Physiol Rev Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Physiol. Rev. 78: 487-545, 1998;
0031-9333/98 $15.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by DEVÉS, R.
Right arrow Articles by BOYD, C. A. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by DEVÉS, R.
Right arrow Articles by BOYD, C. A. R.

PHYSIOLOGICAL REVIEWS   Vol. 78 No. 2 April 1998, pp. 487-545
Copyright ©1998 by the American Physiological Society

Transporters for Cationic Amino Acids in Animal Cells: Discovery, Structure, and Function

R. DEVÉS AND C. A. R. BOYD

Programa de Fisiología y Biofísica, Instituto de Ciencias Biomédicas, Universidad de Chile, Santiago, Chile; and Department of Human Anatomy and Genetics, University of Oxford, Oxford, United Kingdom

I. INTRODUCTION
II. IDENTIFICATION OF DISTINCT TRANSPORT ACTIVITIES FOR CATIONIC AMINO ACIDS
    A. System y+: the Na+-Independent Cationic Amino Acid Transporter
    B. Systems B0,+ and b0,+: the Na+-Dependent and Na+-Independent Broad-Scope Transporters
    C. System y+L: the Cation-Modulated Broad-Scope Transporter
III. IDENTIFICATION OF COMPLEMENTARY DEOXYRIBONUCLEIC ACID SEQUENCES ENCODING PROTEINS INVOLVED IN CATIONIC AMINO ACID TRANSPORT
    A. Cationic Amino Acid-Specific Transporters: the CAT Family
    B. Broad-Scope Amino Acid Transport Proteins: the BAT Family
IV. SPECIFICITY AND MECHANISM
    A. Kinetic Theory and Methodology
    B. Functional Properties of Cationic Amino Acid Transporters
V. PHYSIOLOGY OF CATIONIC AMINO ACID TRANSPORT
    A. Nutrition
    B. Epithelial Transport of Cationic Amino Acids
    C. Role of Cationic Amino Acid Transporters in Cells of the Immune System
    D. Regulation of Cationic Amino Acid Transport
VI. CONCLUDING REMARKS
REFERENCES

    ABSTRACT
Top
References

Devés, R., and C. A. R. Boyd. Transporters for Cationic Amino Acids in Animal Cells: Discovery, Structure, and Function. Physiol. Rev. 78: 487-545, 1998. --- The structure and function of the four cationic amino acid transporters identified in animal cells are discussed. The systems differ in specificity, cation dependence, and physiological role. One of them, system y+, is selective for cationic amino acids, whereas the others (B0,+, b0,+, and y+L) also accept neutral amino acids. In recent years, cDNA clones related to these activities have been isolated. Thus two families of proteins have been identified: 1) CAT or cationic amino acid transporters and 2) BAT or broad-scope transport proteins. In the CAT family, three genes encode for four different isoforms [CAT-1, CAT-2A, CAT-2(B) and CAT-3]; these are ~70-kDa proteins with multiple transmembrane segments (12-14), and despite their structural similarity, they differ in tissue distribution, kinetics, and regulatory properties. System y+ is the expression of the activity of CAT transporters. The BAT family includes two isoforms (rBAT and 4F2hc); these are 59- to 78-kDa proteins with one to four membrane-spanning segments, and it has been proposed that these proteins act as transport regulators. The expression of rBAT and 4F2hc induces system b0,+ and system y+L activity in Xenopus laevis oocytes, respectively. The roles of these transporters in nutrition, endocrinology, nitric oxide biology, and immunology, as well as in the genetic diseases cystinuria and lysinuric protein intolerance, are reviewed. Experimental strategies, which can be used in the kinetic characterization of coexpressed transporters, are also discussed.

    I. INTRODUCTION
Top
Next
References

The transport of cationic amino acids across mammalian cell membranes was, for a long time, thought to occur through a single transporter referred to as system y+. Thus, for more than 20 years, evidence related to mechanistic or physiological questions was interpreted assuming that basic amino acids were transported largely by this system in a number of different cell types (reviewed in Refs. 44, 59, 90, 194, 237, 253).

In the past few years, the use of new molecular and kinetic experimental approaches has unveiled a more complex picture, involving various clearly distinct transporters that show differences in structure, substrate specificity, mechanism, site of expression, and regulation (reviewed in Refs. 18, 52, 123, 125, 127, 129, 157, 158, 159, 208, 218). These findings have raised new questions, as well as the possibility for new interpretations of early observations.

Here we review information generated in different areas in which understanding of cationic amino acid transport is being developed, from the molecular to the kinetic and from the strictly mechanistic to the physiological.

We start by discussing from a historical perspective the functional description of several transport systems. Our aim is not to provide a historical record, but rather to revisit some of the classical papers in the light of the new observations. In general, emphasis has been placed on primary experimental evidence rather than on the original interpretations of the findings at the time when they were reported, and thus we have not held back from showing the data from relevant early studies.

We also present the new molecular evidence that has led to the description of two families of proteins that are involved in cationic amino acid transport (recently reviewed in Refs. 52, 123, 127, 158, 159, 208). This work has given us the primary amino acid sequences of a number of molecules about which nothing was known 7 years ago. In addition, it has suggested their putative membrane topologies, shown their sites of expression, and in some cases, shed light on novel transport mechanisms. We have sought to relate these proteins to the transport systems previously identified on the basis of functional studies, some of which have been carried out over the last three decades (reviewed in Refs. 44, 194), and others only much more recently (64, 222, 226). Then, with both the molecular and kinetic evidence at hand, we discuss the functional and, when possible, the mechanistic properties of the different transporters.

This analysis is preceded by a discussion of the kinetic methodology that can be used, first, to discriminate between transporters that operate in parallel and, second, to disclose the interactions between a transporter and its substrates. This approach should be useful in the design of future experiments, not only in relation to amino acid transport, but in other areas of transport physiology as well. It might be argued that molecular methods will obviate the problems of interpretation, which result from the presence of multiple transporters with overlapping specificities. However, this limitation also applies to transport studies in expression systems, and it is especially important in the design of successful cloning strategies.

Finally, we have reviewed selected areas of physiology in which the function of a given tissue is directly influenced by the activity of one or more cationic amino acid transporters. The biological importance of these is exemplified in relation to a variety of topics including nutrition, endocrinology, immunology, and genetic disease. The discovery that nitric oxide (NO) plays a central role as a mediator of both regulatory and cytotoxic functions and the fact that arginine is its immediate precursor place cationic amino acid transport in a new context (150).

In view of the current widespread availability of bibliographic data bases, we have not attempted a comprehensive review. Instead, we have discussed with detail those observations that appeared more relevant to us, with special emphasis in the experimental strategies that have been used in disclosing the new facts. That is why original experiments are amply shown.

More generally, this review of a tightly defined topic exemplifies the nature of scientific discovery as physiology encounters molecular biology. As the "postsequencing era" in biology and medicine emerges, the subject of this review provides a paradigm for the unpredicted different ways in which gene function has been discovered; we have tried to let the reader see the primary evidence for this.

    II. IDENTIFICATION OF DISTINCT TRANSPORT ACTIVITIES FOR CATIONIC AMINO ACIDS
Top
Previous
Next
References

Traditionally, amino acid transport systems have been classified considering two main criteria: 1) the substrate specificity, i.e., which amino acid or group of amino acids is transported by the system, and 2) the Na+ dependency of the rate of transport, generally defined with reference to the rate measured in the presence of K+ or choline salts. Four transport systems for cationic amino acids have been defined according to these criteria, and they are listed in Table 1.

