Essential amino acid transporter Lat4 (Slc43a2) is required for mouse development

Abstract

Amino acid (AA) uniporter Lat4 (Slc43a2) mediates facilitated diffusion of branched-chain AAs, methionine and phenylalanine, although its physiological role and subcellular localization are not known. We report that Slc43a2 knockout mice were born at expected Mendelian frequency but displayed an ∼10% intrauterine growth retardation and low amniotic fluid AAs, suggesting defective transplacental transport. Postnatal growth was strongly reduced, with premature death occurring within 9 days such that further investigations were made within 3 days of birth. Lat4 immunofluorescence showed a strong basolateral signal in the small intestine, kidney proximal tubule and thick ascending limb epithelial cells of wild-type but not Slc43a2 null littermates and no signal in liver and skeletal muscle. Experiments using Xenopus laevis oocytes demonstrated that Lat4 functioned as a symmetrical low affinity uniporter with a K_0.5 of ∼5 mm for both in- and efflux. Plasma AA concentration was decreased in Slc43a2 null pups, in particular that of non-essential AAs alanine, serine, histidine and proline. Together with an increased level of plasma long chain acylcarnitines and a strong alteration of liver gene expression, this indicates malnutrition. Attempts to rescue pups by decreasing the litter size or by nutrients injected i.p. did not succeed. Radioactively labelled leucine but not lysine given per os accumulated in the small intestine of Slc43a2null pups, suggesting the defective transcellular transport of Lat4 substrates. In summary, Lat4 is a symmetrical uniporter for neutral essential AAs localizing at the basolateral side of (re)absorbing epithelia and is necessary for early nutrition and development.

Key points

• Lat4 (Slc43a2) transports branched-chain amino acids, phenylalanine and methionine, and is expressed in kidney tubule and small intestine epithelial cells.

• Using a new knockout model as a negative control, it is shown that Lat4 is expressed at the basolateral side of small intestine enterocytes and kidney epithelial cells of the proximal tubule, thick ascending limb and distal convoluted tubule.

• In the Xenopus oocyte expression system, Lat4 is shown to function as a uniporter with symmetric intracellular and extracellular apparent affinities for phenylalanine.

• Mice lacking Lat4 display a slight intrauterine growth restriction, postnatal malnutrition and early death, presumably as a result of defective amino acid (re)absorption.

• These results demonstrate the crucial role that the uniporter Lat4 plays for amino acid transport across cellular barriers and mouse development.

Introduction

Dietary protein intake is an essential and tightly regulated process that needs to meet physiological body requirements, especially during highly anabolic periods (Poncet & Taylor, 2013). The enzyme-mediated hydrolysis of dietary proteins in the gastrointestinal lumen produces oligopeptides and individual amino acids (AAs) that are absorbed by moving sequentially across apical and basolateral cell membranes of small intestine enterocytes (Nassl et al. 2011). The zwitterionic nature of neutral AAs prevents the direct crossing of the epithelial lipid bilayer by simple diffusion and imposes the presence of membrane-spanning transporter proteins (Braun et al. 2011). Severe malabsorption and aminoaciduria syndromes have been linked to defects of specific AA transporters, substantiating the essential role that transepithelial transport plays for body AA homeostasis (Verrey et al. 2009; Broer & Palacin, 2011).

The luminal step of neutral AA absorption and reabsorption at the level of the small intestine and kidney, respectively, is largely mediated by the apical transporter B⁰AT1 (SLC6A19). This luminal symporter uses the driving force exerted by the Na⁺ gradient and the membrane potential on the co-transported Na⁺ ion for the concentrative import of its neutral AA substrates (secondary active transport) (Rudnick et al. 2014). Once inside enterocytes or kidney proximal tubule cells, neutral AAs can be effluxed across the basolateral membrane by cooperating AA exchangers and uniporters. Indeed, all neutral AAs can be effluxed by the major basolateral neutral AA anti-porter (obligatory exchanger) LAT2–4F2hc (SLC7A8-SLC3A2). It is, however, important to realize that, for each exported AA, this anti-porter imports another AA. Thus, this anti-porter does not perform a net transmembrane AA efflux (Meier et al. 2002). Unexpectedly, no genetic diseases involving defects of LAT2–4F2hc have been described to date, whereas mutations in the other basolateral anti-porter y⁺LAT1–4F2hc (SLC7A7-SLC3A2), which exchanges preferentially intracellular cationic AAs against neutral AAs and Na⁺, cause lysinuric protein intolerance, an autosomal recessive disease (Borsani et al. 1999). The respective physiological roles of both exchangers were also investigated using knockout (KO) mice. The lack of LAT2 resulted in a mild phenotype with a minor aminoaciduria (Braun et al. 2011), whereas the lack of y⁺LAT1 caused intrauterine growth restriction and resulted in premature death for ∼90% of newborn Slc7a7^−/− homozygous mice (Sperandeo et al. 2007).

To achieve a net efflux of all different AAs from small intestine enterocytes and proximal kidney tubule epithelial cells, transporters are required to mediate the net efflux of some AAs that can then recycle into the cells via the above mentioned anti-porters and thereby drive their efflux activity (Verrey et al. 2009). The T-type aromatic AA uniporter TAT1 (SLC16A10) represents the best characterized such transporter expressed at the basolateral membrane (Ramadan et al. 2006; Ramadan et al. 2007). Using the Xenopus laevis oocyte expression system, it was shown to mediate facilitated diffusion of the essential aromatic AAs phenylalanine, tryptophan and tyrosine with symmetric apparent affinities (i.e. K_m ≈ 30 mm for phenylalanine) and thereby drive the efflux of other LAT2–4F2hc substrates such as the non-essential AA glutamine (Ramadan et al. 2006; Ramadan et al. 2007). Recently, we characterized the physiological role of TAT1 using Slc16a10 deficient mice. Interestingly, these mice presented a mild phenotype with increased plasma, kidney and muscle aromatic AAs, independently of dietary protein quantity, combined with a major aromatic aminoaciduria under a high protein diet (Mariotta et al. 2012). The lack of neurological symptoms and important metabolic dysfunctions, despite the defective intestinal absorption of aromatic AAs such as tyrosine, which is the metabolic precursors of the hormone thyroxine and the neurotransmitter dopamine, suggested a possible compensatory mechanism by (an)other basolateral uniporter(s). We proposed that the uniporter LAT4 (SLC43A2) could play this role.

