UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The influence of transmembrane ion gradients on the flux of weak acids and bases across liposomal systems Redelmeier, Tom E. 1988

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1989_A1 R42.pdf [ 9.1MB ]
Metadata
JSON: 831-1.0098265.json
JSON-LD: 831-1.0098265-ld.json
RDF/XML (Pretty): 831-1.0098265-rdf.xml
RDF/JSON: 831-1.0098265-rdf.json
Turtle: 831-1.0098265-turtle.txt
N-Triples: 831-1.0098265-rdf-ntriples.txt
Original Record: 831-1.0098265-source.json
Full Text
831-1.0098265-fulltext.txt
Citation
831-1.0098265.ris

Full Text

THE INFLUENCE OF TRANSMEMBRANE ION GRADIENTS ON THE F L U X OF WEAK ACIDS AND BASES ACROSS LIPOSOMAL SYSTEMS by TOM E. REDELMEIER A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF BIOCHEMISTRY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA DECEMBER 1988 ®TOM E. REDELMEIER, 1988 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT A primary role of biological membranes is to provide a permeability barrier to ion translocation. This thesis examines how ion gradients and in particular pH gradients influence the net flux of weak acids and bases across membranes. This first requires an understanding of the relationship between transmembrane electrical potentials (A*) and transmembrane pH gradients (ApH). It is demonstrated that an imposed A * or an imposed ApH will drive the net transport of protons across large unilamellar vesicles (LUVs). However, transport of protons does not result in the establishment of the expected electrochemical equilibrium over the time course investigated (8 hr). The results confirm that the stability of the transmembrane electrical potential is primarily influenced by the presence of permeable counterions and not to the high intrinsic proton permeability of the LUVs. The second area of investigation demonstrates that the presence of a high concentration of the weak base TRIS (2-amino-2-hydroxy-methylpropane-l,3-diol) in the external buffer of large unilamellar vesicles exhibiting an imposed A * (inside negative) decreases the apparent A * induced pH gradient. This model membrane system is utilized to examine the mechanism of transport of several lipophilic amines which are of biological interest. Transport of the amines at neutral pH is demonstrated to occur primarily via the neutral unprotonated molecule. In addition, it is demonstrated that the amines do not act as proton lonophores except in the presence of the lipophilic anion tetraphenylboron. The third area of investigation examines influence of transmembrane pH gradients on the net flux of acidic phospholipids. It is demonstrated that a transmembrane pH gradient (inside basic) will drive the net transbilayer transport of phosphatidylglycerol (PG) and phosphatidic acid (PA) but not phosphatidylserine (PS) across LUVs. The observed rate of PG transbilayer transport is several orders of magnitude faster than that previously observed in other model membrane systems. The pH dependence of the kinetics of transport is consistent with a model in which the neutral protonated form of PG is transported in response to pH ii gradients. The primary energetic barrier to transport of the neutral form of PG is the acyl chain component of the model membrane system. iii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS . iv LIST OF TABLES ix LIST OF FIGURES x ABBREVIATIONS xiii ACKNOWLEDGEMENTS xv 1. INTRODUCTION 1.1 Biological Membranes as Permeability Barriers 1 1.2 Structural Model of Membranes ; 1 . 1.3 Rationale for Studying the Permeability Properties of Model System 4 . 1.4 The Chemical Properties of Lipids 5 1.4.1 Structure 5 1.4.2 Acid-base behavior 8 1.5 The Material Properties of Lipid Systems 9 1.5.1 Lipid polymorphism 9 1.5.2 Gel to liquid-crystalline phase transitions 11 1.5.3 The order of a phosholipid bilayer 12 1.5.4 The surface potential of a membrane 13 1.6 Preparation of Model Systems 14 1.6.1 Multilamellar vesicles 14 1.6.2 Small unilamellar vesicles 15 1.6.3 Large unilamellar vesicles 15 1.6.4 Planar lipid membranes 17 iv 1.7 Ion Transport Across Membranes 17 1.7.1 Measuring transmembrane electrical potentials and pH gradients 18 1.7.2 Transport of nonionic solutes 20 1.7.3 Transport of electrolytes 23 1.7.4 Lipid factors which influence ion transport 26 1.8 Summary 27 2. PROTON TRANSPORT IN LARGE UNILAMELLAR VESICLES IN RESPONSE TO ELECTRICAL POTENTIALS AND pH GRADIENTS 2.1 Introduction 29 2.2 Materials and Methods 2.2.1 Materials 33 2.2.2 Preparation of egg phosphatidylcholine 33 2.2.3 Vesicle preparation and characterization 34 2.2.4 Uptake of probes into LUVs 35 2.2.5 Nuclear magnetic resonance 36 2.2.6 Calculation of membrane potentials and pH gradients 36 2.3 Results 2.3.1 Detection of electrical potentials and pH gradients 37 2.3.2 Electrically driven transport of protons across LUVs 40 2.3.3 The influence of the ionic compositiion of the external buffer on proton transport 42 2.3.4 Chemically driven transport of protons across LUVs 44 2.3.5 Ionophore mediated transport of N a + 46 2.4 Discussion 48 v 3. STUDIES ON THE UPTAKE OF LIPOPHILIC AMINES INTO MODEL MEMBRANE SYSTEMS IN RESPONSE TO ION GRADIENTS 3.1 Introduction 54 3.2 Materials and Methods 3.2.1 Materials .....;..;:.;......:..........„.....:.......:. :'. 60 3.2.2 Preparation of vesicles 60 3.2.3 Uptake of probes and drugs into LUVs , 60 3.3 Results 3.3.1 Generation of a model membrane system exhibiting a stable A * and no ApH .. .'„... 61 3.3.2 Uptake of lipophilic amines into model membrane systems 67 3.3.3 Lipophilic cations as proton ionophores 73 3.4 Discussion .76 4. PHOSPHATIDYLGLYCEROL ASYMMETRY IN L A R G E UNILAMELLAR VESICLES INDUCED BY TRANSMEMBRANE pH GRADIENTS 4.1 Introduction 85 4.2 Materials and Methods 4.2.1 Materials .' 89 4.2.2 Preparation of large unilamellar vesicles and establishment of ion gradients : 89 4.2.3 Detection of asymmetry of amine containing lipids 89 vi 4.2.4 Detection of acidic phospholipid asymmetry by ion exchange chromatography 90 4.2.5 Detection of phosphatidylglycerol asymmetry by periodate oxidation 91 4.2.6 Quasielastic light scattering : ... 92 4.2.7 Freeze-fracture 92 4.3 Results 4.3.1 . Phosphatidylserine is not transported in response to transmembrane pH gradients ; ....92 4.3.2 Transbilayer transport of phosphatidylglycerol as detected by ion exchange chromatography 94 4.3.4 Transbilayer transport of phosphatidylglycerol as detected by by periodate oxidation 97 4.3.5 The extent of phosphatidyglycerol asymmetry does not reflect the transmembrane pH gradient 103 4.3.6 Transbilayer asymmetry of other phospholipids which are simple weak acids 108 4.4 Discussion 112 STUDIES ON THE TRANSBILAYER TRANSPORT OF PHOSPHATIDYLGLYCEROL IN LARGE UNILAMELLAR VESICLES 5.1 Introduction : ". 117 5.2 Materials and Methods 5.2.1 Materials :.; .... '. 118 vii 5.2.2 Preparation of lysophosphatidylglycerol 119 5.2.3 Preparation of LUVs 114 5.2.4 Transbilayer transport of PG across LUVs 119 5.2.5 Detection of PG asymmetry 120 5.2.6 Detection of transmembrane pH gradients 121 5.2.7 Measuring the transmembrane distribution of lysophosphatidylcholine 121 5.2.8 Kinetic analysis of. phosphatidylglycerol transport 122 5.3 Results 5.3.1 Kinetic analysis df transport of across LUVs 126 5.3.2 Stability of the basic pH gradient as detected by entrapped pyranine 129 5.3.3 Transport of lysophosphatidylcoline in LUVs exhibiting PG asymmetry 131 5.3.4 The temperature dependence of PG transport 133 5.3.5 The pH dependence of PG transport 135 5.3.6 The influence of lipid composition on PG transport ; 137 5.3.7 Factors which influence the extent of PG asymmetry 141 5.4 Discussion 145 6 SUMMARIZING DISCUSSION 148 BIBLIOGRAPHY 151 viii LIST OF TABLES 1. The names and structures of some common fatty acids 7 2. The uptake of a variety of lipophilic cations in response to either transbilayer electrical potentials or pH gradients 55 3. A comparison of the uptake of the lipophilic cations in LUVs exhibiting a pH gradient (ApH > 2 units), with the chemical properties of the amines 82 4. Lipid factors which influence the rate constant for PG transport 139 5. Lipid factors which influence the extent of PG transport 142 ix LIST OF FIGURES 1. A structural model of biological membranes 2 2. The structure of a phospholipid and commonly occurring head groups 7 3. The relationship between the polymorphic phase behavior of a lipid and its shape 10 4. A * and ApH response to an imposed transmembrane pH gradients 38 5. The correlation between the measurement of pH gradients by [ i 4C]-MeAm redistribution and 3 1 P NMR 38 6. Electrically driven transport of protons as detected by MeAm redistribution 41 7. Electrically driven transport of protons as detected by following the chemical shift of entrapped Pi 41 8. Time course for the development of pH gradients in response to electrical potentials in SUVs 43 9. The influence of external buffer composition on the pH gradients induced in response to electrical potentials in EPC LUVs 43 10. Time course for the development of an electrical potential (inside negative) generated in response to a transmembrane pH gradient in EPC LUVs 45 11. Time course for the development of an electrical potential (inside negative) or a pH gradient (inside acidic) for vesicles experiencing a Na + diffusion potential 47 12. Four putative mechanisms for the transport of amines across membranes 57 13. Time course for the establishment of a measured electrical potential and pH gradient in EPC LUVs exhibiting potassium diffusion potentials 63 x 14. Time course for the establishment of a measured electrical potential and pH gradient in EPC LUVs when Tris is present in the external buffer 64 15. Time course for the establishment of a measured electrical potential and pH gradient in EPC LUVs when TPP + is present in the external buffer 66 16. Time course for the uptake of lipophilic amines into LUVs in the presence of external Tris 69 17. Time course for the establishment of a pH gradient (A) or electrical potential (B) across EPC LUVs in the presence of several lipophilic amines 74 18. Lipophilic amines as proton ionophores 75 19. Transbilayer transport of PS in response to pH gradients 93 20. The influence of ApH on the apparent surface potential of vesicles composed of EPCEPG (9.5:0.5 mol/mol). 95 21. The asymmetry of the periodate assay as demonstrated by the inaccessibility of entrapped 3-phosphoglycerol to periodate oxidation 98 22. Transbilayer transport of EPG in EPCEPG (9.5:0.5 mol/mol) LUVs as detected by periodate oxidation 98 23. Quasielastic light scattering of EPCEPG (9:1 mol/mol) LUVs 100 24. Freeze-fracture electron micrographs of EPCEPG (8:2 mol/mol) LUVs 101 25. The development of PG asymmetry at two temperatures 104 26. Transport of EPG in LUVs in response to acidic and basic pH gradients 106 27. The influence of stearylamine on the extent of PG transport 107 xi 28. The influence of ApH on the apparent surface potential of vesicles composed of EPC:PA (9.5:0.5 mol/mol) 109 29. The influence of ApH on the apparent surface potential of vesicles composed of EPCPI (9.5:0.5 mol/mol) 110 30. The influence of ApH on the apparent surface potential of vesicles composed of EPC:CL (9.5:0.5 mol/mol) I l l 31. Time course of PG transport from the OM to the IM of LUVs in response to a transmembrane pH gradient 127 32. The measurement of the basic pH gradient in LUVs exhibiting transmembrane PG asymmetry by entrapped pyranine ; 130 33. Transport of LPC in response to PG asymmetry 132 34. The influence of temperature on PG transport 134 35. The influence of external pH on the transport of PG 136 xii ABBREVIATIONS USED A A 4 1 2 Change in absorbance at 412 nm Ap Proton motive force ApH Transmembrane pH gradient A * Transmembrane electrochemical potential An" Charged anion B L M Black lipid membrane CAPS 3-(cyclohexylamino) propanesulfonic acid CCCP Carbonyl cyanide m-chlorophenylhydrazone EPPS N-(2-Hydroxyethyl)piperazine-N'-3-propanesulfonic acid ESR Electron spin resonance FATMLVs Frozen and thawed MLVs H n Hexagonal HA neutral amine H 2 A + protonated amine HOCIO3 Perchloric acid HEPES [4-(2-Hydoxyethyl)]-piperazine ethanesulfonic acid IM inner monolayer i-v Current-voltage LUVs Large unilamellar vesicles LUVETs Large unilamellar vesicles by extrusion techniques MES 2-(N-Morpholino)ethanesulfonic acid MLVs Multilamellar vesicles Mo0 3 Molybdenum trioxide MTPP + Methyltriphenylphosphonium xiii NaHSO s Sodium bisulfite Na 2S0 4 Sodium sulfate NMR Nuclear magnetic resonance OM outer monolayer Phospholipids CL Cardiolipin DOPC l,2-dioleoyl-sn-glycero-3-phosphorylcholine DPOPC 1,2-dipalmitoleoyl-s«-glycero-3-phosphorylcholine DPPC 1,2-dipalmitoyl-sn-glycero-3-phosphorylcholine DOPE 1,2-dioleoyl-sw-glycero-3-phosphorylethanolamine DOPG l,2-dioleoyl-s«-glycero-3-phosphorylglycerol EPC (egg) PC derived from hen egg yolk FA Fatty acids LPC Lysophosphatidylcholine MOPC 1 -monooleoyl-sn-glycero-3-phosphorylcholine MOPG 1 -monooleoyl-sM-glycero-3-phosphorylglycerol PA Phosphatidic acid PC Phosphatidylcholine PE Phosphatidylethanolamine PG Phosphatidylglycerol PI Phosphatidylinositol PS Phosphatidylserine SA Stearylamine SM sphingomyelin PL Phospholipid PLM Planar lipid membrane psi Pounds per square inch QELS Quasi-elastic light scattering SD Standard deviation SUVs Small unilamellar vesicles /HBCs transient hydrogen bonded chains (water) TPP + Tetraphenylphosphonium TPB" Tetraphenylboron Tc Gel to liquid-crystalline temperature T LC Thin layer chromatography TRIS 2-amino-2-hydroxy-methylpropane-l,3-diol xiv ACKNOWLEDGEMENTS I would like to acknowledge and thank the members of PRC's laboratory. My experience at UBC has been enjoyable and very rewarding and this is due in no uncertain terms to my colleagues in the lab. In. addition, I would like to specifically thank the following individuals who have played an especially important role during my PhD; Janet Wood, Marcel Bally, Lawrence Mayer, Kim Wong, Mick Hope, Helen Loughrey, Tom Madden, Richard Harrigan, and Kath Quayle. Finally, I would like to thank Pieter Cullis for establishing an environment suitable for the enjoyment of scientific research. His excitement, personal support, humour, and unflagging interest was always appreciated though seldom acknowledged. This work was supported financially by fellowships from the University of British Columbia and by PRC's laboratory. xv To My Mom and Dad Brother and Sister xvi INTRODUCTION 1.1 Biological Membranes as Permeability Barriers Biological membranes provide permeability barriers which serve to separate internal and external environments. This primary function has been attributed to the presence of a lipid bilayer which participates in such diverse events as cellular protection, organization, homeostasis and energy transduction. The recognition of the role of transmembrane electrical potentials and ion gradients in membrane related processes such as energy transduction (Mitchell, 1961; Hatefi, 1985), secondary solute transport (Schuldiner and Kaback, 1975), receptor recycling (Helenius et al., 1983), endocytosis (Mellman et al., 1986) and protein insertion (Wickner, 1983; Eilers and Schatz, 1988) has led to a renewed interest in the nature of the permeability barrier that lipid bilayers represent. Model membrane systems have provided a simple tool for studying the general permeability properties of lipid bilayers. However, the influence of transmembrane ion gradients on the permeability properties of model membrane systems has received little attention. The work presented in this thesis characterizes the influence of transmembrane pH gradients and membrane potentials on the flux of weak acids and bases across model membrane systems. 1.2 A Structural Model of Membranes Our current understanding of the structure of biological membranes was developed in the first part of this century and is summarized by the Fluid Mosaic Model (Singer and Nicholson, 1972). The major features of this model are illustrated in Fig. 1. Biological membranes are composed of lipid (primarily phospholipid and sterol), protein and oligosaccharide. The phospholipid, present in a bilayer form (for a historical perspective see Danielli, 1982), provides not only the major barrier to transmembrane solute movement but also 1 Fig . 1. A structural model of biological membranes. The phospholipid is present in a bilayer form, the protein is present both in an integral and a peripheral association with the lipid and the carbohydrate is seen to be attached covalently to the lipid and the protein. A transbilayer membrane potential is seen to exist across the membrane. Plasma membrane 2 provides the matrix with which membrane proteins are associated. The protein, found both in a "peripheral" and in an "integral" association with the lipid, is involved in such diverse processes as energy transduction, solute transport, transmembrane signaling and anabolic and catabolic metabolism. Finally, the oligosaccharides, attached to the lipid or protein moiety, are known to be receptors for a variety of ligands as well as being involved in cell attachment and recognition. The structural arrangement of membrane components has been an area of intense investigation (for a review see Kleinfeld, 1987; Hui, 1987; Thompson and Tillack, 1985). A fundamental characteristic of the structural arrangement of biological membranes is that the components exhibit rapid lateral movement (for a review see Edidin, 1987). In addition, the protein is thought to have only weak interactions with specific lipid species (for a.review see Devaux and Seigneuret, 1985; Watts, 1987). However, despite the ability of the components to rapidly migrate laterally, and the absence of evidence to suggest a widespread specific interaction between lipid and protein there are indications of some lateral organization of individual components in the membrane (Kleinfeld, 1987; Hui, 1987). For example, evidence deduced from freeze-fracture electron microscopy and photobleach recovery of fluorescent probes has indicated specific structural arrangements of ligand-receptor complexes (for a review see Mellman et al., 1987; Jain, 1983). In addition, microdomains of lipid have been postulated to exist in biological membranes (for a review see Thompson and Tillack, 1985; Jain, 1983). An additional feature of biological membranes is that the major components are asymmetrically distributed across the membrane (see Fig. 1). For example, studies utilizing chemical modification, susceptibility to protease degradation or accessibility to substrate have demonstrated that integral membrane proteins are found specifically oriented with respect to the bilayer (for a review see Etemadi, 1980a). Presumably, this is an important feature of enzymes which are involved in vectorial processes. Maintenance of protein asymmetry after the initial insertion of the protein indicates that the transbilayer reorientation of the protein is extremely, slow. In addition to the asymmetry seen for proteins, asymmetry of lipid is also observed 3 (Etemadi, 1980b; Chapter 4). However, the functional significance of lipid asymmetry is not presently understood. An important feature of the model illustrated in Fig. 1 is that it indicates the presence of a transmembrane electrical potential (A*). A transmembrane electrical potential can be generated across biological membranes by the vectorial transport of a charged ion (or electron) which is not compensated by a movement of another ion. This process is typically protein mediated and is linked to an energy producing event such as hydrolysis of ATP (Na + /K + ATPase; proton ATPase), substrate oxidation (oxidative phosphorylation) or absorption of light (photosynthesis). In addition, similar energy linked vectorial processes will lead to the generation of transmembrane ion gradients. For example, this has been demonstrated for the generation of transmembrane pH gradients in eukaryotic subcellular organelles such as secretory vesicles, , iysosomes, endocytic vesicles and the Golgi apparatus (for a review see Mellman et al., ,1986; Anderson and Orci, 1988). Thus, the asymmetric nature of membranes, in terms of its structural components and in terms of transmembrane ion gradients is an important feature of our present understanding of membranes. 1.3 Rationale for Studying the Permeability Properties of Model Systems Upon hydration, most phospholipids will form closed structures (termed liposomes) which have permeability characteristics qualitatively similar to biological membranes (Deamer and Bramhall, 1986). Since the lipid bilayer is thought to present the major permeability barrier to solute transport across biological membranes, it seems appropriate to use liposomes to model the permeability properties of the more complex biological membranes. Studies utilizing model membrane systems have provided insight into the nature of the energetic barrier that lipid, bilayers represent to solute transport (for example see Deamer and Bramhall, 1986; Fettiplace and Haydon, 1980; Section 1.7). However, the influence of A * and ApH on the passive transport across liposomes of biologically relevant weak acids and bases has received relatively 4 little attention. In this thesis the influence of transmembrane electrical potentials and pH gradients on the transport of solutes across model membrane systems will be investigated. Chapter 2 characterizes the relationship between the flux of protons across LUVs in response to either a transmembrane A * or a transmembrane ApH. The mechanism of transport of lipophilic amines in response to A * and ApH is examined in Chapter 3. Finally, Chapters 4 and 5 discuss the nature of pH driven transbilayer transport of acidic phospholipids across model membrane systems. The remainder of this Introduction will summarize our current understanding of the chemical properties of lipids, the material properties of liposomes and how these factors influence the permeability characteristics of model membrane systems. 1.4 The Chemical Properties of Lipids The chemical properties of lipids have been extensively reviewed elsewhere (for a review see Small, 1985). It is the purpose of this section to give a brief introduction to these properties and detail those factors which may influence the permeability barrier of liposomes. The term lipid refers to substances which are soluble in apolar solvents. They comprise a highly diverse group of substances including fats, phospholipids, fatrsoluble vitamins, sterols and waxes. Here, attention will be directed to commonly occurring membrane lipids such as phospholipids (PL), fatty acids (FA) and sterols since these form the major structural components of prokarytic and eukaryotic membranes (for example see Cullis and Hope, 1985; van Deenen and de Gier, 1974). However, lipids such as diacylglycerol (DG), triacylglycerol (TG), glycolipids and the polyprenoids are important minor components of biological membranes. In turn, they are expected to alter the permeability characteristics of biological membranes. 1.4.1 Structure In Table 1, the names and structures of the most abundant fatty acids are illustrated. 5 They share a common property of all membrane associated lipids in that they are amphiphiles, that is, they contain a hydrophilic head group and a hydrophobic tail. The length and degree of saturation of the acyl chain will markedly influence the hydrophobicity of the molecule (for example see Tanford, 1980) and consequently the material properties of the bilayer (see Section 1.5). However, free fatty acids do not represent a major structural component of membranes. For example, they comprise less than 1% of the total lipid of the erythrocyte plasma membrane (Nelson, 1967). Their primary importance lies in that they provide a major component of phosphoglycerides and other membrane lipids. The phosphoglycerides represent the most important mammalian membrane lipids. The structure of phosphoglycerides, illustrated in Fig. 2, demonstrates the glycerol-3-phosphate backbone with a hydrophilic head group attached via a phosphate ester to the sn3 position of the glycerol and two acyl chains attached via esters to the sn1 and sn2 positions. The structures of the polar head groups are also illustrated in Fig. 2. Note that there is chemical diversity between the polar head groups in terms of the chemical reactivity, charge, polarity, size and hydrogen bonding capacity. The chemical nature of the polar head group will be expected to influence the material properties of the bilayer (Section 1.5). Another important class of lipids is the sphingolipids. Sphingomyelin, the most common mammalian sphingolipid, has a long-chain hydroxylated secondary amine as a backbone, is acylated on the secondary amine and has a phosphocholine esterified to the primary alcohol. Table 1 and Fig. 2 should help to emphasize the chemical diversity that phospholipids represent. It has been estimated that there are well over 100 chemically different phospholipid species in the erythrocyte plasma membrane. There is presently little or no understanding of the functional importance of most of the individual lipid species. The third important class of membrane lipids is the sterols. Cholesterol, the most abundant mammalian sterol, is also amphiphilic though the size of the hydroxyl polar head group is much smaller than that of a phospholipid (Huang and Mason, 1982). The molecule exists as a rigid planar hydrophobic region with an acyl side chain at one end and a small polar 6 Table 1. The names and structures of some common fatty acids No of Carbon Atoms Structural Formula Name Saturated fatty acids 12 C H 3 ( C H 2 ) 1 0 C O 2 H lauric acid 14 C H 3 ( C H 2 ) 1 2 C 0 2 H myristic acid 16 C H 3 ( C H 2 ) 1 4 C 0 2 H palmitic acid 18 C H 3 ( C H 2 ) 1 6 C 0 2 H stearic acid 20 C H 3 ( C H 2 ) l g C 0 2 H arachidic acid Unsaturated fattv acids 16 18 18 18 20 CH 3 (CH 2 ) g CH=CH(CH 2 ) 7 C0 2 H CH 3 (CH 2 ) 7 CH=CH(CH 2 ) 7 C0 2 H CH 3 (CH )4(CH=CHCH ) (CH 2 ) 6 C0 2 H CH 3CH 2(CH=CHCH ) (CH ) 6CO H CH,(CH 2) 4(CH=CHCH,) 4(CH,),CO,H 2'4V 2'2V palmitoleic acid oleic acid linoleic acid linolenic acid arachidonic acid Fig. 2. The structure of a phospholipid and commonly occurring head groups. Ethanolamine (PE) Serine (PS) Glycerol (PG) Glycerol (CL) Myoinositol (PI) 7 head group attached to the other end. It intercalates in the bilayer, such that the polar head group is associated with the fatty acyl carbonyls of the phospholipids and the planar hydrophobic region is associated with the acyl chains (Huang, 1977). It is a major structural component of the plasma membrane of mammalian cells but is virtually absent from the bilayer of such mammalian organelles as the endoplasmic reticulum and the mitochondria (Cullis and Hope, 1985). 1.4.2 Acid-base behavior The common polar head-groups listed above exhibit acidic and basic characteristics. FA, phosphatidic acid (PA), phosphatidylinositol (PI), phosphatidylserine (PS) and cardiolipin (CL) carry a net negative charge at physiological pH whereas phosphatidylcholine (PC) phosphatidylethanolamine (PE) and sphingomyelin (SM) carry no effective net charge: However, PS, PE, PC and SM are zwitterionic and consequently contain charged groups at all pH values, whereas FA, PI, PA, CL and PG are weak acids and therefore carry no effective net charge at pH values below the p K a of the acid. The value of the pK & of the membrane bound carboxyl of FA has been reported to be 7.3 (Rooney et al., 1983) and is expected to be influenced by both the surface potential of the membrane as well as the presence of hydrogen ion donors and acceptors (Cevc et al., 1988). The value of the intrinsic p K a of the PG phosphate has been reported to be 1 (Watts et al., 1978) and is not expected to be considerably different for other phospholipids. The value of the intrinsic pK & of the PE and PS amino groups have been reported to be 9.6 and 9.8 respectively whereas the pK & of the PS carboxyl group has been reported to be 3.6 (Tsui et al., 1986). Cholesterol does not carry a net charge. The actual charge that a phospholipid carries at a given pH will be influenced by the ionic composition of the medium as well as the surface charge of the vesicles (Boggs, 1987). In turn, the charge that a lipid carries will influence the size, polarity, hydrogen bonding capacity and chemical reactivity of the head group. 1.5 The Material Properties of Lipid Systems Chemical properties of lipids refer to the characteristics that may be attributed to the individual molecules (Gruner, 1987) whereas the material properties of lipids refer to the characteristics attributed to the cooperative action of a number of these molecules. Phospholipids will aggregate at low concentrations into higher order structures. The driving force for this macromolecular assembly is termed the hydrophobic effect; that is, the chemical potential of the hydrocarbon is reduced by approximately 3.7 kJ mol"1 per methylene group in going from an aqueous to an apolar environment (Tanford, 1980). It is the material properties of a liposome which are of significance to the study of the permeability characteristics of a lipid bilayer. 1.5.1 Lipid polymorphism Lipid polymorphism will not be covered in detail here, but the interested reader is referred to recent reviews in this area (Cullis et al., 1983; Gruner et al., 1985; Tilcock, 1986). There are many possible structures that lipids can adopt upon hydration including the familiar bilayer, as well as micelles and the hexagonal H n phase. The phase preference of a lipid has been rationalized in terms of its dynamic shape characteristics (Cullis and de Kruijff, 1979). The polymorphic phase behavior that a variety of lipids adopt upon hydration at physiological pH and temperature is illustrated in Fig. 3 along with the putative shape that characterizes the various lipids. Bilayer forming lipids appear to have a cylindrical shape, H n forming lipids have a cone shape and micellar lipids have an inverted cone shape. The ability of lipids to form non-bilayer phases has been used to provide a rationale for the chemical diversity of lipids in terms of their putative role in cellular events such as membrane fusion and transmembrane transport (see for example Cullis et al., 1983). However, the lipid in biological membranes is generally found in a bilayer configuration and consequently it is the material properties of bilayers 9 Fig. 3 The relationship between the polymorphic phase behavior of a lipid and its shape. LIPID PHASE MOLECULAR L Y S O P H O S P H O L I P I D S D E T E R G E N T S S H A P E M I C E L L A R I N V E R T E D C O N E P H O S P H A T I D Y L C H O L I N E S P H I N G O M Y E L I N P H O S P H A T I D Y L S E R I N E P H O P H ATI D Y L I N O S I T O L P H O S P H A T I D Y L G L Y C E R O L P H O S P H A T I D E A C I D C A R D I O L I P I N D I G A L A C T O S Y L D I G L Y C E R I D E HI i B I L A Y E R C Y L I N D R I C A L P H O S P H A T I D Y L E T H A N O L A M I N E ( U N S A T U R A T E D ) C A R D I O L I P I N - C a 2 + P H O S P H A T I D I C A C I D - C a 2 + ( P H < 6 . 0 ) P H O S P H A T I D I C A C I D ( p H O . O ) P H O S P H A T I D Y L S E R I N E ( p H < 4 . 0 ) M O N O G A L A C T O S Y L D I G L Y C E R I D E H E X A G O N A L ( H , , ) A C O N E 10 that will be discussed here. 1.5.2 Gel to liquid-crystalline phase transitions A particularly well studied material property of a bilayer is the gel-liquid crystalline phase transition. Hydrated dispersions of pure phospholipid will undergo sharp gel liquid-crystalline phase.transitions (for a review see Chapman, 1975; Seelig, 1978). The gel phase lipid is characterized as a highly ordered viscous (gel) state whereas the liquid-crystalline phase is characterized as a less ordered (melted) or fluid state. Differential scanning calorimetry (DSC) is the most common technique for characterizing the phase behavior of pure lipid systems in terms of the transition temperature (Tc), the enthalpy and the width of the transition. Other techniques such as X-ray crystallography, NMR and freeze-fracture electron microscopy are used to assign the physical structures to the different phases. The T c of a pure lipid is sensitive to the nature of the polar head group, the acyl chain length and the degree of saturation of the hydrocarbon. For example, the T. of DPPC is 41°C, the T. of DPPS is 54°C and the T. of DPPA is 65°C (for a compilation of data see Silvius, 1982). The increase in T for lipids carrying a net negative charge has been attributed to an increased stability of the gel phase due to the greater electrostatic repulsion for charged head groups. In addition, the hydrogen bonding capabilities of the polar head group appears to play a role in determining the T c (for a discussion see Boggs, 1987). Increasing the acyl chain length of PCs increases the T by approximately 20°C per two-carbon unit whereas inclusion of unsaturated acyl chains markedly decreases the transition temperature (Chapman, 1975). Factors such as the presence of divalent cations or low pH have also been shown to decrease the phase transition temperature of acidic phospholipids which in turn has been attributed to a reduction in surface charge (Watts et al., 1978; Cevc et al., 1981). Cholesterol has been shown to influence markedly the phase behavior of phospholipids. Cholesterol decreases the enthalpy of the transition and above 33 mol% it 11 abolishes T of the lipid as measured by DSC (Chapman, 1975). The importance of the physical state of the lipid in modulating the permeability properties will be discussed in Section 1.7. 