Molecular and functional characterization of organic cation/carnitine transporter family in mice.

Carnitine is essential for beta-oxidation of fatty acids, and a defect of cell membrane transport of carnitine leads to fatal systemic carnitine deficiency. We have already shown that a defect of the organic cation/carnitine transporter OCTN2 is a primary cause of systemic carnitine deficiency. In the present study, we further isolated and characterized new members of the OCTN family, OCTN1 and -3, in mice. All three members were expressed commonly in kidney, and OCTN1 and -2 were also expressed in various tissues, whereas OCTN3 was characterized by predominant expression in testis. When their cDNAs were transfected into HEK293 cells, the cells exhibited transport activity for carnitine and/or the organic cation tetraethylammonium (TEA). Carnitine transport by OCTN1 and OCTN2 was Na(+)-dependent, whereas that by OCTN3 was Na(+)-independent. TEA was transported by OCTN1 and OCTN2 but not by OCTN3. The relative uptake activity ratios of carnitine to TEA were 1.78, 11.3, and 746 for OCTN1, -2, and -3, respectively, suggesting high specificity of OCTN3 for carnitine and significantly lower carnitine transport activity of OCTN1. Thus, OCTN3 is unique in its limited tissue distribution and Na(+)-independent carnitine transport, whereas OCTN1 efficiently transported TEA with minimal expression of carnitine transport activity and may have a different role from other members of the OCTN family.

carnitine transport system that reabsorbs more than 95% of carnitine from the renal glomerular filtrate (2, 4 -6). In 1988, mutant mice that showed a systemic carnitine deficiency (SCD) 1 phenotype were found by Koizumi et al. (7) and were named juvenile visceral steatosis (jvs) mice. jvs mice exhibit fatty liver, hyperammonemia, and hypoglycemia. These symptoms are the phenotype of systemic carnitine deficiency in man (3,8), and it was considered that a defect of the active transport system for carnitine may be related to SCD in human and mouse (3,7). Wu et al. (11) and Tamai et al. (9,10) have isolated and characterized OCTN transporters OCTN1 and OCTN2 in human (9 -11). The latter was shown to be a physiologically essential high-affinity Na ϩ -dependent carnitine transporter by the demonstration that mutations found in SCD patients and jvs mice cause functional alteration of OCTN2 (9,12). Other similar mutations in OCTN2 related to SCD have been reported (13)(14)(15)(16)(17)(18)(19)(20). Interestingly, OCTN1 and -2 in human transported organic cations such as tetraethylammonium (TEA) as well as acylcarnitines (11,(21)(22)(23), suggesting that OCTNs may be important for the transport of xenobiotics and acylated carnitine as well as carnitine itself.
By means of an in vivo disposition kinetic study of carnitine, we demonstrated that the carnitine transporters are absent or functionally deficient in jvs mice because the renal reabsorption, the intestinal absorption, and the distribution to various tissues in these mice are significantly lower than those in wild-type mice (24). However, although functional loss of OCTN2 by mutation was found in jvs mice, the supplementation of carnitine or acetylcarnitine to jvs mice as well as patients improved their pathological symptoms (25)(26)(27). Furthermore, although the tissue distribution of carnitine was decreased in jvs mice, the observed tissue-to-plasma concentration ratios of carnitine in jvs mice are unity or above unity in several tissues (24). Because carnitine is soluble in water and is unlikely to be transported across the cell membrane via passive diffusion, it was strongly suggested that carnitine is concentrated in several tissues of jvs mice by active transport via a transporter other than OCTN2. We preliminarily demonstrated that carnitine was indeed transported by human OCTN1 in addition to OCTN2 (21). Accordingly, it is possible that additional carnitine transporters, which can partially compensate for loss of OCTN2 function, are present in human. Since carnitine metabolism is critical, it is important to clarify the transporters involved in carnitine disposition to understand the physiological roles of carnitine and to identify the * 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 a grant from the Japan Health Sciences Foundation Drug Innovation Project. 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB016257 and AB018436.