LAT4 belongs to the SLC43 family that includes two additional members: LAT3 (SLC43A1), which is a uniporter for large essential neutral AAs and has been shown to be upregulated in liver and skeletal muscle upon starvation and displays only a low mRNA expression level in the small intestine and kidney (Babu et al. 2003; Fukuhara et al. 2007), and EEG1 (SLC43A3), the function of which remains unknown. The latter gene product was shown to be prominently expressed in human liver and heart and, based on immunohistochemistry, to be expressed at the basolateral membrane of human kidney proximal convoluted tubule (Bodoy et al. 2005). LAT4 shows only ∼30% identity with EEG1 and has been shown by Northern blot analysis of human RNA to be highly expressed in placenta, peripheral blood leucocytes and kidney and, additionally, by Northern blot analysis of mouse RNA, to be expressed in the small intestine and brain (Bodoy et al. 2005). Microarray data (www.biogps.org) confirm this expression pattern and additionally show substantial expression in macrophages, microglia and osteoclasts. Based on in situ hybridization, Lat4 was suggested to be expressed in the kidney distal tubule and collecting duct and the images also appeared to be compatible with expression in the proximal tubule. In the small intestine, the signal was localized more to crypt cells (Bodoy et al. 2005). Lat4 mRNA was also detected in cell lines isolated from the mouse kidney proximal tubule, in agreement with the mRNA expression in isolated proximal tubules reported by Cheval et al. (2011). Furthermore, kinetic analysis of phenylalanine uptake in these cells suggested the involvement of Lat4, supporting the possibility that this uniporter plays a role in proximal tubule AA reabsorption.

The preferential substrates of LAT4 comprise only essential AAs, specifically the branched-chain ones (leucine, isoleucine, valine) and additionally phenylalanine and methionine (Bodoy et al. 2005) To test the physiological role of Lat4 in whole-body AA homeostasis and epithelial transport, we generated a global Slc43a2^–/– KO mouse and we now report the analysis of its phenotype.

Methods

Ethical approval

All procedures for mice and X. laevis frogs handling and experimental interventions were performed in accordance with Swiss Animal Welfare laws and approved by the Cantonal Veterinary Office, Zürich.

Generation of Slc43a2/Lat4^−/− KO mice by gene trapping

The Slc43a2^−/− KO mouse was produced by the Phenogenomic Center of Toronto (www.phenogenomics.ca). Embryonic stem cell lines were mutated by insertional mutagenesis using a UPA gene trap vector (Shigeoka et al. 2005) in the sixth intron of the Slc43a2 gene. Cells were prepared for aggregation with diploid embryos to produce chimeric mice. R1 cells were on a 129 mouse background, whereas the germline was tested by crossing the chimera to CD-1 mice and assessing coat colour. The interruption of the gene leads to the production of a truncated transcript presumably subject to nonsense-mediated decay and thus of low abundance. Mice were bred on a C57BL/6 background. All animals were housed under standard conditions and fed a standard diet prior to experimentation.

Antibodies

Anti-mouse Lat4 antibodies targeted to the C-terminal peptide (N-CQLQQKREDSKLFL-C) were raised in rabbit, whereas anti-mouse Tmem27 antibodies targeted to the C-terminal peptide (N-CDPLDMKGGHINDGFLT-C) and anti-mouse B⁰AT1 antibodies targeted to the N-terminal peptide (N-MVRLVLPNPGLEERIC-C) were raised in guinea-pig (Pineda, Berlin, Germany). Anti-mouse sodium-chloride cotransporter (NCC) (Loffing et al. 2004), anti-mouse sodium-potassium-chloride cotransporter 2 (NKCC2) (Kaplan et al. 1996) and anti-mouse MDR1 (Fickert et al. 2001) were previously characterized. For β-actin detection, a commercially available mouse anti-β-actin antibody was used (Sigma-Aldrich, Buchs, Switzerland).

Immunofluorescence

Organs were isolated and immediately frozen in liquid propane. Sections, 7 μm thick, were prepared on Superfrost Plus Menzel slides (Gerhard Menzel GmbH, Braunschweig, Germany). To test for Lat4 localization in NKCC2 expressing kidney tubule segments, 4 μm consecutive sections were prepared. Sections were then fixed with methanol at −20°C for 90 s, rehydrated for 15 min at room temperature (RT) and washed three times for 5 min with PBS at RT. For Lat4, NCC, MDR1, Tmem27 and B°AT1 detection, epitope retrieval was performed by incubating slides in 10 mm sodium citrate (pH 6.0) for 10 min at 98°C with a microwave (Histos 3; Milestone, Shelton, CT, USA). For NKCC2 and Tat1 detection, no epitope retrieval was performed. Sections were then washed again three times for 5 min with PBS at RT and incubated with blocking solution (2% BSA, 0.04% Triton X-100 in PBS, pH 7.4) for 1 h at RT. Primary antibodies were diluted in blocking solution and incubated overnight at 4°C with the respective dilutions: 1:500 for Lat4, 1:1000 for Tmem27, 1:2000 for NKCC2, 1:500 for NCC, 1:500 for Tat1, 1:300 for MDR1 and 1:500 for B°AT1. Sections were then washed three times for 5 min in PBS at RT. Secondary antibodies were applied at 1:500 dilutions for 1 h at RT: Alexa Fluor 488 anti-rabbit, Alexa Fluor 594 anti-rabbit and Alexa Fluor 568 anti-guinea-pig (Molecular Probes, Life Technologies, Grand Island, NY, USA). Staining of nuclei was performed with 4′,6-diamidino-2-phenylindole at 1:500 dilution.

Western blotting on oocyte total membrane lysates

Total membrane lysates of oocytes and western blotting were performed as described previously (Ramadan et al. 2007). For β-actin detection, the antibody was applied at 1:5000 dilution. For Lat4 detection, the antibody was applied at 1:1000 dilution.

Morphological analysis of organs

For isolation of organs, the preparation and section fixation of 7 μm sections was performed as described above. Classical heamatoxylin and eosin staining was performed as described previously (Mariotta et al. 2012).

Cloning of hLAT4 cDNA and cRNA synthesis

Human LAT4 cDNA in pTLN vector was kindly provided by Manuel Palacin (Bodoy et al. 2005). The construct was linearized using the BglII restriction site and cRNA was synthesized using the MEGAscript high-yield in vitro transcription kit (Ambion, Austin, TX, USA) in accordance with the manufacturer's instructions.

Xenopus laevis oocytes preparation and cRNA injection

Stage IV oocytes (Dumont, 1972) were treated with collagenase A for 2–3 h at room temperature in Ca²⁺-free buffer (82.5 mm NaCl, 2 mm MgCl₂, 10 mm Hepes, pH 7.4) and kept at 16°C in modified Barth's solution (88 mm NaCl, 1 mm KCl, 0.82 mm MgSO₄, 0.41 mm CaCl₂, 0.33 mm Ca(NO₃)₂, 2.4 mm NaHCO₃, 10 mm Hepes, 5 mg l⁻¹ gentamicin, 5 mg l⁻¹ doxycycline). Remaining follicular layers were manually removed. Next, 25 ng of LAT4 cRNA was injected and oocytes were incubated for 3 days at 16°C in ND96 solution (96 mm NaCl, 2 mm KCl, 1 mm MgCl₂, 1.8 mm CaCl₂, 5 mm Hepes-Tris, pH 7.4, supplemented with 50 mg l⁻¹ tetracycline).

