Molecular and Functional Identification of Sodium Ion-dependent, High Affinity Human Carnitine Transporter OCTN2*

Primary carnitine deficiency, because of a defect of the tissue plasma membrane carnitine transporters, causes critical symptoms. However, the transporter has not been molecularly identified. In this study, we screened a human kidney cDNA library and assembled a cDNA-encoding OCTN2 as a homologue of the organic cation transporter OCTN1, and then we examined the function of OCTN2 as a carnitine transporter. OCTN2-cDNA encodes a polypeptide of 557 amino acids with 75.8% similarity to OCTN1. Northern blot analysis showed that OCTN2 is strongly expressed in kidney, skeletal muscle, heart, and placenta in adult humans. When OCTN2 was expressed in HEK293 cells, uptake ofl-[3H]carnitine was strongly enhanced in a sodium-dependent manner with K m value of 4.34 μm, whereas typical substrates for previously known organic cation transporters, tetraethylammonium and guanidine, were not good substitutes. OCTN2-mediatedl-[3H]carnitine transport was inhibited by the d-isomer, acetyl-d,l-carnitine, and γ-butyrobetaine with high affinity and by glycinebetaine with lower affinity, whereas choline, β-hydroxybutyric acid, γ-aminobutyric acid, lysine, and taurine were not inhibitory. Because the observed tissue distribution of OCTN2 is consistent with the reported distribution of carnitine transport activity and the functional characteristics of OCTN2 coincide with those reported for plasma membrane carnitine transport, we conclude that OCTN2 is a physiologically important, high affinity sodium-carnitine cotransporter in humans.

Carnitine (3-hydroxy-4-N-trimethylaminobutyric acid) is a small, water soluble molecule that has important physiological roles, including involvement in the ␤-oxidation of fatty acids by facilitating the transport of long chain fatty acids across the mitochondrial inner membrane as their acylcarnitine esters and modulation of intracellular CoA homeostasis (1,2). Carnitine deficiency causes severe pathological symptoms such as cardiomyopathy and muscle weakness (3)(4)(5)(6). Primary carnitine deficiency is thought to be because of a defect of active trans-port of carnitine across plasma membranes, whereas secondary carnitine deficiency seems to be associated with an enzymatic defect, resulting in impaired oxidation of acyl-CoA intermediates in the mitochondria (1)(2)(3). Reduced carnitine concentration in tissue and plasma may be caused by insufficient carnitine uptake activity from plasma and impaired reabsorption in the kidney, respectively (3,(5)(6)(7)(8). Symptoms related to defective carnitine transport have been studied in a carnitine-deficient mutant animal model, juvenile visceral steatosis mouse, which shows several symptoms of primary and/or secondary carnitine deficiency and lacks in high affinity transport activity in several tissues (7, 9 -12).
Although many membrane-physiological studies of carnitine transport mechanisms have been reported (3,(5)(6)(7)(12)(13)(14)(15)(16), it is essential for a clearer understanding of the primary carnitine deficiency to identify the relevant transporter and to functionally characterize the carnitine transport in detail. We have recently cloned and characterized a novel organic cation transporter OCTN1 from human fetal liver (17). OCTN1 was expressed strongly in adult tissues such as kidney, trachea, and bone marrow and weakly in other tissues. When expressed in HEK293 cells, OCTN1 caused significant transport of tetraethylammonium (TEA), 1 a typical organic cation, in a pH-dependent manner. These characteristics strongly suggested that OCTN1 acts as a proton/organic cation antiporter at the renal epithelial apical membrane (17,18). In the present study, we identified a new transporter molecule, OCTN2, with high homology to OCTN1. We cloned full-length cDNA for this putative member of the organic cation transporter family, expressed it in HEK293 cells, and showed that OCTN2 has the characteristics of a high affinity sodium/carnitine cotransporter. Cloning of OCTN2 cDNA and Northern Blot Analysis-A data base search for matches to the cDNA sequence of the OCTN1 gene revealed several genomic cosmid clones (GenBank TM accession numbers L43407, L43408, L46907, L81773, and L43409), derived from human chromosome 5q, that contain sequences highly homologous to OCTN1. Because these genomic sequences do not cover the entire open reading frame for this new gene, which we designated OCTN2 on the basis of its high similarity to OCTN1, we initiated cDNA cloning. From the genomic * This work was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan and by CREST Core Research for Evolutional Science and Technology of the Japan Science and Technology Corporation (JST). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Materials-L-[Methyl-
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB015050.