ʈ To whom correspondence should be addressed: Dept. of Pharmacobio-Dynamics, Faculty of Pharmaceutical Sciences, Kanazawa University, 13-1 Takara-machi, Kanazawa 920-0934, Japan. Cloning of OCTNs cDNA and Tissue Distribution Study-Mouse OCTN2 was cloned as described previously (12). OCTN1 and OCTN3 were cloned as follows. For OCTN1, a data base search using human OCTN1 cDNA sequence as the query identified several mouse Expressed Sequence Tags that show high homology to the query sequence. From these sequences, MONL 1 primer (5Ј-CGCGCCGAATCGCTGAA-TCCTTTC-3Ј) and MONA 4 primer (5Ј-AGGCTTTTGATTTGTTCTGT-TGAG-3Ј) were designed and used for PCR. A cDNA fragment of around 2 kilobase pair was amplified by PCR from mouse kidney-derived cDNA. This fragment was sequenced and confirmed to have the full open reading frame of OCTN1. For OCTN3, in the course of cloning of OCTN2 by 5Ј-RACE (rapid amplification of cDNA ends), we found that a cDNA fragment that was highly homologous to OCTN2, but not identical with either OCTN2 or OCTN1, was co-amplified. As this fragment was considered to be derived from an unidentified OCTN1and OCTN2-related gene, we termed it OCTN3 and cloned its full open reading frame. MONR 2 primer (5Ј-CGCCCACTGCGCCCGAGTATCC-TC-3Ј) and MONB 16 primer (5Ј-ATAGGCAGAGGTGATTCCAAACTT-3Ј) were designed from the sequences of this fragment and mouse OCTN2, respectively. Using these primers a cDNA fragment of extended OCTN3 sequence was amplified by PCR from mouse embryo (17 day)-derived Marathon-Ready TM cDNA (CLONTECH, Palo Alto, CA). Then MONC1 primer (5Ј-GTGCCTTCAGACCTACATTACTTG-3Ј) was designed from the newly obtained sequence and used to clone the 3Ј end of OCTN3 by 3Ј-RACE (rapid amplification of cDNA ends) from mouse testis-derived Marathon-Ready TM cDNA (CLONTECH). The whole sequence of OCTN3 was determined by assembling these sequences. cDNA fragments covering the full open reading frame of each OCTN were re-amplified by PCR and cloned into pcDNA3 vector (Invitrogen, San Diego, CA). Sequence-verified clones were selected and used for expression experiments.
Tissue Distribution Study by RT-PCR and Western Blot Analyses-In RT-PCR analysis, following member-specific primers were used for RT-PCR. OCTN1: MONL1 primer (described above) and MONA 4 primer (described above). OCTN2: MONB 6 primer (5Ј-TGTTTTTCGT-GGGTGTGCTGATGG-3Ј) and MONB 26 primer (5Ј-ACAGAACAGAA-AAGCCCTCAGTCA-3Ј). OCTN3: MONR 2 primer (described above) and MONB 16 (described above). Using appropriate amounts of cDNA of a multiple tissue cDNA (MTC TM ) panel (CLONTECH) derived from various mouse tissues and whole embryo as templates, each gene was amplified, and the expression levels were evaluated. For Western blot analysis, rabbit polyclonal antibodies were raised against synthesized polypeptides of the carboxyl termini of mouse OCTN1, -2, and -3. Amino acid sequences used for OCTN1, -2, and -3 were NH 2 -CGKKST-VSVDREESPKVLIT-COOH, NH 2 -CTRMQKDGEESPTVLKSTAF-C-OOH, and NH 2 -CKESKGNVSRTSRTSEPKGF-COOH, respectively. Mouse tissues were isolated and homogenized in 2 ml of buffer containing 210 mM sucrose, 2 mM ethylene glycol-bis(␤-aminoethyl ether)-N,N,NЈ,NЈ-tetraacetic acid, 40 mM NaCl, 30 mM Hepes, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 M pepstatin-A, 100 M leupeptin, and 2 g/ml aprotinine (pH 7.4) using Polytron homogenizer. Then 800 l of the homogenate was mixed with 750 l of 1.17 M KCl solution containing 58.3 mM tetrasodium pyrophosphate and centrifuged at 230,000 ϫ g for 75 min. The resultant pellet was suspended in 10 mM Tris-HCl and 1 mM EDTA (pH 7.4) and centrifuged at 230,000 ϫ g again. The obtained pellet was suspended again in 600 l of 10 mM Tris-HCl and 1 mM EDTA (pH 7.4) and dispersed ultrasonically. After the addition of 200 l of 16% SDS solution, the solution was mixed and centrifuged at 15,000 ϫ g, and the resultant supernatant was used for Western blot analysis. The sample was separated by 12% SDS-polyacrylamide gel, proteins were transferred to polyvinylidene difluoride membrane Immobilon (Millipore, Bedford, MA), and the membrane was incubated in buffer, 20 mM Tris, 137 mM NaCl, 0.1% Tween 20 (pH 7.5) containing 10% skim milk. The membrane was incubated with respective polyclonal anti-peptide antibodies for 1 to 2 h, rinsed with above buffer without skim milk three times, and incubated with secondary antibody (donkey anti-rabbit IgG, horseradish peroxidase-linked whole antibody (Amersham Pharmacia Biotech). The membrane was washed with the above buffer without skim milk, and the proteins were detected by enhanced chemiluminescence detection method using ECL Plus Western-blotting detection system (Amersham Pharmacia Biotech). Cultured cells transfected with OCTN1, -2, or -3 were obtained as described above, harvested, and treated as the same as described for the tissue sample preparation.
