Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Membrane transporters have significant roles in drug absorption and disposition, as well as physiological functions. Interestingly, many physiological transporters show strict substrate specificity, as observed in transporters for neurotransmitters and nutrients, whereas so-called drug transporters exhibit broader substrate selectivity, as well known in the case of multidrug resistance transporter P-glycoprotein, multidrug resistance-associated proteins, and organic ion transporters such as organic cation transporters (OCTs), organic anion transporters (OATs), and organic anion transporting polypeptides. However, although it is important to clarify the determinants of the substrate selectivity to understand the pharmacological and pharmacokinetic relevance of each transporter, the mechanism by which diverse compounds are recognized as substrates by a single transporter is not well understood.

We previously found in humans a novel transporter, OCTN1 (SLC22A4), that showed structural similarity to OCT and OAT transporter families, and transported organic cations such as tetraethylammonium (TEA) (Tamai et al., 1997; Yabuuchi et al., 1999). Rat and mouse OCTN1 also transported organic cations in an Na⁺-independent manner, as observed with human OCTN1 (Tamai et al., 2000; Wu et al., 2000), indicating that OCTN could be classified as one of the organic cation transporter families. The second member of the OCTN family, OCTN2 (SLC22A5), was suggested to be a pH-dependent organic cation transporter (Wu et al., 1998), but we found that OCTN2 is a physiologically important Na⁺-dependent transporter for carnitine, which is a hydrophilic nutrient essential for -oxidation of long-chain fatty acids in mitochondria (Tamai et al., 1998). Subsequently, OCTN2 was suggested to be a Na⁺-independent organic cation transporter (Ohashi et al., 1999, 2001; Wu et al., 1999). The third member of the OCTN family, OCTN3, so far found only in mouse, exhibited negligible activity for organic cations and was rather specific for carnitine (Tamai et al., 2000). Interestingly, OCTN3 did not necessarily require Na⁺ as a driving force to transport carnitine, and it may function in a distinct manner from OCTN2, although OCTNs show rather high similarity of about 70% or more in amino acid sequences. Human OCTN1 has some activity to transport carnitine (Yabuuchi et al., 1999), but mouse and rat OCTN1 exhibited low or negligible carnitine transport activity (Tamai et al., 2000; Wu et al., 2000). Accordingly, OCTNs may have similar but distinct functional/binding sites to discriminate substrates, and OCTN2, which accepts both carnitine and organic cations, may have multiple or universal binding sites.

As mentioned above, it is noteworthy that transport of organic cations by OCTN2 was Na⁺-independent and the driving force remains to be clarified, whereas Na⁺ was prerequisite for carnitine transport by OCTN2. The mechanistic difference between carnitine and organic cation transport observed in OCTN transporters may be due to the presence of multiple functional sites on OCTN proteins, and it is of interest to identify the binding sites of the substrates and Na⁺ because the presence of multiple functional sites would support the idea that a similar mechanism may be responsible for the polyspecificity observed in drug transporters.

OCTN2 is the causative gene for inherited primary systemic carnitine deficiency (OMIM 212140; systemic carnitine deficiency) and is essential for the reabsorption of carnitine filtered through glomeruli across the renal tubular epithelial cells from urine (Koizumi et al., 1999; Nezu et al., 1999; Yokogawa et al., 1999; Tamai et al., 2001). Several mutated alleles of the OCTN2 gene have been isolated from systemic carnitine deficiency patients and many of them were not functional when expressed in in vitro cultured cells (Lamhonwah and Tein, 1998; Koizumi et al., 1999; Nezu et al., 1999; Seth et al., 1999; Tang et al., 1999; Vaz et al., 1999; Wang et al., 1999, 2002a,c; Wu et al., 1999; Mayatepek et al., 2000. Seth et al. (1999) reported that P478L and Y211F mutations in OCTN2 abolished carnitine transport but did not affect organic cation transport or the Na⁺ activation kinetics. These findings suggested a significant spatial separation of the transport sites for carnitine and organic cations. However, the relationship of the binding sites for carnitine and organic cations remains to be clarified.

