Membrane transport of tetraphenylphosphonium and its homologues through the planar phospholipid bilayer: Concentration dependence and mutually competitive inhibition in membrane passive transport

Membrane Transport of Tetraphenylphosphonium and Its Homologues through the Planar Phospholipid Bilayer: Concentration Dependence and Mutually Competitive Inhibition in Membrane Passive Transport SElJl

MIYAUCHI,ATSUSHIONO, MINORUYOSHIMOTO, AND NAOKIh M O X

Received April 30,1991, from the Department of Biophysics and Physicochemistry, Faculty of Pharmaceutical Sciences, Hokkaido University, Sapporo 060, Japan. Accepted for publication April 29, 1992. Abstract 0 The concentration dependence and mutually competitive

inhibition in the membrane transport of the lipophilic cations tetraphenylphosphonium (TPP') and its homologues were studied by use of a planar phospholipid bilayer. The current-voltage characteristics were analyzed by use of a model developed by Ketterer et al. (J. Membr. Biol. 1971,5, 225245). The conductance at 0 mV [G(O)],which represents the transport ability, increased with an increase in concentration, and the G(0) values revealed saturation at higher concentrations. The relationship between GfO) and concentration was well fitted to a MichealisMenten-type equation with a saturable component {maximum conductance [G(O),,] and concentration yielding one-half of G(O),, [KJ}. The G(O),, and K, values were calculated to be 4.1 x 10-'-242 x lo-' S/cm2and 13489 pM, respectively, depending on the lipophilic cations used. The G(0)values normalized by the corresponding G(O),, values were plotted against the concentrationsnormalized by the corresponding K, values. Normalized curves for TPP+ homologues were all superimposed on a single curve. The mechanismfor the saturation of G(0)values may involve the adsorption of TPP+ homologuesto a planar phospholipid bilayer, because the K,, values calculated in the present study were comparable to those for adsorption to phosphatidylcholine liposomes. We further determined the mutual inhibition of membrane transport by these homologues.Tetraphenylmethylphosphoniumat 1 mM, a concentration threefold higher than the K,value for this compound, reduced the G(0)for other TPP+ homologues by 3&55%, whereas reduction by 10 mM tetraethylammonium was minimal. The type of inhibition was classified as mutually competitive. In experiments on membrane transport, it is often assumed that the observation of saturation and mutually competitive inhibition indicates the presence of a carrier. The present paper, however, presents facts contradicting this assumption: greater caution should be taken in interpreting transport data to indicate carrier-mediated transport.

A carrier-mediated transport system shows the following kinetic featuresl-3: (1)the relationship between the transport rate and concentration displays saturation; (2) the transport process is mutually inhibited by structurally similar compounds; and (3) the system displays a countertransport phenomenon. Many investigations have been performed with whole animals,"7 isolated perfused organs,S-lO isolated cells,11-13 and isolated membrane vesicles1"lg to demonstrate that the transport system under consideration is carrier mediated. Many investigators assume the following criteria to prove the presence of a transport carrier: the saturation phenomena of the transport rate versus concentration andor mutually competitive inhibition. Although these phenomena are characteristics of carrier-mediated transport, are observations of them necessary and satisfactory conditions to confirm such transport? In the present study, we examined the current-voltage (I-V) relationship of the lipophilic cations tetraphenylphosphonium (TPP') and its homologues by use of a planar phospholipid bilayer. The obtained I-V relationship was well fitted to 0022-3549/93/0 100-0027$02.50/0 0 1993. American Pharmaceutical Association

the model developed by Ketterer et al.20; this model assumes that the mechanism of transport of lipid-soluble ions through the bilayer membrane consists of three distinct steps: (1) adsorption to the membrane-solution interface; (2) passage over an energy barrier to the opposite interface; and (3) desorption into the aqueous solution. On the basis of this model, analysis of the obtained I-V relationship yields the parameters governing membrane transport. We found that the transport of TPP' and its homologues displays the kinetic features of saturation, mutually competitive inhibition, and countertransport. It should be noted that these kinetic features are used as criteria for carrier-mediated transport, as described above, although the lipid-soluble transport system for TPP+ and its homologues is evidently passive. Therefore, we conclude that greater caution should be taken in interpreting kinetic data for membrane transport to indicate carrier-mediated transport.