7.5.3 The order of a phospholipid bilayer Recent work has characterized the degree of motion of phospholipids present in a bilayer (for reviews see Davis, 1983; Seelig, 1977). In addition to the lateral motion (in the plane of the bilayer) phospholipids will experience rotational motion (around the long axis of the lipid) and rapid segmental motion (e.g. trans-gauche isomerizations in the acyl chains). The rotational and segmental motion of the phospholipid can be characterized by ESR or 2 H NMR methods. Both methods give rise to an order parameter which characterizes the average orientational order of a probe (spin label or selectively deuterated methyl group) situated in the bilayer. 2 H NMR studies have shown that the first 10 carbons of the acyl chain (going into the bilayer) exhibit similar motional characteristics whereas the remaining carbons (C 1 0 - C 1 6 ) exhibit a decreased average motional order (see for example Seelig, 1977). Thus, amphiphiles exhibit a mobility gradient parallel to the bilayer normal in model membrane systems. It is important to distinguish between the order parameter associated with a particular C - 2 H bond and the fluidity of a membrane. The term fluidity has been used in the literature to signify any changes (temperature, acyl chain length, saturation, cholesterol content) in a membrane which alter the "order" of the bilayer (see for example Hauslay and Stanley, 1982). However, the fluidity of a membrane is formally the reciprocal of the membrane viscosity. As such it is expected to be related to the rate of lateral motion (diffusion) of a lipid or solute in a membrane and not the degree of segmental motion associated with the acyl chains. Consequently, as pointed out by Seelig (1977), there is no simple relationship between the fluidity of a bilayer and the order parameter. For this reason, the permeability properties of model membrane systems will be discussed in terms of an average order parameter rather than . the general concept ofembrane fluidity. 12 1.5.4 The surface potential of a membrane The association of acidic or basic molecules with a membrane will result in the existence of a surface potential at the membrane interface. For example, the presence of the acidic phospholipid PS in a membrane at physiological pH will result in a negative surface potential which in turn will attract cations to the interface of the membrane and repel anions. It is generally accepted that the surface potential profile next to the membrane interface is accounted for by the Gouy-Chapman theory of the diffuse double layer (McLaughlin, 1977; Winiski et al., 1986; Hartsel and Cafiso, 1986; Flewelling and Hubbell, 1986b). For low values of charge density at the interface the magnitude of the electrostatic potential at the interface increases in a linear fashion. The magnitude of the potential at the interface is also sensitive to the presence of counter ions, decreasing with increasing concentration of counter ions. The Gouy-Chapman theory of the diffuse double layer predicts that a membrane composed of 20 mol% acidic lipid will have a surface potential of 60 mV in the presence of 100 mM monovalent cations (McLaughlin, 1977). A surface potential of 60 mV is expected to increase the concentration of cations at the interface by an order of magnitude with respect to the bulk phase, to decrease the interfacial pH by 1 unit and to decrease the interfacial anion concentration by an order of magnitude. Surface potentials are expected to influence the p K a of acids and bases (see also Section L7) which bind to the membrane interface. This gives rise to the concept of apparent and intrinsic pK as. The apparent p K a for a weak acid or base is determined as the pH of the bulk solution when the group is 50% charged whereas the intrinsic p K a is determined as the surface pH when the group is 50% charged (see for example Tsui et al., 1986; Watts et al., 1978). Consequently the presence of a surface charge on a membrane will be expected to influence the relative concentration of the charged and uncharged weak acids or bases present at the interface. The importance of surface potentials on the flux of weak acids and bases will be discussed in more detail in Section 1.7. 13 1.6 Preparation of Model Membrane Systems A number of techniques for the preparation of model membrane systems exist (for a review see Hope et al., 1986; Szoka and Papahadjopoulos, 1980). The choice of the method for the preparation of the model membrane system can influence the permeability properties of the bilayer. This section reviews standard protocols for the preparation of model membrane systems and comments on their advantages and disadvantages in terms of measuring weak acid and base transport in response to transmembrane pH gradients. . 1.6.1 Multilamellar vesicles The first model membrane system, developed by Bangham (1965), involved a gentle dispersion (e.g. vortex mixing) in buffer of lipid which had been dried down from organic solvent onto the walls of a test tube. The resulting vesicles (termed MLVs) are multilamellar, typically have diameters in excess of 1000 nm, are heterogeneous in size, exhibit trapped volumes of approximately 1 pi per /*mol phospholipid and exhibit an unequal distribution of solute across the lamellae (Gruner et al, 1985; Mayer et al., 1985c). MLVs which exhibit an equal distribution of solute entrapped between the lamellae may be made by evaporation of ether from an ether-buffer-lipid mixture (termed Stable Plurilamellar Vesicles, SPLVs) (Gruner et al., 1985) or repeated freezing and thawing of an MLV preparation (termed Frozen and Thawed MLVs, FATMLVs) (Mayer et al., 1985c). Another technique which results in M L V formation involves the dehydration of lipids from an aqueous solution by either freeze-drying or direct vacuum evaporation followed by controlled rehydration (Kirby and Gregoriadis, 1984). These procedures are simple, can be used with any bilayer forming lipid, improve the amount of solute entrapped in the liposomes (ca effective trap volume 10-20 /*1 per /imol phospholipid) as well as the equilibration of solute across the lamellae. However, the major disadvantages to all of these preparations is that the vesicles are multilamellar and are heterogeneous in size. In 14 turn, this complicates the measuring of permeability coefficients for solute transport. There have been a variety of attempts to make a simple unilamellar model membranes which would overcome the disadvantages of the multilamellar vesicles. 1.6.2 Small unilamellar vesicles The first attempts to make homogeneous unilamellar vesicles were based on sonication of MLVs to form limit size vesicles (see for example Huang, 1969). The size of these vesicles (termed small unilamellar vesicles, SUVs) is dependent on the lipid composition varying from 25-30 nm for EPC to 50 nm for cholesterol containing systems (Johnson, 1973). Vesicles of this size range may also be prepared by the French press technique (Barenholz, 1979). Disadvantages to this procedure include vesicle instability, small aqueous trapped volumes (< 0.2 pi per /jmol phospholipid), and the possible presence of lipid oxidation products generated during preparation. Permeability coefficients for ion transport across SUVs are typically 1-2 orders of magnitude smaller than for corresponding larger systems (Deamer and Bramhall, 1986; Perkins and Cafiso, 1986). This observation is not presently understood. In order to overcome the inherent disadvantages of SUVs several procedures have been developed to make unilamellar vesicles of a larger size. 1.6.3 Large unilamellar vesicles Large unilamellar vesicles (LUVs) can be made by ethanol injection (Kremer et al., 1977), ether infusion (Deamer and Bangham, 1976), reverse phase evaporation (Szoka and Papahadjopoulos, 1978), detergent dialysis (see for example Madden, 1986) and medium pressure extrusion (Hope et al., 1985). The first three methods involve the dispersion of lipids in a monomer form in an appropriate organic solvent (for example, ethanol or ether). The lipid is subsequently hydrated by injection of the lipid mixture into a buffer. The organic solvent is 15 evaporated at the time of hydration for the ether method, removed under reduced pressure for the reverse phase evaporation, or is diluted into the buffer for the ethanol injection vesicles. An additional step involving column chromatography is occasionally performed to aid in the further removal of organic solvents. Detergent dialysis techniques involve the detergent induced solubilization of the lipid into micelles and the subsequent removal of the detergent by dialysis. This technique is particularly useful for the incorporation of proteins into model membrane systems. Efficient removal of detergent by this technique requires the use of a detergent with relatively high critical micellar concentration (cmc) such as cholate or octylglucopyranoside. LUVs prepared by all of the above techniques have average diameters of between 50 - 200 nm and trapped volumes of between 1-3 /d per /*mol phospholipid (Szoka and Papahadjopoulos, 1980). However, these techniques have limitations in that they are tedious, may not be applicable to all lipids and may contain residual detergent or organic solvents^  which can alter the permeability properties of the resulting membranes. The technique of choice for making large unilamellar vesicles is medium pressure extrusion (Large Unilamellar Vesicles by Extrusion Techniques, LUVETs). The technique for the extrusion of liposomes is based on the initial observation by Olson and co-workers (1979) who demonstrated that reverse phase vesicles exhibited greater size homogeneity after low pressure extrusion through polycarbonate filters of a defined pore size. This observation was extended by Cullis and co-workers (Hope et al., 1985; Mayer et al., 1985c), who demonstrated that extrusion of MLVs or FATMLVs through polycarbonate filters of a defined pore size ( <100 nm) results in the generation of unilamellar vesicles. For example, EPC V E T 1 0 Q (LUVETs produced by extrusion through 100 nm pore size filters) exhibit average diameters of 90 nm as judged by freeze fracture electron microscopy and quasi elastic light scattering techniques, have trapped volumes of 1.5 /J1 per /xmol phospholipid and are unilamellar as judged by freeze fracture and NMR techniques (Hope et al., 1985). This technique has been discussed in detail elsewhere (Hope et al., 1986). Though this technique represents an important breakthrough in liposome technology there still exists a need for a simple technique for making vesicles which 16 are stable, unilamellar, homogeneous in size but larger in diameter than the conventional LUVETs. 1.6.4 Planar lipid membranes , Planar Lipid Membranes (PLMs) can be made by dissolving lipid in a hydrocarbon solvent and then applying this mixture to a thin aperture separating two aqueous compartments (Mueller et al., 1962; Fettiplace et al., 1974). The solvent collects at the perimeter of the aperture leaving central lipid in a bilayer form. The advantage of this technique is that it allows complete control by the investigator of the two compartments. It is possible to impose transmembrane chemical gradients and monitor the current flux across the lipid bilayer, or alternatively maintain a constant membrane potential across the lipid bilayer and monitor the flux of radiolabeled ions. Disadvantages to this methodology include the relatively small amount of bilayer present in an experiment which decreases the sensitivity of the observation, the possible perturbation of the permeability properties of the bilayer by residual solvent and the potential artifacts due to the interactions of the lipid with the edges of the aperture. 1.7 Ion Transport Across Membranes The Introduction to this point has briefly reviewed the chemical properties of lipids, several of the material properties of liposomes and the current procedures for making model membrane systems. In order to understand the influence of transmembrane ion gradients on the transport of weak acids and bases across model membrane systems it is necessary to briefly review the nature of electrical potentials, acid dissociation constants, and transmembrane pH gradients. Our current understanding of the permeability properties of model membrane systems to non-ionic solutes and electrolytes and the influence of lipid on these properties will be discussed in Sections 1.7.2, 1.7.3 and 1.7.4. 17 1.7.1. Measuring transmembrane electrical potentials and pH gradients A transmembrane electrical potential (A*) is established across a membrane by the transmembrane movement of a charged ion (or electron) which is not compensated for by the movement of another ion. The resulting A * can be calculated from the Nernst equation: (P N a [Na + ] i + P K [ K + ] i + P c l [ C r ] o + . . . ) A * = - RT / nF In (1) ( P N a [ N a + ] o + PK [ K + ] o + P c 1 [ C n i + - ) where R is the gas constant, T is the temperature, F is the Faraday constant, and P is the permeability coefficient for the ions. A * is typically measured across liposomes by determining the ratio of the inside and outside concentrations at equilibrium of a permeable charged ion (see for example Hope et al., 1985; Cafiso and Hubbell, 1978a; Rottenberg, 1979; Chapter 2). A transmembrane electrical potential is expected to drive the net flux of a membrane permeable ion such that at equilibrium the electrochemical potential for the ion (C) is zero: A * = (RT / nF) In (C4 / C o) (2) where the subscripts refer to the inside and outside of the liposome (see for example Rottenberg, 1979). For the case of protons the expected electrochemical equilibrium is reached when: A * = (2.3 RT / F) ApH (3) where the expression 2.3 RT / F is equal to 60 mV at 25 °C. The influence of A * on the transmembrane flux of protons is discussed in more detail in Chapter 2. 18 Many compounds of biological interest have proton donating (acidic) or proton accepting (basic) groups. The equilibrium constant (K a) for a weak acid can be described by the following equation: K a = [H+] [A"]/[HA] (4) where [H+] is the hydrogen ion concentration, [A"] is the concentration of the ionized weak acid and [HA] is the concentration of protonated acid. The p K a is the negative log of the acid dissociation constant. The relationship between the pH of the medium and the relative concentrations of the ionized and protonated weak acid is described by the Henderson-Hasselbach equation: pH = p K a + log ([A"] / [HA]) (5) Similar equations may be used to describe the dissociation of weak bases. Transmembrane pH gradients may be measured by determining the distribution at equilibrium of weak acids and bases (for a review see Rottenberg, 1979). If one assumes that the dissociation constant for a weak acid remains the same on both sides of a membrane then: K a = [H +].[A-]./[HA]. = [H + ] o [A- ] o / [HA] o (6). It is generally accepted that weak acids permeate across membranes as the neutral species (Addanki et al., 1968; Rottenberg, 19.73). Consequently, at equilibrium the concentration of the neutral species will be the same on both sides of the membrane and the distribution of a weak 19 acid (if pK < pH ) will reflect the pH gradient: & O [A-] 0 / [A-]. = [H+]. / [H +] o . ; (7) A major focus of this thesis deals with the flux of weak acids and bases across liposomes in response to transmembrane pH gradients. 7.7.2 Transport of nonionic solutes The net flux of a nonionic solute across a membrane can be described by: J = P A ( C i V C o ) (8) where J is the flux in moles cm2 sec"1, P is the permeability coefficient in cm sec"1, A is the area of the membrane in cm2, and C is the solute concentration on the inside (i) or outside (o) of the enclosed membrane vesicle (for example see Jain, 1980). It is important to note that the driving force for net flux of a non-ionic solute across a membrane is the chemical potential of the solute. Typically, permeability measurements are made by monitoring the net transport of the solute across a bilayer in response to a chemical gradient or by monitoring the release of a radiolabeled solute under conditions where there is'no transmembrane solute chemical potential. For the latter case, the permeability coefficient of a radiolabeled solute entrapped inside a liposome can be calculated from a rearrangement of Equation 8: P = k Vj / A (9) where V. is the internal volume of a vesicle and k is the equilibration constant derived from the slope of a plot of In (Cj(t) / Cj(o)) vs time. For a permeability coefficient of 10"6 cm s"1 20 the half time of release (t^) of a radio-labelled solute from a 100 nm diameter L U V can be calculated to be 0.1 sec whereas for a permeability coefficient of 10"12 cm s"1, t^2 = 13.3 d. Perhaps the most important and most studied non-electrolyte is water. Water transport across model membrane systems has been studied by two independent techniques: the diffusional permeability coefficient (Pd) is determined by following the transport of water by isotope exchange across PLMs under isotonic conditions and the flux permeability coefficient (Pf) is determined by monitoring the osmotically driven increase in volume of a membrane bound compartment such as a liposome (for a review see Fettiplace and Haydon, 1980). For diffusion of water through aqueous pores, the ratio of PfP d is expected to be greater than one due to differences in the resistance to bulk flow versus diffusional flow whereas for diffusion of water through the hydrocarbon the ratio will be expected to be one (Fettiplace and Haydon, 1980). For example, a Pj:Pd ratio of 6 has been reported for diffusion of water across the erythrocyte plasma membrane (Macy and Farmer, 1970). Treatment of the erythrocyte with p-chloromercuribenzosulfonic acid decreases the ratio to unity implying that the presence of diffusional pores is abolished by the chemical treatment. The permeability coefficient for transport of water across model membrane systems is approximately 10"3 - 10"4 cm s"1 and is in reasonably close agreement for both MLVs and PLMs (Deamer and Bramhall, 1986). The activation energy for transport across EPC MLVs has been reported to be 8.3 kcal mol"1 which is considerably greater than that observed for the simple diffusion of water in an aqueous solution (for a compilation of values see Fettiplace and Haydpn, 1980). Factors which would be expected to increase the hydrocarbon chain order of a membrane (above the T ) such as the presence of cholesterol or increasing the degree of saturation and chain length of the hydrocarbon lead to lower water permeability coefficients for liposomes. Two models have been proposed to account for the observations of water permeability across model membrane systems. The solubility-diffusion model stipulates that water, in a monomer form, dissolves in the hydrocarbon of a bilayer and diffuses across the bilayer (for 21 example see Finkelstein and Cass, 1968). An alternative model, termed the transient defect model (discussed in Deamer and Bramhall, 1986), suggests that water permeates through transient defects in the membrane. Presently, there is no clear cut evidence to favour one model over the other and it has been pointed out that the models do not have to be mutually exclusive (Deamer and Bramhall, 1986). Evidence in support of the solubility-diffusion model comes from several observations. The high activation energy for water transport is consistent with the diffusion of water across a wet hydrocarbon. In addition, the ratio of P^Pj measured across PLMs has been estimated to be close to unity implying the absence of diffusional pores (see for example Fettiplace and Haydon, 1980). Cass and Finkelstein (1967) report that there is a reasonable agreement between the observed P d for water transport across PLMs and the theoretical P d calculated from the partition and diffusion coefficient for water transport in a liquid hydrocarbon. Support for the second model stems from recent observation presented by Carruthers and Melchior (1983). Water permeability for liposomes above T is up to 100 fold greater than the permeability below T c for saturated lipids. In addition, the permeability coefficient above T c for water transport is relatively insensitive to the acyl chain length or saturation. Since the acyl chain length and degree of saturation is expected to markedly vary the water content of lipid bilayers by several orders of magnitude (for a discussion see Deamer and Bramhall, 1986) it follows that these observations are not consistent with the solubility-diffusion model. These observations are best accounted for by a model involving transport through hydrated defects. Nonionic solute transport across model membrane systems has also been examined in detail and is reasonably well understood. There is a strong correlation between the permeability coefficient for transport of a molecule and the hydrophobicity of the molecule, a generalization known as Overton's rule (for example see Walter and Gutnecht, 1986). Permeability coefficients for nonionic solute transport range from 10"6 cm s"1 for urea (Walter and Gutnecht, 1986) to 10"10 cm s"1 for glucose (Bresseleers et al., 1984). The rate limiting step for transbilayer transport is thought to be the partitioning bf the molecule into the hydrocarbon (for a review 22 see Diamond and Wright, 1969; Jain, 1980). For a given series of molecules, factors such as molecular size (Lieb and Stein, 1971; Poznansky et al., 1976) or hydrogen bonding capacity (de Gier et al., 1970; Cohen, 1975; McElhaney, 1986) have been suggested to influence the relative transbilayer transport rate but the importance of this effect is small in comparison to the hydrophobic effect (Orbach and Finkelstein, 1980). In general, factors which increase the order parameter of the hydrocarbon chain will in turn decrease the permeability coefficient for transport of nonionic solutes. 1.7.3 Transport of electrolytes The permeability coefficient measured in SUVs for small monovalent cations such as sodium is approximately 10~14 cms' 1 (Hauser et al., 1973) whereas the permeability coefficient for small monovalent anions such as chloride is 10~10 cm s"1 (Toyoshima and Thompson, 1975). Consequently, the energetic barrier to electrolyte transport is much greater than that for nonionic solutes. An anomalous high permeability coefficient of between 10"4 - 10"6 cm sec"1 is observed for protons (or hydroxide ions) and the reasons for this will be discussed in detail in Section 2.1. In addition, a high A * dependent flux has been reported for several lipophilic amines which are of biological interest (Mayer et al., 1985a; Mayer et al., 1988; Bally et al., 1988) and this behavior is discussed in Chapter 3. However, in general the net flux of electrolytes across model membrane systems is extremely low. This Section summarizes our current understanding of this phenomena. There is an inherent difficulty with measurements of permeability coefficients for electrolyte transport across membranes. Net flux of an electrolyte out of a L U V in response to a chemical potential will result, in the absence of any compensating charge movement, in the establishment of an electrical potential. In turn, the established A * will reduce the net flux of the electrolyte out of the LUV. At equilibrium where the transmembrane chemical potential for the ion is equal to and opposite to the established electrical potential, the net flux of the ion is 23 expected to be zero. Thus, permeability measurements for ions must take into account this potential difficulty. In addition, complications may arise because it is often difficult to differentiate unambiguously between electrogenic and non-electrogenic transport of ions which are acids or bases (for example see Toyoshima and Thompson, 1975; Mayer et al., 1988; Bally et al., 1988). The primary energetic barrier to membrane transport of electrolytes is the Born energy barrier (Parsegian, 1969). The energy required to move a monovalent ion with a 0.2 nm diameter from a medium of dielectric constant of 80 (buffer) to a medium of dielectric constant of 2 (interior of the bilayer) is approximately 40 kcal mol"1. Increasing the radius of the ion or increasing the dielectric constant of the bilayer decreases the energetic barrier to solute transport (Dilger, 1979) which in turn increases the net flux of the ion. Flewelling and Hubbell (1986b) have proposed a model to account for the observed differences in the permeability coefficients for anions and cations. They suggest that the transport of electrolytes across membranes will be influenced by three factors in addition to the Born energy described above (Flewelling and Hubbell, 1986a; Flewelling and Hubbell, 1986b). The image energy is a function of the interaction between the charge of the ion after absorbing to a region of low dielectric and the interface which is a region of high dielectric constant (Neumcke and Lauger, 1969). The dipole energy is a function of an organized array of dipoles in the interior of the bilayer. The physical origin of the dipole is unknown but has been suggested to be the ester bonds linking the acyl chain of the phospholipid to the glycerol backbone (McLaughlin, 1977; Flewelling and Hubbell, 1986b). The dipole term gives the interior of the bilayer an estimated positive potential of several hundred mV. This in turn is expected to account for the relatively high permeability coefficients for anion transport as compared to cation transport (for example see Flewelling and Hubbell, 1986a). Finally a neutral energy term takes into account all the nonelectrical interactions that a electrolyte has with a membrane which include hydrophobic, van der Walls and steric factors. 24 Much of the work on electrolyte transport has examined the relationship between the driving force (either electrical or chemical) and the flux of the electrolyte (for example see Anderson and Fuchs, 1975; Haydon and Hladky, 1972; Chapter 2). For the simple case where the potential energy barrier through the bilayer is a linear function of the position of the ion with respect to the bilayer normal (the Goldman constant field assumption) the flux of the ion is expected to be proportional to the size of the driving force (either chemical or electrical). However, when the energy profile is not a simple linear function of the ions position with respect to the bilayer (see Flewelling and Hubbell, 1986b) then the relationship between the flux of the electrolyte and the driving force will be complicated. For example, Nagle (1986) has proposed that the current- voltage (i-v) graphic relationship for proton transport in response to A * may have three possible shapes which correspond to three different models for proton transport (see Section 2.1). Thus the relationship between the driving force for electrolyte transport (the potential) and the flux of the electrolyte (the current) is not necessarily linear (Haydon and Hladky, 1972; Flewelling and Hubbell, 1986b; Nagle, 1986). The actual mechanism of transport of electrolytes is not known, but flux of electrolytes is discussed in terms of the solubility-diffusion model and the transient defect model mentioned in the previous section (for a review see Deamer and Bramhall, 1986). Presently, it is difficult to choose between these two alternatives. Support for the transient defect model stems from observations that the permeability coefficients in LUVs for sodium ions are approximately three orders of magnitude greater than that expected on the basis of Born energy considerations (Hauser et al., 1973). However, since the energetic barrier to ion translocation is not presently well understood this type of argument best serves as a stimulus to further investigation rather than as a definitive statement. 1.7.4 Lipid factors which influence ion transport The lipid composition of a bilayer will influence the bilayer permeability to solutes. 25 Factors which increase the hydrocarbon chain order (low temperature, decreased saturation, increased chain length and higher cholesterol content) will decrease the transbilayer transport of nonionic solutes and electrolytes (for a review see Jain, 1980; Szoka and Papahadjopoulos, 1980). It is not clear whether all of these factors are operating through the same mechanism or indeed whether the order of the hydrocarbon is in fact the material property which should be related to the permeability of ions. For example, the influence of cholesterol on the material properties of a bilayer has been an area of intense investigation. Cholesterol has been shown to influence the hydrocarbon chain order (Davis, 1983), the gel-liquid crystalline phase transition (see Section 1.5.2), the area occupied by the polar head group (Kawato et al., 1978) and has been shown to decrease the amount and the penetration of water into the hydrocarbon (Simon et al., 1982). Factors such as the degree of saturation and the length of the acyl chains will also influence the hydrocarbon chain order (Davis, 1983) but may also influence the solubility of the solute in the bilayer or the packing geometry of the head-group. Consequently, although the correlation between the hydrocarbon chain order and the permeability properties of bilayers is well documented, the precise relationship is not presently understood. Liposomes have been shown to be particularly permeable at their phase transition temperature (Jacobson and Papahadjopoulos, 1975; Jacobson and Papahadjopoulos, 1976). This observation has been interpreted to be in support of the transport defect model (for a discussion see Deamer and Bramhall, 1986). It has been proposed that regions of gel phase lipid coexist in the same bilayer as regions of liquid-crystalline lipid at or close to T c of a pure lipid. This in turn would lead to a hydrated defect. However, there is presently no physical evidence to support lateral phase separation of this type in liposomes. In addition, it is not clear whether the mechanism responsible for ion permeation at the phase transition temperature is operating in liposomes which do not exhibit a T . There has been less work done on the influence of the head group on solute permeation. It is expected that the presence of acidic lipids (FA, PA, PG, PI, PS) will influence the permeability of liposomes to cations due to the presence of a negative surface charge. This has 26 found to be the case for the influence of PA on monovalent cation permeability (Blok et al., 1975) and for PI and PA on K + permeability (Papahadjopoulos and Watkins, 1967) but not for PS on either K + permeability (Papahadjopoulos and Watkins, 1967) or Na + permeability (Hauser et al., 1973). In addition it is expected that the presence of a negative surface potential on a liposome would be expected to influence the relative rates of transport of weak acids and bases. Since the interfacial pH is expected to be lower for liposomes exhibiting a negative surface potential (see for example Tsui et al., 1986) this suggests that the concentration of the neutral membrane permeable form of weak acids would be increased whereas the membrane permeable form of weak bases would be decreased. In turn this suggests that the presence of a negative surface charge on a liposome would be expected to increase the flux of weak acids and decrease the flux of weak bases. However, in general the influence of the polar head group on the transport of ions across model membrane systems is not as significant as the influence of the hydrocarbon chain order. 