Pflügers Archiv

, Volume 447, Issue 5, pp 490–494

The ancillary proteins of HATs: SLC3 family of amino acid transporters


    • Department of Biochemistry and Molecular Biology, Faculty of BiologyUniversity of Barcelona and Parc Cientìfic de Barcelona
  • Yoshikatsu Kanai
    • Department of Pharmacology and ToxicologyKyorin University School of Medicine
The ABC of Solute Carriers Guest Editor: Matthias A. Hediger

DOI: 10.1007/s00424-003-1062-7

Cite this article as:
Palacín, M. & Kanai, Y. Pflugers Arch - Eur J Physiol (2004) 447: 490. doi:10.1007/s00424-003-1062-7


The heteromeric amino acid transporters (HATs) are composed of a light and a heavy subunit linked by a disulfide bridge. The heavy subunits are the SLC3 members (rBAT and 4F2hc), whereas the light subunits are members of the SLC7 family of amino acid transporters. SLC3 proteins are type II membrane glycoproteins (i.e., one single transmembrane domain and the C-terminus located outside the cell) with a bulky extracellular domain that shows homology with α-glucosidases. rBAT heterodimerizes with b0,+AT (SLC7A9) constituting the amino acid transport b0,+, the main system responsible for the apical reabsorption of cystine in kidney. The defect in this system causes cystinuria, the most common primary inherited aminoaciduria. 4F2hc subserves various amino acid transport systems by dimerization with different SLC7 proteins. The main role of SLC3 proteins is to help routing of the holotransporter to the plasma membrane. A working model for the biogenesis of HATs based on recent data on the rBAT/b0,+AT heterodimeric complex is presented. 4F2hc is a multifunctional protein, and in addition to its role in amino acid transport, it may be involved in other cellular functions. Studies on two SLC7 members (Asc-2 and AGT1) demonstrate heterodimerization with unknown heavy subunits.


The SLC3 family is composed of two proteins: rBAT (related to b0,+ amino acid transporter) (SLC3A1) and 4F2hc (the heavy chain of the cell surface antigen 4F2) (SLC3A2) (Table 1). rBAT was identified by expression cloning in Xenopus oocytes [4, 34, 41], where it induces amino acid transport with characteristics of system b0,+, previously described in mouse blastocysts [37]. 4F2hc was originally identified as a lymphocyte activation antigen by means of a monoclonal antibody 4F2 [21]. It was shown that the glycoprotein 4F2hc (~85 kDa) links to a non-glycosylated light chain (~40 kDa) via a disulfide bond to form a heterodimeric 4F2 antigen (CD98) [21]. The homology of 4F2hc [31, 35] with rBAT prompted its expression in oocytes, resulting in the induction of amino acid transport system y+L [3, 42], previously described in erythrocytes [15].
Table 1.

SLC3: the heavy subunits of the heteromeric amino acid transporters

Human gene



Predominant substrates

Transport type /coupling ions

Tissue distribution and cellular/subcellular expression

Link to disease

Human gene locus

Sequence accession ID

Splice variants and their specific features




Heterodimerizes with light subunit bo,+AT (SLC7A9): amino acid transport system bo,+

Exchanger (see details in SLC7 table)

Epithelial cells in kidney and small intestine >liver, pancreas. In epithelial cells, apical plasma membrane

Classic cystinuria type I (genetic defect)






Amino acid transport systems L, y+L, xc- and asc with light subunits SLC7A5–8 and SLC7A10–11

Exchanger (see details in SLC7 table)

Ubiquitous. In epithelial cells, basolateral plasma membrane



The heteromeric amino acid transporters (HATs) are disulfide-bound heterodimers of SLC3 members (heavy subunit) and the corresponding light subunit (some SLC7 members) (see Fig. 2 in review "CATs and HATs: SLC7 family of amino acid transporters"). Excellent reviews on HATs have been published recently [12, 14, 38, 39]. This mini review deals with HATs from the view point of the heavy subunits (SLC3 family), and the review "CATs and HATs: SLC7 family of amino acid transporters" deals with the light subunits.

