Identification of Basic Residues Involved in Drug Export Function of Human Multidrug Resistance-associated Protein 2*

Multidrurg resistance-associated protein 2 (MRP2)/canalicular multispecific organic anion transporter (cMOAT) is involved in the ATP-dependent export of organic anions across the bile canalicular membrane. To identify functional amino acid residues that play essential roles in the substrate transport, each of 13 basic residues around transmembrane regions (TMs) 6–17 were replaced with alanine. Wild type and mutant proteins were expressed in COS-7 cells, and the transport activity was measured as the excretion of glutathione-methylfluorescein. Four mutants, K324A (TM6), K483A (TM9), R1210A (TM16), and R1257A (TM17), showed decreased transport activity, and another mutant, K578A (TM11), showed decreased protein expression. These five mutants were normally delivered to the cell surface similar to the other fully active mutants and wild type MRP2. The importance of TM6, TM16, and TM17 in the transport function of MRP2 is consistent with the previous observation indicating the importance of the corresponding TM1, TM11, and TM12 on P-glycoprotein (Loo, T. W., and Clarke, D. M. (1999)J. Biol. Chem. 274, 35388–35392). Another observation that MRP2 inhibitor, cyclosporine A, failed to inhibit R1230A specifically, indicated the existence of its binding site within TM16.

Transport across the hepatocellular canalicular membrane into the bile is a critical step in the elimination of endogenous and exogenous compounds in mammals (1,2). These compounds are converted into amphiphilic anionic conjugates with glutathione, glucronate, or sulfate by the catalysis of several hepatocellular enzymes (1,2). The excretion of these conjugates into the bile is mediated by the 190-kDa multidrug resistanceassociated protein 2 (MRP2), 1 also termed canalicular multispecific organic anion transporter (cMOAT) (2)(3)(4)(5) or canalicular multidrug resistance-associated protein (cMRP) expressed at canalicular surface of hepatocyte (6,7). The hepatobiliary secretion of such compounds is therefore strongly reduced in mutant rat strains lacking MRP2 expression such as the TR Ϫ rat and the EHBR rat (8 -10). Similarly, a deficiency in human MRP2 expression because of a mutation in the gene causes decreased biliary excretion of bilirubin glucronides, resulting in hereditary conjugated hyperbilirubinemia, known as Dubin-Johnson syndrome (8,11,12).
MRP2 belongs to the ATP-binding cassette transporter superfamily that includes other drug efflux pumps such as the 170 kDa P-glycoprotein (P-gp/MDR1) and the 190-kDa multidrug resistance-associated protein (MRP1) (8). MRP2 is closely related to MRP1 and categorized into CFTR/MRP subfamily, also called ABCC subfamily (13). MRP1, exhibiting a multidrug resistance phenotype in tumor cell lines, shares numerous substrates with MRP2, including glutathione conjugates (14,15). MRP2 is predominantly expressed in canalicular membrane of hepatocyte (14), and also in the kidney, jejunum, and ileum. In contrast to MRP2, MRP1 is expressed in most tissues, but lower in liver, and localized at basolateral site in hepatocyte (6,8,9). This difference enhances the physiological significance of MRP2 on hepatobiliary excretion.
The mechanism of how MRP2 transports the substrates is beyond our sights. In the case of P-gp, predicted transmembrane (TM) regions appeared to be particularly important for its function because the sites for interaction with substrates are embedded in lipid bilayer (16,17). Labeling of P-gp with photoactive analogs of drug substrates (18 -21), cysteine-scanning mutagenesis studies using a thiol-reactive substrate (22), and mutational analysis (23)(24)(25)(26) suggest that TM6, TM11, and TM12 are particularly important for drug-protein interaction. In MRP1-related transporters, the role of the N-terminal transmembrane regions of MRP1 was reported (27,28), but the residues involved in substrate-binding have not been identified.