1.8 Summary This review has attempted to point out the salient features which give membranes their general permeability characteristics. The permeability coefficient for nonionic solutes is primarily influenced by the hydrophobicity of the molecule but other properties including size and hydrogen bonding capacity may also play a role. Net flux of electrolytes across model membrane systems is typically very slow. The net flux will be influenced by both the presence of transmembrane chemical and electrical potentials. The primary energy barrier to transport of ions is the Born energy; however, other factors including the image energy, the dipole energy and neutral energy terms may play a role in the transport of electrolytes across model membrane systems. Factors which increase the hydrocarbon chain order of a membrane will in turn decrease the flux of nonionic solutes and electrolytes across model membrane systems. The polar 27 head group of lipids does not appear to play a major role in determining the flux of solutes across model membrane systems. Chapter 1 of this thesis will investigate the influence of A * and ApH on the flux of protons across LUVs generated by the L U V E T technique. An understanding of the relationship between A * and ApH is a prerequisite to understanding the mechanism of transport of lipophilic amines which are of biological interest. Chapter 2 develops a suitable test system for examining the relationship between A * and ApH and the transmembrane flux of weak bases. The flux of membrane bound weak acids (phospholipids) is examined in Chapter 4. It is demonstrated that a transmembrane pH gradient will drive the net flux of phosphatidylglycerol across a model membrane system. Finally, Chapter 5 demonstrates that the pH dependent transport of PG across model membrane systems occurs via the neutral protonated form. Factors which influence the hydrocarbon chain order of a bilayer are shown to dramatically influence the rate of transbilayer transport. 28 PROTON TRANSPORT IN LARGE UNILAMELLAR VESICLES IN RESPONSE TO ELECTRICAL POTENTIALS AND pH GRADIENTS. 2.1 Introduction The mechanism by which protons or hydroxide ions cross membranes has received considerable recent interest (for reviews see Deamer, 1987; Gutnecht, 1987a; Nagle, 1987; Perkins and Cafiso, 1987b). Protons or their hydroxide equivalents show anomalous transport behavior. Deamer and co-workers (Nichols and Deamer, 1980; Nichols et al., 1980) were the first to show that the permeability coefficients for protons (or hydroxides) were at least 6 orders of magnitude larger (ca 10"5 cm sec"1) than for monovalent cations such as sodium or potassium (ca 10"12 cm sec"1). This behavior is now established for planar lipid bilayers (PLMs) (Gutnecht, 1984), small unilamellar vesicles (SUVs) (Cafiso and Hubbell, 1983; Perkins and Cafiso, 1986), large unilamellar vesicles (LUVs) made by detergent dialysis, ether injection, reverse phase evaporation (Perkins and Cafiso, 1986) and extrusion (Deamer, 1987). Strikingly, the chemically driven flux of protons (or equivalents) is relatively pH independent (Nichols and Deamer, 1980). For example, the flux of protons driven by chemical (pH) gradients varies by less than one order of magnitude for bacterial phosphatidylethanolamine (PE) PLMs over a proton concentration range of 9 orders of magnitude (i.e. from a pH of 2 to a pH of 11) (Gutnecht, 1987). Thus, the observed proton flux is not consistent with a simple diffusion model which would predict a minimum proton flux at neutral pH. Chloroform (Cafiso and Hubbell, 1983), oxidized lipids, (Cafiso and Hubbell, 1983) and fatty acids (Gutnecht, 1984; Gutnecht, 1987c) are all known to increase the transmembrane proton flux. In addition, proton permeability coefficients are sensitive to the size of the vesicles (smaller for SUVs as compared with LUVs) and the size of the pH gradient (increasing with decreasing ApH) (Perkins and Cafiso, 1986). Two models have been proposed to account for these observations; proton transport is facilitated by anionic carriers (Gutnecht, 1984; Gutnecht, 1987a) or proton transport is 29 facilitated by strands of water which penetrate the bilayer (Nichols and Deamer, 1980; Deamer, 1987; Nagle, 1987; Miller, 1988). These models specify predictions which have been examined. Transport by anionic contaminants of lipid preparations would be expected to act like weak-acid proton uncouplers (for a review of weak acid uncouplers see McLaughlin and Dilger, 1980). Thus, the rate determining step for transport would be the transbilayer movement of the charged anion. Agents which modify the dielectric constant or the dipole potential of the membrane (such as phloretin or chlorodecane) would in turn influence the transbilayer proton flux. In addition, this model predicts that the transport would be relatively pH independent below the p K a of the carrier and saturable. For PLMs, proton transport has been shown to be saturable, inhibited by BSA, decreased by phloretin and increased by chlorodecane (Gutnecht, 1987a). All observations are consistent with the former model of proton transport. However, this model does not adequately account for the proton transport observed at acidic pH values in PLMs and the pH profile for proton transport does not resemble the pH dependence of any known uncouplers (for a discussion see Nagle, 1987). In addition, studies on the chemically induced proton translocation of LUVs indicate that it is insensitive to phloretin (Perkins and Cafiso, 1987a) and BSA (Deamer, 1987). Finally, it is unlikely that the levels of putative uncoupler required to increase the proton permeability coefficient by 5 orders of magnitude are present in all liposome preparations. The second mechanism suggests that protons are transported by transient hydrogen bonded (water) chains (tHBC). Protons would be transported along the water strands in an analogous fashion to that proposed for transport through water or ice-crystals (Nagle, 1987). This qualitatively accounts for the high permeability coefficients observed for protons. Support for this model stem from observations which correlate the increased water permeability (ca 100 fold) at the T of liposomes with the increase in the proton permeability (ca 100 fold) which is higher than that observed for monovalent cations (ca 3 fold) (Elamri and Blume, 1983). However, as of yet there is no direct physical evidence for the existence of structured water chains in membranes (see Conrad and Strauss, 1985) and Gutnecht has been unable to 30 demonstrate a strong correlation between water permeability and proton permeability in PLMs (Gutnecht, 1987c). Nagle has proposed 3 mathematical models to describe proton transport along tHBC (Nagle, 1987). The first model predicts that several protons are transported along a water strand which spans the bilayer. The rate limiting step in this model is the turning defect (a pH independent restructuring of the water molecules). The model is distinguished from the other models discussed because the current-voltage (i-v) curves are strongly sublinear for chemically driven (ApH) transport whereas the i-v curves for electrically driven (A*) transport are either strongly superlinear, weakly superlinear or weakly sublinear. The second model predicts that the rate limiting step is the formation of the tHBCs. This model also has a sublinear i-v curve for ApH driven transport but the i-v curves for A * driven transport are sublinear and always greater than the ApH driven transport. Finally, the third model predicts that the tHBCs only cross one monolayer. The rate limiting step is the collision between chains carrying an excess proton in one monolayer with a chain deficient in a proton in the other monolayer. The model is distinguished in that both the i-v curves for ApH and A * driven transport are strongly superlinear. In this model, the ApH driven transport for a given driving force can be larger, than the A * driven transport (Nagle, 1987). However, as of yet there has been little systematic work done in comparing the shape of the i-v curves for proton transport in response to either A * driven transport or ApH driven transport for a given liposomal system. Krishnamoorthy and Hinkle (1984) noted in asolectin SUVs that the i-v curves for A * driven transport were superlinear and for ApH driven transport were linear. However, the amount of transport for a given driving force was greater for ApH than for A * . This is not consistent with the proposed models discussed above. Other workers (Kell and Morris, 1980; O'Shea et al., 1984) also found that the ApH driven i-v curves were linear but other studies (Deamer, 1987) concludes that the i-v curves are superlinear for ApH driven transport. Thus, there is presently no consensus of opinion concerning the shape of the ApH and A * induced i-v curves for proton tranlocation. 31 A consequence of the high proton permeability of membranes is that one would expect proton electrochemical equilibrium to occur across liposomes after imposing either a pH gradient or an electrical potential. This expectation assumes that A * induced proton flux will be equal to ApH induced proton flux when the size of the driving force is the same. A * driven proton translocation (Cafiso and Hubbell, 1983). and ApH driven proton translocation (Perkins and Cafiso, 1986) resulting in proton electrochemical equilibrium (after 100 min) has been observed in SUVs. However as detailed in an earlier Section, the assumption that proton flux will be the same for A * or ApH may not hold true. This may in part explain why some workers (Nichols and Deamer, 1980; Biegel and Gould, 1981; Elamri and Blume, 1983) have been able to demonstrate ApH driven proton flux whereas other workers (Garcia et al., 1984; Konishi et al., 1986) have been unable to demonstrate the expected A * induced proton transport. In this work, the relationship between electrical potentials and pH gradients in vesicle systems has been examined in detail. LUVs generated by extrusion procedures, (Hope et al., 1985) have been employed since they have the distinct advantage that organic solvents or detergents are not required during vesicle preparation. Particular attention is paid to establishing the reliability of measure of electrical potential and pH gradients provided by the probes triphenylphosphonium (TPP+) and methylamine (MeAm). It is shown by two independent techniques for egg phosphatidylcholine (EPC) L U V systems exhibiting a valinomycin induced electrical potential that electrochemical equilibrium is not achieved over an 8 hr time course. Rather, a small, stable, quasi-equilibrium pH gradient is established and the magnitude of this gradient is sensitive to the ionic composition of the external buffer. The lack of equilibrium is revealed by the addition of the proton ionophore CCCP, which results in larger pH gradients corresponding to electrochemical equilibrium. In addition, an instability in the electrochemical gradients noted in the presence of valinomycin and CCCP together is attributed to a synergistic ability of valinomycin and CCCP to facilitate transport of Na + ions. It is suggested that these observations are consistent with the third mathematical model presented by Nagle which 32 suggests that the rate limiting step for proton transport is the recombination of two tHBCs which each span half the bilayer. 2.2 Materials and Methods 2.2.1 Materials Tetraphenylphosphonium bromide (TPP+), valinomycin, carbonyl cyanide m-chlorophenyl hydrazone (CCCP) and all buffers were purchased from Sigma Chemical Co., St. Louis, MO. All radiochemicals were supplied by New England Nuclear. Chloroform, methanol and acetone were purchased from BDH, Vancouver, B.C. Chloroform and methanol were further distilled before use. 2.2.2 Preparation of egg phosphatidylcholine Egg phosphatidylcholine (EPC) was purified by standard procedures or purchased from Avanti Polar Lipids. Briefly, the EPC was prepared from hen egg yolks by preparative silica chromatography. Typically, 36 egg yolk were mixed with 0.5 L of acetone and the resulting precipitate was isolated by filtration using a porous glass scintered filter. The precipitate was washed 3 X with at least 1 L of acetone (total volume) and then extracted 10 X with 2 L (total volume) of chloroform/methanol (1:1, v/v). The resulting crude EPC was then reduced in volume using a rotorary evaporation device and subsequently purified employing silica acid preparative liquid chromatography (PrepLC/System 500, Waters Ass.; using PrepPak-500/silica) with chloroform/methanol/water (60:30:4, v/v) as the mobile phase (Patel and Sparrow, 1978). The resulting lipid was judged to be greater than 99% pure as indicated by iodine vapor stained thin layer chromatography (TLC). The TLC was run on pre-coated silica gel 60 TLC plates (0.25 mm thick; E. Merck, Darmstadt, Germany) using either a base solvent system 33 (chloroform/methanol/ammonia/water; 90:54:5.7:5.4, v/v) or an acid solvent system (chloroform/methanol/acetic acid/water; 25:15:4:2, v/v). EPC purchased from Avanti Polar Lipids was used without further purification. TLC analysis indicated that the lipid was at least 99% pure. 2.2.3 Vesicle preparation and characterization EPC was suspended in chloroform and mixed with an aliquot of [3H]-DPPC (30,000 dpm//imole phospholipid). The chloroform was then removed under a stream of nitrogen and the lipid was exposed to vacuum for at least 2 hr. Multilamellar vesicles (MLVs) were prepared by vortexing the dry EPC lipid film in the presence of the appropriate buffer (50 mg/ml; w/v) for 15 minutes. The lipid dispersion were then frozen and thawed 5 times to obtain equilibrium transmembrane solute distributions (Mayer et al., 1985c). The resulting frozen and thawed MLVs (FATMLVs) were then repeatedly (lOx) extruded through two (stacked) polycarbonate filters (Nucleopore Corp., Pleasanton, CA) of 100 nm pore size using an extrusion device (Sciema Technical Services Ltd., Richmond, B.C.; Lipex Biomembranes, Vancouver, B.C.). The resulting large unilamellar vesicles (LUVs) exhibit trapped volumes of 1.5 pi per /zmole phospholipid and an average diameter as judged by freeze fracture techniques of 90 nm (Hope et al., 1985). Lipid for the preparation of SUVs was treated as described above. Small unilamellar vesicles (SUVs) were prepared by probe tip sonication at 4°C using a Branson microtip sonicator operated intermittently (5 min sonication; 1 min cooling) for 30 min. The resulting SUVs exhibited a trapped volume of 0.16 pi per pmole phospholipid and a mean diameter of 30 nm as detected by quasi-elastic light scattering using a Nicomp Model 270 submicron particle sizer. Trapped volumes were measured as follows: vesicles were prepared as described above in the appropriate buffer which included an aqueous trap marker such as [14C]-inulin or [ 1 4C]-histidine. The entrapped probe was separated from the unentrapped probe by gel filtration 34 chromatography employing Ultragel (AcA 34, L K B , France) for [14C]-inulin or Sephadex G-50 for [14C]-histidine. Aliquots of the eluant were measured for radioactivity using a Packard scintillation counter and assayed for phosphate as described below. The trapped volume was then expressed as pA of entrapped (associated) probe per pinole phospholipid. Phospholipid concentrations were determined by analysis of lipid phosphorus as described previously (Fiske and Subbarow, 1925; Botcher et al., 1961). Aliquots containing between 0.05 and 0.2 pinole phospholipid were digested in 0.6 ml of 70% H C L O s for at least 1 hr. After cooling, 3.5 ml of ammonium molybdate reagent (0.22%, w/v, ammonium molybdate in 2% H 2 S0 4 , w/v) and 0.3 ml of Fiske-Subbarrow reagent (30 gms NaHS0 3, 1 gm Na 2SO g and 0.5 gms bis l-amino-2-napthol-4-sulphonic acid in 200 ml water at 40°C to aid solubilization). Subsequently, the samples were heated for 10 min at 100°C and the absorbance at 815 nm was determined after cooling. The amount of phospholipid was calculated by comparing the absorbance to a standard curve. Vesicle size was routinely determined by quasi elastic light scattering (QELS) (Mayer et al., 1986b) performed on a Nicomp Model 200 Laser Particle Sizer (Nicomp Instruments, Goleta, CA) with a 5 nW Helium-Neon Laser at an exciting wavelength of 632.8 nm. QELS employs a digital autocorrelation to analyze the fluctuations in scattered light intensity generated by the diffusion of particles in solution. The measured diffusion coefficient is used to obtain the average hydrodynamic radius and hence the mean diameter of the particles. 2.2.4 Uptake of probes into LUVs Transmembrane ion gradients were created by preparing LUVs in the presence of the appropriate buffer and then exchanging the untrapped buffer for the buffer of choice employing Sephadex G-50 gel filtration columns. Experiments to monitor electrical potentials were initiated by adding a small volume of these liposomes (to achieve a final phospholipid concentration of 1 /imole phospholipid per ml) to buffer solutions containing the appropriate 35 membrane potential probes and ionophores. Where employed valinomycin was used at a concentration of 0.5-1.0 n% per /xmole of phospholipid and the ionophore CCCP was used at a concentration of 20 pM. Vesicle associated probe was separated from unassociated probe by gel filtration chromatography. Aliquots of vesicles (0.1 ml) were loaded onto 1 ml syringes which had been previously filled with Sephadex G-50 equilibrated and washed in the appropriate buffer. Immediately after loading, the spin-columns were eluted by centrifugation at 500g for at least 3 min. Quantitation of vesicle associated probe was performed by liquid scintillation counting and phospholipid analysis as previously described. 2.2.5 Nuclear magnetic resonance Vesicles for the 3 1 P NMR experiments were prepared in a similar manner as described above except that the final lipid concentration was approximately 20 mM. 3 1 P NMR studies employed a Bruker WP-200 spectrometer operating at 81 MHz. A free induction decay (256 transients) was obtained using 11 ps, 47° pulse, a 1 s interpulse delay and a 10 KHz sweepwidth. An exponential multiplication corresponding to 5 Hz line broadening was applied to the free induction decay prior to Fourier transformation. The pH of the vesicle interior was determined by monitoring the chemical shift of phosphate (Pi) entrapped inside the vesicle and relating the chemical shift to that obtained for standard Pi solutions of known pH (Mayer et al., 1988). 2.2.6 Calculation of membrane potentials and pH gradients The membrane potentials were calculated assuming that the vesicle-associated probe was associated with the internal trapped volume and did not partition into the vesicle membrane. 36 The electrical potential can then be calculated (Rottenberg, 1979) according to : A * (mV) = - 60 log (C./C 0) . (1) where Ci and C o represents interior and exterior TPP + concentrations. The pH gradients may be measured in a similar manner employing methylamine (MeAm) according to the equation : ApH = log ([HA+]./[HA+]o) (2) where H A + represents the vesicle associated (charged) methylamine (Rottenberg, 1979). For the purpose of discussion it should be noted that the proton motive force can be calculated according to: Ap = A * - 60 x ApH (3) 2.3 Results 2.3.1 Detection of electrical potentials and pH gradients Membrane potentials may be measured across liposome membranes by determining the transmembrane distributions of radiolabelled probes (Rottenberg, i979; Hope et al., 1985), ESR probes (Cafiso and Hubbell, 1978a) or fluorescent probes (Rottenberg, 1979). In the case of the radiolabelled probes employed here, it is important to show that correct measures of the membrane potential are obtained. Previously, it has been shown that the equilibrium distribution of 1 4 C labelled methyltriphenylphosphonium ([ 1 4C]-MTPP+) across liposomal systems provides 37 Fig. 4. A * and ApH response to an imposed transmembrane pH gradient. EPC LUVs (1 mM phospholipid) were prepared in a buffer containing 125 mM Na,S0 4, 10 mM glutamic acid, 10 mM MES, (pH with NaOH to 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, 7.5) and the external buffer was exchanged for 125 mM Na 2SO,, 10 mM glutamic acid, 10>mM MES (pH 7.5). The LUVs were incubated in the presence of 20 pM CCCP and the vesicle associated [ 1 4C]-TPP + or [ 1 4C]-MeAm was determined as described in Section 2.2.4. A * (•) and ApH (•) were calculated from the redistribution of the radiolabeled probes according to the method outlined in Section 2.2.6. Imposed pH Gradient (units) Fig. 5. The correlation between the measurement of pH gradients by [ C]-MeAm redistribution and 3 1 P NMR. EPC LUVs (13 mM phospholipid) were prepared in 200 mM Na 2S0 4, 50 mM Mes (pH with H 2 S0 4 to 5.5, 6.0,6.5, 7.0, 7.5) and the external buffer was exchanged by gel filtration chromatography for 250 mM NaCl, 50 mM HEPES (pH 7.5). The pH gradient was calculated by following the redistribution of [ 1 4C]-MeAm or by following the chemical shift of the entrapped phosphate (see Section 2.2.6 and 2.2.5 respectively). 2.0 MeAm ApH (units) 38 an accurate measurement of the actual transmembrane electrical potential (inside negative) as determined from K + ( 4 2 K + ) transmembrane distributions (Hope et al., 1985). Similar results are also observed for 1 4 C labelled T P P + except that the approach to equilibrium for TPP + is considerably faster than that for MTPP + . The ability of [ 1 4C]-TPP + transmembrane distributions to measure transmembrane electrical gradients accurately was further examined here employing L U V systems with transmembrane pH gradients (inside acidic). As shown in Fig. 4, these systems exhibit near theoretical electrical potentials when incubated in the presence of the proton ionophore CCCP which facilitates electrogenic proton ion movement. It should be noted that for large electrical potentials (A* > 150 mV) the TPP + redistribution slightly underestimates the size of the electrical potential. However, at lower values the TPP + response appears to accurately reflect the transmembrane electrical potential. The pH gradients across liposomal systems can be measured by following the redistribution of weak bases or acids which are either radiolabelled (Rottenberg et al., 1979), spin labelled (Cafiso and Hubbell, 1978b), fluorescent (Deamer, 1982) or by following the 3 1 P chemical shift of entrapped phosphate (Mayer et al., 1988). The results of Fig. 4 demonstrate that redistribution of [14C]-methylamine (MeAm) provides an accurate measure of pH gradients imposed across liposomal membranes. As would be expected, the inclusion of the proton ionophore CCCP results in a reduction of the proton motive force (Ap) to zero as the pH gradient measured by MeAm is equal in magnitude to the electrical potential detected by TPP + . In order to demonstrate further that the redistribution of MeAm accurately reflects the actual pH gradient across the vesicle bilayer the MeAm response to an imposed pH gradient was directly compared to that directly measured by following the chemical shift of the entrapped Pi (Mayer et al., 1988). The results of Fig. 5 clearly indicate a close agreement between the interior pH detected by the 3 1 P NMR technique and the MeAm redistribution. The correlation between these two independent methods provides further evidence that the MeAm response accurately reflects pH gradients between the interior and exterior aqueous media. 39 2.3.2 Electrically driven transport of protons across LUVs The results to this stage show that the measured transmembrane distributions of TPP + and Me Am provide accurate measures of electrical potentials and pH gradients in the L U V systems employed here. The next set of experiments were designed to measure the electrical potential and induced pH gradients obtained when LUVs containing a K + buffer (Na+ exterior buffer) were incubated in the presence of valinomycin. As shown in Fig. 6, the addition of the potassium ionophore leads to the generation of an electrical potential of greater than 160 mV (inside negative). This potential is stable over the 4 hr time course (data not shown). As indicated in the introduction, it would be expected that the presence of an electrical potential would drive proton or proton equivalent transport such that Ap would approach zero during the time course of the experiment. This did not prove to be the case. The pH gradient formed over the time course of the experiment reaches a plateau (Fig. 6) at approximately 1 pH unit (or 60 mV). That this plateau does not represent electrochemical equilibrium is demonstrated by the effects of the addition of the proton ionophore CCCP which causes a marked and rapid acidification of the vesicle interior. It is also clear that the presence of the proton ionophore and valinomycin together causes a destabilization of the electrical potential and the pH gradient such that a time dependent dissipation of A * and ApH occurs after the addition of CCCP. The lack of electrochemical equilibrium between the electrical potential and the induced pH gradient was further.demonstrated by an independent technique as shown in Fig. 7. In this experiment the 3 1 P NMR chemical shift of entrapped phosphate was monitored after the establishment of an electrical potential. The time course for the establishment of the pH gradient is similar to that observed employing Me Am (see Fig. 6). After an 8 hr incubation the pH gradient was only 60 mV whereas the measured electrical potential was still greater than 140 mV. 40 Fig. 6. Electrically driven transport of protons as detected by MeAm redistribution. EPC LUVs were prepared in 169 mM potassium glutamate, 20 mM HEPES (pH 7.5) and the external buffer was exchanged for 150 mM NaCl, 20 mM HEPES (pH 7.5). Transmembrane electrical gradients and pH gradients were measured as outlined in Section 2.2.4 and 2.2.6. ApH (•) and Atf (•) were measured following the addition of valinomycin (0.5 pg/nmol phospholipid). At 38 min CCCP was added to the vesicle systems and ApH (A) and A * (O) were monitored. A * (mV) 1 8 0 -120 ApH (units) 60.0 90.0 Time (min) 120.0 150.0 Fig. 7. Electrically driven transport of protons as detected by following the chemical shift of entrapped Pi. Vesicles were prepared in 100 mM potassium glutamate, 50 mM K 2 H P i (pH 7.5) and subsequently the external buffer was exchanged for 150 mM NaCl, 20 mM HEPES (pH 7.5). The electrical potential (•) was measured employing the radiolabeled probe TPP + (see Section 2.2.4). The pH gradient ( A ) was assayed by monitoring the 3 1 P NMR chemical shift of the entrapped Pi (see Section 2.2.5). Valinomycin (0.5 /jg//jmol phospholipid) was added to the vesicles at time zero. 240 Time (hrs) 41 Cafiso and Hubbell (1983), employing spin labelled probes, have demonstrated that the presence of an electrical potential (inside negative) across small (sonicated) unilamellar vesicles resulted in the establishment of a pH gradient (inside acidic) which reached electrochemical equilibrium with the electrical potential within 100 min. Proton or proton equivalent transport driven by imposed pH gradients was also shown to reach electrochemical equilibrium across SUVs (Cafiso and Hubbell, 1982). This observation was later extended to include other liposomal systems including LUVs prepared by ether injection and reverse phase evaporation (Perkins and Cafiso, 1986). These observations contrast with the results presented in Figs. 6 and 7. Consequently, whether the discrepancy could arise from differences in buffers or liposome preparations was examined. Sonicated SUVs were prepared which exhibited a similar size (28 nm diameter) to those employed by Cafiso and Hubbell (1983). A K + chemical gradient of two orders of magnitude was employed. As shown in Fig. 8, this system exhibited a lack of electrochemical equilibrium which was even more pronounced than observed for the L U V system. An imposed electrical gradient of 118 mV did not drive formation of a bulk pH gradient as detected by methylamine accumulation. This could be attributed to a lower proton permeability in SUVs as compared with LUVs as has been observed by others (Perkins and Cafiso, 1986). The subsequent addition of CCCP led to a rapid and marked acidification of the SUV interior, graphically illustrating the lack of electrochemical equilibrium. It may be noted that the instability of the electrochemical gradient in the presence of both valinomycin and CCCP was not as marked as observed for the LUVs. 2.3.3 The influence of the ionic composition of the external buffer on proton transport The influence of the ionic composition of the exterior buffer on electrical potentials and induced pH gradients is illustrated in Fig. 9. These results demonstrate that the external buffer composition does, in fact, have a marked influence on the induced pH gradient. Vesicles 42 Fig. 8. Time course for the development of pH gradients induced in response to electrical potentials in SUVs. Vesicles were prepared by sonication in 250 mM K 2S0 4, 50 mM NaPi (pH 6.8). The external buffer was subsequently exchanged for 247.5 mM Na 2S0 4, 2.5 mM K 2 S 0 4 (pH 6.8) by gel filtration. The detection of pH gradients by MeAm is outlined in Section 2.2.4. The SUVs (1 mM EPC) were incubated in the presence of (O) 1.0 pg valinomycin or (•) 1.0 valinomycin and 20 /iM CCCP. 2.5 ApH (units) 0.0 30.0 60.0 90.0 120.0 150.0 Time (min) Fig. 9. The influence of the external buffer composition on the pH gradients induced in response to electrical potentials in EPC LUVs. Vesicles were prepared in 100 mM K 2S0 4, 1 mM HEPES (pH 7.0) and the external buffer was exchanged for a buffer containing 20 mM histidine, 10 uM K SO^ (pH 7.0) as well as the following salts: (•) 150 mM NaCl, (•) 150 mM LiCl, (•) 150 mM ChohneCl, ( A ) 200 mM sucrose. The pH gradient was detected by following the redistribution of the radiolabeled probe MeAm as described in Section 2.2.4. 2.0 ApH (units) 120.0 150.0 Time (min) 43 prepared in a .K + buffer with low buffering capacity exhibited quite different induced pH gradients when incubated in a variety of external buffers. As expected, the vesicles exhibited similar electrical potentials regardless of the choice of external buffer (data not shown). However vesicles with external buffer containing NaCl exhibited the lowest induced pH gradient (60 mV). In contrast, vesicles with exterior LiCl, CholineCl, or sucrose exhibit larger pH gradients which approach 120 mV at 2 hr. Addition of the proton ionophore CCCP to these vesicle systems led to an additional increase in the pH gradient (data not shown) indicating that complete electrochemical equilibrium was not achieved. An interesting feature of these systems was the fact that the electrochemical gradient was stable in the presence of both valinomycin and CCCP, in contrast to the instability noted in the presence of external NaCl. The external presence of anions such as sulphate or glutamate for chloride did not influence the magnitude of the induced pH gradient when Na + ions were present (data not shown). 2.3.4 Chemically driven transport of protons across LUVs The lack of electrochemical equilibrium between the electrical potential and induced pH gradient in the presence of external Na + led us to examine the relationship between imposed pH gradients and the induced electrical potentials. Points of interest concern whether electrochemical equilibrium is observed and whether Na+.ions influence such equilibrium. The results of Fig. 10A and 1 OB demonstrate that proton equivalent transport can be followed by monitoring the appearance of an electrical potential (inside negative) in an L U V after establishing a pH gradient (inside acidic). In Fig. 10A the proton equivalent transport in the absence of Na + ions is presented. As would be expected, the proton chemical potential across the membrane drives electrogenic proton transport resulting in electrical potentials which approach equilibrium values. The generation of the electrical potential is time dependent and complete electrochemical equilibrium is not reached even at 2 hr. This is in contrast to results 44 Fig. 10. Time course for the development of an electrical potential (inside negative) generated in response to a transmembrane pH gradient in EPC LUVs. Vesicles were prepared in 50 mM citric acid (pH with histidine to 4.0) and the external buffer was exchanged for a buffer composed of (A) 200 mM sucrose, 20 mM histidine (pH 7.0) or (B) 150 mM NaCl, 20 mM histidine (pH 7.0). The electrical potential was detected by following the redistribution of the radiolabelled probe TPP + as described in Section 2.2.4. The LUVs (1 mM EPC) were incubated in the presence of (#) no ionophores, (A) 1.0 /ig valinomycin, (•) 20 / J M CCCP or (•) 1.0 u-% valinomycin and 20 uM CCCP. 2 0 0 A * (mV) 0.0 30.0 60.0 90.0 Time (min) 120.0 1 5 0 . 