Functional Analysis of a Canalicular Multispecific Organic Anion Transporter Cloned from Rat Liver*

Transport of many organic anions across the bile canalicular membrane is mediated by the canalicular multispecific organic anion transporter (cMOAT). Previously, we cloned cDNA that may encode cMOAT from Sprague-Dawley rat liver (Ito, K., Suzuki, H., Hirohashi, T., Kume, K., Shimizu, T., and Sugiyama, Y. (1997)Am. J. Physiol. 272, G16–G22). In the present study, the function of this cloned cDNA was investigated by examining the ATP-dependent uptake ofS-(2,4-dinitrophenyl)-glutathione (DNP-SG) into membrane vesicles isolated from an NIH/3T3 cell line transfected with an expression vector containing the cloned cDNA. Although the membrane vesicles from the control NIH/3T3 cells exhibited endogenous activity in transporting DNP-SG and leukotriene C4 in an ATP-dependent manner, the transfection of cMOAT cDNA resulted in a significant increase in the transport activity for these ligands. The uptake of DNP-SG into membrane vesicles was osmotically sensitive and was stimulated to some extent by other nucleotide triphosphates (GTP, UTP, and CTP) but not by AMP or ADP. TheK m and V max values for the uptake of DNP-SG by the membrane vesicles were 0.175 ± 0.031 μm and 11.0 ± 0.73 pmol/min/mg protein, respectively, for the transfected rat cMOAT and 0.141 ± 0.036 μm and 3.51 ± 0.39 pmol/min/mg protein, respectively, for the endogenous transporter expressed on control NIH/3T3 cells. These results suggest that the product of the previously cloned cDNA has cMOAT activity being able to transport organic anions in an ATP-dependent manner. Alternatively, it is possible that the cDNA product encodes an activator of endogenous transporter since the K m value for DNP-SG was comparable between the vector- and cMOAT-transfected cells. The transport activity found in the control NIH/3T3 cells may be ascribed to mouse cMOAT since Northern blot analysis indicated the presence of a transcript that hybridyzed to the carboxyl-terminal ATP-binding cassette sequence of the murine protein.

Recently, the molecular features of cMOAT have been clarified (20 -22). Focusing on the fact that 1) the substrate specificity of cMOAT is similar to that of human multidrug resistance-associated protein (hMRP) (23)(24)(25) and 2) that the highly conserved ATP-binding cassette (ABC) region is observed among a series of ABC transporters (23), we and others recently cloned cDNA (4,623 base pairs) that may encode cMOAT from SD and Wistar rat liver based on the homology with ABC region of hMRP, respectively (20 -22, 26). Northern blot analysis revealed that the expression of cMOAT is defective both in TR Ϫ and EHBR (20 -22). In addition, studies with antibodies indicated selective loss of the expression of cMOAT from the canalicular membrane in TR Ϫ and EHBR (20,21). Further analysis by Paulusma et al. (20) revealed that a 1-base pair deletion at amino acid 393 resulted in the introduction of the stop codon at amino acid 401 in TR Ϫ rats. We also found that a 1-base pair replacement (G 3 A) at amino acid 855 resulted in the introduction of the premature stop codon in EHBR (22). Since EHBR and TR Ϫ are allelic mutants (27) and both strains exhibit an autosomal recessive inheritance in the biliary excretion of organic anions (18), it was suggested that the impaired expression of this particular protein is related to the pathogenesis of hyperbilirubinemia in the mutant animals.
The impaired expression of cMOAT in a patient suffering from Dubin-Johnson syndrome was also demonstrated immunohistochemically (4,28). Recently, the human homologue of rat cMOAT was cloned from several human tumor cell lines (29). Northern blot analysis suggested that the expression of a transcript which can hybridize with the human cMOAT probe is enhanced in cisplatin-resistant cell lines (29). If we consider the fact that 1) cisplatin is metabolized within the cells to form the glutathione conjugate (30) and 2) that cMOAT can accept many glutathione conjugates as a substrate (1)(2)(3)(4), it is possible * 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) AB008832.
Functional analysis, however, remains to be performed to finally show that the previously cloned cDNA actually encodes a protein with cMOAT activity. In the present study, we established an NIH/3T3 cell line transfected with an expression vector containing cMOAT cDNA and examined the transport activity using membrane vesicles isolated from these transfected cells.
