Identification of Glutathione as a Driving Force and Leukotriene C4 as a Substrate for oatp1, the Hepatic Sinusoidal Organic Solute Transporter*

oatp1 is an hepatic sinusoidal organic anion transporter that mediates uptake of various structurally unrelated organic compounds from blood. The driving force for uptake on oatp1 has not been identified, although a role for bicarbonate has recently been proposed. The present study examined whether oatp1-mediated uptake is energized by efflux (countertransport) of intracellular reduced glutathione (GSH), and whether hydrophobic glutathioneS-conjugates such as leukotriene C4(LTC4) and S-dinitrophenyl glutathione (DNP-SG) form a novel class of substrates for oatp1. Xenopus laevisoocytes injected with the complementary RNA for oapt1 demonstrated higher uptake of 10 nm [3H]LTC4and 50 μm [3H]DNP-SG, and higher efflux of [3H]GSH (2.5 mm endogenous intracellular GSH concentration). The oatp1-stimulated LTC4 and DNP-SG uptake was independent of the Na+ gradient,cis-inhibited by known substrates of this transport protein and by 1 mm GSH, and was saturable, with apparentK m values of 0.27 ± 0.06 and 408 ± 95 μm, respectively. Uptake of [3H]taurocholate, an endogenous substrate of oatp1, was competitively inhibited by DNP-SG. Of significance, oatp1-mediated taurocholate and LTC4 uptake was cis-inhibited and trans-stimulated by GSH, and [3H]GSH efflux was enhanced in the presence of extracellular taurocholate or sulfobromophthalein, indicating that GSH efflux down its large electrochemical gradient provides the driving force for uptake via oatp1. The stoichiometry of GSH/taurocholate exchange was 1:1. These findings identify a new class of substrates for oatp1 and provide evidence for GSH-dependent oatp1-mediated substrate transport.

A major hepatic function is the clearance of a multitude of metabolic products and foreign compounds (xenobiotics) from the blood. Hepatic uptake of organic solutes across the sinusoidal membrane is mediated by multiple transport systems, and four transporters have now been cloned. Ntcp mediates Na ϩ -coupled transport of bile acids (1), oct1 mediates H ϩcoupled uptake of organic cations (2), and oatp1 and oatp2 are related multispecific transporters for which driving force has not yet been identified (3)(4)(5). A recent study indicates that HCO 3 Ϫ efflux may be involved in oatp1-mediated transport (6); however, because intracellular pH is lower than extracellular pH, the HCO 3 Ϫ gradient is unlikely to energize transport of organic anions on oatp1. Indeed, Satlin and co-workers (6) demonstrated that oatp1-mediated cellular uptake of taurocholate was accelerated only when the pH gradient was reversed from the normal physiological condition. Under all other pH gradient conditions examined by these investigators, taurocholate uptake was lower than that seen in controls, contradicting the hypothesis that the bicarbonate gradient energizes oatp1mediated transport.
oatp1-mediated transport is independent of ATP and of transmembrane Na ϩ , K ϩ , or H ϩ gradients (3)(4)(5). Transport is also independent of the Cl Ϫ gradient when measured in the absence of albumin (3). However, oatp1-mediated transport is bidirectional (7), suggesting that solute uptake may be energized by efflux of an intracellular solute. Whether oatp1 actually functions as an antiporter (exchanger) and which intracellular solute may be involved remain to be established.
One solute that could function as an intracellular substrate for oatp1 and related organic anion transporters is reduced glutathione (GSH). This anionic tripeptide is present in high concentrations in virtually all cells, including hepatocytes (ϳ10 mM), whereas plasma concentrations are only about 10 M (8 -10). The high GSH chemical gradient coupled with the negative intracellular membrane potential could provide a substantial driving force for uptake via oatp1.
Export of GSH into the extracellular space is the initial and perhaps limiting step in the turnover of the tripeptide in all mammalian cells (8 -10); however, the transport system or systems that mediate GSH efflux have not been identified. The previously reported cloning of sinusoidal and canalicular GSH transporters has not been confirmed (14) and appears to be an artifact (10,14). Rates of GSH efflux differ between cell types and are roughly equivalent to the rates of synthesis under physiological conditions. The liver is quantitatively the major site of synthesis in the body, and the major supplier of plasma GSH (8 -10). Studies in intact hepatocytes and isolated sinusoidal plasma membrane vesicles indicate that GSH efflux is mediated by a low affinity, ATP-independent, electrogenic carrier (11)(12)(13)(14). Transport is bidirectional, although under physiological conditions unidirectional transport down its concentration gradient is favored, i.e. efflux of GSH from hepatocytes (11)(12)(13)(14).
