, “CPU”... -\ .. ‘ “<' ”emavt'mm‘ THESIS Ll: .. .. ,, a - , .__'A . " .".'. o . r A .' I .3 4 r5 _ , '13-.- ‘ ’ . “-uu‘wlf bi: ‘ - " .“n - ‘ " —‘-- - [I o . f5) "u--__‘g‘—r, UIIV\:« ‘ “ This is to certify that the thesis entitled "An investigation on the interaction of heart drugs beta-adrenergic blockers and calcium blockers with membrane lipids using artificial model membranes" presented by Biao Shi has been accepted towards fulfillment of the requirements for Master of Science degreein Biophysics Major professor z’Z/tea._7 fl}: $3?) H. Ti Tien t / Dae 7/ 0-7639 ‘PVIESI‘J RETURNING MATERIALS: Place in book drop to uaamues remove this checkout from .4--2through3114.....o.......................... 58 Variation of ELM K+ conductance with Gramicidin A concentration in the presence and absence of propranolol-...-.o--oooo........ 62 Variation of ELM K? conductance with valinomycin concentration in the presence and absence of propranolol-.................... 64 The effect of electrolyte concentration on the drug's ability to block valinomycin-mediated K transport...................... 65 TEMPO ESE spectra-effect of various drugs on partitioning of Win “tural lec1th1n lipo‘mcs...OOOOOOOOOOOOOOOOOOOO0.... 75 Variation of TEMPO spectral parameter f in temperature scanning-effect of verapamil on the fluidity of DMPC 11”.“.s0000OOOOOOOOOOOOOOOOOOOO0.0..OOOOOOOOOOOOOOOOOO.COO... 78 TEMPO ESE spectra in temperature scanning-effect of verpamil on partitioning of TEMPO in natural lecithin liposomes......... 79 Variation of TEMPO spectral parameter f in temperature scanning-effect of verapamil on the fluidity of natural 1.:1th1n11p°.“es..00000..OOOOOOOOOOOOOO-OOOOOOO0.00.00.00.00.0 81 Variation of TEMPO spectral parameter f in temperature scanning-effect of verapamil on the fluidity of the liposomes prepared from rat heart lipid extracts......................... 83 Variation of TEMPO spectral parameter E as function of V‘rapn1lconcentrationOOOOOOOOI0.00.0.0...OOOOOOOOOOOOOO0.00.. 84 TEMPO spectral parameter at various pH of natural lecithin liposm salutionO0.0.0.000....0.COOOOOOOOOOOOOOOOOOOOOOO0.000. 85 Double logarithmic plot of ELM K+ conductance vs. protein concentration of sarcolemma solution........................... 93 Staircase-fashioned stepwise increases in ELM K+ conductance ind'JCEd by sarCOJ-em sciatica...0.0.00....OOOOO‘OOOOOOOOOOOOOO. 94 Ca+-dependent stepwise increases in ELM conductance induced by sarc°1msalution..0..00000.0.0000000000000000000000000...... 98 G-V curve-the voltage dependence of the steady state conductance of ELM induced by sarcolemmal K channel........... 99 Stepwise increases in ELM K+ conductance induced by cardiac sarcolemma solution at various applied voltages.....o..........100 (vii) GENERAL INTRODUCTION Cardiovascular diseases (heart diseases) are leading diseases, threatening people's health and life all over the world. Scientists have been making great efforts to develop new drugs to combat them. +2 blockers) and Calcium channel blockers (calcium blockers or Ca beta-adrenergic receptor blockers (beta-adrenergic blockers or beta- blockers) (Figures la and 1b) are two types of effective cardiovascular drugs (heart drugs) that have appeared in the last two decades. They are widely used today to treat a wide variety of heart diseases like angina, heart attack, arrhythmias and hypertension. Calcium blockers are thought to combat heart diseases primarily by preventing Ca+2 entry through the excited myocardiac membranes (Flenken- stein-Grfin, 1984). Because Ca+2 plays a critical role in muscle contraction, the decrease in the availability of Ca+2 slows the pumping rate and declines the contraction tension of the heart, and the blood demand and workload of the heart is reduced. The decrease in the intracellular Ca+2 level also leads the inhibition of the contraction of smooth muscles in the coronary artery walls, and thus allowing them to expand. This, in turn, increases the blood supply to the heart and protects arteries from suddenly clamping and closing off blood to the heart. It is known that beta-adrenergic blockers act in a manifold manner (Shanks, 1984). Firstly, they compete with epinephrine and norepine- l Figure 1a. A'enO‘O' ~- ( (.M {)w—(r‘ u c”; N. w...“ a . I I C 0. com_ rON k/ LabeIanI < )—cHICH,cwr ,CH I I C;N OH Meteors!“ moo-(cs. —Q'—°°‘-C'>O*W I 0‘ no WoI o" 0:.) | W», ”I“ mummy-«ouch,» OxpnnoIoI coats—cw, n N We! ”wWfi: 0-0 OH Malachi I wwwmfi m... ' T" mfi-M—O—WWJ, N’S‘N 01 O N 00-! 004' W - I ‘ I I . Chemical Structures of the beta-adrenergic receptmr blockers. Figureplb. Dmiazem T \ CH'CJ-i N ‘0 (Y'w’ meetdlpine Omit-C scrapie, /\ s H c H 0H,. I H - Perhexmne if“. I C Vorepemll °"~‘° on. 00". m,o—b—L,W3—d-ow, La... Chemical Structures of the calcium channel blockers. 4 phrine for beta-adrenergic receptors and prevent these neurotransmitters from stimulating the target cells (Lefkowitz, 1979). When deprived of such stimulation, the heart muscle pumps blood at a lower rate and requires less oxygen to function satisfactorily. This also helps to decrease the frequency and severity of heart diseases. In addition to the beta-receptor activity-related action, beta-blockers also function by blocking ion transport (Ca...2 and Na+) through the interaction with the cell membranes (Messineo and Katz, 1979). In summary, both types of drugs are delivered to the periphery of target cells through the blood circulation, where they inhibit membrane channels or receptors directly or through other membrane components, exerting their pharmacological effects. These drugs are receiving increasing pharmacological attention, but the precise mechanism of their action in molecular terms remains obscure. Because of the great signficance, they are being investigated extensively by pharmacologists, biochemists and physiologists in different approaches. Theoretically, these studies eventually will permit us to explore the molecular basis of the drug's action. Practi- cally, we will be able to develop a novel approach for rapidly designing and screening more potent and safe heart drugs after our understanding the relationship between the drug's functions and their structures and properties. As a fundamental type of biological macromolecules, lipids have been investigated much less than have proteins and nuclei acids. In the membrane biology, they have been assumed as inertly structural elements and as semipermeable barriers, and their implications for the membrane processes have been relatively ignored (Cullis et al., 1980). The 5 "mosaic fluid model” proposed by Singer and Nicolson (1972) endowed lipids with some influences on the membrane functions through lipid- protein interaction. For instance, the lipid salvation which results in the aggregation,. capping or conformational change of proteins is required for activities of many membrane proteins (Sandermman, 1983). However Singer's concept is not complete, as to where lipids are restricted in providing a matrix in which protein are embedded. If, as argued by ‘many researchers, lipids serve structurally as merely’ a supporting matrix, it would be most likely that a single species of lipid like phosphatidylcholine (PC), or a few species, could meet the requirement for all living membranes (deKruijff et al., 1981). However, the lipid compositions of biological membranes are .extremely hetero- geneous. Human red blood cell membrane alone contains at least 20 different molecular species of PC and a total of some 150-200 chemically different lipid molecules (Golde et al., 1967)! Little is known about the functional role of lipids in. many membrane processes such as transmembrane transport, cell fusion and the interaction with proteins. The recent discoveries of lipid polymorphism (Verkleij et al., 1979) and the liposome helix (Lin et al., 1982) imply that lipids may be much more important in membrane functions than that we expected before. To appreciate the role of lipids, a modification of the fluid mosaic mode1-"metamorphic mosaic" model-was proposed (Cullis et al., 1980). It is widely accepted that the plasma membrane, as being the place where a cell communicates with the external environment and other cells, is the major target site for a variety of drugs (Brasseur et al., 1984). For a large number of drugs, owing to the heterogeneity in chemical structures, the precise mode of interaction with the membrane components 6 (proteins or lipids) remains a matter of debate. It is known that membrane lipids, especially phospholipids can serve as the action sites for certain classes of drugs (Bibl, 1984). In this case, the chemical components and physicochanical properties of membrane lipids are the factors regulating and controlling the biological and pharmacological effects of drugs. One example is local anesthetics, a clear relation- ship between their therapeutic effect and their interaction with lipid components of the membrane has been established (Davio and Low, 1981), even though the detailed mechanism is still obscure. There are different models proposed to explain the action of a drug on the plasma membrane: 1) Drugs act directly through specific protein components (enzymes, receptors and channels) in the membranes without the partici- pation of lipid, and thus the drug's function is little mediated by a change in the property of the lipid matrix (Ohki, 1984). For specific interactions, the drugs usually have molecular structures complementary to the proteins electronically or stereochemically (Phadke et al., 1981). This is probably a common mechanism for many drugs. For instance, some hydrophilic drugs such as the antihistamine, cimetidine (Rooney et al., 1979), work according to this model. They bind to target membrane protein hydrophilically, being analogous to a typical substrate-enzyme interaction in aqueous media. 