Preparation of the Transfected Cell Line-cMOAT cDNA with the shortest 3Ј-UTR length in pBluescript II SK(Ϫ) vector described previously (22) was excised with EcoRI and then inserted into the EcoRI site in the pCXN2 vector. The NIH/3T3 cells, transfected with pCXN2 vector by Lipofectin (Life Technologies), was maintained in the presence of 800 g/ml G418 (Geneticin, Life Technologies) to obtain the colonies. We identified five colonies and determined the expression of cMOAT in each of these using Northern blot analysis. We prepared membrane vesicles from one clone whose cMOAT expression was highest.
Transport Studies-Membrane vesicles were prepared from 2 ϫ 10 8 of the control and transfected NIH/3T3 cells as described previously (32) and were frozen in liquid nitrogen and stored at Ϫ100°C until use. Protein concentrations were determined by the Lowry method. In addition, the orientation of membrane vesicles was determined by examining the nucleotide pyrophosphatase accessibility (33).
The transport study was performed using the rapid filtration technique (5). Briefly, transport medium (10 mM Tris, 250 mM sucrose, 10 mM MgCl 2 , pH 7.4), containing radiolabeled compounds (16 l) with or without unlabeled substrate was preincubated at 37°C for 3 min, and then rapidly mixed with 4 l of membrane vesicle suspension (10 g of protein), with or without 5 mM ATP and ATP-regenerating system (10 mM creatine phosphate, 100 g/ml creatine phosphokinase). In some instances, ATP was replaced by AMP, ADP, GTP, UTP, or CTP. The transport reaction was stopped by the addition of 1 ml of ice-cold buffer containing 250 mM sucrose, 0.1 M NaCl, 10 mM Tris-HCl (pH 7.4). The stopped reaction mixture was filtered through a 0.45-m GVWP filter (Millipore Corp., Bedford, MA) and then washed twice with 5 ml of stop solution. Radioactivity retained on the filter was determined using a liquid scintillation counter (LSC-3500, Aloka Co., Tokyo, Japan).
Northern Blot Analysis-The cDNA fragment containing the aminoterminal ABC region of rat cMOAT (nucleotides 2122-3154) was prepared as described previously (22). The cDNA fragment encoding the carboxyl-terminal ABC region of mouse cMOAT and MRP was amplified from BALB/c mouse liver and lung RNA, respectively, by RT-PCR using degenerated primers as described previously (26). The amplified PCR product was subcloned into the EcoRV site of pBluescript II SK(Ϫ), and then the sequence was determined. This cDNA fragment was excised by digestion with EcoRI and HindIII for use as the probe. Northern hybridization was performed as described previously (26). Filters were exposed to Fuji imaging plates (Fuji Photo Film Co., Ltd., Kanagawa, Japan) for 3 h at room temperature and analyzed by a BAS imaging analyzer (Fuji Photo Film Co., Ltd.).

Northern Blot Analysis of the Expression of Rat cMOAT-
The expression of rat cMOAT in the cells cultured for 10 weeks after transfection was confirmed by Northern blot analysis. As shown in Fig. 1A, rat cMOAT probe hybridized to the NIH/3T3 cells transfected with a vector containing rat cMOAT cDNA, but not to those transfected with the vector. The length of the transcript in cMOAT-transfected NIH/3T3 cells was comparable with the shortest band observed in SD rat liver (22).