In the present study, we found that a Ser467Cys (S467C) mutant of human OCTN2, which lacks carnitine transport function (Koizumi et al., 1999), retains activity for the transport of organic cations. The results of functional analysis indicated that OCTN2 accepts carnitine at a site that overlaps but is not identical with that for organic cations, and an anion recognition site that is important for carnitine recognition is located around transmembrane domain (TMD) 11. Based on the observations, a model of binding sites on OCTN2 is proposed.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials. [Methyl-³H]acetyl-L-carnitine hydrochloride (65 Ci/mmol) and L-[methyl-³H]carnitine hydrochloride (85 Ci/mmol) were purchased from Moravek Biochemicals (Brea, CA). [1-¹⁴C]Tetraethylammonium bromide (2.4 mCi/mmol) and [N-methyl-³H]verapamil hydrochloride (78.6 Ci/mmol) were from PerkinElmer Life Sciences (Boston, MA). Pyridinyl-[5-³H]pyrilamine (28 Ci/mmol) was from Amersham Biosciences UK, Ltd. (Little Chalfont, Buckinghamshire, UK). [9-³H]Quinidine hydrochloride (15 Ci/mmol) was from American Radiolabeled Chemicals (St. Louis, MO). Other reagents were obtained from Sigma-Aldrich (St. Louis, MO), Wako Pure Chemicals (Osaka, Japan), and Nacalai Tesque (Kyoto, Japan) and used without further purification. HEK293 cells were obtained from Japanese Cancer Research Resources Bank (Tokyo, Japan).

Uptake Studies by Transient Expression of OCTN2 in HEK293 Cells. The full-length wild- or S467C-mutant 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 (Tamai et al., 1997). The cells were cultivated in Dulbecco's modified Eagle's medium containing 10% fetal calf serum (Invitrogen, Tokyo, Japan), penicillin, and streptomycin in a humidified incubator at 37°C under 5% CO₂. After 24-h cultivation of the cells in the 15-cm dishes, pcDNA3/OCTN2 or pcDNA3 vector alone (mock) was transfected by adding 20 µg of the plasmid DNA per dish. At 48 h post-transfection, the cells were harvested with a rubber policemen, washed twice, and suspended in the transport medium containing 125 mM NaCl, 4.8 mM KCl, 5.6 mM D-glucose, 1.2 mM CaCl₂, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, and 25 mM HEPES, pH 7.4.

Data Analysis. Initial uptake rates of L-[³H]carnitine, [¹⁴C]TEA, or other radiolabeled compounds were obtained by measuring the uptake at 3, 5, or 30 min, respectively, and the uptake values were usually expressed as the cell-to-medium concentration ratio (C/M) (microliters per milligram of protein/3, 5, or 30 min) obtained by dividing the uptake amount by the concentration of test compounds in the medium. To estimate kinetic parameters for saturable transport and the stoichiometry between Na⁺ and carnitine, the uptake rate was fitted to the following equations by means of nonlinear least-squares regression analysis using WinNonlin (Scientific Consulting, Cary, NC):

(1)

(2)

where v and v' and s are the uptake rates and concentration of substrate (carnitine or TEA), and K_m, V_max, and n are the half-saturation concentration (Michaelis constant), the maximum transport rate, and the Hill coefficient, respectively. All data were expressed as mean ± S.E.M., and statistical analysis was performed by using Student's t test. The criterion of significance was taken to be p < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Na⁺ Dependence of Transport of Carnitine and TEA by OCTN2. Na⁺ dependence of uptakes of L-[³H]carnitine and [¹⁴C]TEA by wild-type OCTN2 was examined after transfection of the cDNA into HEK293 cells. As shown in Fig. 1, A and B, OCTN2-mediated uptake of [¹⁴C]TEA was comparable with and without Na⁺, whereas [³H]carnitine uptake by OCTN2 was significantly decreased in the absence of Na⁺. In the absence of Na⁺, Na⁺ was replaced with N-methylglucamine, and it was confirmed that N-methylglucamine did not interact directly with carnitine on OCTN2 in the separate experiment (data not shown). Accordingly, OCTN2 requires Na⁺ for carnitine transport but not for the transport of the organic cation.