Uptake of AAs in X. laevis oocytes

Oocytes were washed six times with 2 ml of Na⁺-buffer (100 mm NaCl, 2 mm KCl, 1 mm CaCl₂, 1 mm MgCl₂, 10 mm Hepes, pH 7.4) and then pre-warmed for 2 min at 25°C in a water bath. Subsequently, the pre-warming solution was aspirated and 100 μl of Na⁺-buffer containing unlabelled phenylalanine at the indicated concentrations together with [³H]-l-phenylalanine (10 μCi ml⁻¹) as tracer (Hartmann Analytic, Braunschweig, Germany) were added. Oocytes were then incubated for 10 min at 25°C. Thereafter, uptake solution was removed, oocytes were washed six times with 2 ml of ice-cold Na⁺-buffer and separately dissolved in 250 μl of 2% SDS under agitation for at least 30 min. Upon the addition of 3 ml of scintillation fluid (Emulsifier-Safe), radioactivity was counted on a liquid scintillation counter (TRI-CARB 2900TR; Packard Instrument, Meriden, CT, USA). Uptake values were corrected by subtracting the values obtained from non-injected oocytes. Curves corresponding to the Michaelis–Menten equation were fitted to the data sets using GraphPad Prism, version 5.0 (GraphPad Software, San Diego, CA, USA).

Efflux of AAs in X. laevis oocytes

Oocytes were injected with 50 nl of different phenylalanine solutions (10 mm, 25 mm, 50 mm, 100 mm and 200 mm) together with [³H]-l-phenylalanine (0.2 μCi μl⁻¹) as tracer. Assuming an oocyte volume of ∼400 μl (Taylor & Smith, 1987), the resulting oocyte internal phenylalanine concentration upon injection corresponded, respectively, to 1.11 mm, 2.78 mm, 5.56 mm, 11.11 mm and 22.22 mm. Oocytes were subsequently washed four times in ND96 solution and 200 μl of ND96 were given to each oocyte. Efflux was allowed to take place for 10 min at 25°C before recovering the ND96 solution and washing the oocyte four times with ND96. Oocytes lysis and data analysis were conducted as for the uptake of AAs in X. laevis oocytes.

RNA extraction and real-time PCR

Mice at the indicated day after birth were killed by decapitation and organs were immediately rinsed several times with ice-cold PBS prior to being rapidly frozen in liquid propane until further use. Total RNA was extracted and reverse transcription was performed as described previously (Ramadan et al. 2006). For real-time PCR, 10 ng of cDNA template were used and the reaction was set in accordance with the manufacturer's instructions (TaqMan® Universal PCR master mix; Applied Biosystems, Foster City, CA, USA). The abundance of target mRNAs was standardized relative to 18S ribosomal RNA. Primers and probes for Lat2 (Slc7a8), 4F2hc (Slc3a2), y⁺Lat1 (Slc7a7), y⁺Lat2 (Slc7a6), Asct2 (Slc1a5), Tat1 (Slc16a10) and B⁰AT1 (Slc6a19) were described previously (Moret et al. 2007; Mariotta et al. 2012). For Lat4 detection, the primers and probe used were: 5′-GCTGATTGCATATGGAGCAAGTAAC-3′ (forward), 5′-CGAAGTGAACGTCATGCACAT-3′ (reverse) and 5′-CTCTCTGTGCTCATCTTTATCGCCTTGGC-3′. For Eeg1 detection, the primers and probe used were: 5′-GCTACATCTTTGACCGCTTCAA-3′ (forward), 5′-CGCAGGTGTAGAAAAATATGGCTAT-3′ (reverse) and 5′-ACTACTGTGGCCCGCC-3′ (probe). For Lat3 detection, the primers and probe used were: 5′-CCCTGAATGAGAATGCTTCCTT-3′ (forward), 5′-ATGGCATTGGTGAGCTTTTGT-3′ (reverse) and 5′-AGCACCAAGTTCACTAGACCACGCTACCG-3′ (probe).

Isolation of pups at embryonic day 18 (E18) and measurements of AAs in amniotic fluids

Slc43a2^+/− pregnant females were killed by cervical dislocation and amniotic sacs were dissected using standard laparotomy. Amniotic fluids were collected by puncturing the amniotic sacs and AA measurements were performed as described previously (Mariotta et al. 2012). E18 pups were subsequently isolated from their amniotic sacs and weighed.

Plasma AA and acylcarnitine measurements

Amino acids and acylcarnitines were measured after extraction with methanol without derivatization (Neobase Kit; Perkin Elmer, Turku, Finnland) on an Acquity-TQD tandem mass spectrometer (Waters, Milford, MA, USA). Briefly, 3 μl of plasma were spotted onto prepunched plain filter paper (Ahlstom 226; Perkin Elmer) in 96-well plain uncoated microtiter plates. Amino acids and acylcarnitines were extracted with 100 μl of methanolic extraction solution containing stable isotope labelled internal standards by shaking for 45 min at 45°C. Next, 30 μl of extract were then injected with a flow rate of 10 μl min⁻¹ directly into the tandem mass spectrometer and measured afterwards in multiple reaction monitoring mode.

RNA-sequencing analysis of liver

Organs were isolated and RNA was extracted as described above. The quality of the isolated RNA was determined with a Qubit (1.0) Fluorometer (Life Technologies) and a Bioanalyzer 2100 (Agilent, Waldbronn, Germany). Only those samples with a 260 nm/280 nm ratio between 1.8 and 2.1 and a 28S/18S ratio in the range 1.5–2 were processed further. The TruSeq RNA Sample Prep Kit, version 2 (Illumina, Inc., Hayward, CA, USA) was used in the successive steps. Briefly, total RNA samples (100–1000 ng) were polyA enriched and then reverse-transcribed into double-stranded cDNA. The cDNA samples were fragmented, end-repaired and polyadenylated before ligation of TruSeq adapters containing the index for multiplexing. Fragments containing TruSeq adapters on both ends were selectively enriched with PCR. The quality and quantity of the enriched libraries were validated using a Qubit (1.0) Fluorometer and the Caliper GX LabChip GX (Caliper Life Sciences, Inc., Hopkinton, MA, USA). The product is a smear with an average fragment size of ∼260 bp. The libraries were normalized to 10 nM in Tris-Cl 10 mm (pH 8.5) with 0.1% Tween 20. The TruSeq PE Cluster Kit v3-cBot-HS or TruSeq SR Cluster Kit v3-cBot-HS (Illumina, Inc.) was used for cluster generation using 10 pM of pooled normalized libraries on the cBOT. Sequencing were performed on the Illumina HiSeq 2000 paired end at 2 × 101 bp or single end 100 bp using the TruSeq SBS Kit v3-HS (Illumina, Inc.). RNA-sequencing reads were quality-checked with fastqc (www.bioinformatics.babraham.ac.uk/projects/fastqc), which computes various quality metrics for the raw reads. Reads were aligned to the genome and transcriptome with Tophat, version 1.3.3 (ccb.jhu.edu/software/tophat/index.shtml). Before mapping, the low quality ends of the reads were clipped (three bases from the read start and 10 bases from the read end). Tophat, version 1.3.3, was run with default options. The fragment length parameter was set to 100 bases with a standard deviation of 100 bases. Based on these alignments, the distribution of the reads across genomic features was assessed. Isoform expression was quantified with the RSEM algorithm (www.biomedcentral.com/1471-2105/12/323) with the option for estimation of the read start position distribution turned on. Raw data are available on the NCBI BioProject platform (www.ncbi.nlm.nih.gov/bioproject/) under the BioProject ID: PRJNA226691.

