MDR1 P-glycoprotein Reduces Influx of Substrates without Affecting Membrane Potential*

, MDR1 (multidrug resistance) P-glycoprotein (Pgp; ABCB1) decreases intracellular concentrations of structurally diverse drugs. Although Pgp is generally thought to be an efflux transporter, the mechanism of action remains elusive. To determine whether Pgp confers drug resistance through changes in transmembrane potential ( E m ) or ion conductance, we studied electrical currents and drug transport in Pgp-negative MCF-7 cells and MCF-7/ MDR1 stable transfectants that were established and maintained without chemotherapeutic drugs. Although E m and total membrane conductance did not differ between MCF-7 and MCF-7/ MDR1 cells, Pgp reduced unidirectional influx and steady-state cellular content of Tc-Sestamibi, a substrate for MDR1 Pgp, without affecting unidirectional efflux of substrate from cells. Depolarization of membrane potentials with various concentrations of extracellular K (cid:1) in the presence of valinomycin did not inhibit the ability of Pgp to reduce intracellular concentration of Tc-Sestamibi, strongly suggesting that the drug transport activity of MDR1 Pgp is independent of changes in E m or total ion conductance. Tetraphenyl borate, a lipophilic anion, enhanced

MDR1 P-glycoprotein (Pgp), 1 a member of the ATP-binding cassette family of membrane transporters, decreases intracellular concentrations of structurally diverse compounds, many of which are hydrophobic and cationic. Conventionally, Pgp is thought to function as an efflux transporter, although it is unclear whether any experiment has demonstrated unequivocally that Pgp mediates transport of a substrate across a membrane bilayer against its electrochemical gradient.
Several models have been proposed to account for the apparent function and remarkable variety of compounds that are recognized by this protein. Pgp may be an efflux transporter that recognizes substrates within the lipid bilayer ("hydrophobic vacuum cleaner") (1). Pgp has been hypothesized to be a pump with multiple binding sites for different drugs (2), a translocase for lipids (3), or a modifier of vesicular trafficking (4). Another model suggests that MDR1 Pgp indirectly alters partitioning of substrates within cells through effects on transmembrane potential (E m ), intracellular pH, and/or surface potentials and does not directly transport drugs (5). These disparate hypotheses may result from comparisons of data from transfected cells that have not been exposed to chemotherapeutic agents with cell lines in which expression of MDR1 is induced or maintained through exposure to drugs in the MDR phenotype. Pathways of resistance other than Pgp may exist in drug-selected cell lines, potentially confounding identification of properties attributable solely to MDR1. In addition, many studies have used "Pgp-negative" control cells that later were shown to express low levels of endogenous Pgp, further complicating interpretations regarding function of the protein itself.
We investigated MDR1 Pgp transport activity using MCF-7 breast adenocarcinoma cells, which do not express MDR1 (6), and MCF-7/MDR1 cells in which we established and maintained overexpression of MDR1 using a bicistronic vector in the absence of MDR drugs. Effects of Pgp on E m were measured directly by whole-cell patch clamping, and transport activity of the protein was studied with Tc-Sestamibi, an organotechnetium cationic substrate for Pgp (7,8). In the absence of MDR1 Pgp, Tc-Sestamibi accumulates within the mitochondrial matrix of living cells in response to negative mitochondrial inner membrane potentials (⌬) and E m while showing negligible nonspecific binding to lipids and proteins (9 -11). Tc-Sestamibi has no titratable proton, making accumulation within cells independent of intracellular pH (8,12). Using this system, we determined the effects of MDR1 Pgp on substrate content under steady-state and unidirectional influx or efflux conditions. As proposed over 25 years ago (13), we found that the dominant effect of Pgp is to establish a permeability barrier, limiting unidirectional influx of Tc-Sestamibi without affecting E m . EXPERIMENTAL (12). Calf serum was omitted for all electrophysiology experiments. A solution of 142 mM K ϩ and 20 mM Cl Ϫ was prepared by equimolar replacement of potassium methanesulfonate for NaCl (16).