Experimental Section Chemicals-TPP+ was purchased from Dojindo Laboratories (Kumamoto, Japan). Triphenylethylphosphonium (TPEP? and triphenylpropylphosphonium (TPPP+)were purchased from Aldrich Chemical Co., Inc. (Milwaukee, WI). Triphenylmethylphosphonium (TPMP+),triphenylbutylphosphonium(TPBP+),triphenylamylphosphonium (TPAP+), and triphenylhexylphosphonium (TPHP? were purchased from Tokyo Kasei Industry, Ltd. (Tokyo,Japan). Asolectin, N,N-p-phenylenediamine, and tetraethylammonium (TEA+) were purchased from Sigma Chemical Co. (St. Louis, MO). All other reagents were commercial products and were of analytical grade. Purification of Phosphatidylcholine (PC)-Partial purification of PC from asolectin was performed as described by Kagawa and Racker.21 Asolectin (25 g) was washed with 500 mL of dry acetone containing 2.5 mg of an antioxidant (N,N-p-phenylenediamine)at 4 "C. The insoluble fraction was suspended in 100 mL of diethyl ether containing 100 pg of the antioxidant, and the suspension was centrifuged. The soluble fraction was concentrated to dryness under reduced pressure at 4 "C. The dried material was dissolved to a 10% (w/v) concentration in chloroform. This solution was stored at -25 "C in a bottle filled with argon gas. In this form, it remained active in membrane experiments for at least 1year. For the experiments, a 20-pL portion of the stock solution was evaporated with argon gas and dissolved in 100 pL of n-hexane. This solution was designated the lipid bilayer formation solution (PC concentration, 20 mg/mL). Formation of the Planar B i l a y e v A l l procedures described below were carried out with the temperature controlled at 25°C. In accordance with the method originally developed by Montal and Mueller,22 a planar bilayer membrane was formed with PC. The apparatus used was made of Teflon and consisted of two chambers (volume, 2.0 mL each) separated by a septum (membrane kit 5775, Yellow Springs Instruments, Yellow Springs, OH) containing in its center a hole of 1.0 cm (diameter) covered by a thin Teflon film (12.5-pm thick). The film is used for a n oxygen electrode. A planar phospholipid bilayer appeared to be formed at an aperture (diameter, 180-250 pm) in this Teflon film. Both chambers in the apparatus were Journal of Pharmaceutical Sciences I 27 Vol. 82,No. 1, January 1993

filled with 1.5 mL of 50 mM tris(hydroxymethy1)aminomethane (Trid-HCl buffer (pH 7.4). The buffer in each chamber was aspirated with a disposal syringe (Terumo, Tokyo, Japan), and the water level was lowered to just below the aperture. A portion (5 pL) of the lipid bilayer formation solution was added to each chamber, and a monolayer was formed spontaneously at the air-water interface within a few minutes, as described by Schindler.23 After 5 min, the water level in each chamber was gradually raised to above the aperture, and the two monolayers were folded to form a planar bilayer at the aperture. The temperature in each chamber was monitored with a thermocouple. Measurement of Membrane Current-All procedures and measurements were carried out by use of a laboratory-made shield box with the temperature controlled at 25°C. The shield box was fabricated to minimize the noise caused by mechanical vibration or sound. Electricity was passed through a pair of Ag-AgC1 electrodes. The electric current was measured with an I-V converter, which contained an operational amplifier (CEZ-2200; Nihon Kohden, Ltd., Tokyo, Japan) with a 10-Gs1feedback resistor, and was recorded with a pen recorder after filtration through a passive low-pass filter (time constant, 0.1-1 8). For estimation of the membrane capacitance, a square wave voltage of 1mVpp at 60 Hz was applied to one chamber, and the capacitive current across the membrane through the low-pass filter after conversion to voltage by the I-V converter was monitored on an oscilloscope. The capacitive current was linearly proportional to the capacitance up to 500 pF. The membranes were considered appropriate for the subsequent study when the following criteria were fulfilled: the membrane resistance was more than 200 Gs1, and the membrane capacitance was within the range of 550-650 nF/cm2. After 30 min of stabilization, a 20 pL portion of the lipophilic cation solution was added to each chamber. The transmembrane current was determined with voltage increased incrementally (10-mV increments). The phospholipid bilayers used in the present study are good insulators in an aqueous solution for small ions, such as Na+ and K+. Thus, the transmembrane current was due only to the membrane transport of a lipophilic cation used here. We confirmed that various concentrations of salts, such as Na+ and K+, did not alter the I-V relationship (data not shown). Analysis of I-V Curves-We analyzed I-V characteristics by using the model developed by Ketterer et al.20 They assumed that the mechanism by which the transport of lipid-soluble ions through bilayer membranes occurs has three distinct steps: (1)adsorption to the membrane-solution interface; (2) passage over an energy barrier to the opposite interface; and (3) desorption into the aqueous solution. If a voltage V is applied across the membrane, the current observed at the steady state, J O , is determined by the transport of lipidsoluble ion species and is expressed as follo~s24.26:

2zF/3CKtexp(- mu2)sinh(u/2) Jo= 2A exp( - mu2) cosh(d2) + 1

(1)

In eq 1, z is the valency of the lipid-soluble ion; F is the Faraday constant; p is the partition coefficient for the lipid bilayer; C is the lipid-soluble ion concentration in the chamber; Kt is the rate constant for the translocation of the lipid-soluble ion across an energy barrier to the other interface in the absence of external voltage; o is a function of membrane thickness (w = 0.007, based on Table I in Table I-Mlchaells-Menten-TypeKinetlc Parameters' for TPP+ and TPP+ Homologue Membrane Transpolt through the Phosphollpld Bllayer

Substrate TPP+ TPMP+ TPEP+ TPPP+ TPBP+ TPAP+ TPHP+ a f

1 o-8 S/cm2

24f 3 370 f 50 107 f 41 409 f 72 400 2 67 222 4 132 2

57 f 6 4.1 f 0.3 37 f 5 144 f 13 242 231 99 2 12 135 2 12

Computer-calculated SDs. Values were calculated on the basis of

28 I Journal of Pharmaceutical Sciences Vol. 82, No. 1, Januaty 1993

sinh(d2) GW) 2 - [l + 2A exp(-ou2)1 (2) G(0) u 2A exp(-ou2) cosh(d2) + 1 In eq 2, G(V) is conductance and G(0) is conductance at 0 mV. On the basis of previous studies2k26 on the membrane transport of a lipophilic cation, the normalized conductance fits the case of A = 0, indicating that K > > Kt. For low voltages (Id21 < < l), the high-order terms of the series expression for the hyperbolic sine can be neglected; hence, eq 2 is reduced to the following:

GW) = G(0)+ aV2

(3)

In eq 3,

(4) and

On the basis of eq 4, the G(0) value reflects the membrane transport ability in the absence of the membrane potential. Note that Kt corresponds to the mobility of a lipophilic ion through the membrane. The G(0) values were calculated by fitting the conductance curves to eq 3 by use of a nonlinear least-squares method with a microcomputer (PC-9801W, NEC Co., Tokyo, Japan). The G(0) values were not proportional to the concentrations and were therefore analyzed with a double-reciprocal plot (data not shown). This plot revealed a straight line; therefore, we fitted the G(0) data to a MichaelisMenten-type equation with a saturable component: (6) In eq 6, G(O),, and K, represent the maximum conductance (Siemens per square centimeter) and the concentration (micromolar) yielding one-half of G(O),,. The G(O),, and K, values were calculated on the basis of eq 6 with a nonlinear least-squares method. When competitive inhibition by other TPP+ homologues is assumed, the following equation can be used

(7) In eq 7,

(8) In eq 8,.Km:app is the K , value in the presence of the inhibitor, Kiis the inhibition constant, and I is the concentration of the inhibitor. The Km,app value was calculated by fitting the data for the inhibition experiments to eq 7 with a nonlinear least-squares method. The Ki value was then calculated on the basis of eq 8.

