Progressive C-terminal deletions of the renal cystine transporter, NBAT, reveal a novel bimodal pattern of functional expression.

Nearly identical proteins (denoted NAA-Tr, rBAT, D2, NBAT), cloned from mammalian kidneys, induce a largely sodium-independent high-affinity transport system for cystine, basic amino acids, and some neutral amino acids in Xenopus oocytes (system b0,+-like). Mutations in the human NBAT gene have been found in several type I cystinurics. In kidney, NBAT is associated with a second, smaller protein (approximately 45 kDa), and this heterodimer has been proposed to be the minimal functional unit of the renal cystine transporter (Wang, Y., and Tate, S. S. (1995) FEBS Lett. 368, 389-392). To delineate regions minimally required for functional expression in oocytes, we constructed a series of C-terminal truncated mutants of rat kidney NBAT (wild-type (WT), 683 amino acids). Expression of these mutants in oocytes yielded an unusual bimodal pattern for the induction of amino acid transport activity. Thus, initial C-terminal truncations aborted elicitation of transport activity. The next mutant in the series, Delta588-683, exhibited most of the transport-inducing potential inherent in the WT/NBAT. Further deletions again attenuated transport activity. Although both the WT/NBAT and the truncated mutant, Delta588-683, induce qualitatively similar transport systems, the two forms of the protein exhibit contrasting sensitivities toward a point mutation in which the cysteine residue at position 111 was mutated to serine. This mutation did not greatly affect induction of transport by the WT/NBAT; however, the Delta588-683 mutant was inactivated by this mutation. Our data further suggest that cysteine 111 is probably the site of disulfide linkage with an approximately 45-kDa oocyte protein producing a complex equivalent to that seen in kidney membranes.

Plasma membrane transport of amino acids is mediated by several transport systems, Na ϩ -dependent and Na ϩ -independent, with overlapping substrate specificities (1). Mammalian proteins that catalyze, or are related to, many of these transport systems have now been cloned and characterized (1)(2)(3). A majority of these are complex membrane proteins containing from 8 to 12 membrane-spanning domains (MSDs). 1 An exception to this general rule is a family of proteins exemplified by NBAT (also denoted NAA-Tr, rBAT, and D2) (4 -8) and 4F2hc (the heavy chain of a mammalian cell surface antigen, 4F2) (9 -12). These proteins induce transport of basic and certain neutral amino acids when their cRNAs are injected into Xenopus oocytes and are predicted to contain only 1-4 MSDs. For this reason, it has been suggested that NBAT and 4F2hc may not function as transporters themselves but may, in fact, be regulatory/modulatory subunits of larger transporter complexes (2,13). There is, however, no direct evidence that the transport mediated by these proteins involves such complexes; but it is noteworthy that NBAT in the renal and intestinal brush border membranes (BBMs), its primary sites of localization (14 -15), is associated with a smaller protein, the association involving one or more interprotein disulfide bonds (16 -17). An equivalent complex is detected in oocytes injected with NBAT cRNA, and such a heterodimer has been proposed to constitute the minimal functional unit of NBAT-mediated transport (16). In mammalian cells, 4F2hc is also found as part of a heterodimeric complex, 4F2, in which it is disulfide-linked to a smaller subunit (18 -20). However, equivalent heterodimeric complexes have not yet been detected in oocytes injected with 4F2hc cRNA.