Cell Culture, MDR1 Plasmid, and Transfection-MCF-7 cells were cultured in Dulbecco's modified Eagle's medium (Life Technologies, Inc.), 10% fetal bovine serum, and 0.1% penicillin/streptomycin in a 5% CO 2 incubator at 37°C. pGEM3Zf(Ϫ)Xba-MDR1.1 was purchased from American Type Culture Collection and sequenced to confirm wild-type identity. MDR1 was excised with XbaI and subcloned into pBluescript SK (Stratagene, La Jolla, CA). The cDNA for MDR1 was then removed with NotI and BamHI and ligated into the corresponding sites of pIRESneo (CLONTECH, Palo Alto, CA). Cells were transfected with MDR1 or vector using FuGene 6 (Roche Molecular Biochemicals). Clones of transfected cells were isolated and maintained in medium containing 1 mg/ml G418. Drug-sensitive (Pgp-negative) KB 3-1 and MDR (Pgp-positive) KB 8-5 human epidermoid carcinoma cells were grown and maintained as described (17).
Western Blots-Pgp was detected in enriched membrane fractions of cells with monoclonal antibody C219 (Signet Laboratories, Inc., Dedham, MA) as described previously by our laboratory (17).
Immunofluorescence Microscopy-Cells were processed for immunofluorescence microscopy with monoclonal antibody C219 as described for detection of Pgp in KB 8-5 and Chinese hamster ovary cells (18).
Cytotoxicity Assay-72-h cytotoxicity assays with daunomycin and paclitaxel were performed using sulforhodamine B (19). Data are expressed as percent growth relative to cells treated with vehicle only.
Electrophysiology-Whole-cell current-clamp or voltage-clamp experiments were carried out at room temperature using an Axopatch 200B amplifier (20). Micropipettes were pulled from thin-walled glass (WPI Inc., New Haven, CT) on a horizontal puller (Sutter Instrument Co., Novato, CA). Two different solutions were used in pipettes: K INT solution 1 contained 140 mM KCl, 10 mM K-HEPES, and 1 mM K-EGTA, pH 7.35 with KOH; and K INT solution 2 contained 140 mM KCl, 10 mM Na-HEPES, 1 mM K-EGTA, and 2 mM MgATP, pH 7.35 with KOH (10 Ϫ8 to 10 Ϫ9 M free Ca 2ϩ ). The bath solution was serum-free MEBSS without or with drugs as indicated. pClamp Version 6.0 software and a DigiData 1200 converter were used to generate command pulses and collect data. Data were filtered at 5 kHz. Off-line analysis was performed using ClampFit and Microsoft Excel programs. Current-voltage relationships were generated from steady-state currents at 300 ms. Data are presented as means Ϯ S.E.
Cellular Accumulation of 99m Tc-Sestamibi or [ 3 H]Daunomycin-Transport function and modulation of MDR1 Pgp under steady-state conditions were assayed with 99m Tc-Sestamibi (12). Transport assays with 99m Tc-Sestamibi were performed within 3 days of all electrophysiology experiments to confirm functional expression of Pgp in MCF-7/ MDR1 cells. To determine unidirectional influx of 99m Tc-Sestamibi or [ 3 H]daunomycin during short periods of incubation (Յ90 s), the protocol was modified as follows. 1) The concentrations of 99m Tc-Sestamibi and [ 3 H]daunomycin were 40 Ci/ml (5-10 pmol/mCi) and 0.2 Ci/ml (1.9 Ci/mmol), respectively; and 2) coverslips with cells were washed four times with ice-cold MEBSS. Cell-associated tracer is expressed as fmol/mg of cell protein/nM o , where cell content of Tc-Sestamibi or daunomycin (fmol) was normalized to mg of cell protein and extracellular concentration of tracer (nM o ).
For efflux experiments, cells were first incubated for 30 min at 37°C in MEBSS containing 10 Ci/ml 99m Tc-Sestamibi to reach steady-state accumulation of radiotracer (12). Coverslips were blotted briefly (Ϸ5 s) to remove excess buffer, transferred to isotope-free MEBSS (37°C) for various times, and then washed with ice-cold MEBSS. To measure steady-state content of 99m Tc-Sestamibi prior to efflux, incubation in isotope-free buffer was omitted. Accumulation of Tc-Sestamibi was quantified as described above and is expressed as a percentage of steady-state values for each cell line.
Data are expressed as l of intracellular water space/mg of cell protein.