G(O)max,

Km* PM

eq 6.

reference 24); u is reduced voltage, zFV/RT; R is the gas constant; T is absolute temperature; A is the ratio of the Kt value to the K value; and K is the adsorption rate constant for the lipid-soluble ion. In the present experiment, the conductance is measured as a function of the applied voltage. A rearrangement of eq 1to solve for the normalized membrane conductance yields the following24-26:

Results Concentration Dependence of the Kinetic Parameter for Membrane Transport [G(O)]-Figure 1 shows the typical conductance-voltage characteristics at various concentrations. These characteristics revealed parabolic curves, and G(0) was calculated by fitting these curves t o eq 3, with a nonlinear least-squares method. When the G(0) values were plotted against concentrations, a hyperbolic curve demonstrating saturable membrane transport was obtained (Figure

Q

F

2oo

t

<

t

W

"

0

20

Vo4Ptage(AV)

80

100

Figure 1-Voltage

dependence of membrane conductance in the presence of TPHP+(25 "C, pH 7.4). A portion (50 pL) of TPHP+ solution was added to each chamber filled with 1.5 mL of 50 mM Tris-HCI buffer (pH 7.4), and after 5 min of stirring, the transmembrane current was measured. Final TPHP+concentrations were 7 (A),13 (O),and 25 (m) FM.

2). Conductance data were well fitted with the MichaelisMenten-type equation with a saturable component. The G(O),, and K, values were calculated to be 4.1 x 10-8-242 x lo-' S/cm2 and 1 3 4 8 9 pM, respectively (Table I). Figure 3 depicts plots of G(O)/G(O),, against CIK, (C is concentration), which show the relationship between normalized conductance and concentration. All curves for the G(0)concentration characteristics were superimposed on a single line, which represents 1/(1+ x), indicating that the saturation of G(O)may be attributable to the same mechanism. Inhibitory Effect of TPMP+ on TPP+ Homologue Membrane Transport-The inhibitory effect of TPMP+ on TPP+ homologues is summarized in Figure 4. The concentrations of TPMP+ and TPP+ homologues were 1.0 mM and lower than their K, values (20-450 pM), respectively. TPMP+ had a significant inhibitory effect, ranging from 38-55%, on TPP+ homologue membrane transport. On the other hand, 10 mM TEA+ inhibited TPP+ homologue membrane transport only slightly (data not shown). A kinetic study of this inhibitory effect was performed with TPHP+, and the results are shown in Figure 5. TPMP+ inhibited TPHP+ membrane transport in a competitive manner. The inhibition constant for TPMP+ was calculated to be 540 2 144 pM (+. computer-calculated SD),a value that was compatible with the K, value.

300

loo

P 200

400

600

800

1000

Concentration (p M) Figure 24oncentration dependence of the membrane, transport for TPP+ and its homologues. The G(0) value was calculated from fitting conductance-voltage characteristics (Figure 1) to eq 3 with a nonlinear least-squares method. The lines were generated from eq 6 with the G(O),, and K,,,valueslisted in Table I.Key: (0)TPEP+; (0)TPAP+; (0)

TPHP+.

0 0

.0001 .001

.01

.1

1

10

100

C/Km Flgure 3-Relationship between G(0)values normalized by G(O),, and concentrations (C) normalized by K,. The G(O),, and K, values were calculated on the basis of eq 6 with a nonlinear least-squares method. Key: (x) TPP+; (0)TPMP+;(+) TPEP+;(W) TPPP+;(0)TPBP+; (m) TPAP+; (0)TPHP+.

TPPP TPBP TPAP TPHP TPP 323pM 323pM 13.2pM 3.3pM 9.9pM Flgure klnhibitoty effect of TPMP+ at 1.O mM on TPP+ homologue membrane transport at 25°C and pH 7.4. Portions (50 pL each) of TPMP+ and TPP+ homologues were simultaneously added to each chamber filled with 1.5mL of 50 mM Tris-HCI buffer (pH 7.4). After 5 min of stirring, the transmembrane current was measured. TPP+homologue concentrations were lower than their K,,, values. The open and closed columns represent TPP+ homologue membrane transport in the presence and in the absence of 1 mM TPMP', respectively. Numbers represent the percent inhibition by TPMP+.

