Mechanisms of Taurocholate Transport in Canalicular and Basolateral Rat Liver Plasma Membrane Vesicles EVIDENCE FOR AN ELECTROGENIC CANALICULAR ORGANIC

The driving forces for taurocholate transport were determined in highly purified canalicular (cLPM) and basolateral rat liver plasma membrane (LPM) vesicles. Alanine transport was also examined for comparison. Inwardly directed Na’ but not K’ gradients transiently stimulated [SH]taurocholate (1 p ~ ) and [3H]alanine (0.2 mM) uptake into basolateral LPM 3-4- fold above their respective equilibrium values (overshoots). Na’ also stimulated [3H]taurocholate countertransport and tracer exchange in basolateral LPM whereas valino- mycin-induced inside negative K+ diffusion potentials stimulated alanine uptake but had no effect on tauro- cholate uptake. In contrast, in the “right-side out” oriented cLPM vesicles, [SH]taurocholate countertransport and tracer exchange were not dependent on Na+. Efflux of [3H]taurocholate from cLPM was also independent of Na’ and could be trans-stimulated by extravesicular taurocholate. Furthermore, an inside nega- tive valinomycin-mediated K’ diffusion potential inhibited taurocholate uptake into and stimulated tau- rocholate efflux from the cLPM vesicles. These studies provide direct evidence for a ”carrier mediated” and potential-sensitive conductive pathway for the canalicular excretion of taurocholate. In addition, they con- firm the presence of a possibly Nuclear. All other chemicals and reagents (analytical grade) were purchased either from Sigma, P-L Biochemicals, or Calbiochem-Behring Corp. All inorganic chemicals were of reagent grade or of the highest purity available.

has been more clearly defined by direct evidence for a basolateral Na+-taurocholate cotransport system in isolated rat liver plasma membrane vesicles (9)(10)(11). In contrast, the mechanisms and driving forces for bile acid transport across the canalicular or excretory pole, the rate-limiting step in overall transport from blood to bile, remain to be precisely delineated.
In the present study we have directly evaluated and compared the driving forces for taurocholate transport across the two polar membrane domains of hepatocytes. Highly purified blPM' and cLPM rat liver plasma membrane vesicles were simultaneously isolated from the same homogenate by rate zonal and discontinuous sucrose density centrifugation techniques (12). Since some of the characteristics of the basolateral carrier for taurocholate, have been previously defined (9-11) emphasis will be directed toward the canalicular transport system and studies in blLPM vesicles will be included only to the extent that are required for clear demonstration of differences between the taurocholate transport mechanisms in the two isolated LPM subfractions. Experiments with the neutral amino acid alanine, that is transported across blLPM by an electrogenic Na+-alanine cotransport system (13), will also be included to control for both the functional integrity of the isolated LPM subfractions and the effects of artificially induced transmembrane potential changes (e.g. valinomycininduced K' diffusion potentials) on taurocholate transport. Part of this work has previously been presented in preliminary form (14).

Animals
Male Sprague-Dawley rats (Charles River Breeding Laboratories, Inc., Wilmington, MA) weighing 200-250 g were used throughout this study. The animals had free access to water, were fed Purina Rodent Chow ad libitum and were housed in a constant temperature, humidity environment with alternating 12-h light (7 a.m. to 7 p.m.) and dark cycles. Fed animals were regularly killed by decapitation between 7:30 and 8:30 a.m.

Isolation of Canalicular and Basolateral Liver Plasma Membrane Vesicles
The methods for isolation of the cLPM and blLPM subfractions as well as their morphologic and biochemical characterization are described in detail elsewhere (12). In brief, a canalicular enriched "mixed LPM" subfraction was first separated out of a "crude nuclear pellet" by rate zonal flotation (44/36.5%, w/w, sucrose density interface) in the TZ-28 (Sorvall) zonal rotor. After tight homogenization (Type B Dounce homogenizer, 50 up and down strokes) the vesicu-'The abbreviations used are: blLPM, basolateral liver plasma membrane vesicles; cLPM, canalicular liver plasma membrane vesicles; P-face, cytoplasmic surface of freeze fracture membranes; Eface, extracellular surface of freeze fractured membranes; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid.
in Rat Liver Plasma Membrane Vesicles 10615 lated cLPM and blLPM were separated by high speed centrifugation (195,200 X g. , for 3 h) of the mixed LPM through a 3-step sucrose gradient (31, 34, and 38%, w/w). The membranes were collected at 105,000 X gave for 60 min and, except where stated otherwise, were resuspended in a standard membrane suspension medium consisting of 0.25 M sucrose, 0.2 mM CaC12, 5 mM MgSOd, 10 mM Hepes/Tris, pH 7.5. Routinely, the membranes were frozen and stored in liquid nitrogen (-70 "C, protein concentration >3 mg/ml) for up to 2 weeks without loss of transport function for taurocholate or alanine.