Common features of the SLC3 family members

rBAT and 4F2hc share functional and structural properties. Light subunits of HATs need interaction with either rBAT or 4F2hc for routing to the cell plasma membrane, where they form disulfide bound heterodimers (~120–130 kDa) (reviewed in [12, 14, 38, 39] (Fig. 1). Elimination of the disulfide bridge reduces, but does not abolish completely, the expression of the heterodimer in the cell plasma membrane, suggesting non-covalent interactions between the light and heavy subunits [16, 28].
Fig. 1.

Working hypothesis of the biogenesis of the heteromeric amino acid transporters. The light subunit b0,+AT (gray) (I) and the heavy subunit rBAT (green) (II) are inserted in the ER membrane independently. b0,+AT has an active conformation in the absence of rBAT. (III) rBAT and b0,+AT heterodimerize in the ER and the rBAT/b0,+AT complex suffers unknown changes (species in red and blue) and is released from the ER. (IV) The N-glycosylation of the rBAT/b0,+AT complex matures in the Golgi, and finally (V) the complex reaches the apical plasma membrane of epithelial cells. The rBAT/b0,+AT heterodimer is linked by a disulfide bond not shown in the scheme. The experimental evidence for this model is described in the text

SLC3 members seem to be present throughout the animal kingdom (i.e., Metazoa). Six sequences are available for mammalian rBAT (human, rat, mouse, rabbit, dog, and American opossum), sharing 69–80% amino acid identity and six sequences for vertebrate 4F2hc (human, rat, mouse, Chinese hamster, zebrafish, and sea lamprey), sharing 31–77% amino acid identity. Moreover, three hypothetical proteins from Caenorhabditis elegans (TrEMBL accession nos. O45298 and Q9XVU3) and Drosophila melanogaster (Q9VHX9) show homology with vertebrate SLC3 members (~25% amino acid identity).

The rBAT protein (685 amino acid residues for human rBAT) is longer than the 4F2hc protein (529 for human 4F2hc), and they share about 27% amino acid sequence identity. Both proteins show N-glycosylation (~94 kDa and ~85 kDa for the mature glycosylated forms of rBAT and 4F2hc, respectively) and are putatively type II membrane glycoproteins, with the N-terminus intracellular, a single transmembrane (TM) domain, and a bulky extracellular C-terminus. Experimental evidence supports this topology for the N- and C-termini of 4F2hc [18]. The cysteine residue participating in the disulfide bridge with the corresponding light subunit is four to five amino acids away from the TM.

The bulky extracellular domain of SLC3 members shows homology with insect maltase (35–40% amino acid identity) and with bacterial α-glucosidases (~30% amino acid identity). These enzymes belong to the large glycosyl hydrolase family 13 [23], which have a catalytic (β/α)8-barrel (TIM-barrel) as the predominant structural domain [40]. Homology modeling suggests a similar structure for the extracellular domain of rBAT and 4F2hc, but the poor homology with the solved glycosyl hydrolase structures (e.g., Bacillus cereus oligo 1,6-glucosidase (O1,6G); PDB 1uok; [40]) precludes a good secondary structure prediction [12]. Alignment of the extracellular domain of rBAT and 4F2hc with O1,6G suggests that the catalytic residues are not present in the SLC3 members [12]. In agreement with this, Wells and Hediger [41] did not observe α-glucosidase activity for rBAT after expression in Xenopus oocytes.


SLC3A1 is expressed mainly in kidney and small intestine [4, 34, 41], as two mRNA species that represent alternative polyadenylation [25]. The onset of expression of rBAT in kidney is postnatal and full expression occurs after weaning [20]. The rBAT protein is expressed in the apical plasma membrane of epithelial cells of the intestinal mucosa and the renal proximal tubule [20, 30]. rBAT mRNA is also detected in liver, pancreas and brain [4, 5, 24, 34, 41].