Discussion

r

0

00

The purpose of this study was to demonstrate whether membrane transport of TPP+ and its homologues across the planar phospholipid bilayer (artificial membrane) would reveal criteria to demonstrate the presence of a carriermediated transport system: (1)saturation, (2) mutually competitive inhibition, and (3) countertransport. These lipophilic cations have been used to estimate plasma membrane potential changes in Ehrlich ascites cells27,28 and neuroblastoma-glioma hybrid cells29 and in several microorganisms.30-33 These probes are considered to be accumulated by the membrane potential, and the transport system is considered to be a process of passive diffusion through the plasma membrane. In general, one would assume that the membrane transport of these probes would reveal no kinetic feature, such as saturation or mutually competitive inhibition, that a carrier-mediated transport system would possess. As shown in Figure 2, however, the transport of TPP+ and its homologues did show saturation behavior. Furthermore, TPMP+, for which membrane permeability was lowest, inhibited the membrane transport of the other TPP+ homologues by 38-55%, whereas 10 mMTEA' slightly inhibited it. Journal of Pharmaceutical Sciences I 29 Vol. 82, No. 1, January 1993

Artificial planar phospholipid bilayers are now widely studied as models for biological rnembrane~.22~3~.~ The properties of planar bilayers formed from monolayers are identical to those of biological membranes in various aspects: membrane resistance,39.40capacitance,39,41and thickness.42 These models have an advantage in that even minimal transport of an ionic substance can be detected as minimal current by an operational amplifier, which is usually used in patch-clamp experiments. Planar phospholipid bilayers can be used in investigating transmembrane charge transport, such as that of lipid-soluble ions and charged carrier complexes. Furthermore, for passively transported molecules, we can estimate -0.1 0.0 0.1 0.2 0.3 0.4 the membrane permeability in a cell system from the G(0) 1/C (l/pM) values by considering the surface area of the cell. However, Figure 5-Double-reciprocal plot of TPHP+ membrane transport in the this method has some drawbacks: (1) substances without absence (0) and in the presence (0) of 1 mM TPMP+. TPHP+ charges cannot be tested, and (2) transmembrane currents concentrations ranged from 3 to 25 p M . The solid lines were drawn by due to many charged species cannot be distinguished. In the linear regression analysis. The G(O),, values in the absence and in the present study, we were unable to determine the mutual presence of TPMP+ were calculated to be 135 (2 12)and 159 (-t 46) inhibition among homologues for which permeability was x lo-* Slcm2 (2 computer-calculated SDs), respectively. The Km,app high because of the drawbacks described above. Therefore, values in the absence and in the presence of TPMP+were calculated to TPMP+, for which permeability was lowest, was chosen as an be 13 (2 2) and 37 ( 2 5) pM (rt_ computer-calculatedSDs), respectively. inhibitor. In fact, the G(0)value of TPMPf at 1m M is -5% The K, value was then calculated, on the basis of eq 8, to be 540 (? 144) p M (2 computer-calculatedSD). C, concentration. those of the other homologues, and the effect of the TPMP+ transmembrane current on the other homologues was minimal. This type of inhibition was classified as mutually competitive It is clear from the Ketterer mode120.24-26 that lipid-soluble (Figure 5). The kinetic features used as criteria to prove the transport and carrier-mediated transport may display a number of common kinetic features, such as saturation, presence of carrier-mediated transport can be observed in the competitive inhibition, and countertransport. How, then, can artificial phospholipid bilayer, although this membrane one distinguish between these two different mechanisms? The transport is passive. essential difference between a lipid-soluble transport system Another criterion usually used, other than the three described above, is the temperature coefficient (Q10),11-13,34,35 and a carrier-mediated transport system is that the former shows adsorption on both