Transport Study in HEK293 Cells-The full-length OCTN cDNAs were subcloned into the BamHI sites of the expression vector pcDNA3 (Invitrogen, San Diego, CA), and the constructs, pcDNA3/OCTNs, were used to transfect HEK293 cells by means of the calcium phosphate precipitation method as described previously (10). HEK293 cells were routinely grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum (Life Technologies, Inc.), penicillin, and streptomycin in a humidified incubator at 37°C and 5% CO 2 . After 24-h cultivation of HEK293 cells in 10-cm dishes, the cells were transfected with pcDNA3/OCTNs or pcDNA3 vector alone by adding 10 g of the plasmid DNA/ per dish. 48 h after transfection, the cells were harvested and suspended in transport medium containing 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). This suspension and a solution of a radiolabeled test compound in the transport medium were separately incubated at 37°C for 10 min, then transport was initiated by mixing them. At appropriate times, 200-l aliquots of the mixture were withdrawn, and the cells were separated from the transport medium by centrifugal filtration through a layer of a mixture of silicon oil and liquid paraffin with a density of 1.03. Each cell pellet was solubilized in 3 N KOH, then neutralized with HCl, and the associated radioactivity was measured by means of a liquid scintillation counter. HEK293 cells transfected with pcDNA3 vector alone were used to determine the background activity and are designated as Mock cells. Cellular protein content was determined according to the method of Bradford using a protein assay kit (Bio-Rad) and bovine serum albumin as the standard (28). In sodium-free experiments, sodium ions were replaced with N-methylglucamine, and the cells obtained were suspended in sodium-free medium. When transport was measured at acidic or alkaline pH (5.5-8.4), the pH was adjusted appropriately using HCl, NaOH, or KOH.
Data Analysis-Initial uptake rates were usually obtained by measuring the uptake at 3 or 30 min for carnitine and TEA, respectively. The uptake values were expressed as the cell-to-medium concentration (C/M) ratio (l/mg of protein/3 min or 30 min), obtained by dividing the uptake amount in the cells by the concentration of test compound in the medium. To estimate kinetic parameters for saturable transport of carnitine or TEA, the uptake rate 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 or TEA, respectively, and K m and V max are the half-saturation concentration (Michaelis constant) and the maximum transport rate, respectively. All data were expressed as mean 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. Each transport study was repeated at least three times by differently transfected cells, and they essentially showed the same results, so the typical results are shown.

RESULTS
Amino Acid Sequences of Mouse OCTNs-We isolated two new members of the mouse OCTN transporter family, OCTN1 and OCTN3 (Fig. 1a). The sequences of OCTN1 and -3 have been deposited in GenBank TM with the accession numbers AB016257 and AB018436, respectively. The full-length OCTN1, OCTN2, and OCTN3 cDNAs appeared to encode polypeptides of 553, 557, and 564 amino acids and have 73, 86, and 78% similarities with human OCTN2 (9), respectively. The similarities with human OCTN1 were 84, 73, and 67%, respectively (10). Based on the similarities with human OCTN members, we designated the mouse OCTNs as shown in Fig. 1a. OCTNs exhibit about 30% similarity with organic cation transporter OCTs (29 -34) and organic anion transporters OATs (35)(36)(37)(38)(39). The phylogenetic tree of these transporters is shown in Fig. 1b, and it appears that the OCTNs form a distinct family of organic ion transporters.