View larger version (41K): [in this window] [in a new window]   Fig. 1.   Na+ dependence of carnitine (A) and TEA (B) uptake by HEK293 cells transfected with cDNA of wild-type OCTN2. Uptakes of [3H]carnitine (10 nM) and [14C]TEA (100 µM) by HEK293 cells transfected with the pcDNA vector alone (mock, open columns) and the plasmid vector containing cDNA for wild-type OCTN2 (closed columns) were measured for 3 or 30 min, respectively, at 37°C and pH 7.4 in the presence (+Na+) or absence (Na+) of Na+ in the transport buffer. In the absence of Na+, Na+ was replaced with N-methyl-D-glucamine. , significantly different from the uptake by mock cells by Student's t test (p < 0.05). The results are shown as means ± S.E.M. of three determinations.

Uptake of Carnitine and Organic Cations by HEK293 Cells Expressing Wild-Type and S467C-Mutant OCTN2. To compare the transport activities for carnitines and organic cations, we examined the uptake activities for L-[³H]carnitine, [³H]acetyl-L-carnitine, [¹⁴C]TEA, [³H]pyrilamine, [³H]quinidine, and [³H]verapamil in HEK293 cells transfected with cDNA for wild-type OCTN2 or the S467C-mutant OCTN2 found in Japanese people who showed abnormal carnitine concentration in plasma and urine (Table 1) (Koizumi et al., 1999). For L-[³H]carnitine and [³H]acetyl-L-carnitine, the increases in uptake were significantly decreased in S467C-mutant OCTN2-expressing cells to about 10% of those of wild-type OCTN2-expressing cells. In contrast, the increases in uptake of organic cations such as [¹⁴C]TEA, [³H]pyrilamine, [³H]quinidine, and [³H]verapamil by wild-type OCTN2-expressing cells were comparable with those by mutant OCTN2-expressing cells. Accordingly, the S467C mutation caused about 90% decrease in the uptakes of carnitine and acetyl-L-carnitine without any change in the uptakes of organic cations, suggesting that the transmembrane domain 11 containing the S467C mutation is important for carnitine transport but is not involved in the transport of organic cations.

Functional Characterization of Wild-Type and S467C-Mutant OCTN2. The mechanism of the apparent functional alteration of carnitine transport by S467C mutation was examined by kinetic analysis. Eadie-Hofstee plots of the uptakes of carnitine and TEA after subtracting the background uptakes by mock cells gave single straight lines for both wild-type and mutated OCTN2 (Fig. 2, A and B), suggesting the participation of single functional binding sites for carnitine and TEA. The K_m and V_max estimated by nonlinear least-squares regression analysis for the transports of carnitine and TEA by wild-type OCTN2 were 3.5 ± 0.6 µM and 1.6 ± 0.2 nmol/mg of protein/3 min and 291.5 ± 48.2 µM and 2.5 ± 0.2 nmol/mg of protein/30 min, respectively. The K_m and V_max for carnitine and TEA transports by S467C-mutant OCTN2 were 58.0 ± 4.4 µM and 2.4 ± 0.1 nmol/mg of protein/3 min and 355.0 ± 33.1 µM and 4.5 ± 0.2 nmol/mg of protein/30 min, respectively. The uptake efficiencies evaluated in terms of V_max/K_m of wild-type OCTN2 for carnitine and TEA were 452 µl/mg of protein/3 min and 8.7 µl/mg of protein/30 min, respectively. The V_max/K_m values of S467C-mutant OCTN2 for carnitine and TEA were 41.0 µl/mg of protein/3 min and 12.6 µl/mg of protein/30 min, respectively. Thus, a significant decrease in V_max/K_m for carnitine transport was observed in the mutant OCTN2 to about 10% of that of wild-type OCTN2, and it was apparently caused by an increase in K_m value from 3.5 to 58 µM. On the other hand, no decrease, but rather an increase was observed in V_max/K_m for TEA in the mutated OCTN2. These results indicate that the 90% loss of carnitine transport by S467C-mutant OCTN2 (Table 1) was due to the decrease in the affinity of the OCTN2 protein for carnitine.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Concentration dependence of carnitine (A) and TEA (B) uptake by HEK293 cells transfected with wild-type and S467C-mutant OCTN2. Uptakes of carnitine and TEA were measured for 3 or 30 min at 37°C in transport buffer, pH 7.4, respectively. The results are shown as Eadie-Hofstee plots of the saturable components after subtracting the background uptakes evaluated by using mock cells. Lines represent the saturable uptake of wild-type (closed circles) and S467C-mutant (open circles) OCTN2 obtained from nonlinear least-squares regression analysis of the OCTN2-mediated uptake.