Characterization of the Isolated LPM Subfractions
The degree of purification of cLPM and blLPM was extensively analyzed by intracellular and plasma membrane marker enzyme activities (12). These studies indicated minor contamination of both LPM subfractions with intracellular organelles and virtually complete separation of blLPM from cLPM as reflected by the absence of Na+K+-ATPase, and glucagon-stimulatable adenylate cyclase activities, or intact secretory component in cLPM (12). In contrast, the blLPM subfraction was contaminated with cLPM by approximately 10% (12). Transmission electron microscopy revealed that both cLPM and blLPM are composed of membrane vesicles, although blLPM still contain some unbroken lateral membrane sheets (12). All cLPM subfractions in this study were routinely tested for Na+K+-ATPase activity (15) and only those cLPM preparations without any detectable Na+K+-ATPase activity were used for the transport studies (approximately four out of five membrane preparations). Protein was determined according to Lowry et al. (16) using bovine serum albumin as a standard.

Freeze Fracture Analysis
I n Vivo-To define the particle density on the P-face and E-face surface of canalicular microvilli, freeze fracture replicas of bile canaliculi were prepared as previously described in specimens of liver of untreated rats (17) using undirectional shadowing at an angle of 45".
In Vitro-50 pg of frozen, quick thawed (37 "C), and rehomogenized cLPM vesicles (30 up and down strokes with a tight Dounce homogenizer) were fixed in 1% glutaraldehyde cryoprotected with glycerol to a final concentration of 25% (v/v) for 1-1.5 h at 4 "C. Samples were frozen in Freon 22 cooled by liquid nitrogen at its melting point and processed as described (11) using a Balzers freeze etch unit, BAF 301 (Balzers Hudson, NH). The replicas were examined with a Carl Zeiss, Inc., 10B transmission electron microscope. A circular test system of known area (18) was used for analysis of the particle density in intact canalicular microvilli (Fig. 1). In the isolated cLPM vesicles the total number of particles present on the fractured profiles was counted since the particles exhibited an uneven surface distribution (Fig. 1E). The approximate surface area was determined from the measured diameters assuming a planar configuration of the profiles. To reduce the error introduced by this assumption, only cast shadow-free vesicles, e.g. profiles hit by the 45" shadow casting tangent point and the vesicle pole (18), were included in the final analysis (Table I). Several freshly isolated membrane preparations were also analyzed and established that particles were not lost during storage of cLPM vesicles in liquid nitrogen.

Transport Studies
Frozen membrane suspensions were quickly thawed by immersion of the tubes in a 37 "C water bath, diluted to the desired protein concentration and again vesiculated with a tight Dounce homogenizer (Type B; 30 up and down strokes), and placed on ice. In experiments where preloading of the vesicles with taurocholate was required, the unlabeled (Table 11) or labeled (Figs. 7 and 8) compound was added before vesiculation of the thawed membranes. After tight homogenization these membranes were also preincubated at 25 "C for 20-30 min before placing them on ice to ensure maximal intravesicular concentrations of taurocholate.
Transmembrane transport of [3H]taurocholic acid and [3H]alanine were measured by a rapid Millipore filtration technique (Millipore Cow., Bedford, MA). Membrane suspensions (40-80 pg of protein in 20 pl) were preincubated at 25 "C for at least 5 min. Uptake studies were initiated by the addition of 80 p1 of incubation medium whereas tracer efflux was determined by addition of 180 pl of incubation medium. The exact composition of the media is given in the figure and table legends of the individual experiments. After incubation for the indicated time intervals, transport was terminated by the addition of 3.5 ml of ice-cold stop solution (100 mM KCl, 100 mM sucrose, 0.2 mM CaC12, 5 mM MgS04, 10 mM Hepes/Tris, pH 7.5). Membrane vesicle-associated ligand was separated from free ligand by immediate rapid filtration (1 ml/s) through a 0.45-pm Millipore filter (HAWP) which had been presoaked in cold deionized water and in the case of taurocholate, additionally prefiltered with 3 ml of 1 mM taurocholate to diminish nonspecific filter binding. The filter was washed twice with 3.5 ml of stop solution, dissolved in 6 ml of Redi-solvHP (Beckman Instruments, Inc.) and counted in a Beckman LS 7000 liquid scintillation counter. Nonspecific binding to the membranes was determined in each experiment by addition of cold incubation solution and cold stop solution to 20 pl of membrane suspension kept at 0-4 "C. This blank was subtracted from all determinations except for efflux studies where the blank values represented 100%. Unless otherwise indicated all incubations were performed in triplicates and all observations confirmed with two or more separate membrane preparations.