 
View this table:
[in this window] [in a new window]
  		TABLE 1.   First description of cationic amino acid transport systems

Our definition of a "cationic amino acid transport system" applies to transporters exhibiting affinities and translocation rates for cationic amino acids, which are higher than or equivalent to those for other types of amino acids. Thus transporters that interact only weakly with dibasic amino acids, such as systems ASC (254) and asc (74), will be excluded from any further analysis.

Only one of the four systems (y+) listed in Table 1 is selective for cationic amino acids (although it does present weak interaction with neutral amino acids) (48, 51, 239). The other three systems (B0,+, b0,+, and y+L) accept a wider range of substrates, including cationic and neutral amino acids; they differ, however, in their interactions with inorganic monovalent ions. System B0,+ is Na+ dependent and will not operate at an appreciable rate, unless this cation is present (226). System b0,+ is Na+ independent, and it functions in the presence of Na+, K+, choline, or Li+ salts (222). Finally, system y+L exhibits a more complex pattern in its cation interaction; whereas the transport of lysine through this system is unaffected by Na+ replacement, the affinity of system y+L toward neutral amino acids is dramatically decreased if Na+ in the medium is substituted with K+ (64).

In this section, we examine the first evidence that led to the identification of these four transport systems.

A. System y+: the Na+-Independent Cationic Amino Acid Transporter

The concept of system y+ as a Na+-independent transporter for cationic amino acids, which interacts less strongly with neutral amino acids (but only when Na+ is present), originated from the early work of Christensen and co-workers (48, 49) with Ehrlich cells and reticulocytes. Afterward, it became the paradigm for cationic amino acid transport and was extended to nearly all other cells and tissues analyzed (see Ref. 237 for a comprehensive review). The more recent evidence showing that some transport systems are able to recognize cationic and neutral amino acids with comparably high affinity, whereas others exhibit a marked preference for cationic amino acids (64, 222, 226), poses the question as to whether the earlier observations reflect the operation of a single transport entity (system y+). It is possible that the properties originally attributed to system y+ represent the sum of at least two separate transport activities, differing in their affinities toward neutral amino acids.

The importance of this question calls for a reexamination of the early data, especially considering that 30 years ago the existence of multiple transport systems for cationic amino acids was also considered, although later rejected, by Christensen and co-workers (42, 50) in studies with Ehrlich cells.

In the original paper describing the entry of labeled lysine into Ehrlich cells, Christensen (42) noted that lysine influx was partially, but substantially (60-70%), inhibited by phenylalanine, whereas lysine was able to eliminate a small portion of phenylalanine uptake (20%) (Fig. 1). It was concluded that there was a "lysine accepting system" that could also recognize neutral amino acids with specificity resembling that of system L for neutral amino acids (leucine, methionine, and phenylalanine); thus the authors referred to this system as a "positive-charge tolerant variant of the L system" and tentatively named it "system L+." In a subsequent paper (51), it was shown that at least two pathways for cationic amino acids could be distinguished: the former "lysine-accepting agency" (L+) and a second route, which resisted inhibition by phenylalanine, referred to as the "lysine-preferring agency." The latter system was by definition selective for cationic amino acids and was named "Ly+."


View larger version (14K):
[in this window]
[in a new window]
 

View larger version (21K):
[in this window]
[in a new window]
  		FIG. 1.   Cationic and neutral amino acids compete for transport in Ehrlich cells. A: inhibition of L-[14C]lysine (1 mM) influx by phenylalanine. B: inhibition of L-[14C]phenylalanine (1 mM) influx by lysine. Arrow indicates additional inhibition produced by 10 mM aminoisobutyric acid (AIB) and L-phenylalanine. External medium was Krebs-Ringer bicarbonate solution (pH 7.4, 37°C). [Modified from Christensen (42).]

The investigation of lysine transport was then extended to rabbit reticulocytes (48). A saturable transport process for lysine, arginine, and ornithine could also be demonstrated in these cells, but the observations differed in two important respects from those in Ehrlich cells: 1) the strength of the interaction of this system with neutral amino acids such as leucine, phenylalanine, and methionine was much lower [the apparent inhibition constants (Ki) for these amino acids in reticulocytes were ~50 mM]; and 2) cationic amino acids did not affect the influx of neutral amino acids. The authors concluded, correctly, that there was no basis to propose the existence of a system whose receptor site did not differentiate between cationic and neutral side chains. However, instead of recognizing the differences between the transport phenotypes of Ehrlich cells and reticulocytes, an attempt to unify the observations in these two experimental systems was made. This led to the proposal that the movement of cationic amino acids across these membranes occurred via a single transport system (system Ly+) that was able to bind cationic and neutral amino acids (50). Interestingly, in both types of cells, the association with neutral amino acids was reduced when Na+ in the medium was replaced by choline (48, 50), but as discussed in section IVB4, they showed differences in their interactions with K+ (213).

We now reexamine the evidence that led Christensen and co-workers to reject their original hypothesis which stated 1) that two transport systems for cationic amino acids, differing in their specificities toward neutral amino acids, were necessary to account for the transport properties of Ehrlich cells; and 2) that one of these systems was able to transport with equivalent efficiency both neutral and cationic amino acids.

First, the inhibition of lysine influx by amino acids with nonpolar side chains of medium size (phenylalanine, leucine, and methionine) was initially associated with the operation of system L (51). In a logical next step, the interaction between cationic substrates and the recently discovered cyclic amino acid analog 2-amino-endo-bicyclo[2.2.1]heptane-2-carboxylic acid (BCH), a specific substrate for system L, was investigated (50). It was found that the influx of BCH into Ehrlich cells was not affected by lysine. It was concluded (correctly) that "cationic amino acids have exceedingly little effect on uptake by the L system" (cited from Ref. 50). However, the experiment did not give evidence against a system that would transport lysine, methionine, leucine, and phenylalanine (but not BCH), and this possibility was not recognized.

Second, homoarginine was found to be a weak inhibitor of phenylalanine fluxes, affecting 3-8% of the flux, although most of this effect was already apparent at 3 mM homoarginine (50). Phenylalanine, on the other hand, inhibited a larger fraction of homoarginine flux (60%) (49). The lack of correspondence in the magnitude of these two effects was interpreted as reflecting the inability of neutral amino acids to interact with the cationic amino acid transporter. Again, although the experiment showed that cationic amino acids did not affect the major route for neutral amino acid entry (presumably system L), it did not rule out the existence of a broad-scope system (such as L+) in Ehrlich cells.

Independent evidence for a transport system with marked selectivity toward cationic amino acids (as found for reticulocytes) is to be found in the work of White et al. (239) with cultured human fibroblasts. It was here that the term system y+ was introduced to indicate that the system did not serve for lysine exclusively, but also for arginine, ornithine, and other cationic analogs. Arginine influx into human fibroblasts was seen to be independent of Na+ and pH and was inhibited by neutral amino acids with low affinity (Fig. 2). Homoserine was significantly more effective than serine or glutamine, and these interactions were undetectable when Na+ was replaced by choline. This behavior resembles more closely the observations in reticulocytes than those in the Ehrlich cells. Further evidence suggesting that the main pathways for cationic amino acids in Ehrlich cells and rabbit reticulocytes do not represent the same activity comes from a study of the structural selectivity in the interaction of neutral amino acids and alkali metal ions with the cationic transporters in the two cell types (213). These observations are discussed in section IVB.