Determination of radiolabelled leucine and lysine distribution after oral administration

Mice at day 3 after birth were starved for 30 min in a humidified chamber on a warming plate set at 37°C. Subsequently, 5 μl of radiolabelled AA solution (1 mm leucine, [³H]-l-leucine 0.2 μCi μl⁻¹, 1 mm lysine, [¹⁴C]-l-lysine 0.04 μCi μl⁻¹) was administered and the mouse was placed in the chamber again for 30 min. Thereafter, mice were killed by decapitation and plasma was collected as described before. Organs were isolated, rinsed several times with ice-cold PBS and weighed. Subsequently, organs were lysed in Solvable (Perkin Elmer) overnight at 50°C and bleached with 200 μl of 30% H₂O₂. Radioactivity was determined upon addition of 12 ml of Ultima Gold scintillation liquid (Perkin Elmer) as described before (Mariotta et al., 2012). AA accumulation within each organ was normalized to organ weight.

Blood glucose measurement

Mice at day 2 were killed by decapitation and blood glucose was measured with reactive test stripes Accu-Check (Roche Diagnostics, Mannheim, Germany).

Results

Lat4 protein absence in Slc43a2^−/− mice and localization in wild-type mice epithelia

The expression and localization of the Lat4 protein was investigated using a new antibody that was validated by western blotting (data not shown). Immunofluorescence images of the small intestine and kidney proximal tubule of 3-day-old pups confirmed the absence of Lat4 protein in Slc43a2 null pups compared to wild-type littermates (Fig.1A–D). The integrity of the KO tissues showing no Lat4 signal was demonstrated by co-labelling with specific markers; for example, Tmem27 for kidney and MDR1 for liver (Fig.1A, B, G and H). The lack of Lat4 signal in Slc43a2 null mice additionally demonstrated the specificity of the immunofluorescence signals that localized to the basolateral membrane of kidney tubule (Fig.1A) and small intestine (Fig.1C) epithelial cells. As expected based on previous mRNA data (Bodoy et al. 2005), Lat4 expression was not detected in skeletal muscle and liver of wild-type and KO pups (Fig.1E–H).

Figure 1.

Confirmation of Lat4 deletion and validation of Lat4 antibody by immunofluorescence on tissue of wild-type and Lat4 KO pups

A basolateral signal is seen with Lat4 antibody (green) in the kidney (A) and small intestine (C) of 3-day-old wild-type pups, whereas no such signal is visible on sections from KO littermates (B and D). No signal (green) is observed in skeletal muscle (E) and liver (G) of wild-type pups in addition to the background also visible in the KO mice (F and H). As a luminal marker, Tmem27 (red) was used for kidney proximal tubule cells (A and B) and MDR1 was used for the canalicular membrane of liver hepatocytes (red) (G and H). Scale bars = 20 μm.

A strong Lat4-specific labelling was detected on small intestine villi at the basolateral membrane of enterocytes, whereas no clearly specific Lat4 signal was visible in the crypts (Fig.2A and B). To investigate the localization of Lat4 along the mouse kidney nephron in more detail, we performed co-stainings or stainings of consecutive sections with tubule-specific markers. The strong expression of Lat4 in the proximal tubule was confirmed by labelling of the luminal SLC6-transporter accessory protein Tmem27 (Fig.2C) and the aromatic AA uniporter Tat1 (consecutive sections Fig.2D and E) in the same cells (Danilczyk et al. 2006; Ramadan et al. 2006). A weaker expression of Lat4 was detected in the distal convoluted tubule, which was co-stained with apical NCC antibody (Fig.2F), whereas a strong expression of Lat4 was revealed in the thick-ascending limb of the loop of Henle, which was identified by NKCC2 labelling in a consecutive section (Fig.2G and H) (Wagner et al. 2008). Taken together, our immunofluorescence data show that Lat4 protein is expressed at the basolateral membrane of small intestine enterocytes, the kidney proximal tubule, thick ascending limb and, to a minor extent, of distal convoluted tubule epithelial cells.

Figure 2.

Localization of Lat4 on small intestine and kidney sections by immunofluorescence

A, on this cross-section of duodenal villi, Lat4 signal (green) localizes to the basolateral membrane of the enterocytes that are counterstained for the luminal transporter B⁰AT1 (red). B, on a sagittal section of duodenal villi, Lat4 signal (green) is visible all along the villi (V) but not in crypts (C). In both (A) and (B), 4′,6-diamidino-2-phenylindole (blue) was used to label the nuclei. On kidney sections, Lat4 (green) signal is detectable in the proximal tubule (PT) which is identified in (C) by the luminal protein Tmem27 and by the basolateral uniporter Tat1 (pink) on consecutive sections (D and E). F, labelling of the luminal transporter NCC (red) identifies the distal convoluted tubule (DCT), which displays a weaker Lat4 signal compared to the proximal tube. G and H, a strong Lat4 signal (green) is visible in the thick ascending limb (TAL), which is identified in consecutive sections by the luminal transporter NKCC2 (red). Scale bar = 20 μm.