Sidedness of cLPM Vesicles
The orientation of the isolated cLPM vesicles was determined by freeze fracture analysis (19) in order to know whether the direction of the flux observed i n vitro is identical with that occurring in vivo. Following the general principle that the P-face of cellular membranes exhibit a higher particle density than the external membrane leaflet (E-face; 20), we first defined the particle densities on Pand E-faces of canalicular microvilli in vivo (Fig. 1A). Since in situ the microvilli all exhibit right-side out configuration, P-and E-faces are represented by convex-particle enriched and concave-particle poor membrane areas, respectively. A circular test system of known area (70 nm in diameter; Fig. 1A) was randomly assigned to a total of 250 cast shadow-free canalicular microvilli and the enclosed number of particles determined (18). This analysis revealed 2538 +764 particles per Fm2 (mean f S.D.) on P-faces and 398 f 427 particles per pm2 on E-faces.
Based on two standard deviations, P-faces were counted as every fractured cLPM vesicle exhibiting a particle density exceeding 1250 per pm2 (upper limit for E-faces). Using these criteria, P-faces represented about 42% of all cast shadowfree vesicles (Table I), which is close to the theoretically expected value of 50%. The particle density on these P-faced vesicles was then analyzed (see "Experimental Procedures") and the convex-and concave-oriented vesicles were defined as "right-side out" or "inside out'' vesicles, respectively. By these criteria 77% of the vesicles in the cLPM subfraction demonstrate right-side out configuration in which the extravesicular membrane face corresponds to the bile lumina1 surface in uiuo. This predominantly right-side out configuration of the cLPM vesicles is in agreement with the sidedness of cLPM vesicles recently isolated and determined by other techniques (21). Thus, canalicular rat liver microvilli are also predominantly oriented right-side out when fragmented into membrane vesicles like brush-border membranes from rat kidney cortex and rat small intestine (19).

Effects of Nu+ and K' on Alanine and Taurocholate Uptake into blLPM and cLPM Vesicles
To test for the functional integrity of solute transport we first measured the uptake of alanine and taurocholate into both blLPM and cLPM vesicles. In agreement with recently .. I conclude that the cLPM transport systems should remain functionally intact as well.