Intracellular K ϩ Concentration-Cells were plated as described above for accumulation of 99m Tc-Sestamibi. Coverslips without cells were incubated in the same medium for use as controls, and values for these blank coverslips were subtracted from data for samples with cells. Cells were washed with phosphate-buffered saline prepared without potassium and extracted with 1% nitric acid for 30 min. Samples were microwaved to prepare for analysis, and each sample run included a 1% nitric acid solvent blank and an external standard. K ϩ content was determined by an inductively coupled plasma atomic emission spectrometer (IRIS Advantage Duo-View system, Thermo Jarrell Ash, Franklin, MA) using a method that scanned each sample three times and provided a data value with 1 level of confidence. 2 Data for g of K ϩ were normalized to mg of cell protein, and intracellular K ϩ concentration ([K ϩ ] i ) in each cell line was calculated by dividing K ϩ content by intracellular water space.

RESULTS
Characterization of MCF-7/MDR1 Cells-Because cell lines derived from stepwise selection with MDR drugs may have multiple mechanisms of drug resistance (22), we stably transfected MCF-7 cells with MDR1 and established clonal cell lines that express Pgp without use of MDR drugs. Pgp was detected specifically by Western blotting in several different clones that expressed the protein at levels comparable to or greater than KB 8-5 cells (Fig. 1A), a cell line derived from selection in colchicine (23). A clone (termed 3-4) with higher amounts of Pgp than in KB 8-5 cells was selected for further characterization. In the remainder of report, this clonal cell line is referred to as MCF-7/MDR1 cells. Pgp localized predominantly to the plasma membrane in MCF-7/MDR1 cells, as determined by immunofluorescence microscopy (Fig. 1B). No immunodetectable Pgp was present in control MCF-7 cells, as determined by Western blotting or immunofluorescence ( Fig. 1A and data not shown). Function of transfected MDR1 Pgp initially was characterized with Tc-Sestamibi. Compared with parental cells, steady-state accumulation of Tc-Sestamibi after 30 min of incubation was almost 70-fold less in MCF-7/MDR1 cells (55 Ϯ 11 versus 0.8 Ϯ 0.1 fmol/mg of cell protein/nM o , respectively). Radiotracer content in MDR1 transfectants increased to control values when cells were incubated with a saturating dose of LY335979 (1 M) (Fig. 1C), a specific inhibitor of MDR1 Pgp (15). LY335979 did not enhance accumulation of Tc-Sestamibi in MCF-7 cells, providing further evidence that this cell line does not express MDR1 Pgp (6,24). Previously, our laboratory has shown that KB 8-5 cells accumulate ϳ50-fold less Tc-Sestamibi than Pgp-negative parental KB 3-1 cells (17). Thus, these functional data are consistent with differences in relative expression of Pgp in MCF-7/MDR1 and KB 8-5 cells. Furthermore, to verify that transfected MDR1 conferred multidrug resistance to MCF-7/MDR1 cells, we performed 72-h cytotoxicity assays with doxorubicin and paclitaxel, two validated substrates for Pgp (15,25). Compared with control cells, MCF-7/ MDR1 cells were also ϳ100-fold more resistant to doxorubicin and at least 50-fold more resistant to paclitaxel, as determined by 72-h cytotoxicity assays (Fig. 1, D and E). Overall, these data demonstrate that functional MDR1 Pgp was expressed and localized correctly, conferring MDR to transfected MCF-7 cells. As determined by differences in accumulation of Tc-Sestamibi between MCF-7/MDR1 and control cells, function of Pgp was stable over at least 7 months of continuous culture (data not shown). Therefore, these matched cell lines provided an appropriate system for biophysical analysis of MDR1 Pgp without the confounding effects of prior exposure to MDR drugs.