Tissue Distribution of OCTNs-To compare the tissue distributions of OCTN1, -2, and -3, we performed RT-PCR and Western blot analysis in mouse tissues. Figs. 2, a and b, show the tissue distribution of OCTNs determined by RT-PCR and Western blot analyses, respectively. In RT-PCR analysis of adult mouse tissues, OCTN1 and OCTN2 were clearly expressed in kidney, liver, and testis and weakly in other tissues. They were also found in embryo. On the other hand, OCTN3 was strongly expressed in testis and weakly in kidney but was hardly expressed in other tissues. Antibodies were raised against the divergent COOH-terminal portions of each protein. The polyclonal antibodies were purified by affinity chromatography and were specific for OCTN1, -2, and -3, respectively, with no crossreaction detectable on immunoblots as shown by the immunodetection of the cell lysates of HEK293 cells transfected with mouse OCTN1, -2, or -3 (Fig. 2b, left four lanes). Furthermore, HEK293 cells expressed all of OCTN1, -2, and -3 in significant amounts adequate for functional analysis in the following transport experiments. All antibodies recognized proteins of 64, 70, and 82 kDa for OCTN1, 70 and 80 kDa for OCTN2, and 54 and 60 kDa for OCTN3 in immunoblot of membranes from various mouse tissues. The estimated size of OCTN1, -2, and -3 from the deduced amino acid sequences were 62,287, 62,776, and 63,317, respectively. There are variations in molecular size in the detected bands, and they are close to the estimated sizes. This is presumably ascribed to the variation in post-translational modification, degradation, or glycosylation. Expression profile of OCTN2 protein was broad, and strong expression was observed in kidney. OCTN1 protein was also detected in several tissues, and kidney showed strong expression. OCTN3 was characterized by strong expression in testis and weakly in kidney without expression in other tissues. Therefore, most of the tissue expression profiles were comparable between OCTN genes and their products.
Functional Characterization of OCTNs Expressed in HEK293 Cells-Since human OCTN1 and OCTN2 transported TEA in a Na ϩ -independent manner and carnitine in a Na ϩ -dependent manner (9 -11, 22), we studied the transport characteristics of mouse OCTN family members for both of carnitine and TEA by transient expression in HEK293 cells (Figs. 3, a and b). Uptakes of [ 3 H]carnitine were increased linearly up to 10 min in OCTN2-and OCTN3-transfected HEK293 cells. In OCTN1-transfected cells, [ 3 H]carnitine uptake was very slight, but the uptake activity expressed was statistically higher than that by Mock-transfected cells after  analyses (b). a, specific primers used were described under "Experimental Procedures." The results were analyzed by agarose electrophoresis. G3PDH, glyceraldehyde-3-phosphate dehydrogenase. Sk., skeletal. b, anti-peptide antibodies specific to each OCTN member were used. Cell lysate from HEK293 cells of wild type and transfected with each cDNA of OC-TNs were prepared as described under "Experimental Procedures" and applied on the gel at 6, 7, 3, and 8 g of protein for HEK, HEK-OCTN1, -2, and -3, respectively. Tissues homogenate prepared from mice brain, lung, heart, liver, kidney, and testis were applied on gel (10 g of protein each) and analyzed. Upper, middle, and lower panels show the results from OCTN1, -2, and -3, respectively. Molecular mass is shown in both sides in kDa. 10 min. Because the same results were obtained by differently transfected cells, OCTN1 should have low, but existing, carnitine transport activity. Uptake of [ 14 C]TEA was increased up to 30 min by OCTN1-and OCTN2-transfected cells, whereas its uptake by OCTN3-transfected cells was negligible over 60 min. Comparison of relative transport activities of three OCTNs by correcting the expressed protein amount obtained in Western blot analysis was difficult due to the variation of titers among antibodies, although they recognize each protein specifically (Fig. 2b). Therefore, to avoid the difference of the expression level of respective protein in cells and to allow comparison of the substrate preference among the three members, the relative transport activities for carnitine and TEA were evaluated as summarized in Table I. Here, because OCTN3 essentially is unlikely to transport TEA, we used the apparently expressed transport activity as the difference between OCTN3-transfected and Mock cells. The ratios obtained by dividing the uptake of [ 3 H]carnitine (10 nM) by that of [ 14 C]TEA (100 M) at 10 min were 1.78, 11.3, and Ͼ746 for OCTN1, -2, and -3, respectively. The result suggests that OCTN3 is highly specific for carni-tine, OCTN1 has a preference for TEA, and OCTN2 has intermediate specificity. Fig. 4 shows the Na ϩ dependence of OCTNs-mediated [ 3 H]carnitine uptake. OCTN1-and OCTN2-mediated [ 3 H]carnitine uptakes were much greater in NaCl-containing medium than in N-methyl-D-glucamine chloride-containing medium (i.e. in the absence of Na ϩ ). However, OCTN3-mediated [ 3 H]carnitine uptake was not significantly different between the presence and the absence of Na ϩ . So, OCTN3 seems to have very distinct functional properties from the other OCTNs. In addition, the pH dependence of OCTN2-and OCTN3-mediated [ 3 H]carnitine uptake is shown in Figs. 5, a and b. When the pH in the transport medium was acidic, pH 5.5 and 6.0, OCTN2mediated [ 3  The K m value of TEA uptake via OCTN1 and OCTN2 were 452 Ϯ 41.0 and 215.7 Ϯ 22.7 M, respectively, and the V max values were 11.2 Ϯ 0.75 and 6.02 Ϯ 0.25 nmol/mg of protein/30 min, respectively. The kinetic parameters of carnitine uptake by OCTN1 and TEA uptake by OCTN3 could not be determined because of the difficulty in analyzing them due to the low or negligible transport activity.