Although both NBAT and 4F2hc elicit amino acid transport activity in Xenopus oocytes, their precise functions in mammalian cells have still not been defined, although accumulating data indicate that NBAT is involved in renal and intestinal cystine transport. Thus, several mutations in the human NBAT gene have been found in a substantial number of patients with cystinuria (21)(22)(23)(24)(25), an autosomal recessive disease in which excessive amounts of cystine and basic amino acids are excreted in urine (26,27). Recent studies indicate that mutations in NBAT occur only in type I cystinurics but not in Types II and III (28 -30). It, therefore, appears that other gene products in addition to NBAT are also involved in renal cystine transport. We have speculated that one of these might be the protein complexed to NBAT in renal and intestinal membranes (16). The transport systems induced in oocytes by NBAT and 4F2hc exhibit some similarities and certain significant differences. For example, NBAT induces a high-affinity, largely Na ϩindependent uptake system for basic amino acids, cystine, and certain neutral amino acids, a system similar to b 0,ϩ , a Na ϩindependent system first detected in mouse blastocysts (31). 4F2hc, in contrast, induces a transport system resembling system y ϩ L, i.e., Na ϩ -independent uptake of basic amino acids and largely Na ϩ -dependent uptake of neutral amino acids (9 -12). The similarities between the transport induced by the two proteins could be a reflection of the fact that alignment of their sequences reveals a significant degree of identity and similarity (12). Interestingly, a report recently appeared claiming that a truncated form of human NBAT, in which 175 residues from * This work was supported by a grant from the National Institutes of Health (DK51218). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed: Dept. of Biochemistry, Cornell University Medical College, 1300 York Ave., New York, NY 10021. Tel.: 212-746-6410; Fax: 212-746-8875; E-mail: sstate@mail. med.cornell.edu. 1 The abbreviations and trivial names used are: MSD, membranespanning domain; NBAT, neutral and basic amino acid transporter (NAA-Tr; rBAT); BBM, brush border membrane; MSH, 2-mercaptoethanol; WT, wild-type; ⌬588 -683, etc., denote the truncated forms of the C-terminal tail were deleted, induced amino acid transport in oocytes that was qualitatively similar to that elicited by 4F2hc, i.e., system y ϩ L-like (32). To define more closely the structural boundaries at which this switch in specificity occurs and also to elucidate the domains essential for its functional expression, we generated mutants of rat kidney NBAT containing progressive C-terminal deletions. A striking outcome of our studies is that the expression of the truncated mutants in oocytes shows a bimodal pattern for the induction of amino acid transport activity. A preliminary account of this work has appeared as an abstract (48).
EXPERIMENTAL PROCEDURES 3 H-labeled amino acids were from Amersham Pharmacia Biotech. Oocyte-positive Xenopus laevis females were from Nasco (Fort Atkinson, WI). The Riboprobe RNA transcription kit and T7 RNA polymerase were from Promega, and the RNaid kit for the purification of cRNA was from Bio 101, Inc. Restriction enzymes used in this study were purchased from New England Biolabs Inc. and Promega.
Cloning of rat kidney NBAT/cDNA has been described previously (4). The cDNA (cloned into the SalI and NotI sites of pSPORT1 (pSPORT/ NBAT)) contains a 2049-base pair open reading frame (683 amino acids) within a 2259-base pair sequence. For cRNA synthesis, the plasmid was linearized with NotI and transcribed in vitro with T7 RNA polymerase in the presence of the GpppG cap, using the protocol supplied with the Riboprobe transcription kit. The RNA transcripts were purified using the RNaid kit. Production and properties of the anti-peptide antibodies directed against NBAT have been described previously (33). In this study, we used the immunoglobulin fractions from antisera directed against NBAT sequences 357-375 (Ab357) and 527-539 (Ab527) (see Fig. 1). Details of the SDS-PAGE and the ECL Western blotting (Amersham Pharmacia Biotech) procedures have been described (16).
Construction of Mutants- Fig. 1 schematically depicts the proposed 4 MSD topological model for NBAT and the sites of C-terminal truncations (labeled T1 through T5). The truncations were effected by single base changes in pSPORT/NBAT which converted codons for selected amino acid residues to stop codons. The mutations were introduced using reagents and protocol supplied with the QuikChange Site-Directed Mutagenesis Kit from Stratagene and pairs of custom-synthesized complementary mutagenic oligonucleotides (from GeneLink, Thornwood, NY). The deletions (⌬) effected, along with the amino acid residues, codons for which were mutated to stop codons, were as follows: ⌬658 -683, K658X; ⌬615-683, K615X; ⌬588 -683, K588X; ⌬566 -683, W566X; and ⌬508 -683, W508X. Fragments from the mutant cDNAs between the restriction sites for XhoI (base 1393 of the cDNA) and NotI (downstream of cDNA) were purified by agar-gel electrophoresis and ligated into the wild-type (WT) pSPORT/NBAT from which the XhoI/ NotI fragment had been deleted. Mutations were confirmed by complete sequencing of the XhoI/NotI cassettes.