In this study, we tested whether the basic residues around TM6 -TM17 of MRP2 participate in binding of substrates. We constructed 13 mutants by site-directed mutagenesis, and these mutants were expressed in COS-7 cells to test transport activity. The results suggested that Lys 324 , Lys 483 , Arg 1210 , and Arg 1257 are involved in substrate-protein interaction, whereas Lys 578 may be involved in stable expression of MRP2.
Cells and Vectors-Human MRP2 expression vector, pCI-neo/MRP2, and CHO-K1 cells expressing human MRP2 (29), CHO/MRP2, were kindly provided by Dr. Kuwano (Kyusyu Universisty, Kyusyu, Japan). In brief, NotI-tagged cDNA fragment encompassing the whole open reading frame of human MRP2 was cloned in the NotI site of mamma-lian expression vector pCI-neo (Promega). CHO-K1 cells were transfected with pCI-neo/MRP2 and the stable clone CHO/MRP2 was selected by 800 g/ml G418. For mutagenesis, we constructed pCI-neo/ MRP2X, which contains a deletion of XhoI site in the multicloning site of pCI-neo, and pCI-neo/MRP2V, which contains I448V and L653V mutations by overlapping polymerase chain reaction mutagenesis method to create Sse8387I and MluI sites, respectively.
Site-directed Mutagenesis-The seven mutants of MRP2, K316A,  K324A, K329A, H439A, K483A, K578A, and R590A, were generated by the overlapping polymerase chain reaction method. The other six mutants, R1023A, H1042A, R1100A, R1210A, R1230A, and R1257A, were generated by the method of Kunkel (30). For this, the 1758-base pair ApaI-Eco81I fragment of pCI-neo/MRP2, which encodes the C-terminal half encompassing 2449 -4207 amino acids of MRP2, was subcloned into pUC118EA, which has ApaI and Eco81I sites in the multicloning site, and mutations were introduced. Then the original ApaI-Eco81I fragment of pCI-neo/MRP2 vector was exchanged with the mutated one. Each mutation was confirmed by restriction enzyme analysis and DNA sequencing.
Cell Culture-CHO-K1 and CHO/MRP2 cells were cultured in Eagle's minimum essential medium (Life Technologies, Inc.) containing 10% fetal bovine serum, 100 g/ml kanamycin sulfate, and 100 U/ml penicillin in 5% CO 2 at 37°C. To keep MRP2 expression in CHO/MRP2, G418 was added at final concentration of 800 g/ml in the medium. COS-7 cells were routinely maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum in 5% CO 2 at 37°C.
Transient Expression of MRP2 in COS-7 Cells-The expression vectors, pCI-neo, pCI-neo/MRP2, pCI-neo/MRP2V, pCI-neo/MRP2X, and MRP2 mutants were transiently transfected into COS-7 cells using LipofectAMINE. Briefly, 4 ϫ 10 5 cells/well were seeded in 6-well plates, and 24 h later, cells were washed and overlaid with 1 ml of serum-free Dulbecco's modified Eagle's medium containing 1 g of supercoiled DNA and 6 l of LipofectAMINE. After 24 h of incubation, the medium was replaced with 2.5 ml of Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum.
Glutathione-methylfluorescein (GS-MF) and Glutathion-S-bimane (GS-B) Transport Assay-Excretion measurements were made 50 h after transfection. The excretion assay medium used in the assay was composed of 130 mM NaCl, 5 mM KCl, 1 mM MgSO 4 , 1.3 mM CaCl 2 , 1.2 mM KH 2 PO 4 , 19.7 mM HEPES, 5 mM glucose, pH 7.4. Cells were washed once with ice-cold excretion assay medium and incubated at 4°C for 60 min with ice-cold excretion assay medium containing 5 M CMFDA or 100 M mBCl. Thereafter, the cells were washed twice with ice-cold excretion assay medium and incubated at 37°C for 12 min in excretion assay medium. In CsA-directed inhibition assay, cells were incubated with culture medium containing 10 M CsA at 37°C for 30 min before preloading of CMFDA. 10 M CsA was also included in preloading and excretion period. At designated time, 100-l aliquots of the medium were collected, and the fluorescence of GS-MF and GS-B was determined by measuring the fluorescence at excitation of 490 nm and emission of 520 nm and at excitation of 380 nm and emission of 461 nm, respectively (SPECTRA MAX GEMINI; Molecular Devices). At the end of the experiment, the cells were solubilized by adding 1 ml of 0.1% Triton X-100 in PBS, and the fluorescence of cell lysate was also measured. The total amount of cell proteins was determined by the method of Bradford (31) using bovine serum albumin as a standard.
