The functional expression of sodium-dependent bile acid transport in Madin-Darby canine kidney cells transfected with the cDNA for microsomal epoxide hydrolase.

Previous studies have suggested that the enzyme microsomal epoxide hydrolase (mEH) is able to mediate sodium-dependent transport of bile acids such as taurocholate into hepatocytes (von Dippe, P., Amoui, M., Alves, C., and Levy, D. (1993) Am. J. Physiol. 264, G528-G534). In order to characterize directly the putative transport properties of the enzyme, a pCB6 vector containing the cDNA for this protein (pCB6-mEH) was transfected into Madin-Darby canine kidney (MDCK) cells, and stable transformants were isolated that could express mEH at levels comparable with the levels expressed in hepatocytes. Sodium-dependent transport of taurocholate was shown to be dependent on the expression of mEH and to be inhibited by the bile acid transport inhibitor 4,4′-diisothiocyanostilbene-2,2′disulfonic acid (DIDS), as well as by other bile acids. Kinetic analysis of this system indicated a Km of 26.3 μM and a Vmax of 117 pmol/mg protein/min. The Km value is essentially the same as that observed in intact hepatocytes. The transfected MDCK cells also exhibited sodium-dependent transport of cholate at levels 150% of taurocholate in contrast to hepatocytes where cholate transport is only 30% of taurocholate levels, suggesting that total hepatocyte bile acid transport is a function of multiple transport systems with different substrate specificities, where mEH preferentially transports cholate. This hypothesis is further supported by the observation that a monoclonal antibody that partially protects (26%) taurocholate transport from inhibition by DIDS in hepatocytes provides almost complete protection (88%) from DIDS inhibition of hepatocyte cholate transport, suggesting that taurocholate is also taken up by an alternative system not recognized by this antibody. Additional support for this concept is provided by the observation that the taurocholate transport system is almost completely protected (92%) from DIDS inhibition by this antibody in MDCK cells that express mEH as the only bile acid transporter. These results demonstrate that mEH is expressed on the surface of hepatocytes as well as on transfected MDCK cells and is able to mediate sodium-dependent transport of taurocholate and cholate.

Previous studies have suggested that the enzyme microsomal epoxide hydrolase (mEH) is able to mediate sodium-dependent transport of bile acids such as taurocholate into hepatocytes (von Dippe, P., Amoui, M., Alves, C., and Levy, D. (1993) Am. J. Physiol. 264, G528 -G534). In order to characterize directly the putative transport properties of the enzyme, a pCB6 vector containing the cDNA for this protein (pCB6-mEH) was transfected into Madin-Darby canine kidney (MDCK) cells, and stable transformants were isolated that could express mEH at levels comparable with the levels expressed in hepatocytes. Sodium-dependent transport of taurocholate was shown to be dependent on the expression of mEH and to be inhibited by the bile acid transport inhibitor 4,4-diisothiocyanostilbene-2,2disulfonic acid (DIDS), as well as by other bile acids. Kinetic analysis of this system indicated a K m of 26.3 M and a V max of 117 pmol/mg protein/min. The K m value is essentially the same as that observed in intact hepatocytes. The transfected MDCK cells also exhibited sodium-dependent transport of cholate at levels 150% of taurocholate in contrast to hepatocytes where cholate transport is only 30% of taurocholate levels, suggesting that total hepatocyte bile acid transport is a function of multiple transport systems with different substrate specificities, where mEH preferentially transports cholate. This hypothesis is further supported by the observation that a monoclonal antibody that partially protects (26%) taurocholate transport from inhibition by DIDS in hepatocytes provides almost complete protection (88%) from DIDS inhibition of hepatocyte cholate transport, suggesting that taurocholate is also taken up by an alternative system not recognized by this antibody. Additional support for this concept is provided by the observation that the taurocholate transport system is almost completely protected (92%) from DIDS inhibition by this antibody in MDCK cells that express mEH as the only bile acid transporter. These results demonstrate that mEH is expressed on the surface of hepatocytes as well as on transfected MDCK cells and is able to mediate sodiumdependent transport of taurocholate and cholate.