sides of the membrane, so that which is determined by the activation energy in the membrane. We recently determined the temperature dependency molecules in the two surrounding solutions have simultaneous access to the transport pathway. In contrast, the binding of the membrane transport of these cations (unpublished data). The Arrhenius plots for TPP+ and its homologues site of a carrier is accessible from only one side of the revealed a straight line, and the activation energies were membrane at a time and alternates between the two sides. calculated to be -15-30 kcal/mol. The Qlo values correspondThus, a carrier necessarily involves a mobile or transitional ing to these activation energies ranged from -3-5. These element, both of which are described in the excellent review values, however, suggest the presence of a carrier-mediated by LeFevre.1 In either event, the restriction that solutes on both sides of the membrane do not have simultaneous access transport system,34although this membrane transport mechanism is passive. Therefore, these Qlo values may not repreto the binding site is the sine qua non of the carrier hypothesis sent a sufficient criterion, either. that distinguishes this transport mechanism from the lipidsoluble transport mechanism. Criteria for distinguishing An important question is why these kinetic features, which between the lipid-soluble transport system and the carrierare present in a carrier-mediated transport system, were observed despite the absence of a carrier. As stated above, the mediated transport system should be direct consequences of this fundamental difference. In this sense, the phenomenon of transport of lipophilic cations is well described by the model “uphill”countertransport is the most widely accepted positive developed by Ketterer et a l . 2 0 According to this model,20 lipidsoluble ions are transported through the bilayer membrane via criterion for the implication of a carrier.1-3 If one unequivothree distinct processes: (1) adsorption to the membranecally demonstrates the presence of a carrier-mediated transsolution interface; (2) passage over an energy barrier to the port system by using only a kinetic approach, he or she will certainly find the phenomenon of uphill countertransport. opposite interface; and (3) desorption into the aqueous solution. It should be noted that ion species adsorbed to the membrane can only pass over the energy barrier inside the membrane. If References and Notes one takes into account the fact that the number of ions adsorbed 1. LeFevre, P. In Current Topics in Membranes and Transport; to the membrane-solution interface is limited, the saturation Bronner, F.; Kleinzeller, A., Eds.; Academic Press: New York, behavior is easily understood. In fact, the study by Demura et 1975; Vol. 7, p 109. al.36 with PC liposomes demonstrated that these probes showed 2. Schults, S. G. Basic Principles of Membrane Transport; Carnbridge University Press: Cambridge, 1980; pp 95-115. Langmuir adsorption, indicating the limitation in the number 3. Stein, W. D. The Movement of MoleculesAcross Cell Membranes; of lipophilic cations that can be adsorbed to the membraneAcademic Press: Orlando. FL. 1986: DD 363474. solution interface. According to Andersen and Fuchs,24 it is 4. Scharschmidt,B. F.; Waggoner, J. G.: Berk, P. D. J. Clin. Invest. likely that the adsorbed ions are localized in the layer of the 1975,56, 1280-1292. polar head groups of the lipid molecules. Whereas the hydro5. Nightingale, C. H.; Greene, D. S.; Quintiliani, R. J. Pharm. Sci. 1975, 64, 1899-1927. carbon tails have a tendency to remain as closely packed as 6. Iga, T.; Klaassen, C. D. J. Pharmacol. Exp. Ther. 1979, 211, possible for reasons of energy, the head groups are pushed aside 690-697. by the adsorbed ions. Furthermore, when one considers that 7. Tsuji, A.; Terasaki, T.; Tamai, I.; Takeda, K. J. Pharmacol. Exp. Langmuir adsorption to the bilayer shows some specificity, the Ther. 1990,253, 315-320. mutually competitive inhibition and countertransport can also 8. Addison, J. M.; Burston, D.; Dalrymple, J. A.; Matthews, D. M.; be easily understood. Payne, J. W.; Sleisenger, M. H.; Wilkinson, S. Clin. Sci. Mol. 6r