The C111S (Cys-111 to Ser-111) mutant of the WT/NBAT (NBAT/ C111S) was constructed by site-directed mutagenesis using the PCR overlapping extension procedure with primers containing the desired mutation (34,35). The Cys to Ser substitution was used because this substitution is conservative and not likely in and of itself to distort the conformational structure of the protein (36,37). The mutagenic primers were 5Ј-CCAAAAT[C]CCTTGACTGGTGG-3Ј and its reverse complement (the mutated nucleotide is shown within square brackets). The mutation was carried out in two steps. In the first step, the fragment to be overlapped was amplified in separate reactions in which the sense mutated primer was paired with a downstream unmutated antisense primer (5Ј-GTCCAGCTTCTCTTGGATACC-3Ј, complementary to bases 459 -479 in NBAT cDNA), and the antisense mutated primer was paired with an upstream vector sequence (5Ј-TAATACGACTCACTAT-AGGG-3Ј, T7 promoter primer). In both reactions, 660 ng of pSPORT/ NBAT served as a template. In the second step, a small amount of each product from the first reactions (purified by agarose-gel electrophoresis), together with the T7 promoter primer and the downstream wildtype antisense primer, was used for the PCR reaction (using a Gene-Amp PCR kit from Perkin Elmer). The PCR product was purified by agar-gel electrophoresis, digested with SalI/BglII, and ligated by T4 DNA ligase into the pSPORT/NBAT from which the SalI/BglII fragment had been deleted (BglII restricts NBAT cDNA at nucleotide 402, i.e. downstream of Cys-111). The mutation was confirmed by doublestranded DNA sequencing.
The C111S mutant of the ⌬588 -683 truncated form of NBAT (⌬588 -683/C111S) was made as follows: the fragment obtained by treatment of the ⌬588 -683 mutant with BglII and NotI was purified and ligated into NBAT/C111S plasmid from which the BglII/NotI cassette had been deleted.
DNA sequencing was performed either by the Protein/DNA Technology Center of the Rockefeller University, New York, NY or by the DNA Sequencing Facility at the BioResearch Center of Cornell University, Ithaca, NY.
Expression of NBAT and Its Mutants in Xenopus laevis Oocytes-Isolation and maintenance of oocytes and amino acid uptake measurements have been described previously (4). The oocytes were injected with 45 nl of the desired cRNA solution. The rate of uptake of amino acids was measured at 20°C in a medium containing 0.015 M Tris-HCl, pH 7.5, and either 0.1 M NaCl (for transport in presence of Na ϩ ) or 0.2 M sucrose (for transport in absence of Na ϩ ), and 50 M (unless otherwise stated) of the 3 H-labeled amino acid (L-arginine, as a representative of the basic amino acid substrates, and either L-leucine or L-phenylalanine as a neutral amino acid substrate). Note that previous studies have shown that transport of L-cystine into NBAT cRNA-injected oocytes displays characteristics similar to those seen for either Arg, Leu, or Phe (6, 7). Generally, six oocytes were incubated in 1 ml of the transport medium. The uptake is expressed as picomoles of amino acid transported into each oocyte (after subtraction of uptake into uninjected oocytes as controls; we have found that amino acid uptake rates with uninjected oocytes were similar to those obtained with water-injected oocytes) and is presented as the average of individual measurements on five to six oocytes Ϯ S.E.