Study of Intracellular Accumulation of GS-MF-Cells seeded on coverslips were washed once with excretion assay medium and incubated at 37°C for 15 min with excretion assay medium containing 2.5 M CMFDA. Thereafter the cells were washed twice with excretion assay medium and incubated at 37°C for 30 min. To visualize GS-MF accumulation, coverslips were mounted on slide glass and examined with a fluorescence microscope (model IX70; Olympus, Tokyo, Japan).
Anti-MRP2 Antiserum-Anti-MRP2 antiserum was established by the immunization of rabbit (New Zealand White) with FYFMAKEAGI-ENVNSTKF peptide of MRP2 corresponding to C-terminal amino acids 1528 -1545.
Immunoblot Analysis of MRP2 Expression-COS-7 cells transfected with MRP2 expression vector were disrupted in RIPA buffer (10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 10 m/ml aprotinin) and centrifuged at 15,000 rpm for 20 min. The supernatant was recovered, and the protein concentration was determined by the method of Bradford (31). 50 g of protein was applied to SDS-polyacrylamide gel electrophoresis of 7.5% gel concentration. After electroblotting to polyvinylidene fluoride membrane, the membrane was blocked with 3% skim milk in Tris-buffered saline with 0.1% Tween 20 (TBST) and further probed with 500-dilution of anti-MRP2 antiserum diluted with 3% skim milk in TBST. Alkaline phosphatase-conjugated goat anti-rabbit antibody at 1,000-dilution was used as a secondary antibody, and the signal was detected with chromogenic substrate, 5-bromo-6-chloro-3-indolyl phosphate, and nitroblue tetrazolium.
The Immunofluorescence Cytochemical Detection of MRP2 Proteins-The cells cultured on coverslips were fixed in PBS containing 4% paraformaldehyde for 30 min at room temperature. They were washed with PBS and were permeablized with 0.1% Triton X-100 in PBS. After washing, blocking was made with 2% normal goat serum in 0.1% bovine serum albumin in PBS for 30 min at room temperature. They were then incubated for 1 h at room temperature with 100 l of rabbit anti-MRP2 C terminus antiserum diluted 1:100 with PBS and then washed and incubated for 1 h at room temperature with 100 l of fluorescein isothiocyanate-conjugated goat anti-rabbit IgG diluted 1:100 with PBS. Then the cells were mounted on slide glass and examined with a fluorescence microscope (model IX70; Olympus, Tokyo, Japan) or confocal laser-scanning microscope (model LSM510; Carl Zeiss Co., Ltd.)

Visualization of Intracellular Accumulation of GS-MF with
Fluorescence Microscope-We have examined MRP2-mediated excretion of a fluorescent substrate, GS-MF. GS-MF is produced by sequential modification of CMFDA within the cell cytosol. At first CMFDA, which can pass through the plasma membrane, is hydrolyzed by cytoplasmic esterase, and then the metabolite was conjugated with glutathione by glutathione S-transferase (32). The resulting GS-MF is now dependent on specific transporter like MRP2 to go across the membrane. This was shown in Fig. 1. COS-7 cells transfected either with pCIneo/MRP2 or pCI-neo, CHO-K1, and CHO/MRP2 were seeded on coverslips and incubated with 2.5 M CMFDA. After washing, the cells were further incubated with CMFDA-free medium. In CHO/MRP2 and COS-7 cells transfected with pCIneo/MRP2, intracellular fluorescence was apparently lower than that of control CHO-K1 and vector-transfected COS-7 cells (Fig. 1). This result reveals that GS-MF is actively excreted by pCI-neo/MRP2 transfection in these cell systems.