The present study also examined whether leukotriene C 4 (LTC 4 ) 1 is a substrate for oatp1. The cysteinyl leukotrienes are potent mediators of inflammatory responses (15), and the liver is the major site of their removal from the blood (15-18). The mechanism for their hepatic uptake is unknown, although evidence from isolated hepatocytes and perfused rat liver studies indicate that uptake is mediated by oapt1 (17)(18)(19). LTC 4 uptake into isolated rat hepatocytes and plasma membrane vesicles is independent of the Na ϩ gradient, is inhibited by substrates of oatp1, and exhibits an apparent K m of 0.20 M (19). Once taken up into hepatocytes, LTC 4 is transported across the canalicular membrane into bile by the ATP-dependent transporter mrp2 (which is also called cMRP or cMOAT). Hepatic uptake of another glutathione S-conjugate (2,4-dinitrophenylglutathione, DNP-SG) also displays characteristics of oatp1mediated transport (20). To examine the hypothesis that oatp1 mediates uptake of these glutathione S-conjugates, the present study measured LTC 4 and DNP-SG uptake in Xenopus laevis oocytes injected with oatp1 complementary RNA (cRNA). The results demonstrate that LTC 4 and DNP-SG are substrates for this multispecific organic solute transporter, and that hepatic uptake is energized by countertransport of intracellular GSH.
Synthesis of oatp1 and Ntcp cRNA-The cDNAs for the rat organic anion transporting polypeptide (oatp1) and the rat Na ϩ /taurocholate cotransporter (Ntcp) were prepared as described previously (1,3). Capped cRNA was transcribed in vitro with T3 or T7 RNA polymerase (Ambion, Austin, TX), and the reactions were terminated by digestion of the DNA templates with RNase-free DNase I. cRNAs were precipitated with lithium chloride and stored at Ϫ70°C until used for oocyte microinjection.
Isolation of Rat Liver mRNA-Total RNA from rat liver was prepared using the guanidinium thiocyanate/cesium chloride method (25), and mRNA was purified by chromatography with oligo(dT)-cellulose using an mRNA purification kit (Amersham Pharmacia Biotech).
Preparation and Microinjection of X. laevis Oocytes-Isolation of X. laevis oocytes was performed as described by Goldin (26) and as described previously by our laboratory (27). Frogs were anesthetized by immersion for 15 min in ice-cold water containing 0.3% tricaine (Sigma). Oocytes were removed from the ovary and washed with Ca 2ϩ -free OR-2 solution (in mM: 82.5 NaCl, 2 KCl, 1 MgCl 2 , and HEPES-Tris, pH 7.5) and incubated at room temperature with gentle shaking for 90 min in OR-2 solution supplemented with 2 mg/ml collagenase (Sigma type IA). Oocytes were transferred to fresh collagenase solution after the first 45 min of incubation. Collagenase was removed by extensive washing in OR-2 solution at room temperature. Stage V and VI defolliculated oocytes were selected and incubated at 18°C in modified Barth's solution (in mM: 88 NaCl, 1 KCl, 2.4 NaHCO 3 , 0.82 MgSO 4 , 0.33 Ca(NO 3 ) 2 , 0.41 CaCl 2 , and 20 HEPES-Tris, pH 7.5), supplemented with penicillin (100 units/ml) and streptomycin (100 g/ml). After 3-4 h of incubation, healthy oocytes were injected with 50 nl of oatp1 cRNA (0.5-10 ng/ oocyte), 50 nl of Ntcp cRNA (0.5 ng/oocyte), or 50 ng of rat liver mRNA. Control oocytes were injected with a corresponding volume of sterile H 2 O. Injected oocytes were cultured at 18°C with a daily change of modified Barth's medium. Healthy oocytes with a clean brown animal half and a distinct equator line were selected for experiments.  (27). The uptake was stopped by adding 2.5 ml of ice-cold modified Barth's solution and oocytes washed three times each with 2.5 ml of ice-cold modified Barth's solution. Two oocytes each were dissolved in a polypropylene scintillation vial with 0.2 ml of 10% sodium dodecyl sulfate, and counted in a Packard model 4530 scintillation spectrometer after addition of 5 ml of Opti-Fluor (Packard Instruments, Downers Grove, IL). Taurocholate Uptake into Oocytes-Oocytes were injected with either 0.5 ng of Ntcp cRNA or oatp1 cRNA. Na ϩ -dependent taurocholate (10 M) uptake was determined at 25°C in 100 l in modified Barth's solution, supplemented with 2.1-3.5 Ci of [ 3 H]taurocholate. Uptake was terminated by addition of ice-cold stop solution, which was modified Barth's solution containing 1 mM unlabeled taurocholate in order to reduce unspecific binding of tracer taurocholate (28). Inhibitors were dissolved in the uptake buffer at the indicated concentrations.