View larger version (29K):
[in this window]
[in a new window]
 

View larger version (24K):
[in this window]
[in a new window]
  		FIG. 2.   Cationic amino acid influx into cultured human fibroblasts is inhibited by unlabeled cationic and neutral amino acids with different affinities. A: L-[14C]lysine (2.4 µM) uptake was measured in presence of L-arginine (black-square), L-lysine (open circle ), L-ornithine (black-triangle), and L-diaminobutyrate (DAB; black-diamond ). B: L-[14C]arginine (100 µM) uptake was measured in presence of L-serine (black-diamond ), L-glutamine (*), L-homoserine (black-triangle), and L-arginine (black-square). External medium contained modified Earle's balanced salt solution (pH 7.4, 37°C). [Modified from White et al. (239).]

We believe, therefore, that the current idea that cationic amino acid transport in Ehrlich cells, reticulocytes, and fibroblasts occurs through a single transport system (239) is not consistent with the experimental data. We propose that the designation "system y+" be reserved for the transport of cationic amino acids via a Na+-independent route showing marked specificity for positively charged substrates. This system would account for most of the flux in the case of reticulocytes and fibroblasts, but only a smaller fraction of the flux in the case of Ehrlich cells.

As discussed in sections IIB and IIC, independent evidence, obtained more recently, supports the notion that distinct cationic amino acid transporters differing markedly in their interactions with neutral amino acids coexist in various cell types (64, 222, 226).

B. Systems B0,+ and b0,+: the Na+-Dependent and Na+-Independent Broad-Scope Transporters

In the course of transport studies in mouse blastocysts, Van Winkle and co-workers (222, 226) identified two novel transport systems carrying both neutral and cationic amino acids, but differing in their requirements for Na+: system B0,+ is Na+ dependent, and system b0,+ is Na+ independent.

The existence of system B0,+ became apparent when measuring the effect of various amino acids (including arginine and lysine) on alanine fluxes. Only ~1% of alanine uptake (at 50 µM) was Na+ independent (as compared with fluxes measured in Li+ and choline) (226). Lysine was found to be an effective competitive inhibitor of L-alanine uptake, and the kinetic parameters for these two amino acids, measured either directly or in transport-inhibition experiments, were consistent with the conclusion that they shared the same pathway (Table 2). Other cationic amino acids were also found to be good inhibitors of alanine influx, and the transporter was shown to interact with bicyclic amino acid analogs, such as 3-amino-endo-bicyclo[3.2.1]octane-3-carboxylic acid (BCO) and BCH. The difference of these interactions from those exhibited by any of the previously described systems (A, ASC, L, or y+) justified the designation of system B0,+ as a newly identified transporter.

 
View this table:
[in this window] [in a new window]
  		TABLE 2.   Kinetic parameters for the transport of L-lysine, L-alanine, and BCO through system B0,+ in mouse blastocysts

During these studies, it was observed that a substantial portion of mediated lysine influx persisted in the absence of Na+, suggesting that a Na+-independent system may be operating in parallel to the Na+-dependent system B0,+. This possibility was explored by Van Winkle et al. (222) in a later paper. It was found that the lysine flux measured in a Li+-containing medium was substantially inhibited by 10 mM arginine and leucine. In a mirror experiment, leucine influx (in the presence of choline) was found to be inhibited by 10 mM lysine and arginine. Cyclic amino acid analogs (BCH, BCO) did not affect the rate under the same conditions. The analysis of the effect of varying concentrations of unlabeled amino acids on the flux of radioactive lysine or leucine in (Li+ medium) (Fig. 3) led to the identification of a Na+-independent and broad-scope transporter that showed high affinity for leucine and lysine. The transporter was designated system b0,+ and accounted for most of the leucine (88%) and lysine (98%) flux at low concentration (1 µM). A "cation-preferring" system was also identified that carried only a small fraction (2%) of the lysine flux (226).


View larger version (23K):
[in this window]
[in a new window]
 

View larger version (25K):
[in this window]
[in a new window]
  		FIG. 3.   Mutual inhibition between cationic and neutral amino acids in rat blastocysts. Uptake of L-[3H]lysine (0.87 µM) (A) or L-[3H]leucine (0.8 µM) (B) in presence of various unlabeled amino acids. External medium was phosphate-buffered LiCl (57 mM Li2HPO4 , 18 mM KH2PO4 , 46 mM LiCl, and 1 mg/ml bovine serum albumin (pH 7.1, 37°C). [Modified from Van Winkle et al. (222).]

C. System y+L: the Cation-Modulated Broad-Scope Transporter

More recently, a kinetic study of the partial inhibition of lysine influx by neutral amino acids led Devés et al. (64) to identify a novel transport system for cationic amino acids in human erythrocytes. The transporter was designated system y+L. Neutral amino acids were found to inhibit lysine (1 µM) entry, with the effect reaching a maximum at ~50% of the original flux (Fig. 4A). On the basis of a kinetic analysis of the effect of leucine on lysine influx (at varying substrate concentrations) (Fig. 4B) (see sect. IVA), it was shown that lysine enters the erythrocytes through two transporters: 1) a high-affinity low-capacity transporter, which recognizes leucine and lysine with comparable affinities (system y+L), and 2) a lower affinity high-capacity transporter, which is specific for cationic amino acids (system y+) (64). This proposal subsequently received support from inactivation experiments with the sulfhydryl reagent N-ethylmaleimide (NEM), which selectively inhibited system y+, thus permitting the study of system y+L in isolation (63).


View larger version (15K):
[in this window]
[in a new window]
  		FIG. 4.   L-Lysine entry is partially inhibited by L-leucine added to cis-side. L- [14C]lysine influx was measured in presence of varying concentrations of unlabeled L-leucine. A: L-lysine concentration was 1 µM. Curve was calculated on basis of Equation 3 simplified for [s]/Km << 1 (see sect. IVA). Analysis assumes two parallel systems (y+ and y+L); calculated inhibition constants for two systems are as follows (in mM): K i y+ L , 0.022 ± 0.003; K i y+, 30.36 ± 7.9. Dashed line was drawn for a single system with an inhibition constant for leucine of 80 µM. B: lysine concentration was 1, 10, or 100 µM. Curves for 10 and 100 µM were calculated by inserting into Equation 2 (see sect. IVA) lysine transport parameters estimated independently: K my+, 112 µM; K my+ L , 13.9 µM; Vmax y+ L / Vmax y+, 0.12 (see sect. IVA). External medium contained 150 mM NaCl, 4 mM KCl, and 5 mM sodium phosphate buffer (pH 6.8, 37°C). [Redrawn from Devés et al. (64).]

The substrate specificity of system y+L, in the presence of Na+, resembles that of system b0,+ described in section IIB, but the two systems differ in their interactions with monovalent inorganic cations. Whereas system b0,+ is Na+ independent, the specificity of system y+L varies depending on the ionic composition of the medium. In Na+ or Li+ medium, system y+L interacts strongly with both neutral and cationic amino acids, but if these ions are replaced by K+, the affinity for neutral amino acids is dramatically decreased (9, 63). This behavior is illustrated in Figure 5. It is very important for the correct identification of system y+L to keep in mind that, as shown in Figure 5, inset, at sufficiently high concentrations of neutral amino acids, inhibition will be observed in the absence of Na+. Thus the concentrations of amino acids used when searching for this activity must be appropriate.