Quantitative Assay of MRP2-mediated Transport Activity-The excretion of GS-MF and GS-B from intact cells or MRP2transfected cells was examined by using fluorescent substrate, CMFDA, and mBCl. Like GS-MF, GS-B is also a fluorescent glutathione conjugate synthesized within the cells from mBCl. Cells preloaded with 5 M CMFDA or 100 M mBCl were incubated in medium without these fluorescent substrate, and the excreted GS-MF or GS-B in the medium was quantitated every 3 min. In CHO/MRP2 and COS-7 cells transfected with pCI-neo/MRP2, the excretion rate of GS-MF into the medium was 3.7-and 3.2-fold higher than that of the control CHO-K1 and COS-7 cells with vector alone (t ϭ 12 min), respectively (Fig. 2). The excretion of GS-B into the medium from CHO/ MRP2 and COS-7 cells transfected with pCI-neo/MRP2 was 3.5-and 2.6-fold higher than that of the control cells (t ϭ 12 min), respectively (Fig. 3). Because of the excretion of GS-MF and GS-B during preloading period with their precursors at 4°C, the amount of accumulated GS-MF and GS-B in the cells expressing MRP2 was about 30% less than that of control cells at the starting point (t ϭ 0 min) or after incubation period (t ϭ 9 or 12 min) of excretion measurement (data not shown). Therefore, the excreted amount of GS-MF and GS-B was normalized by the amount of GS-MF and GS-B remained in the cells after excretion measurement (t ϭ 9 or 12 min).
Mutagenesis and Expression in COS-7 Cells-It was previously characterized that P-gp transports wide variety of cationic substrates (33). The mechanism of how those substrates could be recognized by P-gp was speculated to be due to an interaction of negatively charged amino acid residues in transmembrane region of P-gp with cationic substrates (34). Conversely, it is supposed that positively charged amino acid residues are important in anionic substrate transporting MRP2.
To explore this possibility, we constructed MRP2 mutants that are carrying mutations at basic amino acid residues around transmembrane regions by site-directed mutagenesis. We replaced each of 13 basic residues (Lys 316 , Lys 324 , Lys 329 , His 439 , Lys 483 , Lys 578 , Arg 590 , Arg 1023 , His 1042 , Arg 1100 , Arg 1210 , Arg 1230 , and Arg 1257 ) around transmembrane regions from TM6 to TM17 with alanine (Fig. 4). The wild type or mutant MRP2 was transiently expressed in COS-7 cells, and the expression amount was compared by immunoblot analysis (Fig.  5) and immunocytostaining (Fig. 6) with rabbit anti-MRP2 C terminus antiserum. The expression amount of each mutants was consistent in both immunoblot and immunocytochemical analyses. Except for K578A, all of the mutants were expressed at nearly the same level when compared with wild type MRP2. Only the expression of K578A was faint in both immunoblot and immunocytochemistry.

GS-MF Excretion of Mutant MRP2
Proteins-First of all, we examined the excretion of GS-MF from COS-7 cells transfected with pCI-neo/MRP2V, which contains conservative amino acid changes as I448V and L653V, or pCI-neo/MRP2X, which contains a small deletion at the multi-cloning site of the vector. MRP2X-transfected cells retained almost the same GS-MF excretion activity with wild type MRP2, and MRP2V-transfected cells also retained the comparable activity (70% of wild type MRP2), indicating that the mutation did not severely affect the MRP2 function (Fig. 7). Next, we studied transport activity of MRP2 mutant proteins. Fig. 8 shows the excretion of GS-MF from COS-7 cells transfected with MRP2 mutants of TM6 -TM11. In K324A and K483A mutants, the excretion of GS-MF decreased about 40% compared with MRP2V. As to K578A mutant, GS-MF excretion decreased to the level of control (Fig.  8); however, expression level of this mutant also decreased (Figs. 5 and 6). Fig. 9 also presents the excretion of GS-MF from COS-7 cells transfected with MRP2 mutants of TM12-TM17. In R1210A and R1257A mutants, the excretion of GS-MF decreased approximately to the level of control, but expression level was comparable with MRP2X (Figs. 5 and 6). As shown in Fig. 10, we observed with confocal fluorescence microscopy that these mutant MRP2s including K578A were expressed at the cell surface. Therefore, the defect in the transport function of these mutants was clearly due to the functional defect except for K578A. The decrease in transport activity of K578A was mainly due to the decrease in the expression level.