2) The final target of some drugs is the membrane proteins, but they affect the proteins only secondarily to their action upon the lipid matrix (Trudell, 1977). This lipid-protein cooperation model is thought to be quite common, and different ways are envisaged to link it to an effect of protein functions. Besides the lipid salvation mentioned 7 before, in which the activity of a protein is sensitive to the fluidity of either bulk lipids or annular lipids surrounding it, the binding of a charged drug may alter the charge state of the membrane. The electrical perturbation may change the gating states of channels, functional conformation of enzymes and affinity of receptors to ligands. IAn example is the histamine H1 antagonist methydilazine (Lee, 1982) in which positively charged molecules show extensive lipid binding, and repel the cationic histamine molecules away from the membrane surface, thus reduce agonist availability in the vicinity of histamine receptor. Another alternative is the membrane pathway model. Herbette (1984) reported that propranolol, a beta-adrenergic blocker, has to partition into the lipid bilayer before searching the binding site on the membrane protein receptor by the lateral diffusion. The action site of a drug can also be at the lipid-protein interface. In the "lipid shape, concept" (Hornby and Cullis, 1981), some membrane channels (e.g. Na+ channel) are presumed to be protein-lipid complexes and the existence of lipids of a certain shape is necessary for the channel to acquire a conductive configuration. When the lipids are displaced by drug molecules of different shapes, the protein pore will be closed because of the loss of the ”braces” supporting a proper architecture. 3) Membrane lipids play the predominant role in the drug's action. This not only refers to the lipid facilitation of the travel of many drugs acting intracellularly though the membrane. A classical example of the direct implication of lipids was proposed by Seeman (1972). In his membrane expansion theory for uncharged anesthetics, the drug molecules penetrate into the hydrophobic core of the lipid bilayer, expanding critical regions in the membrane and thus preventing the ionic 8 permeability. The antibiotic property of that polymyxin may also lie in the change in the membrane permeability due to the distortion of lipid bilayer (Hartmann et al., 1977). Electron microscopy demonstrated polymyxin induced separation of the bound lipids and free lipids and the appearance of domain structurals, leading to the formation of structural breaches on the separated lipid domains. The discovery of non-bilayer structures presents a new approach by which a drug may act through membrane lipids. These structures, like inverted micella and hexagonal (HII) phase, have been proved to participate actively in some important membrane process such as the membrane fusion (Hui et al., 1981) and transmembrane transport of divalent ions (Verkleij et al., 1982). Some drugs, e.g.. the antitumor agent, adriamycin (Goormaghtiph, 1982), are found to interfere with the transition between bilayer and non-bilayer lipid phases. Of course, the variety of drug-membrane interactions can not be covered by such a few simple models. For instance, some drugs may be nondiscriminative by membrane components. They are directed to an interface between an aqueous solution and macromolecular structures on the membrane, regardless of proteins or lipids (Shibuta et al., 1984). Considering the complexity of biological membranes and the diversity of the structures and properties, we should say that all these models (and others) are possible. Different drugs, needless to say, function in different mechanisms, and a single species of drug may function in different ways, depending on the target membrane, the local physiolo- gical conditions and the drug's concentration and state. The action of the heart drugs on the membranes is far from fully understood. For beta-adrenergic blockers there are two distinct 9 effects: specific and non-specific. In the specific effect, beta- blockers bind to beta-adrenergic receptors directly (Lefkowitz, 1979), leading to the attenuation of the response to the beta-agonist stimulation. In the non-specific effect, they cause a functional perturbation of the biological membranes. The target sites for the non-specific actions are believed to be lipid components (Hellenbrecht et al., 1973). The drug molecules partition into the bulk lipid matrix and decrease ion flux through interaction with membrane lipids, leading to the cardiac effects (Messineo and Katz, 1979). There has been a large body of evidence that the inhibition of Ca+2 uptake in cardiac cells by beta-blockers was independent of their beta-blocking activity, only representing the lipid effect (Noack et al., 1978). For example, d-isomers of these agents, which are devoid of beta-blocking activity, +2 uptake (Evangelista et al., 1981). are equally effective against Ca These membrane lipid effects were suggested to account for the pacemaker and negative inotropic properties of beta-blockers (Rauls and Baker, 1979), and contribute the antiarrhythmic and antiangina function of these agents (Shank, 1984). The lipid effect of beta-blockers also influences blood vessels. As we know, excessive reactivity of blood platelets may lead to atherosclerotic vascular diseases. PrOpranolol and some beta-blockers are found to be capable of normalizing hyper- active platelet aggregation in patients with coronary artery diseases (Frishman et al., 1974) and relieving angina pectoris (Weskler, 1977). This anti-aggregation effect correlates well with lipid solubility but not other properties of these agents (Iwamura et al., 1983), and Kerry et a1. (1984) prOposed that it is, most likely, due to the interaction of the drug with membrane phospholipids, which causes membrane disorgan- 10 ization and thus directly impairs the affinity of the receptor for 2 agonists and/or the function of the Ca+ ionophore. The mechanism of the action of calcium. blockers remains ‘more obscure. Their potency, specificity' and subtle structural-activity relations imply that the pharmacological function of Ca+2 blockers, at least some of them, is more relevant to the binding to recognition sites at or near Ca+2 channels. Such binding sites have been reported to be 2+ found. on tissue extracts containing Ca channel activities. For example, Norman et al. (1983) reported a binding site for dihydropyridin Ca+2 blockers located on a glycoprotein macromolecule of Mt. 200,000. On the other hand, some investigators believe that at least some Ca+2 blockers function by purely physico-chemical interaction with the cell membranes (Towart and Schrsmm, 1984). It is suggested that Ca+2 blockers may act at more than one site, rather than solely on the slow +2 2 channels (Church and Zsoter, 1980). Some Ca+ blockers, e.g. Ca verapamil, have intracellular sites. They are facilitated to travel through the plasma membrane by lipid components (Pang et al., 1984), and then modulate Ca+2 level inside the cells through intracellular components such the channels on the sarcoplasmic reticulum (SR) (Galvin et al., 1982) or a Ca+2 binding protein like calmodulin (Epstein et al., 1982). Whether or not the interaction of Ca+2 blockers with membrane lipids is related to their Ca+2 blocking activities has not been studied and is uncertain. There have been some reports that the depression of Ca+2 uptake may also result from the general perturbation of the membrane lipids (Fairhurst et al., 1980). Galenhofen and Hermstein + (1975) observed two mechanisms in the depression of Ca 2 transport by methylverapamil (D600): 1) the depression due to a specific protein at 11 low drug concentration, and 2) the depression due to the lipid effect at high drug's concentrations. Recently Erdreich and Rahamimoff (1984) found that verapamil affected Rafi-Ca+2 antiport system derived from +2 heart sarcolemmal vesicles. The inhibition of Ca uptake by verapamil could be reversed by adding an excess of PC, and so it is possible that the drug acts in the lipid phase and not just on the transporting molecules. Calcium blockers so far examined are all appreciably membrane- active agents (Sasaki, 1984). For instance, they are able to inhibit the platelet aggregation through a modification of physicochemical prOperties of membrane phospholipids (Kiyomoto, 1980). Is the potent +2 Ca +2 blockade by Ca blockers a result of potentiation of the nonspecific membrane effect, and is the membrane-active property +2 +2 prerequisite to the potent Ca blocking activity of the known Ca blockers? There have been no answers. It is easy to understand the exertion of all beta-adrenergic blockers on beta-receptors in terms of their similar chemical structures. But the molecular architectures of the so- called ”Ca” blockers", verapamil, diltiazem, dihydropyridines and others, have no outstanding common features (Henry, 1980) except that all possess a bulk hydrophobic mass with a secondary or tertiary nitrogen in the periphery. Usually, a wide spectrum of chemical structures means a lack of stereospecificity in the drug's action, and it may be indicative of the interaction of the drugs with membrane lipids rather than with specific proteins (Sasaki, 1980). This may be also explained by the existance of the different populations of Ca+2 channels or binding sites whose affinities are different for various +2 Ca blockers. This concept has been applied to interpret the 12 histological selectivity of Ca+2 blockers (Hof, 1984). As an example, Nachshen and Blaustein (1979) suggested that Ca+2 channels in neurons 2 blockers than their equivalents in myocardiac are less sensitive to Ca+ cells. Since it appears that the interaction with membrane lipids repre- sents an aspect of the action of the heart drugs, the detailed knowledge about it will be useful. Even if this interaction does not contribute much to their therapeutic functions but causes some side effects, the knowledge will help us to design new drugs devoid of the adverse effects which may restrict the clinical use of the drugs. As a possible application in this research, we can develop liposome carries, which are affinitive to both the drugs and the biological membranes, to deliver the drugs to the target tissues more efficiently (Finkelstein et al., 1978). Complex structural and environmental factors make biological membranes in ‘3132_ not readily amenable to the investigation of a membrane process in physical-chemical terms. Model membranes such as planar bilayer lipid membrane (BLM) and lipid microvesicles (liposomes) provide good research tools for this goal. Besides the simplicity which make analysis easier, BLM simulates some of the important characteris- tics of biological membranes (Tien, 1985): it exists as ”liquid-like” ultrathin structure associated with two coexisting liquid interfaces, and it is capable of separating dissimilar aqueous phases and functioning in a vectorial or directional manner. These advantages have shown the practical value of ELM in the area of drug research (Nelson et al., 1984). 