A Unique Multifunctional Transporter Translocates Estradiol-17β-Glucuronide in Rat Liver Microsomal Vesicles*

A wide array of drugs, xenobiotics, and endogenous compounds undergo detoxification by conjugation with glucuronic acid in the liver via the action of UDP-glucuronosyltransferases. The mechanism whereby glucuronides, generated by this enzyme system in the lumen of the endoplasmic reticulum (ER), are exported to the cytosol prior to excretion is unknown. We examined this process in purified rat liver microsomes using a rapid filtration technique and [3H]estradiol-17β-d-glucuronide ([3H]E217βG) as model substrate. Time-dependent uptake of intact [3H]E217βG was observed and shrinkage of ER vesicles by raffinose lowered the steady-state level of [3H]E217βG accumulation. In addition, rapid efflux of [3H]E217βG from rat liver microsomal vesicles suggested that the transport process is bidirectional. Microsomal uptake was saturable with an apparentKm and V max of 3.29 ± 0.58 μm and 0.19 ± 0.02 nmol·min−1·mg protein−1, respectively. Transport of [3H]E217βG was inhibited by the anion transport inhibitors 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid and probenecid. Specificity of the transport process was investigated by studying thecis-inhibitory effect of anionic metabolites, as well as substrates of the plasma membrane multidrug resistance-associated proteins on the uptake of [3H]E217βG. Collectively, these data are indicative of a novel multifunctional and bidirectional protein carrier for E217βG and other anionic compounds in the hepatic ER. This intracellular membrane transporter may contribute to the phenomenon of multidrug resistance.

the microsomal UDP-glucuronosyltransferases (UGTs), is therefore primarily a detoxification reaction, which occurs largely in the liver. Glucuronides, generated in the ER lumen, are subsequently exported from the hepatocytes by conjugate export pumps, which are members of the multidrug resistanceassociated protein (Mrp) and the organic anion transport protein (OATP) families (1)(2)(3)(4). The substrate specificity of Mrp1 and Mrp2 is similar and includes glucuronoconjugates, but the subcellular localization of these transporters is distinct (5). Mrp1 is localized in the basolateral membrane of polarized cells, such as hepatocytes, whereas Mrp2 exhibits the feature of a multispecific organic anion transporter, acting at the level of the bile canalicular (apical) membrane. Expanding numbers of other transporters for organic anions, including glucuronides, are being identified as shown by the recent characterization of rat Mrp3 (6), human OATP2 (2), and OATP8 (4). None of the glucuronide transporters identified to date, however, have been shown to be functional in intracellular membranes.
Analysis of the topology of UGTs suggests that the bulk of the enzyme protein, in particular the active site, conceptually subdivided into aglycone and donor substrate binding sites, is located in the cisternal lumen of the ER (7). Access of the donor substrate UDP-GlcUA to the UGT active site is driven by a carrier-mediated process, which has been functionally characterized (8 -12), but its identity remains obscure. Whereas the putative existence of active transport processes, which reduce the cytoplasmic accumulation of glucuronides has been an area of intense scrutiny, little is known about the initial pathway of these nascent metabolites as they are generated within the lumen of the ER. It is notable that glucuronides are amphipathic compounds, and the glucuronic acid moiety (pK a is ϳ3-4) (13) is likely to be in an anionic form when generated in the ER lumen. This observation supports the requirement for a transport process across the ER membrane, as observed in plasma membranes. Of interest is a study by Waddell et al. (14) using intact microsomes prepared from an infant exhibiting conjugated hyperbilirubinemia, which indicated that, when compared with a matched control, bilirubin-glucuronide export from the ER was defective. In contrast, 1-naphthol-glucuronide efflux was unaffected, indicating differential sorting between glucuronoconjugates. A variety of other organic compounds of physiological importance, such as reduced glutathione, dihydroascorbic acid, glucose 6-phosphate, glucose, UDP-sugars, and phosphatidylcholine are also transported across the ER membrane (15)(16)(17)(18)(19)(20), but the identification of the proteins involved lags behind that of plasma membrane transporters.
Despite the fact that it is a critical step in the overall excretory process of nascent glucuronides, the mechanism of glucuronide transport across the hepatic ER has not been characterized. Here we demonstrate uptake, accumulation, as well as * This work was supported in part by National Institutes of Health Grant DK-36887. A preliminary report of this study was presented at Digestive Diseases Week 2000, in San Diego, May 2000 and was published in abstract form in (2000) Gastroenterology 118, (Abstr. A934). 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.
efflux, of radiolabeled estradiol-glucuronide in rat liver microsomes. The data support the concept of membrane transporters of broad substrate specificity in the ER similar to those in the plasma membranes, which are involved in the excretion of glucuronides.