Mutual Inhibitory Kinetics of Uptakes of Carnitine and TEA by OCTN2. Mutual inhibitory kinetics between TEA and carnitine were studied. Figure 3, A and C, show uptakes of increasing concentrations of TEA in the presence and absence of a constant concentration of carnitine (50 µM). The Eadie-Hofstee plot of TEA uptake exhibited a single straight line with a K_m value of 304 ± 38.9 µM. In the presence of 50 µM carnitine, the apparent K_m value was increased to 1540 ± 789 µM, whereas no significant change in V_max was observed (5.01 ± 0.32 versus 5.95 ± 2.30 nmol/mg of protein/30 min in the absence and presence of carnitine, respectively), showing competitive inhibitory kinetics. The K_i of carnitine for TEA transport was 26.8 ± 4.01 µM, which was about 7 times larger than the K_m of carnitine as described below (3.92 µM). Figure 3, B and D, shows uptakes of increasing concentrations of carnitine in the presence and absence of a constant concentration of TEA (500 µM). The evaluated K_m and V_max values for carnitine were 3.92 ± 0.16 µM and 1.70 ± 0.02 nmol/mg of protein/3 min, respectively, in the absence of TEA. TEA exhibited competitive and noncompetitive mixed type inhibition for OCTN2-mediated uptake of carnitine with a slight increase in apparent K_m (5.88 ± 0.35 µM) and a slight decrease in V_max (1.29 ± 0.03 nmol/mg of protein/3 min), compared with those in the absence of TEA. By secondary plot analysis, the apparent K_i value of TEA for carnitine transport was estimated to be 426.0 ± 18.7 µM, which is close to the K_m of TEA itself (304 µM). Carnitine competitively inhibited TEA uptake, but the inhibitory effect of TEA on carnitine uptake could not be explained in terms of simple competitive kinetics. In addition, K_i value of carnitine (26.8 µM) on TEA uptake was significantly higher than K_m of carnitine (3.92 µM), whereas K_i of TEA (426 µM) was relatively close to K_m of TEA (304 µM). These results suggested that the functional sites of carnitine and TEA on OCTN2 protein are very close to each other but are not identical.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Kinetic analysis of mutual inhibitory effects on uptakes of TEA (A and C) and carnitine (B and D) by HEK293 cells transfected with wild-type OCTN2. TEA uptake was measured for 30 min at 37°C in the presence or absence of unlabeled carnitine (50 µM). Carnitine uptake was measured for 3 min at 37°C in the presence or absence of unlabeled TEA (500 µM). The results were analyzed by means of Lineweaver-Burk (A and B) and Eadie-Hofstee (C and D) plots. The data were obtained by subtraction of the uptakes by mock-transfected HEK293 cells from those by OCTN2-transfected HEK293 cells.

Inhibitory Effects of Cationic and Anionic Compounds on Uptakes of Carnitine and TEA by OCTN2. To clarify the selectivity of OCTN2, the inhibitory effects of various cationic and anionic compounds on OCTN2-mediated uptakes of L-[³H]carnitine and [¹⁴C]TEA were examined (Table 2). Verapamil showed marked inhibitory effects on both L-[³H]carnitine and [¹⁴C]TEA uptake via OCTN2. Cationic compounds such as tetramethylammonium, TEA, and tetrabutylammonium also showed significant inhibitory effects (p < 0.05). Interestingly, the anionic compounds bromosulfophthalein (BSP), probenecid, and valproate had weak but significant inhibitory effects on OCTN2-mediated uptake of L-[³H]carnitine at 0.1 or 0.5 mM. Furthermore, probenecid and valproate significantly inhibited OCTN2-mediated [¹⁴C]TEA uptake at the higher concentration of 5 mM, whereas BSP exhibited comparable inhibitory potency to that observed in the inhibition of carnitine uptake. It is likely that OCTN2 recognizes organic anions as well as organic cations, whereas carnitine transport by OCTN2 may be more susceptible to anionic compounds than organic cation transport.