View larger version (18K):
[in this window]
[in a new window]
  		FIG. 5.   Interaction of system y+L with leucine, but not lysine, is affected by inorganic monovalent cations in human erythrocytes. L-[14C]lysine (1 µM) influx in presence of varying concentrations of unlabeled L-lysine (A) or L-leucine (B). External medium (pH 6.8, 37°C) contained either Na+ (154 mM NaCl, 5mM sodium phosphate), K+ (154 mM KCl, 5 mM potassium phosphate), or Li+ (154 mM LiCl, 5 mM potassium phosphate). Calculated inhibition constants are as follows (in µM): Ki Lys , 9.5 ± 0.67 (Na+), 8.65 ± 0.33 (K+); Ki Leu, 10.7 ± 0.7 (Na+), 983 ± 80.7 (K+), 4.52 ± 0.26 (Li+). [Redrawn from Devés et al. (63).]

    III. IDENTIFICATION OF COMPLEMENTARY DEOXYRIBONUCLEIC ACID SEQUENCES ENCODING PROTEINS INVOLVED IN CATIONIC AMINO ACID TRANSPORT
Top
Previous
Next
References

The experimental possibilities made available by recombinant DNA technology have led in recent years to the molecular identification of various proteins involved in the transport of cationic amino acids. Products of two types of genes have been identified: 1) the cationic amino acid transporters (CAT), encompassing four homologous proteins [CAT-1, CAT-2A, CAT-2(B), and CAT-3] that accept arginine, lysine, and ornithine as substrates (reviewed in Refs. 52, 103, 123, 125, 127) and 2) the broad-scope amino acid transport proteins (BAT), with two members (rBAT/D2 and 4F2hc), whose expression induces transport of both cationic and neutral amino acids in Xenopus laevis oocytes. The acronym rBAT was originally developed to represent "related to b0,+ amino acid transporter." It has been suggested that the proteins in this group are not transporters but are transport activators. For this reason, they are here referred to as "transport proteins" (reviewed in Refs. 18, 157-159).

In this section we discuss the structural properties of these proteins and briefly mention some of their distinctive functional characteristics. The kinetic properties of the transport activities induced by these proteins are examined with more detail in section IV.

A. Cationic Amino Acid-Specific Transporters: the CAT Family

1. mCAT-1: the constitutive and widely distributed cationic amino transporter

A) IDENTIFICATION. In 1991, Kim et al. (110) and Wang et al. (230) discovered that, when expressed in Xenopus oocytes, the membrane receptor for ecotropic murine leukemia viruses (ecoR) induced cationic amino acid transport. Albritton et al. (6) had previously cloned the cDNA encoding ecoR by transfection of mouse DNA into human cells and selection for susceptibility to ecoR infection. It was noticed that the predicted transmembrane topology of the protein encoded by ecoR was homologous to that of various membrane transport proteins in Saccharomyces cerevisiae including the gene products Can1 (97), Ctr (153), and Hip1 (207), which are arginine, choline, and histidine transporters, respectively. As expected from this finding, injection of mRNA, prepared from in vitro transcription of ecoR cDNA into Xenopus oocytes, was able to induce amino acid transport activity (110, 230). Thus expression of a functional receptor resulted in the selective induction of arginine, lysine, and ornithine uptake. The overall transport properties and the expression pattern of ecoR were compatible with system y+. To reflect its cellular function, the virus receptor was renamed mCAT-1 (mouse cationic amino acid transporter) (53).

Alternative names have been used to describe the genes encoding the mouse and human CAT-1 proteins or the gene product itself (see Table 3). In this review, CAT-1 is used throughout.

 
View this table:
[in this window] [in a new window]
  		TABLE 3.   Properties of the CAT transporters

B) STRUCTURE. The mCAT-1 transporter is a single polypeptide of 622 amino acids and a relative molecular mass of 67 kDa. On the basis of sequence analysis, the protein was initially predicted to contain 14 membrane-spanning domains (6). With the consideration that the putative NH2 terminus lacked a leader sequence, a structural model was proposed that placed the NH2 terminus on the cytoplasmic side and assumed that every hydrophobic domain was membrane spanning. As pointed out by Closs (52), this model is supported by the finding of similarities in the localization of the transmembrane segments 1 through 7 of mCAT-1 and the protein permeases for yeast and fungi (203).

An alternative model assuming 12 potentially membrane-spanning segments has been proposed (107, 179). This model considers the evolutionary relationship of mCAT-1 with members of the APC family of transporters (arginine, polyamines, choline) as well as the conserved nature of structural features to make the prediction. The permeases in this family contain 9-12 hydrophobic domains and have been shown to exhibit ~20% identity to mCAT-1 (179). The model is consistent with the proposal that the mCAT-1 sequence may have arisen from duplication of a primordial gene that encoded a six transmembrane sequence, since transmembrane region 1 exhibits some homology with transmembrane region 7 as do transmembrane regions 6 and 12. Similarities have also been observed between the extracellular loops 3 and 6 (107).

As discussed by Closs (52), evidence supporting the 14 transmembrane model has also come from immunostaining of unpermeabilized cells with antibodies raised against peptides present in the third and fourth putative extracellular loops in the 14 transmembrane model (244). This finding argues against the 12 transmembrane model, for in this case, the corresponding peptide is located intracellularly. Support for the 14 transmembrane model has also been sought in analyses of N-glycosylation (reviewed in Ref. 52). Two putative N-linked glycosylation sites (Asn-223 and Asn-229) conserved in all CAT proteins are present in the third extracellular loop. Another potential glycosylation site (Asn-373) is exposed extracellularly in the 12 transmembrane segment model but intracellularly in the 14 transmembrane segment model. Mutation of Asn-223 and Asn-229 to histidine results in a protein that migrates with the same mobility as the N-glycosidase-treated wild-type mCAT-1. This suggests that Asn-373 is nonglycosylated (although not necessarily intracellular). In both models, the viral binding site is located in the third extracellular loop of mCAT-1 (5, 251).

The complete amino acid sequence and the putative organization of mCAT-1 in the membrane (according to the 12- and 14-transmembrane segment models) are shown in Figure 6.


View larger version (52K):
[in this window]
[in a new window]
 

View larger version (31K):
[in this window]
[in a new window]
 

View larger version (12K):
[in this window]
[in a new window]
  		FIG. 6.   CAT transporters. A: deduced amino acid sequence and alignment of mouse cationic amino acid transporters mCAT-1 and mCAT-2A. NH2-terminal residues are numbered 1 [initiating methionine (M) in both sequences] and COOH-terminal amino acids as residues 622 (mCAT-1) or 657 (mCAT-2A). Homology is shown as follows: a vertical solid line shows identity; two dots show strong chemical similarity of residues; one dot shows similarity; no symbol shows residues that are dissimilar (or are not aligned in optimization procedure). Horizontal lines (designated TM for transmembrane) indicate sequences that (on basis of their hydrophobicity) are predicted to span the membrane, and they are numbered I-XIV. Dotted line for TM VII and TM X indicates particular uncertainty as to whether these 2 sequences do traverse membrane (see Fig. 6B and text). Symbol (psi ) shown above some asparagine residues (N) indicates potential sites for N-linked glycosylation. Shown in bold (residue 107 or 109 of mCAT-1 or mCAT-2A) is conserved glutamic acid residue (E), which was shown to be essential for transport and sequence YGE (residues 235-237 of mCAT-1) known to be required for viral binding (ecoR). B: 2 alternative models for orientation of mCAT-1 in membrane. Model A: 14 transmembrane segment model (numbered as shown on sequence). Model B: 12 transmembrane segment model, in this model sequences corresponding to TM VII and X are not membrane spanning. Third loop contains viral binding site that is represented by a pale dashed line. Glycosylated sites are represented by Y. Dark dashed line indicates domain (41 amino acids) that determines affinity for substrate; this sequence is extracellular in 14 TM model and intracellular in 12 TM model [Redrawn from Closs (52).] C: a comparison of divergent sequence in 2 protein isoforms from alternately spliced mCAT-2 RNA with equivalent mCAT-1 sequence. Asterisks indicate residues that are identical in all 3 CAT proteins. [Redrawn from MacLeod et al. (125).]