Preparation and Characterization of Rat Liver Microsomal Vesicles-Microsomes were prepared from livers of 24 h-fasted male Harlan Sprague Dawley rats (220 -250 g), as described previously (10). The ER vesicles were immediately frozen in liquid nitrogen and kept at Ϫ74°C. Protein concentration was determined using the method of Bradford (21) with bovine serum albumin as the protein standard. Mannose-6phosphatase latency assay (22,23) was used to determine the integrity of the microsomal preparations, which was consistently greater than 95%. Vesicle integrity was also confirmed by determination of 4MUglucuronidating activity of rat liver microsomes; the vesicles were treated for 30 min on ice with digitonin at a detergent to protein ratio of 2 (w/w). A control in which detergent was omitted was run in parallel. Glucuronidation of 4MU (10 g of microsomal protein) was assayed as described (24), and latency of the enzyme activity was determined on the assumption that maximal activity is achieved in microsomes, which have lost vesicular integrity (25). Untreated microsomal vesicles were 94 -96% latent, consistent with mannose-6-phosphatase assay.
Purity of the ER vesicles was assessed by measurement of plasma membrane (Na ϩ , K ϩ , and Mg 2ϩ -ATPases) and Golgi membrane (ovomucoid galactosyltransferase) marker enzyme activities, as we have reported previously (10,12,26), and were in accordance with our work and other (11) earlier reports of only minor contamination.
[ 3 H]E 2 17␤G Uptake in Liver Microsomal Vesicles-[ 3 H]E 2 17␤G influx was determined at 25°C using a rapid filtration technique suitable for the detection of internalized radionuclide in rat liver microsomes (vide supra). [ 3 H]E 2 17␤G (15-30 mCi/mmol unless otherwise stated) was resuspended in a "cytosol-like" buffer, which consisted of 100 mM KCl, 20 mM NaCl, 5 mM MgCl 2 , 25 mM HEPES, 1 mM NaH 2 PO 4 , 1 mM EGTA, 0.5 mM CaCl 2 , and 5 mM NaN 3 (pH 7.2) (27) and was adjusted to the desired final concentration in E 2 17␤G indicated in the figures. In some experiments, ATP (1-5 mM) and a creatine phosphate/creatine kinase ATP regenerating system (27) were included in the incubation medium. Uptake was initiated by the addition of microsomes brought to a final concentration of 2.5 mg/ml protein (100-l final volume). Transport was stopped at indicated times by the addition of 4 ml of ice-cold cytosol-like buffer, followed by immediate filtration (FH225V, Hoeffer Scientific Instruments, San Francisco, CA). Filters were additionally washed twice with 5 ml of ice-cold cytosol-like buffer, solubilized in 10 ml of Filtron X and counted for radionuclide incorporation in a Beckman LS 5000 TD liquid scintillation counter (Beckman Instruments Inc., Palo Alto, CA). Transport was distinguished from binding (including nonspecific binding), by including in each experiment a control, in which microsomes had been previously treated with the pore-forming reagent alamethicin (0.1 mg/mg microsomal protein) (15). Uptake was calculated as the alamethicin-releasable content of the vesicles.
[ 3 H]E 2 17␤G Efflux in Liver Microsomal Vesicles-Outward flux of internalized [ 3 H]E 2 17␤G was studied by first preloading the radionuclide into rat liver microsomes (25 mg of protein/ml) for 15 min under the experimental conditions described for uptake. Preloaded vesicles were then promptly diluted 40-fold in cytosol-like buffer, adjusted to 25°C or 4°C, and the remaining internalized radionuclide detected by rapid filtration, as described above, at various times after dilution (see figures).