sequence, two primers (631RT S4 5Ј-GTGCTGTTGGGCTCCTTCATT-TCA-3Ј and 631RT A1 5Ј-AGCTGCATGAAGAGAAGGACACTG-3Ј) were prepared and used in reverse transcription-polymerase chain reaction of human kidney-derived cDNA. This afforded a 900-base pair 32 cDNA fragment of OCTN2. Screening of a human kidney cDNA library with this fragment as the probe yielded overlapping, longer clones that provided additional sequences. A primer (631R S6 5Ј-AGCATCCT-GTCTCCCTACTTCGTT-3Ј) designed from the new sequence was used to amplify the 3Ј portion of OCTN2 by 3Ј rapid amplification of cDNA ends using human kidney Marathon-Ready TM cDNA (CLONTECH). Finally, the full coding sequence of OCTN2 was obtained by assembling these sequences. To assess OCTN2 expression in human tissues, an OCTN2 cDNA fragment, amplified with the 631RT S4 and 631RT A1 primers, was labeled with [␣-32 P]dCTP and subjected to Northern blotting with poly(A) ϩ RNA from a wide range of normal human tissues and cancer cell lines (CLONTECH). Hybridization was carried out in Ex-pressHyb hybridization solution (CLONTECH) at 68°C for 1 h. Membranes were then washed in 2ϫ SSC containing 0.1% SDS at room temperature for 60 min and finally in 0.1ϫ SSC, 0.1% SDS at 50°C for 20 min twice. Transport Study in HEK293 Cells-The full-length OCTN2 cDNA was subcloned into the BamHI sites of the expression vector pcDNA3, and the construct, pcDNA3/OCTN2 was used to transfect HEK293 cells by means of the calcium phosphate precipitation method as described previously (17). The cells were cultivated in Dulbecco's modified Eagle's medium containing 10% fetal calf serum (Life Technologies, Inc., Tokyo, Japan), 100 units/ml penicillin, and 100 g/ml streptomycin in tissue culture dishes in a humidified incubator at 37°C under 5% CO 2 for 24 h and then transfected with pcDNA3 plasmid carrying full-length OCTN2 cDNA or with the pcDNA3 plasmid vector alone. At 48 h after transfection, the cells were harvested by scraping with a rubber policemen and suspended in the medium for transport study, which consisted of 125 mM NaCl, 4.8 mM KCl, 5.6 mM D-glucose, 1.2 mM CaCl 2 , 1.2 mM KH 2 PO 4 , 1.2 mM MgSO 4 , and 25 mM HEPES (pH 7.4). The cell suspension and transport medium containing a radiolabeled test compound were preincubated separately for 20 min and then mixed to initiate uptake. At appropriate times, 200-l aliquots of the mixture were withdrawn, and the cells were separated from the transport medium by centrifugation in a microtube containing a silicon oil and liquid paraffin mixture with a density of 1.03. The resultant cell pellets were solubilized in 3 N KOH and then neutralized with HCl, and the associated radioactivity was quantitated in a liquid scintillation counter (Aloka, Tokyo, Japan). Cellular protein content was determined according to the method of Bradford (19) using a Bio-Rad protein assay kit. When sodium ions were replaced with other cations, the obtained cells were suspended in sodium-free medium. The composition of sodium-free medium was the same as that of the above transport medium except that the sodium chloride was replaced isotonically with potassium chloride, choline chloride, N-methylglucamine chloride, or lithium chloride.
Usually initial uptake rates were obtained by measuring the uptake at 3 min. To estimate kinetic parameters for saturable transport, the uptake rate (v) was fitted to the following equation by means of nonlinear least squares regression analysis using WinNonlin (Scientific Consulting Inc., Cary, NC). v ϭ V max ϫ s/(K m ϩ s), where v and s are the uptake rate and concentration of carnitine, respectively, and K m and V max are the half-saturation concentration (Michaelis constant) and maximum transport rate, respectively. All data were expressed as the means Ϯ S.E., and statistical analysis was performed by use of Student's t test. The criterion of significance was taken to be p Ͻ 0.05.

