Mutations in SLC6A19, encoding B⁰AT1, cause Hartnup disorder

Abstract

Hartnup disorder, an autosomal recessive defect named after an English family described in 1956 (ref. 1), results from impaired transport of neutral amino acids across epithelial cells in renal proximal tubules and intestinal mucosa. Symptoms include transient manifestations of pellagra (rashes), cerebellar ataxia and psychosis^1,2. Using homozygosity mapping in the original family in whom Hartnup disorder was discovered, we confirmed that the critical region for one causative gene was located on chromosome 5p15 (ref. 3). This region is homologous to the area of mouse chromosome 13 that encodes the sodium-dependent amino acid transporter B⁰AT1 (ref. 4). We isolated the human homolog of B⁰AT1, called SLC6A19, and determined its size and molecular organization. We then identified mutations in SLC6A19 in members of the original family in whom Hartnup disorder was discovered and of three Japanese families. The protein product of SLC6A19, the Hartnup transporter, is expressed primarily in intestine and renal proximal tubule and functions as a neutral amino acid transporter.

Main

Despite molecular characterization of other proximal tubule transporters, the neutral amino acid carrier defective in Hartnup disorder (OMIM 2345000) has resisted genetic identification². We carried out homozygosity mapping and fine mapping in ten members of two consanguineous families (the siblings in whom Hartnup disorder was originally discovered¹; family A; Fig. 1a) and in siblings from the US⁵ (family B; Fig. 1a). We found linkage of Hartnup disorder to 5p15 only in family A, with a maximum combined multipoint lod score of 2.31 at 11.24 cM (P = 0.01). This confirmed our previous results showing linkage to chromosome 5p15 (ref. 3). In family B, we obtained a maximum multipoint lod score of −2.40 at 15.81 cM.

Figure 1: Families with Hartnup disorder and topology of B⁰AT1 and its mutations.

(a) Families with Hartnup disorder. In family A (in whom Hartnup disorder was originally discovered), a splice site mutation in SLC6A19 segregates with Hartnup disorder. Asterisks indicate individuals from whom DNA was available for analysis. (b) Depiction of B⁰AT1 in the plasma membrane. Mutations are indicated in red; arrowheads indicate splice sites.

We simultaneously pursued two mouse monoamine transporter-related orphan genes, Slc6a18 (also called Xtrp2; ref. 6) and Slc6a19 (encoding B⁰AT1; ref. 4). These members of the SLC6 family of transporters map to the mouse chromosomal region that is homologous to human chromosome 5p15. Both Slc6a18 and Slc6a19 showed abundant expression in mouse kidney, as assessed by real time RT-PCR (Fig. 2a). Immunohistochemistry confirmed expression of mouse B⁰AT1 at the brush border of small intestine (data not shown) and kidney proximal tubule cells (Fig. 2b).

Figure 2: Expression of B⁰AT1 and Slc6a18 in mouse tissues and human tissue expression profile for the Hartnup transporter SLC6A19 (encoding B⁰AT1).

(a) Relative expression ratios from real-time RT-PCR expression profiles of mouse Slc6a19 (black bars) and mouse Slc6a18 (gray bars). Error bars represent s.e.m. (b) Immunofluorescence of B⁰AT1 (red) and of the basolateral protein 4F2hc (green) indicates the luminal expression of B⁰AT1 in kidney proximal tubule (highest in segment S1). G, glomerulus. The proximal tubule cells of mouse kidney start in the glomerular capsule. (c) PCR amplification of SLC6A19 cDNA in a panel of human tissues. SLC6A19 was highly expressed in kidney and small intestine, and weakly expressed in pancreas. (d) SLC6A19 cDNA was amplified from cells isolated from all segments of human proximal tubule (S1/S2 and S3), but not from the glomerulus, medullary thick ascending limb (mTAL), cortical thick ascending limb (cTAL) or distal tubule (DT).

The human homolog, B⁰AT1, is encoded by the predicted locus SLC6A19, with a 2,022-bp open reading frame. PCR amplification using human kidney cDNA produced a 1,905-bp product with 100% identity to SLC6A19 sequence. We next determined the genomic organization of SLC6A19, which has a stop codon 28 bases before the ATG in the 5′ untranslated region. SLC6A19 has 12 coding exons. The B⁰AT1 protein contains 634 amino acids and 12 predicted transmembrane regions (Fig. 1b). In a panel of human cDNAs, we detected robust expression of SLC6A19 in kidney and small intestine, with minimal expression in pancreas (Fig. 2c). SLC6A19 was also expressed in stomach, liver, duodenum and ileocecum (data not shown). In contrast, human SLC6A18 was abundantly expressed in human kidney but not in human intestine (data not shown). These findings indicated that SLC6A19 was the primary candidate for the gene causing Hartnup disorder.