GSH Efflux in Oocytes-Oocytes injected with either water or oatp1 cRNA were reinjected with 50 nl of [ 3 H]GSH (0.5-1 Ci/l) and allowed to recover for 30 min in modified Barth's solution with 0.5 mM acivicin. Oocytes were washed twice with 2.5 ml of modified Barth's solution prior to efflux studies. Efflux was measured at 25°C in 200 l of modified Barth's solution in the presence or absence of 50 mM GSH, 50 M taurocholate, or 1 M BSP. Efflux was terminated at 5, 20, or 65 min by removing the medium and counting it separately from the oocytes. The 5-min values were used as a background correction and were subtracted from the 20-and 65-min values to give 15-and 60-min efflux rates.
Changing Intracellular GSH Concentrations in Oocytes-To increase GSH, oocytes were injected with 50 nl of different GSH stock solutions (e.g. 220 mM GSH stock, for an increase of ϳ20 mM in the oocytes). Oocytes were incubated at 25°C for approximately 30 min before they were used in the experiment. To decrease GSH, oocytes were incubated in modified Barth's solution containing 5 mM buthionine sulfoximine and 5 M 1-chloro-2,4-dinitrobenzene for 18 h prior to uptake experiments. GSH content of the oocyte was measured as described previously (27).

RESULTS
X. laevis oocytes injected with the cRNA for either oatp1 or Ntcp, the Na ϩ -taurocholate cotransporting polypeptide, showed enhanced uptake of taurocholate (Fig. 1A), confirming previous results (1,3,4). Oocytes injected with oatp1 cRNA also demonstrated enhanced uptake of 10 nM LTC 4 and 50 M DNP-SG, when compared with oocytes injected with water or total liver mRNA (Fig. 1B); however, the extent of stimulation of LTC 4 and DNP-SG transport was lower than that of taurocholate. In contrast, oocytes injected with Ntcp-cRNA did not show enhanced transport of these glutathione S-conjugates (Fig. 1B).
The oatp1-mediated uptake of [ 3 (Fig. 2), and was independent of extracellular Na ϩ and Cl Ϫ concentrations (Fig. 3). Uptake was unaffected when Na ϩ in the culture medium was replaced with choline ϩ , or when Cl Ϫ was replaced with gluconate Ϫ (Fig. 3). Isosmotic replacement of NaCl with sucrose also failed to affect LTC 4 and DNP-SG uptake. These results with LTC 4 and DNP-SG are similar to those reported previously for taurocholate and BSP transport on oatp1 (3,4).

H]LTC 4 and [ 3 H]DNP-SG increased with time over a 4-h interval
The concentration dependence of uptake was examined in oocytes incubated with increasing concentrations of LTC 4 (10 nM to 500 nM) and DNP-SG (1 M to 1 mM), (Fig. 4). It was not possible to examine lower concentrations of LTC 4 because of limitations imposed by the specific activity of the isotope, nor was it feasible to examine higher LTC 4 concentrations due to its limited solubility in aqueous medium. Uptake of LTC 4 and DNP-SG was saturable, with apparent K m values of 0.27 Ϯ 0.06 and 408 Ϯ 95 M (mean Ϯ S.D.), and V max values of 40 Ϯ 35 fmol⅐oocyte Ϫ1 ⅐h Ϫ1 and 4.7 Ϯ 0.9 pmol⅐oocyte Ϫ1 ⅐h Ϫ1 , respectively (Fig. 4). This K m value for LTC 4 in oatp1-expressing oocytes is comparable to that measured in isolated rat hepatocytes, 0.20 M (19), consistent with a role for oatp1 in mediating LTC 4 uptake.