Na⁺ Dependence of Inhibitory Effects of Valproate on TEA Uptake by OCTN2. We further analyzed the inhibitory effect of anions on OCTN2 using valproate, which has an alkyl residue and a carboxyl moiety within the molecule, as an inhibitor. The influence of Na⁺ on the inhibitory effect of valproate on the TEA transport was examined because valproate showed a stronger inhibitory effect on carnitine transport than on TEA transport, as described above (Table 2). The inhibitory effect of increasing concentrations of valproate on the OCTN2-mediated [¹⁴C]TEA uptake was examined in the presence and absence of extracellular Na⁺ (Fig. 4). The inhibitory effect of valproate on [¹⁴C]TEA uptake was concentration-dependent, and was about 2-fold stronger in the presence of Na⁺. Accordingly, Na⁺ seems to modify the affinity of valproate for the TEA binding site on OCTN2.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Na+ dependence of inhibitory effect of valproate on TEA uptake by HEK293 cells transfected with wild-type OCTN2. Uptakes of [14C]TEA (100 µM) were measured for 30 min at 37°C in transport buffer, pH 7.4, in the absence or presence of increasing concentrations of valproate from 0.5 to 10 mM. The results are shown after subtracting the uptakes by mock cells from those by the OCTN2-expressing cells in the presence (closed circles) or absence (open circles) of Na+ in the transport buffer. Uptake in the absence of Na+ was determined by replacing Na+ with N-methyl-D-glucamine in the transport buffer. The results are the means ± S.E.M. of three determinations. , significantly different from the uptake in the presence of Na+ by Student's t test (p < 0.05).

Na⁺ Concentration Dependence of Carnitine Uptake by S467C-Mutant OCTN2. To investigate the cause of the decrease of carnitine transport activity in the S467C-mutant OCTN2, the effect of Na⁺ was compared with that in wild-type OCTN2. Figure 5 shows the Na⁺ concentration dependence of L-[³H]carnitine uptake by S467C-mutant OCTN2 and mock cells. The uptake rate of L-[³H]carnitine after subtracting the uptake by mock cells from that by S467C-mutant OCTN2-expressing cells increased with increasing concentration of Na⁺, and the carnitine uptake exhibited a simple hyperbolic curve as the Na⁺ concentration was increased. The estimated Hill coefficient (n) and K_m(Na⁺) of S467C-mutant OCTN2 according to eq. 2 were 1.32 ± 0.43 and 11.5 ± 3.8 mM, respectively. These values are comparable with previously reported values [n and K_m(Na⁺) of wild-type OCTN2, 0.93 and 18.5 mM, respectively] (Ohashi et al., 1999). Accordingly, the recognition of Na⁺ by wild-type and S467C-mutant OCTN2 protein in relation to carnitine transport is similar. In Fig. 5, slight Na⁺ dependence in carnitine uptake was also observed in mock cells. This observation may be due to the native uptake activity in HEK293 cells by OCTN2 or similar transporters as reported previously (Scaglia et al., 1999).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Extracellular Na+ concentration dependence of L-[3H]carnitine uptake by HEK293 cells transfected with S467C-mutant OCTN2. Uptake of L-[3H]carnitine (10 nM) by HEK293 cells transfected with S467C-mutant OCTN2 (closed circles) and pcDNA3 vector alone (open triangles) was measured for 3 min at 37°C. Osmolality of the medium was adjusted by adding appropriate concentrations of N-methyl-D-glucamine chloride to make the sum of NaCl and N-methyl-D-glucamine chloride concentration in the medium 125 mM. The results are the means ± S.E.M. of three determinations.

Inhibitory Effect of Valproate on TEA Uptake by Wild-Type and S467C-Mutant OCTN2. Because Na⁺ binding to OCTN2 directly affected the recognition of valproate by OCTN2 (Fig. 4), the inhibitory effects of valproate on TEA uptake by wild-type and S467C-mutant OCTN2 were studied (Fig. 6). Herein, it was confirmed that TEA transport by S467C-mutant OCTN2 was Na⁺-independent as the same as observed in wild-type OCTN2 (data not shown). Valproate (5 mM) inhibited TEA uptake by wild-type and S467C-mutant OCTN2 to 45.3 ± 6.87 and 76.7 ± 0.47% of the control values, respectively. A marked difference in sensitivity to inhibition by valproate was observed between wild-type and S467C-mutant OCTN2. Because the existence of anion-, cation-, and Na⁺-binding sites was suggested by the mutual inhibition kinetics and Na⁺ dependence of carnitine uptake by OCTN2, this result indicates the anion binding site close to the Na⁺-binding site on OCTN2 may be affected by the S467C mutation.