Characterization of Proteins Expressed in Oocytes-Oocyte membranes were isolated as described previously (16). The membrane pellet (from two to four oocytes) was suspended either in 25 l of 0.125 M Tris-HCl buffer (pH 6.8) containing 4% SDS (Tris-SDS) (nonreducing conditions) or in 25 l of the Tris-SDS buffer containing 5% 2-mercaptoethanol (MSH) (reducing conditions) and heated at 100°C for 5 min. The SDS extracts were then subjected to SDS-PAGE (10 or 12% gels) followed by Western analyses. NBAT and its mutants, blotted onto the nitrocellulose membranes, were detected using anti-NBAT antibodies, Ab357 and Ab527, either used individually or in combination, and horseradish peroxidase-conjugated donkey anti-rabbit IgG, from Amersham Pharmacia Biotech, followed by the ECL Western blot detection procedure as described in the protocol supplied by them.
Detection of Expressed Proteins on Oocyte Surface by Confocal Microscopy-Three days after the injection of cRNA, 4 -6 oocytes were suspended in 1 ml of PBS containing 3.3% paraformaldehyde and incubated at 20°C for 15 min. Uninjected oocytes were similarly treated and served as controls. Oocytes were extensively rinsed with PBS and kept overnight at 4°C in 1 ml of PBS containing 2% non-fat dry milk. The oocytes were then washed and incubated at 20°C for 1 h in 2 ml of PBS containing 0.5% dry milk and Ab357 (1:1,000, final dilution). After thorough washings, the oocytes were incubated in 2 ml of PBS containing rhodamine-conjugated goat anti-rabbit IgG (from Pierce; 1:400, final dilution) to label the surface-bound anti-NBAT antibodies. The labeled oocyte was then placed on a 1-cm 2 glass slide. Sequential focal planes of the fluorescence-labeled oocyte were imaged with an MRC600 laser scanning confocal unit (Bio-Rad, Hercules, CA) attached to a Zeiss Axiovert microscope. The oocyte was oriented such that its vegetal (white) hemisphere faced the light source. The oocyte was excited with the 514-nm line from a 25-mW Argon ion laser, and standard rhodamine optics were employed. A ϫ10 objective with a 0.25 numerical aperture was used to obtain an optical section every 20 m. (See Ref. 38 for details regarding confocal microscopy.)

Expression and Characterization of C-terminal Truncated
Mutants of NBAT-The proposed four MSD topological model for NBAT is schematically depicted in Fig. 1 (adapted from Ref. 33). The sites at which the truncations were effected are labeled T1 through T5. Fig. 2 is a representative Western blot of an SDS gel, run under reducing conditions, showing the monomers of the WT/NBAT and its truncated forms in membranes of oocytes injected with the appropriate cRNA (10 ng/oocyte). Blot A was probed with Ab357 directed against an epitope preserved in all of the truncated forms of NBAT. Blot B was probed with Ab527 which is directed against an epitope deleted from the mutant ⌬508 -683. Therefore, the band corresponding to this mutant is not seen (lane 2), whereas the other mutants are readily observed. Ab527 also reacts with an approximately 62-kDa protein in uninjected as well as the injected oocytes. This protein, thus, appears to be an endogenous oocyte membrane protein whose identity is as yet unknown. As noted previously (39), the WT/NBAT monomers migrate as two closely spaced bands (approximately 80 and 83 kDa, respectively) (lane 1 in each of the two blots), shown to be because of differential glycosylation. The monomeric species of the trun-cated forms exhibit the approximate expected mass, ⌬508 -683, the smallest of the mutants being approximately 58 kDa (this form lacks two of the five potential N-glycosylation sites of WT/NBAT). Noticeable also is a progressive decrease in the amount of expressed protein with longer truncations. Fig. 3 shows the induced Arg uptake activity in oocytes at 3 and 5 days after injection of the cRNAs. A surprising bimodal pattern for the induction of transport activity is revealed. Deletion of as few as 25 C-terminal amino acids virtually abolished the ability of the resultant truncated form (⌬658 -683 mutant) to induce amino acid transport. The next in the series of truncated mutants (⌬615-683) was also devoid of transport activity. However, further deletion of 27 amino acids (⌬588 -683) restored the ability of the mutant to induce transport (to about 75% of that elicited by the WT/NBAT). Note that in this truncated form, the fourth putative MSD has been deleted. Further truncations resulting in the ⌬566 -683 and ⌬508 -683 forms again abrogate the ability to induce transport. It is noteworthy that the mutation to a stop codon that generates the ⌬566 -683 form also disrupts the third heptad repeat in the leucine zipper motif (residues 545 to 572 in the WT/NBAT), an ␣-helical coiled-coil structure shown to be involved in oligomerization of a wide variety of proteins (see Ref. 40 and references contained therein). A parallel bimodal pattern of expression of transport activity was also seen when either Phe or Leu was used as a transport substrate instead of Arg (data not shown).
