, arginine, and lysine were all severely impaired in ⌬vba2 mutant vesicles (Fig. 4, A-C, closed circles). Tyrosine uptake also decreased, but to a lesser extent (Fig. 4D). In ⌬vba3 mutant vesicles, uptake of histidine and lysine, but not arginine or tyrosine, was impaired (Fig. 4, A-C,  closed triangles). Interestingly, a slight decrease in calcium uptake was observed in VBA3, rather than VBA2, mutant vesicles (Fig. 4E). We concluded that the three genes, VBA1, VBA2, and VBA3, chosen as the candidate transporters in experiments with intact cells are involved in vacuolar uptake of basic amino acids and that VBA1 and VBA3 mediate histidine and lysine uptake, whereas VBA2 mediates histidine, arginine, and lysine uptake. To determine participation of these VBA genes in total uptake activities in vacuole, we further examined vesicular uptake of mutants deleted in pairs or all three VBA genes (Fig. 5). More or less 20% of the uptake activities of histidine, arginine, and lysine by the parent were all retained in the triple mutant ⌬vba1⌬vba2⌬vba3 (⌬ymr088c::URA3MX4 ⌬ybr293w::KanMX4 ⌬ycl069w::NatMX4), suggesting the presence of the other system(s) for basic amino acids. Furthermore, tyrosine uptake largely decreased in deleted strains of the VBA1 and VBA2 genes (Fig. 5D). It is also noteworthy that the impairment in calcium transport of the VBA3 mutant was enhanced by combination with a deletion of the VBA1 gene (Fig. 5E).
In sum, our results defined a new VBA transporter family within the multidrug permease family of S. cerevisiae. These VBA transporters facilitate the uptake of basic amino acids by vacuoles.

DISCUSSION
Active transport systems for amino acids operate in yeast vacuoles. The multiplicity of such systems has been demonstrated by the kinetics of amino acid uptake by vacuolar membrane vesicles (4). Although the AVT genes involved in the transport of glutamine (asparagine), isoleucine (leucine), and tyrosine into vacuoles were recently reported (7), the genes for the transport of basic amino acids have not been determined. In the current study, we found three VBA genes encoding proteins primarily involved in transporting basic amino acids (histidine, lysine, and arginine) into S. cerevisiae vacuoles. These VBA transporters represent the main route for uptake of basic amino acids into vacuoles under standard culture conditions. Vba1p and Vba3p transport histidine and lysine, whereas Vba2p transports histidine, arginine, and lysine. Vba2p is also involved in tyrosine transport (Figs. 4D and 5D). Vba3p likely has broad substrate specificity not only for amino acids but also for calcium (Fig. 4E) and other cationic metabolites such as choline or polyamine (23). Another issue to be interpreted concerns why the impairment in calcium transport was enhanced by combination with the absence of Vba1p (Figs. 4E and 5E). Further investigation is required to understand the details of substrate specificity and the physiological roles of VBA transporters in the vacuolar compartment.
(VMA1) was defective for the uptake of lysine, arginine, histidine and tyrosine, but not glutamate, whose uptake was unaffected. Uptake of isoleucine and phenylalanine was ϳ50% relative to the control parent strain (data not shown). Although the effect of the VMA1 mutation may not always result solely in the inability to generate the proton electrochemical gradient across the vacuolar membrane, these results suggested that amino acid uptake by intact cells primarily reflects transport into vacuoles. Candidate vacuolar amino acid transporter genes were selected based on this notion. The results obtained from intact cells were consistent with those for vacuolar vesicles. We observed a transient uptake of histidine by ⌬vba3 (⌬ycl069w) and ⌬vba2 (⌬ybr293w) cells (Fig. 1B), indicating efflux from cells of ninhydrin-reactive material-possibly histidine or another amino acid(s). It has been reported that spermidine is essential for key steps in cell metabolism such as translation initiation (24); when overaccumulated, spermidine is extruded from cells via Tpo1p on the plasma membrane (Tpo1 is a member of the major facilitator superfamily) (25). Therefore it is possible that a pathway involving extrusion of amino acids from the cytoplasm helps to balance the amino acid pool within cells. Although the details of the phenomenon shown in Fig. 1B are unclear, Vba3p and Vba2p, not likely Vba1p (Fig. 1B), may regulate in some fashion amino acid extrusion from cells.
Genes that are phylogenetically related to VBA genes are depicted in Fig. 6. S. pombe fnx1 (Fig. 6A), induced upon nitrogen starvation, is involved in resistance to 3-amino-1,2,4-triazole and 4-nitroquinoline N-oxide (20). The substrate specificity of drug resistance of fnx1 is similar to that of S. cerevisiae Atr1p (20). Although its subcellular localization in S. pombe has not been determined, fnx1 is assumed to function at the plasma membrane (20). Azr1p (encoded by YGR224w, 613 residues), with 13 putative transmembrane spans, reportedly is the plasma membrane protein involved in resistance to azoles such as ketoconazole and fluconazole and to acetic acid (21). Sge1p (encoded by YPR198w, 543 residues), with 14 putative transmembrane spans, is involved in resistance to crystal violet and 10-N-nonyl acridine orange and is purportedly an integral plasma membrane protein (22). Although Vba1p, Vba2p, and Vba3p transporters may function at the vacuolar membrane, we investigated their involvement in drug resistance in S. cerevisiae. However, ⌬vba1, ⌬vba2, ⌬vba3, and parent BY4741 cells showed no difference in sensitivity to the following drugs up to the maximum concentra- tion that allowed cell growth: crystal violet (2 g/ml), 3-amino-1,2,4-triazole (100 M), 4-nitroquinoline N-oxide (2 M), and ethidium bromide (10 mM) (data not shown). The functions of Vba4p (encoded by YDR119w) and Vba5p (encoded by YKR105c) have not been reported. The uptake of histidine, lysine, or arginine in intact ⌬vba4 (⌬ydr119w) and ⌬vba5 (⌬kr105c) mutant cells was not remarkably different from that of the parent BY4741. However, global analysis of GFP fusion protein localization in S. cerevisiae (26) indicates that Vba4p (768 residues) with 14 putative transmembrane spans is localized to the vacuolar membrane, likely being involved in the transport of basic amino acids by vacuoles. Furthermore, whereas VBA1 gene expression is induced upon nitrogen starvation, VBA4 gene expression is repressed (Saccharomyces genome data base; www.yeastgenome. org/). The function and subcellular localization of Vba5p (582 residues, with 14 putative transmembrane spans) is important in relation to Vba3p, because most of their sequences are conserved except for the amino-terminal part. It was recently reported that Btn1p, an ortholog of the human Batten disease gene CLN3, is involved in the uptake of arginine into S. cerevisiae vacuoles (27). The amounts of arginine and lysine in the vacuolar pool severely decreased in the ⌬btn1 mutant (27). Interestingly, Btn1p is not a member of the multidrug permease family. We now consider that the basal uptake activities in vacuoles for basic amino acids observed by the triple mutant (Fig. 5) arise from Btn1p as well as the VBA gene products Vba4p and Vba5p. We are currently characterizing the vacuolar amino acid transporter at the molecular level with regard to its multiplicity and regulatory mechanisms, with the goal of understanding how vacuoles contribute to nitrogen metabolism.