C) FUNCTION. Injection of mCAT-1 cRNA into X. laevis oocytes was found to induce cationic amino acid uptake (110, 230). Transport activity was followed by measuring either uptake of radioactively labeled substrates (Fig. 7) or cationic amino acid-specific inward currents (Fig. 8). In both functional expression assays, transport of L-arginine, L-ornithine, and L-lysine was found to be Na+ independent and saturable [Michaelis constant (Km) values 70-100 µM]. Radiolabeled arginine uptake (100 µM) was inhibited by unlabeled L-cationic amino acids, but not by L-neutral amino acids (glutamine, serine, homoserine, and 2-methylaminoisobutyric acid) or D-isomers (0-3 mM) (110). Consistently, neutral amino acids were seen to induce currents only at very high concentrations and provided Na+ was present. The Km for L-cysteine was 24.7 mM and that for L-homoserine was ~10 mM (230).


View larger version (14K):
[in this window]
[in a new window]
 

View larger version (13K):
[in this window]
[in a new window]
  		FIG. 7.   Injection of mCAT-1 mRNA induces saturable L-[3H]arginine uptake into Xenopus oocytes. A: Michaelis-Menten plot, Vmax = 190 pmol·oocyte-1·h-1, Km = 0.07 mM. Uptake solution contained (in mM) 100 NaCl, 2 KCl, 1 MgCl2 , 10 HEPES, and 50 Tris, pH 7.5. B: uptake of various 3H- or 14C-radiolabeled amino acids (0.1 mM) in oocytes injected with mCAT-1 mRNA (solid bars) or water (open bars). Uptake of cationic amino acids was also measured with choline chloride (100 mM) substituting for sodium chloride. [Modified from Kim et al. (110).]


View larger version (20K):
[in this window]
[in a new window]
  		FIG. 8.   A: lysine evokes inward currents in Xenopus oocytes previously injected with mCAT-1 mRNA. L-Lysine concentrations are indicated above each record (in mM); bars indicate period of application. Holding potential was -60 mV. B: currents induced by L-lysine (1 mM), but not by L-homoserine (10 mM), are Na+ independent. Na+ was substituted with equimolar Tris. C: currents induced by L-arginine, L-lysine, and L-ornithine are saturable. Peak currents evoked by a given concentration of amino acid were normalized to current evoked by 10 mM L-arginine in same oocyte. [Modified from Wang et. al. (230).]

Increase in the binding of gp70 (the murine ecotropic viral coat protein that binds to the receptor) was also observed in injected oocytes, showing that expression of a functional receptor protein coincides with the increase in L-arginine uptake (110, 230).

D) TISSUE DISTRIBUTION. Expression of mCAT-1 has been found in all tissues investigated with the exception of the liver. Two transporter transcripts (7 and 7.9 kb) that result from use of alternative polyadenylation sites have been reported. Greatest expression was found in the testis (which synthesizes large amounts of arginine- and lysine-rich proteins), but transcripts were also present in bone marrow, brain, stomach, spleen, kidney, lung, ovary, uterus, large and small intestine, thymus, heart, and in trace amounts in skeletal muscle and skin (102, 110). Three different mCAT-1 transcripts were found in heart (102), and the smallest (~6 kb) appears to be unique to cardiac muscle. The absence of mCAT-1 in liver is supported by the failure of murine ecotropic retroviruses to infect this tissue (56, 92). The tissue distribution of mCAT-1 is listed in Table 4.

 
View this table:
[in this window] [in a new window]
  		TABLE 4.   Expression of mCAT transcripts in mouse or rat tissues

E) HUMAN CAT-1 AND OTHER SPECIES. A human cDNA, homologous to the murine ecotropic retroviral receptor, was cloned by Yoshimoto et al. (249) from a T-cell line. The human cDNA (originally called H13) presents 87.6% amino acid identity to its murine counterpart, and its sequence predicts a protein with 629 amino acids (~68 kDa). The gene is ubiquitously expressed in human tissues except in liver. The genomic DNA was also isolated from a cosmid library derived from human lymphocytes, and its organization was elucidated; it was mapped to chromosome 13 (249). This gene was cloned independently by Albritton et al. (4), designated ATRC1 and mapped to 13q(12-14).

More recently, a rat homolog was also cloned (170, 245), and it exhibits 95% homology to the mouse gene.

2. mCAT-2: the tissue-specific and regulated cationic amino acid transporters

Two transporters that are structurally homologous to mCAT-1 but differ in either kinetic properties, tissue distribution, or regulated expression have been identified. These proteins have been designated mCAT-2 or mCAT-2A (first identified in hepatocytes) (53, 108) and mCAT-2 or mCAT-2(B) (first identified in macrophages and lymphocytes) (54, 102, 108, 124). Here we use the acronym mCAT-2A to refer to the hepatocyte isoform and mCAT-2(B) to the macrophage and lymphocyte isoforms, as recently proposed by Closs (52).

A) IDENTIFICATION AND STRUCTURE. In 1990, MacLeod et al. (126), using a subtraction-differential screening assay, were able to identify several novel cDNA clones from two closely related murine lymphoma cell lines with distinct degrees of phenotypic differentiation. One of these genes was named Tea (T-cell early activation receptor) because it was induced early in the response of normal T cells to mitogens; this gene was mapped to chromosome 8 (124). The gene was not expressed in quiescent splenic T cells but was induced 16-fold when the T cells were activated with the mitogen concanavalin A. The cDNA revealed significant homology with only one protein, the mouse ecotropic retroviral receptor (ecoR) (whose transport function was unknown at the time), but it was substantially shorter. In a subsequent report, the original cDNA sequence for Tea, which had been truncated at the NH2 terminus, was corrected (179) and shown to encode a protein 658 amino acids long (i.e., containing 205 amino acids more than the original version). Independently, this protein, which showed 61% amino acid sequence identity to mCAT-1, was found to be expressed in lipopolysaccharide-stimulated murine macrophages (55). We refer to this other protein as mCAT-2(B).

Using a Tea probe to screen a mouse liver cDNA library (known not to contain mCAT-1) under stringent conditions, Closs et al. (53) were able to identify a clone encoding for a protein showing 97% identity with Tea, which was designated mCAT-2A. The deduced sequences of the two proteins differ only by 20 amino acids within a stretch of 41 amino acids (Fig. 6). Because a single genomic fragment contains both exons, it has been proposed that the two isoforms [mCAT-2A and mCAT-2(B)] result from mutually exclusive alternate splicing of the primary transcript. They have been proposed to contain the same number of transmembrane segments (12 or 14 depending on the analysis) (53, 179).