Inhibition experiments revealed that BSP, taurocholate, and bilirubin ditaurate, substrates for oatp1, are strong inhibitors of LTC 4 and DNP-SG transport in oocytes injected with oatp1 cRNA (Table I). Conversely, DNP-SG inhibited [ 3 H]taurocholate uptake, and the inhibition was competitive in nature (Fig. 5). The K m for taurocholate uptake in the absence of DNP-SG was 38 Ϯ 8 M (mean Ϯ S.D., n ϭ 3; Fig. 5), a value comparable to that reported previously (3,4). The K m was increased to 88 Ϯ 18 M in the presence of 100 M DNP-SG, whereas the V max was unchanged (3.5 Ϯ 0.4 and 4.7 Ϯ 1.5 pmol⅐oocyte Ϫ1. h Ϫ1 , respectively; Fig. 5), indicating competitive inhibition. Mutual inhibition of transport provides strong evidence for a shared transport system. 4,4Ј-Diisothiocyanatostilbene-2,2Ј-disulfonic acid, ouabain, and other glutathione S-conjugates such as BSP-SG (0.5 mM) and S-(p-chlorophenacyl)-glutathione (1 mM) also inhibited uptake of LTC 4 and DNP-SG (Table I). LTC 4 uptake was decreased to 43% and 23% of control in the presence of BSP-SG and S-(p-chlorophenacyl)-glutathione, respectively. DNP-SG uptake was also inhibited by these glutathione S-conjugates, but the extent of inhibition was lower (Table I). These data indicate that bulky glutathione S-conjugates may be substrates for oatp1.
To test the possibility that GSH is a substrate for oatp1, [ 3 H]GSH efflux was measured in oocytes expressing oatp1. The endogenous GSH transport rate in oocytes was high (Table II), as reported previously (27), but was even higher in oatp1expressing oocytes (Table II), consistent with the hypothesis that GSH is transported on oatp1. The presence of 50 mM unlabeled GSH in the efflux medium further accelerated [ 3 H]GSH efflux in both control and oatp1-expressing oocytes (Table II)  tively; Table II), consistent with GSH/organic anion exchange on oatp1.
To estimate the stoichiometry of GSH/organic anion exchange on oatp1, the amount of taurocholate-stimulated [ 3 H]GSH efflux (Table II) Table II, using similar batches of oocytes that were incubated under nearly identical conditions. As shown in Table II Additional evidence for GSH/organic anion exchange on oatp1 was provided by studies that measured taurocholate and LTC 4 uptake in oocytes with differing intracellular GSH concentrations (Fig. 6). Taurocholate uptake in oocytes with low (ϳ0.3 mM) intracellular GSH was 46% of control oocytes (ϳ2.5 mM endogenous GSH; Fig. 6A). When the intracellular GSH concentration was raised to ϳ20 mM there was a significant trans-stimulation of taurocholate uptake to 155% of control values, suggesting that GSH is a driving force for oatp1. In contrast, oatp1-expressing oocytes loaded with 20 mM amounts of either N-acetylcysteine, L-glutamate, or glutarate did not demonstrate enhanced uptake of [ 3 H]taurocholate (Fig. 6A), indicating that trans-stimulation is selective for GSH. Uptake of LTC 4 in oatp1-expressing oocytes was also inhibited by GSH depletion, and trans-stimulated by high intracellular GSH (Fig.  6B).
The relation between oatp1-mediated taurocholate uptake and intracellular GSH concentration is illustrated in Fig. 7. Taurocholate uptake reached a maximum at about 10 mM GSH and was not further accelerated at higher intracellular GSH concentrations, indicating a saturable transport mechanism. If one assumes a one-for-one exchange of extracellular taurocholate for intracellular GSH and Michaelis-Menten kinetics, these data indicate that the apparent K m for GSH is approximately 3 mM; however, this is only an approximation because of the indirect nature of the measurement. Given that intracellular GSH concentrations in hepatocytes are ϳ10 mM, the transporter should always be saturated under physiological conditions. However, GSH may become limiting for oatp1 function under conditions of extensive GSH depletion.