A more detailed investigation of the transport induced in oocytes by NBAT has shown that this transport is in fact an exchange transport in which uptake of a cationic amino acid is enhanced by the outward flux of neutral amino acids and uptake of a neutral amino acid is stimulated by counter-transport of a cationic substrate (41,42). It was, therefore, essential to determine whether the bimodal pattern of expression also applies to the exchange transport activity. cRNA-injected oocytes were pre-loaded with either [ 3 H]Arg or [ 3 H]Leu. Stimulation of the Arg and Leu efflux by addition of unlabeled Leu and Arg, respectively, to the extracellular medium was then determined (Fig. 4). With both substrates, the stimulated efflux from oocytes expressing the truncated mutant, ⌬588 -683, was about 70% of that seen with the oocytes expressing the WT/ NBAT. No difference in the rates of efflux of either Arg or Leu were seen between the uninjected oocytes and the oocytes ex-  (33). Approximate end points of the MSDs are numbered. Putative N-glycosylation sites are marked by the symbol "Y". Heavy bars represent the epitopes recognized by Ab357 and Ab527 (A and B, respectively). T1-T5 represent location of the amino acid residues mutated to stop codons in the construction of the C-terminal truncation mutants (⌬658 -683, ⌬615-683, ⌬588 -683, ⌬566 -683, and ⌬508 -683, respectively). The X symbols denote the approximate position of a leucine zipper motif in NBAT sequence (note that the line representing the protein is not drawn to scale). pressing the inactive truncated mutants; also, no significant stimulation of efflux was observed upon addition of an extracellular amino acid to these oocytes (Fig. 4, solid circles). The efflux rates observed in these oocytes were about 1/7th-1/6th of the stimulated rates seen with either the WT/NBAT or the ⌬588 -683 mutant expressing oocytes. Thus, the bimodal pattern of functional expression of the C-terminal truncated forms also applies to the exchange transport. Fig. 5 shows that both the WT/NBAT and its ⌬588 -683 mutant induce the same transport phenotype in oocytes (system b 0,ϩ -like). Thus, a large fraction of the induced uptake of both Arg and Phe (as well as of Leu) with both forms of this protein is Na ϩ -independent. As also noted in several previous studies (4,8,13,23), the NBAT-induced transport is stimulated about 20 -30% by Na ϩ (thus, the term "system b 0,ϩ -like" used to describe the NBAT-induced transport). Whether this is a consequence of induction in oocytes of more than one amino acid transport system by NBAT is not clear at present. It is pertinent, however, that the transport activities induced by the ⌬588 -683 mutant are also similarly affected by Na ϩ . With both forms of the protein, the uptake of Arg is inhibited by either Phe or Leu, and the uptake of Phe is inhibited by Arg. No significant difference was found in the K m values for Arg (K m values, 20 to 30 M) and Leu (K m values, 30 to 40 M) between the WT/NBAT and its ⌬588 -683 mutant. However, the transport V max for the truncated form with either Arg, Leu, or Phe was consistently about 75% of that observed for the WT/NBAT. In separate experiments (data not shown), we showed that the rate of induced transport in oocytes by the WT/NBAT cRNA reaches a maximum value between 2.5 and 5 ng of cRNA per oocyte although the amount of protein expressed in oocyte membranes continues to increase with increasing amounts of injected cRNA (each group of oocytes in this experiment were injected with 1, 2.5, 5, 10, 25, and 50 ng of cRNA/oocyte, respectively). This divergence between induced transport and the amount of protein expressed has been noted previously (39,43). To explain this phenomenon, it has been proposed that an endogenous oocyte protein is recruited by the newly synthesized NBAT subunits to produce a transport-competent complex. If so, the upper limit of induced transport should be determined by the amount of the recruited oocyte protein and not by the total NBAT protein synthesized. With the truncated form, ⌬588 -683, also, we found that the maximum transport was attained between 2.5 and 5 ng cRNA/oocyte (data not shown). Therefore, for kinetic measurements, the oocytes were injected with amounts of each cRNA (10 ng/oocyte) that were well above the amounts necessary for induction of maximum transport activity. Thus, the difference in the transport V max observed between the two forms must be a reflection of the inherent differences between the WT and the ⌬588 -683 mutant and not a consequence of any minor differences in the amount of the two proteins synthesized in oocytes.