3
When the cLPM vesicles were studied in similar fashions, late was transiently accumulated 1.5-fold above equilibrium values. As previously observed with a mixed LPM fraction (11) the extent of the Na+-stimulated taurocholate uptake was proportional to the size of the imposed Na+ gradients (25, 50, and 100 mM in Fig. 2B, cLPM). However, in contrast to the studies with blLPM, taurocholate transport into cLPM was not stimulated by Na' when the ion was equilibrated on both sides of the membrane. Higher equilibrium uptakes (60 min) were observed for both alanine and taurocholate in cLPM compared to blLPM vesicles (Fig. 2). This is explained by differences in intravesicular volumes in the two LPM subfractions, which are approximately 2-fold higher in cLPM (1.9 +: 0.2 pl X mg" of protein; mean f S.D., n = 6) than blLPM (0.9 f 0.2 p1 x mg" of protein) as calculated from values for intravesicular volumes were also obtained with microvilli (A) and isolated canalicular LPM vesicles (B) for alanine (2.3 f 0.2 pl X mg" of protein and 1.3 f 0.1 pl X determination of the sidedness of cLPM vesicles. A, the particle mg" of protein for cLPM and blLPM, respectively) whereas density on P-and E-faces of the microvilli were determined by the apparent intravesicular volume of distribution for tauroe applying a circular test system of known area (18); x 60,648. B, cast shadow (CS) free, P-faced vesicles were classified as right-side out or inside out depending on their convex or concave orientation, respec-x mg" of Protein and 5*5 * 0-5 x mg" of protein in cLPM tively. A cast shadow is defined as the sharp edged shadow of the and blLPM respectively, confirming a high degree of tauroprofile edge cast down into the cavity or up to the pole of the concave cholate binding to or within the isolated membrane vesicles or convex vesicle profile; X 28,158. (10,11).
" was amounting to 15*5 ' 2'9 reported findings (9-11, 13) both alanine and taurocholate uptakes into blLPM were markedly stimulated by an inwardly directed Na+ gradient (3-4-fold intravesicular accumulation above equilibrium values; Fig. 2, A and B; blLPM). In contrast, 5-6-fold lower initial rates of uptake were observed in the presence of K+gradients or when Na' and K+ were equilibrated on both sides of the membranes. However, even under ion-equilibrated conditions, initial uptake rates of both alanine and taurocholate were consistently higher in the presence of Na' than with K' . These findings establish that true Na+ coupled cotransport systems exist in basolateral rat liver plasma membranes for the transport of both alanine and Binding of Taurocholate to cLPM Vesicles The extent of taurocholate binding was first evaluated by determining the effect of the medium osmolarity on taurocholate uptake at equilibrium (60 min). Alanine and D-glucose were also included for comparison. As illustrated in Fig. 3 the equilibrium uptakes of all three substrates decreased linearly as the osmolarity (concentration of cellobiose) increased and the vesicles diminished in size. Only the extrapolated regression line of D-glucose demonstrated zero uptake at infinite osmolarity indicating that D-glucose does not bind significantly to cLPM vesicles. In contrast, in three separate experiments 15-25'36 of the alanine and 45-65% of taurocholate bound to cLPM vesicles at equilibrium. Despite the relatively large proportion of binding, [3H]taurocholate could not be displaced by a 100-fold excess of cold taurocholate in the stop medium. This strongly suggests that taurocholate binds predominantly to the inside of the membrane vesicles, e.g. after uptake of taurocholate into the intravesicular space, a conclusion supported by the findings that taurocholate uptake did not occur at 0 "C whereas at 25 and 37 "C, intravesicular taurocholate increased linearly with time ( Fig. 4).

Temperature Dependency of Taurocholate Uptake in cLPM
Vesicles Increasing the temperature from 0 to 37 "C increased initial uptake rates for taurocholate in the presence of both NaCl and KC1 gradients (Fig. 4). Except at 0 "C, NaCl gradients promoted approximately 4-6-fold higher taurocholate uptake

I/[CELLOBIOSE (moles/ liter)]
rates for taurocholate efflux are temperature-sensitive as well 500 (Fig. 7). rates than KC1 gradients as expected. Furthermore, maximal intravesicular accumulation of taurocholate occurred earlier at 37 "C (9 s) then at 25 "C (>15 s) suggesting that initial stimulating taurocholate uptake distinguish whether the cations create a vesicle inside positive diffusion potential or whether there is a cation selective cotransport system or charge barrier within the canalicular membrane. In order to examine these possibilities we first looked for evidence for taurocholate countertransport and tracer exchange in both blLPM and cLPM.

Countertransport (Trans-stimulation) of Taurocholate Uptake into blLPM and cLPM Vesicles
The membrane vesicles of both subfractions were preloaded with 20 FM unlabeled taurocholate as described under "EXperimental Procedures." Initial uptake rates taurocholate were determined in the presence of out to in Na' and K+ gradients. NO; was used as a highly permeable anion    to minimize the development of transmembrane electrical potential differences which might exert stimulatory or inhibitory effects on taurocholate uptake. As demonstrated in Table I1 (part A ) , in blLPM uesicles the intravesicular taurocholate (20 PM) stimulated tracer taurocholate uptake (countertransport) only in the presence of an inwardly directed Na' gradient. In contrast, in c L P M countertransport was observed both in the presence of Na' and K' gradients suggesting the presence of a Na+ independent canalicular taurocholate anion "carrier."