Kinetic properties of LAT4 expressed in X. laevis oocytes

Originally, Lat4 was suggested to function as a uniporter based on the observation that it facilitated the efflux of phenylalanine from X. laevis oocytes independent of extracellular AAs (Bodoy et al. 2005). To determine the kinetic properties of Lat4, we performed a series of uptake and efflux rate measurements with different concentrations of phenylalanine using the X. laevis oocytes expression system (Fig.3). Fitting Michaelis–Menten kinetics curves to the experimental phenylalanine uptake results, we derived the Michaelian constant (K_0.5) of 5.2 ± 0.5 mm (R² = 0.89). This value corresponded to a low-affinity transport comparable to that previously reported by Bodoy et al. (2005) but no additional high-affinity/low capacity component was detected (Fig.3A). To determine the efflux kinetic properties, oocytes were injected with increasing concentrations of phenylalanine and the resulting intracellular concentration of phenylalanine was estimated assuming an oocyte volume of 400 nl, as described previously (Meier et al. 2002). Fitting the Michaelis–Menten kinetics to the efflux values measured after 10 min yielded a low-affinity K_0.5 of 6.0 ± 1.4 mm (R² = 0.66), which was comparable to the apparent affinity value obtained for the uptake (Fig.3B). Based on these results, we concluded that Lat4 shows symmetric intracellular and extracellular apparent affinities for phenylalanine.

Figure 3.

Concentration-dependence of l-phenylalanine uptake and efflux in X. laevis oocytes expressing human LAT4

Oocytes were injected with 25 ng of hSLC43A2 cRNA and incubated for 3 days. A, concentration-dependent uptakes were performed for 10 min with nine different concentrations of phenylalanine; n = 8–24 oocytes pooled from three independent experiments. B, for the efflux experiments, oocytes were injected with 50 nl of phenylalanine at five different concentrations; n = 12–14 oocytes pooled from three independent experiments. Data are the are mean ± SEM. Curves corresponding to the Michaelis–Menten equation were fitted to the experimental data using a non-linear regression routine (GraphPrism, version 5.0).

Intrauterine growth restriction and reduced amniotic fluid AA concentrations in Slc43a2^−/− mice

Lat4 is known to be expressed in mouse placenta (Bodoy et al. 2005), where it might play an important role in AA transport, in particular by cooperating with the anti-porter Lat2–4F2hc (Cleal et al. 2011). To investigate the role of Lat4 in placental AA transport and intrauterine development, we analysed fetuses isolated from Slc43a2^+/− inter se crossed mice at E18 (Fig.4). Slc43a2^−/− fetuses displayed a weight reduction of ∼10% compared to Slc43a2^+/+ and Slc43a2^+/− littermates (E18 mean weight of ∼0.82 g for Slc43a2^−/− versus 0.93 g for Slc43a2^+/+) (Fig.4B) but no additional obvious phenotypical sign (Fig.4A). This reduced intrauterine growth did not result in prenatal lethality because the birth statistics of Slc43a2^−/− mice did not differ from the expected Mendelian ratio (P = 0.48, χ² test with N > 100 pups). To determine the effect of Lat4 deletion on placental AA transport, we measured the AA concentration in the amniotic fluid of individual pups (Fig.4C). Interestingly, Slc43a2^−/− mice displayed a marked reduction (i.e. > 50% compared to Slc43a2^+/+) in the concentration of almost all AAs, with the exception of aspartate and glutamate that did not show a statistically significant change. Measurements of other parameters, such as glucose and total protein concentration or osmolarity, did not show any difference (data not shown), pointing towards a specific defect of fetal AA supply. Therefore, we concluded that the lack of Lat4 results in a defective fetal AA delivery, which is not only restricted to Lat4-specific substrates but includes most AAs. These findings suggested that the defective Lat4 transport had led prenatally to a general decrease of AA-related transport, resulting in reduced intrauterine growth.

Figure 4.

Effect of Slc43a2 deletion on intrauterine growth and AA concentrations in amniotic fluid at E18

A, representative image of E18 Slc43a2^+/+ and Slc43a2^−/− fetuses with corresponding weights. Scale bar = 5 mm. B, AA concentrations in amniotic fluids (C). Results were normalized to Slc43a2^+/+ (1.0) and represent the mean ± SEM of data pooled from two independent experiments: n = 7 for Slc43a2^+/+ mice, n = 15 for Slc43a2^+/− mice and n = 4 for Slc43a2^−/− mice. Two-way ANOVA and Dunnet's multiple comparison post hoc test for (B) and Bonferroni post hoc test for (C): **P < 0.01. NS, non-significant.

Growth defect, metabolic alterations and early postnatal lethality of Slc43a2^−/− mice

Despite the reduced intrauterine growth reported above, Slc43a2^−/− pups showed no pathological sign at birth. They displayed a normal skin colour and the presence of a white area at the left side of the abdomen (i.e. the so-called ‘milk spot’ confirmed normal milk delivery by the mother). However, monitoring of the postnatal weight revealed that Slc43a2^−/− pups gained only very little weight compared to littermates (Fig.5A and B). Furthermore, they all died within few days after birth, as shown in the Kaplan–Meier survival analysis (Fig.5C). Indeed, more than 50% of the observed Slc43a2^−/− pups died until postnatal day 7 and mortality reached 100% at day 10.

Figure 5.

Postnatal growth retardation and premature death of Slc43a2^−/− mice

A, representative image of a Slc43a2^−/− mouse (top) versus a Slc43a2^+/+mouse (bottom) at day 7 after birth. B, growth curves for Slc43a2^+/+, Slc43a2^+/- and Slc43a2^−/− mice; represented as the mean ± SEM: n = 7 for Slc43a2^+/+, n = 22 for Slc43a2^+/− and n = 7 for Slc43a2^−/−; comparison of Slc43a2^+/+ and Slc43a2^−/−. Two-way ANOVA and Bonferroni post hoc test: **P < 0.01, ***P < 0.001; the cross represents the death of all Slc43a2^−/− mice. C, survival curves for the same mice as (B).

To obtain more information about the possible pathophysiological mechanisms underlying the early lethality, we measured the concentration of AAs in the plasma of 3-day-old Slc43a2^−/− pups and compared this with age-matched littermates (Fig.6A). We observed a major decrease of certain AAs in Slc43a2^−/− mice compared to Slc43a2^+/+. The highest relative difference was observed for non-essential AAs that are not substrates of Lat4 (e.g. proline with ∼75% reduction followed by histidine, serine and alanine). Also Lat4 substrates appeared to be reduced in Slc43a2^−/− mice [methionine strongly (> 50%) and branched-chain AAs less], although their decrease was not statistically significant. The only Lat4 substrate that did not appear to be decreased was phenylalanine. Thus, in analogy to the previous AA measurements in amniotic fluids, this blood AA pattern does not reflect a specific lack of Lat4 substrates but results probably from the interplay of transport and metabolic pathways that are altered in the Slc43a2^−/− mice. In addition to the decrease in plasma AAs, we detected an increase of plasma acylcarnitine-conjugated fatty acids, in particular of dicarboxylic acids and long chain unsaturated fatty acids, which also highlighted a state of malnutrition (Fig.6C and D). With regard to glucose homeostasis, Slc43a2^−/− mice killed at day 2 showed a high blood concentration variability compared to Slc43a2^+/+ and Slc43a2^+/− mice (Fig.6B). Based on the hypothesis that this variability was the result of the different feeding status of single animals and that Slc43a2^−/− pups might display a defect in glucose disposal (impaired insulin release) and gluconeogenesis, we starved the pups for 30 min before death and blood collection (Fig.6E). Indeed, under these conditions, all Slc43a2^−/− mice displayed a reduction in blood glucose concentration of ∼50% compared to their wild-type littermates. For this and also all other tested parameters, heterozygous Slc43a2^+/− mice showed no difference compared to their Slc43a2^+/+ littermates.