In 1999, b0,+AT (SLC7A9) was identified as the light subunit that co-expresses with rBAT system b0,+ amino acid transport activity [9, 22, 29]. This system mediates the exchange of cystine and dibasic amino acids with neutral amino acids [11]. rBAT and b0,+AT heterodimerize in renal brush-border membranes [19]. Quantitative co-immunoprecipitation revealed that all b0,+AT present in renal brush-border preparations heterodimerizes with rBAT, but not all rBAT forms heterodimers with b0,+AT [19]. This strongly suggested additional light subunits for rBAT, not yet identified, in renal brush-border membranes. b0,+AT protein has a gradient of expression along the proximal tubule (higher in the convoluted than in the straight part) [29], which is complementary to that of rBAT [20], suggesting that these additional subunits have a similar distribution to rBAT along the proximal tubule.

System b0,+ (i.e., the rBAT/b0,+AT heterodimer) is defective in cystinuria. This is the most common primary inherited aminoaciduria, characterized by hyperexcretion of dibasic amino acids and cystine in urine. The low solubility of cystine causes the formation of cystine calculi (reviewed in [27]). Cystinuria is a recessive disease, and it has been subdivided into phenotypes I (silent heterozygotes) and non-I (moderate to high hyperexcretion of cystine and dibasic amino acids in heterozygotes). Mutations in SLC3A1 cause cystinuria type I [7, 8], whereas mutations in SLC7A9 cause cystinuria non-type I, and some cases of cystinuria type I (see the review "CATs and HATs: SLC7 family of amino acid transporters"). Eighty-four percent of the cystinuria patients studied are explained by mutations in either SLC3A1 or SLC7A9 genes [13], leaving little room for new cystinuria genes. In cystinuria patients the reabsorption of cystine could be almost null, suggesting that the amino acid transport system b0,+ (i.e., the rBAT/b0,+AT heterodimer) is the main, if not the only, apical reabsorption system for cystine [19]. At present, >60 cystinuria-specific mutations (>30 missense mutations) in SLC3A1 have been identified (reviewed in [27]). Most of the missense mutations involve amino acid residues within the glucosidase-like extracellular domain. All cystinuria-specific rBAT mutations functionally analyzed showed trafficking defects, in agreement with the role of rBAT in the routing of rBAT/b0,+AT heterodimer to the plasma membrane (reviewed in [12]).

Recent studies on the rBAT/b0,+AT heterodimeric complex have led to the proposal of a working model for the biogenesis of HATs (Fig. 1): (1) b0,+AT is required for the maturation of the N-glycosylation of rBAT [1, 32], suggesting that heterodimerization occurs in the ER; (2) the stability of rBAT is increased by b0,+AT, whereas the stability of the latter seems to be independent of rBAT [1, 32]; and (3) reconstitution of system b0,+ transport activity in liposomes showed that b0,+AT is fully functional in the absence of rBAT [32]. This demonstrates that rBAT is not necessary for the proper insertion and folding of b0,+AT within the ER membrane, and that b0,+AT is the "catalytic subunit". Further research is needed to assess whether rBAT modifies the transport properties of rBAT/b0,+AT heterodimeric complex.


4F2hc is a ubiquitous multifunctional protein. In contrast to the rBAT /b0,+AT heterodimeric complex (i.e., system b0,+) in which mutations of both heavy and light subunits are involved in cystinuria (see above), mutations of 4F2hc have not been identified even in diseases such as lysinuric protein intolerance caused by the functional disruption of y+LAT-1 (SLC7A9) transporters, which require 4F2hc for their functional expression. The genetic defect of 4F2hc is believed to be lethal, because 4F2hc is the heavy subunit of several amino acid transporters (e.g., systems L, y+L, xc and asc) and it is also involved in a variety of cell functions as indicated below.