Uptake of [ 3 H]DNP-SG and [ 3 H]LTC 4 into Membrane
Vesicles-The enrichment of leucine amino peptidase was 8.1-and 6.6-fold in plasma membrane vesicles from cMOAT-transfected and vector-transfected cells relative to the cell homogenate, respectively. The sideness of the membrane vesicles was also comparable between these cells; 36 and 34% of the membrane vesicles were inside-out for cMOAT-transfected and vectortransfected NIH/3T3 cells, respectively. Fig. 2 4 in an ATP-dependent manner, the stimulating effect of ATP was greater in the cMOAT-transfected NIH/3T3 cells (Fig. 2, A and B). The clearance for the initial uptake of DNP-SG into cMOAT-transfected cells was 9.85 Ϯ 0.42 l/ min/mg of protein (mean Ϯ S.E.; n ϭ 3), which is significantly (p Ͻ 0.05) higher than that observed in NIH/3T3 cells without any plasmid (3.15 Ϯ 0.075 l/min/mg of protein; n ϭ 3) or vector-transfected NIH/3T3 cells (2.66 Ϯ 0.22 l/min/mg of protein; n ϭ 3). The clearance for the uptake of [ 3 H]DNP-SG into membrane vesicles isolated from NIH/3T3 cells not transfected with any plasmid (3.15 Ϯ 0.075 l/min/mg of protein; n ϭ 3) was not significantly different from that observed for the vector-transfected NIH/3T3 cells (2.66 Ϯ 0.22 l/min/mg of protein; n ϭ 3). Moreover, we found that the expression of rat cMOAT in the transfected cells was reduced during storage in liquid N 2 ; after thawing, expression of the transcript in the transfected cells fell below the limit of detection. In accordance with the reduced expression of transfected cMOAT, the uptake of DNP-SG into membrane vesicles from thawed cells was significantly (p Ͻ 0.05) reduced to 3.13 Ϯ 0.22 l/min/mg of protein (n ϭ 3), a figure not significantly different from that obtained in parental and vector-transfected NIH/3T3 cells.
In addition, the uptake of [ 3 H]DNP-SG into membrane vesicles from cMOAT-transfected and vector-transfected cells was reduced by increasing the sucrose concentration in the medium. The y-intercept for relationship between the amount of DNP-SG associated with the vesicles versus the reciprocal of the sucrose concentration in the medium was almost 0. GTP, CTP, and UTP stimulated the uptake of [ 3 H]DNP-SG into membrane vesicles isolated from cMOAT-transfected cells to 38.3, 41.2, and 44.4% of that observed in the presence of ATP. In the same manner, the uptake of [ 3 H]DNP-SG into membrane vesicles from vector-transfected cells in the presence of GTP, CTP, and UTP was 55.6, 54.4, and 56.6% of that observed in the presence of ATP. In contrast, no effect of ADP or AMP was observed for both membrane vesicle preparations. Vana- The ATP-dependent uptake of [ 3 H]DNP-SG into membrane vesicles was saturable (Fig. 3). The nonlinear regression analysis of the vector-transfected cells revealed that the uptake can be described by a saturable (K m ϭ 0.141 Ϯ 0.036 M, V max ϭ 3.51 Ϯ 0.39 pmol/min/mg of protein) and a non-saturable component with a clearance of 0.228 Ϯ 0.030 l/min/mg of protein.
The ability of rat cMOAT to transport [ 3 H]DNP-SG was determined by subtracting the uptake in the vector-transfected cells from that in the cMOAT-transfected cells (Fig. 3). The K m and V max values for the transport of [ 3 H]DNP-SG by rat cMOAT were 0.175 Ϯ 0.031 M and 11.0 Ϯ 0.73 pmol/min/mg of protein, respectively.
Northern Blot Analysis of the Expression of Endogenous Transporters-Expression of MRP/cMOAT related proteins in the control NIH/3T3 cells was examined by Northern blot analysis. The amplified cDNA fragment encoding carboxyl-terminal ABC region of mouse cMOAT (367 base pairs) (Fig. 4) exhibited 92.6 and 94.3% homology at the cDNA and deduced amino acid level with rat cMOAT (20 -22), respectively, and hybridyzed with the poly(A) ϩ RNA from NIH/3T3 cells to produce the ϳ6-kilobase band (Fig. 1B). The deduced amino acid sequence of the amplified cDNA fragment of mouse MRP was the same as that reported previously (34). The homology in the carboxylterminal ABC region between mouse MRP and mouse cMOAT was 68.0 and 78.0% at the cDNA and deduced amino acid level, respectively. Northern blot indicated that the expression of mouse MRP in NIH/3T3 cells was below the detection limit (data not shown). DISCUSSION In the present study, we examined the function of the product of the recently cloned rat cDNA, whose expression is defective in EHBR and TR Ϫ (20 -22), by examining the transport of typical substrates for cMOAT in the cDNA transfected cells. Since 1) the uptake of [ 3 H]DNP-SG into membrane vesicles from the cMOAT-transfected cells was stimulated to a greater extent by ATP compared with that from the vector-transfected cells ( Fig. 2A) and 2) since the expression of rat cMOAT in the cMOAT-transfected cells was confirmed by Northern blot analysis, it was concluded that rat cMOAT activity is associated with its expression at an mRNA level. Uptake of DNP-SG and LTC 4 into membrane vesicles isolated from the control NIH/ 3T3 cells was stimulated by ATP, suggesting the presence of endogenous ABC transporters for these glutathione conjugates (Fig. 2). To estimate the function of rat cMOAT, therefore, the uptake in control cells should be subtracted from that in cMOAT-transfected NIH/3T3 cells. Since the uptake of DNP-SG into membrane vesicles was comparable between the parent and vector-transfected NIH/3T3 cells (see "Results"), the vector introduction may not affect the expression of endog-enous transporters.