Numerous studies have been concerned with characterizing the hepatocyte membrane transport systems that mediate the sodium-dependent uptake of organic anions such as bile acids. These compounds play a critical role in numerous physiological processes such as digestion, biliary excretion of cholesterol and metabolites of exogenous species such as carcinogens, and regulation of cholesterol metabolism (1,2). These systems have been studied in perfused liver, intact hepatocytes, and sinusoidal plasma membrane vesicles. Kinetic studies have suggested the parallel functioning of several distinct sodium-dependent organic anion transporters with different substrate specificity profiles (3)(4)(5).
Two unrelated proteins with similar apparent molecular masses (ϳ50 kDa) have been proposed as putative sodium-dependent bile acid transporters. One of these proteins (Ntcp) was identified using expression cloning in Xenopus laevis oocytes and was shown to mediate sodium-dependent taurocholate transport in oocytes as well as in transfected COS-7 cells with kinetic properties consistent with those observed in hepatocytes (6,7). Studies utilizing photoaffinity labeling procedures (8,9), monoclonal antibodies (mAb) 1 (10), proteoliposome reconstitution (11), and correlation of transport with protein expression (12) indicated the involvement of an additional protein in the transport process. This protein was isolated and subsequently shown to be indistinguishable from the enzyme microsomal epoxide hydrolase (mEH) (EC 3.3.2.3) (13), which also plays a central role in the metabolism of polycyclic aromatic hydrocarbon carcinogens (14). It is pertinent to note that several other proteins such as dihydrodiol dehydrogenase (15), glutathione S-transferase (15,16), and members of the cytochrome P-450 enzyme superfamily (17,18) are also involved together with mEH in the metabolism of polycyclic aromatic hydrocarbon carcinogens, as well as mediating a number of physiological processes by binding various bile acids or bile acid derivatives.
The expression of mEH on the hepatocyte sinusoidal surface in addition to its well documented location in the endoplasmic reticulum raises a number of provocative questions concerning the mechanism of membrane protein targeting. Evidence to support the surface location of this protein was first suggested by studies using a variety of enzymatic markers (19) and subsequently confirmed by the demonstration that a mAb reacting with mEH could block labeling of sinusoidal plasma membranes in intact hepatocyte as well as partially protect hepatocytes from inhibition of taurocholate transport by the anion transport inhibitor, 4,4Ј-diisothiocyanostilbene-2,2Ј disulfonic acid (DIDS) (10). Confocal immunofluorescence analysis of intact hepatocytes using a mAb against mEH also demonstrated * This investigation was supported by Grant DK 25836 from the National Institutes of Health. 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 expression of this protein on the hepatocyte surface. 2 In addition to mEH, several cytochrome P-450 enzymes have been shown to be expressed on the surface of hepatocytes as well as the endoplasmic reticulum (20 -22) with an orientation based on epitope mapping that would suggest the existence of two alternate topologies in the endoplasmic reticulum. Recent studies have demonstrated directly or indirectly that several proteins such as mEH (23), P-glycoprotein (24,25), several cytochrome P-450 enzymes (20), and ductin (26) can exhibit two or more distinct topological orientations generated from a single population of nascent chains. Alternate protein topologies may thus result in differential targeting of these alternate forms, resulting in the expression of some proteins at multiple cell sites.
In order to demonstrate directly the transport properties of mEH, which were suggested by previous studies on hepatocytes, and to assess its function in the absence of Ntcp, we have employed cDNA transfection procedures, which have established that this bifunctional protein, when expressed on the plasma membrane, is able to mediate the sodium-dependent uptake of taurocholate and cholate into transfected Madin-Darby canine kidney (MDCK) cells.

EXPERIMENTAL PROCEDURES
Construction of Expression Vector pCB6-mEH-The cDNA for mEH was prepared from rat liver poly(A) ϩ mRNA using reverse transcriptase and the polymerase chain reaction amplification procedure (27) and subsequently ligated into the BglII/KpnI sites of the pCB6 vector under the control of the butyrate-inducible cytomegalovirus promoter for expression in MDCK cells. Synthetic oligonucleotide primers for the polymerase chain reaction were targeted toward sequences in the 5Ј-and 3Ј-untranslated regions of mEH beginning at position 85 and ending at position 1582 in the published cDNA sequence (28). The primers were also designed to introduce new restriction sites for subsequent use in the construction of expression vector pCB6-mEH by standard techniques (29). The fidelity of vector constructs was verified by restriction mapping and by DNA sequencing of the cDNA insert using the dideoxy procedure (29).