Effects of MDR1 Pgp on Resting E m and Macroscopic Conductance-Previous studies suggest that MDR1 Pgp (26) and other ATP-binding cassette transporters (27) reduce E m in cells. To determine directly whether MDR1 Pgp alone affects resting E m , we compared the electrical properties of the cell lines by conventional whole-cell patch clamping. Using K INT solution 1 (no ATP) in the pipette and MEBSS in the bath solution, E m did not differ between cell lines, measuring Ϫ36.4 Ϯ 6.0 and Ϫ32.9 Ϯ 3.7 mV (n ϭ 7 and 9; p Ͼ 0.25) in control and MCF-7/MDR1 cells, respectively ( Fig. 2A). Because ATP-dependent K ϩ channels have been shown to influence E m of MCF-7 cells (28), we also determined E m using a pipette solution (K INT solution 2) that contained ATP. E m of MCF-7 and MCF-7/MDR1 cells were Ϫ42.8 Ϯ 6.6 and Ϫ47.6 Ϯ 3.3 mV (n ϭ 3; p Ͼ 0.48), respectively, using K INT solution 2 . For both cell lines, E m was more negative as measured with K INT solution 2, although differences were significant only for MCF-7/ MDR1 cells (p Ͻ 0.01). However, E m did not differ between control cells and Pgp transfectants under either experimental condition. In MEBSS, an ϳ2-log difference in E m would be required to account for the almost 70-fold reduction of Tc-Sestamibi in the Pgp-expressing line (Fig. 1), assuming a constant ⌬ in both cell lines. We also did not detect differences in macroscopic currents between MCF-7/MDR1 and control cells as measured with a pipette solution of either K INT solution 1 or 2 (Fig. 2, B and C; and data not shown). Measurements of E m and macroscopic currents were stable for at least 2 min, implying that the patch clamp did not disrupt membrane integrity and allow leakage of ions or macromolecules from cells. In addition, no difference in E m was measured between Pgp-negative KB 3-1 cells and Pgp-positive KB 8-5 cells (data not shown), for which colchicine (an MDR drug) was used in the latter to select and maintain expression of MDR1 Pgp.
If MDR1 Pgp were to confer MDR through alterations in E m , specific inhibitors of Pgp should hyperpolarize E m only in cells that express Pgp. To test this possibility, whole-cell patch clamping was performed with saturating doses of either LY335979 (1 M) or GF120918 (300 nM), another potent inhibitor of MDR1 Pgp (14). Although these inhibitors blocked drug transport by MDR1 Pgp within seconds after addition to cells (see below), neither E m nor macroscopic conductance was affected in either cell line during 2 min of monitoring (data not shown). Overall, the data directly demonstrate that expression of functional MDR1 Pgp without selection in MDR drugs does not alter basal E m or membrane conductance. We performed transport assays with the K ϩ ionophore valinomycin added to standard buffer (5.4 mM extracellular K ϩ ) (Fig. 3A). Because intramitochondrial and cytosolic K ϩ concentrations are approximately equal (30), these conditions were predicted to depolarize ⌬ toward zero and to hyperpolarize E m toward the K ϩ reversal potential. In MCF-7 cells, radiotracer content decreased from 43.2 Ϯ 4.0 to 2.3 Ϯ 0.6 fmol/mg of cell protein/nM o in the absence and presence of valinomycin (1 g/ml), respectively, consistent with depolarization of ⌬ as the dominant determinant for reduction of the accumulation of hydrophobic cationic compounds in these cells (9,12,31). Steady-state accumulation of Tc-Sestamibi in MCF-7/MDR1 cells was 0.8 Ϯ 0.1 fmol/mg of cell protein/nM o and could not be detected above background when valinomycin was added to 5.4 mM K ϩ buffer (threshold of detection of Ϸ0.1 fmol/mg of cell protein/nM o ). Addition of GF120918 (300 nM) or LY335979 (1 M) increased net accumulation of Tc-Sestamibi in MCF-7/ MDR1 cells to values observed in parental cells, whereas control cells were unaffected (Fig. 3A). These results indicate that MDR1 Pgp does not reduce cell content of Tc-Sestamibi by decreasing steady-state ⌬.