View larger version (52K): [in this window] [in a new window]   Fig. 6.   Inhibitory effect of valproate on TEA uptake by HEK293 cells transfected with wild-type and S467C-mutant OCTN2. The uptakes of [14C]TEA (100 µM) were measured for 30 min at 37°C in transport buffer, pH 7.4 in the absence (closed columns) or presence (open columns) of 5 mM valproate. The results show the values after subtracting the uptakes by mock cells from those by the OCTN2-expressing HEK293 cells. The results are the means ± S.E.M. of three determinations. , significantly different from the inhibitory effect in wild-type OCTN2 by Student's t test (p < 0.05).

Sequence Alignments of OAT, OCTN, and OCT in TMD 11 and the Hydrophilic Loop between TMD 10 and 11. Because carnitine transport by S467C-mutant OCTN2 as well as P478L-mutant OCTN2 (Seth et al., 1999) was altered without any effect on TEA transport, and these amino acids are located in the TMD 11 region on OCTN2 protein, we hypothesized that the anion recognition site or pocket may be located in or near TMD 11. Accordingly, we compared amino acid sequences among SLC22A families, including OAT, OCT, and OCTN transporter (Fig. 7). Sequence alignments showed that there is one conserved basic amino acid, arginine (R), in the middle of TMD 11 in the OAT family and OCTN family. It is likely that the substrate recognition characteristics of TMD 11 in the OCTN family are similar to those in the OAT family. The OCT family has a conserved acidic amino acid, aspartate (D) in the corresponding position. Wang et al. (2000b) reported that mutation of E452K in OCTN2 influenced the affinity of Na⁺ to OCTN2 protein. The OCTN family exhibits about 30% similarity with the OAT and OCT families, which are Na⁺-independent organic solute transporters. Therefore, we also compared the amino acid sequences of the OAT, OCTN, and OCT families at this position. E452 is located on the hydrophilic loop between TMD 10 and 11 on OCTN2, and glutamate (E) is commonly located on related positions in all OATs, OCTs, and the other OCTNs. These results indicated that E452 is neither relevant to cation or anion differentiation, nor directly related to the Na⁺-binding site of OCTN2.

View larger version (48K): [in this window] [in a new window]   Fig. 7.   Multiple alignments of transmembrane domain 11 and the hydrophilic loop between TMD 10 and 11 in the OAT, OCTN and OCT families. Arrowheads show key amino acid positions. The position of TMD 11 for human OCTN2 is surrounded with a square. The numbering of amino acid residues is based on that of human OCTN2.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have already demonstrated the multifunctionality of OCTN2, which transports carnitine in an Na⁺-dependent manner and organic cations in an Na⁺-independent manner (Fig. 1; Tamai et al., 1998, 2000; Ohashi et al., 1999, 2001). We also reported a similar situation in NPT1 transporter, which transports inorganic phosphate and organic anions such as p-aminohippuric acid in Na⁺-dependent and Na⁺-independent manners, respectively (Yabuuchi et al., 1998; Uchino et al., 2000). The characteristic that different substrates are transported by different driving forces is interesting because it may be one of the mechanisms of the polyspecificity of drug transporters. For the purpose of clarification of such multifunctionality of OCTN transporters, herein we used a mutant of OCTN2 that has a single amino acid change; this mutant was found in Japanese subjects who showed abnormal carnitine concentration in plasma and urine in our previous study (Koizumi et al., 1999). The mutant OCTN2 from serine 467 to cysteine (S467C) had little carnitine transport activity, but retains organic cation transport activity (Table 1). A similar observation was reported in P478L mutant of human and rat OCTN2 and in Y211F mutant of rat OCTN2, suggesting that there are distinct functional sites for organic cations and carnitine on OCTN2 (Seth et al., 1999). In the present study, based on the functional characterization of wild and S467C-mutant OCTN2 transporter, the selective reduction of carnitine transport by S467C mutation was ascribed to the change in affinity of the site to recognize an anionic moiety that is specific for carnitine but not for organic cation.