C111S Mutants of WT/NBAT and Its Truncated Form, ⌬588 -683-As noted (see above) NBAT displays significant homology with the heavy chain (4F2hc) of the cell surface antigen 4F2. Furthermore, both NBAT and 4F2hc possess a cysteine residue in very similar location, i.e., just C-terminal to the first MSD (residues 111 and 103, respectively, in rat NBAT (4) and 4F2hc (11)). NBAT and 4F2hc from other species also possess a Cys in a similar location, leading to the speculation that this conserved Cys might be involved in disulfide linkages between the two proteins and their respective associated subunits (12). To test this hypothesis in NBAT, we mutated Cys-111 to Ser in both the WT/NBAT and its truncated form, ⌬588 -683.
The transport of Arg and Leu induced in oocytes by the C111S mutant of WT/NBAT (NBAT/C111S) was qualitatively similar to that induced by the WT/NBAT (Table I). The time course of expression of transport by the two cRNAs was also similar, and the mutation had virtually no effect on the apparent transport K m values for either Arg or Leu. The rate of uptake (as well as the transport V max ) in the mutant cRNAinjected oocytes for each of the two amino acids was, however, consistently about 30% lower than the corresponding rates in the WT cRNA-injected oocytes. Oocytes injected with equal amounts of the WT/NBAT and NBAT/C111S cRNA express approximately the same amount of the respective proteins (Fig.  6C). Significant differences were, however, observed in the nature of the oligomeric species produced by the WT/NBAT and the NBAT/C111S mutant. Thus, when SDS-PAGE was carried out under nonreducing conditions (Fig. 6A), a number of high molecular weight species (170 kDa and greater) containing NBAT were detected in oocytes injected with the WT/NBAT cRNA (Fig. 6A, lane 2). Some of these are absent from membranes of the NBAT/C111S cRNA-injected oocytes (Fig. 6A,  lane 3). These oligomeric species are disulfide-linked because none of them can be detected when the SDS-PAGE was carried out following treatment of oocyte membranes with SDS in the presence of MSH (Fig. 6C, lanes 1 and 2). The 170 -180 kDa bands and those greater than 220 kDa are most likely, as indicated previously (16,17), homopolymers of the NBAT subunit. In the case of the NBAT/C111S mutant, the most prominent oligomer band is about 170 kDa (Fig. 6A, lane 3), a size expected for a homodimer of the NBAT subunit. More interesting, however, is the finding that a species with a molecular mass of about 125 kDa, prominent in WT/NBAT cRNA-injected oocytes, is not detected in NBAT/C111S cRNA-injected oocytes (Fig. 6A, lanes 2 and 3, closed arrowhead). As suggested previously (16), this species has the size consistent with that expected for the disulfide-linked heterodimer (equivalent to those seen in the kidney and intestinal BBMs) containing an NBAT subunit and a protein of about 45 kDa. 2 The apparent inability to form this species from the C111S mutant suggests that Cys-111 is most likely necessary for the disulfide linkage between the two subunits of this heterodimer. If such a heterodimer is indeed the minimal functional unit of the NBATinduced transport, then one must conclude that a disulfide linkage between the two subunits is not a prerequisite for the transport activity because the activity was not abolished by the

TABLE I Characteristics of the amino acid transport expressed in oocytes by
NBAT and its mutant forms Each group of oocytes was injected with the cRNA for either the WT/NBAT, its C111S mutant, the truncated mutant, ⌬588 -683, or its C111S mutant (⌬588 -683/C111S) (10 ng/oocyte). The uptake of Arg and Leu was determined 3 days after the injection of the cRNA either in Tris-sucrose (ϪNa ϩ ) or in Tris-NaCl (ϩNa ϩ ) (see "Experimental Procedures"). The results shown are averages of uptake into six oocytes from each group for a typical experiment (after subtraction of uptake into uninjected oocytes). The standard error in these experiments ranged from Ϯ8 to 12%. C111S mutation. Presumably, the subunits may still interact via other, noncovalent, interactions to generate the active complex. The higher V max for the WT/NBAT, however, suggests that the disulfide linkage between the two subunits might fine tune the complex for optimum activity.