To further define the role of oatp1 in GSH efflux, additional studies examined taurocholate uptake in the presence of 1 mM methionine, cystathionine, or dithiothreitol, compounds that are known to alter sinusoidal GSH release in mammalian hepatocytes (13, 29 -32). In intact hepatocytes, methionine and cystathionine are effective trans-inhibitors of sinusoidal GSH efflux, and dithiothreitol a trans-stimulator of GSH efflux (13, 29 -32). However, these agents had no effect on oatp1-mediated taurocholate uptake (data not shown), indicating that oatp1 is not sensitive to these agents and therefore may not be the major mechanism of GSH release from hepatocytes. DISCUSSION The present study demonstrates that LTC 4 and DNP-SG are substrates for the hepatic sinusoidal organic solute transporter oatp1, substantiating previous findings in rat hepatocytes and isolated liver plasma membrane vesicles (18 -20). Uptake of these glutathione S-conjugates into oatp1-expressing oocytes was saturable, inhibited by known substrates of oatp1, and was independent of the Na ϩ and Cl Ϫ gradients. The apparent K m for LTC 4 uptake in oatp1-expressing oocytes, 0.27 M (Fig. 4), is similar to that measured in hepatocytes, 0.20 M (19), consistent with hepatic uptake on oatp1. The transport characteristics of DNP-SG in oocytes expressing oatp1 are also similar to those reported in perfused rat liver; uptake was comparatively low, Na ϩ -independent, and inhibited by substrates for oatp1 (20). LTC 4 and DNP-SG may of course also be substrates for additional hepatic solute transporters, including other members of the oatp family, and this will need to be evaluated.
The present findings with LTC 4 and DNP-SG differ from previous studies with oatp1 (3,4), which failed to find an effect of LTC 4 and DNP-SG on [ 35 S]BSP uptake in oatp1-expressing oocytes. The reason for this apparent discrepancy is most likely related to the kinetics of transport of BSP versus the glutathione S-conjugates on oatp1. BSP is one of the best substrate for this transporter, with a K m of 1.5 M, whereas DNP-SG has both a lower affinity (K m of 408 M) and transport capacity (V max of 4.7 pmol⅐oocyte Ϫ1. h Ϫ1 for DNP-SG (Fig. 4B), versus 9 pmol⅐oocyte Ϫ1. h Ϫ1 for BSP (3)). Thus, high concentrations of DNP-SG are needed to elicit significant inhibition of BSP uptake, but these concentrations were not used in previous studies. In contrast, LTC 4 has a higher affinity (K m of 0.27 M), but its use as an inhibitor is limited by its very low water solubility. Maximum LTC 4 concentrations in aqueous medium are on the order of 0.5 M, a concentration that is expected to have minimal effects on BSP uptake via oatp1.
Our results also provide direct evidence that intracellular GSH is a driving force for solute uptake on oatp1. First, the observation that glutathione S-conjugates are substrates for  4 and DNP-SG uptake into X. laevis oocytes expressing oatp1 Oocytes were injected with either water or oatp1 cRNA (5 ng/oocyte for LTC 4 and 0.5 ng/oocyte for DNP-SG). Uptake was measured under conditions described under "Experimental Procedures." Values are means Ϯ S.D. of three or four oocyte preparations, and are corrected for water-injected oocytes. All results are significantly different from control (p Ͻ 0.01, using unpaired Student's t test), except GSSG.  oatp1 indicates that the GSH moiety may be a determining factor in substrate recognition. Second, GSH cis-inhibited LTC 4 , DNP-SG, and taurocholate uptake, indicating a shared transport mechanism. Third, oatp1-mediated taurocholate uptake was trans-stimulated by GSH, but not by glutarate, Lglutamate, or N-acetylcysteine, providing strong evidence for GSH/taurocholate exchange. The relation between intracellular GSH and oatp1-mediated taurocholate uptake (Fig. 7), indicates that transport is maximal at GSH concentrations that are comparable to those of intact hepatocytes, ϳ10 mM, and gradually declines at lower GSH concentrations. Thus, oatp1 function may be compromised by GSH availability if GSH levels are significantly decreased. Fourth, [ 3 H]GSH efflux was higher in oatp1-expressing oocytes, and was further accelerated in the presence of extracellular taurocholate, BSP, and high concentrations of GSH, providing strong evidence for GSH/organic anion exchange on oatp1. Fifth, a comparison of the amount of GSH released and taurocholate taken up on oatp1 indicates one-for-one exchange of these monovalent anions.