13 The plan of the study of the heart drugs using artificial model membranes are in two steps: 1) Investigate the drug's action on the unmodified model membranes (membranes without the native proteins). 2) Isolate the protein channels and receptors from living tissues, incorporate them into the model membranes and investigate the drug's action on the reconstituted membranes. In the first step, we have extensively and systematically studied the interaction of the two types of heart drugs (beta-adrenergic blockers and calcium blockers) with membrane lipids using unmodified BLMs and liposomes, as well as model membranes modified with artificial ionOphore substances. The experimental results showed: 1) Many beta-blockers and Ca+2 blockers produced large transmem- brane potentials. The magnitudes of the induced potentials are correlated with the lipid-solubility and the membrane effects of the drugs. 2) Many of the drugs increased the membrane fluidity regardless of the physical state of the lipid bilayer. The fluidizing effects are also correlated to the lipophilicities and the membrane effects of the drugs. 3) These drugs inhibitied the carrier-mediated cation transport but not the channel-mediated cation transport through BLM. This inhibitory effect is mainly due to an electric perturbation of the membrane elicited by the drugs. These facts suggest that the heart drugs can induce several alterations of physicochemical prOperties of the membrane through their interaction with lipids. This lipid effect may be involved in or may 14 influence some of the biological and pharmacological functions of the drugs directly or through lipid-protein interactions. It should be pointed out here that we only focus on the drug's action at the lipid level in this thesis. It does not mean, in any sense, that the role of membrane proteins should be ignored. On the contrary, proteins, as major performers of the membrane function, are responsible for the pharmacological effects, especially for more specific effects of many drugs. Lipids, not being informational macrobiological molecules, appear not to confer much specificity on a biological system according to today's concept. In the second step, we 'will investigate the interaction between the drugs and the ‘related membrane proteins. So far we have reconstituted voltage-operated potassium channel and Ca+2-dependent channel from rat cardiac sarcolemma into am. We feel that we can not delineate the mechanisms of the drug's action unless we make it clear in the different aspects and at the different levels. PAHI mmuc-mwcznnmmmmmoum The interaction of a drug with a lipid bilayer can be expected to produce various effects; and changes in the fluidity and electrical properties are the most common effects with important consequences for a wide variety of membrane processes (Lee, 1978). Electrical perturbation of BLMs is conveniently monitored by the measurement of electrical parameters, such. as membrane potential, resistance, capacitance and dielectric breakdown using BLM (Tien, 1974). This strategy has been -ployed with various types of drugs. For instance, the influence of anesthetics on the membrane potential and resistance was studied on BLM by several groups (e.g., Tsofina et al., 1978). Schlieper and Medda (1980) used unplanar and unbilayer model membrane systems to investigate several beta-adrenergic blockers. There have been few experiments done with calcium blockers on the unmodified BLM. In the following experi- ment both beta-blockers and Ca+2 blockers were found to induce transmem- brane potential but showed little influence on membrane resistance and capacitance. 15 16 .IEIIIIALS AND METHODS Experimental Arrangement A schematic diagram showing the experimental arrangement for the formation and electrical measurement of BLM is drawn in Figure 2. It mainly consists of: 1) Outer container - made from a solid Lucite block. with 2 adjacent drilled holes (one playing the role of outer chamber) and with 2 glass windows, one for illumination and another for observation. 2) Teflon septum - a 10 ml Teflon beaker with a small aperture punched through it. The diameter of the aperture was measured with a calibrated reticle and its area is 19.0 mmz. 3) Stirrers -» a pair' of magnetic stirring bars, one in each chamber. 4) Calomal electrodes - a pair of KCl-saturated calomel electrodes are used to provide electrical contact with the membrane. 5) A pair of glass electrodes (Mankson), used for monitoring the solution pH and also for electrical measurement. 6) Digit pH meter (Larza). 7) High impedance electrometer (Keithly 610C). 8) Low impedance picoammeter (Keithley 41). 9) Low level capacitance meter (ICE/Electronic Model 1-6). 10) External variable voltage source. 11) Standard resistor box - it can be selected as 0 or 105 to 10 by a factor of 10. 12) Chart recorder (Omniscope). 13) X-Y recorder (MFE plotamatic 715M). 17 Figure 2. Schematic diagranlof eXperimental arrangement used in BLM studies. 18 £3 23.; 2.3. =Z~3l£ 2 T1483..- :29 1.33....- m L. W. m... 3L . . 1 \ Bye—ml» fl. ........._:.....a. a... L: gain—Q. .g Ens-«3: 32828 =- 3:3- 3... a. l l -‘ I I to huaflntu 28-.— , e I \L, ” “ ‘IIIL .. a". E _ 19 BL! Forming Solution and Formation of BL]! The membrane forming solution (PS) consisted of 3.12 egg phospha- tidylcholine (PC), 1.7% E.coli phosphatidylethanolamine (PE), 1.1% bovine phosphatidylserine (PS) and 1.01 oxidized cholesterol (DC) in W/V except for that described specially in some experiments. This formula gives a stable BLM with long lifetime (0.5 - 2 hrs) and is suitable for our measurements. Before application, all phospolipid-organic solvent solutions were evaporated, first in a stream of argon and then in vacuum. The procedure for oxidation of cholesterol followed Tien et a1. (1966). The dried phospholipid and oxidized cholesterol were dissolved in n-octane or n-decane. The BLM was formed by usual technique (Tien, 1974) on the small hole (of Teflon cup with the aid of a Hamilton repeating microsyringe. The formation of ”black” membrane was monitored using a stereoscopic microscope. Electrical Measurement After the membrane had turned "black” a small volume of drug stock was injected into the inside of the Teflon cup with a micropipette and the same volume of the bulk solution was withdrawn for excluding any change due to hydrostatic pressure difference. The drug stock was prepared by dissolving a drug in the same buffer (0.01 M NaCl, 5 mm tris-glycine, pH 6.5 except for that described specially) used as the BLM bathing solution (BS), and the pH of the stock was readjusted so that the pH of bathing solution was kept constant after the addition of the drug, avoiding any interference from H+ concentration gradient across the membrane. The introduction of the drug solution was followed by stirring with magnetic bars for a certain time to make the bulk solution homogeneous quickly. 20 The BLM separated two bathing solutions in which two electrodes were submerged. The side (Teflon beaker) to which drugs were added was designated as the inside (c_i_s_ side), and electrode in the £13 side was connected to the measurement instruments. The electrode in the outer chamber (outside or 2333 side) was connected to virtual ground, serving as a reference electrode. BLM resistance was measured by a voltage divider circuit (Tien, 1974). A known voltage (V1) was applied across the .BLM equivalent RC circuit, which is parallel to electrometer, and a standard resistor (R1) in series, and the BLM resistance Rm was calculated from the electrometer reading of Vm (i.e., membrane poten- tial) with the equation, m 1V -V (I) The membrane potential was measured by disconnecting the power source and 31 from the circuit. The time course was recorded with a chart recorder. In the pH effect experiment, pH in the bulk solution was monitored accurately by a pair of glass electrodes connected to a pH meter. The BLM formation system and electrodes were protected by a Faraday cage from the external noise. All the measurements were performed at room temperature (22 : 1°C) and were repeated 2-3 times. RESULTS Many of the drugs tested (Figure 1), both beta-adrenergic blockers and calcium blockers, were able to induce large transmembrane potential while being added to the bulk solution on the gig side. This potential was due to the effect of drugs with BLM itself rather than the hydro- static or mechanical perturbation which might arise at the moment of the 21 addition of chemicals and while stirring. This was demonstrated by a control in which the injection of an electrolyte solution, instead of drug stock, did not produce the effects. The potential has minus sign with respect to the is side to which the drugs were added. And its magnitude is dependent upon the drug's concentration, the composition and concentration of the forming solution, and the pH, ionic species and strength of the bathing solution. 