In striking contrast to the findings described above, the C111S mutation in the ⌬588 -683 truncated form (⌬588 -683/ C111S) almost completely abolished its ability to induce amino acid transport (Table I), although there was virtually no difference in the level of expression of the two mutant proteins (Fig.  6C, lanes 3 and 4). The truncated forms exhibited a reduced tendency to form disulfide-linked oligomers in contrast to the corresponding full-length proteins (compare Fig. 6B with 6A). There is also a significant increase in the intensity of a band at about 150 kDa (presumably a homodimer) in the ⌬588 -683/ C111S cRNA-injected oocytes (Fig. 6B, lane 2 versus lane 1). A weak band is seen at about 116 kDa in ⌬588 -683 cRNAinjected oocytes (Fig. 6B, lane 1). This is the size expected for a heterodimer between the ⌬588 -683 monomer and an oocyte protein of about 45 kDa (equivalent to the 125-kDa species seen in WT/NBAT cRNA-injected oocytes, Fig. 6A, lane 2). This band is barely detectable in ⌬588 -683/C111S cRNA-injected oocytes (Fig. 6B, lane 2).
Immunocytochemical localization studies on intact oocytes using confocal microscopy revealed that the WT/NBAT, the ⌬588 -683 truncated form, and their respective C111S mutants are expressed to about the same density on the external surface of oocytes injected with equal amounts of the respective cRNA (Fig. 7). The inability of ⌬588 -683/C111S mutant to induce transport cannot, therefore, be because of any defect in its targeting to the oocyte surface.

DISCUSSION
The most striking result of our studies is that progressive C-terminal truncations of rat NBAT are associated with a bimodal pattern for the expression of amino acid transport in Xenopus oocytes. The C-terminal deletions first abrogate the ability of the truncated forms to induce amino acid transport. Further deletion restored this function. The phenotypical characteristics of the transport induced by this mutant (uptake into oocytes as well as exchange transport) resembled those of the transport induced by the WT/NBAT. Further C-terminal deletions from this truncated form again attenuated the transport activity. These results tentatively identify two regions of NBAT which appear to be critical for its functional expression: one, the 26 residues at the C terminus (residues 658 to 683); and two, a cassette of approximately 43 amino acids containing the leucine zipper motif followed by the stretch of amino acids between this motif and the start of the fourth MSD (residues 545 to 587) (see Fig. 1). These two regions surround the fourth MSD, deletion of which reverses the inhibitory effect of truncation of the cytoplasmic C-terminal tail.
Detailed understanding of the structural basis for the observed bimodal pattern of expression requires further studies. Nevertheless, our observations suggest that elements of the C-terminal tail provide the guiding force for holding NBAT in an active conformation, perhaps by providing the sites for interaction with other regions of NBAT itself or with the associated protein in the proposed functional complex. Deletion of this C-terminal tail presumably removes this influence, forcing the truncated mutant into a nonproductive conformation. This inactive structure could be stabilized by the tethering influence of the fourth MSD because deletion of this domain restores the ability to induce transport. The fact that disruption of the leucine zipper motif leads to inactivation suggests a possible role of this motif in the functional expression of NBAT. Further studies are needed to assess the importance of this motif.