Additional evidence for GSH/organic solute exchange on oatp1 comes from previous studies of GSH transport in rat liver sinusoidal membrane vesicles (11). GSH transport in sinusoidal vesicles was cis-inhibited and trans-stimulated by BSP-SG, a substrate for oatp1, suggesting that GSH efflux may drive the uptake of organic anions into hepatocytes (11). Studies with canalicular membrane vesicles have also demonstrated transstimulation between GSH and certain organic solutes (33,34), suggesting that GSH/organic solute exchange may be a general mechanism.
To date, two major families of ATP-independent organic anion transporters have been identified. One family comprises the organic anion/dicarboxylate exchangers, and includes kidney oat1 (or roat1) (35)(36)(37). Cellular uptake of organic anions
The present findings with oatp1 raise the possibility that GSH exchange may be a common mechanism for this family of transporters. However, it is important to note that, despite the high sequence homology among the oatp-related transporters, there are major differences in substrate specificity and possibly in energy coupling. For example, oatp1 and oat-k1 share 72% amino acid identity, yet oat-k1 is unable to transport either taurocholate or LTC 4 (38), whereas oatp1 accepts both as substrates (Ref. 3 and present study). Additional studies are needed to characterize the substrate specificity and driving forces for the oatp-related transporters.
The overall contribution of oatp1 to sinusoidal GSH efflux is unknown, but it is unlikely to be the predominant mechanism of GSH release in hepatocytes. Studies characterizing sinusoidal GSH efflux reveal that methionine and cystathionine transinhibit GSH efflux in hepatocytes (29,30,32), and dithiothreitol trans-stimulates GSH efflux (13,31). In the present study, methionine, cystathionine, and dithiothreitol did not affect oatp1-mediated taurocholate transport, indicating that this putative GSH efflux mechanism is not altered. Taken together, these findings are consistent with the presence of at least two GSH transport systems in the sinusoidal membrane (Fig. 8): a putative GSH transporter (gsht) that is sensitive to methionine, cystathionine, and dithiothreitol; and oatp1, which is unaffected by these compounds. trans-Stimulation of GSH efflux by BSP or BSP-SG may occur through oatp1, whereas cis-inhibition may occur through both systems. It is plausible that other oatp1-like transporters that have recently been cloned may also be contributing to GSH efflux (5), and this possibility is currently being examined in our laboratory.
Previous studies have shown that sinusoidal GSH release in perfused rat liver is inhibited by BSP, rose bengal, indocyanine green, and bilirubin (41), substrates of oatp1. Inhibition occurs from an intracellular site and is reversible (41); inhibition is overcome when these anions are cleared from the cell by transport across the canalicular membrane into bile. This cis-inhibition of the putative sinusoidal GSH transporter by some oatp1 substrates may be of physiologic significance in that it would maximize intracellular GSH concentrations for organic anion uptake on oatp1 (Fig. 8). Interestingly, transport of these same organic anions into bile is mediated by mrp2, a transporter that also appears to function optimally when intracellular GSH concentrations are high (42)(43)(44). Thus, the inhibition of sinusoidal GSH efflux by oatp1 substrates, and the resulting higher cellular GSH concentrations, may accelerate both hepatic uptake and biliary excretion of these compounds. The net effect is faster hepatobiliary clearance of chemicals, many of which are potentially toxic (Fig. 8).
In conclusion, the present study demonstrates that GSH is a driving force for oatp1 and that glutathione S-conjugates are a novel class of substrates for this transporter. oatp1 may therefore contribute to sinusoidal GSH release and the regulation of GSH turnover in hepatocytes. The data are consistent with the presence of multiple GSH efflux mechanisms, and explain several observations regarding GSH efflux characteristics in intact hepatocytes. However, confirmation of this model awaits the molecular characterization of additional GSH transport proteins. and gsht mediate GSH efflux in the sinusoidal membrane of the hepatocyte. gsht is trans-inhibited by methionine, stimulated by dithiothreitol, and cis-inhibited by BSP and other organic anions that are substrates for oatp1. oatp1-mediated uptake of organic anions, including BSP, is energized by GSH efflux. mrp2, which is located on the canalicular membrane, mediates ATP-dependent efflux of organic anions into the bile canaliculus. mrp2 is stimulated by intracellular GSH.