1. Dose-lespouse Relation Three calcium blockers and seven. beta-adrenergic blockers were examined by measuring the potential across BLM bathed in 0.01 M NaCl buffer (5 mM tris-glycin pH 6.5) (Figures 3 and 4). The drugs began to show their effect at about 10'”6 M of the final concentration in the bathing solution and could induce quite high transmembrane potentials 3 (~50 to -100 mV) at 10- M. The exceptions were sotalol and atenolol which produced small potentials (about -10 my at 10-.3 M). In the cases of all drugs, no significant change in current was measured while the potential was developing. The dose-response experiment was performed by the successive addition of the drug solution without rupture of the membrane after the potential no longer changed. The curves in Figures 3 and 4 showed a quasi-linear relation between the induced potential and drug's concen- tration over the range from 10-6 to 10--3 M for the drugs. The capacities for inducing the potentials for various drugs are different. This can be estimated with the values of the induced potential at the same drug concentration and salt concentration (Table 1). It reflects the different abilities of drugs to interact with the membrane lipids. -)00 I -30 .. : 3 -so - E I E -40 - E o E x -20 — 1/ o I Q :1 A ‘3 ./E/ l l 10"5 . 10‘5 10" 10-3 Drug Concentration (M) Figure 3. Variation of the induced potential with the concentration of eta-blockers: (0) propranolol, (O) labetalol, (X) oxprenolol, (A) metoprolol, (®) pindolol, (I) atenolol and (0 ) sotalol. 23 -80 +— C> -eo . E E -40 - I § 5 -20 - I I h] 1 l 10" 10'5 10" 10"3 0mg Concenttation (M) Figure 4. Variation of the induced potentials with the concentrations of calcium blockers: (.)verapami1, (O ) nicardipin and ( x )diltiazem. 24 Table l: The potentials induced by various drugs at 1.0mM in 1.0mM NaCl in BLM of various composition. (A) potentials (B) potentials (mv) on PC+ (mv) on PC+ Drug PE+OC+PS BLM PE+OC BLM (mv) Ratio B/A verapamil -75.0 :_3.0 -45.5 :_2.5 0.61 nicardipin -61.5 diltiazem -49.0 :,5.0 -37.0 1 1.0 0.76 propranolol -96.5 :_4.5 -53.0 :_2.0 0.55 labetalol -73.5 :,3.0 -50.0 3.2.0 0.68 oxprenolol -70.0.1 1.5 -48.0 1.4.0 0.64 metoprolol ~64.0 -38.0 0.59 pindolol -55.0 atenolol - 9.5 1 2.5 solalol ~ 7.0 - 5.0 0.71 Some values are the average of the potentials from two repeats + stand- ard deviation of the mean. 25 While developing the transmembrane potential, the drugs showed little influence on BLM resistance (Figure 5). The slight apparent decreases in the membrane resistance in Figure 5 were due to not only the introduction of the drugs but also to a.spontaneous change during the lifetime of BLM which at time is difficult to avoid completely. The results for the heart drugs can be compared with the behavior of 2-2-dinitrophynel (DNP), a well known phosphorylation uncoupler. DNP also generated a transmembrane potential (Lea and Croghan, 1969). We measured about +45 my potential across BLM in presence of 10'"3 M DNP in gig side, but the development was accompanied by a tremendous decrease (3 orders) in BLM resistance. 2. Effects of Ionic Strength on Drug-Induced Potential Verapamil, a representative of calcium blockers, which is widely used in dilating blood vessels, was chosen for further study in the series of experiments described below. The ionic strength in bulk solution affected the induced transmembrane potential in a unique way. In the experiment shown in ' Figure 6, verapamil concentration was fixed at 10"3 M. Up to 0.1 M of electrolyte KCl, the potential had practically no change with the variation of the salt concentration. Over 0.1 M, it was depressed dramatically with the increase in KCl concentration until it was almost totally eliminated by highly concentrated electrolyte (about 1 M). This phenomenon was also observed with solutions of other monovalent metal cations (Na+ and Li+). And it was true for other drugs tested. Some results are shown in Figure 7. Exceptions were sotalol and atenolol, where the effect of ionic strength was not significant, because their 10:0 10’ .7 3 8: V 10' C i 2 i 3 HF :- m‘ Figure 5. 26 Preptaeolol o — ‘0 Vtmemil G P D 0" p— U \\\D L 1 1 1 "-6 10" no“ 10" Chemical Beecesmtien (n) Variation of BLM resistance with the concentrations of the drugs and DNP. ( O ) verapimil, ( O ) propranolol, ( O ) DNP. Membrane solution in DNP experiment was the same as that in the drug experiment. The bathing solution in DNP experiment: 0.1MNaCl. ‘40 lrooseenbmo Potential (ml) -20 Figure 6. 27 on n: + 00 + 9:, a. lloCl /]l/‘ I PC+PE+OC+ PS, io help/i 1.." if :I c PE+OC. io Neil ",x ’ T * I”’ 'c * I I, ’ 0" ,‘oo""* I x" PC+PE+0¢. ioCoClz 1 1 P’ 1.0 0.1 0.01 0.001 0 Concentration ot Electrolyte (l) The transmembrane potentials induced by veraparmil 0.001M on neutral BLM and PS-containing BLM under various ionic strength and species. PC + PE + 00 + PS BLM was formed in NaCl (H) and CaC12 (On-O) bathing solution. PC + PE + 00 BLM was formed in NaC1 (O--O) and CaClz (O---O) solution. -so~ -‘o h Iraaseeebmo Potential (at!) 28 d,- / Olovenolol. m n + on + w/Jki { Pintlolol, *w*** +411 1%-. 0xo'reaolol, Fe + P: #- oc ' —"“""””'*_""“"'--+--«If-----i I” , " r' Pindolol rc / x . +rs+oc _,* -o-uc a ' -“" n a ,— -2° - I I ’- ~ JV 9’ ', , (’ /. :0'”‘ ;“___—————-""""""r/ . . :: . LII 10" in" 10-3 0 IaCl Concentration (I) Figure 7. The transmembrane potentials induced by oxprenolol (O), pindolol (x), and sotalol (I) on neutral BLM and PS- containing BLM at various ionic strength: PC + PE + 00 + PS BLM ( ) and PC + PE + 0C BLM (---- ). Drug concentration was 0.001M. 29 induced potentials are too low to be much affected, and such small changes are difficult to measure accurately. 3. Effects of Ionic Species on Drug-Induced Potential The ionic species is another important factor modulating the drug's effect. Divalent metal cations, especially Ca+2, are distinct from monovalent cations in influencing the drug-induced potential (Figure 6). When electrolytes were administered at the same level, verapamil produced the much smaller potentials across the BLM bathed in CaCl2 than those bathed in NaCl or KCl solution. Furthermore, one can see the dissimilarity in the curves of potential vs. salt concentraton in Ca"-2 and in Na+ bulk solution. The induced potential decreased significantly +2 even though not as greatly in the presence of Ca but experimentally showed little change in the presence of Na+. These facts imply an antagonism of Ca"-2 against the drug molecules. This antagonizing effect was found with BLMs of different formula, but it seems more prominent on PS-containing BLM. Other divalent cations like Mg+2 also diminished the drug-induced potential, but not a potently as Ca+2. 4. Effects of Lipid Composition Upon Drug-Induced Potential These heart drugs generally induced much higher potential on the BLM containing phosphatidylserine (PS) than on the BLM without PS (Figures 6 and 7). To evaluate the influence of PS, the induced potential was measured on two types of BLM namely 1) PC + PE + OC + PS, 3 M NaC1. The two and 2) PC or + co at 10"3 M verapamil and 10‘ potentials and their ratio are listed in Table l which shows that the potentials on BLM without P8 are only 50-75! of BLM containing P8. P8 is a negatively charged phospholipid owing to '000" group of serine residue (pK -2.2 ) and thus it tends to increase the charge 3O density on the polar region of the membrane. To confirm the correlation of the charge with the induced potential, the acidic lipids cardiolipin and ganglioside were used to substitute PS in the forming solution. As expected, these 2 acidic lipids also enhanced the verapamil's effect, although not effectively as PS (Figure 8). The correlation was supported strongly by the following quantitative experiment in which we prepared a series of forming solution in such a way that PC and OC were fixed at the usual level (3.11 and 1.02), but PS concentration was varied from 0.01-1.01. The measurement revealed that the potential induced by verapamil increased as a linear function of PS content in the forming solutions (Figure 9). Contrary to this, neutral phospholipid PC of the high concentraton diminished the drug-elicited potential as shown in a similar experiment (Figure 10) where 0C was 0.11 and PC varied from 0.1 to 102 in the forming solutions. 5. Effects of pH on Drug-Induced Potential Figure 11 describes diagramatically the induced-potential as a function of the pH of the bathing solution. A distinguishing feature in the pH effect curve is that the maximum potential appeared at about pH 8 on PS-containing BUM. On the neutral BLM, the peak value shifted to the right (higher pH) slightly. The pH values of both are close to pK of verapamil (8.45) (Lflllmann and Wehling, 1979). The potentials decreased when pH is changed to a more alkaline or a more acidic milieu. Other heart drugs gave the similar potential vs. pH curves. For instance, metaprolol has pK of 9.68 (Newton and Kenza, 1978) and its peak potential was obtained at around pH 9.0 (Figure 11). These facts 31 R+PE+OC+PS I/ \ff—I +/ N+PETOC+GNIIHOSMC E-so - i 4’ ‘f’ +/ \IKI g-“ - t/ rc+rr+oc -20 _ .i. J .444 1 .___JLill4 l to" 10'2 10°3 ” o IaCl Concentration ( I I Figure 8. The transmembrane potential induced by propranolol 0.001M on BLM of various composition: PC + PE-+OC + PS BLM (0), PC + PE + 00 + ganglioside BLM (- ),Tc + PE + 0c BLM (o). The concentrations of PC, PE, and 00 were as usual. PS or ganglioside concentration was 1.12. 32 ~30 - -60 .— 2 .3 2;; 3 -4o - i a. S 5 .-.= .: -20 h- .1 ! J 0:01 0.1 1,0 PS Concentration (“4% ) Figure 9. Variation of the induced potential with P8 content in BLM forming solution. The forming solution was composed of 3.5% PC, 0.52 0C plus PS. The drug used was verapamil (0.001M). 33 -60— ; E I: '40 '— E 4’. C i 15 5 I m :5 -2o - i: 4 I I 0.1 1.0 10.0 PC Concentration ("4%) Figure 10. Variation of the induced potential with PC concentration in BLM forming solution. The forming solution was composed of 0.12 00 plus PC. The drug used was verapamil (0.001M). 