Tracer Exchange of Taurocholate in blLPM and cLPM
If transmembrane taurocholate transport is directly coupled with the transport of Na' (Na+-taurocholate cotransport), uptake across the vesicular membrane should occur at faster rates when Na+ rather than K+ is equilibrated on each side of the vesicle membrane (tracer exchange). The data in Table I1 (part B ) demonstrate that under Na+-and K+-equilibrated conditions, Na' indeed stimulated the uptake of the extravesicular tracer into taurocholate preloaded (20 PM) blLPM vesicles. In contrast, in identical experiments no differences in taurocholate tracer exchange rates beteen Na+ and K+ were observed in cLPM vesicles. Thus, the "Na+ effect" on uptake rates of taurocholate in cLPM cannot be attributed to a coupled Na+-taurocholate cotransport system. Rather it appears that the observed stimulation of taurocholate transport into cLPM vesicles by Na+ and other cation gradients (Figs.  2, 4, and 5) might result from creation of positive diffusion Taurocholate Transport in Rat Liver Plasma Membrane Vesicles 10619 potentials inside the vesicles or from effects of Na+ on negatively charged barriers within the membrane. Therefore, we next investigated the effects of membrane potential changes on the transmembrane taurocholate transport in cLPM.

Effects of the Electrical Membrane Potential on Taurocholate Transport
Previous studies with plasma membrane vesicles from rat liver have demonstrated that the hepatic uptake of alanine is Na+ dependent and stimulated by a negative intravesicular electrical membrane potential (13). Therefore, alanine uptakes were included in this part of the study as a reference for assessment of the effectiveness of either anions or the K+ ionophore valinomycin in inducing transmembrane potential changes.
Effects of Anions on Na+-stimulated Taurocholate Uptake into cLPM Vesicles-Variable changes in the transmembrane potential may be induced by selecting anions with different membrane permeability characteristics. As demonstrated in Table I11 initial uptakes of alanine in the presence of an inwardly directed Na+ gradient decreased with decreasing membrane permeation by the accompanying anions (SCN-> NO? = C1-> SO:-> gluconate-). Since thiocyanate usually diffuses into vesicles more rapidly and gluconate more slowly than Na+, these data indicate that uptake of alanine is stimulated by the relatively more negative intravesicular membrane potential. In contrast, using identical conditions, the Na+-stimulated taurocholate uptake is inhibited in the presence of the more permeant anions thiocyanate and nitrate compared to the less permeant anions sulfate and gluconate. As previously observed in the mixed LPM (ll), highest initial uptake rates were consistently found with chloride suggesting that this anion exerts an additional, but as yet undetermined stimulatory effect on transcanalicular transport of taurocholate. Nevertheless, the overall comparison of the effect of anion substitution on alanine and taurocholate uptakes in cLPM indicate that in the presence of Na+, uptake of alanine occurs as a cation whereas under similar conditions taurocholate is transported as an anion.

Effects of Valinomycin-induced K+-diffusion Potentials (Inside Negative) on Alanine and Taurocholate Uptake into blLPM and cLPM Vesicles-To
more directly investigate the effects of the electrical membrane potential on alanine and taurocholate transport, blLPM and cLPM were preloaded with K gluconate and uptake rates of both solutes were studied in the presence and absence of the K+ ionophore valinomycin. Under these conditions and without any K+ present outside the vesicles (e.g. in the incubation medium) valinomycin permits a rapid out-diffusion of K+ thereby .creating a transient negative potential within the vesicles. As shown in Fig.  6A, an inside negative membrane potential enhances the uptake of alanine into blLPM above control values as expected in the presence of an inwardly directed Na' gradient (13). However, the same phenomenon was also seen in cLPM indicating the occurrence of an electrogenic Na'-alanine cotransport in canalicular plasma membranes as well. In the absence of Na+ (tetramethylammonium substitution) the transport of the neutral alanine was diminished and not affected by the valinomycin-induced transmembrane potential changes. In contrast, under Na+-free conditions, the uptake of the anionic taurocholate was consistently inhibited by the negative intravesicular membrane potential both in blLPM as well as cLPM (Fig. 6B). However, in the presence of inwardly directed Na' gradients, taurocholate uptake was distinctly different in the two LPM subfractions. In these experiments, uptake of taurocholate into blLPM vesicles was unchanged by valinomycin whereas its transport into cLPM vesicles was inhibited (Fig.6B). These results are consistent with the anion substitution experiments (Table 111) and indicate that, whereas the basolateral Na+-taurocholate co-

Taurocholate Transport in Rat Liver Plasma
Membrane Vesicles transport appears electroneutral (ll), taurocholate is invariably transported as an anion across the bile canalicular membrane whether Na' is present or absent in the incubation system. The latter findings complement preliminary reports by others demonstrating stimulation of taurocholate uptake into cLPM vesicles by valinomycin-induced inside positive K' diffusion potentials (22).