Cell Culture and Transfection Procedures-MDCK (NLB-2) cells were purchased from the American Type Culture Collection and grown at 37°C in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. MDCK cells were transfected with pCB6-mEH using calcium phosphate precipitation procedures (30). Stable transformants were obtained by selecting in the presence of G-418 (500 g/ml). After 3 weeks numerous resistant clones were obtained that expressed mEH. The transfected cells were subsequently treated with sodium butyrate (10 mM) for 20 h as described previously (31) to increase the expression of mEH. Transfected MDCK cells expressing tyrosine aminotransferase (32) instead of mEH as well as untransfected cells were used as controls.
Transport of Taurocholate and Cholate by MDCK Cells and Hepatocytes-MDCK cells were removed from tissue culture dishes (100 mm) by initial treatment with 4.5 ml of a 0.002% trypsin solution in phosphate-buffered saline (PBS), pH 7.4, for 5 min at 37°C, followed by washing with 2 ml of medium/fetal bovine serum followed by 10 ml of PBS. The cells were then treated with 4.5 ml of EDTA (0.07-0.15 mM) in PBS for 5 min at 37°C so that the cells could be detached by moderate pipetting of this solution on the surface of the monolayer. The amount of EDTA used was dependent on the particular clone and on the time under culture conditions. Detached cells were washed in PBS, 10 mM MgCl 2 , 2 mM CaCl 2 by centrifugation for 2 min at 1000 ϫ g, and the resultant pellets were suspended in either 140 mM NaCl or 140 mM choline chloride in 5 mM MgCl 2 , 2 mM CaCl 2 , 10 mM Tris, pH 7.4. The combined pellets derived from five plates were suspended in 800 l of either of these buffers. Uptake was measured by adding 125 l of buffer containing sodium or choline chloride with various amounts of [ 3 H]taurocholate or [ 3 H]cholate as well as unlabeled bile acid to 125 l of buffer containing approximately 2 ϫ 10 6 MDCK cells. Transport was measured at 37°C for the indicated time periods at various concentrations. Uptake was terminated by the rapid addition of 1 ml of buffer at 4°C, and the cell suspension was isolated by rapid centrifugation through light mineral oil and dibutyl phthalate (1:4). The surface of the pellet was washed with 0.4 ml of buffer followed by suspension of the pellet in 0.3 ml of 0.1 N NaOH, and the associated radioactivity was evaluated. These values were corrected for radioactivity associated with the cell pellet when cells in 1 ml of buffer at 4°C were added to the labeled substrate and immediately centrifuged. Inhibition studies using bile acids were affected by the simultaneous addition of these compounds with the particular substrate. DIDS inhibition was evaluated by preincubating the cells with this reagent (100 M) for 20 min at 37°C prior to uptake measurements. The effect of mAb 25A-3 (1 mg/ml) on DIDS inhibition of bile acid transport was evaluated by preincubating the cells with the antibody for 30 min at 37°C prior to the addition of DIDS.
Hepatocytes were obtained from the livers of Sprague-Dawley rats using a collagenase perfusion technique (35). Cell viability throughout the course of the experiments was estimated by trypan blue exclusion. All studies were carried out on hepatocyte suspensions with an 85-93% viability. The effects of DIDS and mAb 25A-3 on the transport of taurocholate and cholate were determined as described previously (10).