We kinetically analyzed the uptakes of carnitine and TEA by wild and S467C-mutant OCTN2. This mutation did not cause change in the kinetic parameters for TEA transport (Fig. 2). However, the K_m value of carnitine for S467C-mutant OCTN2 was about 15-fold that for the wild type, resulting in a reduction of transport efficiency to about 10% of that of wild-type OCTN2. Because the decrease of carnitine uptake by S467C mutation was explained by the increase in K_m, S467 must be important to determine the affinity for carnitine but not for organic cations. Mutual inhibition kinetics (Fig. 3) revealed that carnitine competitively inhibited TEA uptake, but TEA did not show completely competitive inhibition kinetics for carnitine uptake. Furthermore, the K_i value of carnitine for TEA transport (26.8 µM) was about 7 times higher than the K_m value of carnitine itself (3.92 µM), whereas the K_i value of TEA for carnitine transport (426 µM) was comparable to the K_m value of TEA (304 µM). We observed similar mutual inhibitory kinetics in mouse OCTN2 (data not shown). Seth et al. (1999) reported that carnitine transport via rat and human OCTN2 was competitively inhibited by TEA, suggesting an interaction on a common binding site on OCTN2, although they did not describe the effect of carnitine on TEA transport. These observations suggest that the binding sites for TEA and carnitine overlap, but are not identical, on OCTN2. Based on the finding that the K_i value of carnitine for TEA transport is higher than the K_m value of carnitine, we speculate that the binding site of TEA is not completely shared with carnitine, and a conformational change of OCTN2 due to the prior binding of TEA may affect carnitine binding to OCTN2. Furthermore, mixed type inhibition of carnitine transport by TEA, with comparable K_m and K_i values for TEA, indicated that the binding site for carnitine is partly shared with TEA. Accordingly, the kinetic analysis suggests that the TEA binding site is located within the carnitine binding site.

The selectivity of inhibitors on TEA and carnitine transport by OCTN2 (Table 2) indicates that OCTN2 recognizes both cationic and anionic charges. However, relatively lower inhibitory potencies of anionic probenecid and valproate on TEA uptake than on carnitine uptake were observed, whereas cationic TEA and verapamil exhibited comparable inhibitory effects on uptake of both compounds. Herein, verapamil, a cationic substrate of OCTN2, strongly inhibited the uptakes of both compounds. Verapamil was reported to be a noncompetitive inhibitor of carnitine transport in human fibroblasts (Scaglia et al., 1999). The mechanism of strong inhibition of verapamil on OCTN2 has not been clarified yet and it may include both competitive and noncompetitive inhibition. These results suggest that the binding sites for TEA and carnitine should have distinct affinity for anionic compounds, but they have a common binding site for cationic compounds. Accordingly, an anionic moiety of the substrate/inhibitor may increase the affinity to the part of the carnitine binding site that is not involved in the binding of organic cations. This seems reasonable because carnitine, but not TEA, has an anionic moiety within the molecule.

A significant difference in the functionality of OCTN2 toward TEA and carnitine lies in the Na⁺ dependence. Therefore, we examined the effect of Na⁺ to further support the presence of differential binding sites for TEA and carnitine. Wu et al. (1999) and we (Ohashi et al., 2001) have already demonstrated that carnitine exhibits higher affinity to OCTN2 in the presence of Na⁺, whereas TEA does not. Valproate showed weak but significant inhibitory effects on the uptakes of carnitine and TEA (Table 2), and its chemical structure is similar to that of carnitine in part. Valproate showed concentration-dependent inhibition of TEA uptake both in the absence and presence of Na⁺, and its inhibitory potency was high in the presence of Na⁺ (Fig. 4). It was also demonstrated that cephaloridine, a zwitterionic -lactam antibiotic, has a Na⁺-dependent inhibitory effect on TEA uptake by rat OCTN2 (Ganapathy et al., 2000). Accordingly, Na⁺ may be required for the higher affinity binding of anionic and zwitterionic compounds to OCTN2 by activating anion recognition site.