The uptake of [ 3 H]DNP-SG was osmotically sensitive (see "Results"), suggesting that a large part of the accumulation by vesicles can be accounted for by transport into the intravesicular space, but not by binding to the vesicle surface. GTP, CTP, and UTP could also enhance the uptake of DNP-SG by cMOAT to some extent (see "Results"), which was consistent with hMRP (35). Vanadate was effective in reducing the ATP-stimulated uptake of DNP-SG, which was consistent with the observations in CMVs (8) and in hMRP (32).
Kinetic analysis revealed that the K m of rat cMOAT was 0.175 M (see "Results"), which was more than 10-fold lower than that reported for the uptake of DNP-SG by CMVs; using rat CMVs, Kobayashi et al. tively, it is also plausible that the cDNA product encodes an activator of endogenous transporter since the K m value of DNP-SG was comparable between the vector-and cMOATtransfected NIH/3T3 cells. Such protein-protein interaction has been demonstrated on the plasma membrane. For example, Inagaki et al. (39) reported that the co-expression of ATP-dependent K ϩ -channel activity (␤ cell inward rectifier (BIR)) and sulfonylurea receptor (SUR) is required for BIR activity in COS cells. They found that COS-1 cells transfected with BIR alone or SUR alone did not exhibit this function (39).
In the present study, we found the presence of ATP-dependent transport of DNP-SG and LTC 4 in control NIH/3T3 cells. The presence of such endogenous activity on mouse plasma membrane has been reported previously; Saxena and Henderson (37) found that DNP-SG is taken up into membrane vesicles from L1210 cells via high (K m ϭ 0.63 M) and low (K m ϭ 450 M) affinity systems, the former being inhibited by LTC 4 with a K i value of 0.20 M. Northern blot analysis using a mouse MRP probe suggested that the expression of MRP in NIH/3T3 cells was minimal (see "Results"). In contrast, a transcript which hybridizes to the carboxyl-terminal ABC region of mouse cMOAT was observed at the position which was same as that in mouse liver (Fig. 1), suggesting that the mouse cMOAT rather than MRP may be responsible for the endogenous activity in transporting DNP-SG in NIH/3T3 cells.
In conclusion, we have shown that the product of the previously cloned cDNA (20 -22) has the ability to transport glutathione conjugates in an ATP-dependent manner, which is the most important characteristic of cMOAT. Together with the previous finding that the expression of the cloned cDNA 1) is predominantly observed in the liver among all the tissues examined (20,22), 2) is almost exclusively observed on the bile canalicular membrane (20,21), and 3) is hereditarily defective in both allelic mutant rat strains (EHBR and TR Ϫ ) (20 -22) and humans (28), this leads us to conclude that the defective expression of this transporter is the pathogenesis for the jaundice in the Dubin-Johnson syndrome found in humans.  4. Alignment of the nucleotide and deduced amino acid sequence of the cDNA probe for mouse cMOAT and mouse MRP. Carboxyl-terminal ABC region of the mouse cMOAT and mouse MRP were amplified from BALB/c mouse liver and lung, respectively, using degenerated PCR primers described previously (26). The sequence between forward and reverse primers is listed in this figure. Asterisks represent the consensus sequence. Dots represent the amino acids homologous to each other.