30 I Journal of Pharmaceutical Sciences Vol. 82, No. 1, January 1993

Med. 1975,49,313-322. 9. Reichen, J.; Paumgartner, G. Am. J.Physiol. 1976,231,734-742. 10. Wolkoff, A. W.; Goresky, C. A.; Sellin, J.; Gatmaitan, 2.;Arias, I. M. Am. J. Physiol. 1979,236,E638-E648. 11. Schwarz, L.; Burr, R.; Schwenk, M.; P f d , E.;Greim, H. Eur. J. Biochem. 1975,55,617-623. 12. Schwenk, M.; Burr, R.; Schwarz, L.; PfafF, E. Eur. J. Biochem. 1976,64,189-197. 13. Tsuji, A.; Terasaki, T.; Takanosu, K.; Tamai, I.; Nakashima, E. Biochem. Pharmacol. 1986,35,151-158. 14. Ganapathy, V.;Leibach, F. H. J. Biol. Chem. 1983,258,1418914192. 15. Inui, K.; Takano, M.; Okano, T.; Hori, R. J.Pharmacol. Exp. Ther. 1985,233,181-185. 16. Tamai, I.; Tsuji, A.; Kin, Y. J.Pharmacol. Exp. Ther. 1988,246, 338-344. _ _ _ -~~ 17. Okano, T.; Inui, K.; Maegawa, H.; Takano, M.; Hori, R. J. Biol. Chem. 1986,261,14130-14134. 18. Tsuji, A.; Terasaki, T.; Tamai, I.; Hirooka, H. J. Pharmacol. Exp. Ther. 1987,241,594-601. 19. Berk, P. D.; Potter, B. J.; Stremmel, W. Hepatology 1987, 7, 165-176. 20. Ketterer, B.; Neumcke, B.; h u g e r , P. J. Membr. Biol. 1971,5, 225-245. 21. Kagawa, Y.; Racker, E. J. Biol. Chem. 1971,246,5477-5487. 22. Montal, M.; Mueller, P. Proc. Natl. Acad. Sci. U.S.A. 1972,69, 3561-3566. 23. Schindler, H. FEBS Lett. 1980,122,77-79. 24. h d e r s e n , 0.S.;Fuchs, M. Biophys. J. 1975,15,795-830. 25. Smejtek, P.; Paulis-Illangasekare, M. Biophys. J. 1979, 26,

441466. 26. Pickar, A.D.; Benz, R. J.Membr. Biol. 1978,44,353-376. 27. Heinz, E.;Geck, P.; Pietrzyk, C. Ann. N.Y. Acad. Sci. 1975,264, 428-441. 28. Johnstone, R. M. Biochim. Biophys. Acta 1978,512,550-556. 29. Lichtenstein, D.; Kaback, H. R.;Blume, A. J. Proc. Natl. Acad. Sci. U.S.A. 1979,76,650-654. 30. Nicholls. D. G.Eur. J. Biochem. 1974,50.305-315. 31. Rottenberg, H. Bioenergetics 1975,7,61-74. 32. Miller, A. G.; Budd, K. J. Bacteriol. 1976,128,741-748. 33. Kamo, N.; Kobatake, Y. Methods Enzymol. 1986,125,4658. 34. Christensen, H. N. Biological Transport,2d ed.;N. A. Benjamin: Reading. MA. 1975:vv 107-165. 35. SugiyGa, Y:; Kimura, S.; Lin, J. H.; Izukura, M.; Awazu, S.; Hanano, M. J. Pharm. Sci. 1983,72,871376. 36. Demura, M.; Kamo, N.; Kobatake, Y. Biochim. Biophys. Acta 1987,894,303-308. 37. Singer, S.J.; Nicolson, G. L. Science 1972,175,720-731. 38. Chapman, D. In Cell Membranes: Biochemistry, Cell Biology, and Pathology; Weissmann, D.; Claiborne, R., Eds.; Hospital Practice: New York, 1975;pp 13-22. 39. Mueller, P.; Rudin, D. 0. In Current Topics in Bioenergetics; Sanadi, D. R., Ed.; Academic Press Inc.: New York, 1969;Vol. 3, pp 157-249. 40. Hanai,T.; Haydon, D. A.; Taylor, J. J. Theor. Biol. 1965, 9, 433-443. 41. Fettiplace, R.; Andrews, D. M.; Hydon, D. A. J. Membr. Biol. 1971,5,277-296. 42. Levine, Y. K.; Wilkins, M. H. F. Nature (London) New Biol. 1971, 230,69-72.

Journal of Pharmaceutical Sciences I 31 Vol. 82,No. 1, January 1993