0 45 obtained for vesicles prepared by reverse phase techniques which indicated electrochemical equilibrium at 20 min (Perkins and Cafiso, 1986). Fig. 10A also indicates the effect of ionophores on proton equivalent transport. The addition of valinomycin increases proton transport significantly as has also been observed elsewhere (Rossignol et al., 1982). The addition of the proton ionophore CCCP leads to a stable electrochemical equilibrium. The presence of both CCCP and valinomycin does not affect this stability. The induced electrical potentials observed when Na + ions are present in the external medium are shown in Fig. 10b. Given the previous results indicating that Na + reduces the induced ApH observed in response to A * , it is perhaps surprising that the presence of Na + ions does not affect the rate of proton equivalent transport as detected by the induced electrical potential. Similar results were also obtained when Na + ions were present in the interior buffer (data not shown). Again, complete electrochemical equilibrium is not reached in this system at 2 hr. Addition of valinomycin causes a slight increase in the transport of proton equivalents but there is no evidence that components of the buffer are transported in a manner which leads to decay of the electrochemical gradients. Addition of the proton ionophore CCCP results in the generation of a stable electrochemical equilibrium. The addition of both valinomycin and CCCP also leads to electrochemical equilibrium, however, a slow time dependent dissipation of the electrochemical gradient is subsequently observed. This dissipation is not seen in Fig. 10A indicating that it is caused by the NaCl buffer. 2.3.5 Ionophore mediated transport of Na+ Two observations made in the course of this work indicate that, in the presence of exterior Na + , the presence of both CCCP and valinomycin causes dissipation of the electrochemical gradients. In the case of valinomycin induced electrical potentials observed in LUVs with interior K + and exterior Na + buffers, the addition of CCCP results not only in electrochemical equilibrium but also in a rapid subsequent decrease of A * and ApH. This 46 Fig. 11. Time course for the development of an electrical potential (inside negative) or a pH gradient (inside acidic) for LUVs experiencing a Na + diffusion potential in the presence of a variety of ionophores. Vesicles were prepared in 100 mM Na 2S0 4, 1 mM HEPES (pH 7.0) and the external buffer was exchanged for a buffer composed of 200 mM sucrose, 20 mM histidine, (pH 7.0). The electrical potential (A) and pH gradient (B) were detected as outlined in Section 2.2.4. The LUVs (1 mM EPC) were incubated in the presence of (o) no ionophores, (•) 1.0 y.g valinomycin, (A) 20 /iM CCCP or (•) 1.0 /xg valinomycin and 20 /*M CCCP. 150 47 behavior is not observed in the presence of exterior ions such as choline or L i + . Alternatively, in the systems where an imposed ApH leads to an induced A * , the presence of valinomycin and CCCP also results in a rapid decrease in ApH and A * when Na + is present, but not when other external buffers are employed (Fig. 10). These observations suggest that CCCP and valinomycin may act synergistically to transport Na + ions, thus dissipating electrochemical gradients. In order to test this possibility, the ability of these ionophores to induce electrical potentials in systems exhibiting Na + chemical gradients was investigated. LUVs were prepared in the presence of a Na + containing buffer and the external buffer was then exchanged for a sucrose containing buffer. The flux of Na + can be followed by measuring the generation of an electrical potential or pH gradient (as detected by TPP + or MeAm respectively). As would be expected, in the absence of any ionophores there is no net movement of Na + and consequently no net generation of an electrical potential. As shown in Fig. 11, the presence of valinomycin or CCCP alone does not increase the flux of Na + ions as detected by the establishment of a membrane potential. However, the presence of both valinomycin and CCCP causes the establishment of an electrical potential (inside negative) and pH gradient (inside acidic). This experiment demonstrates that valinomycin and CCCP can act synergistically to transport N a + ions. It is not clear whether this is electrogenic or electroneutral transport. V 2.4 Discussion The results presented in this work provide new information regarding the transport of protons or proton equivalents across membranes. There are three major points of interest. The first concerns the accuracy of the electrical potentials and pH gradients reported by the radiolabeled probes triphenylphosphonium (TPP+) and methylamine (MeAm). The second point concerns the lack of electrochemical equilibrium between A * and the induced ApH, particularly in the presence of external Na + . Finally, the observations that CCCP and valinomycin together can facilitate N a + transport are discussed. 48 There are three lines of evidence supporting the reliability of the electrical potential and pH gradients reported by TPP + and MeAm. First, the electrical potential reported by TPP + for L U V systems exhibiting imposed pH gradients (in the presence of CCCP) correlates well to the theoretical values expected for electrochemical equilibrium. This also supports previous work (Hope et al., 1985) on valinomycin induced electrical potentials in L U V systems containing K + ions where it was found that the electrical potential reported by M T P P + correlated well with theoretical values calculated from the K + ( 4 2 K + ) transmembrane concentration gradients. Second, in the case of MeAm, the pH gradient reported correlates well with imposed transmembrane pH gradients. Further, the pH gradients measured by MeAm agree closely with those measured by monitoring the 3 1 P NMR chemical shift of entrapped Pi. A final point is that the electrochemical gradient reported by MeAm and TPP* are internally self consistent. Thus in systems where electrochemical equilibrium is expected, the pH gradients and electrical potentials obtained are consistent with a proton motive force of zero. This is observed in LUVs with imposed pH gradient in the presence of CCCP (Fig. 10A) and in LUVs with K + ion gradients in the presence of valinomycin and CCCP (Fig. 6).. The observed accuracy of the electrochemical gradients reported by TPP + and MeAm would not be expected a priori for the L U V systems employed here. This is mainly due to their small size and the lipophilic character of the probe molecules. For example, LUVs would be expected to have a ratio of internal aqueous volume to lipid volume of 2.3, assuming a bilayer thickness of 5 nm and a diameter of 90 nm. Given the lipophilic character of membrane potential and pH probes (Demura et al., 1985a; Demura et al., 1985b; Demura et al., 1987; Lee and Forte, 1980; Cools and Janssen, 1986), a large proportion of the LUV-associated probes would be expected to be associated with the bilayer rather than contained in the interior aqueous compartment. In the case of the membrane potential probes this would be expected to lead to overestimation of the actual electrical potential. As detailed here, such deviations are not observed. The reasons for this are not understood, but may be related to a greater ability to 49 remove the lipid associated as opposed to the entrapped probe during the chromatographic procedure used here to separate.LUVs from untrapped material. The lack of electrochemical equilibrium between membrane potentials developed by valinomycin in LUVs containing a K + buffer and the induced pH gradients is now discuussed. As detailed here, this effect is particularly obvious when the external buffer contains Na + and the lack of equilibrium is unambiguously demonstrated by the much larger pH gradients induced in the presence of CCCP. Particular points of interest concern the discrepancy between these'results and previous studies by other workers, the possible mechanisms which could result in such a lack of equilibrium and the role of nominally non-permeating ions in influencing the magnitude of induced pH gradients. Previous work employing SUVs (Cafiso and Hubbell, 1983) and spin labelled probes for the detection of electrical potentials and pH gradients indicated that pH gradients induced in response to membrane potentials come to electrochemical equilibrium within 100 min. In our hands, SUVs with a similar buffer composition did not exhibit a detectable induced ApH after 2 hr. This discrepancy cannot be attributed to the small size of the vesicles or the buffer composition. Similarly, Perkins and Cafiso (1986) have shown for LUVs produced by the reverse phase technique (Szoka and Papahadjopoulos, 1978) that an electrical potential corresponding to electrochemical equilibrium is observed within 20 min after establishing a transmembrane pH gradient. This contrasts with the observations presented here where complete equilibrium is not reached even after 2 hr. Possible reasons for these.discrepancies could include an influence, of the membrane associated spin labelled probes on proton or proton equivalent permeability or some small differences in the properties of the lipid or composition of the bilayer. The LUVs produced by extrusion have the important advantage that organic solvents or detergents are not employed for their synthesis. The possibility that residual organic solvents or lipid oxidation products are present which increase proton permeability (Cafiso and Hubbell, 1982; Gutknecht, 1987) cannot be easily dismissed. It should also be noted that other workers (Garcia et al., 1984; Konishi et al., 1986) noted a lack of electrically driven proton transport 50 which they attributed to a threshold effect where an electrical potential of at least 60 mV is required to drive proton or proton equivalent transport. The physical basis for an inequivalence between electrical potentials and induced transmembrane pH gradient is clearly of major interest. The small quasi-equilibrium induced pH gradient of 1 pH unit (60 mV) or less induced in response to an electrical potential of 150 mV or more suggests an inequivalence between proton fluxes induced in response to proton gradients and those induced in response to electrical potentials. In particular these results suggest that a lower proton flux is obtained in response to electrical potentials than is obtained in response to chemical potentials of equal magnitude. Such behavior has been reported previously. For example, Krishnamoorthy and Hinkle (1984) observed a proton equivalent flux of 2.5 nmol/min/mg phospholipid for a chemical driving force of 60 mV in asolectin vesicles. The corresponding flux for a 60 mV electrical potential was only 0.5 nmol/min/mg phospholipid, whereas the flux for a 160 mV electrical driving force was 2.2 nmol/min/mg phospholipid. This would suggest that a pseudo equilibrium condition could be achieved in these vesicles when an electrical potential of 160 mV was offset by a pH gradient of only 60 mV. There is some theoretical basis for the suggestion that proton or proton equivalent flux in response to chemical gradients can be larger than for an electrical potential. In particular, among the models for transmembrane proton transport by transient hydrogen bonded water (Nagle, 1987) such observations could be consistent with a model where proton equivalent conduction is mediated by a transient hydrogen bonded chain (tHBC) which stretches across one phospholipid monolayer. This model, originally suggested by Deamer and Nichols (1983) appears to be the only model analyzed by Nagle (1987) that could account for the observed quasi-equilibrium condition. Other models suggest that proton equivalent transport in response to electrical potentials is equal to or greater than proton equivalent transport in response to chemical gradients. However, this model does not provide a clear understanding of the smaller induced pH gradients observed when Na + ions are present in the external buffer. In this model the rate 51 limiting step for proton translocation in response to electrical potentials is the recombination of tHBC, and it is difficult to see how the presence of Na + ions should slow down this recombination. One possibility is that the presence of Na + ions in the external buffer facilitate the movement of OH" ions in response to the chemical driving force. However, one would then expect to see an effect of Na + ions on the stability of a pH gradient, which was not observed (see Fig. 10b.) This leaves open the possibility that Na + ions can alter the concentration of tHBC, or that Na + ions exert their effect on OH" ion movement in response to electrical potentials only at neutral or basic pH values. The final topic of discussion involves the striking observation that valinomycin and CCCP can catalyze the transport of Na + ions across a liposome in a synergistic manner. Ternary complexes have been described for a variety of uncouplers bf oxidative phosphorylation (Yoshikawa and Terada, 1981; Blok et.al., 1975; Castaing et al., 1986). Blok and co-workers first suggested that valinomycin can form a ternary complex with K + and SCN" anion (Blok et al., 1975). Since then Yoshikawa and Terada have suggested that valinomycin and K + can form ternary complexes (1:1:1 ratio) with a variety of weak acid uncouplers such as FCCP and 2,4-dinitrophenol (Yoshikawa and Terada, 1981). These complexes will accelerate the transport of both protons and K + ions in a synergistic manner. Finally, Castaing and co-workers (Castaing et al., 1986) have demonstrated that FCCP can accelerate the transport of both K + and Na + ions across LUVs by a macropolycyclic complexing agent (222)C10-cryptand. These observations are consistent with the possibility that CCCP, valinomycin and Na + ions can form a ternary complex which can synergistically promote net Na + transport. In summary, the results presented here show that radiolabelled probes of membrane potential (TPP+) and pH gradient (MeAm) provide surprisingly accurate measures of electrical potentials and pH gradients in L U V systems. These probes have been employed, to demonstrate that, particularly in the presence of external Na + ions, the induced pH gradients observed in response to K + diffusion potentials are appreciably smaller than would be expected on the basis of electrochemical equilibrium. It is suggested that this reflects a higher proton or proton 52 equivalent flux in response to proton chemical gradients than for nominally equivalent electrical potentials. Finally, it is demonstrated that the K + ionophore valinomycin and the proton ionophore CGCP can transport Na + ions in a synergistic manner. It is suggested that valinomycin, CCCP and Na + ions can form a ternary, membrane permeable complex. 53 STUDIES ON THE UPTAKE OF LIPOPHILIC AMINES INTO MODEL MEMBRANE SYSTEMS IN RESPONSE TO ION GRADIENTS 3.1 Introduction The ability of ammonia to cross biological membranes is well documented (Croft, 1967; Rottenberg and Grunwald, 1972b). This molecule is generally accepted to cross membranes in a neutral form leading to transmembrane concentration gradients which reflect the transmembrane proton gradients. Indeed most alkylamines are thought to cross membranes in an analogous manner (Gaenssien and McCarty, 1971; Bar-On and Degani, 1985; Ritchie and Gibson, 1987; Hope and Cullis, 1987). This has led to the development of a variety of techniques for the measurement of pH gradients across membranes which include monitoring the distribution of the (radiolabelled) probe methylamine (see Chapter 2; Rottenberg, 1979), the fluorescent probe 9-aminoacridine (Deamer et al., 1972; Fioler et al., 1974; Schuldiner et al., 1978; Nichols et al., 1980) and the ESR probe N-tempoyl-n-hexyl amine (Cafiso and Hubbell, 1978). All of these techniques are based on the widely accepted ability of these amines to cross membranes in a neutral form such that at equilibrium their distribution across the membrane will reflect the transmembrane pH gradient. The behavior of simple amines contrasts with the behavior of probes of electrical potentials which can cross membranes in a charged form. Electrical potential probes such as safranine o (Bally et al., 1985), tetraphenylphosphonium (see Chapter 2) and diS-C 3 (Konishi et al., 1986) share similar properties with the simple amines listed above in that they are both lipophilic and charged at physiological pH. However, unlike the amine probes they can respond to electrical potentials. This different behavior is attributed to charge derealization on the electrical potential probes which results in a more lipid soluble molecule (McLaughlin and Dilger, 1980; Flewelling and Hubbell, 1986b). 54 Table 2. The uptake of a variety of lipophilic cations in response to either transbilayer electrical potentials or pH gradients. Electrical potentials (A* > 100 mV) were initiated by adding valinomycin to LUVs exhibiting a potassium/sodium diffusion potential. Transmembrane pH gradients (ApH > 2 units) were established by entrapping a low pH buffer and subsequently exchanging the external buffer by gel filtration chromatography for a high pH buffer. The % uptake represents the maximum amount of drug associated with the vesicles which were incubated at a lipid:drug ratio of 10:1 (mol/mol). % % Lipophilic Cation Uptake Ref.1 Uptake Ref.1 A * ApH Dibucaine 50 1 >95 6 Doxorubicin >95 2 >95 6 Chlorpromazine . 35 3 50 6 Timolol 12 5 50 6 Propranolol 30 5 >95 6 Dopamine 50 . 4 90 6 Serotonin 20 4 40. 6 Epinephrine 20 / 4 20 5 Methylamine <1 5 50 5 1. The reference numbers correspond to: 1 Mayer et al., 1985a 2 Mayer et al., 1985b 3 Bally et al, 1985 4 Bally et al., 1988 5 Redelmeier, unpublished observations 6 Madden et al., unpublished observations 55 A large variety of pharmaceuticals are lipophilic amines with basic pK as. Recent work in this laboratory has been aimed at characterizing uptake of such agents into liposomal systems exhibiting transbilayer electrical potentials (inside negative) or transbilayer pH gradients (inside acidic). The data summarized in Table 2 indicate that an imposed electrical potential or an imposed pH gradient results in the association of the local anaesthetic dibucaine (Mayer et al., 1985a; Mayer et al., 1988), the antischizophrenic agent chlorpromazine (Bally et al., 1985), the antineoplastic agent doxorubicin (Mayer et al., 1985b; Mayer et al., 1986a), the biological amines epinephrine and dopamine (Bally et al., 1988; see also Nichols and Deamer, 1976) and the adrenergic y9-blockers propranalol and timolol (unpublished observations). Indeed, most of the pharmaceuticals which are simple amines respond in this manner. The ability of amines to respond to both electrical potentials and pH gradients suggests a mechanism where amines can cross membranes in both a charged and uncharged form (Mayer et al., 1988; Bally et al., 1988). However, the presence of acidic pH gradients in vesicles which exhibit electrical potentials (see Chapter 2) makes it difficult to demonstrate unambiguously that transport has occurred via the charged form. In addition, significant differences exist in the ability of amines to be accumulated into LUVs in response to electrical potentials and pH gradients and differences exist in the extent of ApH driven accumulation of the various amines. Thus, the mechanism of transport of lipophilic amines across LUVs in addition to the nature of the association of the amines with the LUVs is not presently understood. Four models for the transport of lipophilic amines across membranes have been proposed and these are illustrated in Fig. 12. Fig. 12a embodies the proposal that the neutral amine is the permeating species. This is consistent with the widely accepted understanding of the transport of pH probes (for example, see Deamer et al., 1972; Cafiso and Hubbell, 1978). There have been fewer studies on the mechanism of transport of lipophilic amines which are of biological or pharmaceutical interest. Gutnecht and Walter (1981b) examined the transport of several biological amines across PLMs and observed only significant transport of the uncharged amine. They note however, that the flux of the amine across the unstirred layer (USL) of the PLMs 56 Fig. 12. Four putative mechanisms for the transport of amines across membranes. HA represents the neutral form of the amine, H 2 A + represents the charged form and H + represents free protons. Mechanism A suggests that lipophilic amines cross membranes in a neutral form. Mechanism B, C and D suggests that transport occurs via the charged ion. The charged ion in B is transported by itself, in C, transport occurs via an anion pair (An") and D, transport occurs via a complex between the charged amine and a neutral amine. A M E M B R A N E IN O U T H 2 A + HA + H + B M E M B R A N E IN O U T H + + HA H 2 A H 2 A + H 2 A + IN H 2 A + + A n " H 2 A + A n " O U T H 2 A + A n " H 2 A + + A n " IN H A H 2 A + H 2 A + H A O U T H 2 A + H A H 0 A H A 57 may be rate limiting at high pH values. In turn, this suggests a role for the charged amine in transport across the USL. Further evidence in support of transport of amines via the uncharged molecule has been provided by Madden and co-workers (1988) who examined the influence of electrical potentials (inside negative) on the transport of several lipophilic cations (dopamine, doxorubicin and propranalol) in cytochrome-oxidase containing proteoliposomes. The second model to account for transport of lipophilic amines across membranes (Fig. 12b) suggests that transbilayer transport is mediated by the charged ionic form. Observations in support of this model come from studies on the influence of electrical potentials on the transport of lipophilic amines in LUVs (Mayer et al., 1985a; Mayer et al., 1986a; Mayer et al., 1988; Bally et al., 1988). First, the extent of uptake of the local anaesthetic dibucaine and the biological amine dopamine into LUVs exhibiting large (greater than 100 mV) negative electrical potentials (created by K + diffusion potentials) but relatively small induced pH gradients (1 unit) was much greater than vesicles where only a pH gradients Of 1 unit had been imposed (Mayer et al., 1988; Bally et al., 1988). This suggests a role for the,electrical potential in determining the uptake of lipophilic amines into these vesicles. Second, if amine uptake depends upon a pH gradient formed in response to an electrical potential, then an increase in the buffering capacity of the vesicle interior (which would decrease the rate of formation of the pH gradient) should slow down the rate of uptake of the lipophilic amine. However, the extent of uptake of both dopamine and dibucaine into vesicle systems is insensitive to the buffering capacity of the vesicles (Mayer et al., 1988: Bally et al., 1988). Finally, Mayer and co-workers (1988) have shown that the quaternary amine derivative of dibucaine, which would only be able to respond to an electrical potential and not a pH gradient, can be accumulated in response to an electrical potential. Two other possible mechanisms illustrated in Fig. 12c and 12d are variations of the second mechanism. Fig. 12c indicates a mechanism originally suggested by Garlid and Nakashima (1983), where the permeating species is a neutral complex of the protonated amine and a cotransported anion. Support for this model stem from the observations that the 58 uncoupling ability of local anaesthetics requires the presence of the salt of a strong acid. Finally, the mechanism of Fig. 12d (McLaughlin, 1975) suggests that the permeating species is a complex between a neutral amine and a charged amine. This accounts for the observation that the conductance of PLMs varies with the square of tetracaine concentration and is maximal close to the pK of the amine. The mechanisms illustrated in Fig. 12 are not mutually exclusive. If amines are transported in both the ionic and nonionic form, then one would expect that the lipophilic amine would act as a proton ionophore (for a review on uncouplers see Dilger and McLaughlin, 1980). However, since neutral molecules which are general anaesthetics are also found to act as uncouplers (Rottenberg, 1983; Barchfeld and Deamer, 1985; Bangham and Hill, 1985) it is important to be able to distinguish between the two independent models for uncoupling. The former model stipulates that amines can be transported in a charged and uncharged form (see Fig. 12) whereas the latter model stipulates that anaesthetics can act to dissipate ion gradients through a generalized increase in ion permeability (Bangham and Hill, 1985). In practice, it is difficult to distinguish between these two mechanisms of uncoupling. This may, in part, explain the controversy over why some workers (Azzi and Scarpa, 1967; Jung and Brierley, 1979; Dabadie et al., 1987) have been unable to demonstrate that amines (which are local anaesthetics) can act to uncouple state 4 mitochondria whereas other workers find that local anaesthetics do uncouple mitochondria (Massari and Pozzan, 1976; Garlid and Nakashima, 1983; Dabadie et al., 1987). This work attempts to resolve further the question of whether simple amines can cross membranes in both a charged and an uncharged form. A model system exhibiting a stable electrical potential and no measurable pH gradient is developed, characterized and employed to investigate the uptake of a variety of lipophilic amines, with the objective of determining the uptake mechanism. It is shown that the rapid uptake of the lipophilic amines (timolol, dibucaine, chlorpromazine and doxorubicin) is correlated with the presence of large transmembrane pH gradients, consistent with the transport of the neutral form of these drugs. 59 Significant uptake of the lipophilic amines studied here is observed in the presence of an . electrical potential and the absence of any measurable pH gradient. The possibility that this uptake represents charged transport is discussed in terms of the observation that the lipophilic amines do not act as uncouplers under these conditions. It is suggested that these observations can be accounted for by a model where only the neutral form of the amine is transported across the bilayer at a significant rate under physiological conditions. 3.2 Materials and Methods 3.2.1 Materials Timolol maleate, doxorubicin hydrochloride, chlorpromazine hydrochloride, dibucaine hydrochloride and all probes, buffers and ionophores were purchased from Sigma Chemical Company, St. Louis, MO. [3H]-dipalmitoyl phosphatidylcholine (DPPC), [ 1 4C]-tetraphenylphosphonium ion (TPP+) and [3H]-chlorpromazine were all purchased from New England Nuclear. [14C]-timolol was a generous gift of The Liposome Co., Princeton, New Jersey. Egg phosphatidylcholine (EPC) was purified as described in Section 2.2.2. 3.2.2 Preparation of LUVs Vesicles were prepared and characterized as previously described in Section 2.2.3. 3.2.3 Uptake of probes and drugs into vesicles Transmembrane K + ion gradients were created by preparing LUVs in the presence of 100 mM K 2 S 0 4 , 1 mM HEPES (pH 7.0) (unless otherwise stated in the Figure legends). These LUVs were then passed over a Sephadex G-50 gel filtration column equilibrated with the 60 appropriate buffer. Experiments to monitor the uptake of drugs or membrane potential probes were initiated by adding a small volume of concentrated vesicles (to achieve a final phospholipid concentration of 1 mM) to buffer solutions containing the appropriate drugs, membrane potential probes and ionophores. Unless otherwise stated in the Figure legends, the drug concentrations used were 100 pM. Quantification of vesicle associated drug and probe following appropriate incubation periods was performed by passing 0.1 ml aliquots of the sample over 1.0 ml Sephadex G-50 mini columns to separate free from vesicle associated probe (Section 2.2.4). The quantification of vesicle associated radiolabelled chlorpromazine and timolol was performed in a Packard 2000CA scintillation counter. The quantification of vesicle associated doxorubicin and dibucaine was performed by suspending an aliquot in 0.5% Triton X-100 and measuring the fluorescence employing a Perkin-Elmer spectrofluorometer (Mayer et al., 1985a; Mayer et al., 1986a). The excitation wavelength for doxorubicin was 490 rim and the emission was recorded at a wavelength of 590 nm whereas the excitation wavelength for dibucaine was 330 nm and the emission was recorded at a wavelength of 416 nm. Lipid phosphate was determined as outlined in Section 2.2.3. The electrical potential and pH gradients were calculated as described in Section 2.2.6. 3.3 Results 3.3.1 Generation of a model membrane system exhibiting a stable A * and no ApH In order to test whether one or more of the mechanisms illustrated in Fig. 12 are involved in the accumulation of lipophilic amines in response to ion gradients, appropriate test systems are required. Ideally, such systems would display a A * and no ApH, or a ApH and no A * . Presently available systems do not display these characteristics. In particular, the results of Chapter 2 demonstrate that for LUVs with K + / N a + transmembrane gradients of 4 orders of magnitude ( K + inside), the addition of valinomycin results in a large electrical gradient (A* > 61 150 mV) but only a small induced ApH (ApH < 1 unit). Alternatively, imposed pH gradients generate a time dependent formation of an electrical potential. Thus, all of these systems exhibit an appreciable A * and ApH, which complicates investigations of the mechanism of uptake of lipophilic amines. A fundamental complication of the. above systems concerns the variety of ion species present, which include Na + , K + and CI". For example, the presence of Na + appears to alter the extent of proton transport (Chapter 2) whereas chloride would be expected to be membrane permeable under these conditions (Toyoshima and Thompson, 1975). A simple buffer system was therefore adopted, consisting of 100 mM K 2 S 0 4 , 1 mM HEPES (pH 7.0) in the interior buffer and 200 mM sucrose, 10 /xM K 2 S 0 4 , 20 mM histidine (pH 7.0) in the exterior buffer. As indicated in Fig. 13, these systems exhibit different ionophore induced A * and ApH responses than observed for vesicles experiencing K + / N a + gradients (Chapter 2). First, as illustrated in Fig. 13A, the addition of the potassium ionophore valinomycin to this system leads to the generation of a stable electrical potential Of 150 mV, and an induced ApH gradient of 1.6 units (100 mV). As has been shown before, this is appreciably larger than the ApH observed when Na + is in the exterior medium. The results of Fig. 13B illustrate that this system is not at electrochemical equilibrium as the presence of the proton ionophore CCCP in addition to the potassium ionophore valinomycin leads to the generation of a larger (approximately equilibrium) pH gradient of 2.5 units (145 mV). In contrast to the situation where Na + ions are present in the external buffer (Chapter 2), here it is demonstrated that the electrochemical gradient is stable over the time course of the 2 hr experiment. This is consistent with the suggestion made in Chapter 2 that valinomycin and CCCP dissipate the electrical potential by increasing Na + permeability. The results of Fig. 13C demonstrate that the presence of nigericin, an ionophore which catalyzes the exchange of potassium ions for protons, leads to the rapid establishment of a pH gradient of 2.5 units. Surprisingly, this ionophore does not create a pH gradient when Na + ions are present in the external buffer (data not shown) indicating that this ionophore will catalyze significant potassium-sodium exchange. The slower evolution of A * to achieve 62 Fig. 13. Time course of the establishment of a measured electrical potential (•) and pH gradient (•) in EPC LUVs exhibiting potassium diffusion potentials. EPC LUVs were prepared according to established procedures (see Section 3.2.2) in 100 mM K 2 S 0 4 , 1 mM HEPES (pH 7.0) and the untrapped buffer exchanged for 200 mM sucrose, 10 pM K 2 S 0 4 > 2 0 m M histidine (pH 7.0). The experiment was carried out at 2 mM EPC, 10 uM TPP + , 10 uM MeAm and (A) 0.5 /ig valinomycin per pmol phospholipid; (B) 0.5 /ig valinomycin per pmol phospholipid and 20 /iM CCCP; (C) 1 /iM nigericin. 