Transfection of MDCK Cells and Expression of mEH-
The expression vector pCB6-mEH was transfected into MDCK cells, and stable transformants were selected based on G-418 resistance. Clones were tested for the expression of mEH by immunoprecipitation with mAb 25D-1 as shown in Fig. 1. Cells transfected with pCB6 lacking the mEH insert (Fig. 1, lane c) or untransfected cells (data not shown) did not express mEH. Transfection with PCB6-mEH followed by induction with butyrate gave the correct immunoprecipitate (Fig. 1, lane b). Butyrate treatment resulted in an approximately 10-fold increase in mEH expression to levels comparable with those found in hepatocytes (Fig. 1, lane a). In addition to the close correspondence of mEH levels and molecular weight in MDCK cells compared with hepatocytes, the specific enzymatic activity of mEH was also shown to be 85% of that observed in hepatocytes. Several of these clones were then tested for sodiumdependent bile acid transport capacity. The functional integrity of the transfected cells was shown to be extremely sensitive to the amounts of EDTA and trypsin used to remove the cells from the plate prior to transport measurements.
Bile Acid Transport Analysis-Cells transfected with pCB6 without the mEH insert as described above resulted in negligible sodium-dependent transport as shown in Fig. 2 (lane 1). Similar results were obtained with pCB6-mEH transfected cells at 4°C or at 37°C without butyrate induction and when using untransfected MDCK cells (data not shown). Untrans- fected cells did exhibit an undefined bile acid association that was identical in the presence and the absence of Na ϩ (20 and 4 pmol/mg/min for 5 and 1 M taurocholate and 8 pmol/mg/min for 1 M cholate) and was also insensitive to DIDS or bile acid derivatives. Transfected MDCK cells, in the absence of Na ϩ , also exhibited this background, which was unaffected by DIDS, bile acid derivatives, and serum albumin. The functional expression of sodium-dependent transport of taurocholate by MDCK cells was dependent on the expression of mEH (Fig. 2,  lane 2). Treatment of these cells with 0.2% trypsin for 10 min at 37°C resulted in a 90% loss of transport capacity (data not shown). Sodium-dependent taurocholate transport was also significantly inhibited by the bile acid transport inhibitor DIDS (Fig. 2, lane 4) and by other bile acids where cholate was a considerably more effective inhibitor than taurochenodeoxycholate (Fig. 3). The inhibitory efficiency of these bile acids on taurocholate transport is reversed from that observed in isolated hepatocytes (5) or COS-7 cells transfected with Ntcp (6), suggesting the existence of a different substrate profile for Ntcp and mEH. Taurocholate uptake was also significantly inhibited by bovine serum albumin (75%) (Fig. 2, lane 5) in contrast to the stimulatory effect of this protein on taurocholate uptake mediated by Ntcp (4). A similar inhibitory effect by bovine serum albumin has, however, been observed for bumetanide uptake in oocytes, mediated by a protein that also transports taurocholate but is different from Ntcp (4). The relationship of the bumetanide transport protein and mEH remains to be elucidated. The time course for sodium-dependent uptake of taurocholate by the transfected MDCK cells, shown in Fig. 4 is linear for at least 30 s. The initial transport rates (15 s) as a function of substrate concentration (1-50 M) are shown in Fig.  5. Analysis of these results on an Eadie-Scatchard plot indicated a K m of 26.3 M and a V max of 117 pmol/mg protein/min. This K m value is essentially the same as that observed for taurocholate uptake in hepatocytes (1) as well as in oocytes and COS-7 cells expressing Ntcp (6,7).
In addition to mediating taurocholate transport, MDCK cells transfected with pCB6-mEH exhibited sodium-dependent transport of cholate, which could also be inhibited by DIDS (Fig. 6). In contrast to hepatocytes, where sodium-dependent cholate uptake is approximately 30% of taurocholate uptake (3,5), cholate transport in the MDCK cell system (Fig. 6, lane 2) was 50% higher than taurocholate transport as shown in Fig. 3 (lane 1). Estimates from available data in the literature suggest that Ntcp expressed in oocytes primarily transports taurocholate (36). These observations would indicate that the bile acid transport characteristics observed in hepatocytes may thus be comprised of contributions from Ntcp as well as mEH and possibly other as yet to be defined transport systems.