34 -30 .. * Verapamil. PC+PE+OC+PS ~60w- 2 3.. IotOprolm, g PC+Pe+oc+ss 5 11' 3-40:- 3 m C . g Veraperml, e 3 ; P¢+ seq-ac -20.. J i 5 i 2.0 4.0 6.0 8.0 PI! is The Iatltint Solution: Figure 11. Variation of the induced potential as a function of pH of the bathing solution. Potential induced by 0.001M verapamil on PC + P: + 00 + PS BLM (e) and on PC + PE + 00 BLM (o) and potential induced by 0.001M metaprolol on PC + PE 4» 00 + Ps BLM (0). 35 indicate that the uncharged forms of these drugs seem to be the func- tional species responsible for the transmembrane potential. Generally, the pH giving peak potentials in our experiments does not quite agree with the pK of drugs in aqueous solution reported in literatures, often being 0.5-1.0 unit lower than the latter. This difference can be explained as follows: firstly, pH at the interface between the membrane and soluton is not in accord with that in the bulk solution. According to the Boltzmann distribution of a charged particle in a varying field, H+ concentrations at the membrane surface [H+]s) and in bulk solution ([H+]b) are related by the equation (Lee 1977), -F wo/R'r (n+1, - (n+1, e (2) Correspondingly, 1’0 0 RT (3) where we is surface charge potential on the membrane. It can be readily seen that for negatively charged BLM (wo < 0), the pH at the membrane surface is always lower than that in bulk solution. In our case, 1 137 / c), pH at assumingipo - -0.020 V (calculated fromwo - 51.4 Sinh- the membrane surface is 0.35 pH unit lower). Secondly, as pointed out by Schreirer et al. (1984), the pK value for a ligand in the membrane phase may differ from that in aqueous phase if the ligand or/and the membrane possess ionizable groups. DISCUSSION 1. Origin of the Transmembrane Potential The origin of a transmembrane potential on BLM was discussed in detail by Tflen (1974). An observed transmembrane potential in a BLM 36 system may be due to one or a combination of the followings: 1) Diffusion potential which is associated with the movement of charged particles through BLM under a concentration gradient. 2) Surface charge potential which arises as a result of the absorption and redistribution of charge on the membrane surface. 3) Dipole potential which originates from a change in dipoles of membrane lipids or absorbed ligands. In our experiment, if the transmembrane potential is due to the diffusion of charged particles-protonated drug molecules or protons in this case, it would be accompanied by a large decrease in the membrane resistance (Foster and McLaughlin, 1974), but such change was ‘not observed. More convincingly, a diffusion potential is expected to be depressed around pK of a drug where its molecules exist predominately as a neutral form. This is in direct contradiction. with the actual observation in pH effect where the potential had its maximum around the pK (Figure 11). It is true that some drugs are found to travel through BLM, but their diffusion species are the neutral molecules. This is different from DNP which was shown to travel through BLM in a negatively charged form ‘HA; (Finkelstein, 1970), and its induced potential. is mainly a diffusion component which is followed by a tremendous decline in BLM resistance (Figure 5). The potential induced by DNP can be predicted by the Nernst equation, which is used to describe the diffusion potential when a concentration gradient of DNP is across BLM, [Cl RT out so - —-ZF 1n __[C]in (4) But in the presence of the concentration gradient of the drugs, the potential developed did not follow equation (4). For example, for a 37 10-fold difference in verapamil concentration. on two sides, only' a potential of ~15 mV, instead of the theoretical value of the diffusion potential ~58 mV, was measured. All the facts seem to rule out the diffusion component as a major part of the drug-induced potential. In the case where there is no charge diffusion, a transmembrane potential must be related to a difference between two potentials at bifaces. The surface charge potential is usually suggested as a preferential candidate for the potential induced by amphiphalic drugs (Singer, 1977). However this component, which is based on the Couy~ Chapman double electrical layer model (Ohki, 1976), can not totally account for our experimental data. When Couy's theory is applied to a EU! system, we have (Haydon and Myers, 1973), d . N2 (zearc)1/2 Shh-1 ZNG (5) F ZRT For unit-unit electrolyte simply, ZKT -1 136.60 "’6 ' T 31“" (“T—4 (5) where WC: Couy's surface potential (mv) K: Baltzman constant T: Temperature (OK) e: Electronic charge a: Surface charge density (3-1) c: Electrolyte concentration in bulk solution (M) Taking 60 32 as average area for lipid molecule (Fettiplace et al., 1971),the surface charge density on BLM was estimated as 1/450 32 from PS percentage in total lipids (142). By using equation (6), we calculated the Gouy potentials at the different concentrations of 38 unit-unit electrolytes and plotted a theoretical curve (Figure 12). It is clear that the calculated surface charge potential does not correspond to the experimental data. At low salt concentration, the measured value is much smaller than expected by the theory. More outstandingly, in contradiction with what was predicted by Couy's model, the drug-induced potential remained constant practically, rather decreased while salt concentration ranged from 0-0.1 M. Therefore some factors other than the Gouy component should be taken into considers~ tion. A jump in membrane potential can also arise from an alteration of the oriented dipole in the polar group region of BLM (Hladky and Haydon, 1973). Generally an overall surface potential (Calvani potential) is the sum of Couy's potential WC and d1P013 potential (Tien, 1985). 0 I the + 4nu d (7) where n is the number of dipole per cmz, Vd the overall dipole moment of the constituent molecules in BLM. The dipole potential can be changed by means of the insertion of either the charged (Haydon and Myers, 1973) or neutral molecules (Anderson et al., 1976) into the membrane. The interaction of neutral molecules with zwitterionic lipids may alter the dipole potential by changing either the single dipole moment or the orientation of the dipoles. The following facts support the hypothesis that the drug elicits the transmembrane potential through affecting the dipole potential: 1) Unlike Gouy potential, the dipole potential is independent of electrolyte concentration (Hladky and Haydon, 1973) and this is seen in our experiment (Figures 6 and 7). ISO - 100 - :: 3 i E i ‘3 so - I 39 Couy's Potential Propranolol fi—o—qf—o / .—_/F— . Voraoaunil Figure 12. l l l J K 1 1.0 0.1 10-2 10-3 Concentration ot l-l Electrolyte (a) Comparison of Couy's surface charge potential and the drug- induced transmembrane potential: the potential induced by 0.001M verapamil (0) and 0.001M propanolol (O) on PC + PE 4» 0C + PS BLM. Couy's potential was calculated using ezqua~ tion (5), assuming the surface charge density of 1/4SOA ‘ 4O 2) Unlike Couy's potential, the dipole potential is located within polar-layer region, extending very little, if at all, into the aqueous phase adjacent to the membrane (McLaughlin, 1977), so that a ligand needs to penetrate into the bilayer to alter the dipole. Thus, the higher dipole potential depends on the availability of more lipid- soluble ligands. This is coincident with the observation in the experiment on pH effect. From the arguments above, the origin of the transmembrane potential can be partially explained as follows: the drug molecules alter, the orientated dipole of polar phospholipids on the side on which they are present, and so make the dipoles on the two sides asymmetrical, and the elimination of the symmetry of dipoles results in the observed potential across BLM. 2. Relation Between the Induced Potential and the Drug's Membrane Effect As summarized in Table 2, the order of the magnitudes for drug~ induced potentials is parallel to the order of the drug's lipOphilici~ ties. Pcorr in the table stands for the correlated partition coefficient of a drug in octanol-HZO system, which usually served as a measure of the hydrophobility of a chemical. Pcorr is calculated from the apparent partition coefficient (Papp) with equation (Wang and Tien, 1980), 2cm - Papp/(lra) (8) Here is the ionization degree of a drug, and a. l/1+ antilog (pK-pH) (9) The further analysis was done by plotting log P against the potential (Figure 13). Linear regression of the data by least square fit showed a 41 Table 2: Comparison of the drug-induced potentials and lipophilicities. Induced Poten- tial on PC+PE+ Drug PKa log Pap? log Pear: oc.ps BLM propranolola 9.45 1.24 3.29 ~96.5 labetalol“ 9.45 1.13 3.18 -73.5 oxprenololb ' 0.118 2.17 -7o.o metoprolola 9.68 ~0.25 2.04 ~64.0 pindololb ~0.328 1.73 ~55.0 sotalold -1.45 0.60 -7.0 atenolola 9.45 ~1.62 0.43 ~9.5 verapamilc 8.45 1.70 2.74 -75.0 nicardipin -6l.5 diltiazem ~49.0 Pap is the apparent partition coefficient and Pcorr the corrected parfation coefficient of a drug in octanol/water system (see text). The values of PXa, log P are taken from (a) Wang and Lien (1980), (b) Harada et a1. (1981). tg¥’Lu11nmnn and Wehling (1979), (d) Hellenbrecht et al. (1972). '100 ‘80 anl l an an I a a transmembrane Potential 0. Figure 13. 42 Ate Sot 1 I o 13 10 $0 '09 Penn Repression.curve-corre1ation between the induced potentials and partition coefficient of the drugs. Pcorr is the corrected partition of a drug in octanel/water system. The data is treated by least squares fit and the regression curve is plotted. The curve has regression coefficient of ~0.96l (p<0.05). 