180-(mV) Time (min) 63 Fig. 14. Time course of the establishment of a measured electrical potential (•) and pH gradient (•) in EPC LUVs when TRIS is present in the external buffer. EPC LUVs were prepared according to established procedures (see Section 3.2.2) in 100 mM K 2 S 0 4 , 1 mM HEPES (pH 7.0) and the untrapped buffer exchanged for 200 mM sucrose, 10 /iM K 2 S 0 4 , 20 mM HEPES / TRIS (pH 7.0). The experiment was carried out at 2 mM EPC, 10 uM TPP + , 10 uM MeAm and (A) 0.5 /ig valinomycin per nmol phospholipid; (B) 0.5 /ig valinomycin per /umol phospholipid and 20 /iM CCCP;. (C) 1 uM nigericin. A * (mV) -- 3 ApH (units) -- 1 ApH 2 (units) -- 3 - 2 -- 1 150 ApH (units) Time (min) 64. electrochemical equilibrium in the system which exhibits a pH gradient may be attributed to the electrogenic transport of K + ions either by nigericin or to the transport of proton equivalents in response to a chemical gradient. In summary, the results of Fig. 13 demonstrate that ionophores of various types can be employed to establish electrical potentials and pH gradients of varying magnitude. However, due to the high intrinsic proton permeability of lipid bilayers, no single system exhibits a large A * without a pH gradient or vice-versa. The results illustrated in Fig. 14 demonstrate the development of a model membrane system which exhibits a stable A * and no observed ApH. The presence of TRIS (ca 10 mM) in the external buffer of LUVs exhibiting a K + diffusion potential is shown to dissipate effectively transmembrane pH gradients. In the presence of valinomycin (Fig. 14A) a stable A * of 160 mV (inside negative) is established which compares favorably with that observed in Fig. 13A. However, no A * induced pH gradient is detected by MeAm redistribution. The presence of TRIS, a weak base, in the external buffer of LUVs exhibiting a stable A * dissipates the induced pH gradient. This observation has been previously documented by Sone and co-workers (1976). They propose a model whereby TRIS is transported across the membrane in a neutral form and dissipates transmembrane pH gradients upon protonation. The stability of the membrane potential in this system is expected to depend upon the net inward flux of protons (or equivalents) since it is expected that potassium release in this system will depend upon the transmembrane movement of a counter ion. The results of Fig. 14B are consistent with this hypothesis indicating that the presence of the proton ionophore CCCP (as well as valinomycin) causes a time dependent dissipation of the electrical potential. The presence of the proton ionophore also leads to electrochemical equilibrium for all time points (see also Chapter 2). Thus, the system demonstrated in Fig. 14A is a particular useful tool for determining whether agents can act as proton ionophores since an increase in the net flux of protons will lead to a dissipation of the electrical potential. A further test of the ability of TRIS to dissipate pH gradients was conducted to determine whether the presence of specific ionophores was required. The results of Fig. 14C 65 Fig . 15. Time course of the establishment of a measured electrical potential (•) and pH gradient (•) in E P C L U V s when T P P + is present in the external buffer. E P C L U V s were prepared according to established procedures (see Section 3.2.2) in 100 m M K 2 S 0 4 , 1 m M H E P E S (pH 7.0) and the untrapped buffer exchanged for 200 m M sucrose, 10 pM K 2 S 0 4 , 20 m M H E P E S (pH 7.0). The experiment was carried out at 2 m M E P C , 5 m M T P P + , 10 pM M e A m and (a) 0.5 /ig valinomycin per /xmol phospholipid; (b) 0.5 ug valinomycin per pmol phospholipid and 20 pM C C C P ; (c) 1 / J M nigericin. A * (mV) 180 120 A p H (units) -- 1 A * (mV) 1B0--1 2 0 -60 B - 3 -H-- f -A p H (units) -- 1 A * (mV) 1 8 0 -120-A p H (units) 120 150 Time (min) 66 demonstrate that a potassium diffusion potential will drive the establishment of a pH gradient (inside acidic) in the presence of nigericin a potassium/proton exchanger. The observed instability of the pH gradient is similar to that demonstrated in Fig. 14B suggesting that the putative movement of TRIS does not require the presence of specific ionophores. Tris is also able to dissipate imposed pH gradients in the absence of ionophores (data not shown) indicating that the transmembrane flux of TRIS does not require the presence of specific ionophores. Attempts to generate a model membrane system which exhibited a stable ApH without an induced A * employed the permanently charged lipophilic cation TPP + . The results presented in Fig. 15A demonstrate that the presence of TPP + (5 mM) in the external buffer of LUVs exhibiting a K + diffusion potential in the presence of valinomycin caused a dissipation of the observed electrical potential from 150 mV (Fig. 14A) to 70 mV. An induced pH gradient was not observed in this system. The presence of the proton ionophore CCCP in addition to valinomycin to this system (Fig. 15B) caused a slight increase in the observed pH gradient but no difference in the observed electrical potential. The results of Fig. 15C demonstrate that the presence of the potassium-proton exchanger nigericin results in the establishment of a large pH gradient (inside acidic). However, an induced electrical potential is developed despite the presence of the permeable charged cation in the external buffer. The pH gradient is observed to be unstable which may be due to a generalized increase in the permeability of vesicles at this ratio of TPP + to phospholipid (5:1 mol:mol). In summary, a model membrane system exhibiting a pH gradient and no A * was not observed in the presence of a high concentration of external T p p + 3.3.2 Uptake of lipophilic amines into model membrane systems The results to this stage indicate two approaches for determining whether lipophilic amines can cross membranes in a charged or uncharged form. The first approach is to correlate the accumulation of the lipophilic amine with either the presence of a ApH (neutral transport) 67 or the presence of a A * (charged transport). The second approach argues that amines which are transported in the charged and uncharged forms will act like proton ionophores and consequently dissipate the electrical potential in the system presented in Fig. 14A. These two approaches will be used to determine the mechanism of transport of several lipophilic amines which are of pharmaceutical interest. The transport of the pharmaceuticals will be compared with that of MeAm - a model lipophilic amine which is generally accepted to be transported as the non-ionic species; The results of Fig. 16A support the commonly accepted model for MeAm permeability (Chapter 2; Rottenberg, 1979). No significant accumulation of MeAm is observed in LUVs which exhibit a stable electrical potential but no measurable pH gradient. Significant accumulation of MeAm is observed (27 nmoles per /jmol phospholipid) in the presence of both valinomycin and CCCP, consistent with the generation of a pH gradient (inside acidic). This accumulation into the LUVs is transient and within 2 hrs the MeAm has been totally released from the vesicles. Finally, nigericin also causes a similar transient association of MeAm with the vesicles. Particular attention should be paid to the observations that MeAm is taken up in the presence of a pH gradient and that full release of MeAm is observed upon dissipation of the electrochemical gradient. Timolol, a ^-adrenergic blocker, is a secondary amine (pK a 9.2) which is used for the treatment of glaucoma. It is known to have only a limited local anaesthetic action which has been suggested to be due to its low octanol:buffer partition coefficient (Herbette et al., 1983). The results of Fig. 16B demonstrate that the association of timolol parallels that observed for MeAm. An accumulation of 3 nmol of timolol per'/unol phospholipid is observed over 60 min in a model membrane system which exhibits a stable electrical potential (A* > 150 mV) and no pH gradient (ApH < 20 mV). This represents an average velocity over the first 60 min for the putative net flux of the charged amine of 25 x 10"6 nmol/min.cm2 (assuming that 1 /miol of phospholipid has a surface area of .2000 cm2). However, the presence of valinomycin and CCCP results in a much more rapid association of timolol with the LUVs (29 nmol per /xmol 68 Fig. 16. Time course for the uptake of lipophilic amines into LUVs in the presence of external TRIS. All conditions are outlined in Fig. 14. Initial concentrations of lipophilic amines are (A) 100 /iM MeAm; (B) 100 /iM timolol; The symbols represent uptake in the presence of ( • ) 0.5 /ig valinomycin per /imol phospholipid; (•) 0.5 /xg valinomycin per /xmol phospholipid and 20 /iM CCCP; ( A ) 1 /iM nigericin. 100 Uptake of MeAm (nmol/lmol PL) 6 0 -100 Uptake of Timolol (nmol/lmol PL) 150 Time (min) 69 Fig. 16. Time course for the uptake of lipophilic amines into LUVs in the presence of external TRIS. All conditions are outlined in Fig. 14. (C) 100 ./iM dibucaine HCl; (D) 100 uM chlorpromazine HCl; (E) 100 /iM doxorubicin HCl. The symbols represent uptake in the presence of ( • ) 0.5 /ig valinomycin per /imol phospholipid; ( • ) 0.5 /xg valinomycin per /imol phospholipid and 20 /iM CCCP; ( A ) 1 /iM nigericin. 100 Uptake of Dibucaine (nmol/lmol PL) Uptake of Chlorpromazine (nmol/lmol PL) Uptake of Doxorubicin (nmol/lmol PL) 60 90 Time (min) 150 70 phospholipid is vesicle-associated after 2 min) representing an average velocity for the transport of the neutral form of the amine of 7 x 10"3 nmol/min.cm2. Note that this is a significant underestimation of the initial velocity of the neutral form of the amine but a more accurate estimation is beyond the scope of the technique used for the isolation of the LUVs. The association of timolol with LUVs in the presence of the ionophore nigericin parallels that seen for the model membrane system containing valinomycin and CCCP. Thus, the rate of flux at neutral pH of the uncharged form of the timolol is at least 300 fold greater than that for the charged flux of the amine even though the p K a for timolol is 9.2. The extent of the ApH induced association of timolol (29 nmol per /xmol phospholipid) with the LUVs is similar to that observed for MeAm (Fig. 16A). This association with the LUVs is transient, and the release bf timolol from the LUVs parallels the dissipation of the electrical potential and pH gradient (data not shown but see Fig. 14B for reference). These results appear consistent with the hypothesis that transport of timolol occurs primarily in the neutral form of the amine near physiological pH. Dibucaine, a local anaesthetic, is a tertiary amine (pK a 8.5), which has been observed to be accumulated in LUVs exhibiting a large A * (inside negative) and a small induced ApH (Table 2; Mayer et al., 1985a) and LUVs exhibiting a large ApH (Table 2; Mayer et al., 1988). The results of Fig. 16C demonstrate that dibucaine is slowly accumulated in vesicles exhibiting a stable electrical potential and no induced pH gradient. This represents an average velocity for the putative flux of the charged amine of approximately 167 x 10"6 nmol/min.cm2 over the first 60 min. The net flux of dibucaine is significantly faster than that observed for timolol or MeAm under similar conditions. However, the presence of valinomycin and CCCP or nigericin accelerates the association of dibucaine with the LUVs (va y 20 x 10"3 nmol/min.cm2), consistent with the much more rapid uptake of the neutral species in response to a pH gradient. The extent of uptake of dibucaine (80 nmol per /xmol phospholipid) into LUVs exhibiting a ApH of greater than 2 units exceeds that observed for MeAm (Fig. 16A) or timolol (Fig. 16B). Despite the eventual dissipation of both the pH gradient and the electrical potential 71 in these model membrane systems, approximately 50 nmol of dibucaine per nmol phospholipid remains associated with the LUVs over the time course of the experiment. These observations imply that dibucaine is interacting with LUVs in a fundamentally different manner than either MeAm or timolol. The results of Fig. 16D and Fig. 16E demonstrate similar experiments examining the association of the anti-schizophrenic agent chlorpromazine and the anti-neoplastic agent doxorubicin with model membrane systems exhibiting ion gradients. The results indicate that both chlorpromazine and doxorubicin slowly associate with the LUVs in the presence of valinomycin and TRIS - a model membrane system which exhibits a stable electrical potential and no observable pH gradient. However, the presence of a pH gradient of greater than 2 units leads to a rapid association of the drugs with the LUVs (35 nmol per /imol phospholipid for chlorpromazine or 85 nmol per nmol phospholipid for doxorubicin). In addition, although some drug is released as the pH gradient and electrical potential is dissipated across the model membrane systems, significant levels of drug remain associated with the LUVs after the 2 hr time course. These results parallel the observations of the association of dibucaine with the model membrane systems. In summary, for the drugs studied here, the rapid association of the drug with the LUVs is correlated with the presence of a pH gradient (inside acidic) whereas the presence of an electrical potential (inside negative) leads to a much slower association of the drugs with the model membrane systems. This is consistent with the higher permeability of the neutral form of the amine. It is notable that the amount of chlorpromazine, dibucaine and doxorubicin associated with the vesicles in response to ApH far exceeds that observed for MeAm or timolol. All of the drugs are released more slowly than the initial rate of uptake and in fact considerable association of chlorpromazine, dibucaine and doxorubicin with the vesicles is observed after a complete dissipation of the electrochemical gradient. The accumulation of several lipophilic amines in LUVs which exhibited a large electrical potential and no pH gradient suggest that these amines can act as proton ionophores. This 72 question is of interest since amines which have local anaesthetic properties have been suggested to be uncouplers of oxidative phosphorylation (Garlid and Nakashima, 1983; Dabadie et al., 1987). In the L U V system illustrated in Fig. 16 amines acting as proton ionophores would generate a A * induced pH gradient if the flux of the protonated amine is greater than the flux of the TRIS. Alternatively, if the flux of the charged amine is greater than the net inward flux of protons then an instability of the membrane potential would be seen. However, under the conditions of Fig. 16B-E the drugs did not alter either the establishment of the pH gradient or the stability of the membrane potential (data not shown) indicating that the putative flux of the amines in these model membrane systems does not lead to a proton ionophoric activity. 3.3.3 Lipophilic amines as proton ionophores The ability of lipophilic amines to act as uncOuplers is expected to depend upon the ratio of drug to lipid since hydrophobic molecules such as general anaesthetics have been shown to alter membrane permeability in a concentration dependent manner (see for example Barchfeld and Deamer, 1985). The results of Fig. 17 demonstrate that lipophilic amines can increase the permeability of liposomes when incubated at high drug:lipid ratios (0.5:1 mol/mol). In Fig. 17A, the influence of lipophilic amines on the A * induced formation of ApH is examined. Consistent with the high permeability of the neutral form of the amines, all of the drugs decrease the formation of the A * induced pH gradient. This behavior is similar to that observed for the influence of the weak base TRIS on the A * induced formation of ApH in LUVs (see Fig. 14). The results presented in Fig. 17B demonstrate that under these conditions (drug:lipid 0.5:1 mol/mol) the membrane potential of the LUVs is observed to be unstable indicating that a significant increase in the permeability of the model membrane system has occurred. In addition, at higher drug:lipid ratios or for higher values of external pH all the drugs cause a further increase in the instability of the membrane potential (data not shown). This latter observation suggests a role for the neutral form of the amine in altering the apparent membrane 73 Fig. 17. Time course for the establishment of the pH gradient (A) or electrical potential (B) across EPC LUVs in the presence of several lipophilic amines. LUVs containing 100 mM K 2S0 4, 1 mM sodium MES (pH 6.0) were passed down a G-50 Sephadex gel filtration column equilibrated in 200 mM sucrose, 10 /iM K 2S0 4, 20 mM histidine (pH 6.0). The liposomes were then incubated in the presence of no added drug (•), 500 /iM timolol (A.), 500 uM dibucaine HCL (•), 500 /iM chlorpromazine HCL (•) or 500 /iM doxorubicin HCL (y) and 0.5 ug valinomycin per /imol phospholipid. The pH gradient (A) or electrical potential (B) was detected as outlined in Section 3.2.3. 74 Fig. 18. Lipophilic amines as proton ionophores. EPC LUVs containing 100 mM K 2 S 0 4 , 1 mM HEPES (pH 7.0) with 200 mM sucrose, 20 mM HEPES / TRIS (pH 7.0) in the external medium were incubated in the presence of 10 (iM TPB" and 100 /iM MeAm (•), 100 pM Timolol ( A ) , 100 /iM chlorpromazine HCL ( T ) , 100 /iM dibucaine HCL (•) or 100 /iM doxorubicin HCL (•). The electrical potential was monitored in the presence of 10 /iM MeAm, 10 /iM TPP + and 0.5 /ig valinomycin per /imol phospholipid. The electrical potential was determined according to the procedure outlined in Section 3.2.3. (mV) A*150-2 0 0 -100-T 50 0 0 30 60 90 120 150 Time (min) 7 5 permeability of the model membrane systems though the exact mechanism is not presently clear. These experiments demonstrate that the ratio of drug to lipid is an important factor in determining whether the presence of a hydrophobic amine will influence the permeability properties of model membrane systems. As detailed previously, the uncoupling ability of an amine has been proposed to depend upon the salt of a strong acid (Garlid and Nakashima, 1983). The results of Fig. 18 indicate that amines which have local anaesthetic properties are able to act as proton ionophores in LUVs. Addition of TPB" (a salt of a strong acid) causes an increase in the instability of the A * in the presence of external TRIS (for a comparison see Fig. 14). A further increase in the instability of the A * is observed in the presence of the local anaesthetic dibucaine and the anti-schizophrenic agent chlorpromazine whereas the /J-blocker timolol or the anti-neoplastic agent doxorubicin appear not to influence the stability of the electrical potential. The dissipation of . the electrical potential is consistent with the ability of amines with local anaesthetic properties to form an ion pair with TPB" which facilitates transport of the charged amine across membranes. 3.4 Discussion The results presented in this chapter provide new information concerning the transport of amines across model membrane systems. There are three points of interest which will be discussed in turn. The first point concerns the ability of TRIS to dissipate A * induced pH gradients, the second point concerns the mechanism of transport of the amines in response to transmembrane pH gradients and electrical potentials and finally the ability of amines to act as proton ionophores is discussed. Examining the influence of A * or ApH on the association of drugs with liposomes is complicated by the inherent high permeability of liposomes to protons (Chapter 2; Deamer, 1987) and to the influence of the drugs themselves on the permeability properties of the model 76 systems (Barchfeld and Deamer, 1985; Bally et al., 1988). In particular, a model membrane system exhibiting a large stable A * and no induced pH gradient would be a useful tool for examining the interaction of weak bases with model membrane systems. Work presented in this chapter demonstrates the development of such a model membrane system. It is shown here that the presence of a relatively high (ca 10 mM) concentration of the weak base TRIS in the external buffer of vesicles exhibiting large (> 150 mV) K + diffusion potentials effectively dissipates the measured A * induced pH gradient. TRIS is also able to dissipate pH gradients formed in the presence of valinomycin and CCCP or nigericin. This is entirely consistent with previous reports indicating that TRIS can dissipate pH gradients in proteoliposomes (Sone et al., 1976) and kidney cells (Steels and Boulpaep, 1987). This system is particularly useful for examining the mechanism of transport of lipophilic amines across LUVs. Work presented in this chapter appears to support the mechanism in which amines are transported primarily as the neutral form. Support for this model comes from several observations. Rapid transport (v a y = 7-20 x 10"3 nmol min"1 cm2) of the lipophilic amines is observed in the presence but not in the absence (v a y = 2-17 x 10"5 nmol min"1 cm2) of a pH gradient. The membrane potential of LUVs exhibiting a potassium diffusion potential in the presence of valinomycin and TRIS is stable in the presence of relatively low concentrations of amine (100 /iM) indicating that the flux of the charged amine does not accelerate the flux of protons in response to A * . Finally, amines were able to dissipate A * induced formation of pH gradients even at acidic pH (pH 6.0) where most of the amine is in the neutral form. These observations are consistent with the rapid transport of the non-ionic form of the molecule and demonstrate that transport via the charged form is very slow in comparison. Difficulties with this simple mechanism stem from two observations: the extent of uptake into vesicles exhibiting pH gradients of 2.6 units is much greater for several amines than predicted on the basis of the Henderson-Hasselbach equilibrium, and that the release of the lipophilic amines is not entirely correlated with the dissipation of the pH gradient. These difficulties will be discussed in turn. 77 The first point concerns the amount of drug accumulated by LUVs exhibiting pH gradients of greater than 2 units (inside acidic). The influence of transmembrane pH gradients on the expected transmembrane concentration gradient of amines is derived from the following considerations. Equation 1 describes the acid dissociation constant (K a) for a simple amine. K a = ([HA] [H+]) / [H 2 A + ] (1) where [H 2 A + ] and [HA] refer to the charged and neutral form of the amine respectively and [H+] refers to the hydrogen ion concentration. If the equilibrium constant for the amine is the same for the vesicle associated amine and the free amine then: [H 2 A + ] o / ([HA]o [H+]o) = [H 2A+]j / ([HA]; [H+].) (2) where the subscripts i and o refer to the inside and outside of the vesicle respectively. For the case where [H 2 A + ] Q » [HAQ] (pK& » pHQ) and where the concentration of the neutral membrane permeable form is the same on both sides of the membrane then it follows that: [H+]j / [H +]Q = [H 2A +]. / [H 2 A + ] o (3) Since the K a « [H +] i 5 the vesicle associated amine is (for the purposes of this argument) equal to [H 2A +].. However, it should be noted that where p K a = or < pH Q then: [H+]j / [H +]Q > [H 2A +]. / [H 2 A + ] Q (4) Simple amines such as the radiolabeled probe MeAm have been shown to respond to transmembrane pH gradients in this manner (Chapter 2). For example, in L U V systems exhibiting pH gradients of 2.6 units, 25 nmol of methylamine per /wnol phospholipid (initial 78 concentration of 100 / iM,. l mM phospholipid) is accumulated. The ^-blocker timolol is also seen to be accumulated to a similar extent (29 nmol per /imol phospholipid) in response to transmembrane pH gradients whereas chlorpromazine (43 nmol per /imol phospholipid), dibucaine (80 nmol per /imol phospholipid) and doxorubicin (88 nmol per /imol phospholipid) are all accumulated to a much greater extent. The physical basis for these observations is clearly of interest. The above analysis (Henderson-Hasselbach equilibrium) is "an oversimplification. The correlation between the pH gradient and the amount of vesicle associated amine (Fig. 4, Fig:5) does not unequivocally demonstrate that the entrapped MeAm is associated with the aqueous internal compartment. Lipophilic amines may also be associated with the internal bilayer of the L U V or with other structures such as chelates and precipitates. For example, the extent of interaction of a lipophilic amine between the aqueous compartment and the inner leaflet of the membrane will be influenced by the lipid buffer partition coefficient of the amine, the p K a of the free and membrane bound form of the amine, the pH of the medium, the pH at the interface of the membrane and the location of the charge of the amine with respect to the interface (Westman et al., 1982). Consequently, it is difficulty to predict the exact nature of the interaction of amines with LUVs in response to pH gradients. However, it is possible to predict whether changes in the partition coefficient, the p K a or the solubility of the amines could account for the observed differences in the ApH association of the amines with the LUVs. The extent of association of the amines with LUVs in response to pH gradients (2.6 units) and the relevant chemical properties of the amines are summarized in Table 3. The lipophilic amines listed here have pK as which are basic for the soluble form. It is expected that the membrane bound form of the amine will have a lower p K a For example, the p K a of membrane associated tetracaine has been shown to be lower than the soluble form due to the higher interfacial pH at surfaces with significant positive potentials (Westman et al., 1982). However, on the basis of the preceding analysis, a p K a which is within 1 unit of the pH of the external medium is expected to decrease, not increase the extent of association of an amine in 79 response to a pH gradient (equation 4; see also Deamer et al., 1972). Thus, the observed difference in the accumulation of lipophilic amines into LUVs can not be accounted for by differences in the pK as of the L U V associated amines. Intuitively, one would expect a correlation between the hydrophobicity of the drug and the extent of accumulation into LUVs. This presumes that a significant partitioning of the amine into the inner leaflet of the bilayer will take place. Typically, the hydrophobicity of drugs is characterized by the partition coefficient between an aqueous phase (buffer) and an organic phase (octanol). The octanolibuffer partition coefficient of the drugs are listed in Table 3. Clearly, there is no direct relationship between the amount of association of the amines with the LUVs exhibiting pH gradients and the hydrophobicity of the molecules as judged by their octanol:buffer partition coefficient. Timolol has the lowest partition coefficient and has the lowest ApH dependent accumulation. However, doxorubicin is taken up to a greater extent than chlorpromazine. In addition, chlorpromazine is accumulated less effectively than dibucaine despite being more hydrophobic. Thus, there is no simple relationship between hydrophobicity of the drug and the ApH dependent accumulation. Perhaps the reason the hydrophobicity of the molecule does not have a major influence of the extent of association of the drugs is due to the membrane saturation effect detailed by Flewelling and Hubbell, (1986a) and Frezatti and co-workers (1986). For example, TPP + and TPB" have binding affinities for EPC bilayers which differ by 4 orders of magnitude but the maximum observed binding density for the two molecules is approximately 1 molecule of probe per 100 phospholipid molecules (Flewelling and Hubbell, 1986a). This would correspond to 5 nmol of drug per /imol phospholipid which is clearly below the amount of accumulation observed in the systems examined here. Thus, the p K a and the hydrophobicity of dibucaine, chlorpromazine and doxorubicin do not account for the extent of accumulation observed in the presence of a pH gradient. The third possible explanation, as noted above, allows for the possibility that the drugs may form alternative structures after associating with the liposomes. As indicated in Table 3, dibucaine, chlorpromazine and doxorubicin have only a limited solubility in the internal buffer 80 whereas MeAm and timolol are observed to be much more soluble. Under similar conditions (pH and concentration of drug) it has been suggested that dibucaine and chlorpromazine will form micelles (Mayer et al., 1988; Luxnat and Galla, 1986) and that doxorubicin will form dimers and higher ordered structures which are stabilized by stacking interactions (Eksborg, 1977). These structures could account for the extent of uptake of these drugs into liposomal systems in response to pH gradients. However, there is presently no structural evidence to support this hypothesis. A striking observation made during the course of this study was that several lipophilic cations (dibucaine, chlorpromazine and doxorubicin) were not released from LUVs after complete dissipation of the electrical potential and pH gradient. Presently, the mechanism by which the amines are retained in the liposomes is unclear. Dalmark and Storm (1981) have previously noted that passive permeability of doxorubicin across erythrocytes is influenced by the concentration of the amine. Increasing concentrations of doxorubicin led to a decreased passive flux across the membrane. The rationale for this behavior assumed that stacking interactions which are favoured at relatively high concentrations of amine decrease the amount of the membrane permeable monomer. An analogous mechanism could be operating for vesicle entrapped doxorubicin, chlorpromazine and dibucaine. The putative structures mentioned in the previous section would be expected to lower the concentration of the membrane permeable form. In turn, this would be expected to slow the rate of release of the amines in the absence of a pH gradient. However, this mechanism does not make any specific predictions concerning the relative amount of doxorubicin, dibucaine and chlorpromazine associated with the LUVs after dissipation of the pH gradient. The final area of discussion concerns the putative proton ionophoric action of ,the amines. Clearly, the results presented in Section 3.3.2 indicate that the amines do not act as proton ionophores when incubated at a drug:lipid ratio of 10:1. This implies that the flux of the charged amine is slower than the flux of protons in this system. This is consistent with the previous observations concerning the influence of local anaesthetics on mitochondria (Garlid and 81 Table 3. A comparison of the uptake of the lipophilic cations in LUVs exhibiting a pH gradient (ApH > 2 units) with the chemical properties of the amines. Lipophilic Amine Uptake1 nmol//»mol PL log Partition Coefficient2 (Ref)3 . pK (Ref)3 Solubility4 (mM) Methylamine 25 NA 10.5 (1) > 50 Timolol 29 0.7 (2) 9.2 (3) > 50 Dibucaine 81 4.3(4) 8.5 (5) < 10 Chlorpromazine 45 5.2 (4) 9.4(5) < 10 Doxorubicin 85 1.1 (6) 8.2(5) < 10 1. Uptake of lipophilic cations in the presence of a stable pH gradient (ApH > 150 mV) is expressed in nmol per /imol phospholipid. 2. Refers to the log of the octanol/buffer partition coefficient at neutral pH. 3. The reference numbers correspond to: 1 Rottenberg, 1979 2 Herbette et al., 1986. 3 Raymond and Born, 1986. , 4 Hansen and Elkins, 1971. 5 Newton and Kluza, 1978. 6 Goldman et al, 1978. 4. The maximum solubility of drugs was determined in 100 mM K 2 S 0 4 , 20 mM HEPES, (pH 7.0). 82 Nakashima, 1983). However, it is also demonstrated (Section 3.3.2) that at higher drug:lipid ratios (0.5:1) the more hydrophobic amines (chlorpromazine, dibucaine and doxorubicin) can increase the permeability of the liposome such that the electrical potential is unstable. This influence may well be mediated by the neutral form of the drug since the effect is pronounced at high pH and the neutral form would be expected to have a higher buffenlipid partition coefficient. It is perhaps not surprising that neutral hydrophobic molecules can increase the permeability of liposomes since general anaesthetics (which are neutral hydrophobic molecules) have been shown to increase the permeability of liposomes to protons and potassium iOns (Barchfeld and Deamer, 1985). Finally, the local anaesthetics dibucaine and chlorpromazine show proton ionophoric action in the presence of TPB"; an observation consistent with that previously documented (Garlid and Nakashima, 1983). The physiological relevance of these observations is not presently clear. It has been proposed that lipophilic amines can respond to transmembrane electrical potentials (inside negative) and as such would be expected to accumulate in the inner leaflet of membranes exhibiting the appropriate membrane potential (Mayer et al., 1985a; Bally et al., 1985). However, the work presented in this Chapter argues that the flux at physiological pH is primarily in the neutral form. Consequently, it is expected that amines will accumulate in organelles which exhibit acidic pH gradients such as chromaffin granules, endocytic compartments or lysosomes (for a review of acidic organelles see Anderson and Orci, 1988) rather than into organelles which exhibit negative potentials (such as the mitochondria). The extent to which this property is involved in their pharmacological action is not presently understood. In addition it has been proposed that lipophilic amines which are local anaesthetics can act as uncouplers in the presence of the salt of a strong acid (Garlid and Nakashima, 1983). However, the requirements for this mechanism are quite specific. Possible candidates for the lipophilic anion are TPB" and SCNT but are not particularly physiological. The presence of high concentrations (10 mM) of physiological salts such as chloride or sulphate did not influence the 83 A * induced flux of chlorpromazine (data not shown). Consequently, the physiological relevance of this mechanism is not clear. It appears more likely that lipophilic amines which are local anaesthetics will be accumulated in acidic organelles as discussed above. Since, as a class they are quite hydrophobic this would suggest that they would in turn be concentrated in the internal leaflet of these organelles. It is possible that these properties are in turn related to their pharmacological action. 