Previous studies have demonstrated that mAb 25A-3, although incapable of directly inhibiting taurocholate transport, was able to protect hepatocytes from DIDS inhibition of tauro- cholate uptake by 26% (10). This low protection value suggests that there may be other transporters, such as Ntcp, that mediate taurocholate transport but that are not recognized by this antibody. To further probe this possibility, the MDCK cells transfected with pCB6-mEH and exhibiting sodium-dependent taurocholate transport capacity were also pretreated with mAb 25A-3. As shown in Table I, the observed DIDS inhibition (also shown in Fig. 2, lane 4) was now decreased by 92%, demonstrating that the antibody was effective in protecting the cell in question from DIDS inhibition, and in the case where mEH is the only functioning transporter, protection is essentially complete. In hepatocytes, where multiple transporters exist, protection is less pronounced ( Table I). The above results suggest that the mEH system may be the primary system for sodiumdependent cholate transport in hepatocytes. In this regard, the ability of mAb 25A-3 to block DIDS inhibition of hepatocyte cholate transport was also investigated. As shown in Table I, this antibody was able to protect cholate transport by 88% as compared with 26% for taurocholate, further supporting the concept that sodium-dependent cholate transport is mediated primarily by mEH. DISCUSSION Previous studies have suggested that mEH is able to mediate sodium-dependent bile acid transport into hepatocytes (10 -13). In order to further establish this functional role, MDCK cells were transfected with pCB6-mEH, which resulted in permanent transformants that expressed mEH (Fig. 1), as well as sodium-dependent bile acid transport (Fig. 2). These properties are not found in untransfected MDCK cells or in cells transfected without the mEH insert and are greatly reduced in cells expressing low levels of mEH prior to butyrate induction. Transport was shown to be stimulated by Na ϩ and significantly inhibited by the bile acid transport inhibitor DIDS as well as other bile acids (Figs. 2 and 3). The physiological significance of this transporter is supported by the Eadie-Scatchard analysis of the substrate concentration curve (Fig. 5), which indicated a K m value of 26.3 M, which is similar to those values previously reported for hepatocytes (1) as well as for COS-7 cells transfected with Ntcp (6). In contrast, the substrate specificities of these two systems are different, because mEH preferentially transports cholate (Fig. 6), whereas Ntcp transports primarily taurocholate (36).
The possibility of more than one sodium-dependent transport system has also been supported by kinetic studies, suggesting the functional contribution of two sodium-dependent taurocholate transport systems, one of which also transports cholate (3,5). Studies utilizing X. laevis oocytes have also suggested the existence of a mRNA fraction that upon oocyte injection results in the expression of a protein that mediates the sodium-dependent uptake of cholate and bumetanide as well as taurocholate. Oocytes expressing Ntcp, in contrast, preferentially transport taurocholate, have a reduced transport capacity for cholate and a negligible uptake of bumetanide, clearly establishing the existence of two distinct transporters (4). Further evidence that the observed sodium-dependent uptake of bile acids is a result of the contribution from more than one transport system is derived from the antibody protection experiments described in Table I. The partial protective effect (26%) of mAb 25A-3 on the DIDS inhibition of taurocholate transport in hepatocytes suggests the parallel functioning of another transport system that is not affected by this antibody. In contrast, mAb 25A-3 afforded at least 92% protection from DIDS inhibition of taurocholate transport in the MDCK cells, which establishes that mEH is expressed on the surface of MDCK cells following transfection, with a topology in this system that is the same as in hepatocytes as adjudged by the recognition of the 25A-3 epitope on the cell surface. In addition, the high level of effectiveness of this antibody in protecting taurocholate transport from DIDS inhibition when mEH is the only functional bile acid carrier, suggests that the 26% protection observed in hepatocytes may in fact roughly approximate the contribution of mEH to the total uptake of taurocholate. The large protective effect of this antibody (88%) on hepatocyte cholate transport is also in agreement with the observed preferential uptake of this bile acid by mEH and suggests that this system appears to be the primary physiological transporter for cholate and possibly other, as yet to be defined, substrates, complementing the physiological role of Ntcp as the primary transporter of taurocholate.