43 good correlation between the two parameters (r - ~0.96l, p < 0.05). This correlation indicates that the drug's action occurs on the menbrane rather than the aqueous ”unstirred layer” adjacent to the membrane surface. In other words, the observed potential results from a direct interaction between drug molecules and membrane lipids. This correla- tion also suggests that this interaction is mainly governed by hydrOphobic effect, even though a contribution from electrostatic force can not be totally excluded. In the present experiment, many of beta-adrenergic blockers but not all were able to produce rather high transmembrane potential. It is not difficult to link the relative potential-inducing capacities of these compounds to their chemical structures. Most of the beta-blockers tested are the analogs of propranolol, being amphiphilic molecules containing a hydrOphobic ring system and a hydrophilic side chain (Figure 1a). Investigation using NMR (Kulkarni et al., 1979), ESR (Phadke et al., 1981) and neutron diffraction (Herbette, 1983) have revealed that propranolol has an extended configuration while inserting into a lipid bilayer. Its naphatalene moiety partitions into the hydrocarbon core, and the amine side chain is positioned in the head group region. This intercalation allows it to disturb the oriented dipoles of phospholipids and reduce the surface charge, resulting in a potential difference across 311:. This interpretation is valid for lobetalol, oxprenolol and metoprolol. On the other hand, sotalol and atenolol are also ”propranolol-like" drugs but they bear additional hydrophilic groups attached to their aromatic rings. It seems plausible to ascribe the poor potential-inducing ability of these two agents to this merely structural dissimilarity: the hydrophilic group prevents 44 the drug molecules from penetrating into lipid phase. This speculation is strongly reinforced by the application of timolol. It resembles the hydrophilic property of sotalol and atenolol by containing polar ring moieties and so failed to induce a large transmembrane potential. Beta-adrenergic blockers can be subgrouped into two classes according to their action sites. Some of them (e.g., propranolol) possess both beta-receptor blocking activity and general membrane effects, and some (e.g., timolol) show strong beta-receptor blocking activity but minimal general membrane effect. Usually the latter relieve the symptoms of heart diseases more specifically but less potently at the same dose. It has been proved that the order of the inhibitory effects of beta-blockers on Ca+2 uptake (Temple at al., 1974) and the negative inotropic effects of these agents (Harada et al., 1981) followed the order of their abilities to perturb the membrane lipids. For instance, in Messineo and Katz's (1979) experiment the velocity of Ca+2 influx to RS vesicles was eliminated 502 by 0.65 mM prOpranolol but required 11 mM timolol for the same effect. Herbette et al. (1981) reported that on SR membrane the magnitude of the inhibition of Ca"-2 uptake by beta-blockers depended on both lipid solubility of the drugs and the physical state of fatty acid chains of membrane lipids. On the basis of neutron diffraction analysis, they suggested that the drugs interfere with the function of Ca+2 channel protein indirectly via the bulk lipid matrix, and other sites of drug interaction, namely with the lipid annulus and directly with channel protein, do not occur. The induced potential in our experiment is clearly correlated with the ability of the drugs to block Ca+2 uptake and other membrane effects. Propranolol and exprenolol, as the most powerful inhibitor of 45 Ca+2 uptake (Temple et al., 1974) produced the higher potentials. The poor inhibitors sotalol atenolol and timolol induced the small change in the transmembrane potentials. The middle size potentials correspond the status of metoprolol and pindolol as the moderate inhibitors. Thus the measurement of the induced transmembrane potential on BLM provides a method of estimating the potency and selectivity of beta-blockers. There is little known about the effect of calcium blockers on electrical properties of BLM. In this experiment, the members of this group that were used all tended to generate a potential across BLM. Some of Ca+2 blockers like verapamil may interact wih membrane lipids in a mechanism similar with that of beta-blockers, but it is dangerous to extrapolate this to all compounds of Ca+2 blockers because of their diverse chemical structures. Since many heart drugs show capacity of developing transmembrane potentials, a question. arises logically as to whether or not this electrostatic effect is implicated in the drug's function. Such a mechanism. has been employed in the blockade of ionic channels by beta-adrenergic blockers (Schlieper and Medda, 1980). It is generally agreed that calcium blockers inhibit Ca“.2 influx through acting on voltage-Operated channels (VOC) (Lanvin et al., 1983), but relatively little is known about these channels in terms of molecular insight. They are supposed to be protein pores with negatively charged groups which are particularly intriguing for divalent cations and serve as selectors to distinguish between different cations. They are also endowed with "voltage sensors" which confer voltage dependence on channel opening and closing (Triggle, 1982). The precise mechanism for which heart drugs regulate the gating of VOC is not clear. It is known 46 2 that Ca+ blockers do not act on VOC by directly plugging the channel pore in a manner analogous to tetrodotoxin (TTX) blockade of Na+ channels (Janis et al., 1984). Therefore, they may bind allosterically to a channel and trigger a conformational change, making the channel insensitive to voltage. As another possibility, they may create an electrostatic field in the environment of VOC to influence the gating of the channel. Biological membranes are ultrathin structures with less than 100 3 thickness (Robertson, 1981), so that a drug can build up an extremely high electrical field (105 V/cm) by means of developing a membrane potential, which is enough for changing the gating of VOC through moving "sensors” in the direction of the field. The calculation made by Lundstr8m (1977) indicated that the electrical potential due to the surface charge and dipoles is of the same order of the actual voltage difference across the excitable membrane. Recently Erdreich and Rahamimoff (1984) reported that part of the inhibitory effect of verapamil on Ca+2 uptake by cardiac sarcolemma is due to its action on the membrane potential, even though it is not known how this effect is exerted. The speculation will be challenged by an argument about the specificity of the drug's action. There are several groups of membrane active drugs which can induce electrical change on the membranes. Why 2 blockers inhibit Ca.)2 uptake potently but others do not? We do do 0.." not know whether the selectivity of drugs is only determined by their action on the specific proteins or if they can be also discriminated by lipid components in an unknown manner, even the latter seems not very meaningful in the today's concept. As far as the induced potential is concerned, there is no fundamental difference between beta-blockers and 47 calcium blockers (Figures 3 and 4, and Table 1) because the BLMs used are pure lipid bilayers, lacking of specific protein channels and receptors for the two groups of drugs. The results from such model membranes are not sufficient to explore the action of the drugs on the related proteins directly or via lipid-protein interaction. 3. Facilitation of P8 in Drug's Action In our experiment, all the heart drugs induced 0.5 - 1 fold higher potential on PS-containing BLM than on neutral BLM (Table 1). One can simply ascribe this enhancement to the presence of the net charge on the former. pK values of the drugs tested was between pH 8 and pH 10, and thus a considerable proportion of the drug molecules in the experimental pH (6.5 ~7.0) are protonated. These protonated drug molecules neutralize the negative charge on gig side of the PS-containing BLM, thus creating a greater potential difference than those on neutral BLM. The net negative charge on PS-containing BLM also facilitates the positively charged drug in penetrating into the bilayer to alter the dipole potential. However this interpretation is only partially true. As discussed before, the interaction between the drugs and membrane lipids is mainly hydrophobic and the drug species for developing potential is mainly in the neutral form. In Figure 11, we can see that the ratio of the potentials on PS-containing and neutral BLM appeared to be approximately the same at high pH and low pH, indicating that the enhancement of potential due to P8 has little dependence on the charge state of the drug's molecules. In other words, the electrostatic interaction bewtween the drugs and lipids is not the ‘major factor regulating the potential. 48 Liillmann and Wehling (1979) measured the binding of varous amphiphathic drugs (including verapamil, prOpranolol, metoprolol and atenolol used in this experiment) to liposomes composed of different polar lipids and they found an affinity order of PC < PE < PS. They suggested that the preferential binding of drugs to PC-containing membrane is mainly due to the formation of a non-tight packing membrane structure. Owing to a large hydration radius of serine residue, PS bears a larger polar head group compared with PC. Thus lipid molecules are kept distant from each other by the steric hindrance and electro- static repulsion between adjacent serine molecules (Rojus and Tobius, 1965), and there is a large area not occupied by lipids but available for the intercalation of drug molecules. Surewicz and Leyko (1981) reported that the partition coefficient of prOpranolol in PS bilayer is more than 20 times higher than in PC bilayer. The steric effect due to the intrinsic lipid-lipid interacton can explain why not only charged but also uncharged drug species are easier to insert into the PS-containing BLM. This steric effect model is also suppoted by the experiment in Figure 9 where the drug-induced potential decreased at high PC-concentration which is favorable to the formation of a tight-packing structure (Schlieper and Medda, 1980). 