Effect of Anion Transport Inhibitors on the Uptake of [ 3 H]E 2 17␤G in Liver Microsomal Vesicles-The reversible effect of furosemide and probenecid was evaluated by the addition of these general anion transport inhibitors to the incubation medium containing intact microsomes in cytosol-like buffer, and the initial uptake rate was determined 30 s after the addition of [ 3 H]E 2 17␤G at 25°C. The effect of the transmembrane anion transport inhibitors DIDS and SITS on the uptake of estradiol-glucuronide also was evaluated. The compounds, solubilized in dimethyl sulfoxide (Me 2 SO, 2% v/v), were preincubated for 10 min at 25°C with intact rat liver microsomes (45 mg of protein/ml), prior to 18-fold dilution in cytosol-like buffer at 25°C, and uptake was measured at 30 s as described above. Uptake was evaluated in the absence of inhibitors (Me 2 SO alone, 2% v/v) to permit expression of the inhibitory potency as percent of control. To determine the effect of DIDS on the efflux of [ 3 H]E 2 17␤G, microsomes were first preloaded with [ 3 H]E 2 17␤G for 15 min at room temperature, as described above, followed by incubation in the presence of 5 mM DIDS (inhibition assay) or Me 2 SO (control) for 10 min at 25°C. Microsomes were subsequently diluted 40-fold in cytosol-like buffer and time-dependent efflux was immediately determined at 25°C, as described above.
cis Inhibition Studies of [ 3 H]E 2 17␤G Uptake in Liver Microsomal Vesicles-The effect of putative competitive inhibitors was assessed on estradiol-glucuronide uptake at 30 s and 25°C with addition of each compound at the desired final concentration, as indicated in Table I. All compounds examined in these competition experiments were also assessed for chemiluminescence, but were found not to interfere with scintillation counting.

Effect of Changing Osmolarity on [ 3 H]E 2 17␤G Accumulation in Liver Microsomal
Vesicles-The effect of changing osmolarity of the incubation medium was studied by measuring the uptake of estradiol-glucuronide after addition of the membrane-impermeant trisaccharide raffinose, under equilibrium conditions. Briefly, incorporation of the radionuclide and time-dependent uptake were determined under the experimental conditions described above. After 11 min incubation, steady-state accumulation of [ 3 H]E 2 17␤G was evident, and the intravesicular volume was subsequently reduced by the addition of raffinose (500 mM final concentration, dissolved in cytosol-like buffer). Time-dependent uptake of radionuclide was determined for an additional 15 min after addition of raffinose and compared with that in a control experiment (cytosol-like buffer without raffinose).

Analysis of [ 3 H]E 2 17␤G Purity and Stability in the Presence of Microsomes-[ 3 H]E 2 17␤G
stock solution was analyzed, as described below, by thin-layer chromatography (preadsorbent layer TLC plates, J.T. Baker, Inc, Phillipsburg, NJ) and was found to be Ͼ99% pure. Stability of this metabolite under the experimental conditions of transport measurement was determined by incubating 50 M [ 3 H]E 2 17␤G with intact rat liver microsomes in cytosol-like buffer, under the transport conditions described above, and the final volume reduced to 20 l, using a TLC system, as described (28,29). At indicated times (0 -60 min), 20 l of ice-cold ethanol was added to the incubation medium, and the tubes were immediately centrifuged at 14,000 rpm for 10 min at 4°C in a benchtop centrifuge. In addition, 50 M [ 3 H]E 2 17␤G was incubated with 10 units of Escherichia coli ␤-glucuronidase in cytosol-like buffer for 1 h, mixed with 20 l of ice-cold ethanol, centrifuged, and loaded on TLC, to provide a positive control for [ 3 H]E 2 17␤G hydrolysis by ␤-glucuronidase. Radiolabeled compounds were detected by autoradiography after 5-8 day exposure. TLC bands were also scraped into vials and mixed in a solution of 0.5 ml of methanol and 5 ml of Ecoscint A (National Diagnostics, Atlanta, GA) prior to counting for radioactivity.
Statistics-Experiments were performed at least in triplicate, with each of the values of a single set of experiments corresponding to the mean of a minimum of 2-3 determinations Ϯ S.E. Mean values were compared using the Student's t test.

RESULTS
As with other glucuronoconjugates, estradiol-glucuronide is enzymatically generated within the lumen of the ER and an outward movement of glucuronides from the ER to the cytosol prior to cytosolic clearance by specific plasma membrane export pumps is currently postulated. In the present studies, we have evaluated both inward and outward flux of intact estradiolglucuronide in intact rat liver microsomes. Efflux studies were reduced to a limited set of experiments for practical reasons. It is generally accepted that transmembrane solute movement may be studied in both directions, although one particular direction of this movement may be of principal physiological relevance. For instance, the sialic acid uptake process has been studied in rat liver lysosomes, whereas the physiological net flux of this acidic sugar is outwardly directed (30). Similarly, both influx and efflux of reduced glutathione (16), glucose (31), and phosphatidylcholine (18) have been demonstrated in intact rat liver microsomes.