During development, mEH is expressed at 10% of adult levels at day 19 of gestation as assessed by immunoprecipitation procedures with mAb 25D-1 and is shown to reach adult levels approximately 50 -60 days after birth (12). This time course is also closely paralleled by the expression of mEH enzymatic activity and the increase in taurocholate transport capacity (37)(38)(39). A complementary study (40) exploring the identity of the sodium-dependent bile acid transporter was performed based on the observed ontogenic expression of a protein whose apparent molecular mass (48 kDa) and isoelectric point (approximately 9) were very close, if not indistinguishable from mEH (apparent molecular mass, 49 kDa; pI, 8.2). This protein was first detected at day 20 of gestation, and its intensity increased during post natal development at a rate corresponding to the ontogenic expression of sodium-dependent taurocholate transport activity. Monospecific, polyclonal antibodies raised against this protein directly inhibited transport by 50%, whereas proteoliposome reconstitution also implicated this protein in the transport process. Additional independent  evidence demonstrates that the photoaffinity labeling of a 50-kDa protein utilizing a photoreactive analog of taurocholate was significantly reduced from the levels observed in adult hepatocytes when using hepatocyte plasma membranes from 9-day-old rats (41). Taken together, our studies as well as those of Suchy and co-workers and Ziegler et al. (40,41) strongly implicate a 48 -50-kDa protein whose level of expression increases over a period of approximately 50 -60 days without a change in its molecular mass. The identity of the 48-kDa protein reported by Suchy has, however, not been published. This same group has reported an additional study (42) on the ontogenic expression of a sodium-dependent bile acid transporter represented by Ntcp as elucidated by Meier and coworkers (7), which is clearly not related to mEH or to the 48-kDa protein originally reported by Suchy and co-workers (40). Of particular note in this report is the fact that during development, Ntcp is expressed with a molecular mass of 39 kDa and is only converted to the adult form (50 kDa) by further glycosylation at some time after 4 weeks, clearly differentiating this processing from that observed for mEH. Surprisingly, the level of expression of Ntcp is 82% of adult levels on postnatal day 1 and reaches adult levels within the first week when the level of sodium-dependent taurocholate transport is only 25% of adult levels. Because there is no apparent change in the molecular weight or level of expression of Ntcp during the first 4 weeks, whereas transport capacity (V max ) increases dramatically during this time period (38), it appears that there is a significant discordance between the expression of Ntcp and sodium-dependent taurocholate transport during the first month of postnatal development. In contrast, the functional expression of sodium-dependent bile acid transport on the apical surface of the rat ileum during postnatal development (1-6 weeks) closely parallels the expression of the recently cloned ileum transporter for these substrates (43). A similar relationship is also observed for a hepatic sinusoidal sodium-independent organic anion transporter. 3 The relative contributions of mEH and Ntcp in mediating the uptake of taurocholate as well as other bile acids and nonbile acid substrates in different physiological states remains to be established. An antibody against Ntcp that could directly demonstrate the functional role of Ntcp in hepatocytes, as we have reported in this and other studies for mEH, would be useful in establishing their respective physiological roles.
Recently a study has been published that suggests an inability to correlate the expression of mEH in transfected BHK cells with the appearance of mediated transport for taurocholate, cholate, or bumetanide (44). Obtaining negative results in a particular heterologous expression system does not necessarily disprove a thesis. In the case of MDCK transfectants, numerous technical details needed to be established before a functional system was obtained. In this regard, we have demonstrated the sensitivity of transfected MDCK cells to trypsin treatment. Because Honscha et al. (44) have used this enzyme, albeit at an unstated concentration and time period, to remove the BHK cells from the plate prior to measuring transport, their procedures may have led to the inactivation of this transport system. Furthermore they have (a) failed to establish whether mEH was expressed on the surface of the transfected BHK cells; (b) reported that the enzymatic activity of mEH expressed in BHK cells was only 12% of the activity observed in hepatocytes; and (c) contradicted previous reports of one of the authors concerning the surface localization of mEH in hepatocytes (19). Their conclusions must therefore be questioned.
In conclusion, the results presented in this paper strongly support our previous studies and directly demonstrate the ability of mEH to mediate the sodium-dependent transport of taurocholate and cholate into MDCK cells and to participate with Ntcp in the net uptake of these substrates in hepatocytes. The membrane topology and cell targeting of this protein as well as its interaction with nonbile acid substrates such as steroids and carcinogens is under investigation.