Substrate specificity of OCTN2-and OCTN3-mediated carnitine uptake was examined in terms of the inhibitory effect on the initial uptake of [ 3 H]carnitine (Table II). Uptakes of [ 3 H]carnitine by OCTN2 and OCTN3 were significantly inhibited by unlabeled carnitine (5 M), acylcarnitines (5 M), betaines (5, 50 or 500 M), and TEA (500 M). In the case of OCTN2, choline (500 M), an endogenous organic cation, reduced the uptake of [ 3 H]carnitine. However, lysine or ␥-aminobutyrate was not inhibitory. In the case of OCTN3, stronger inhibitory effects of L-carnitine and acylcarnitines were observed compared with the case of OCTN2, whereas glycinebetaine, choline, and TEA showed weaker inhibitory effects. These results suggest that OCTN3 has a higher specificity than OCTN2 for carnitine and related compounds, or in other words, OCTN2 has broader range of specificity than does OCTN3. DISCUSSION In the present study, we isolated two new members of the carnitine/organic cation transporter OCTN family in mice, mouse OCTN1 and -3 (Fig. 1a), and compared their tissue distributions and functional properties with those of the previously reported OCTN2 to characterize them and to elucidate their physiological roles.
As shown in the phylogenetic tree in Fig. 1b, mouse OCTNs seem to form an organic ion transporter family distinct from the organic cation (OCT) and organic anion (OAT) transporter families. By the RT-PCR method, OCTN1 and OCTN2 were widely expressed in all of adult tissues examined in variable amounts, whereas OCTN3 was expressed only in testis strongly and weakly in kidney (Fig. 2a). OCTN2 has been determined to be directly related to SCD (12), and in the  present study it was confirmed by the strong expression of the protein in various tissues by RT-PCR and Western blot analysis (Fig. 2b). Especially, strong expression of OCTN2 protein in kidney demonstrated its contribution to the reabsorption of carnitine from urine after glomerular filtration to maintain carnitine content in body. OCTN1 protein was also detected strongly in kidney, whereas the expression of OCTN3 was very low. These tissue distribution profiles were rather comparable between expressions of the genes and their products. OCTN1 and -2 were continuously expressed in embryonic tissues from 7 to 17 days after birth, whereas OCTN3 was detected only in 7-day embryo. Thus, the tissue distributions of OCTN family members are different, and the strong expression of OCTN3 in testis and limited expression in embryo may reflect a specific role different from those of other members, as discussed below. Skeletal muscle and heart accumulate carnitine. RT-PCR showed low expression of both OCTN1 and -2 in heart, and they were also detected strongly by Western blot analysis. Although by RT-PCR OCTN1 and -2 were expressed in skeletal muscle, we failed to detect the signals of them by Western blot analysis (data not shown). Recently, similar but differential characteristics of carnitine transport from OCTN2 was reported by using isolated membrane vesicles in rat skeletal muscle, suggesting that OCTN1 or others distinct from OCTN2 participate in muscle in rats (40). Therefore, since human OCTN2 is strongly expressed in skeletal muscle (9), mouse and rat may have different regulatory mechanisms of carnitine disposition from human in terms of carnitine transport in skeletal muscle.