Time Course and the Effect of Changes in the Medium Osmolarity on [ 3 H]E 2 17␤G Uptake in Liver Microsomal Membranes-We first evaluated the internalization of [ 3 H]E 2 17␤G
into microsomes at 25°C. Incubation of 50 M [ 3 H]E 2 17␤G with rat liver microsomal vesicles resulted in time-dependent radionuclide uptake and incorporation or membrane binding (Fig.  1). Radionuclide incorporation was markedly reduced when microsomes were previously incubated with the pore-forming reagent alamethicin (results not shown), and hence all uptake activities, including those in Fig. 1, were expressed as the difference between total radionuclide incorporation and membrane-bound incorporation detected when the microsomal vesicles were incubated with alamethicin. Initial uptake activity was linear for the first 40 s and achieved equilibrium after 5 min, which was then maintained for a period up to 40 min ( Fig.  1 and results not shown). Initial uptake was subsequently determined after 30-s incubation under these experimental conditions.
To further characterize the transport process, time-dependent uptake of [ 3 H]E 2 17␤G into rat liver microsomes was measured until a steady-state level of accumulation was achieved and the effect of vesicle shrinkage on [ 3 H]E 2 17␤G accumulation was evaluated. The membrane-impermeant trisaccharide, raffinose, was added at equilibrium (11 min after initiation of influx), and the results compared with uptake under the same conditions, but in the absence of raffinose. The data shown in Fig. 1 (Fig. 2), excluding any possible breakdown of the translocated substrate.
To distinguish possible transport activity arising from a minor fraction of plasma membrane vesicles contaminating the ER fraction, we studied the effect of digitonin on uptake activity. Digitonin has been shown to selectively permeabilize plasma membranes at low concentration and to disrupt the integrity of the ER membrane at higher concentration. Incubation of digitonin with a preparation of intact microsomes, under conditions which have been shown to permeabilize the plasma membrane (20 g of digitonin/mg protein, Ref. 31), did not exert any effect on the time-dependent uptake of [ 3 H]E 2 17␤G (results not shown). As expected, a concentration-dependent decrease in [ 3 H]E 2 17␤G uptake was evident at higher digitonin concentrations (Fig. 3). Parallel to this reduction in uptake, we observed a concomitant and progressive activation of 4MU glucuronidation catalyzed by lumen-oriented UGTs (Fig. 3) (Fig. 4). This rate of efflux was markedly diminished when the experiment was performed at 4°C. Inclusion of 1-5 mM ATP and an ATPregenerating system (27) did not further enhance E 2 17␤G uptake or efflux (results not shown).

Kinetic Parameters of [ 3 H]E 2 17␤G
Transport-Further characterization of this transport process was achieved by study of the concentration dependence of uptake. Uptake was determined at 30 s for concentrations ranging from 1 to 20 M. Saturation kinetics were observed and a Lineweaver-Burk plot of the data supported Michaelis-Menten kinetics with an apparent K m and V max of 3.29 Ϯ 0.58 M and 0.19 Ϯ 0.02 nmol⅐min Ϫ 1⅐mg Ϫ1 protein (Fig. 5). The K m value for E 2 17␤G is half that for the rat plasma membrane MRP2 (33), about 20fold lower than that for rat MRP3 (6), and comparable to that for rat Oatp1 (34).