Efflux of Taurocholate from cLPM Vesicles
Since the majority of cLPM vesicles exhibit right-side out configuration (Table I), efflux from the vesicles rather than uptake represents the physiologic direction of transcanalicular solute transport. We therefore measured taurocholate efflux from cLPM vesicles to determine if we could verify the conclusions drawn from the uptake studies. As demonstrated in Fig. 7A efflux rates from taurocholate preloaded (10 phi) cLPM vesicles were temperature-sensitive, unchanged in the presence of in to out gradients of either Na' or K' and stimulated by the simultaneous presence of 50 PM of unlabeled taurocholate on the outside of the vesicles. This countertransport of [3H]taurocholate efflux was also observed at similar rates with either in to out Na+ or K' gradients. Furthermore, taurocholate efflux was directly stimulated by a valinomycininduced intravesicular negative K' diffusion potential (Fig.  7B). The alternative possibility of an alkalinization of the extravesicular medium through a K+/H' exchange, which then could stimulate taurocholate efflux through an additional OH-/taurocholate anion exchange mechanism, was excluded by the findings that an outside alkaline pH gradient (pH 7.6 out/6.O in) did not stimulate taurocholate efflux (data not shown). The data therefore are consistent with a conductive carrier-mediated pathway for the canalicular excretion of the taurocholate anion.