When the OCTNs were expressed in HEK293 cells, significantly increased uptake of carnitine was observed with all three OCTNs, although the expressed activity was very small in the case of OCTN1 (Fig. 3). Rat OCTN1 hardly transports carnitine (41), and the carnitine transport activity of human OCTN1 was significantly lower than that of human OCTN2 (21). In addition, the specificity of carnitine transport of mouse OCTN1 in relation to the organic cation TEA was the lowest among the three members (Table I). These results suggest that OCTN1 has differential role from OCTN2 or -3 by transporting other compounds such as organic cations as substrate. Transport of carnitine by rat OCTN2, which was also termed CT1, showed similar affinity (25 M) (42) to that of mouse OCTN2 (22.1 M). Furthermore, human OCTN2-mediated carnitine transport was also high affinity (4.34 M) (9), so OCTN2 appears to be the major Na ϩ -dependent carnitine transporter common to all these species. Uptake activity for TEA was observed in mouse OCTN1-and OCTN2-expressing cells, whereas TEA transport by OCTN3 was negligible (Fig. 3). In addition, K m of OCTN3 for carnitine was smaller (2.99 M) than that of OCTN2 (22.1 M), and the effects of inhibitors on OCTN3 were specific compared with OCTN2, showing a higher specificity of mouse OCTN3 for carnitine. Furthermore, OCTN1 and OCTN2 exhibited Na ϩ -dependent carnitine uptake, whereas OCTN3 did not show Na ϩ dependence (Fig. 4). These observations strongly imply that mouse OCTN3 is a different type of carnitine transporter from OCTN1 or -2, showing higher specificity for carnitine, Na ϩ independence, and a narrow tissue distribution. Transport of TEA via mouse OCTN1 and OCTN2 exhibited lower affinity, with K m values of 452.3 and 215.7 M, respectively. Thus, mouse OCTN1 and -2 accept both carnitine and TEA as substrates but have higher affinity for carnitine than for the organic cation TEA. Interestingly, mouse OCTN1 and -2 transport carnitine in a Na ϩ -dependent manner, whereas TEA transport by them was Na ϩindependent. Accordingly, mouse OCTN1 and OCTN2 appear to be multispecific transporters, mediating organic cation transport as well as carnitine transport by different mechanisms, as has been demonstrated in the human counterparts (9 -11, 21-23).
Detection of OCTN1 gene in liver is in contrast with the absence of human OCTN1 in liver (10), demonstrating a species difference. Cultured human hepatoma HLF cells expressed OCTN2 strongly and OCTN1 slightly (43). In in vitro studies of carnitine uptake by hepatocytes isolated from jvs mice, which genetically lack OCTN2-mediated carnitine transport, a significant decrease of accumulation of carnitine compared with that by wild-type mice was seen (44). Furthermore, when we studied in vivo carnitine disposition in jvs mice, the accumulation of  carnitine in liver was significantly decreased in comparison with that of wild-type mice (24). Accordingly, the major Na ϩdependent carnitine transporter in liver may be OCTN2, and the contribution of OCTN1 is minor. The reasons for this are the very low activity of carnitine transport by OCTN1 (Fig. 3) and presumably lower expression in liver as observed in the Western blot analysis (Fig. 2b). This is supported by the recent finding of negligible transport of carnitine by rat OCTN1 (41). Similarly, a significant contribution of OCTN2 to reabsorption of carnitine in kidney was demonstrated in human and jvs mice (12)(13)(14)(15)(16)(17)(18)(19)24) despite strong expression of OCTN1 in mouse kidney (Fig. 2, a and b). Very recently, it was reported that targeted deletion of a region around OCTN1 gene as well as OCTN2 gene caused many abnormalities related to lipid metabolism in mice (45). Interestingly, the phenotypic abnormalities were not improved by carnitine administration. The symptoms shown in OCTN2-defective jvs mice were improved by carnitine administration (25). Taken all together, OCTN1 should play a distinct role from OCTN2 and -3 in the disposition of carnitine or related compounds.