Wang et al. (2000b) reported that the E452K mutation in OCTN2 affected the affinity for Na⁺ but not carnitine. This observation indicates that E452, located in the hydrophilic loop between TMD 10 and 11, is involved in Na⁺ recognition or flux. However, E452 on OCTN2 is conserved among OCTN, OAT, and OCT transporters that are commonly classified into the SLC22A family and show Na⁺-independent transport activity (Fig. 7). Accordingly, the Na⁺-binding site might be spatially close to E452, and the mutation E452K may influence the size or shape of the Na⁺-binding pocket on OCTN2. In the present study, we analyzed the Na⁺ activation kinetics of carnitine transport by the S467C mutant because it is located close to E452. However, no significant change of K_m(Na⁺) or Hill coefficient in the S467C mutant was observed compared with wild-type OCTN2 (Ohashi et al., 1999). Accordingly, S467C does not affect the binding of Na⁺, and this supports the abnormality of the anion recognition site in the S467C mutant, as discussed above. So, we examined the inhibitory effect of valproate on TEA uptake in the presence of Na⁺. The inhibitory effect of valproate on TEA uptake by wild-type OCTN2 was stronger than that in the case of the S467C mutant. These results indicated that S467C caused functional alteration of the anion recognition site, which may be located in the TMD 11 region of OCTN2.

Recently, it was reported that the conserved basic amino acid residues lysine 370 and arginine 454 of rat OAT3 are important for organic anion transport (Feng et al., 2001), and acidic amino acid aspartate 475 in rat OCT1 is involved in organic cation transport (Gorboulev et al., 1999). Because OCTNs have structural similarity with these transporters, the functional site on OCTN2 might be related to these amino acid residues. As aligned in Fig. 7, R471, which is located in TMD11 of OCTN2 and close to S467, is conserved at the corresponding position of R454 in OAT3, whereas it is replaced with aspartate (D475) in OCTs. Because mutation of OAT and OCT corresponding to R471 of OCTN2 caused alteration of functionality (Gorboulev et al., 1999; Feng et al., 2001), TMD 11 must be important to determine the affinity for substrates. Mutations of S467C and P478L (Seth et al., 1999) in OCTN2 may influence the size or shape of the anion recognition pocket, leading to functional alteration of anion recognition of OCTN2. Based on these observations, the molecular mechanism of OCTN2 is postulated (Fig. 8). The positive charge of TEA is recognized by the cation binding site, and other factors such the length of the side chain or hydrophobicity decide the affinity. The negative charge of valproate is recognized by the anion binding site, located presumably in TMD 11, and activated by Na⁺ binding to OCTN2, and the isopropyl residue interacts with the side chains of TEA. Finally, the positive and negative charges of carnitine are recognized by the cation and anion binding sites, respectively, then high-affinity binding of carnitine to OCTN2 is achieved by Na⁺-dependent activation, which is closely related to the anion binding site. This binding model for substrates well explains the OCTN2 characteristics of Na⁺-dependent recognition of zwitterionic and anionic compounds and Na⁺-independent recognition of organic cation.

View larger version (34K): [in this window] [in a new window]   Fig. 8.   Model of binding sites on OCTN2 protein for organic solutes and Na+. The proposed mechanism of substrate recognition of OCTN2 is schematically shown. The binding site is composed of three sites: a site for charged cationic moieties, a site for charged anionic moieties, and a site for Na+. Herein, TEA, valproate, and carnitine are drawn as models of cationic, anionic, and zwitterionic substrates, respectively. The specific binding of Na+ to its binding site can activate a nearby anion recognition site on OCTN2 protein. The cation recognition site can accept the cationic moiety without activation by Na+. Thus, this model explains both Na+-coupled carnitine transport and Na+-uncoupled organic cation transport.

In conclusion, by using the S467C mutant of OCTN2, we obtained evidence that OCTN2 may have functionally distinct binding sites for carnitine and organic cations, and the anion recognition site that is specific for carnitine is located in the TMD 11 region. These findings help to provide a basis for the molecular mechanism of polyspecificity of OCTN2, and similar mechanisms may underlie the structure-activity relationships of other drug transporters.