DISCUSSION
In the present study we used simultaneously isolated and highly purified basolateral and canalicular rat liver plasma membrane vesicles to directly evaluate the driving forces for taurocholate transport across these two polar membrane domains of hepatocytes. The data obtained with blLPM confirm that the sinusoidal or basolateral uptake of taurocholate into hepatocytes is a secondary active transport process driven by the out to in Na' gradient ( Fig. 2) and mediated by an apparently electroneutral Na+-taurocholate cotransport system ( Fig. 6; Table 11;. In contrast, excretion of taurocholate from the cells into bile canaliculi, a process that is directly mimicked by taurocholate efflux from the isolated right-side out cLPM vesicles (Table I), is mediated by a Na' independent anion carrier and appears to be driven by the physiologic intracellular negative membrane potential (Fig.  7). Finally, the present study also shows evidence for an electrogenic Na+-alanine cotransport system in cLPM (Figs. 1 and 6) that may reflect a potentially important mechanism for back reabsorption of metabolically precious neutral amino acids from bile canaliculi.
A high degree of purification of the two separated vesicular LPM subfractions is crucial for the correct assignment of the presented in vitro data to the two surface domains of hepatocytes. This question has been dealt with extensively in a previous validation of the subcellular fractionation procedure adopted in the present investigation (12). These earlier studies have shown that the presence of Na'K'-ATPase activity in TIME (minutes) cLPM is a sensitive criterion to test for eventual crosscontamination with basolateral membrane fragments and indicated that cLPM with undetectable Na'K+-ATPase activity are maximally contaminated with blLPM to an extent of 0.7%. Since in the present study only Na+K+-ATPase activity free cLPM preparations were used for functional studies, it can be assumed that the results obtained with this LPM subfraction indeed reflect bile canalicular transport processes in uiuo. In contrast, the blLPM subfraction is consistently contaminated to an extent of approximately 10% with canalicular membrane components (12). Since the cLPM vesicles exhibit a 2-fold higher intravesicular volume per milligram of protein compared to blLPM vesicles (see "Results") it must be assumed that around 20% of the total intravesicular space in blLPM is enclosed by bile canalicular membranes. Nevertheless, because cLPM and blLPM are simultaneously isolated from the same homogenate and cLPM are virtually free of basolateral contaminants, functional phenomena that are exclusively present in blLPM can be correctly assigned to the basolateral pole of hepatocytes in uiuo.
Recent studies from this and other laboratories have demonstrated direct evidence for a basolateral Na+-taurocholate cotransport system in sinusoidal rat liver plasma membrane vesicles isolated by different techniques (9)(10)(11). The present findings of a Na' gradient induced transient accumulation of taurocholate within the blLPM vesicles (Fig. 2) as well as the Na+-stimulated countertransport (trans-stimulation) and tracer exchange of taurocholate (Table 11) confirm these earlier reports and establish that the basolateral uptake of taurocholate into hepatocytes is driven by the out to in electrochemical Na' gradient via an obligatory coupled Na+taurocholate symport system. However, similar to rat intestinal brush-border membrane vesicles (23, 24) the rheogenicity of this hepatocellular Na' taurocholate symport remains controversial (9)(10)(11). Thus, no effects of valinomycin-induced K' diffusion potentials (inside negative) on Na'-stimulated taurocholate uptake were observed in blLPM isolated in this study (Fig. 6) as well as in our earlier studies with the mixed LPM fraction (1 l), suggesting electroneutral Na+-coupled taurocholate uptake, whereas others have provided evidence for an electrogenic (positively charged) Na+ taurocholate cotransport in sinusoidal plasma membrane vesicles (9,10). Since our blLPM subfraction is still contaminated with cLPM (see above) an electrogenic basolateral transport system for taurocholate might be masked. However, the data clearly demonstrate that in blLPM the Na+-stimulated taurocholate uptake can overcome an inside negative membrane potential (Fig. 6) indicating that the physiologic out to in Na' gradient can drive uphill transport of the anionic taurocholate against an otherwise unfavorable electrochemical gradient (intracellular negativity; [taurocholate]intracellular > [taurocho-In contrast, in cLPM the Na+-stimulated taurocholate uptake was clearly inhibited by an intravesicular negative membrane potential created either by permeant anions (Table 111) or by valinomycin-induced rapid outdiffusion of K' (Fig. 6). Furthermore, efflux of taurocholate from the cLPM vesicles was stimulated by an artifically imposed intravesicular negativity (Fig. 7). Thus, transfer of taurocholate across the bile canalicular membrane appears to be invariably accompanied by transfer of a negative charge. Since efflux from cLPM vesicles represents the physiologic direction of solute movement in uiuo (Fig. 1, Table I probably also of other bile acids. However, calculations based on the Nernst equation suggest that the electrical potential cannot be the only driving force for the movement of bile acids across the canalicular membrane since it would account for only a 3-instead of a 10-fold concentration difference (1,2). Additional driving forces might include extrusion of bile acid loaded vesicles into the canalicular lumen through exocytosis (27) and/or other direct energy dependent carriermediated transport mechanisms (28).
Our present observations that [3H]taurocholate uptake in cLPM exhibits saturation while efflux demonstrates countertransport (Fig. 7) strongly suggest that an integral bile acid anion carrier is present within the bile canalicular membrane. This conclusion is consistent with binding studies and photoaffinity labeling of isolated bile canalicular enriched plasma membrane subfractions using either bile acids (29) or photolabile bile acid analogues (30), respectively. Similar bile acid binding proteins have been demonstrated in basolateral as well as canalicular plasma membranes (30) supporting our findings that Na+ stimulates initial taurocholate uptake in both blLPM and cLPM (Figs. 2 and 5) despite intrinsic biochemical differences in the two transport systems as discussed above. We therefore speculate that a similar but differentially gated and asymmetrically functioning bile acid carrier may be present at the two polar domains of hepatocytes. Further experimental validation must include isolation and functional reconstitution of the putative transport protein(s) present in basolateral (31, 32) as well as canalicular rat liver plasma membranes.
Our studies also provide evidence for electrogenic Na+dependent alanine uptake into cLPM vesicles (Figs. 2 and 7, Table 111). Preliminary observations suggest that this canalicular Na+-stimulated alanine uptake exhibits countertransport and is not inhibitable by the amino acid analog 2-methylaminoisobutyric acid.' These criteria have been reported to be characteristic for the so called "ASC-system" in intact cells including rat hepatocytes (33,34). Thus, the ASC-system rather than the "A-system" may be present at hepatic canalicular liver plasma membranes, where it could primarily serve to reabsorb cysteine, a breakdown product of canalicular secreted reduced glutathione. The intriguing possibility that the various amino acid transport systems are differentially distributed on the two polar surface domains of hepatocytes is under further investigation in this laboratory.
In summary, our findings demonstrate that transhepatocyte excretion of taurocholate is a secondary active transport driven by the electrochemical potential difference for Na' across the basolateral membranes (Na+-taurocholate cotransport) and by the electrical potential difference across the bile canalicular membranes. The canalicular excretory step is mediated by an anion carrier the molecular characteristics of which remain to be exactly defined. The presented biochemical differences between blLPM and cLPM extend the previously reported evidence for functional polarity of the various hepatocellular surface domains (12) and confirm that the isolated basolateral and canalicular LPM subfractions are suitable for further evaluation of the physiologic polarization of the liver cell.