84 PHOSPHOLIPID ASYMMETRY IN LARGE UNILAMELLAR VESICLES INDUCED BY TRANSMEMBRANE pH GRADIENTS 4.1 Introduction The transbilayer phospholipid distribution in biological membranes is frequently observed to be asymmetric. For example, this has been demonstrated in the erythrocyte plasma membrane (for review see Zwaal, 1978; Op den Kamp, 1979; Etemadi, 1980b), the platelet plasma membrane (Chap et al., 1977), rat liver endoplasmic reticulum (Higgins and Pigott, 1982; Sleight and Pagano, 1985) the Golgi complex (van Meer et al., 1987) and for beef heart inner mitochondria membrane (Kreba et al., 1978). Two general conclusions have been reached concerning the nature of phospholipid asymmetry in biological membranes. First, in the plasma membrane of the erythrocyte, lymphocyte or platelet, the amino containing phospholipids phosphatidylserine (PS) and phophatidylethanolamine (PE) are found preferentially in the inner leaflet (IM) while the choline containing lipids phophatidylcholine (PC) and sphingomyelin (SM) are located preferentially in the outer leaflet (OM). The second general conclusion states that the rate of transbilayer movement in membranes involved in the biosynthesis of lipids (i.e. the endoplasmic reticulum) is considerably more rapid than observed for membranes not involved in lipid biosynthesis (for a review see Dawidowicz, 1987). However, even the relatively slow (t^2 = 13-24 hr) transbilayer movement of PC in the erythrocyte would be expected to dissipate lipid asymmetry over the lifetime of the cell (105 d) (van Deenen, 1981). Consequently, the mechanism whereby transbilayer phospholipid asymmetry is created and maintained is being investigated. Recently, Devaux and co-workers have demonstrated that spin labelled analogues of the amino containing lipids PS and PE will be preferentially transported from the OM to the IM of the erythrocyte plasma membrane (Seigneuret and Devaux, 1984; Zachowski et al., 1986), the platelet plasma membrane (Sune et al., 1987) and the lymphocyte plasma membrane (Zachowski 85 et al., 1987). The amino lipid transport requires intracellular ATP, is inhibited by N-ethylmaleimide, vanadate and intracellular C a 2 + (Zachowski et al., 1986). The transport of the PS analogue is saturable and inhibited by the PE analogue but not by lyso-PS, PC or SM analogues. These observations are consistent with the hypothesis that amino-lipid transport is mediated by an amino-lipid "flippase". The general conclusions concerning the putative role of an enzyme in amino lipid transport reached by Devaux and co-workers have been supported by other workers using short chain saturated analogues of PS (Daleke and Huestis, 1985) or radiolabeled analogues of PS, PE and PC (Tilley, et al., 1986). In addition, Schroit and co-workers (1987) have recently identified a 31 K D erythrocyte protein which is labelled by a photoactivatable PS analogue. A 31 K D peptide is also labelled by sulfhydryl reagents which inhibit PS transport in the erythrocyte (Connor and Schroit, 1988). The previous observation by Devaux and co-workers and the results of the chemical cross linking studies provide strong evidence that amino lipid transport is mediated by proteins in the erythrocyte. However, there is no evidence to suggest that transport of the choline lipids is mediated by a protein. The transport of choline lipid analogues in the erythrocyte is not saturable, is not competitively inhibited by the amino lipid analogues, does not require intracellular ATP and is not inhibited by N-ethylmaleimide (Zachowski et al., 1986). Thus, SM and PC are thought to be transported in a protein independent manner (Zachowski et al., 1986). The driving force for the asymmetry of the PC across the bilayer has been suggested to be asymmetry of the amino lipids (Zachowski et al., 1985). Transport of phospholipids across the endoplasmic reticulum is considered to be fast in comparison to the erythrocyte membrane (t^2 of minutes) (Zilversmit and Hughes, 1977; van den Besselaar et al., 1978). Bell and co-workers have examined the transbilayer transport of a short chain (water soluble) radiolabeled analogue of PC in isolated microsomal membranes. The transport is sensitive to trypsin, chemical reagents and is competitively inhibited by structural analogues (Bishop and Bell, 1985; Kawashima and Bell, 1987) which suggests the involvement of a specific lipid "flippase". Backer and Dawidowicz (1987) have been able to reconstitute a long 86. chain PC translocating activity into liposomes. However, this activity was not sensitive to proteases, heat denaturation or N-ethylmaleimide and consequently does not appear to be directly related to the work by Bell. In addition to the work involving nascent lipid transport in the endoplasmic reticulum Rothman and Kennedy (1977) have demonstrated that newly synthesized PE in the cytoplasmic membrane of B. megaterium is rapidly transported across the cytoplasmic membrane (t^2 < 3 min). Thus, there is good evidence to suggest that transport of lipids across membranes involved in lipid biosynthesis is fast in comparison to membranes not involved in lipid biosynthesis. Another mechanism which has been proposed to account for the rapid transbilayer transport of phospholipids in membranes is related to the formation of nonbilayer structures (Cullis and de Kruijff, 1978). PC transport across the erythrocyte plasma membrane (Classen et al., 1987; Tournois et al., 1987), rat liver endoplasmic reticulum (Van Duijn et al., 1986) and model membrane systems (Gerritsen et al., 1980a) has been correlated with the presence of nonbilayer structures which have, in some cases, been tentatively identified as H n or inverted micelles. However, this dpes not preclude the possibility that lipid transport is a protein mediated process. Phospholipid distribution and translocation across model membrane systems has also been studied. In SUVs, a curvature induced asymmetry is often observed. For example, PC SUVs containing low mole fractions of PE display an asymmetry in which PE favours the OM (Litman, 1974; Lentz and Litman, 1978; Nordlund et al., 1981). The distribution of acidic phospholipids is more controversial: at low mole fractions, PS and PG have been reported to favour the OM of SUVs (Lentz et al., 1982; Nordlund et al., 1981; Massari et al., 1978) or the IM (Kumar and Gupta, 1984; Berden et al., 1975; Barsukov et al., 1980). However, there is general agreement in the literature that curvature induced asymmetry is only relevant to limit size vesicles (Nordlund et al., 1981; Kumar and Gupta, 1984; Hope and Cullis, 1987). In general, the transport of lipid across model membrane systems has been shown to be very slow (t1/2 of hours) (for a recent review see Op den Kamp, 1979; Etemadi, 1980b). 87 Attempts have been made to determine what factors modulate the transbilayer movement of lipids in model membrane systems. Factors such as the reconstitution of erythrocyte intrinsic membrane proteins (de Kruijff, et al., 1978; Gerritsen et al., 1980b), the effect of the gel to liquid crystalline phase transition (de Kruijff and Zoelen, 1978), lipid phase separation between PCs of different chain length (de Kruijff and Wirtz, 1977), the presence of phospholipase D (de Kruijff and Baken, 1978), the presence of detergents (Kramer et al., 1981) and lipid oxidation products (Shaw.and Thompson, 1982) have all been shown to increase the transbilayer movement of lipids in model membrane systems. However, the rate of transbilayer transport of phospholipids in these membranes is still considerably slower than that observed in biosynthetic membranes such as the endoplasmic reticulum or the erythrocyte plasma membrane. Recently, the influence of transmembrane pH gradients on the transport of fatty acids (FA), stearylamine (SA) and sphingosine in LUVs has been investigated (Hope et al., 1987). Perhaps not surprisingly, given the observations that fatty acids (FA) can act as proton ionophores, (Gutnecht, 1987c) it was found that a pH gradient (inside basic / outside acidic) led to a transbilayer transport of fatty acids from the OM to the IM of the liposomes. SA and sphingosine were transported in a similar manner except that they were accumulated into the IM of vesicles exhibiting the reverse pH gradient (inside acidic/outside basic). This transport was notable in that it was both rapid ( t^ < 2 min) and complete (> 95% asymmetry for both FA and SA). This Chapter describes the influence of transmembrane pH gradients on the, transbilayer transport of acidic phospholipids across LUVs. It is demonstrated that an asymmetric distribution of phosphatidylglycerol (PG) and phosphatidic acid (PA) can be induced by a pH gradient (inside basic/outside acidic) at elevated temperatures. Zwitterionic phospholipids, such as the amino-lipids phosphatidylserine (PS) and phosphatidylethanolamine (PE) remain equally distributed across the bilayer of LUVs under similar conditions. 88 4.2 Materials and Methods 4.2.1 Materials Egg phosphatidylcholine (EPC), dioleolphosphatidylcholine (DOPC), egg phosphatidylglycerol (EPG), beef heart cardiolipin (CL), egg phosphatidic acid (EPA) and soya phosphatidylinositol (SPI) were obtained from Avanti Polar Lipids and used without further purification. [3H]-dipalmitoylphosphatidylcholine ([3H]-DPPC) was purchased from New England Nuclear. 4.2.2 Preparation of large unilamellar vesicles and establishment of ion gradients Large unilamellar vesicles (LUVs) were prepared in the appropriate buffer as previously described (Section 2.2.2). Gel filtration was performed as described in Section 2.2.3. 4.2.3 Detection of asymmetry of amine containing lipids PS, PE, SA asymmetry across LUVs was assayed by trinitrobenzenesulforiic acid (TNBS) labelling according to a previously published procedure (Hope and Cullis, 1987). TNBS will label primary amines yielding a phenylated derivative which absorbs at 420 nm. LUVs are sufficiently impermeable to TNBS (Hope and Cullis, 1987) that the in situ labelling of an amino-lipid in LUVs in the presence and absence of a detergent will give an indication of the asymmetry of the lipid. However, the rapid (t^2 < 1 min) transbilayer movement of SA makes unreliable the detection of less than complete SA asymmetry by this method since the labelling of the external SA has been shown to disturb the transbilayer distribution (data not shown). Typically, the asymmetry of a amino containing lipid is determined as follows. The sample cuvette in a double beam Shimadzu UV-160 spectrophotometer contained an aliquot of vesicles 89 (ca 1 mg phospholipid / ml with 10% amino-lipid) in the presence of 200 mM Na 2S0 4, 20 mM Pi, (pH 8.5) for SA and PE containing LUVs and 200 mM Na 2S0 4, 20 mM borate, (pH 9.5) for PS containing LUVs. The reference cuvette contained a similar concentration of vesicles without the amino-lipid. The absorbance at 420 nm was monitored after the addition of an aliquot of TNBS (final concentration of 0.5 pM) until a plateau representing 50% labelling is reached. Subsequently, an aliquot of Triton X-100 is added sufficient to disrupt the vesicles and the absorbance was again monitored until a plateau value had been reached. The asymmetry can be calculated as the A 4 2 0 in the absence of detergent divided by the A 4 2 0 in the presence of detergent. PE or PS containing LUVs in the absence of a pH gradient give 50% labelling under these conditions (Hope and Cullis, 1987). 4.2.4 Detection of acidic phospholipid asymmetry by ion,exchange chromatography LUVs of the appropriate lipid composition were prepared in 200 mM sodium HEPES, (pH 8.0) or 150 mM sodium citrate, (pH 4.5) and the external buffer was exchanged using gel filtration chromatography for a buffer of low ionic strength containing either 300 mM sucrose, 20 mM glutamic acid, (pH 4.5) or 300 mM sucrose, 20 HEPES, (pH 8.0). After establishing the ion gradients, (pH 8.0/4.5; 8.0/8.0 ; 4.5/4.5) vesicles were incubated at 20°C or 60°C for 30 min. Vesicles were loaded onto a 1.5 by 15 cm DEAE sephacel column (Pharmacia) previously equilibrated with the low ionic strength, and 1 ml fractions were collected. The column was eluted with 2 column volumes of the low ionic strength buffer and the remaining liposomes were eluted with a high ionic strength buffer containing 0.5 M NaCl. Fractions of the eluant were counted by liquid scintillation counting in order to follow the elution profile of the liposomes. 90 4.2.5 Detection of phosphatidylglycerol asymmetry by periodate oxidation Vesicles with transmembrane pH gradients were prepared as described above. Asymmetry of EPG was determined by a modification of a chemical assay developed for the' detection of PG asymmetry in SUVs (Lentz et al., 1980; Lentz et al., 1982). This assay is based on the asymmetric periodate oxidation of the terminal vicinal hydroxyl group of PG. The exposed PG was determined as follows: 0.8 ml of vesicles (10-20 mg/ml phospholipid) were added to 5.0 ml of 100 mM sodium acetate, 100 mM citrate (pH 6.0) and the total volume of the reaction mixture was made up to 9.0 ml with water. The oxidation was initiated by the addition of 1 ml of freshly prepared 100 mM sodium periodate. At the indicated time intervals, the oxidation was quenched by adding a 1 ml aliquot of the reaction mixture to a Pyrex tube containing 100 'pi of 1 M sodium arsenite in IN H 2 S0 4 . The formaldehyde produced from the oxidation of the terminal vicinal hydroxyls was determined by the Hantzsch reaction (Nash, 1953). At least 15 min after the oxidation had been quenched, 50 pi of 200 mM sodium cholate was added to disrupt the vesicles and 1 ml of Nash reagent (15 g ammonium acetate, 300 pi glacial acetic acid, 200 /ilacetylacetone in 100 ml HjO) was added for the detection of the formaldehyde. The samples were then capped, heated at 60°C for 10 min and the absorbance was read at 412 nm using a Shimadzu UV-160 spectrometer. Total PG was determined by including the appropriate amount of sodium cholate during the oxidation procedure and excluding it from the latter procedure for the detection of formaldehyde. Control experiments indicated that this assay is linear up to 400 nmoles of total PG (data not shown). Typically less than 200 nmoles of total PG was assayed in this manner. The absorbance for the background (100% EPC vesicles) was less than 5% of the absorbance for the total PG under these conditions. Data is expressed as the % of total PG available for oxidation after correcting for background absorbance. 91 4.2.6 Quasi-elastic light scattering Vesicles were prepared as described above. Quasi-elastic light scattering measurements to determine vesicle size were performed as previously described in Section 2.2.3. 4.2.7 Freeze-fracture Freeze-fracture was performed on appropriate samples containing 25% (v/v) glycerol as a cryoprotectant. The samples were quickly frozen in a liquid freon slush. Freeze-fracture studies were performed on a Balzers BAF 400D apparatus, and replicas were viewed employing a JEOL Model JEM-1200 EX electron microscope. 4.3 Results 4.3.1 Phosphatidylserine is not transported in response to transmembrane pH gradients It has been previously demonstrated that simple lipids which are weak bases or acids can equilibrate across the bilayer of LUVs in a manner that reflects the transmembrane pH gradient (Hope and Cullis, 1987). The transbilayer movement of lipids which are simple weak acids or bases in response to pH gradients is presumed to occur via the neutral membrane permeable species (see for example Chapter 3). Consequently, it was of interest to extend this previous work and examine the influence of transmembrane pH gradients on the transbilayer movement of phospholipids which are weak acids. Phosphatidylserine (PS), a zwitterionic acidic phospholipid can be. negatively charged (pH > 4)j net neutral (3 < pH < 4) or positively charged (pH < 3) (Tsui et al., 1986). If these ionized states exhibit different transbilayer transport rates then an applied transmembrane ApH will result in an asymmetric distribution of PS between the IM and OM of model membrane systems. 92 Fig. 19. Transbilayer transport of PS in response tb pH gradients. EPGPS (9.5:0.5 mol/mol) LUVs were prepared in 300 mM sodium phosphate (pH 8.0) and the external buffer was exchanged for (o) 200 mM sodium sulfate, 20 mM phosphate (pH 8.0) or (•) 200 mM sodium sulfate, 20 mM citrate (pH 4.0). The vesicles were incubated for the appropriate time period at 60°C and the % available PS was determined by the procedure outlined in Section 4.2.3. 100 80 % 60 AVAILABLE PS 40--2 0 -0 0 30 60 90 120 150 TIME (min) 93 The transport of PS in response to imposed pH gradients was examined in LUVs composed of EPC and a low mole fraction of PS. The rationale for picking these conditions was as follows: LUVs were chosen since they do not exhibit curvature induced asymmetry, have high entrapped volumes and have the distinct advantage that no solvent or detergent is used during their preparation (Hope et al., 1985) which may perturb the stability of the transmembrane ion gradients (Cafiso and Hubbell, 1983) or the transport of the lipid (Kramer et al., 1981). EPC was chosen as the lipid matrix because it represents a relatively unsaturated lipid which should facilitate transmembrane transport and finally, the PS was present as a low mole fraction in order to avoid possible packing constraints experienced by lipid redistribution. At an external pH of 4, a significant portion of the PS is net neutral since the intrinsic pK & of the carboxyl is 3.6 (Tsui et al., 1986). If this species is able to be transported at a faster rate than the negatively charged species (internal pH 8.0) then a transmembrane asymmetry of PS will develop (IM > OM). The results of Fig. 19 demonstrate that incubation of LUVs exhibiting a transmembrane pH gradient (pH,8.0 inside, pH 4.0, outside) at elevated temperatures (60°C) results in no net movement of PS from the OM to the IM- Similar behavior is observed for the zwitterionic lipid phosphatidylethanolamine (PE) (Hope et al., 1988), arguing that the rate of transbilayer transport for the zwitterionic phospholipids PS and PE does not vary significantly with pH. 4.3.2 Transbilayer transport of phosphatidylglycerol as detected by ion exchange chromatography It is reasonable to conclude from the results presented above that the presence of a charge on the phospholipid head group slows transbilayer movement of the neutral form to an insignificant rate. This is consistent with previous reports indicating that the polarity of the head group is an important factor in determining the rate of transbilayer transport Of phospholipids (Homan and Pownall, 1988). Consequently, it was of interest to examine the transbilayer movement of the acidic phospholipid phosphatidylglycerol (PG) which has a single 94 Fig. 20. The influence of ApH on the apparent surface potential of vesicles composed of EPCEPG (9.5:0.5 mol/mol). Each panel represents elution profiles from a DEAE Sephacel ion exchange column. Panel A represents EPC LUVs (pH 8.0/4.5) which have been incubated at 25°C for 30 min prior to column loading. Panel B represents EPCEPG (9.5:0.5 mol/mol) LUVs with (•) a pH gradient (pH 8.0/4.5) which had been incubated at 25°C for 30 min. Panel, C represents EPCEPG (9.5:0.5 mol/mol) LUVs with (•) or without (•) a pH gradient (pH 8.0/8.0) which had been incubated at 60°C for 30 min. The columns were eluted first with a low ionic strength elution buffer composed of either 300 mM sucrose, 20 mM HEPES (pH 8.0) or 300 mM sucrose, 20 mM citrate (pH 4.5) and after fraction 11 were eluted with a high ionic strength elution buffer composed of 0.5 NaCl. 100 % RECOVERY OF LIPOSOMES 100 RECOVERY OF LIPOSOMES 10 15 20 25 % RECOVERY OF LIPOSOMES FRACTION NUMBER 95 ionizable group. The single charge present on this molecule arises from the ionization of the phosphate group which has an intrinsic p K a of approximately 1 (Watts et al., 1978). Thus, at pH 4, a small but significant proportion of the PG molecules will be uncharged due to protonation of the phosphate. Development of PG asymmetry in LUVs was assayed by ion exchange chromatography which has been previously used to demonstrate FA asymmetry in EPC vesicles (Hope and Cullis, 1987). The principle behind this assay is that LUVs which exhibit a significant negative surface potential will bind to an anion exchange column (DEAE) under low salt conditions whereas in the absence of a surface potential LUVs elute in the void volume of the DEAE column. Consequently, this assay will detect a decrease in surface charge due to the transport of an acidic lipid from the OM to the IM of LUVs. The results of Fig. 20A demonstrate that EPC vesicles, which have no apparent surface potential do not bind to an anion exchange (DEAE) column under low salt or high salt conditions. EPG:EPC (9.5:0.5 mol/mol) vesicles prepared with an imposed pH gradient (inside basic / outside acidic) bind to the ion exchange columns under low ionic strength conditions but not under high ionic strength conditions (Fig. 20B) indicating that the external leaflet of the liposomes has a significant negative surface potential. Control experiments demonstrate that vesicles with hp pH gradient (pH 4.5/4.5 or pH 8.0/8.0 inside and out) exhibit similar behavior to the vesicles of Fig. 20B (data not shown). The data of Fig. 20C clearly demonstrate that an incubation of the vesicles at 60°C for 30 min prior to loading the vesicles onto the column causes a decrease in the apparent negative surface potential of the vesicles exhibiting a transmembrane pH gradient. This is consistent with a pH dependent transport of PG from the OM to the IM of the LUVs. In the absence of a ApH, incubation of the LUVs at 60°C did not influence the apparent surface charge pf the vesicles. Control experiments demonstrated that these changes in surface potential arising from the putative PG transport can be reversed by a dissipation of the imposed pH gradient at elevated temperatures (data not shown). In addition, no significant degradation of the PG under these conditions could be detected by TLC (data not shown). Thus, these results support the proposal that PG 96 equilibrates across the bilayer in response to transmembrane pH gradients. 4.3.4 Transbilayer transport of phosphatidylglycerol as detected by periodate oxidation The results of Fig. 20 clearly demonstrate an asymmetric distribution of EPG in LUVs that exhibit a pH gradient at elevated temperatures. To obtain quantitative information on the amount of EPG in the OM a second specific assay was employed in which the terminal vicinal hydroxyl of EPG is oxidized to formaldehyde using periodate (Lentz al., 1980). In turn, the formaldehyde produced by the oxidation of PG is detected by the Hantzsch reaction (Nash, 1953). In order to establish that only the PG in the external leaflet is oxidized under the conditions of the assay the following experiment was performed. EPC LUVs containing 25 mM 3-phosphoglycerol (3-P-G) were prepared and the external 3-P-G was removed using standard gel filtration techniques. Under these conditions, the vesicles are impermeable to 3-P-G and consequently the oxidation of 3-P-G by periodate is an indication of the asymmetry of the. periodate assay. The results of Fig. 21a demonstrate that in the absence of detergent, significant levels of 3-P-G are not oxidized over the first 15 min whereas in the presence of detergent the 3-phosphoglycerol is completely oxidized within 6 min. Thus, periodate is impermeable to model membrane systems under these conditions and consequently the periodate oxidation of vicinal hydroxyls represents a simple, useful assay for the detection of glycerol asymmetry. The results of Fig. 22 demonstrate a typical time course for the in situ oxidation of EPG in EPC:EPG LUVs (9.5:0.5 mol/mol). In the absence of detergent, 50% of the available hydroxyl i i . • groups are oxidized within 12 min. This is consistent with the 50:50 distribution of phospholipid between OM and IM of LUVs (Hope et al., 1985). In addition, since the oxidation is complete within 12 min it is expected that significant periodate oxidation of PG in the internal leaflet has not occurred. A similar time course was observed for the oxidation of EPCEPG (9.5:0.5. mol/mol) LUVs which were prepared with a transmembrane pH gradient (inside basic / outside acidic) and incubated at 20°C for 30 min (data not shown). However, incubation of LUVs with 97 Fig. 21. The asymmetry of the periodate assay as demonstrated by the inaccessibility of entrapped 3-phosphoglycerol to periodate oxidation. EPC LUVs were prepared in 25 mM 3-phosphoglycerol, 50 mM citrate, 50 mM acetate (pH 6.0) and the external buffer exchanged for 50 mM citrate, 50 mM acetate (pH 6.0) using gel filtration chromatography. The oxidation of the entrapped 3-phosphoglycerol (3-P-G) was performed as described in Section 4.2.5 in the presence (•) and absence of 10 mM sodium cholate (O). % A V A I L A B L E 3-P-G 100 TIME (min) Fig. 22. Transbilayer transport of EPG in EPCEPG (9.5:0.5 mol/mol) LUVs as detected by periodate oxidation. Vesicles were prepared in 300 mM phosphate (pH 8.0) and the external buffer exchanged for either (•) 100 mM sodium sulfate, 10 mM citrate (pH 4.0) or (A) 100 mM sodium sulfate, 5 mM phosphate (pH 8.0). Total PG (•) was assayed by oxidation in the presence of 10 mM sodium cholate. The vesicles were incubated at 60°C for 30 min prior to assaying for PG asymmetry according to the procedure outlined in Section 4.2.5. ?8 a transmembrane pH gradient (8.0 inside / 4.0 outside) for 30 min at 60°C resulted in a decrease in the PG available for periodate oxidation (Fig. 22). Under these conditions, 85% of the OM PG has been transported to the IM of the LUVs. This is consistent with the results of the ion exchange assay indicating that PG is transported in model membrane systems in response to pH gradients. The results to this point indicate that the presence of transmembrane pH gradients at elevated temperatures results in a decrease in the surface charge of EPG containing LUVs and a decrease in the EPG available for periodate oxidation. Fusion of LUVs to form larger multilamellar liposomes would also show a concomitant decrease in total OM,phospholipid and consequently could account for these observations. Three experiments were conducted to rule. out this possibility. Generation of multilamellar.structures was monitored by quasi-elastic light scattering and freeze-fracture techniques and the ability of transmembrane pH gradients (inside acidic) to drive transport of PG from the IM to the OM was examined. These experiments will be discussed in turn. A convenient method for determining the size distribution of LUVs is quasi-elastic light scattering (Pletcher et al., 1980; Mayer et al., 1986b). The results of Fig. 23a demonstrate a typical quasi-elastic light scattering spectrum for EPCEPG. (9:1 mol/mol) vesicles. The spectrum indicate that the LUVs exhibit a mean diameter of 97 nm (S.D.. 23 nm). Since the size estimate of EPC LUVs by quasi-elastic light scattering techniques is 135 nm (S.D. 35 nm) (Mayer et al., 1986b), it is apparent that inclusion of the acidic lipid decreases the size of the LUVs. The reasons for this are not presently understood. The results demonstrated in Fig. 24B indicate that a 30 min incubation at 60°C does not significantly change the size distribution (mean diameter 93 nm; S.D. 23 nm) of EPCEPG (9:1 mol/mol) LUVs exhibiting a ApH. This clearly indicates that significant fusion of LUVs to form MLVs is not taking place under these conditions. Freeze-fracture electron microscopy of LUVs can also be used to give direct evidence of fusion events (Nayar et al., 1982). Freeze-fracture electron micrographs of EPCEPG (8:2 mol/mol) vesicles are illustrated in Fig. 24. The absence of cross fractures in Fig. 24A 99 Fig. 23. Quasi-elastic light scattering of EPC:EPG (9:1 mol/mol) LUVs. LUVs were prepared in 300 mM sodium phosphate (pH 8.0) and the external buffer was exchanged for 200 mM sodium sulfate, 20 mM citrate (pH 4.5). The vesicles were incubated for (A) 30 min at 25°C or (B) 30 min at 60°C. QEL was performed as described in Section 4.2.6. RELATIVE INTENSITY 120 100 40 90 140 190 240 290 DIAMETER ( n m ) 340 390 120 100--RELATIVE 80 INTENSITY 390 DIAMETER ( n m ) 100 Fig. 24. Freeze-fracture electron micrographs of EPGEPG (8:2 mol/mol) LUVs. LUVs were prepared in 300 mM sodium citrate (pH 4.0) and the external buffer exchanged for 200 mM sodium HEPES (pH 7.5). The vesicles were then incubated for (A) 30 min at 25°C or (B) 30 min at 45°C prior to performing the freeze-fracture electron microscopy (see Section 4.2.7). 101 demonstrates the unilamellar character of these vesicles. The results of Fig 24B illustrate the electron micrograph of EPCEPG (8:2 mol/mol) vesicles which have been incubated with a pH gradient (inside acid / outside basic) for 30 min at 45°C - conditions which have been shown to promote PG redistribution (data not shown). The absence of an increase in cross fractures or an increase in the number of small vesicles indicates that a significant restructuring of the LUVs has not taken place. Thus, the results of the freeze-fracture studies taken together with the results of the experiments involving quasi-elastic light scattering indicate that significant fusion is not taking place in LUVs exposed to pH gradients at elevated temperatures. In addition, the results demonstrate that the development of PG asymmetry does not lead to changes in the morphological characteristics of the LUVs as judged by freeze-fracture techniques. Finally, transbilayer transport of PG can be demonstrated in response to both acidic pH gradients (inside acidic / outside basic) as well as basic pH gradients (Fig. 22). The results of Fig. 26a demonstrate a time course for the periodate oxidation of EPG in EPCEPG (9:1 mol/mol) vesicles which had been incubated with an imposed pH gradient (pH 4.0 inside / 8.0 outside) prior to the oxidation. Control experiments demonstrate that incubation of LUVs at 25°C for 30 min prior to the oxidation assay results in oxidation of 56% of the available PG. This is consistent with an approximately symmetrical transbilayer distribution of PG. Prior incubation of the vesicles in the absence of a pH gradient gave a similar pattern of oxidation (data not shown). EPCEPG (9:1 mol/mol) LUVs with an imposed ApH (pH 4.0 inside / 8.0 outside) which have been incubated for 30 min at 60°C prior to the oxidation assay are asymmetric. Approximately 83% of the EPG is available for periodate oxidation in the absence of detergent. This indicates that approximately 34% of the IM PG has been transported to the OM of the LUVs which represents a net movement of 3.4% of the total IM phospholipid to the OM. The results of Fig. 26b demonstrate the amount of EPG transported for the basic pH gradient (pH 8.0 inside / 4.0 outside). Again, 56% of the PG is available for oxidation in LUVs which have been incubated at 25°C for 30 min prior to the oxidation assay. Incubation of the 102 vesicles at elevated temperatures (30 min, 60°C) results in transport of EPG from the OM to the IM of the LUVs. Thus, the extent of transport of PG in response to the acidic pH gradient (34%) is observed to be approximately the same as the extent of transport observed for the basic pH gradients (33%). It is difficult to make further direct comparisons between these two model membrane systems since the stability of the pH gradient is different for the two conditions (data not shown). 