Inhibition of GS-MF Excretion by CsA in MRP2 Mutants-We further examined whether basic residues of transmembrane regions are involved in CsA-directed inhibition of the substrate transporting activity of MRP2. Cells expressing active MRP2 mutants were incubated with 10 M CsA, and extrusion of GS-MF was measured in the existence of 10 M CsA. In R1230A mutant, the MRP2-mediated transport activity was not inhibited by CsA, whereas the other active mutants were suppressed on their transport activity up to 50% (Fig. 11). DISCUSSION According to the fact that MRP2 is involved in excretion of amphiphilic anionic conjugates, it has been predicted that positively charged amino acid residues in transmembrane helices are involved in substrate binding of MRP2. MRP1 and its related transporters have 17 transmembrane helices; however, the N-terminal five helices (TM1-TM5) have been shown to be dispensable for substrate translocating activity on MRP1 (28). In conjunction with other drug-extruding transporters, the core unit essential for transport activity seems limited on the tandem repeated structure of six-transmembrane helices and a nucleotide-binding domain. Hence, we tested whether the basic residues around transmembrane regions of TM6 -TM17 of MRP2 participate in binding of substrates by site-directed mutagenesis and transport activity assay.
First, we established a method for assaying transport activity of MRP2 by using fluorescent substrates, GS-MF and GS-B, which are glutathione-conjugated metabolites derived from CMFDA and mBCl, respectively. In CHO-K1 and COS-7 cells transfected with a MRP2 expression vector, the excretion of GS-MF and GS-B into the medium was 3-4-fold higher than that of the cells transfected with vector alone (Figs. 2 and 3). Although Evers et al. (35) reported that in nonpolarized cells exogenous MRP2 was not efficiently expressed at cell surface, there was no difference observed between wild type and the mutant MRP2s in their cell surface expression. The results indicated that the method developed here is useful to measure the transport activity of MRP2 and its mutants in these living cells.
For mutagenesis, we constructed human MRP2 expression vectors, pCI-neo/MRP2V and pCI-neo/MRP2X. pCI-neo/ MRP2X-transfected COS-7 cells retained almost the same GS-MF excretion activity with the wild type, and pCI-neo/ MRP2V-transfected COS-7 cells also retained the comparable activity, 70% of the wild type (Fig. 7). pCI-neo/MRP2V contains mutations of I448V and L653V, which were located in TM8 and NBD1, respectively. In patients with Dubin-Johnson syndrome, loss of function mutations was occasionally found in NBD1 (12). Likewise, the mutation of NBD1 in pCI-neo/ MRP2V might cause slightly decreased activity of MRP2Vmediated transport.