We sequenced SLC6A19 in six families with typical neutral aminoaciduria (Fig. 1a) and identified five mutations in SLC6A19 (deletion, nonsense, missense and splice site) in four families (Table 1). Family A had a homozygous splice site mutation, IVS8 + 2T → G, that segregated with the phenotype. Family C (Japanese) had two affected individuals who were homozygous with respect to 884–885delTG in exon 3. A single affected member of family D (Japanese) had a homozygous missense mutation in exon 1, 169C → T, changing the conserved Arg57 to a cysteine residue. This occurred in the first membrane segment of B⁰AT1 (Fig. 1b) and was not present in 100 Japanese control alleles. We identified an individual with asymptomatic Hartnup disorder in family E (Japanese; Fig. 1a) by neonatal mass screening. He was compound heterozygous with respect to c340delC in exon 2 and the nonsense mutation 682–683AC → TA in exon 5, with termination at amino acid 228. Families B and F, American sibships with neutral aminoaciduria^5,7, lacked mutations in exons 1–12 of SLC6A19 and in the gene's splice sites. This finding, and the absence of linkage to chromosome 5p15 in family B, suggests that there is locus heterogeneity. Other genes, such as those encoding a complementary transporter protein, a putative associated protein, a subunit or an efflux transporter at the basolateral side of proximal tubule cells, may be involved in transepithelial neutral amino acid transport. These possibilities have precedent in the proximal tubule transport of dibasic amino acids, which is impaired in cystinuria by defects in either subunit of the heterodimeric apical cystine transporter rBAT-b^0,+AT or, in the case of lysinuric protein intolerance, in the basolateral transporter 4F2hc-y⁺LAT1 (refs. 8,9).

Table 1 Mutations in SLC6A19 associated with Hartnup disorder

The cellular expression of SLC6A19 was consistent with its putative role in plasma membrane transport. We amplified total cDNA from different portions of human renal tubules by PCR using SLC6A19 primers. Expression was apparent in all segments of the proximal tubules, but not in the glomerulus, thick ascending limb or distal tubule (Fig. 2d). Mouse B⁰AT1 transported neutral amino acids in Xenopus laevis oocytes; the K_m for leucine was ∼0.63 mM, and transport was sodium-dependent and specific for neutral amino acids⁴. We expressed human B⁰AT1 in oocytes and found sodium-dependent neutral amino acid transport with a K_m for leucine of 1.2 mM (Fig. 3a–c). Transport was abolished by the missense mutation R57C in family D (Fig. 3d).

Figure 3: Function of SLC6A19 (encoding B⁰AT1).

(a) B⁰AT1-mediated transport of L-leucine (1 mM) was sodium-dependent. (b) The transport of ¹⁴C-labeled amino acids (1 mM) mediated by hB⁰AT1 demonstrated substrate selectivity for neutral amino acids. (c) B⁰AT1-mediated transport of ¹⁴C-labeled leucine was saturable and conformed to Michaelis-Menten kinetics, with K_m of 1.2 mM and V_max of 7.7 pmol per oocyte per min. Inset, Eadie-Hofstee plot of leucine uptake. (d) The R57C mutant of B⁰AT1 (found in family D) yielded minimal uptake of ¹⁴C-labeled leucine (1 mM). Error bars represent s.e.m.

On the basis of mapping data, tissue and cellular expression, functional studies, and mutation identification in the family in whom Hartnup disorder was originally discovered and in three other families, we conclude that B⁰AT1 is the Hartnup transporter, joining the transporters in the plasma membrane that are involved in human disease¹⁰. Similar results and conclusions were found in Australian individuals with Hartnup disorder¹¹.

Methods

Mapping.

We obtained genomic DNA from whole blood or as published previously³ from members of families A, B, C, D and E after obtaining informed consent under a protocol approved by the National Human Genome Research Institute Institutional Review Board³. For family F, we extracted DNA from cells obtained from the Coriell Human Mutant Cell Repository. A genome scan using automated fluorescent microsatellite analysis of 2,011 markers, with an average marker density of 2 cM throughout the genome and an average marker heterozygosity of ∼0.75, was done at deCODE Genetics. For fine mapping between D5S1981 and D5S464, we used 11 polymorphic markers with an average spacing of 426 kb. We carried out linkage analyses using mapping by homozygosity with the software HOMOZ that calculates multipoint lod scores in pedigrees with inbreeding loops. We carried out two-point and multipoint parametric and nonparametric analyses of linkage using LINKAGE, with the speed improvement implemented in FASTLINK and GENEHUNTER, respectively^12,13,14. We used the high-performance computational capabilities of the SGI Origin 2000 system at the Center for Information Technology of the National Institutes of Health.

Sequencing and mutation analysis.

We chose intronic primers (Invitrogen) to sequence exons and splice sites of each exon. Primer sequences are available on request. We separated the PCR products of each exon by electrophoresis on agarose gels and removed specific bands for DNA isolation. We sequenced this DNA in both directions using a Beckman Coulter CEQ8000 following the manufacturer's protocol or as published previously³.

RNA extraction and reverse transcription.