4.3.5 The extent of phosphatidylglycerol asymmetry does not reflect the transmembrane pH gradient The results to this point indicate that rapid PG transport occurs in model membrane systems in response to pH gradients at elevated temperatures. The influence of temperature on the rate of transport is examined in Fig. 25. At elevated temperatures (60°C) transport of PG is observed to occur rapidly. The maximum amount of asymmetry is observed within 10 min and is stable over the time course of the experiment. However, the maximum transbilayer distribution of PG (6:1 IM:OM) is considerably less then expected for the transbilayer distribution of a weak acid in response to a ApH of 2 units (100:1). Transbilayer transport of PG is observed to be temperature dependent. At 37°C, transbilayer transport of EPG is observed to occur at a much slower rate. However, the results demonstrate that significant transport of PG is observed at 37°C. It is important to note that both the rate of pH dependent transport of EPG and the extent of transbilayer asymmetry for EPG is considerably less than observed for stearic acid (Hope and Cullis, 1987). The influence of temperature oh the rate of pH dependent transbilayer transport of PG in LUVs is examined in more detail in Chapter 5. The reason why,the extent of pH driven PG asymmetry is limited is not understood at this point. A likely explanation is that packing constraints caused by the accumulation of EPG in one monolayer prevent the further transport of EPG since this would require an expansion in surface area that cannot be accommodated in LUVs with diameters on the order of 0.1 pm. For 103 Fig. 25. The development of PG asymmetry at two temperatures! EPCEPG (pH 9.5:0.5) LUVs were prepared in 300 mM sodium phosphate (pH 8.0) and the external buffer was exchanged for 200 mM sodium sulfate, 20 mM citrate (pH 4.0). The vesicles were then incubated at 37°C (•) or 60°C (A) for the indicated time period. Vesicles were cooled to 20°C prior to detection of asymmetry by the periodate method (see Section 4.2.5). 100--% 75 AVAILABLE PG 50 25-• 0 0 15 30 45 60 75 T I M E (min) 104 example, it has been shown that erythrocytes undergo radical shape changes in response to a very small increase (perhaps as little as 1%) in the surface area of one monolayer over the other (Ferrel et al., 1985). The maximum asymmetry observed for the PG containing vesicle systems described here would be equivalent to a 3-4% increase in surface area of the monolayer containing excess EPG and an equivalent decrease in the donor monolayer (assuming there is no change in the packing density of each monolayer). In the absence of compensating movements of PC, the packing-restraints imposed by these large changes in surface area might be sufficient to limit EPG asymmetry. Stearic acid and stearylamine would be expected to have a considerably smaller head group area and consequently the putative packing constraints generated from the movement of lipid would be expected to have a smaller influence. The possibility that PG asymmetry is limited by packing constraints was examined in the following manner. It was argued that the putative packing constraints could be relieved by the transport of a membrane bound weak base (stearylamine) which is expected to be transported in the opposite direction to the weak acid. The influence of SA on the extent of PG transport was examined in LUVs composed of EPC:SA:Chol:EPG (2:2:4:2 mol/mol) . Cholesterol was included in the LUVs in order to increase the stability of the pH gradient and to help maintain the asymmetry of the periodate assay (data not shown). The results of Fig 27a demonstrate the time course for periodate oxidation of. EPGChoLEPG (4:4:2 mol/mol) LUVs which have had a prior incubation with a pH gradient (pH 4.0 inside / 8.0 outside) for 30 min at either 25°C or 60°C. In the absence of heating, the vesicles exhibit no PG asymmetry which is consistent with the observations presented in Fig. 26a. However, an incubation of the LUVs at at 60°C for 30 min caused transport of EPG from the IM to the OM. However, the inclusion of SA in these model membrane systems had no apparent influence on the extent of PG asymmetry. The. results of Fig. 27a clearly demonstrate that the inclusion of a membrane bound weak base in the LUVs had no influence on the extent of PG asymmetry observed at equilibrium. It is interesting to directly compare the extent of asymmetry of the weak base with that of the weak acid in the model membrane system. The results of Fig. 27b demonstrate that 105 Fig. 26. Transport of EPG in LUVs in response to acidic and basic pH gradients. EPC:EPG (9:1 mol/mol) LUVs were prepared in the presence of (A) 300 mM sodium citrate (pH 4.0) or (B) 300 mM sodium phosphate (pH 8.0) and the external buffer was exchanged by standard gel filtration techniques to 200 mM sodium sulfate, 20 mM phosphate (pH 8.0) or 200 mM sodium sulfate, 20 mM citrate (pH 4.0). The vesicles were incubated either at (A) 25°C or 60°C ( A ) for 30 min. prior to the detection of PG asymmetry (see Section 4.2.5). The total PG content (•) was determined by oxidation in the presence of 10 mM sodium cholate. 0 5 10 15 20 25 30 TIME (min) 106 Fig. 27. The influence of stearylamine on the extent of PG transport. EPC:Chol:EPG (4:4:2 mol/mol) or EPC:Chol:SA:EPG (2:4:2:2 mol/mol) LUVs were prepared in 300 mM sodium citrate (pH 4.0) and the external buffer was exchanged by standard gel filtration techniques to 200 mM sodium sulfate, 20 mM phosphate (pH 8.0). Periodate oxidation of LUVs (A) was performed as described in Section 4.2.5. TNBS labelling of the SA (B) was performed as described in Section 4.2.3. The symbols represent (A ) incubation of SA containing LUVs for 30 min at 25°C, ( A ) incubation of SA containing LUVs at 60°C for 30 min, (•) incubation of EPC:Chol:EPG LUVs at 60°C for 30 min and (•) oxidation in the presence of detergent. A TIME (min) 107 the asymmetry of SA in LUVs composed of EPC:EPG:Chol:SA (2:4:2:2 mokmol) can be followed by TNBS labelling of the exposed primary amine. The presence of detergent sufficient to solubilize the vesicles causes rapid labelling of all of the available SA. In the presence of a pH gradient (pH 4.0 inside / 8.0 outside) LUVs which have had a prior incubation at 25°C for 30 min demonstrate no detectable SA labelling (data not shown but see Hope and Cullis, 1986) indicating complete SA asymmetry. A prior incubation of these LUVs for 30 min at elevated temperatures (60°C) does not change the asymmetry of SA since no SA can be labelled by TNBS under these conditions (Fig. 27b). Thus, the extent of asymmetry of SA (> 95%) is much greater than that observed for PG (75%) under apparent equilibrium conditions for both molecules. In addition, the transmembrane pH gradient may be measured by examining the distribution of radiolabeled MeAm (Chapter 2). It was observed that the vesicles of Fig. 27 had a pH gradient of greater than 100 mV or 1.6 pH units (data not shown). Thus, the extent of PG asymmetry does not appear to be limited by the degradation of the pH gradient and does not completely reflect the size of the pH gradient. This point will be discussed in more detail in Chapter 5. 4.3.6 Transbilayer asymmetry of other phospholipids which are simple weak acids The results to this point extend previous observations demonstrating that lipids which are weak acids will redistribute across LUVs in response to pH gradients (Hope and Cullis, 1987). The rate of transport and the extent of transport of the lipids is sensitive to the nature of the polar head group (see also Homan and Pownall, 1988). For example, stearic acid asymmetry (> 95%) is established across LUVs within 2 min at room temperature whereas PG asymmetry is established only slowly at 37°C and the extent of asymmetry is less than expected on the basis of redistribution of weak acids (Fig. 25). Several other phospholipids which are simple weak acids may also be transported in a similar manner. Phosphatidic acid is the most likely candidate since the polarity of the head group is less than that of PG. In addition, the rate of transbilayer transport of a short chain fluorescent analogue of PA has been shown to be pH dependent 108 Fig. 28. The influence of ApH on the apparent surface potential of vesicles composed of DOPGPA (9.5:0.5 mol/mol). Each panel represents elution profiles from a DEAE Sephacel ion exchange column. The low ionic strength elution buffer was composed of either 300 mM sucrose, 20 mM HEPES (pH 8.0) or 300 mM sucrose, 20 mM citrate (pH 4.0). The columns were eluted after fraction 11 with a high ionic strength elution buffer composed of 0.5 NaCl. Panel A represents EPGPA (9.5:0.5 mol/mol) LUVs with (•) a pH gradient (pH 8.0/4.0) which had been incubated at 25°C for 30 min. Panel B represents EPGPA (9.5:0.5 mol/mol) LUVs with (•) or without (•) a pH gradient (pH 8.0/8.0) which had been incubated at 60°C for 30 min. RECOVERY OF LIPOSOMES 100 80-60--40--20 A • • 5 10 15 20 FRACTION NUMBER 25 100 80-% 60 + RECOVERY OF LIPOSOMES 40 4 20-5 10 15 20 FRACTION NUMBER 25 109 Fig. 29. The influence of ApH on the apparent surface potential of vesicles composed of EPGPI (9.5:0.5 mol/mol). Each panel represents elution profiles from a DEAE Sephacel ion exchange column. The low ionic strength elution buffer was composed of 300 mM sucrose, mM citrate (pH 4.0). The columns were eluted after fraction 11 with a high ionic strength elution buffer composed of 0.5 NaCl. Panel A represents EPGPI (9.5:0.5 mol/mol) LUVs incubated with a pH gradient (pH 8.0/4.0) at 25°C for 30 min. Panel B represents EPC:PI (9.5:0.5 mol/mol) LUVs incubated with a pH gradient (pH 8.0/4.0) at 60°C for 30 min. 100 80-% 60 RECOVERY OF LIPOSOMES 40 20-5 10 15 20 FRACTION NUMBER RECOVERY OF LIPOSOMES 100 80--5 10 15 20 FRACTION NUMBER 110 Fig. 30. The influence of ApH on the apparent surface potential of vesicles composed of EPC:CL (9.5:0.5 mol/mol). Each panel represents elution profiles from a DEAE Sephacel ion exchange column; The low ionic strength elution buffer was composed of 300 mM sucrose, 20 mM citrate (pH 4.0). The columns were eluted after fraction 11 with a high ionic strength elution buffer composed of 0.5 NaCl. Panel A represents EPGCL (9.5:6.5 mol/mol) LUVs incubated with a pH gradient (pH 8.0/4.0) at 25°C for 30 min. Panel B represents EPGCL (9.5:0.5 mol/mol) LUVs incubated with a pH gradient (pH 8.0/4.0) at 60°C for 30 min. RECOVERY OF LIPOSOMES 100 8 0 -60 40 2 0 - -A 5 10 15 20 FRACTION NUMBER 25 100 RECOVERY OF LIPOSOMES 5 10 15 20 FRACTION NUMBER 111 (Homan and Pownall, 1988). Cardiolipin (CL) which consists of two esterified PA molecules and phosphatidylinositol (PI) which has a hydroxylated carbohydrate esterified to a PA molecule may also be transported in this fashion though the polarity of these head groups is expected to be considerably greater than that of PG and may provide a barrier to transport. A quantitative asymmetry assay is not presently available for the detection of phospholipid asymmetry for these lipids. However, the ion exchange technique (Fig. 20) can be used to screen for acidic lipid asymmetry. LUVs composed of EPC and 5 mol% of acidic phospholipid were prepared with an internal pH of 8.0 and the external buffer was exchanged for a buffer of low ionic strength (pH 4.0). The elution profiles from DEAE Sephacel columns of vesicles containing PA, PI and CL are shown in Fig. 28, Fig. 29 and Fig. 30. Vesicles with no surface charge (EPC) elute in the void volume of DEAE Sephacel columns (Fig. 20A). Vesicles containing 5 mol% PA (Fig. 28a) bind to the column under low salt conditions indicating the presence of a significant negative surface potential. After a prior incubation at 60°C for 30 min, LUVs exhibiting a pH gradient do not bind the column under low salt conditions indicating a significant reduction in surface charge. This is consistent with the transport of PA from the OM to the IM of the LUVs in response to a transmembrane pH gradient. The results of Fig. 29 (PI) and Fig. 30 (CL) demonstrate that the presence of a pH gradient and elevated temperatures does not cause a decrease in the apparent surface charge as detected by ion exchange chromatography. Given the limitations of the ion exchange technique, it is not possible to conclude unambiguously whether PI and CL are undergoing transbilayer movement. A more detailed examination of the response of these lipids to transbilayer pH gradients will require the development of a quantitative assay. 4.4 Discussion The results of this chapter provide new evidence concerning the transbilayer transport across model membrane systems of phospholipids which are simple weak acids. There are four 112 points of interest which include the rate of transport of membrane bound weak acids, the mechanism of transport of acidic phospholipids, the extent of phospholipid asymmetry observed. in response to pH gradients and the biological significance of the pH dependent transbilayer transport of acidic phospholipids. These points will be discussed in turn. To date phospholipid transport has been observed to be very slow ( t^ < 24 hrs) in model membrane systems (for reviews see Op den Kamp, 1979; Etemadi, 1980b). The exact reasons for this have not been clearly understood until recently. In recent studies, Bishop and Bell (1985) compared the rate of transport of a sulfhydryl analogue of diacylglycerol (DG) (phosphatidylthioglycerol) with the rate of transport of a sulfhydryl analogue of PA (dioleoylthioglycerol) across model membrane systems. Since, the rate of transport of the DG analogue was considerably faster than the PA analogue they concluded that the primary energetic barrier to lipid transport was the polar head grOup. Further evidence in support of the suggestion that the polar head group is the primary energetic barrier to transport has come from studies examining the rate of transport of several short chain fluorescent phospholipid analogues in SUVs (Homan and Pownall, 1988). These workers found a good correlation between the rate of transport of the analogues and the polarity of the head group as measured by retention on a reverse-phase HPLC though PA and PG were transported faster than predicted on the basis of simple polarity considerations. The work presented in this chapter supports this general conclusion. The pH dependence of transport is consistent with the idea that the polarity of the head group is the primary factor in determining the rate of transport of a membrane bound weak acid. The polarity of the head group is in turn influenced by the degree of ionization of the head group. Consequently, PC which has a quaternary amine and the amino lipids PS and PE which are zwitterionic are likely to be transported slowly due to the presence of permanently charged residues. The degree of ionization of the head group is expected to be influenced by the pH at the interface of the liposomes and the binding of cations to the acidic residues of the polar head group. In turn, this suggests that the presence of cations will influence the rate of transbilayer transport of 113 phospholipids. There is some support for this idea. In SUVs, the curvature induced asymmetry of PG can be markedly influenced by the presence of the divalent cations C o 2 + and M n 2 + added after the generation of the liposomes (Lentz et al., 1982). This suggests that the divalent cations are mediating the transport of the phospholipids across the bilayer. In addition, the presence of PA (Nayar et al., 1984; Serham et al., 1980) and the presence of CL (Smaal et al., 1987) in liposomes has been shown to markedly increase the permeability of the liposomes to C a 2 + . This suggests that the divalent cations are being transported via ion pairs with the acidic lipid. Thus, studies utilizing model membranes systems have provided evidence to suggest that the polarity of the polar head group is the major energetic barrier to transport and that it can be influenced by biologically relevant factors such as pH and divalent cation binding. The second area of discussion involves the mechanism of transport of PG in this system. In light of the previous observations concerning transport of FA (Hope and Cullis, 1987) and the results presented here, it is suggested that phospholipids which are simple weak acids will distribute in membranes according to the transmembrane pH gradient. This suggests that the permeable form of the weak acid is the protonated molecule. There is some support for this model. Homan and Pownall examined the transport of short chain fluorescent analogues of PA and found that transmembrane transport was considerable faster at acidic pH values. There is no evidence presented in this work which deals specifically with the possibility that other lipids may be involved in the transbilayer transport of PG. Two mechanisms have been previously suggested to account for the transbilayer transport of phospholipids across membranes. Both mechanisms involve the cooperative action of a number of phospholipids to account for the rapid transport of lipid across bilayers. The first mechanism suggests that lipid transport in one direction is coupled to transport in the reverse direction (flip-flop) (Bretscher, 1973) whereas the second mechanism suggests that non-bilayer structures such as lipidic particles are involved in the rapid transbilayer movement of lipid across the membrane (Cullis and de Kruijff, 1979; Gerritsen et al., 1980a; Smaal et al., 1987). Since transport of PC is 114 typically very slow in model membrane systems ( t^ < 24 hr) (for a review see Op den Kamp, 1979) it is unlikely that PG transport in this system is accompanied by transport of PC in the reverse direction. Indeed, the presence of a lipid which is transported in the reverse direction (SA) had no influence on the extent of PG asymmetry. Consequently it is not necessary to invoke the the cooperative action of a number of phospholipids to account for the rapid transport presented in this Chapter. Since PG can be transported across LUVs in response to pH gradients, it is somewhat surprising that the extent of asymmetry does not fully reflect the size of the pH gradient. This does not appear to be a kinetic effect since the maximum amount of asymmetry observed occurs at 10 min (60°C) and is stable for the time course of the experiment. The size of the pH gradient also does not appear to limit the extent of asymmetry since a pH gradient in excess of 100 mV (1.6 units) can be measured after the maximum amount of asymmetry has been observed. Finally, in a model membrane system which contained both SA and PG, the extent of asymmetry for SA is observed to be much greater than that for PG. It may be proposed that the extent of the pH dependent PG asymmetry is limited by a packing constraint imposed by the movement of the lipid. Since the size of the polar head group for SA and FA is expected to be much smaller than that of a phospholipid, it is likely that the putative packing constraint generated by transport of SA or FA across LUVs would be considerably less than for that for a phospholipid. In addition, it is unlikely that movement of PC in these model membrane systems would relieve the putative packing pressure. However, there is presently no evidence to support the possibility that packing constraints are important. For example, the presence of SA in liposomes had no influence on the extent of PG asymmetry developed in response to transmembrane pH gradients. Previous work has demonstrated that trisialoganglioside can be incorporated into preformed LUVs by a spontaneous exchange from micelles (Feigner et al., 1981). The maximum amount of trisialoganglioside incorporated asymmetrically into the OM of the LUVs is 10-12% of total lipid indicating that the OM can accept up to 20% more lipid. Moreover experiments demonstrate that the ganglioside 115 intercalates into the vesicles, does not disturb the transmembrane distribution of PC and that the vesicle integrity is maintained. This strongly suggests that membranes have a surprising capacity to accept mass transfer of lipid and indicates that it is unlikely that asymmetry of PG is being limited by packing constraints. The importance of the pH dependent "transbilayer transport of PG in model membrane systems to membrane dynamics in vivo is not clear at the present time. It is worth noting that transbilayer pH gradients have been observed across a variety of mammalian organelles including mitochondria (inside basic), lysosomes (inside acidic) and endosomes (inside acidic) (for a review see Anderson and Orci, 1988). The results presented in this chapter suggest that the passive transport of PG and PA.across biological membranes will be influenced by the existence of transmembrane pH gradients. Note that the transbilayer transport of acidic phospholipids is not only a mechanism for the establishment of transmembrane phospholipid asymmetry but will also be expected to play a role in the maintenance of transmembrane acidic phospholipid asymmetry. Thus, in the absence of any specific transport system the acidic phospholipids PG and PA are expected to be found preferentially on the basic side of. membranes exhibiting pH gradients. The relevance of the pH depndent transport of acidic phospholipids to transport of CL and PI is not presently clear. PG arid CL are found . preferentially in the inner leaflet of the plasma membrane of Clostridium butyricum (Goldfine et al., 1982; Johnston and Goldfine, 1985). A pH gradient of greater than 2 units exists across this membrane (inside basic) (H. Goldfine, personal communication) which in turn could be the driving force for lipid asymmetry. In addition, transbilayer lipid asymmetry has been reported for the inner mitochondria membrane with the majority of the simple weak acid CL found preferentially in the IM (Kreba et al., 1979). However, it is unlikely that the asymmetry of zwitterionic phospholipids such as PS, PE and PC are transported spontaneously in response to transbilayer pH gradients in vivo. Thus, the importance of this work lies in establishing that phospholipids which are simple weak acids can be transported across model membrane systems much more rapidly than previously reported. 116 STUDIES ON THE TRANSBILAYER TRANSPORT OF PHOSPHATIDYLGLYCEROL IN LARGE UNILAMELLAR VESICLES 5.1 Introduction Transbilayer asymmetry of membrane components appears to be a ubiquitous phenomenon (for a review see Etemadi, 1980a; Etemadi, 1980b; Chapter 4). One mechanism which can account for this observation involves the asymmetric incorporation of components into membranes at the time of synthesis coupled with a slow half-life of transbilayer transport. For example, this mechanism could account for the asymmetry of mammalian sphingomyelin and glycosphingolipids. In this case, asymmetry is established during synthesis of the lipid in the Golgi complex and the orientation in membranes is maintained during the lifetime of the lipid due to the slow rates of transbilayer transport (van Meer et al., 1987; Simons and Verta, 1987). However, this mechanism does not appear to account for transbilayer glycerolipid asymmetry. Glycerolipid transport across membranes involved in the biosynthesis of lipids has been shown to be very fast ( t^ of minutes) (Zilversmit and Hughes, 1977; Rothman and Kennedy, 1977). In addition, transbilayer transport of glycerolipids in non-biosynthetic membranes such as the erythrocyte plasma membrane (for example see Seigneuret and Devaux, 1984; Tilley et al., 1986) occurs sufficiently rapidly (t^2 2-24 hrs) that phospholipid asymmetry generated during lipid synthesis would be expected to be dissipated over the lifetime of the cell (105 days). Consequently, the general mechanism by which transbilayer phospholipid asymmetry is established and maintained is under current investigation (see Chapter 4). Studies examining the transbilayer transport of lipids across model membrane systems provide insight into the nature of this problem. For example, phospholipid transport across model membrane systems is typically very slow (t^2 of days) (for a review see Op den Kamp, 1979). The rate of transport appears to be dependent upon the nature of the polar headgroup (Bishop and Bell, 1985; Homan and Pownall, 1988). Recently, it has been established that simple 117 lipids including fatty acids (FA), stearylamine (SA), sphingosine (Hope et al., 1987), as well as the phospholipids, phosphatidylglycerol (PG) and phosphatide acid (PA) (Chapter 4; Hope et al., 1988) are rapidly transported in LUVs in response to pH gradients. The zwitterionic phospholipids phosphatidylserine (PS) and phosphatidylethanolamine (PE) are not transported under similar conditions. It has been suggested that the pH dependent transport of lipids across model membrane systems may be important in vivo. In particular, the pH dependence of transport may be involved in the establishment and maintenance of lipid asymmetry across biological membranes exhibiting transmembrane pH gradients. This work examines in more detail the factors which influence the pH dependent transport of the simple weak acid PG in model membrane systems. The transbilayer transport of PG is treated as an exponential decay to an observed equilibrium value. The apparent equilibrium value is less than expected for the redistribution of a simple weak acid in response to a pH gradient of greater than 2 units. In addition, the extent of asymmetry at its apparent equilibrium is influenced by the PG content of the vesicles and the presence of phosphatidic acid (PA) but not to the presence of phosphatidylserine (PS) or divalent cations. The apparent rate constant for PG transbilayer transport is influenced by temperature, pH, cholesterol content of the LUVs, the saturation and chain length of the host lipid in the LUVs and the surface charge of the vesicles. The evidence presented supports a model in which PG is transported in response to ApH across model membrane systems as the neutral molecule. 5.2 Materials and Methods 5.2.1 Materials Phospholipids were purchased from,Avanti Polar Lipids and used without further purification. Cholesterol was purchased from Sigma Chemical Co, St Louis, MO. Radiochemicals 118 were purchased from New England Nuclear. Pyranine was purchased from Molecular Probes, Eugene, Oregon. All other buffers and chemicals were of analytical grade. 5.2.2. Preparation of lysophosphatidylglycerol Monooleoylphosphatidylglycerol (MOPG) was prepared by the following procedure. Dioleoylphosphatidylglycerol (DOPG) (1 g) was dissolved to clarity in 60 ml ether containing 0.6 ml water. Phospholipase A 2 (Croatalus adamenteus, Sigma Chem. Co., St Louis, MO.) dissolved in 2.5 ml 200 mM sodium borate (pH 7.4) was added to the ether solution and the mixture was agitated for 1 hr at 30°C. The extent of the reaction was monitored by TLC using the acid solvent system described in Section 2.2.2. Upon completion, the mixture was divided into two GSA bottles and 150 ml of ether was added to each bottle. The resulting precipitate of MOPG was then concentrated by low speed centrifugation (500g). The precipitate was resuspended in a minimum volume of ethanol and further precipitations (by the addition of 200 ml of ice cold ether) were carried out until the MOPG was judged greater than 95% pure by TLC. 5.2.3 Preparation of LUVs • Vesicles were prepared in the appropriate buffer as outlined in Section 2.2.2. Transmembrane ion gradients were established across LUVs by gel filtration as outlined in Section 2.2.3 5.2.4 Transbilayer transport of PG across LUVs Vesicles prepared in 300 mM sodium EPPS (pH 9.0) were passed down a G-50 gel . filtration column equilibrated in 200 mM Na 2S0 4, 1 mM EPPS (pH 9.0). Aliquots (80 p\) of the eluant (10-20 mg / ml phospholipid) were then added to Pyrex tubes which were incubated at 119 the appropriate temperature for 5 min. Typically, the transport of the PG was initiated by adding an aliquot (80 /il) of a buffer consisting of 100 mM sodium sulphate, 100 mM sodium citrate (pH 4.0) (45°C) to establish the transmembrane pH gradient. Lipid redistribution was stopped by an addition of 0.7 ml of ice cold buffer consisting of 100 mM sodium acetate, 100 mM citrate (pH 6.0). These conditions inhibit the transport of PG in response to basic pH gradients for three reasons: the temperature drop reduces the rate of transport, the acetic acid acts to dissipate the pH gradient (inside basic) and the higher external pH reduces the concentration of the membrane permeable neutral form of PG. The samples were then refrigerated (4°C) until analyzed for PG asymmetry. The extent of PG asymmetry is stable at 4°C for at least 24 hr (data not shown). 5.2.5 Detection of transbilayer PG asymmetry Asymmetry of PG was determined by the periodate oxidation assay described in Section 4.2.5. Analysis of PG asymmetry in vesicles was performed at 25°C except for DPPC containing vesicles which were oxidized above the lipid phase transition temperature at 45°C. The PG present in the OM was determined by oxidation of the exposed PG. Oxidation was initiated by the addition of 100 y\ of freshly prepared 100 mM sodium periodate and quenched after 12 min by the addition of 100 /il of 1 M sodium arsenite in 0.5 M H 2 S0 4 . Formaldehyde produced by the oxidation of the PG was detected by the Hantzsh reaction (Section 4.2.5.). Control experiments demonstrated that a 12 min oxidation of the PG in vesicles exhibiting no pH gradient resulted in oxidation of 50% of the PG present in the sample. Control experiments also demonstrated that entrapped 3-phosphoglycerol is not oxidized under these standard conditions indicating that the periodate only has access to the outer membrane (OM) PG (see Fig. 25). 120 5.2.6 Detection of transmembrane pH gradients EPCEPG (9:1 mol/mol) LUVs were prepared in 300 mM sodium EPPS (pH 9.0) as described above except that the fluorescent pH indicator pyranine (1 mM) (Clement and Gould, 1981) was included in the buffer. The external buffer was then exchanged for 150 mM Na 2S0 4, 1 mM EPPS (pH 9.0) using gel filtration as described above. Control experiments demonstrated that when pyranine was added to preformed vesicles it was entirely removed during gel filtration. Thus, the pyranine associated with the vesicles is entrapped and is expected to report the internal pH of the vesicles. The fluorescence response of the entrapped pyranine to changes in pH was then calibrated. Excitation was performed at 460 nm and the emission was recorded at 510 nm using a SIM-AMINCO SPF-500C spectrofluorometer. Vesicles were diluted (ca 1 mM phospholipid) in 150 mM Na 2S0 4, 20 mM EPPS, 20 mM HEPES, 20 mM PIPES, 20 mM MES which had been adjusted to a range of pH values (pH 5.0 to 9.0) with NaOH. Nigericin (final concentration of 100 nM) and gramicidin (final concentration of 10 p,g per ml) were added to the vesicle solution to facilitate equilibration of transmembrane proton gradients. A standard curve was developed calibrating the fluorescent response of the entrapped pyranine to changes in the external pH in the presence of protonionophores. The internal pH of vesicles exhibiting a transmembrane pH gradient (pH 9.0 inside / pH 4.0 outside) was monitored with time by following the fluorescence response of the entrapped pyranine and converting it to an internal pH by comparison with the calibration curve. 5.2.7 Measuring the transmembrane distribution of lysophosphatidylcholine EPCEPG (9:1 mol/mol) LUVs (10 mg phospholipid per ml) were prepared as described above except that trace amounts of [3H]-DPPC (0.1 p.C\ per ml) and [ 1 4C]-LPC (0.1 p,Ci per ml) were included with the lipid before solvent evaporation. Transmembrane pH gradients (pH 9.0 121 inside / pH 4.0 outside) were established as described above and the sample was incubated at 45°C for the indicated time period. The incubation was stopped by the addition of cold 100 mM sodium acetate, 100 mM sodium citrate, (pH 6.0) containing 50 mg per ml fatty acid-free bovine serum albumin (BSA), (Sigma Chem. Co). BSA was included in the stop buffer in order to remove the LPC from the OM of the LUVs. The vesicles were then separated from the BSA containing LPC on a sepharose 4-B gel filtration column. Control experiments demonstrated that in the absence of BSA all of the [ 1 4C]-LPC eluted with the liposomes. Recovery of liposomes on the column was typically greater than 95%. Fractions of the eluant were counted on a Packard 2000CA liquid scintillation counter to determine the amount of LPC associated with the IM of the vesicles. 5.2.8 Kinetic analysis of phosphatidylglycerol transport Weak acids are thought to be transported across lipid bilayers in response to pH gradients as the protonated (neutral) species (for example see Gutnecht and Walter, 1981b). Consequently, the net inward flux (J n e t) of the weak acid PG is expected to be a function of the concentration gradient of the neutral species, the membrane area and the permeability coefficient. Jnet - d N ( A ) 0 t 0 t A * - - P A n ([AH]Q - [AH].) (1) where N(A) o t o t is the total number of PG molecules in the exterior monolayer, [AH] is the (surface) concentration of the neutral form of the acid, P is the permeability coefficient, A m is the area of the membrane and the subscripts o and i are the OM and IM respectively. 