B) FUNCTION. Injection of mRNA derived from these clones into Xenopus oocytes induced, as expected, Na+-independent and stereoselective cationic amino acid transport (53, 55, 102). However, in spite of the very high structural homology, the mCAT-2A and mCAT-2(B) transporters have been found to differ not only in their tissue distribution, but also in their affinities for the substrate and possibly in other kinetic properties, such as their sensitivity toward trans-stimulation. Whereas in the case of mCAT-2(B) the Km values for arginine, lysine, and ornithine were in the same range as those measured for mCAT-1 (55, 102), the transporter isolated from liver cDNA library, mCAT-2A, exhibited a marked lower affinity (Km = 2-5 mM) for its cationic substrates and appears to be less responsive to trans-stimulation (see sect. IVB1 ) (53). This finding coincides with the observations, in hepatocytes, in which saturable arginine transport could not be demonstrated, suggesting that arginine crosses the membrane through a low-affinity route (238).

As discussed in a following section, chimeric proteins have been constructed, and it has been shown that the proteins in this family contain domains that can function independently to control specific aspects of transport and retroviral receptor function (55, 108). Table 3 summarizes the properties of the different transporters in the CAT family.

C) TISSUE DISTRIBUTION. Whereas mCAT-1 is expressed nearly ubiquitously (except in the liver), mCAT-2A and mCAT-2(B) expression is restricted to a more limited number of tissues and cell types. The low-affinity isoform mCAT-2A is found in liver, skeletal muscle (most abundantly posttrauma), skin, ovary, and stomach, and it is absent from brain, large and small intestine, and kidney and quiescent or activated splenocytes. The high-affinity isoform mCAT-2(B) is expressed in lung, brain, activated macrophages and splenocytes, ovary, skeletal muscle, and testis, and it is absent from small intestine, kidney, liver, and resting macrophages and splenocytes (127). The patterns of expression of the mCAT proteins are shown in Table 4.

Recent studies have shown that the expression of these different tissue-specific isoforms may be regulated physiologically. Thus, during liver regeneration after partial hepatectomy, it has been shown (12) that the CAT-1 isoform (rather than CAT-2A) is expressed and can be induced additionally in response to hormones (insulin, glucocorticoids). The pattern of the induction of this gene is that of a classical delayed early response gene. Thus induction is cycloheximide sensitive (i.e., itself requires protein synthesis). There is also posttranscriptional regulation of CAT-1 gene expression and rapid turnover of the induced mRNA, suggesting tight regulation of expression of the protein by a number of separate mechanisms, some of which are tissue specific.

D) HUMAN CDNA. Human sequences for the CAT-2 transporters have also been reported by Closs et al. (54). Analyses of the deduced amino acid sequences of hCAT-2A and hCAT-2(B) demonstrated 90.9% identity with the respective murine proteins. In their functional domains (42 amino acids), the human CAT-2 transporters differ only by one residue from the respective mouse proteins. As expected, the kinetic properties observed for the three isoforms of the transporter encoded in the murine CAT genes were also seen in the human. Hoshide et al. (99) showed that hCAT-2 mapped to chromosomal location 8p21.3. In keeping with data discussed in section V, there were not any mutations in this gene associated with the inherited metabolic disorder lysinuric protein intolerance (LPI).

3. rCAT-3: the brain-specific cationic amino acid transporter

Recently, homology screening of a rat brain cDNA library, with a mCAT-1 probe, led Hosokawa et al. (100) to the identification of a cDNA (rCAT-3) encoding a novel member of the murine CAT family whose expression is restricted to the brain. The gene product has 619 amino acids and a calculated molecular mass of ~67 kDa. The predicted amino acid sequence of rCAT-3 shows 53-58% identity with those of the other members of the CAT family. Transient expression of rCAT3 into COS-7 cells induced system y+ activity. Arginine uptake (100 nM) was Na+ independent, not affected by neutral amino acids (1 mM), and inhibited by K+-induced depolarization. The Km for arginine was 103 µM.

B. Broad-Scope Amino Acid Transport Proteins: the BAT Family

1. rBAT and D2: the Na+-independent broad-scope amino acid transport proteins

A) EXPRESSION CLONING. In 1992, a novel class of amino acid transport-related proteins was identified in three separate laboratories by expression cloning from rabbit and rat kidney cDNA libraries. These proteins were designated "neutral amino acid transporter" (209) or D2 (rat kidney) (234) and rBAT (rabbit kidney) (20).

In the first of these studies, Tate et al. (209) failed to recognize the ability of the protein to induce cationic amino acid transport, and it was thought to represent a neutral amino acid transporter. It was noticed, however, that its specificity toward neutral amino acids was not identical to that of system L. Subsequently, Bertran et al. (20) and Wells and Hediger (234) reported the independent cloning of one identical and one homologous sequence and were able to identify this protein as responsible for the induction of cationic and neutral amino acid transport with a specificity analogous to that reported for system b0,+ in rat blastocysts (Fig. 9).


View larger version (23K):
[in this window]
[in a new window]
  		FIG. 9.   Uptake of L-[3H]arginine (A) and L-[3H]leucine (B) induced by rabbit broad-scope amino acid transporter protein (rBAT), in Xenopus oocytes, is inhibited by various neutral and cationic amino acids. Concentration of unlabeled amino acids was 5 mM, except in case of cystine (200 µM). Concentration of labeled amino acid was 50 µM. External medium contained (in mM) 100 choline chloride, 2 KCl, 1 CaCl2 , 1 MgCl2 , and 10 HEPES/Tris, pH 7.5. [Redrawn from Bertran et al. (20).]

Interestingly, in addition to inducing the uptake of various cationic and neutral amino acids, injection of D2 or rBAT cRNA into Xenopus oocytes was found to induce cystine uptake 200- to 400-fold over control values. Cationic amino acids, in excess, inhibited the uptake of neutral amino acids, and likewise, neutral amino acids inhibited the uptake of cationic amino acids (Fig. 9). The transport of L-leucine, L-alanine, and L-lysine (at 15 µM) through D2 was unaffected by Na+ replacement with N-methylglucosamine (234).


View larger version (41K):
[in this window]
[in a new window]
 

View larger version (44K):
[in this window]
[in a new window]
 

View larger version (37K):
[in this window]
[in a new window]
  		FIG. 10.   BAT transport proteins. Comparison of amino acid sequence of rat D2 and 4F2hc. NH2-terminal methionine residues are numbered 1 and COOH-terminal amino acids as residues 683 (D2) and 527 (4F2hc). Similarity between 2 sequences is indicated by symbols between optimally aligned sequences (see Fig. 6). Two alternative membrane topologies have been proposed for D2, assuming either 1 TM domain or 4 TM domains. Four putative transmembrane sequences are indicated (TM I to TM IV) by horizontal lines; these are dotted for TMs II, III, and IV, indicating relative uncertainty as to whether these sequences do in fact cross membrane. Potential N-glycosylation sites are indicated (psi ). B: diagram represents proteins D2 and 4F2hc as type II membrane glycoproteins; a putative light subunit is also shown to right. cDNA encoding this protein has not been cloned, but there is evidence for association of heavy chain with a lighter subunit through disulfide bridges (see text). [Redrawn from Palacín (157).] C: 4 transmembrane segment model for D2. Approximate end points of membrane-spanning domains are numbered, and potential glycosylation sites are marked (Y). Heavy bars represent epitopes against which antipeptide antibodies were made in a study of membrane topology of these proteins [Redrawn from Mosckovitz et al. (141).]

B) STRUCTURE. D2 and rBAT are single polypeptides of 683 and 677 amino acids, respectively (79% identity). The transporter exists in glycosylated form (84 and 87 kDa), and the unglycosylated form has a molecular mass of 78 kDa. Surprisingly, in contrast to most transport proteins, hydrophobicity analysis does not predict a large number of membrane-spanning segments (Fig. 10). Two alternative membrane arrangements have been proposed. One has a single transmembrane domain (20, 234) and the other four such segments (209).