Although the phenotypic characteristics of amino acid transport induced by the WT/NBAT and ⌬588 -683 mutant are similar, a notable difference between the two forms is revealed when one of their cysteine residues (Cys-111) is mutated to serine. In WT/NBAT, this mutation reduces transport activity by about 30% but has no effect on substrate K m values. Comparison of the disulfide-linked oligomers produced from the WT/NBAT and its C111S mutant indicates that Cys-111 is critical for the formation of a disulfide-linked heterodimer (about 125 kDa), equivalent to that found in renal BBMs and which has been proposed to be the functional unit of NBATmediated transport (16). However, the disulfide-linkage involving Cys-111 does not appear to be obligatory for induction of transport activity because the C111S mutant is still active. In contrast, however, the C111S mutation in the ⌬588 -683 form completely abolishes its transport activity. One possible explanation for this finding might be that, lacking C111, the ⌬588 -683/C111S mutant cannot form a stable complex with the accessory subunit. Another possibility is that the WT and the truncated subunits associate with different oocyte proteins and that the role of Cys-111 in the complex originating from ⌬588 -683 subunit is different from its role in the complex containing the WT subunit. Detailed understanding of the molecular basis for the findings reported here must await cloning and characterization of the NBAT-associated protein(s).
Our proposal that the transport induced in oocytes by NBAT requires its association with endogenous oocyte proteins (16) has parallels in other transport systems. Thus, formation of functional potassium channels in oocytes from expressed mammalian MinK and GIRK1 subunits appears to involve their interaction with proteins of oocyte origin (44,45). Similarly, generation of an active Na ϩ /K ϩ ATPase from injected ␤ subunit cRNA alone has been shown to be dependent on association of the newly synthesized ␤ subunits with pre-existing oocyte ␣ subunits (46). Of relevance to amino acid transport is the proposal that expression of more than one kinetically distinguishable transport process in oocytes by the murine cationic amino acid transporter (mCAT) is because of its association with more than one type of oocyte protein (47). Furthermore, induction of amino acid transport in oocytes by 4F2hc may also involve its association with an oocyte protein.
As noted above, Miyamoto et al. (32) recently reported that a ⌬511-685 deletion mutant of human kidney NBAT induced a system y ϩ L-like transport in oocytes akin to that induced by 4F2hc. The equivalent deletion mutant of rat NBAT, ⌬508 -683, in contrast, was unable to induce any noticeable transport activity in oocytes. These conflicting data cannot readily be ascribed to differences in primary structures of human and rat NBATs because the two proteins are highly homologous, displaying greater than 80% amino acid sequence identity and nearly 90% similarity (6,7), and induce similar amino acid transport phenotypes in oocytes. It should be noted that Miyamoto et al. did not study the effects of shorter C-terminal truncations of human NBAT. It would be of interest to determine whether a bimodal pattern of functional expression is also observed with the human NBAT.
The relevance of our findings with the truncated forms of NBAT to cystinuria cannot be assessed at present. However, it is of interest to note that one of the mutations detected in the NBAT gene of a cystinuric involved deletion of the nucleotide A at position 1749 (22). This frameshift mutation creates a stop signal 14 codons downstream of the deletion, which should result in a truncated form of human NBAT missing the 88 C-terminal amino acids. Note that the site of truncation, thus, falls in the midst of the fourth MSD, making this truncated mutant approximately equivalent to the active ⌬588 -683 mutant of rat NBAT described in our study. It would be of interest to determine whether such a truncated form of human NBAT would be active in inducing transport in oocytes. Other mutations leading to premature stop codons in human NBAT gene have also been described (E483X (25) and R270X (22)). The truncated products of such mutant genes, based on our studies, would almost certainly be inactive. Several mutations detected in the human NBAT gene are clustered in the C-terminal tail (21,22), which our studies have shown to be critical for expression of transport activity in oocytes. The effect of such mutations on expression of transport in oocytes needs to be evaluated.