MDR1 Pgp Affects Accumulation of Tc-Sestamibi Even in the
To depolarize both E m and ⌬ to Ϸ0 mV, we equilibrated [K ϩ ] i and extracellular K ϩ concentration by incubating cells in 142 mM K ϩ and 20 mM Cl Ϫ buffer containing valinomycin (1 g/ml). Cl Ϫ was reduced to prevent high KCl buffer-induced increases in cell volume mediated by K ϩ /Cl Ϫ cotransporters expressed in mammalian cells (32). Under isoelectric conditions, steady-state content of Tc-Sestamibi in MCF-7 cells was reduced to 0.92 Ϯ 0.07 fmol/mg of cell protein/nM o , which is 2.5-fold less than that observed in buffer containing 5.4 mM K ϩ (Fig. 3B). By comparison, accumulation of Tc-Sestamibi in MCF-7/MDR1 cells was reduced to background levels. Thus, MDR1 Pgp either transported tracer out of the cells against a significant concentration gradient to levels well below those expected for passive distribution into intracellular water space (calculated as 3.3 fmol/mg of cell protein/nM o ) or produced a diffusion barrier that prevented equilibration of an inwardly directed concentration gradient for the tracer. When Pgp was inhibited with GF120918 (300 nM) or LY335979 (1 M), accumulation of Tc-Sestamibi in MCF-7/MDR1 cells increased to that observed in MCF-7 cells. Thus, although these data demonstrate that net content of Tc-Sestamibi was reduced in fully depolarized cells, MDR1 Pgp lowered cell content of radiotracer to an amount less than that produced by reductions in membrane potentials alone.
Effects of MDR1 Pgp on Unidirectional Influx and Efflux of Substrate-Steady-state reduction in cell content of Tc-Sestamibi produced by Pgp could be due to alterations in influx (permeability) and/or efflux (active transport). To determine the relative contribution of each component, we first quantified cell-associated Tc-Sestamibi over the initial 90 s of exposure to radiotracer. Uptake of Tc-Sestamibi in control cells was linear throughout this period and did not reach steady state until Ϸ30 min, indicating that these early time points represent unidirectional influx of radiotracer (Fig. 4A) content of Tc-Sestamibi in MCF-7/MDR1 cells was ϳ7-fold less than in control cells. Furthermore, the effects of MDR1 Pgp were completely reversed within 10 s by GF120918 (300 nM) or LY335979 (1 M) (Fig. 4, B and C), providing additional evidence that MDR1 Pgp markedly reduces influx of Tc-Sestamibi. To determine whether Pgp also limits unidirectional influx of a different substrate, we measured net content of daunomycin at time points between 10 and 90 s in these cells. Similar to Tc-Sestamibi, cell content of daunomycin in MCF-7/ MDR1 cells was less than in control cells at all time points, although differences did not become statistically significant until 30 s of influx (Fig. 4D). However, unlike Tc-Sestamibi, cell-associated daunomycin in the MDR1 transfectants did not reach a plateau during the initial 90 s of incubation, which may be due to partitioning of anthracyclines into lipid bilayers (33).
To determine whether MDR1 Pgp also affects unidirectional efflux of Tc-Sestamibi, cells were incubated with Tc-Sestamibi for 30 min to reach steady-state levels of radiotracer and then transferred to isotope-free buffer (zero trans-conditions). During the initial 30 s of efflux, cell-associated Tc-Sestamibi was not affected by Pgp, as evidenced by an ϳ30% decrease in radiotracer in both cell lines (Fig. 5A) fectants, respectively. Comparable results were observed when parental cells were loaded with 40-fold less radiotracer than MCF-7/MDR1 cells to achieve similar amounts of cell-associated Tc-Sestamibi at the start of efflux (data not shown). When incubations in isotope-free buffer were extended to longer periods of time, significant differences between cell lines were not detected until 5 min and were maintained over the ensuing 30 min (Fig. 5B). Control cells retained 67.1 Ϯ 8.4 and 26.2 Ϯ 4.7% of the initial content of Tc-Sestamibi after 5 and 30 min of efflux, respectively, compared with 22.2 Ϯ 2.4 and 7.9 Ϯ 2.2% in MCF-7/MDR1 cells, respectively. During these longer incubations, LY335979 (1 M) increased cell content of Tc-Sestamibi in MCF-7/MDR1 cells to levels observed in parental cells (Fig. 5B). However, these extended incubation times yielded data reflecting the combined effects of efflux and contaminating influx (re-entry) of substrate into cells and no longer isolated unidirectional efflux kinetics. Overall, these studies demonstrate that MDR1 Pgp has no significant effect on initial efflux of Tc-Sestamibi from MCF-7/MDR1 cells.