To gain insight into the substrate-binding site of MRP2, we constructed 13 mutants, each of which was replacing basic amino acid residue by alanine. Except for a mutant K578A, the protein expression level of these mutants in COS-7 cell was comparable with that of wild type (Figs. 5 and 6). Among these mutants, K578A, R1210A, and R1257A were the most influenced by mutation concerning substrate translocating activity FIG. 6. Fluorescent immunocytostaining of MRP2 and its mutants. COS-7 cells transfected either with wild type or mutant MRP2 expression vector were exposed to the rabbit anti-MRP2 C terminus antiserum after permeabilization with 0.1% Triton X-100. Cells were then incubated with fluorescein isothiocyanate-conjugated goat anti-rabbit antibody and were photographed using a fluorescence microscope. The photograph was taken under the same conditions. Exposure time was 0.13 s. ( Figs. 8 and 9). The pattern of cytochemical staining of these mutants was not changed with wild type MRP2 (Fig. 10). These results suggested that Arg 1210 and Arg 1257 are important residues for substrate transport activity of MRP2. As to K578A, the excretion of GS-MF and protein expression was coincidentally lowered (Figs. 5, 6, and 8). Therefore, it was thought that Lys 578 has a role in protein expression rather than substrate translocation. In K324A and K483A mutants, GS-MF excretion also moderately decreased (Fig. 8). Protein expression level (Figs. 5 and 6) and membrane localization (Fig. 10) were nor-mal in these mutants, so that it is indicated that Lys 324 and Lys 483 are also involved in substrate transport function of MRP2. Lys 324 , Lys 483 , Arg 1210 , and Arg 1257 are located in TM6, TM9, TM16, and TM17, respectively. In the case of P-gp, TM1, TM6, TM11, and TM12 are important in recognition and binding of substrate (22,36,37). TM6, TM16, and TM17 of MRP2 correspond to TM1, TM11, and TM12 of P-gp. Together with these findings, it is suggested that Lys 324 , Lys 483 , Arg 1210 , and Arg 1257 of MRP2 are constituent of substrate binding site. Although there is a difference in the chemistry of residues involved in substrate transport, it is intriguing to speculate that the functional transmembrane helices of MRP2 and P-gp for substrate binding are almost common.
Next, we examined whether basic residues of TMs are also involved in CsA inhibition. CsA is known to inhibit MRP2 transport activity through binding to but not to be transported by MRP2 (38). The previous study has shown that 10 M CsA causes drug accumulation in MRP2 expressing LLC-PK1 cells as much as the level of control LLC-PK1 (29). Hence, we tried to see the effect of 10 M CsA on mutant MRP2 expressors. In most active mutants excretion of GS-MF was inhibited by CsA as well as wild type MRP2, whereas in R1230A mutant excretion of GS-MF was not influenced (Fig. 11). This result clearly indicated that Arg 1230 is an essential residue for the inhibition of CsA, either through direct binding of CsA or coordinating to block substrate translocation with CsA bound on another site. Interestingly, in contrast to the fact that both Arg 1210 and Arg 1257 are located at intracellular side of TM16 and TM17, respectively, Arg 1230 is located at extracellular side of TM16. So far determined in P-gp, important residues are located at intracellular side (39). The unique location of Arg 1230 may indicate a novel functional relationship on CsA inhibition site and substrate translocation pathway of MRP2.
In summary, the results of this study showed that the basic residues in TM6, TM9, TM16, and TM17 are involved in substrate binding and a residue in TM11 is involved in stable expression of MRP2. This is the first indication that transmembrane basic residues are truly important for substrate translocation on MRP2. Like P-gp, these functional residues of MRP2 are located in intracellular side of the corresponding transmembrane helices. In this study, we used glutathione conjugate as a substrate for MRP2. Further studies are required to determine the role of basic residues by using other substrates, glucronate, and sulfate conjugates.
FIG. 10. Immunolocalization in COS-7 by confocal laser scanning microscope. COS-7 cells transfected either with expression vector for wild type MRP2 or K324A, K483A, K578A, R1210A, and R1257A mutants were immunostained using anti-MRP2 C terminus antiserum, and their localization was observed by confocal laser scanning microscope. A 1.2-m-thick section was observed. The ring-like staining seemed MRP2 expression at plasma membrane and cytoplasmic periphery of the membrane.
FIG. 11. Effect of CsA on the excretion of GS-MF from active COS-7/ MRP2 mutants. COS-7/MRP2 mutants, which retain transport activity of GS-MF, were incubated at 37°C with (q) or without (E) 10 M CsA. After 30 min, the cells were incubated at 37°C with (q) or without 10 M CsA (E) and 5 M CMFDA for 60 min. Thereafter, the cells were washed and incubated at 37°C with excretion medium with (q) or without 10 M CsA (E) for 9 min. The amount of GS-MF in the medium was divided by the amount of intracellular GS-MF at 9 min. The experiments were performed in triplicate.