4F2hc heterodimeric complexes are sorted to the basolateral membrane, whereas the rBAT /b0,+AT heterodimeric complex is sorted to the apical membrane of the epithelial cells [1, 2]. Therefore, it is proposed that 4F2hc and rBAT are endowed with the ability to recognize their own partner and also with the signals to be sorted to their destination in epithelial cells. Although the molecular nature of these properties is unknown, some studies have addressed these issues [6]. Truncated 4F2hc consisting of the cytosolic N-terminus and the TM is sufficient to promote LAT1 expression at the membrane surface, whereas the extracellular domain of 4F2hc is necessary for the recognition of other light chains such as LAT2 and y+LAT2. Thus, it is suggested that different domains of 4F2hc are necessary for association with different light chains.

As mentioned above, 4F2hc is a multifunctional protein. It is also involved in a variety of activities such as cell activation, cell growth and cell adhesion. Thus, 4F2hc expression is up-regulated in cancers and activated lymphocytic cells, suggesting a role for 4F2hc in cell growth and malignant transformation (reviewed in [12, 14]). It was also shown that anti-4F2hc antibodies suppressed the growth of cancer cells, and 4F2hc over-expression in NIH3T3 cells resulted in their malignant transformation [33, 43]. Whether these effects are due to the role of 4F2hc supporting amino acid transport or to an unidentified role triggering intracellular signaling pathways that lead to cell growth stimulation and eventually malignant transformation is unknown. In the last instance, it has been suggested that 4F2hc plays a role in integrin function. The dynamic regulation of integrin affinity for ligands in response to cellular signals is central to integrin function. This process is mediated through integrin cytoplasmic domains. 4F2hc was identified as a protein involved in integrin signaling and regulating integrin activation by its capacity to complement dominant suppression in an expression cloning scheme [17]. It was suggested that the N-terminus domain of 4F2hc is associated intracellularly with the C-terminus domain of integrin β-subunits. Moreover, 4F2hc plays an essential role in the cell fusion induced by virus infection of eukaryotic cells. In fact two molecules that regulate cell fusion have been identified and designated fusion regulatory protein-1 (FRP-1) and FRP-2. FRP-1 and FRP-2 turned out to be heterodimeric 4F2 antigen and integrin α3 subunit respectively (reviewed in [36]). Therefore, it is reasonable to speculate that 4F2hc is involved in complex cellular signaling involving multiple pathways related to cell growth, cell adhesion and malignant transformation.

Are there other SLC3 members?

Two members of the SLC7 family, Asc-2 and AGT1, appear as heterodimers in the plasma membrane but they do not co-localize with 4F2hc and rBAT in vivo [10, 26]. Furthermore, the co-expression of these light subunits with rBAT or 4F2hc did not induce amino acid transport through the cell membrane. The ability of 4F2hc and rBAT to bring the light subunits of HATs to the plasma membrane was used to generate fusion proteins with Asc-2 or AGT1 that fully express amino acid transport activity [10, 26]. The fusion proteins made with 4F2hc or rBAT exhibited basically identical substrate selectivity and kinetic properties in each case, further supporting the hypothesis that the heavy subunits are not involved in the transport function itself of the corresponding HAT holotransporter. Homology search through the genome databases did not result in the identification of new SLC3 members. Further research is needed to assess whether the heavy subunits of Asc-2 and AGT1 are distant homologs of the SLC3 family or belong to a new family of heavy chains of HATs.


We thank Robin Rycroft for editing the manuscript. The laboratory of M.P. is supported by the Spanish Dirección General de Investigación Científica y Técnica Research Grant PM 99/0172, the Comissionat per a Universitats i Recerca and Grant 2001SGR00118 from the Generalitat de Catalunya (Spain), and the Instituto de Salud Carlos III network G03/054 (Spain). The laboratory of Y.K. is supported by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan, the Japan Society for the Promotion of Science, and the Promotion and Mutual Aid Corporation for Private Schools of Japan.

Copyright information

© Springer-Verlag  2004