Figure 6.

Impact of Slc43a2 deletion on plasma AAs, acylcarnitines and glucose

Plasma AAs (A), long-chain unsaturated acylcarnitines (C) and dicarboxylacylcarnitines (DAs) (D). Represented are the mean ± SEM: n = 3 for Slc43a2^+/+ (black bars) and n = 6 for Slc43a2^−/− (white bars). Two-way ANOVA and Bonferroni post hoc test: *P < 0.05, **P < 0.001. B and E, blood glucose measurement from mice at day 2 without (B) and with 30 min of starvation (E) prior to sample collection. Represented are the mean ± SEM. Student's t test: **P < 0.01.

Because Slc43a2^−/− pups displayed a failure to grow and signs of malnutrition, we attempted to rescue them. We considered that an early defect in weight gain might represent a competitive disadvantage for obtaining breastfeeding within a large litter. Accordingly, we decreased this potential disadvantage by removing all Slc43a2^+/− littermates on day 3, leaving only an equal number of Slc43a2^+/+ and Slc43a2^−/− pups. However, Slc43a2^−/− mice neither grew, nor survived better, even by reducing the litter size (data not shown). Moreover, we attempted to compensate malnutrition by s.c. administration of an AA-containing solution (Aminoven 10%, Fresenius KABI, 80 μl injection twice a day); however, this intervention resulted in no increase in weight gain or the lifespan of Slc43a2^−/− mice.

Liver morphology and gene expression of Slc43a2 null pups

Because the liver plays a central role in nutrients metabolism, we investigated the possible impact of Slc43a2 defect on its morphology and gene expression pattern. Histological staining indeed showed sites of leucocyte infiltration in its periportal regions (Fig.7B). Liver RNA-sequencing identified 16 078 transcripts as ‘present’ (Tophat software), with 1433 displaying a statistically significant (P < 0.01) differential expression (Fig.7A). From these transcripts, the expression of 144 was changed by more than three-fold (77 increased and 67 decreased). Among the upregulated genes, many are known to be involved in liver regeneration (Dmbt1; Bisgaard et al. 2002), cell proliferation (Htr2b; Soll et al. 2010), tissue remodelling (Mep1a and Mep1b; Sterchi et al. 2008) and detoxification (Gpx6 and Ggt6; Gridley et al. 2011; Heisterkamp et al. 2008). Interestingly, the transcripts encoding acetyl-CoA carboxylase β (Acab), involved in lipogenesis (Postic & Girard, 2008), and glucokinase (Gck), involved in liver gluconeogenesis (Massa et al. 2011), were also upregulated. By contrast, genes involved in fatty acids biosynthesis, such as Fas, Srbp1 and Scd1 were not changed significantly. As expected based on the leucocyte infiltration, Cyp2a5 and Ceacam3 expression was increased (Sipowicz et al. 1997). The downregulation of the transcripts Car3, Dio3 and Inmt, as well as of genes involved in hepatobiliary cholesterol excretion (Abcg5 and Abcg8; Back et al. 2013) and the circadian clock (Dbp; Ando et al. 2011) similarly suggest major metabolic changes, inflammation, toxicity and regeneration at the level of the liver (Hsieh et al. 2009; Dudek et al. 2013) that might explain in part the increase in plasma fatty acids and the selective decrease of plasma AAs.

Figure 7.

Effect of Slc43a2 deletion on liver and kidney

A, heat map summary of differential gene expression pattern in Slc43a2^−/− liver. RNA was isolated from mice at day 2 and each column represents a different mouse. Upregulated genes are indicated in red; downregulated genes are indicated in blue. B, heamatoxylin and eosin staining of liver. Left: liver section of Slc43a2^−/− KO mouse. Right: liver section of Slc43a2^+/+ mouse. *Leucocyte infiltration in the periportal region of Slc43a2^−/− mouse liver. Scale bar = 50 μm. C, gene expression of selected AA transporters in the kidney of Slc43a2^+/+ (black bars) and Slc43a2^−/− (white bars). Represented are the mean ± SEM: n = 3. Two-way ANOVA and Bonferroni post hoc test: **P < 0.01.

At the level of the kidney, which also plays an important role for the systemic metabolism of AAs (Makrides et al. 2014), we tested the expression of AA transporters but did not identify any major alterations in Slc43a2^−/− mice (Fig.7C). Interestingly, other AA transporters of the SLC43 family (i.e. Lat3 and Eeg1) showed no differential expression in Slc43a2^−/− mice; therefore, the lack of Lat4 did not appear to result in any compensatory regulation of AA transporters with similar substrate affinities.

Accumulation of leucine in the proximal part of the small intestine

The major metabolic changes described above suggested that the early death of Slc43a2^−/− mice might be a result of the status of malnutrition. Because we showed that Lat4 was expressed at the basolateral membrane of enterocytes (Fig.2), we hypothesized that this malnutrition might result from a defective intestinal AA absorption. To test this hypothesis, we starved pups for 30 min and then fed them with 5 μl of an AA solution containing 1 mm of each leucine and lysine with corresponding ³H and ¹⁴C tracers. Leucine was chosen as Lat4 substrate because it is not transported by the other basolateral uniporter Tat1, whereas lysine was chosen because it is not transported by Lat4. Upon administration of this AA mixture, we monitored the appearance of the labelled AAs in plasma and their accumulation in different organs after 30 min (Fig.8). Pups lacking Lat4 displayed a marked two-fold increased retention of leucine, but not of lysine, in the first part of the small intestine compared to wild-type littermates (Fig.8A and C). This selective retention of the Lat4 substrate leucine in the small intestine of Slc43a2^−/− pups suggested its intracellular retention in the absence of basolateral Lat4. This selective retention was not associated with a visible morphological difference as observed on haematoxylin and eosin stained small intestine sections. It also did not lead to differences in terms of AA appearance in plasma or levels in other organs after 30 min (Fig.8B and D).

Figure 8.

Distribution of radiolabelled l-leucine and l-lysine in different organs after oral administration

Mice were starved 30 min prior to oral administration of radiolabelled AA solution. Leucine (A) and (C) lysine accumulation in different organs normalized by weight. Leucine (B) and (D) lysine concentration in plasma 30 min after oral administration (Slc43a2^+/+: black bars, Slc43a2^−/−: white bars). The small intestine of the pups was divided into three equals parts from proximal to distal: SI1, SI2 and SI3. Represented are the mean ± SEM pooled from three independent experiments: n = 6 for Slc43a2^+/+ mice and n = 8 for Slc43a2^−/− mice. Two-way ANOVA and Bonferroni post hoc test: ***P < 0.001.