Although MDR1 Pgp did not affect the resting E m of cells as measured by whole-cell patch clamping, we considered that Pgp could directly or indirectly alter the dipole potential within the plasma membrane, producing a more positive intramembranous potential and limiting influx of cationic substrates like Tc-Sestamibi. Accordingly, TPB, a lipophilic anion that imposes a negative dipole potential within lipid bilayers (34), should reduce the effective intramembranous potential and preferentially enhance accumulation of Tc-Sestamibi in MCF-7/MDR1 cells. Previously, we showed that TPB enhances accumulation of Tc-Sestamibi in cultured cells, although the effects of MDR1 Pgp were not determined (12). In the presence of TPB, Tc-Sestamibi content after 30 s of influx increased in both cell lines in a concentration-dependent manner (Fig. 6A). However, the -fold change in accumulation of radiotracer with TPB was significantly higher in Pgp-expressing cells. Influx of Tc-Sestamibi increased by 50-and 360-fold in control and MCF-7/MDR1 cells, respectively, when 30 M TPB was added to the buffer (Fig. 6B). At concentrations Ͼ30 M TPB, accumulation of radiotracer decreased in both cell lines, likely due to toxicity (data not shown). When Pgp was inhibited with LY335979 (1 M), influx of Tc-Sestamibi in the presence of TPB did not differ between MCF-7/MDR1 and parental cells, further indicating that functional MDR1 Pgp mediated the differential effects of TPB on accumulation of radiotracer. DISCUSSION We have critically examined potential mechanisms of substrate transport by MDR1 Pgp using stringent reagents and tools to determine the effects mediated solely by Pgp. Our results demonstrate that MDR1 Pgp can confer MDR without affecting resting E m or macroscopic conductance of the plasma membrane in cells not previously exposed to MDR drugs. Conversely, under conditions that depolarized E m and ⌬, MDR1 Pgp reduced net cell content of Tc-Sestamibi to levels below those produced by external manipulation of membrane potentials alone, showing that transport of Tc-Sestamibi by Pgp is not dependent on changes in transmembrane potential. Furthermore, specific inhibitors of MDR1 Pgp blocked transport activity without altering E m or macroscopic conductance of Pgp-positive or Pgp-negative cells.
MDR1 Pgp has been associated with a variety of perturbations in ion conductance. For example, enhanced Na ϩ currents were measured in cell lines selected for expression of Pgp through drug exposure (35,36), but changes in Na ϩ conductance were not coupled to drug transport (36). Despite changes in Na ϩ currents, E m did not differ between drug-sensitive and drug-selected MDR cells, although the effects of MDR1 potentially could have been masked by the effects of selection in chemotherapeutic drugs. Prior studies with LR73 cells reported that the resting E m of Ϫ46 mV in these cells was reduced by ϳ20 mV in cells transfected with MDR1; these measurements were made fluorometrically with K ϩ /valinomycin null point titration (37,38). In the present study, E m in control MCF-7 cells was Ϫ43 mV, as determined by whole-cell patch clamping with ATP in the pipette solution (K INT solution 2), and was not significantly different in MDR1 transfectants. These data argue strongly against a hypothesis that Pgp depolarizes only cells that are above a threshold value for E m . Furthermore, the magnitude of depolarization attributed to MDR1 Pgp in previous studies cannot account for the large differences in accumulation of Tc-Sestamibi between control and MCF-7/MDR1 cells.
Ion channel activities have been associated with expression of MDR1 Pgp in response to various stimuli. Pgp has been reported to mediate ATP-and Na ϩ -dependent Cl Ϫ /H ϩ antiport in response to rapid changes in extracellular ion gradients, thus alkalinizing cells and contributing to MDR (39). MDR1 Pgp has also been reported to function as, or regulate, a swelling-activated Cl Ϫ channel (40). Our study did not investigate ion conductances stimulated by changes in ion gradients or channel activity during hyposmotic stress, but our data indicate that these conditions are not necessary to activate drug transport. Although our data do not exclude possible channel activity associated with MDR1 Pgp under selected conditions, the results show that Pgp does not change steady-state currents in MCF-7 cells and imply that such alterations are not essential for drug resistance.