4. Antagonism of Ca“'2 Against Drug's Action The antagonism of divalent metal cations, especially Ca+2, against beta-blockers (Schlieper and Steiner, 1982) and Ca+2 blockers (Fairhurst et al., 1980) has been known for some time. The drug-induced inhibition of Ca+2~dependent responses in cardiac and smooth muscles can be over- come by the elevation of Ca+2 itself. Since Ca+2 does not block the binding of the drugs to Ca+2 channels (Krafte et al., 1985), the 49 antagonism may be related to other membrane components. Langer (1980) reported that the force of heart contraction depends on the physical state of glycolipids and glycoprotein attached to lipid bilayer, which act as binding sites for Ca+2 with their sialic acid residues. Negatively charged lipids like PS, PI and cardiolipin are known to be also binding sites for divalent cations (Barrett, 1981). There is a suggestion that Ca+2 must first bind to these sites before it crosses the membrane via channels. Indeed Ca+2 is found to compete for the same sites on phospholipids with cationic drug molecules (Browning and Akutsu, 1982). 2 The role of membrane lipids in the interaction between Ca+ and heart drugs is confirmed by our experiment in terms of the depression of the induced potential in the presence of Ca+2. In the neutral BLM, the antagonism is the result of electrical screening, and in PS~containing membrane, a relatively prominent effect may be caused by a steric hindrance rather than the reduction of negative surface charge, as pointed out by Schleiper and Steiner (1983). Ca+2 is known to be able to bridge between a pair of PS molecules (Gregory and Cinsherg, 1984), and in this way it fills the intermolecular space between PS and blocks the entrance to the interior of the lipid bilayer. X-ray diffraction showed that the spacing of PS-Na+ complex is 788 but the spacing of PS~Ca+2 complex is only 53 X in PS-membrane. In order to test the steric effect, we designed the experiment shown in Figure 14. The same quantity of drug was added into Ca+2 and K+ solution of the same normal concentration separated by BLM, and in this case the apparent potential should be intermediate between the potentials develOped at two halves of bilayers. The induced transmembrane potential was found to be larger on A -4o« 3 i E ‘5 0 ‘5 G. g -20 ;. E ID E V! F. .2 Figure 14. 50 o PC+PE+OC+PS A V 0‘0 o’o/PW- PE +oc o/ / L J _I L l 0.1 10.2 ' ”‘3 Concentration of Electrolyte ( N ) Verapamil-induced potentials across BLM separating different electrolyte solutions: KCl solution in c_is side and CaCl in £13.29. side. The electrolyte solutions were added into the distilled water bulk solution after the formation of BLM and their final concentrations were equal in normality. The potential was induced by 0.001M verapmil on BLM composed of PC + PE + 00 + PS (0) or only PC + PE +0C (O). 51 PS-PC-OC BLM than. on the PC-OC BLM, implying the dissimilarity' of 2 antagonizing mechanism of Ca+ and Rf on the former and the similarity on the latter. This is difficult to be interpreted solely by screening 2 and Na+ should have approximately the same effect because by which c9 there is no significant difference in charge density on the membrane surface in both cases. Pm II EFFECTS OF MEGS OI mesons-mm ”SPORT As mentioned before, the reconstitution experiment is the best way to explore the mechanism of the membrane channels using BLM. Ca+2 channels have been incorporated into BLM recently in several labora- tories (e.g., Ehrlich et al., 1984; Coronado and Affolter, 1985). Prior to the reconstitution experiment, we used ”artificial” ionophore substances to probe the effect of heart drugs on membrane transport. These substances, most of which are antibiotical products of micro- organisms, have been widely employed to study membrane transport processes (Tien, 1985) because their mechanism models have been well established. Of course, caution should be observed while extrapolating the results of such experiments to the events on the biological membranes. However if we consider the drug's action through membrane lipids, there might be some resemblances between the drug's effect upon the artificial ionophores and native transport proteins, so that the ionophore-probing investigation may offer some clues for understanding the real events that happen in vivo. As examples of such applications, nonactin, a monovalent metal cation carrier, was used to study local anesthetics (McLaughlin, 1975), and the Ca+2 carrier, A23187, was used 52 S3 to study calcium blockers (Malaisse, 1979) and beta-adrenergic blockers (Weksler et al., 1977). Valinomycin and gramicidin A are well-known antibiotic ionophores. They facilitate the translocation of monovalent cations across the membrane in different mechanisms and they were chosen for this experi- ment. As a mobile carrier, valinomycin forms a complex with K+ and diffuses back and forth through the lipid bilayers. Gramicidin A, an outstanding representative of channel formers, increases the membrane permeability to ions by forming a structural pore spanning lipid bilayer. MATERIALS All) mm The measuring apparatus is basically the same as that described in Part I. The steady-state current am) through the BLM was measured using an electrometer or low impedance picomneter while a voltage (Vm) was applied. The membrane conductance (cm) was calculated from the current-voltage (I-V) curve recorded by an x-y recorder: (10) <'H a m To ensure that Gm had reached a stable value, all I-V curves were recorded at 10 minutes after the addition of chemical. Valinomycin and gramicidin were dissolved in 992 ethanol. The volmne of the antibiotics stock added was less than 0.22 of the total volume. of the bulk solution to avoid the perturbation of BLM due to ETOH, which was proved with the appropriate. 54 IBSULTS l. The Effect of Propranolol on Valinomycinelediated K+ transport The addition of 10‘8 M valinomycin into 0.1 M KCl buffer solution (5 mM HEPES pH 6.5) on the gig side increased the conductance of BLM (52 natural lecithin plus 12 DC in n-decane) by 2-3 orders (from 10..9 (idem.1 to 10.7 to 10-6n-lcm-1). After K+ conductance (GK+) no longer changed, a certain amount of propranolol was introduced and the membrane current was decreased. I‘V curves at different drug concentrations are plotted in Figure 15, whose slopes give the membrane conductance GK+° In the controls performed with propranolol of the same concentration but without valinomycin, GK+ did not change or increased slightly, con- firming that the decrease in GK+ resulted from the inhibition of valinomycin-mediated translocation. . Over the range of 10-4 to 10..2 M, GK+ in log scale appeared as a linear function of propranolol concentration. At the higher concentra- tions, the saturation of the 'drug's effect was observed. Two things should be noticed here: 1) The maximum reduction of Ci" caused by propranolol was generally about 10-fold. The membrane conductance could not be restored to its original level before the addition of valino- mycin. Because propranolol was in a great excess of valinomycin with molar ratio of 104 - 106:1, it is not likely that propranolol inhibits K+ transport by interacting with valinomycin directly, which is expected to abolish K+ current totally. 2) The absolute concentration of propranolol for blocking K+ transport is rather high (about 10-4 M), 2-3 orders higher than those for affecting the specific proteins but at the same order as those for perturbing membrane lipids. These facts suggest Figure 15. 55 Current-voltage curves (I-V curves) or valinomycin-mediated K transport at various propranolol concentrations. Membrane solution: 5.01 natural. lecithin (plant) plus 1.02 0C. Bathing solution: 0.01M KCl, SmM HEPES (pH 6.5). Valinomy- cin concentration: 1.0 x 10' M. Propranolol concentrations are shown in the figure. 56 A as: e» $225 3 2 . q 3 72 x3 7: x S A ,5 e, 32.; 3 - 1 9.2 auuqmau ( j. 01) “1| wanna 57 that propranolol, most likely, interferes with the carrier's function through effects on the lipid bilayer. The experiment was repeated with 2,4-dinitrophynol (DNP) to examine the effect of propranolol on a negatively charged permeant. At a concentration of 10.5 M DNP reduced the membrane conductance by about 1 order. Unlike the valinomycin-K+ complexes, the diffusion. of DNP through the BLM was accelerated by propranolol (Figure 16). The drug influences the transport of two species in an opposite and symmetrical manner. 2. Effects of Other Drugs om Valinomycineledisted Kf Transport Several Ca+2 blockers and beta-blockers were tested using valino- mycin as probe (Figure 16). Verapamil, diltiazem, exprenolol and metoprolol were found to be capable of depressing K+ conductance, in more or less degree, whereas sotalol, atenolol and timolol lacked this ability. Again, here we see a parallel relation between the depression of GK+ and the drug's lipid effect. To clarify this, we measured valinomycin-dependent GK+ in the absence and presence of the drugs and design them as Go and 01 respectively. The ratio Go/Gl serves as a parameter to evaluate the ability of a drug to inhibit 23... 2.2+ =2... 1‘02 affected on membrane conductance by altering the jump frequency and probably also the duration of the open state. (This was not measured.) 4) It was reported that the osmotic gradient and Ca+2 are two essential factors for the incorporation of many sarcolemmal or SR ' vesicle into BLM (Miller, 1978). But in our experiment, the fusion of the channel preparation did not require such osmotic gradient across BLM. Ca+2 in the bathing solution showed neither positive nor negative effect on K+ conductance, which was proved by adding an excess of the Ca".2 chelater HGTA. It may be that the K+ channel in our experiment is different from Ca+2-dependencent K+ channel reported elsewhere (Miller, 1978; Latorre et al., 1982; Coronado and Latorre, 1982). Haas (1982) reported that certain sarcolemmal K+ channels are inhibited by verapamil, but the blockade of K+ conductance by the heart drugs has not been observed on the BIM modified with our cardiac sarcolema preparations. So far we have been not succeeded in reconsti- tuting Ca“,-2 channels. This may be due to the biochemical isolation procedure, but the electric measurement also renains a problen. +2 channel currents are more difficult to be detected Generally Ca because of the lower rate. In the most cases it has been studied by measuring single channel current using the voltage clamp technique (which is not available for us at present). More efforts will be made to toward study in the future. onuosum BIBLIOGRAPHY Akiyama, S. and H.I. Igish, 1979. Japan. J. Pharmacol. 