The first proposal postulates that the rBAT/D2 cDNA encodes a type II membrane glycoprotein, containing a cytoplasmic NH2 terminus, a single transmembrane segment [residues 87-107 (D2); 81-102 (rBAT)], and a N-glycosylated extracellular COOH terminus; there are seven putative N-glycosylation sites in rBAT and D2. Homology has been found between the COOH-terminal domain and a family of carbohydrate-metabolizing enzymes and related proteins that lack the membrane-spanning domain (20, 234). This family includes a maltase-like gene of the mosquito Aedes aegypti (34% identity, 55% similarity) (101) and an oligo-1,6-glucosidase of Bacillus cereus (30% identity, 49% similarity) (233). In the case of the rat isoform (D2), an aspartic acid, which is considered to be part of the catalytic site of amylases and alpha -glucosidases, is conserved. However, catalytic activity could not be demonstrated (234).

In addition, rBAT/D2 shows amino acid sequence homology (~30% amino acid identity) with a type II membrane glycoprotein, the human and murine cell surface antigen 4F2 heavy chain (4F2hc). The two proteins share four amino acid sequence fragments (10-18 amino acid residues) that are highly conserved (67-80% identity) (30, 115, 162, 171, 211). The 4F2 surface antigen is composed of two chains (heavy and light) that are linked by a disulfide bond (see sect. IVB2 ).

The alternative proposal (141, 209) postulates that the D2 protein has four sequences (each 23-28 amino acids), which are potential transmembrane domains. The second and fourth putative transmembrane domains (residues 390-413 and 588-610, respectively) include charged amino acid residues.

The two alternative models were tested by Mosckovitz et al. (141) using two different methods. In the first approach, the surface of COS-7 cells, expressing D2, was probed with antibodies directed against putative extracellular and cytoplasmic domains (see scheme in Fig. 10C ). Antibodies were added in the presence or absence of Triton X-100 to permeabilize the cells and were then visualized with fluorescent labeling. In the second approach, right-side-out rat renal brush-border membrane vesicles were subjected to limited surface proteolysis by papain, and the fragments were subsequently probed with several site-specific antibodies to determine approximately where the cleavages occurred. Cleavage of D2 with papain produces several fragments that remain firmly embedded in the membrane, and thus both approximations yielded results consistent with the four membrane-spanning topological model (141, reviewed in Ref. 208). The unusual membrane topology that has been predicted for the broad-scope amino acid transport proteins has led to the proposal that rBAT and D2 are not the complete transporters. Three possibilities have been suggested: 1) the carrier could be a homoligomer, 2) the cloned polypeptide could be part of a heteroligomeric carrier that can associate with silent transporters in the oocyte, or 3) it could be a specific activator of an endogenous carrier. Wells and Hediger (234) have suggested that these proteins may undergo dimerization mediated by leucine zippers. In addition, a cysteine residue that may play a role in the heterodimeric structure of the surface antigen 4F2 is conserved in the case of rBAT and D2. This residue is located just COOH terminal to the transmembrane region.

Recently, Wang and Tate (231) obtained evidence suggesting that in the rat kidney and jejunal brush-border membrane, D2 is found in association with a 50-kDa protein, and the association has been shown to involve one or more interprotein disulfide bonds. When brush-border membranes were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis under nonreducing conditions, D2 was seen to migrate partly (15-50%) as a 135-kDa species; this fraction increased to 90% when NEM was included in the buffer. Because the D2 cDNA clone encodes a 85-kDa subunit, it follows that the formation of 135-kDa species must involve association of D2 proteins with ~50-kDa proteins of oocyte origin. Analogous results were obtained in rabbit kidney brush-border membranes.

As discussed by Wang and Tate (231), the proposed oligomeric organization may explain why the expression of D2 in COS-7 cells failed to induce amino acid transport activity (142). It is possible that the 50-kDa or similar proteins required for functionally competent heterodimers are absent in COS-7 cells. This is consistent with the Western analysis of membranes from D2 cDNA-transfected COS-7 cells, which under nonreducing conditions did not show D2 species in the 135- to 140-kDa range (although intense bands corresponding to the D2 monomers were visible). On the contrary, a small but significant amount of disulfide-linked oligomers (135-140 kDa) has been detected in oocytes injected with D2/cRNa, in addition to D2 monomers and higher aggregates (180-200 kDa), the latter presumably representing disulfide-linked homodimers of D2 subunits (142, 231, reviewed in Ref. 208). As pointed out by Tate (208), if the formation of such a heterodimer is a prerequisite for expression of transport activity, the upper limit of transport must then depend on the availability of 50-kDa protein in oocyte. This would explain why, in D2-injected oocytes, 10-fold increase in total D2 (24-48 h after injection) produces only ~1.5-fold increase in the rate of leucine uptake (142).

Further insight into the role of rBAT has come from an analysis of two mutants of the human gene (M467T and M467K) which, as discussed below, are responsible for the genetic defect cystinuria type I (32). Chillarón et al. (41) showed that both mutants display an intracellular trafficking defect that impairs their transport to the oocyte surface and maintains them in the intracellular location, probably the endoplasmic reticulum. However, a clear lack of correlation between the amount of rBAT protein in the oocyte surface and the induced amino acid transport was observed. Despite the low amount of M467T protein reaching the plasma membrane, the transport activity eventually (at 7 days) became the same as in the wild-type injected oocytes. It was concluded that rBAT is necessary, but not sufficient, for amino acid transport activity.

Recently, to evaluate whether rBAT functions as a component of an amino acid transporter or as a transport activator, COOH-deletion mutants of the human gene were prepared and assayed for their ability to stimulate transport. Several deletions (Delta 141-685, Delta 261-685, and Delta 511-685) were made by Miyamoto et al. (137). Mutant cRNA was synthesized and injected into Xenopus oocytes, and the uptake of cystine, arginine, and leucine was determined. The wild-type cRNA resulted in the induction of system b0,+-like amino acid transport. Of the mutants, only Delta 511-685 stimulated transport, but the activity in this case matched that of system y+L (induced by 4F2hc); that is, mutant Delta 511-685 stimulated the transport of arginine and leucine (not of cystine), and only the transport of leucine was Na+ dependent. Unfortunately, the authors refer to this activity as y+-like. It is interesting to notice that the Delta 511-685 deletion in rBAT renders a protein of a similar size to the single transmembrane-protein 4F2hc. This led to the proposal that both rBAT and 4F2hc are likely to have a single transmembrane segment and further that the COOH terminal of rBAT is important in determining its specificity as a transport activator.

In another recent study, Peter et al. (165) suggested that the D2-stimulated activities of both neutral and cationic amino acids appear to consist of, at least, two distinct pathways and that these activities bear functional similarities to transporters in native oocytes that might be activated by formation of complexes with the D2 protein.

C) TISSUE DISTRIBUTION. Clones of D2 and rBAT hybridized intensely to a species of 2.2 kb and weakly to another of 4.4 kb in mRNA from kidney and intestine (rat and rabbit, respectively) (20, 209, 234). A 2.2-kb species was also seen in mRNA from a kidney-derived cell line LLC-PK1 (234). Because the relative intensities of the bands do not change with stringency, they appear to reflect the use of different polyadenylation sites. At low stringency and after long exposure, bands were seen in pancreas, liver (2.2 kb), heart, brain (2.4 kb), and lung. No hybridization was seen to mRNA from skeletal muscle (20, 209, 234).