Since mouse OCTN1 exhibited very low carnitine transport activity, the substrate specificity of only mouse OCTN2 and -3 was examined by observing the inhibitory effect of various compounds on carnitine transport (Table II). Strong inhibitory effects were observed with carnitine-related compounds, including acylcarnitines and ␥-butyrobetaine, whereas lysine, ␥-aminobutyrate, or choline showed negligible or weak inhibitory effect; these are consistent with observations in human and rat OCTN2-mediated carnitine transport (9,23,42). Derivatives or precursor of carnitine showed stronger inhibitory effects on carnitine transport mediated by OCTN3 than that by OCTN2, whereas choline, TEA and glycinebetaine were stronger inhibitors of OCTN2 than OCTN3. Accordingly, OCTN3 is highly specific and has higher affinity for carnitine, and OCTN2 has a rather broader specificity. However, since both commonly accept carnitine and related compounds as substrates and are present in kidney, they may have a common role in the reabsorption of carnitine by mediating Na ϩ -dependent and -independent transport, respectively. The other point of interest is the pH dependence observed in carnitine transport by OCTN2 but not by OCTN3 (Fig. 6). Earlier studies suggested the existence of a specific carnitine transport system in kidney, which showed Na ϩ and pH dependences (46,47). The luminal side of renal tubules is acidic, which may impair the activity to some extent. At present, the physiological significance of the apparent pH dependence observed in OCTN2mediated carnitine transport is not clear.
As discussed above, the contribution of OCTN3 to the reabsorption of carnitine in the kidney may not be great. Therefore mouse OCTN3 was suggested to have a different role from OCTN1 and OCTN2 on the basis of the absence of Na ϩ dependence, high specificity for carnitine, and limited tissue distribution (primarily in testis and a low level in kidney). Spermatozoa are produced in the testis, mature by post-gonadal modifications in the epididymis, and acquire fertilizing ability (48 -50). In the caudal epididymis and the epididymal fluid, extremely high levels of carnitine are observed, and the presence of a specific carnitine transport system was suggested (2,49). These observations suggest that carnitine is important in the development and postgonadal maturation of spermatozoa. It was also shown that OCTN3 mRNA is strongly expressed in testis, and K m for carnitine uptake via OCTN3 (2.99 M) is similar to that of the previously reported Na ϩ -independent high affinity carnitine transport system (K m , 25 M) found in proximal regions of the rat epididymis (51) despite differential experimental methods. In epididymal tubules from rat, carnitine trans-port showed stereospecificity, with higher affinity for L-carnitine compared with the D isomer (48). We compared the inhibitory effects of L-and D-carnitine on mouse OCTN3. Both of the isomers at 5 M showed significant inhibition of L-[ 3 H]carnitine transport (29.5 Ϯ 5.6% of control for the L isomer and 48.5 Ϯ 3.9% for the D isomer), but the D isomer was less potent. This result implies that mouse OCTN3 is involved in carnitine transport to testis and/or epididymis for the functional development and maturation of spermatozoa. Acylcarnitines were more inhibitory than carnitine itself on [ 3 H]carnitine transport by mouse OCTN3 (Table II). The K m value of acetyl-L-carnitine for mouse OCTN3 was 0.63 Ϯ 0.43 M (data not shown), which is lower than the value of carnitine (2.99 M). Therefore, acylcarnitine may be a better substrate for OCTN3. So far, no counterpart of mouse OCTN3 has been found in human, and therefore, mice may have a distinct physiology from human in carnitine disposition and/or carnitinerelated nutritional status. Alternatively, human OCTN1 and/or human OCTN2 may take the role of mouse OCTN3, because both human OCTN1 and -2 were present in testis (9,10). Further studies on the role of the OCTN transporter in testis and epididymis are needed.
In conclusion, OCTN1 and OCTN3 were cloned from mouse as novel members of the OCTN transporter family. Comparative studies of tissue distributions and functional characteristics indicated that all the OCTNs play physiologically important roles in carnitine transport, although OCTN1 showed very low activity of carnitine transport. OCTN2 seems to be the most physiologically important high affinity Na ϩ -dependent carnitine transporter, operating for the reabsorption of carnitine from urine as well as playing a major role in tissue distribution, whereas OCTN3 may be important in testis. OCTN1 and OCTN3 may contribute to reabsorption of carnitine in kidney but were supposed to have other roles due to low activity of OCTN1 and highly specific expression of OCTN3 in testis. OCTN3 is a unique transporter with high specificity for carnitine, Na ϩ -independent transport activity, and predominant expression in testis. Furthermore, OCTN1 and -2 may be involved in the distribution and elimination of organic cations in several tissues. Accordingly, OCTN transporters seem to exhibit multifunctionality as carnitine and organic cation transport systems.