Influence of Anion Transport Inhibitors on [ 3 H]E 2 17␤G Transport in Rat Liver Microsomes-To further characterize the specificity of the transporter for anionic species, we sought to evaluate the effect of the anion transport inhibitors (disulfonic stilbenes) SITS and DIDS on the movement of estradiolglucuronide across the ER membrane (Fig. 6). About 50% inhibition of [ 3 H]E 2 17␤G uptake was observed when the microsomes were incubated for 10 min in the presence of 5 mM DIDS. In contrast, SITS did not exhibit any inhibitory effect at the same concentration (Fig. 6A). As we previously reported for UDP-glucuronic acid uptake in rat liver microsomes, this difference may be due to the presence of 2 isothiocyanate groups in DIDS, which may contribute to inactivation by an efficient cross-linkage of two lysyl residues of the transporter (12). Probenecid, another anion transport inhibitor, inhibited the uptake of [ 3 H]E 2 17␤G by 25% at a concentration of 1 mM (results not shown). Collectively, these data support the concept of a carrier-mediated process for anionic substrates, such as estradiol-glucuronide, in the hepatic ER membrane.  Carrier-mediated transport of glucuronides across the ER membrane constitutes a key element in the overall excretory process of glucuronoconjugates, in conjunction with transport across plasma membranes. The existence of such a mechanism to extrude nascent glucuronides generated within the ER lumen has been postulated, but direct experimental evidence has been lacking. An exchange mechanism between UDP-GlcUA, the essential cosubstrate required for glucuronidation (which is synthesized in the cytosol) and glucuronides generated within the ER lumen has been proposed (35). Waddell et al. (14) observed the accumulation of bilirubin glucuronides in the lumen of liver microsomes prepared from a jaundiced patient. Interestingly, naphthol-glucuronide efflux was observed in the same microsomes, suggesting that the mechanisms of outward movement of these two glucuronides may be distinct. This may reflect a multiplicity of transport processes in the ER, as has been demonstrated in the plasma membrane (for recent review see Ref. 36). Additionally, it has been shown recently that, in contrast to model ER phospholipid membranes, bilirubin diglucuronide is able to readily cross native microsomal membranes; this observation supports the presence of a protein-mediated, facilitated flux for this metabolite (37). However, a detailed analysis of the transport mechanism for glucuronides in liver microsomes has not been reported, and hence definitive evidence for the involvement of protein carrier(s) in this process is lacking. Here, we provide an extensive characterization of glucuronide transport in rat liver microsomes using tritium-labeled estradiol-glucuronide as a model substrate. This metab-olite, generated by UGTs from the endogenous substrate estradiol by conjugation to glucuronic acid within the lumen of the ER, has been implicated in the pathophysiology of cholestasis (38). Using this glucuronoconjugate, we provide a detailed analysis of the kinetic mechanism involved in the intracellular endoplasmic reticular transport of glucuronides.
First, we documented the stability of E 2 17␤G under the experimental conditions employed in the transport assay, and thus demonstrated that microsomal-associated metabolic breakdown of E 2 17␤G does not account for the transmembrane transport and accumulation of this glucuronide. Second, it was essential to demonstrate that the E 2 17␤G transport was not the result of membrane contamination of the ER fraction, because this glucurononconjugate is also transported in plasma membrane inside-out vesicles (6, 39 -41). This possibility was excluded using the following approach. Incubation of ER membranes with digitonin under defined experimental conditions, which resulted in selective permeabilization of plasma membranes (31), had no measurable effect on [ 3 H]E 2 17␤G transport. Moreover, increasing the digitonin/protein weight ratio, which results in gradual disruption of the microsomal membrane, as shown by activation of 4MU glucuronidation (catalyzed by luminal ER-bound UGTs), resulted in concomitant reduction of [ 3 H]E 2 17␤G uptake (Fig. 3). Thus, we have clearly documented that [ 3 H]E 2 17␤G transport activity detected in the present study is attributable to a protein carrier in the ER, functionally distinct from those in the plasma membranes.
We then proceeded to further characterize the function of this novel transport mechanism in the ER membrane. We observed time and temperature-dependent radionuclide incorporation into rat liver microsomes. Transport was distinguished from membrane binding by systematic inclusion in parallel experiments of the pore-forming reagent alamethicin (42,43). Of note is the ATP-independence of the transport process. Transport is saturable (K m ϭ 3.29 Ϯ 0.58 M), in accordance with the contribution of a protein carrier for this glucuronide, and not merely simple binding or a diffusion process. Sensitivity of [ 3 H]E 2 17␤G accumulation in microsomal vesicles to osmotic change induced by raffinose provided additional evidence for a carrier-mediated process. cis inhibition studies (Table I) suggest a substrate specificity directed toward selected sulfo-and glucuronoconjugates. The lack of an inhibitory effect on the uptake of [ 3 H]E 2 17␤G by UDP-glucuronic acid and glucuronic acid, and the potent inhibition observed in the presence of some sulfoconjugates, indicates that the specificity is not driven by the glucuronic acid moiety. The inhibition appears to be linked to the presence of an anionic moiety, which is required, but is not sufficient to obtain potent inhibition, as shown by the lack of effect of some of the sulfoconjugates and glucuronides examined. For instance, this is the case for acetaminophenol-glucuronide, which did not significantly inhibit [ 3 H]E 2 17␤G transport, although we have observed that this compound is transported across the ER membrane (44). This observation supports the presence of additional glucuronide transporters of different specificity in the ER membrane, as previously suggested (14).