The site of expression of D2 in the kidney was studied by Kanai et al. (104) by a combination of in situ hybridization and immunocytochemistry with antibodies that differentially recognize specific segments of proximal tubule. The D2 antisense RNA hybridized in the same segments that were strongly positive for anti-ectoATPase, but negative both for carbonic anhydrase type IV and the glucose transporter GLUT-2. These findings show that, in the renal proximal tubule, D2 is strongly expressed more distally in the S3 segment, with only weak expression in segments S1 and S2. The signal is absent in all other parts. The S3 expression thus coincides with transport activity for cystine and other amino acids as characterized by microperfusion studies. Renal proximal tubular brush-border membrane localization of D2 was independently confirmed in two other laboratories (81, 167). A less intense labeling was seen in the brush-border membranes of jejunal epithelial cells (167).

As reported by Pickel et al. (167), an unexpected finding of these immunocytochemical localization studies was that, in the intestine, intense D2-specific immunolabeling was seen in a selected population of enteroendocrine cells and enteric neurons. The neuronal labeling was localized within dense-core vesicles in axon terminals apposed to the basal lamina near fenestrated blood vessels. In an extension of this study, Nirenberg et al. (154) reported highly granular immunolabeling for D2 in the chromaffin and ganglion cells of rat adrenal medulla. In addition, labeled varicose processes were detected in brain stem and spinal cord nuclei. Ultrastructural examination of the nuclei of the solitary tract of rats showed that D2 was localized predominantly in axon terminals. Further work is required to characterize the anti-D2 reactive protein in these cells.

D) HUMAN RBAT AND OTHER SPECIES. A human rBAT cDNA was isolated by screening a human kidney cortex cDNA library for expression of Na+-independent transport of L-arginine in Xenopus oocytes (19). The clone also induces L-cystine and L-leucine uptake, and it is found in kidney, small intestine, pancreas, and liver. The isolated clone (685 amino acids) encodes a 78-kDa protein 85 and 80% identical to rBAT and D2, respectively. In another report, the human rBAT clone was isolated by low-stringency screening of a human kidney cDNA library using the radiolabeled D2 insert as a probe (118). Southern blot analysis of genomic DNA from a panel of mouse-human somatic cell hybrids showed that the gene for human rBAT resides on chromosome 2 (118). Since the discovery that mutations in human rBAT are responsible for one form of cystinuria, there has been an intense study of this human gene (reviewed in Ref. 159); this work is reviewed in section V.

In a recent report, Mora et al. (140) have shown that OK cells (a "proximal tubular-like" cell line, derived from opossum kidney) express a rBAT transcript. Poly(A)+ RNA from OK cells induced system b0,+-like transport activity in oocytes, and the induction could be suppressed by hybrid depletion with human rBAT antisense oligonucleotides. A polymerase chain reaction-amplified cDNA fragment (~700 bp) from OK cell RNA was shown to correspond to an rBAT protein fragment 65-69% identical to those from human, rabbit, and rat kidneys. To show that the rBAT protein is functionally related to this transport activity, OK cells were transfected with human rBAT antisense and sense sequences. Transfection with rBAT antisense, but not with rBAT sense, resulted in the specific reduction of rBAT mRNA expression and b0,+-like transport activity. The activity was further localized to the apical pole of confluent OK cells (140). The properties of the rBAT and D2 proteins are summarized in Table 5.

 
View this table:
[in this window] [in a new window]
  		TABLE 5.   Properties of the BAT transport proteins

2. 4F2hc: the cation-modulated and broad-scope amino acid transport protein

The homology (~30% identity) detected between the broad-scope amino acid transport inducing proteins, rBAT and D2, and the heavy chain of the cell surface antigen 4F2hc (Fig. 10) prompted the investigation of the possible involvement of this protein in transport (17, 235).

The 4F2 cell surface antigen is induced during the process of cellular activation and remains at constant level in exponentially growing cells (162). It was originally identified by the production of a mouse monoclonal antibody (mAb4F2) against human T-cell lines (93, 95) and mapped to human chromosome 11 (166). The cDNA clones for the human, mouse, and rat heavy chain antigens (4F2hc) have been isolated and are 75% identical at the amino acid level (30, 115, 162, 171, 211). The 4F2hc gene is a member of the family of growth-related genes characterized by a low level of constitutive expression in resting cells and a high level of expression after cell activation or neoplastic transformation (246, 247). It has been recently proposed that 4F2hc is identical to the fusion protein-1 (FRP-1) that regulates virus-mediated cell fusion of monocytes. The FRP-1 is principally detected in concanavalin A- or interleukin-2-stimulated lymphocytes, although it is scanty in resting lymphocytes (155). The pattern of expression of 4F2 in malignant cells (squamous cell carcinomas) has been correlated with the tumor-spreading pattern, differentiation, and metastatic behavior (183).

A) STRUCTURE. The 4F2 surface antigen is constituted by a 85-kDa glycosylated heavy chain covalently linked to a 41-kDa highly hydrophobic light chain. The heavy chain is a type II membrane glycoprotein with cytoplasmic NH2 terminus, a single transmembrane segment, and glycosylated extracellular COOH terminus (115, 162, 171, 211). Northern analysis has demonstrated that its expression is widely distributed in mouse tissues (162).

The rBAT/D2 and 4F2hc show a very similar localization of the single putative transmembrane domain within their sequences (Fig. 10). Neither rBAT nor 4F2hc has the conserved aspartic (or glutamic) acid residues, which are part of the catalytic site for the homologous amylases and alpha -glucosidases and which are conserved in D2.

B) FUNCTION. As expected, Bertran et al. (17) and Wells et al. (235) found that the injection of 4F2hc mRNA (human) induced a broad-scope amino acid transport activity. However, the expression of 4F2hc produces a smaller increase in basal amino acid uptake (3- to 4-fold) than the expression of rBAT. The uptake of arginine, lysine, ornithine, leucine, but not cystine was stimulated [Km for L-arginine was 43 µM (235)]. Interestingly, the Na+ dependence of cationic and neutral amino acid uptake was found to differ. Whereas lysine and arginine were transported at the same rate in the presence or absence of Na+, the interaction with neutral amino acids was markedly Na+ dependent. Thus uptake of L-arginine (15 µM), in 4F2hc-injected oocytes, was strongly inhibited by 2 mM L-leucine, L-methionine, or L-homoserine in the presence of Na+, but not in its absence (N-methyl-D-glucosamine) (Fig. 11). The substrate specificity induced by 4F2hc corresponds to that of system y+L (9, 64, 72, 73).


View larger version (17K):
[in this window]
[in a new window]
  		FIG. 11.   Uptake of neutral amino acids, but not cationic amino acids, in 4F2hc-injected Xenopus laevis oocytes is Na+ dependent. A: uptake of various 14C-labeled amino acids (15 µM) was measured in presence of 100 mM Na+ or 100 mM N-methyl-D-glucosamine (open bars). B: effect of 2 mM homoserine (HS), leucine (Leu), and methionine (Met) on L-[14C]arginine (15 µM) influx. Experiments were performed in presence of either 100 mM Na+ (solid bars) or 100 mM N-methyl-D-glucosamine (open bars). [Modified from Wells et al. (235).]

C) TISSUE DISTRIBUTION. Northern blot analyses have demonstrated that 4F2hc gene is expressed at relatively high levels in adult testis, lung, brain, kidney,