Discussion

Lat4 is considered a ‘system L’ transporter because it transports large neutral AAs independent of Na⁺. This functional denomination, however, does not include other important transport characteristics such as transport type and affinity that clearly differentiate Lat4 from the classical system L transporters Lat1 and Lat2 (Lat1–4F2hc/Slc7a5-Slc3a2 and Lat2–4F2hc/Slc7a8-Slc3a2). Indeed, as originally suggested by Bodoy et al. (Bodoy et al. 2005), Lat4 functions as a uniporter (facilitated diffusion pathway) and we show here that it does this with a low, symmetrical apparent affinity (Fig.3). These functional characteristics contrast with those of the classical system L transporters Lat1 and Lat2 that function as asymmetric anti-porters with high apparent affinity for the uptake of their substrates (Meier et al. 2002). Additionally, the selectivity of Lat4 is more limited and includes only essential AAs, specifically the branched-chain ones plus phenylalanine and methionine. Because the in vivo role of this AA transporter has yet not been investigated, we further analysed its localization and tested its physiological function in the present study using a newly generated general KO mouse model.

The main observation reported here is that mice defective in Lat4 suffer from a small intrauterine growth restriction, a major postnatal growth defect and premature death within a few days after birth, despite normal feeding behaviour suggested by the ‘milk spot’. The question is what functional defect is provoked by the lack of Lat4, which leads to this developmental problem? The phenotype does not appear to depend on the genetic background because it was not modified by backcrossing for six generations with C57BL/6 or S129 mice. Despite the fact that attempts to rescue the pups by litter size reduction (decreased competition) or parenteral nutritional supplementation did not succeed, this lethal phenotype is probably largely the result of a combination of defective placental and, in particular, intestinal AA transport that leads to severe postnatal malnutrition.

Lat4 is a basolateral uniporter for essential neutral AAs in the small intestine and kidney tubule

A first crucial aspect towards understanding the function of Lat4 is its tissue and subcellular localization. In the original study by Bodoy et al. (2005), which reported its identification and mRNA expression in human and mouse tissues, Lat4 mRNA was shown to be present at high levels in placenta, kidney and small intestine, as well as at lower levels in several other tissues and in a proximal tubule cell line. Additionally, expression data from available microarray analysis (www.biogps.com) indicate a strong expression in macrophages, microglia and osteoclasts. As previously hypothesized, we now show, using a newly generated antibody, that Lat4 localizes to the basolateral membrane of AA transporting cells of small intestine and kidney proximal tubule. Furthermore, we confirm the selectivity of the staining at these locations using the KO mouse model. These images show that Lat4 is expressed in the small intestine at a much higher level in the villi than in the crypts, similar to other AA transporters involved in transepithelial transport (Fig.2A and B) (Dave et al. 2004). These data strongly support the hypothesis that Lat4 plays an important role for the (re)absorption of AAs. Interestingly, in the kidney, in addition to a strong labelling in the proximal tubule (Fig.2C–E), there is also a strong basolateral labelling in the thick ascending limb of the loop of Henle and a weaker signal in the distal convoluted tubule (Fig.2F–H). The low level or absence of Lat4 protein expression in liver and skeletal muscle suggested by previous mRNA data is also confirmed by our immunofluorescence results (Fig.1E–H).

As demonstrated in the present study with expression experiments in X. laevis oocytes, Lat4 is a symmetrical uniporter with a low affinity for its substrates (Fig.3). Thus, at physiological substrate concentrations, its directional transport activity is proportional to the transmembrane substrate concentration difference. This allows Lat4, if expressed at a sufficiently high level, to mediate the equilibration of its AA substrates (branched-chain AAs, phenylalanine and methionine) between the extracellular space and the cytosol. It can thus mediate a directional transport of its substrates across the plasma membrane; for example, their basolateral efflux from small intestine and kidney proximal tubule cells after their luminal uptake from the intestinal lumen or primary urine. We postulate that this directional basolateral efflux transport also represents a recycling pathway for AAs that are taken up across the same basolateral membranes with high affinity by the parallel localized heterodimeric anti-porters Slc7a8-Slc3a2/Lat2–4F2hc and Slc7a7-Slc3a2/y⁺Lat1–4F2hc. Such a high affinity uptake of essential AAs by these anti-porters allows them to efflux in exchange non-essential AAs present mostly at much a higher concentration than the essential AAs in these epithelial cells. Thus, Lat4 may exert a quantitative control of the Lat2 and y⁺Lat1 anti-porter efflux activity by recycling the essential AAs that they take up in exchange to the outside. Correspondingly, the expectation is that the lack of Lat4 could affect the transepithelial transport of other, in particular non-essential and cationic AAs. Such a functional relationship/cooperation with anti-porters for the efflux of non-essential AAs has previously been demonstrated in the X. laevis oocyte expression system and in vivo for the aromatic AA transporter Tat1, another essential AA uniporter expressed at the basolateral membrane of transporting epithelial cells (Ramadan et al. 2007; Mariotta et al. 2012). However, in the present case, functional experiments aiming to test the impact of Lat4 defect on the transport of AAs were strongly limited by the fact that Lat4 mice did not develop normally and died a few days after birth.

Concerning Lat4 localization, an intriguing observation was that Lat4 is expressed at a high level also in the kidney thick ascending limb, as previously reported at the mRNA level (Cheval et al. 2011). We postulated that the basolateral expression of Lat4 in these cells plays a role with respect to energy metabolism and/or for the synthesis of the highly abundant secreted surface protein uromodulin (Glaudemans et al. 2014) rather than for the transepithelial transport of AAs. These cells might thus import branched-chain AAs and methionine via Lat4 for use in protein synthesis and as catabolic substrates for the production of ATP or to drive the influx of non-essential AAs, such as glutamine, which then function as metabolic fuel (opposite net transport direction compared to transepithelial transport).