Our data show that intact microsomes are capable of both internalizing and transporting [ 3 H]E 2 17␤G of the vesicles, indicating that translocation is bidirectional. Similar phenomena have also been observed for the transport of glucose, phosphatidylcholine, and reduced glutathione in intact rat liver microsomes (16 -18). In vivo, outward transport of glucuronides is physiologically relevant because the nascent species are generated within the ER lumen prior to excretion. Although the physiological relevance of inward transport has not been clearly established, these influx data are of particular interest in light of our previous observations on the glucuronidation of bilirubin, the end product of heme catabolism: Bilirubin can form mono-and di-glucuronides through one or two of its aglycone carboxyl groups. Studies in intact rat liver microsomal vesicles underscored the transmembrane transport of bilirubin monoglucuronide from the cytosol to the luminal side of the ER, with subsequent glucuronidation leading to the rapid efflux of bilirubin diglucuronide (45). Thus (some) glucuronides remaining in the cytosol may be taken up again into the ER lumen via the transporter described in this study, to undergo further metabolism, such as additional glucuronidation (45) or hydrolysis by microsomal ␤-glucuronidases (32).
Collectively, our results demonstrate that in rat liver, the translocation of [ 3 H]E 2 17␤G and possibly of other anionic compounds across the ER membrane is facilitated by a transporter(s). The primary function of this transport process(es) is the export of nascent glucuronides from the ER lumen to the cytosol. Estradiol-glucuronide translocation by plasma membrane export pumps has been described previously (6, 39 -41). However, in the present study, we observed that neither ATP nor GSH were effectors of the microsomal transport activity. Thus, a major difference in the catalytic mechanism between rat liver plasma membrane transporters for estradiol-glucuronide and the ER transporter is the lack of ATP dependence of the latter. On the other hand, whereas transport by plasma membrane OATP1 is not ATP-dependent, it may be energized by counter-transport of GSH, a cis-inhibitor and trans-stimulator of rat OATP1 transport function (46). In contrast, GSH did not inhibit the uptake activity of either rat or human OATP2 (47,48). Because [ 3 H]E 2 17␤G transport in microsomes was unaffected by the presence of GSH (Table I), these observations support a translocation mechanism for the efflux of estradiol-glucuronide from the ER lumen, which is distinct from those in the plasma membrane, e.g. OATP1, MRPs, MDR1. The ER transporter identified in this study may share some features with rat OATP2 (47), but this requires more investigation with further characterization of OATP2.
Immunostaining with anti-MRP antibodies has suggested the presence of efflux pumps in intracellular compartments, possibly the ER, in addition to the plasma membrane (49,50). In rat liver, at least two transporters of the MRP family, MRP2 (40,41) and MRP3 (6) are able to translocate estradiol-glucuronide across the plasma membrane. It is unclear at present whether these observations indicate the presence of active Mrp (or another transporter with partial overlapping primary sequence) in the ER membrane, or reflect ER-protein trafficking and post-translational modification prior to insertion into the plasma membrane. As proposed previously, it is feasible that the transporter(s) involved may be related in their primary sequences to those of the plasma membrane transporter(s) superfamily (14). Examples of membrane transporters for specific hydrophilic compounds, which are functionally active in both plasma and ER membranes have been documented, but to the best of our knowledge, there is no indication as to whether their primary structures are closely related, because of the lack of sequence information on intracellular transporters. In the case of chloride channels, however, sequence similarity between plasma and ER channels suggests that they may belong to a homologous gene family or possibly arise from alternative mRNA splicing (51).
Plasma membrane transporters and drug-metabolizing enzymes induced by chemotherapeutics contribute to the phenotype of multidrug resistance. The expression level and transport activity of the ER transporter(s) identified in our study may modulate glucuronide disposition, and thereby may affect the overall toxicity and activity of glucuronides or glucu-ronidated compounds. Thus, the ER transport process may contribute to the multifactorial resistance mechanism by facilitating the extrusion of anticancer drugs, which undergo glucuronidation (such as camptothecin and anthracycline derivatives). Moreover, this ER transporter may be of particular relevance for those compounds where glucuronidation leads to bioactivation (52), i.e. resulting in modulation of the impact of bioactive glucuronides on cell function, as observed for instance with estradiol-glucuronide, the marker substrate utilized in our study and a potent cholestatic estrogen metabolite (38). The availability of biochemical tools, such as antibodies, inhibitors, and photoaffinity labels employed during the past decade to characterize multidrug resistance gene products and Mrp proteins could potentially be exploited to identify the ER glucuronide transporter(s) described in the present studies and to define its contribution to the process of glucuronide excretion.