Amino Acid Sequence and Tissue Distribution of Human
OCTN2-The full-length OCTN2 cDNA appeared to encode a polypeptide of 557 amino acids and have 75.8% similarity with human OCTN1 (17) (Fig. 1A). Human OCTN2 is predicted to have twelve putative membrane-spanning domains by hydropathy analysis according to TopPred 2 (20) as well as three N-glycosylation sites and six protein kinase C phosphorylation sites. The presence of twelve membrane spanning domains agrees with that of previously known membrane transporters (21). Like human OCTN1, OCTN2 has a unique sugar transport protein signature (17). Comparison of the amino acid sequence with those of other organic ion transporters revealed that human OCTN2 has similarity with rat OCT1 (32.5%) (22), rat OCT2 (33.6%) (23), human OCT1 (33.1%) (24), human OCT2 (33.1%) (25), and rat OAT1 (28.4%) (26). No significant similarity was observed with human oatp (27) or monoamine neurotransmitter transporters such as serotonin transporter (28) or monoamine transporter (29). These data indicate that OCTN2 may be a member of the organic cation transporter family. Fig. 2 shows the tissue distribution of human OCTN2 examined by Northern blotting analysis. In fetal tissues, OCTN2 was expressed strongly in kidney and weakly in liver, lung, and brain. In adult, it was expressed strongly in kidney, skeletal muscle, placenta, heart, prostate, and thyroid and weakly in pancreas, liver, lung, brain, small intestine, uterus, thymus, adrenal gland, trachea, spinal cord, and several other tissues. Interestingly, specific bands corresponding to human OCTN2 were also detected in human tumor-derived cell lines, including melanoma G361, lung carcinoma A549, colorectal carcinoma SW480, chronic myelogenous leukemia K562, and carcinoma of cervix HeLa S3. The distribution in tumor cell lines is very similar to that of human OCTN1. On the other hand, the distributions of OCTN2 in normal fetal and adult tissues are different from those of OCTN1, especially as regards the significant expression of OCTN2 in fetal kidney and in skeletal muscle, liver, placenta, heart, prostate, and thyroid in adults, although OCTN1 and OCTN2 both exhibit broad tissue distribution.
Functional Analysis of OCTN2 Expressed in HEK293 Cells-Because human OCTN1 transported the organic cation TEA in a pH-dependent manner when expressed in HEK293 cells in our previous study (17), we expressed human OCTN2 in the same cells and measured the uptake of cationic compounds for comparison with that of OCTN1. Although TEA is a good substrate of OCTN1 (17) OCTN2 (B). A, conserved amino acid residues in OCTN1 and OCTN2 are highlighted. Potential N-glycosylation sites, protein kinase C phosphorylation sites, and the sugar transporter protein signature sequence motif of OCTN2 are indicated by *, #, and ϩ, respectively. B, the twelve putative membrane-spanning domains are numbered on the hydropathy plot. and 263 Ϯ 12 in nontransfected cells at the TEA concentration of 60 M and at 3 min). Because guanidine was suggested to be transported by a different transporter than that for TEA across the apical membrane of renal tubular epithelial cells (30), we examined the uptake of [ 14 C]guanidine in human OCTN1-or OCTN2-expressing cells. However, neither OCTN1 (106 Ϯ 5.6 pmol/mg protein/3 min at 10 M guanidine) nor OCTN2 (95.1 Ϯ 11.8 pmol/mg protein/3 min at 10 M guanidine) exhibited increased guanidine transport activity compared with nontransfected cells (98.4 Ϯ 5.9 pmol/mg protein/3 min at 10 M guanidine). Zwitterionic carnitine is reabsorbed in the kidney via an active transport mechanism, although the molecular identity of the transporter remains to be established. Accordingly, we examined the transport of carnitine in the present study. Although OCTN1 showed slight but significant uptake of L-[ 3 H]carnitine (2.5-fold increased uptake compared with nontransfected cells), a very large uptake of L-[ 3 H]carnitine was seen with the human OCTN2-expressing HEK293 cells, as described below. Fig. 3A shows the time course of the uptake of L-[ 3 H]carnitine by HEK293 cells transfected with OCTN2 or with the expression vector pcDNA3 alone, in the presence or absence of sodium ions. Uptake of L-[ 3 H]carnitine was significantly increased by OCTN2 transfection both in the presence and absence of sodium ions in the transport medium. The uptake of L-[ 3 H]carnitine was particularly high in OCTN2-transfected cells in the presence of sodium ions, and it appears that OCTN2 is a sodium ion-dependent carnitine transporter. Because the cells transfected with expression vector alone showed a slight but significant increase of L-[ 3 H]carnitine uptake in the presence of sodium ions compared with that in the absence of sodium ions, HEK293 cells themselves seem to have a weak activity of sodium ion-dependent carnitine transport. This is not surprising considering that HEK293 cells were originally derived from human embryonic kidney. Because sodium iondependent and OCTN2-mediated uptake of carnitine increased linearly up to 5 min, initial uptake rate of carnitine was determined at 3 min to characterize the transporter in all subsequent studies.