Prenatal and postnatal failure to grow, malnutrition and death of Slc43a2^−/− KO mice

The global disruption of the AA uniporter Lat4 highlighted the pivotal importance of this AA transporter in prenatal and postnatal development. Slc43a2^−/− KO mice displayed defective growth before birth with an ∼10% reduction of weight at E18 compared to wild-type littermates. The reported defective growth showed phenotypical analogies with other KO mouse models of AA transporters, such as y⁺Lat1 and CAT-1 (Xie et al. 2005; Sperandeo et al. 2007), suggesting that Lat4 should be also considered together with the aforementioned AA transporters as a very important player in placental AA transport. Interestingly, Lat4 dysfunction resulted not only in a reduction of its AA substrates in the amniotic fluid, but also in a general decrease of all AAs, except for aspartate and glutamate. Thus, defective Lat4 function causes pleiotropic consequences that are not limited to a lack of Lat4-specific substrates but encompass many other AAs. The reduced weight and reduced concentration of AAs in the amniotic fluid neither resulted in prenatal lethality, nor disrupted the postnatal feeding behaviour as suggested by the presence of a normal ‘milk spot’. However, as shown by the postnatal growth curve and the concentration of different metabolites in plasma of 3-day-old mice, it appeared they display a defect in enteric AA absorption that could be the origin of the postnatal growth defect and also contribute to a major extent to the premature lethality of Slc43a2^−/− mice.

The metabolic status of the Slc43a2^−/− pups already 3 days after birth showed many analogies with starvation. Initially, and unexpectedly, methionine was the only Lat4 substrate to be highly decreased in their plasma, whereas branched-chain AAs and phenylalanine showed only a minor reduction. By contrast, many non-essential AAs were markedly decreased, mostly alanine and proline. However, based on the suggested function of Lat4 as a ‘gate keeper’ for the epithelial basolateral efflux of non-essential AAs via anti-porters (see above), it is not unexpected that a Slc43a2^−/−-induced defect in AA absorption may lead to a catabolic state in which the decrease in non-essential AAs is more marked than that of essential ones (Fig.9).

Figure 9.

Schematic representation of apical and basolateral AA transporter cooperation in proximal tubule kidney cell

Neutral AAs are imported apically mostly via the symporter B⁰AT1 (SLC6A19) and are effluxed across the basolateral membrane mostly by the anti-porter LAT2–4F2hc and the uniporters TAT1 and LAT4. We postulate that essential neutral AAs are equilibrated between the intracellular compartment and the extracellular space via the two low affinity uniporters TAT1 and LAT4 and may recycle into the cell via anti-porters and thereby drive the efflux of other AAs, in particular of (non)-essential neutral AAs via LAT2–4F2hc and cationic AAs via y⁺LAT1–4F2hc. Therefore, a defect in the uniporter(s) LAT4 and/or TAT1 is suggested to impact not only on the basolateral efflux of their own AA substrates, but also on that of the other substrates of the anti-porters LAT2–4F2hc and y⁺LAT1–4F2hc.

Non-essential AAs are also specifically decreased in fasted children (Wolfsdorf et al. 1982) and upon prolonged endurance exercise, particularly alanine and proline that are central for AA metabolism and gluconeogenesis (Ji et al. 1987; Huq et al. 1993). Also, the decrease in tyrosine, histidine and serine can be explained by their catabolic usage to provide substrates for the Krebs cycle. The increase in plasma acylcarnitine-conjugated dicarboxylic and long-chain unsaturated fatty acids in Slc43a2^−/− mice is an additional metabolic fingerprint indicating that the Slc43a2^−/− pups are in a state of starvation (Costa et al. 1999; Hunt & Alexson, 2002). Additionally, histological examination and the gene expression analysis of the liver, by indicating processes of inflammation, regeneration and alteration of glucose and fatty acids metabolism, further demonstrated that Slc43a2 null pups suffered from malnutrition. Indeed, an alteration in glucose metabolism, which is a hallmark of nutritional defects in children (Bandsma et al. 2010; Spoelstra et al. 2012), was also present in Slc43a2^−/− mice, as shown by the marked decrease in blood glucose concentration measured upon 30 min of starvation.

The hypothesis that Lat4 impacts on AA absorption was further supported by our observation that leucine administered orally was retained in the proximal part of the small intestine of Slc43a2^−/− mice, unlike lysine, which is not a Lat4 substrate. This suggests that luminally absorbed leucine accumulated in small intestine enterocytes from which it could not be efficiently released basolaterally. However, no concomitant alteration in plasma appearance or organ accumulation of leucine was observed in these mice, such that a possible major defect in AA absorption from ingested food could not be verified. It is, however, possible that the non-physiological administration of two AAs at a high concentration (1 mm each) to a large extent bypassed a defect preventing the sufficient absorption from mother's milk. An additional untested possibility is that the AAs permanently secreted by the liver and the pancreas into the intestine lumen, as well as the AAs resulting from enterocyte shedding, were not efficiently reabsorbed, leading finally to a substantial AAs loss into the faeces.

An interesting observation is that, out of five KO mouse models lacking AA transporters expressed in the small intestine and kidney proximal tubule, the KO of Lat4 is only the second one that leads to a lethal phenotype. Neither the lack of luminal B⁰AT1 transporter, nor the lack of the basolateral l-type anti-porter Lat2 (Slc7a8) and of the aromatic AA uniporter Tat1 (Slc16a10) prevents normal life and reproduction under laboratory conditions. The KO of Lat4 is the first AA transporter defect that, despite a normal Mendelian birth ratio, leads to a severe postnatal malnutrition with early death. Many of the observations reported in the present study suggest that the primary cause of Slc43a2^−/− pups malnutrition may be a defect in AA absorption from the gut. However, it was not formally excluded that the observed malnutrition syndrome was partly the result of an insufficient suckling drive or another alteration preventing the appropriate use of mother's milk as a nutrition source.

In addition, we considered the possibility that Lat4 deficiency, because of its high expression in macrophages and microglia (www.biogps.org), could impact on the survival of the pups by affecting the immune system. However, transgenic mice with defective hematopoietic mononuclear phagocytic cells, such as Csfr1^−/− KO (Dai et al. 2002) or op/op mice (Wiktor-Jedrzejczak et al. 1990), have been shown to survive beyond weaning, clearly indicating that the lethality of Lat4 KO mice is not the result of an immune defect.

In conclusion, we provide the first characterization of a total Slc43a2 KO mouse model. The small intrauterine growth retardation and the postnatal malnutrition phenotype leading to an early death indicate that Lat4 is an essential AA transporter required for the proper placental AA transport and postnatal nutrient absorption.

Glossary

Abbreviations

AA

amino acid

DA

dicarboxylacylcarnitine

E

embryonic day

KO

knockout

NCC

sodium-chloride cotransporter

NKCC2

sodium-potassium-chloride cotransporter 2

RT

room temperature

Additional information

Competing interests

All authors declare that there are no competing financial, personal or professional interests that could be construed to have influenced this paper.

Author contributions

A.G., L.M. and F.V contributed to the conception and design of the experiment. A.G., L.M., L.B., B.H., R.F., S.M.R.C. and F.V. were involved in the collection, analysis and interpretation of data. A.G. and F.V. drafted and revised the article. All authors approved the final version submitted for publication.

Funding

The laboratory of FV is supported by Swiss NSF grant 31-130471/1 and the National Centre of Competence in Research (NCCR) Kidney.CH.