29, 144-146. Altas, D. and M. Alder, 1981. Proc. Natl. Acad. Sci. USA 78, 1237-1241. Anderson, 0.3., G. Finkelstein, J. Katz and A. Cass, 1976. J. Gen. Physiol. 67, 749-771. Azzi, A. and C. Montecucco, 1977. Horizons Biochem. Biophys. 4, 267-294. Barritt, G.J., 1981. TIES. 6, 322-325. Brasseur, H., J.M. Ruysschaert and P. Chatelain, 1984. Biochem. Pharmacol. 33, 1015-1021. Browning, J.I. and H. Akutsu, 1982. Biochim. Biophys. Acta 684, 172-178. Case, G.D., J.M. Vanderkooi and A. Scarpa, 1974. Arch. Biochem. Canvin, C., R. Londzenhiser and C. VanBreemn, 1983. Annu. Rev. Pharmacol. Toxicol. 23, 373-396. Chapmann, D. and J. Urbina, 1974. J. Biol. Chem. 249, 2512-2521. Church, J. and T.T. Zsoter, 1980. Can. J. Physiol. Pharmacol. 58, 254-264. Cohen, F.S., J. Zimmerberg and A. Finkelstein, 1980. J. Gen. Physiol. 75, 251-270. Coronado, R. and H. Affolter, 1985. BiOphys. J. 47, 434a. Cullis, P.R., B. de Kruiff, M.J. Hope, R. Mayer and S.L. Schimid, 1980. Can. J. Biochem. 58, 1091-1100. Davio, S.R. and P.S. Low, 1981. Biochim. BiOphys. Acta 644, 157-164. Ehrlich, B.H., A. Finkelstein, M. Forte and C. Hung, 1984. Science 225, 427-428. Eibl, H., 1984. Angewandte 23, 257-328. Engelman, D.M., 1970. J. Mol. Biol. 47, 115. Epstein P.M., K. Fiss, R. Hachisu and D.M. Adrenyak, 1982. Biochem. Biophys. Res. Comun. 105, 1142. 103 104 Erdreich, A. and H. Rahamimoff, 1984. Biochem. Pharmacol. 33, 2315-2323. Fairhurst, A.S., M.L. Whittaker and F.J. Ehlert, 1980. Biochem. Pharmacol. 29, 155-162. Fettiplace, H., D.M. Andrews and D.A. Haydon, 1971. J. Membrane. Biol. 5, 277-296. Finkelstein, A., 1970. Biochim. Biophys. Acta 205, 1-6. Finkelstien, A. and A. Cass, 1968. J. Gen. Physiol. 52, 1455. Finkelstein, M. and G. Weissmann, 1978. J. Lipid. Res. 19, 289-303. Fleckenstein-Grfin, C., M. Frey and A. Fleckenstein, 1984. Trends Pharmacol. Sci. 5, 283-286. Folch, J., M. Lees and G.H. Sloane Stanly, 1957. J. Biol. Chem. 226, 497-509. Foster, M. and S. McLaughlin, 1974. J. Membr. Biol. 17, 155-180. Frishman, W.H., B. Weskler, J.P. Christodoulin, C. Smithen. and T. Killip, 1974. Circulation 50, 887-898. Galvin, R.A., N. Pearson, F.C. Messineo and. A.M§ Katz, 1982. J. Cardiovasc. Pharmac. 4, 935. Godin, D.V., N.T. Wan and J.M. Tuchek, 1976. Biochim. BiOphys. Acta 436, 757-773. Godin, D.V., T. An and M.E. Garnett, 1979. J. Mbl. Cell. Cardiol. 11, 261-274. Galenhofen, K. and N. Hermstein, 1975. Blood Vessels 12, 21. Goldman, R., T. Facchinetti, D. Bach, A. Raz and M. Shinitzky, 1978. Biochem. Biophys. Acta 512, 254-269. Goldman, 8.8. and R.W. Albers, 1978. J. Biol. Chem. 248, 867-874. Gomperts, B.D., 1976. In The Plasma Membran: Model for Structure and Function, Academic Press. Goormaphtigh, H., M. vanderbranden, J.M. Ruysschaert and B. de Kruijff, 1982. Biochim. BiOphys. Acta 685, 137-143. Gregory, D.P. and L. Ginsberg, 1984. Biochim. Biophys. Acta 769, 238-244. Harada, S., T. Ban, T. Fujita and A. Koshiro, 1981. Arch. Int. Pharmacodyn. 252, 262-271. 105 Hartmann, W., H-J. Galla and E. Sackmann, 1977. FEBS Lett. 78, 169. Haydon, D.A. and v.3. Myers, 1973. Biochim. Biophys. Acta 307, 429-433. Hellenbrecht, D., B. Lemmer, G. Wiethold and H. Grobocker, 1973. Naunyn-Schmiedeberg's Arch. Pharmacol. 277, 221-226. Herbette, L., A.M. Katz and J.M. Sturtevant, 1983. Mol. Pharmacol. 24, 259-264. Herbette, L., J.G. Sarmiento and D.G. Rhodes, 1984. Biophys. J. 45, 312a. Hladky, S.B. and D.A. Haydon, 1973. Biohchim. Biophys. Acta 318, 454-468. ° Hof, R.P., 1984. Trends Pharmacol. Sci. 5, 100-102. Hornby, A.P. and P.R. Cullis, 1981. Biochim. BiOphys. Acta 647, Hui, S.W., R.P. Stewart, L.T. Boni and P.L. Yeagle, 1981. Science 212, 912-923. Iwamura, M., T. Ishimori, M. Makino, K. Yasuda, A. Izumi and N. Himori, 1983. Japan J. Pharmacol. 33, 219-226. Janis, R.A., D. Rampe, J.G. Sarmiento and D.J. Triggle, 1984. Biochem. Biophys. Res. Comm. 121, 317-323. Jones, G.R., S.W. Maddock and H.R. Beach, 1980. J. Biol. Chem. 255, 9971-9980. Kass, R.S., 1982. J. Pharmacol. Exp. Ther. 223, 446. Kerry, R.M., M.C. Schutton and R.B. Wallis, 1984. Biochem. Pharmacol. 33, 2615-2622. Kiyomoto, A., 1980. Inhibition of platate aggregation. Diltiazem Kyoto Conference, Vol. II, pp. 80-91, Tanabe Seiyaku Co., Ltd., Osaka. Krafte, D., B. Coplin, P. Bennett and R.S. Kass, 1985. Biophysics J. 47, 512a. Krasne, S., G. Eisenman and G. Szabo, 1971. Science 174, 412-415. de Kruijff, B., P.R. Cullis and A.J. Verkleij, 1980. TIBS 5, 79-81. Kursch, B., H. Lilllmann and K. Mohr, 1983. Biochem. Pharmacol. 17, 2589-2594. Laeuger, Richter, and Lesslauger, 1967. Ber. Bunsengers. Phys. Chem. 71, 906. 106 Lagercrantz, C., T. Larsson, L. Sfirskog and P. Lincoln, 1984. Biochem. Pharmacol. 33, 1851-1857. Langer, G.A., 1980. J. Molec. Cell. Cardiol. 12, 231. Latorre, R., C. Vergara and C. Hidalgo, 1982. Proc. Natl. Acad. Sci., USA 79, 805-809. Lea, E.J.A. and P.C. Croghan, 1969. J. Membr. Biol. 1, 225-227. Lee, A.G., 1977. Biochim. BiOphys. Acta 237-281. Lee, A.G., 1977a. Mol. Pharmacol. 13, 474-484. Lee, A.G., 1982. Trends Pharmacol. Sci. 3, 145-147. Lee, A.G., 1976. Nature 262, 545-548. Lefkowitz, 1979. Ann. Intern. Med. 91, 450-458. Lin, K.C., H.M. Weis and H.M. McConnell, 1982. Nature 296, 164-165. Lfillman, H. and M. Wehling, 1979. Biochem. Pharmacol. 28, 3409-3415. Lfiundstrhm, I., 1977. FEBS Letters 83, 7-9. McIlhenny H.M., 1971. J. Med. Chem. 14, 1178. McLaughlin, 8., 1975. In Molecular Mechanism of Anesthesia,¥grggress in Anesthesiolm, Vol. 1 (H.R. Fink, Ed.), pp. 193-200, New York, Raven Press. ‘ McLaughlin 8., 1977. In Current Topics in Membrane and Transport, Vol. 9 (F. Bronner, A. Kleinzeller, eds.), pp. 71-144, New York, Academic Press. Messineo, F.C. and .A.M. ‘Katz, 1979. J. Cardiovasc. Pharmacol. 1, Miller, Co, 1978. J. Membr. 31010 40, 1-230 Miller, C. and E. Racket, 1976. J. Membr. Biol. 30, 283-300. Murphree, S.A., T. Tritton, P.L. Smith and A.C. Sartorelli, 1981. Biochim. BiOphys. Acta. 649, 317-324. Mueller, P. and D.O. Rudin, 1967. Biochem. BiOphys. Res. Commun. 26, 398-404. Mueller, P. and D.O. Rudin, 1968. Nature 217, 713-719. Nachshen, D.A. and M.P. Blaustein, 1979. Mol. Pharmacol. 16, 579-586. Newton, D.W. and R.B. Kenza, 1978. Drug Intell. Clin. Pharm. 12, 546. 107 Noack, E., M. Kurzmack, S. Verjovski-Almeida and G. Inesi, 1978. J. Pharmacol. Exp. Ther. 206, 281-288. Ohki, 8., 1970. Biochim. BiOphys. Acta 219, 18-27. Ohki, 8., 1976. Frog. Surf. Membr. Sci. 10, 117. Ohki, 8., et al., 1980. In Proceedings of the the Symposium on Steric Effect in Biomolecules, PP. 183-193. Ohki, 8., 1984. Biochim. Biophys. Acta 777, 55-66. Pang, D.C. and N. Sperelakis, 1984. Biochem. Pharmacol. 33, 821-826. Papahadjopoulos, D., K. Jacobson, G. Posts and G. Shepherd, 1975. Biochim. BiOphys. Acta 394, 504-519. Phadke, RmSe, N.V. Kumar, R.V. Hosur, VFW. Kulkarni, A. Saran and G. Goril, 1981. Int. J. Quantum Chem. 20, 85-92. Rauls, D.0. and J.K. Baker, 1979. J. Med. Chem. 22, 81-86. Rojus, E. and T.M. Tobias, 1965. Biochim. Biophys. Acta 94, 394-404. Rooney, E.K., M.C. Cove and A.G. Lee, 1979. Biochem. Pharmacol. 28, 2199-2205. Sandermann, H., 1983. TIBS 8, 408-411. Sasaki, Y., T. Morita and S. Takeyama, 1984. Japan J. Pharmacol. 34, Schlieper, P. and P.R. Medda, 1980. Naunym-Schmiedeberg's Arch. Pharmacol. 310, 195-202. Schlieper, P. and R. Steins, 1983. Biochem. Pharmacol. 32, 799-804. Schreier, S., W.A. Frezatti, P.S. Jr. Aranjo, H. Chaimorich and I.M. Cuccoria, 1984. Biochim. Bipohys. Acta 769, 231-237. Seaman, P., 1972. Pharmacol. Rev. 24, 583-655. Shanks, R.G., 1984. Trends Pharmacol. Sci. 405-409. Shibuta, A., Y. Suezaki, H. Kamaya and I. Ueda, 1984. Biochim. BiOphy. Acta 772, 383-392. Shimshick, E.J. and H.M. McConnel, 1973. Biochemistry 12, 2351. Singer, M.A., 1977. Biochem. Pharmacol. 26, 51-57. Singer, S.J. and G.L. Nicolson, 1972. Science 175, 720-731. 108 Srivastava S., R. Phadke, G. Govil, H. Kranck, M.A. Rixmontel and D.Vasilescu, 1983. Physiological Chem. and Phys. and Medical mm 15, 469-477. Surewicz, W.K. and W. Leyko, 1981. Biochim. Biophys. Acta 643, 387-397. Szabo, C., G. Eisenman. and S.M. Ciani, 1969. J} Membr. Biol. 1, 346-382. Temple, D.M., W. Hasselbach and M. Makinose, 1974. Naunym-Schmiede- berg's Arch. Pharmacol. 282, 187-194. Thayer, S.A., M. Welcome, A. Chabra and A.S. Fairhurst, 1985. Biochem. Pharmacol. 34, 175-180. Tien, H.T., 1974. In Bilayer Lipid Membrane (BLM): Theory and Practice, Marcel Dekker, Inc., New York. Tien, H.T., 1985. Progress in Surface Science, in press. Tien, H.T. and Diana, 1967. Nature 215, 1199. Tien, H.T., S. Carbone and E.A. Dawidowicz, 1966. Nature 212, 718-719. Trauble, H. and H. Eibl, 1974. Proc. Natl. Acad. Sci. USA 71, 214-219. Triggle, D.J., 1982. In Biochemical Pharmacology of Calcium Blockers (S.F. Plain and R. Zelis, eds.), Urban, MD, pp. 121:134. Trudell, J.R., 1977. Anesthesiology 46, 5-10. Tsofina, L.M., Z.K.M. Khashayer and M.A. Dadayan, 1978. Biofizika 23, 257-260. Van Golde, L.M.G., V. Tomasi and L.L.M. Van Deenen, 1967. Chem. Phys. Lipids 1, 282-293. Vertleij, A.J., C. Mombers, J. Leunissen-Bijvelt and P.H.J.Th. Ververgaert, 1979. Nature 162-163. Verkleij, A.J., B. Humbel, B. Studer and M. Miller, 1985. Biochim. BiOphys. Acta 812, 591-594. Verkleij, A.J., R. DeMaagd, T. Leunissen-Bijvelt and B. deKruijtt, 1982. Biochim. Biophys. Acta 684, 255-262. Vuilleumier, P., P. Gazzati, E. Carafoli and W. Simon, 1977. Biochim. Biophys. Acta 467, 12-18. Wang, P.H. and E.J. Lien, 1980. J. Pharmaceutical Science 69, 662-668. Wang, S.M., et al., 1983. Kexue Tongbao 28, 60-62. Weskler, B.B., M. Gillick and J. Pink, 1977. Blood 49, 185-186. 109 Zubay, C., 1983. in Biochemistry, Addison-Wesley Publishing Comp., pp. 648-652. “'n’n'nnnMinnmy“