The present data show that MDR1 Pgp reduces intracellular content of Tc-Sestamibi by limiting influx without affecting unidirectional efflux. Most studies of MDR1 Pgp have reported that the transporter both reduces influx and enhances efflux of substrates, although exceptions have been described in which Pgp affects only influx (41) or efflux (42) of compounds. As indicated by data from our efflux protocol, substrate content in non-polarized MCF-7/MDR1 cells was reduced only at time points beyond the experimental conditions of unidirectional efflux. Differences between our data and previous studies could be due to prior determinations at time points beyond the period of unidirectional efflux, the presence of multiple mechanisms of MDR in drug-selected cells, use of substrates with varying amounts of partitioning into membranes or trapping in subcellular compartments, or disparities between the effects of Pgp in non-polarized versus polarized cells. Nevertheless, our system shows that transfection of MDR1 functionally establishes a permeability barrier to influx of Tc-Sestamibi. The "membrane vacuum cleaner" model for Pgp in effect predicts the same behavior (43). However, our previously published data show an identical in-to-out content ratio for Tc-Sestamibi in Pgp-expressing cells when assayed over a 7-log range of extracellular tracer concentration (pM to 10 M) both in the absence and presence of the MDR modulator quinidine (12). The lack of sigmoidal activation with no evidence of saturable behavior at high substrate concentrations is attributed best to a process dominated by diffusion rather than enzyme kinetics (39). As a whole, these data favor a model whereby Pgp establishes a permeability barrier to influx of substrates.
The mechanism through which Pgp establishes a permeability barrier remains to be identified. TPB preferentially enhanced influx of Tc-Sestamibi in MCF-7/MDR1 cells, suggesting that Pgp may act by increasing the net positive dipole potential within the plasma membrane. The dipole potential is manifest between the hydrophobic interior of a bilayer and water molecules immediately adjacent to lipid head groups (44). In bilayers of phosphatidylcholine, the dipole potential within the membrane is calculated to be approximately ϩ280 mV (45), which impedes diffusion of hydrophobic cationic compounds such as Tc-Sestamibi. If MDR1 Pgp increased the intramembranous dipole potential to more positive values, influx of Tc-Sestamibi or other positively charged hydrophobic compounds would be retarded. TPB, a hydrophobic anion, is predicted to associate with phospholipid membranes near the polar head groups, thereby reducing intramembranous dipole potentials and enhancing the kinetics of cation translocation (46). Ion pairing effects on membrane solubility of substrates also may contribute to the effects of TPB.
MDR1 Pgp potentially could affect intramembranous dipole potentials by altering the distribution of lipids, such as sphingomyelin (47) and glucosylceramide (48), between inner and outer leaflets of the plasma membrane. Pgp localizes to low density membrane domains (18,49), in which the magnitude of the dipole potential is increased due to enrichment with cholesterol (50). Interestingly, although TPB behaves as a modulator of Pgp, tetraphenylphosphonium, a hydrophobic cation that is the counterpart of TPB, is a substrate for the transporter (51). Verapamil, a classic inhibitor of MDR1 Pgp, has also been shown to decrease the intramembranous dipole potential in lipid vesicles (52), which would directly enhance influx kinetics of hydrophobic cations like Tc-Sestamibi. Consistent with this model, the effects of TPB on dipole potential would be expected to affect the kinetics of transmembrane permeation, but not the final steady state (46). Accordingly, the maximal effects of TPB on cell content of Tc-Sestamibi were identical between Pgp-expressing and Pgp-non-expressing cells. Another mechanism related to this model is that altering the dipole potential indirectly inhibits Pgp transport by chang-ing lipid-protein interactions on the intramembranous surface of the transporter. Previous work has shown that function of MDR1 Pgp is affected significantly by changes at the lipidprotein interface (53,54).
In summary, we have shown that MDR1 Pgp significantly decreased influx of an organotechnetium cationic substrate without affecting resting E m . However, Pgp did not alter unidirectional efflux of Tc-Sestamibi. These data suggest that the dominant function of MDR1 Pgp in non-polarized cells is to produce a permeability barrier for Tc-Sestamibi, perhaps by altering the dipole potential within plasma membranes. These results provide features that should be incorporated into any comprehensive model of the mechanism of action for this transporter.