We killed male C57BL6 mice by intraperitoneal injection of ketamine and xylazine followed by cervical dislocation. We then collected tissues and rapidly froze them until further use. We collected stomach, duodenum, jejunum, ileum and colon, rinsed them several times with ice-cold phosphate buffered saline (PBS; pH 7.4), scraped off the mucosa cell layers on ice and rapidly froze them. For total RNA extraction, we thawed tissues in RLT buffer of the RNeasy Mini Kit (Qiagen) containing 10 μl ml⁻¹ of β-mercaptoethanol (Sigma) and homogenized them on ice. RNA was bound to columns and treated with DNase for 15 min at 25 °C to reduce genomic DNA contamination. We used 100 ng of RNA as template for reverse transcription with TaqMan Reverse Transcription Kit (Applied Biosystems) in the presence of 2.5 μM of random hexamer primers.

Real-time PCR.

We carried out real-time PCR using cDNA template and TaqMan Universal PCR master mix (Applied Biosystems). We chose primers to result in amplicons of 70–100 bp that span intron-exon boundaries. We labeled probes with reporter dye FAM at the 5′ end and the quencher dye TAMRA at the 3′ end (Microsynth). Reactions were done in 96-well optical reaction plates using a Prism 7700 cycler (Applied Biosystems). After 10 min at 95 °C, we carried out 40 cycles of 95 °C for 15 s and 60 °C for 1 min with auto-ramp time. For data analysis, we set the threshold to 0.06 (value in the linear range of amplification curves). All reactions were run in duplicate. We calculated the abundance of target mRNA relative to a reference mRNA (GAPD). Assuming an efficiency of 2 (relative increase in template mRNA required to decrease by 1 the number of cycles), we calculated relative expression ratios as R = 2^{(Ct(GAPD)−Ct(test))}, where Ct is the cycle number at the threshold and test is the tested mRNAs.

Immunohistochemistry.

We anesthetized male C57BL6 mice (10–12 weeks old) with ketamine and xylazine and perfused them through the left ventricle with PBS followed by a buffered paraformaldehyde solution (4%, pH 7). We removed kidneys and small intestine, flushed them with fixative solution, cut them and embedded them in Tissue-Tek V.I.P. medium (Sakura) just before freezing in liquid nitrogen. We cut 5-μm serial sections with a cryostat and collected them on polylysine-coated slides. We incubated sections for 5 min in 0.1% SDS, washed them three times and then incubated them with the primary antibody (diluted in PBS, 0.04% Triton X-100) for 1 h at room temperature. We again washed sections three times and then incubated them with the secondary antibody (diluted in PBS, 0.04% Triton X-100) for 1 h at room temperature. After washing them five times, we mounted the sections using DAKO mounting medium and viewed them with a Leica SP1 UV CLSM confocal microscope. We used rabbit antibody to mouse B⁰AT1 (antigen peptide NH₂-MVRLVLPNPGLEERIC-CONH₂; serum, dilution 1:200) and goat antibody to mouse 4F2hc (Santa Cruz Biotech, 1:400) as primary antibodies and donkey antibody to rabbit (1:1,000) and donkey antibody to goat (1:500; Molecular Probes) as secondary antibodies.

Segmental expression in human kidney.

We carried out segment-specific dissection according to a previously described protocol¹⁵. We extracted total RNA and reverse transcribed it to generate segment-specific cDNA as described¹⁶.

Functional studies.

We carried out RT-PCR of SLC6A19 cDNA (primer sequences available on request) on the first-strand cDNA generated from human kidney poly(A)⁺ RNA (Clontech). We cloned the PCR fragment into pCR2.1-TOPO vector (Invitrogen). For the expression in oocytes, we excised the SLC6A19 cDNA with EcoRI and subcloned it into pcDNA3.1(+) vector (Invitrogen). We carried out site-directed mutagenesis (R57C) of SLC6A19 cDNA in the pcDNA3.1(+) plasmid using the QuikChange kit (Stratagene). We checked the accuracy of the cDNAs by direct sequencing. We injected X. laevis oocytes with 25 ng of cRNA synthesized in vitro from SLC6A19 cDNA in the pcDNA3.1(+) plasmid linearized with NotI. We carried out in vitro transcription using the T7 mMESSAGE mMACHINE Kit (Ambion) and polyadenylation of the cRNA using the Poly(A) Tailing Kit (Ambion). We measured transport after 2–3 d of expression as described¹⁷. We measured uptake of ¹⁴C-labeled amino acids in the regular uptake solution (100 mM NaCl, 2 mM KCl, 1.8 mM CaCl₂, 1 mM MgCl₂ and 5 mM HEPES buffer (pH 7.4)) or in a solution in which NaCl was replaced by choline-Cl, containing 2.5 μCi ml⁻¹ of radioactively labeled compounds and 0.03–2 mM of amino acids. We measured uptake (pmol per oocyte per min) in triplicate for 30 min.

GenBank accession numbers.

SLC6A19 open reading frame, XM_291120; SLC6A19 cDNA sequence, AY596807.