To confirm the driving force for OCTN2-mediated carnitine transport, the effect of sodium ion replacement with various cations on carnitine transport was examined. When sodium was replaced with N-methylglucamine, potassium, or choline, initial uptake of L-[ 3 H]carnitine was almost abolished, whereas substitution with lithium resulted in retention of a weak uptake activity (Fig. 3B). When the temperature of the transport assay was decreased to 4°C, the initial uptake was decreased to 3.15 Ϯ 0.50% of that at 37°C, demonstrating a significant temperature dependence of transport with an estimated activation energy of 17.9 kcal/mol.
The concentration dependence of L-carnitine transport was examined to estimate the half-saturation concentration of OCTN2-mediated carnitine transport (Fig. 4A). OCTN2-mediated transport was saturable, and an Eadie-Hofstee plot showed a single straight line (Fig. 4B), demonstrating the presence of a single functional site on the OCTN2 protein. The half-saturation concentration and maximum transport activity estimated by nonlinear least squares regression analysis using WinNonlin (SCI, Cary, NC) were 4.34 Ϯ 0.602 M and 2.94 Ϯ 0.12 nmol/mg protein/3 min, respectively.
Specificity of OCTN2-mediated carnitine transport was examined in terms of the inhibitory effect on the initial uptake of L-[ 3 H]carnitine. As is clearly shown in Fig. 4C, structurally analogous compounds, D-carnitine, acetyl-D,L-carnitine, and ␥-butyrobetaine reduced L-[ 3 H]carnitine uptake at 5 and 50 M. Because L-carnitine demonstrated more potent inhibition than the D-isomer at low concentration (5 M) OCTN2 seems to have stereospecificity. Glycinebetaine less effectively reduced the uptake of L-[ 3 H]carnitine, and others examined, including ␥-aminobutyric acid, choline, ␤-hydroxybutyrate, lysine, and taurine, were not inhibitory. Accordingly, the structural requirement of OCTN2-mediated carnitine transport is rather strict.

DISCUSSION
Carnitine is normally maintained at a steady level in the blood, suggesting its physiological importance (1, 2). Although carnitine is biosynthesized in liver and brain (31), a significant amount of carnitine is also obtained from the diet via carriermediated transport across the intestinal epithelial cell membranes (32,33) and is retained in the body through reabsorption in the kidney via active transport across the renal tubular epithelial cell membrane (1,7,14,15). Furthermore, many studies have demonstrated that the tissues that extensively accumulate carnitine, such as skeletal muscle, heart, liver, and epididymis, take up and/or release carnitine via specialized transport mechanisms to maintain the steady-state tissue concentration (1). Although many of these membrane physiological studies suggested the participation of multiple sodium ion-dependent transporters with high and low affinities in carnitine movement across tissue plasma membranes (1,13,15), no such transporter has yet been identified at the molecular level. Accordingly, to achieve a better understanding of the biological and physiological roles of carnitine, as well as carnitine-related pathological states, it is essential to identify the carnitine transporter(s). We have previously cloned the human organic cation transporter OCTN1, which may participate at least partially in proton/organic cation antiport at the renal apical membranes, and characterized it by measuring the uptake of the typical organic cation TEA by OCTN1-transfected HEK293 cells (17). In the present study, we succeeded in obtaining cDNA of a second member of the human OCTN family, OCTN2, which has a high similarity to OCTN1, and found that OCTN2 has many of the characteristics of a high affinity, sodium iondependent carnitine transporter.
The idea that human OCTN2 is a sodium ion-dependent carnitine transporter is supported by the specific tissue distribution and the result of functional expression in HEK293 cells. Most adult tissues that highly express OCTN2, including skeletal muscle, kidney, placenta, and heart, have been reported to take up carnitine via a sodium ion-dependent, carrier-mediated transport mechanism (1, 15, 16, 34 -36). As regards fetal tissues, we previously found that mouse embryo fibroblasts take up carnitine in a sodium ion-dependent manner with a halfsaturation concentration of 5.5 M, although the distribution of carnitine transport activity in fetal tissues was not established (13). Tissues that have apparently low affinity carnitine transporters, such as liver, brain, and intestine, with apparent halfsaturation concentrations between 0.2 mM and 10 mM (1,33,37,38), showed low expression of OCTN2. The distribution of OCTN2 is different from that of OCTN1, which has 75.8% sequence similarity with OCTN2 and coincides well with the functional distribution of sodium ion-dependent, high affinity carnitine transport activity as discussed above. Furthermore, OCTN2 hardly transported TEA, a good substrate of OCTN1, or guanidine, a substrate of the second renal organic cation transporter, which is distinct from that for TEA (30), OCTN2 seems likely to have some physiological role other than the renal excretion of organic cations.
When OCTN2 was expressed in HEK293 cells, a high uptake of L-[ 3 H]carnitine was observed in the presence of sodium ions (Fig. 3, A and B). Because the cellular volume of HEK293 cells obtained from the accumulation of 3 H 2 O is 6.3 l/mg protein (17), L-[ 3 H]carnitine apparently accumulated to the extent of about 320-fold at the steady-state in the cells by utilizing an inside-directed sodium ion gradient as the driving force. Lithium ions partially retained uphill transport (Fig. 3B) in a manner comparable with the carnitine transport obtained using rat renal brush border membrane vesicles (15) and human placental choriocarcinoma cells (16), which suggests that lithium ions are partially accepted as the cation for cotransport with carnitine. When sodium was replaced with choline, carnitine uptake was specifically eliminated (Fig. 3B). Although 500 FIG. 4. Concentration dependence (A and B) and substrate specificity (C) of carnitine transport by OCTN2 expressed in HEK293 cells. A, uptake of carnitine at various concentrations was measured at pH 7.4 and 37°C for 3 min. B, Eadie-Hofstee plot of the saturable component. Dotted and solid lines represent the apparently nonsaturable uptake obtained from cells transfected with the pcDNA3 plasmid vector alone (triangles) and the saturable uptake obtained from nonlinear least squares regression analysis of the OCTN2-mediated uptake after subtracting the apparently nonmediated uptake (shown by triangles) from the total uptake (open circles), respectively. C, inhibition by test compounds at various concentrations (in M shown in Fig.) was measured in the same way at 3 min at the L-[ 3 H]carnitine concentration (10 nM). All the results are shown as the means and S.E. of three determinations, and * indicates a significant difference from the control studied in the absence of any inhibitor. When S.E. was smaller than the symbols, it was not shown. GABA, ␥-aminobutyric acid.
M choline was not inhibitory (Fig. 4B), choline may have a low affinity to compete with carnitine binding to OCTN2.
The half-saturation concentration of L-carnitine uptake by OCTN2 was estimated to be 4.34 M, which is very similar to the values observed for high affinity carnitine transport in membrane physiological studies in kidney, skeletal muscle, heart, placenta, and cultured fibroblasts (1, 2, 5, 6, 15, 16, 34 -36), tissues that exhibited high expression of OCTN2 in Northern blot analysis (Fig. 2). Carnitine transport in these tissues was reported to be significantly inhibited by the Disomer of carnitine, acetylcarnitine, and ␥-butyrobetaine in a stereospecific manner (6,15,16,35,36). Furthermore, glycinebetaine and choline were low affinity inhibitors and ␥-aminobutyric acid was not inhibitory (14 -16, 36). These previously reported substrate specificity characteristics exactly coincide with the properties of OCTN2. We conclude that OCTN2 is a high affinity, sodium ion-dependent carnitine transporter expressed in several tissues, including kidney, skeletal muscle, heart, and placenta.
The amino acid sequence of human OCTN2 is very similar to that of human OCTN1, although OCTN1 exhibited only a low carnitine transport activity and had the functional characteristics of a proton/organic cation antiporter (17). Furthermore, TEA, a good substrate for OCTN1, was not transported well by OCTN2. Such significant differences of substrate specificity and driving force for transport despite the similarity in amino acid sequence may suggest that the binding or recognition specificity for substrates and cotransported ions on the transporter proteins is determined in a very limited region and that OCTN proteins have both common structures as membrane transporters and distinct small regions for the recognition of the substrate and cotransported ions. It will be interesting to identify the essential amino acid sequence of the functional binding sites by constructing chimeric proteins of OCTN1 and OCTN2. The strong expression of OCTN2 in human-derived tumor cells is similar to that of OCTN1 (14). This may reflect up-regulation of expression of the gene in malignancy to meet a higher requirement for L-carnitine.
In conclusion, human OCTN2 was cloned as a new member of the family of organic cation transporters. Studies of its tissue distribution and its functional expression in HEK293 cells indicated that OCTN2 is a physiologically important, high affinity carnitine transporter that shows significant sodium ion dependence. The identification of this carnitine transporter should contribute to a better understanding of the physiological and biochemical functions of carnitine, as well as to the development of measures to treat primary carnitine deficiency. In addition, isolation of the counterpart of OCTN2 in mice and comparison with juvenile visceral steatosis mice, which show abnormal carnitine metabolism and related diseases, may help to identify the principal causes of carnitine deficiency syndromes (7, 9 -12).