122 From the law of conservation of mass N(A) o t o t / A m = (N(A")o + N(AH)o) / A m = [A"]o + [AH] o (2) and from the acid dissociation constant K 8 = [A- ] o [H + ] o / [AH] o (3) Rearranging equation 3 and substituting into equation 2 gives N(A) o t o t / A m = (1 + K a / [H+]o) [AH] o (4) Substituting equation 4 into equation 1 gives (dN(A)Q t o t / dt) / Am = (1 + K a / [H+]Q) d[AH]o / dt (5) If it is assumed that K a » [H +] o, [H+]j « [H +] o and that the concentration of the PG in the IM and OM are the same initially (t = 0) then (dN(A) o t o t / it) I A m = (K. / [H+]Q) d[AH]o / dt = - P [AH1 (6) Therefore: d[AHl / dt = - P [H +l [AH] / K a J 0 / o Jo ' a = - k [AH] (7) 123 where k = [H +] o P / K a (8) An analytical solution to equation 7 gives: [AH(t)]o = [AH(o)]o e"kt + C . (9) Since [AH(t)] o is proportional to [A"(t)]o: [A"(t)]o = [A"(0)]o e"kt + C (10) If it is assumed that the exponential decay is to some equilibrium value [A~(eq)]: [A>q)] =C (11) then substituting equation 11 into equation 10 gives: ([A-(t)] o - [A-(eq)]o) / [A - (0)] o = e"kt (12) In ([A-(t)] o - [A-(eq)]o) / [A"(o)]o = -k t (13) Consequently, a plot of In ([A"(t)] - [A"(eq)]o) / [A~(o)]o vs t should yield a straight line with slope k and units of t"1. This analysis is essentially the same as that used for the analysis of the spontaneous exchange of lipids between vesicles (see for example Nichols and Pagano, 1981; Brown and Thompson, 1987; Homan and Pownall, 1988). 124 Rate parameters were determined by applying a nonlinear least-square analysis to the data (see also Johnson and Frasier, 1985) using a commercially available plotting program (Sigma-Plot, Jandel Scientific, 1986). The best fit to the data was determined after evaluating the data using a one-exponential function with a constant equilibrium value. Attempts to treat the data as either a two exponential decay (two rate constants) or as a second order reaction (with respect to [PG]2) did not improve the fit of the corresponding curves to the data. Consequently all of the kinetic data was treated as a first order decay. It should be noted that the reliability of this approach is influenced by the accuracy of the determination of the equilibrium value. For the case where initial rates were fast (t^2 < 15 min) the equilibrium value could accurately be determined by examining the value at 4-5 half-lives (60 min). However, where comparisons were made between vesicles exhibiting widely differing rate constants (such as cholesterol containing systems), it was assumed that the equilibrium value for PG redistribution was constant for all the lipid preparations. This assumption was made necessary due to the inherent uncertainties of measuring equilibrium values for liposomal preparations which exhibit transport half-lives of several hours. In Section 5.3.7 it is shown that this assumption is valid for the cholesterol containing systems. The measurement of rate constants for transport across model membrane systems is plagued by the relatively high standard deviations of these measurements. For example, the permeability coefficients for proton translocation across membranes typically have standard deviations of between 30-50% of the permeability coefficients (see for example Nichols and Deamer, 1980; Perkins and Cafiso, 1986). Difficulties stem from potential differences in the lipid preparations, the presence of oxidation products, differences in the morphological characteristics of the liposomal preparations and other potential experimental artefacts. A potential serious problem derived from the above analysis for the determination of rate constants lies in the accurate determination of the equilibrium value. For example, a 5% error in the determination of the equilibrium value will result in a error of 14% in the determination of the rate constant (Johnson and Frasier, 1985). In our hands, six independent estimations of the 125 rate constant for translocation of EPG in EPC (9:1 mol/mol) using different batches of lipid over a six month period gave an estimate of 6.0 x 10"2 min"1 with a standard deviation of 8.0 x 10"3 min"1. For this reason, the data presented in Table 4 and 5 should be interpreted in terms of qualitative trends rather than as absolute measurements. The 90% confidence interval for the above experiment is 7.0 x 10"3 min"1 and consequently differences between rate constants of less than this should be interpreted with caution. 5.3 Results. 5.3.1 Kinetic analysis of transport of PG across LUVs Previous work has indicated that weak acids and bases will redistribute across LUVs in response to transmembrane pH gradients (Mayer et al., 1988; Bally et al., 1988; Hope et al., 1987; Hope et al., 1988; Chapter 3; Chapter 4). The general requirements for this transport process are that the molecules are simple weak acids or bases, that an adequate proportion are in the neutral form and that the neutral form of the molecule is reasonably lipid soluble. These general requirements can be satisfied by phospholipids such as phosphatidylglycerol (PG) and phosphatidic acid (PA) but cannot be satisfied for the zwitterionic phospholipids phosphatidylserine (PS), phosphatidylethanolamine (PE) or phosphatidylcholine (PC) (Chapter 4; Hope et al., 1988). In this Chapter, factors which influence the pH driven transport of PG in large unilamellar vesicles are characterized with the objective of determining the transport mechanism. The results of Fig. 31A illustrate that a pH gradient (inside basic) can drive the accumulation of PG into the inner monolayer (IM) of EPCEPG (9:1) LUVs. Approximately 25% of the PG in the exterior leaflet (OM) is transported within 50 min to the interior leaflet (IM). this corresponds to transport of approximately 5% of the total OM lipid to the IM under 126 Fig. 31 A. Time course of PG transport from the OM to the IM of LUVs in response to a transmembrane pH gradient. EPCEPG (9:1 mol/mol) LUVs were prepared in 300 mM EPPS (pH 9.0) and the external buffer was exchanged for 150 mM sodium sulphate, 1 mM EPPS (pH 9.0). The transport at 45°C was initiated by the addition of an aliquot of 100 mM sodium citrate (pH 4.0) and transport was stopped by adding an aliquot of cold 100 mM sodium citrate, acetate (pH 6.0). The detection of PG asymmetry was determined as described in Section 5.2.3. % AVAILABLE PG TIME (min) Fig. 3IB. A plot of the data of Fig. 31A as described in Section 5.2.8 indicating that the transport can be described by an exponential decay. In ([PG(t)]o - [PG(eq)}o) [PG(0)1 TIME (min) 127 conditions where one would not expect transport of IM lipid (EPC) to the OM. The transport of PG demonstrated in Fig. 31A is considerably slower than that previously observed for other membrane bound weak acids such as stearic acid where greater than 95% asymmetry is observed after a five min incubation at 25°C (Hope and Cullis, 1987). Presumably, the asymmetry of stearic acid is established at a faster rate than PG because of differences between the polar head groups. The transport of PG demonstrated in Fig. 31A is sufficiently rapid that it is expected to be important under physiological conditions (see also Chapter 4). The presence of a transmembrane pH gradient of greater than 2 units is expected to cause the redistribution of a weak acid such that a concentration gradient of greater than 100 fold is achieved. The degree of asymmetry may well be influenced by the location of the weak acid since the presence in the membrane would be expected to lead to a surface charge asymmetry. However, previous work (Hope et al., 1987) has indicated that greater than 95% of stearic acid is sequestered in LUVs in response to a transmembrane pH gradient of greater than 2 units (inside basic). The data detailed: in Fig. 31A demonstate that the PG transport plateaus by 50 min at an equilibrium inside-outside ratio of 4:1. Thus, the degree of asymmetry is considerably less than that expected for the transbilayer distribution of a membrane weak acid • (> 100:1) in response to a bulk pH gradient of > 2 units. The reasons for the lack.of complete . asymmetry are not presently clear but factors which influence the extent of PG transport will be discussed in Section 5.3.7. Transport of weak acids across model membrane systems is expected to occur via the neutral membrane permeable form (for example see Gutnecht and Walter, 1981b). The net flux is expected to depend upon the concentration gradient of the neutral permeating species and consequently the net flux, of a weak acid in response to a transmembrane pH gradient is expected to decay exponentially to zero. Treating the data illustrated in Fig. 31A by the kinetic analysis outlined in Section 5.2.8. using an equilibrium value ([A"(eqj]o) of 100:1 (IM:OM) did not lead to a linear relationship between In ([A"(t)] - [A"(eq)]o) / [A"(o)]o and t. However, the results illustrated in Fig. 3IB demonstrate that the initial rate of PG transport follows an 128 exponential decay if A"(eq)]o is estimated as the extent of PG transport at 4 half-lives (60 min). The rate constant of the decay (k) can be calculated from the slope of the line to be 0.06 min"1 (t-i/2 =17 min). Treating the data in this manner does not rule out the possibility that a process with a significantly slower rate constant is occurring in this system (for example see Section 5.3.2 or Section 5.3.3). 5.3.2 Stability of the basic pH gradient as detected by entrapped pyranine It is important to be able to demonstrate that the pH gradient which drives the asymmetry illustrated in Fig. 31 has not dissipated over the time course of the experiment. Experiments utilizing spin columns (see Chapter 2) for detection of vesicle associated radiolabelled acetate (Rottenberg, 1979) as a pH probe for LUVs exhibiting a pH gradient (inside basic) did not give accurate estimations of the pH gradient (data not shown). The reasons for this are not presently clear but are probably related to the leakage of acetate out of the LUVs during the column isolation procedure. Consequently, a second assay based on the pH dependence of the impermeable fluorescent dye pyranine was used for the detection of the transmembrane pH gradient (inside basic) (Elamri and Blume, 1981; Deamer and Bramhall, 1987). The fluorescent response of the dye is sensitive to the protonation state of a hydroxyl group which has a p K a of 7.2 when entrapped inside vesicles. A calibration curve relating the fluorescence response of the entrapped pyranine to the pH of the external medium (in the presence of ionophores to facilitate proton equilibration) can be established and used as an estimation of the pH gradient. The results of Fig. 32 demonstrate the time dependence of the change in the internal pH after a pH gradient (pH 9 inside / pH 4 outside) has been established. Over the 1 hr time course at 45°C the internal pH is seen to be dissipated from an initial pH of 9 to a pH of 6.5. The degradation of the pH gradient appears to be biphasic with a rapid degradation occurring during the first two minutes and a slower degradation occurring over the rest of the time course. In order for a pH gradient to be dissipated in this fashion, significant 129 -w Fig. 32. The measurement of the basic pH gradient by entrapped pyranine. EPCEPG (9:1 mol/mol) LUVs were prepared in 1 mM pyranine, 300 mM sodium EPPS (pH 9.0) and the external buffer was exchanged for 150 mM sodium sulfate, 1 mM sodium EPPS (pH 9.0). A standard curve was prepared as described in Section 5.2.6 relating the external pH (in the presence of ionophores) with the fluorescence response of the entrapped pyranine. The incubation at 45°C was initiated by adding an aliquot of 100 mM sodium citrate (pH 4.0) and the internal pH was monitored over the indicated time period. 9 T 5-4 J : 1 i | 0 30 60 90 120 TIME (min) 130 counter ion diffusion must occur in order to dissipate the established positive potential (for a discussion see Redelmeier et al., 1988; Chapter 2). For example, the presence of high concentrations of a membrane permeable weak acid in the external buffer, membrane permeable weak base in the internal buffer or the combination of high proton permeability plus a membrane permeable anion in the external buffer or a membrane permeable cation in the internal buffer would be expected to dissipate this pH gradient. Presently, the reason for the biphasic degradation of the pH gradient in LUVs exhibiting basic pH gradients is not presently understood. The important point established is that the equilibrium PG distribution observed at 60 min is not due to the degradation of the transmembrane pH gradient. This agrees with the previous observation (Chapter 4) that the pH gradient (inside acidic) measured across EPC:Chol:SA:EPG LUVs at the apparent PG equilibrium is much greater than reflected by the transmembrane PG distribution. 5.3.3 Transport of lysophosphatidylcholine in LUVs exhibiting PG asymmetry It is of interest to examine whether the transport of PG in a model membrane system will drive the transport of another lipid. This is of particular interest since it has been proposed that the net transport of PC in the plasma membrane of the mammalian erythrocyte is driven by the asymmetry of the amine containing lipids (Zachowski et al., 1987). In addition, it is of interest to directly examine whether other lipid species may be transported concurrently with PG in a mechanism analogous to the flip-flop mechanism suggested by Bretscher (1973) or to the mechanism involving the formation of non-bilayer structures (Cullis and de Kruijff, 1979; Smaal et al., 1987). Since the amount of transport can be expected to be very low under these conditions it was not possible to measure transport of EPC or an analogue since an adequately sensitive assay for monitoring EPC transport is not available. However, transport of radiolabelled lysophosphatidylcholine (LPC) can be readily monitored by comparing the amount of LPC which can be extracted from LUVs by BSA to that which cannot be removed. 131 Fig. 33A. Transport of LPC in response to PG asymmetry. E P C E P G (9:1 mol/mol) LUVs containing trace amounts of [ 3H]-DPPC (•) and [ 1 4C]-LPC ( A ) were prepared in 300 mM sodium EPPS (pH 9.0) and the external buffer was exchanged for 150 mM sodium sulfate, 1 mM sodium EPPS (pH 9.0). The transport at 45°C was initiated by adding an aliquot of 100 mM sodium citrate (pH 4.0) and the transport was stopped by adding cold 100 mM sodium acetate, 100 mM citrate (pH 6.0) containing 50 mg per ml of fatty acid free BSA. The asymmetry of LPC was determined as described in Section 5.2.7. Fig. 33A represents an elution profile of a sepharose 4-B gel filtration column used to separate LUVs from the included BSA. % Recovery of Lipid 60 4 5 - -30 15 • A 1 • • \ / -\ • • • • 5 10 15 20 COLUMN FRACTION 25 Fig. 33B. A time course of LPC transport in response to PG asymmetry. The experimental procedure is outlined in the legend of Fig. 33A. Transport in vesicles with PG asymmetry (•) and without (O) PG asymmetry is monitored. 100 % LPC Available for Extraction 150 132 The results of Fig. 32 demonstrate the elution profile of EPCEPG (9:1 mol/mol) LUVs containing radiolabeled LPC for a CL-2B gel filtration column. Radiolabeled LPC co-elutes with the liposomes as monitored by radiolabeled DPPC in the absence of BSA in the external buffer (data not shown). However, inclusion of BSA in the external buffer causes 57% of the [ 1 4C]-LPC to elute in the included volume of the column without influencing the elution profile of the EPC vesicles. This coincides with the elution profile of the BSA (data not shown). The amount of LPC that can be extracted from the vesicles is insensitive to the amount of BSA present in the external buffer indicating that all of the LPC has been removed from the OM of the vesicles, under these conditions. This suggests that the transbilayer distribution of LPC in the LUVs is 57% in the OM and 43% in the IM. The results of Fig. 33B demonstrate that the amount of LPC which can be extracted by BSA from EPCEPG (9:1 mol/mol) LUVs is not altered by the movement of PG from the OM to the IM. Over the time course of the experiment there is no significant difference in the amount of LPC which can be extracted from vesicles exhibiting PG asymmetry as compared with vesicles exhibiting no PG asymmetry. This provides indirect evidence to support the suggestion that PG is not being transported into the IM concurrently with transport of another lipid out of the IM. However, it does not directly rule out the possibility that EPC is being transported under these conditions. 5.3.4 The temperature dependence of PG transport The results to this point demonstrate that PG transport across LUVs in response to transmembrane pH gradients can be described by an exponential decay to an equilibrium distribution. The observed equilibrium PG distribution is not influenced by degradation of the pH gradient and no other lipids appear to be transported in the LUVs in response to movement of PG. Factors which influence the rate constant for PG translocation were examined to clarify the mechanism of transport. The results of Fig. 34A demonstrate the temperature dependence of PG transport. The temperature dependence is remarkable for a simple diffusion process for the 133 Fig. 34A„ The influence of temperature on PG transport. EPCEPG (9:1 mol/mol) LUVs were prepared in 300 mM sodium EPPS (pH 9.0) and the external buffer was exchanged for 150 mM sodium sulfate, 1 mM sodium EPPS (pH 9.0). The transport was initiated by adding an aliquot of 100 mM sodium citrate (pH 4.5) and the transport was stopped by adding cold 100 mM sodium acetate, 100 mM citrate (pH 6.0). The PG asymmetry was determined as described in Section 5.2.5. and analyzed as described in Section 5.2.8. Transport was followed at (•) 30°C, (A), 37°C, (•) 45°C, (•) 52°C, and (•) 59°C. In ([PG(t)] o - [PG(eq)]o) [PG(0)1 Fig. 34B An Arrhenius plot of the data given in Fig. 34A. log Rate Constant -2.0 -4.0 3.5 103 / T ( ° K _ 1 ) 134 rate increases three fold for every 7°C rise in temperature. It is unlikely that the temperature dependence is due to an influence of temperature on the p K a of the external buffer since the p K a 2 of citrate changes less than 0.04 units for a 30°C change in temperature (Lange's Handbook of Chemistry, 1979). It is less clear whether temperature will influence the p K a of PG, however the p K & x of soluble Pi is not markedly influenced by temperature, increasing 0.04 units for every 10°C rise in temperature (Lange's Handbook of Chemistry). Consequently, the observed temperature dependence of the rate constant is likely a reflection of the energetic barrier to PG transbilayer transport. The results of Fig. 34B demonstrate an Arrhenius plot of the data described in Fig. 34A. The plot of the log of the rate constant vs the reciprocal of the temperature is linear over the temperature range studied. The plot can be used to calculate an activation energy of 31 kcal mol"1 which compares favorably with that reported by Homan and Pownall (1988) (34 kcal mol" 1) for the transbilayer transport of a short chain fluorescent analogue of PG in SUVs. The significance of the high activation energy for transport of PG across bilayers will be discussed in Section 5.3.6. 5.3.5 The pH dependence of PG transport Weak acids and bases are thought to permeate across membranes as the neutral species (Hope and Cullis, 1987). Nonetheless, there have been reports that weak bases can cross model membranes as the charged species (Mayer et al., 1985; Mayer et al., 1988; Bally et al., 1988). Furthermore, it has been suggested that FA can act as protonionophores in model membrane systems, which suggests that they can move across the membrane in both the charged as well as the neutral form (Gutnecht and Walter, 1981; Gutnecht, 1987b). In turn, this may be due to the ability of anions to more readily permeate across model membrane systems than cations of similar size and charge (Flewelling and Hubbell, 1986b). Examining the pH dependence of the PG transport across LUVs should distinguish between the two mechanisms of transport. One 135 Fig. 35A. The influence of external pH on PG transport. EPCEPG (9:1 mol/mol) LUVs were prepared in 300 mM sodium EPPS (pH 9.0) and the external buffer was exchanged for 200 mM sodium sulfate, 1 mM sodium EPPS (pH 9.0). The transport was initiated by adding an aliquot of 100 mM sodium citrate which had been adjusted to the indicated pH. The transport was stopped by adding cold 100 mM sodium acetate, 100 mM citrate (pH 6.0). The PG asymmetry was determined as described in Section 5.2.5. and analyzed as described in Section 5.2.8. Transport was followed at pH 5.0 (•), 4.5 (A) , 4.0 (•), 3.5 (T) and 3.0 (•). In ([PG(t)]o -[PG(eq)]o) [PG(0)L Fig. 35B A log plot of the data given in Fig. 35A. log Rate Constant E X T E R N A L pH 136 would expect that the rate of transport of the PG would be markedly influenced by the pH of the external buffer if the PG is transported as the neutral species since the rate constant (k) is expected to be proportional to [H +] o (see equation 8 of Section 5.2.8.). However, transport of the charged species would not be expected to be influenced by the pH of the external medium if transport occurred via the charged form. The results of Fig. 35A demonstrate the influence of the external pH on the rate of transport of PG. The rate of transport appears to be inversely proportional to the pH of the external medium. For example, decreasing the external pH from 4 to 3 increases the rate of transport 10 fold which is to be expected if transport occurs via the neutral species. The results of Fig. 35B demonstrate that there is a linear relationship between the log of the rate constant for transport and the pH of the external medium. Thus, the pH dependence of the transport is consistent with the transport of the neutral protonated species. The relatively high permeability coefficient observed for transport of the neutral species in comparison with that for the charged species presumable reflects the large energetic barrier for transport of charged compounds. 5.3.6 The influence of lipid composition on PG transport A model describing the transport of nonionic solutes across lipid bilayers must take into account four independent steps; permeation through the aqueous phase including any unstirred layer, penetration of the interface, permeation across the hydrocarbon and a subsequent permeation through the opposite interface (for example see de Gier et al., 1971; Diamond and Katz, 1969). For the.case of transport of a phospholipid the model may be simplified to the latter three steps. The high activation energy for translocation of PG is consistent with the rate determining step being the penetration of the interface by the polar head group since it is unlikely that simple diffusion across a bilayer would have an activation energy of 30 kcal mol"1. However, it has been previously demonstrated that transport of neutral molecules across lipid bilayers have been found to be sensitive to a number of lipid related factors. For example, 137 increasing the saturation or chain length of the lipid hydrocarbon decreases the rate of transport of small neutral molecules such as glycerol without altering the apparent activation energy for transport (de Gier et al., 1970). Inclusion of cholesterol in model membrane systems has been shown to increase the rate of transport of molecules across membranes which would normally be in the gel phase but markedly slows down the, rate of transport of neutral molecules across membranes which would normally be in the liquid-crystalline phase (Bittman and Blau, 1972; Carruthers and Melchior, 1983). Anomalously high rates of transport are observed at the main gel to liquid crystalline transition temperature (Papahadjopoulos et al., 1973). Finally, the inclusion of different polar head groups in the membrane may be expected to influence the rate of non-electrolyte transport due to changes in vesicle surface charge, hydration and to changes in the order parameter of the acyl chain though there have been no systematic studies in this regard. Consequently, it is of interest to examine the influence of different lipid compositions on the rate of transport of PG in model membrane systems. Table 4 summarizes the influence of a variety of lipid compositions on the apparent rate constant for PG transport in LUVs. Comparison of the transport rates for EPG in dipalmitoylphosphatidylcholine (DPPC; CI6) vesicles, a relatively saturated host lipid with those for an unsaturated lipid vesicles dipalmitoleoylphosphatidylcholine (DPOPC; CI6:1 A 9 ) indicates that the degree of saturation of the hydrocarbon influences the rate of PG transport. Above the main transition temperature, EPG is transported in the DPPC matrix at approximately 55% of the rate of the unsaturated lipid system DPOPC. In addition, the transport of EPG is sensitive to the length of the acyl chain of the host matrix, decreasing with increasing acyl chain length. Thus, transport of EPG in DOPC (CI8:1 A 9 ) was approximately 51% slower than the transport in the DPOPC (CI6:1 A 9 ) . Similar influences of hydrocarbon chain length and saturation have been observed for the passive diffusion of neutral molecules such as water (Fettiplace, 1978; Carruthers and Melchior, 1983), glucose (de Gier et al., 1970), and other small sugars (McElhaney, 1986) across model membrane systems. 138 Table 4. Lipid factors which influence the rate constant for PG transport. LUVs were prepared in 300 mM sodium EPPS (pH 9.0) and the external buffer was exchanged for 150 mM sodium sulfate, 1 mM sodium EPPS (pH 9.0) . The transport at 45°C was initiated by adding an aliquot of 100 mM sodium citrate (pH 4.0) unless otherwise indicated and the transport was stopped by adding cold 100 mM sodium acetate, 100 mM citrate (pH 6.0). The PG asymmetry was determined as described in Section 5.2.5 and the kinetic analysis was performed as described in Section 5.2.8. The data is the mean of two independent experiments in which the rate constant was determined with no less than eight time points performed in duplicate. Rate Lipid Temp. Constant Composition (°C) (x 102) (min -1) DPPGEPG 9:1 50 5.3 DPOPGEPG 9:1 50 19.2 DOPGEPG 9:1 50 9.7 DOPGDOPG 9:1 45 5.1 DOPGMOPGDOPG 8:1:1 45 6:5 DOPGMOPG 9:1 45 6.2 EPGEPG 9:1 45 5.6 EPC:Chol:EPG 8:1:1 45 5.9 EPC:Chol:EPG 7:2:1 45 4.8 EPGChokEPG 6:3:1 45 1.9 EPC:Chol:EPG 5:4:1 45 11 EPGEPG 9:1 45 4.6 EPC:PS:EPG 8:1:1 45 5.7 EPC:PS:EPG 7:2:1 45 7.0 EPGEPG 9.5:.5 45 5.6 EPGEPG 9:1 45 6.5 EPGEPG 8:2 45 7.8 . EPGEPG 7:3 45 8.7 139 The influence of lyso phospholipids on the permeability properties of lipid vesicles has . been previously investigated (Mandersloot et al., 1975; Ralston et al., 1980). Since LPC forms micelles upon isolation, it is generally assumed that their presence in lipid bilayers will increase the permeability properties of model membrane systems (Hauser, 1987). The data in Table 4 support this suggestion, though the increase in the rate constant of PG translocation is not large. The presence of 10 mol% MOPC in DOPC vesicles significantly increases the rate of DOPG transport in DOPC by 27%. The mechanism by which LPC increases the permeability of model membrane systems is not presently understood but may be related to changes in the order of the lipid hydrocarbon. It is of interest to examine directly whether changes in the acyl chain component of the PG would alter the rate of PG translocation. The results of Table 4 demonstrate that transport of MOPG in DOPC (9:1 mol/mol) is similar to transport of DOPG in a DOPCMOPGDOPG mixture (8:1:1 mol/mol). The similarity in transport rates between the MOPG and the DOPG indicates that the primary energetic barrier to PG transport is the polar head group of the molecule rather than the acyl chain component. This agrees with previously published results indicating that the length of the acyl chain component does not markedly influence the rate of transbilayer transport of short chain fluorescent analogues of phospholipids (Homan and Pownall, 1988). The results of Table 4 also demonstrate that cholesterol strongly influences the rate of EPG translocation in EPC vesicles. The influence of cholesterol on the translocation rate is not, observed to vary in a linear fashion with the presence of cholesterol in the LUVs. Below 20 mol% cholesterol, PG translocation is reduced by a maximum of only 16%. However, increasing the cholesterol content of the LUVs above 20 mol% dramatically inhibits transport. This compares favorably with the result previously demonstrated for the inhibition of transport across model membrane systems of neutral molecules such as water (Carruthers and Melchoir, 1983), glycerol (de Gier et al., 1971) and other small molecules (Cohen, 1975a). The exact mechanism of inhibition is not presently understood. The presence of cholesterol in model 140 membrane systems is known to increase the order parameter of lipid in the liquid crystalline phase (for a review see Davis, 1983), decrease the average head group area (LeCuyer and Dervichian, 1969), increase the bilayer thickness (LeCuyer and Dervichian, 1969) and decrease the penetration of the water in the hydrocarbon (Simon et al., 1982). It is not presently clear whether the decrease in permeability that is seen when cholesterol is titrated into vesicles is a function of one or more of these material properties. In this regard the results indicating that cholesterol inhibited transport of PG in a non-linear fashion are particularly interesting. Recently, Finean and Hutchinson (1988) have reported that cholesterol influences the physical state of the DOPC MLVs (as judged by the T. of the lipid and the lipid packing dimension) in a non-linear fashion. Below 20 mol% cholesterol there is relatively little difference in these parameters whereas above 20 mol% cholesterol there is an approximately linear change in these parameters. Consequently, the results of the influence of cholesterol on the rate of PG transport parallel the physical state of the bilayer. The transport of PG in response to pH gradients is expected to be sensitive to the interfacial pH of the vesicles. This arises due to the well characterized influence of surface potential on the interfacial pH of vesicles which in turn influences the observed p K a of lipid head groups (Tsui et al., 1986; Watts et al., 1978). Titration of bovine brain PS into EPCEPG vesicles is expected to increase the negative surface charge of the vesicles in this manner (Winiski et al., 1986). PS is not expected to be transported under these conditions (Chapter 4). The results of Table 4 indicate that the presence of 20 mol% PS in the EPC matrix slightly increased the rate of PG transport. This is consistent with a slight increase in the negative surface charge of the vesicles which in turn would lead to an decrease in the interfacial pH. Consistent with this hypothesis is the observation that increasing the concentration of PG in the vesicles from 5 mol% to 30 mol% causes an increased rate of transport (Table 4). However, it should be noted that bovine brain PS is relatively unsaturated which would also be expected to increase the rate of transport. However, the results are consistent with the transport of the 141 neutral form of PG where factors which decrease the interfacial pH of the vesicles would increase the rate of transport of PG. 5.3.7 Factors which influence the extent of PG asymmetry The results to this point demonstrate that PG is transported across model membrane systems in response to transmembrane pH gradients. The transport follows an exponential decay to an equilibrium value. However, the equilibrium value for transmembrane PG asymmetry is not equal to that expected for the transmembrane distribution of a weak acid in response to a pH gradient of greater than 2 units. In Chapter 4 it is demonstrated that the weak base stearyl amine (SA) had no influence on the extent of asymmetry of PG. This suggests that packing constraints are not important in determining the extent of PG transport. The results of Table 5 examine the influence of a several other factors which may be expected to alter the PG distribution at equilibrium. Increasing the amount of PG in the vesicles from 5 mol% to 30 mol% increases the amount of PG transported from 2.7% to 7.8% of the OM phospholipid. Thus, increasing the PG content of the LUVs 6 fold increases the amount of PG transported only 3 fold. The internal to external ratio of PG which one would expect to remain constant for a given pH gradient, falls from 5:1 for the 5% PG to 2.4:1 for 30 mol% PG. This is not consistent with the simple redistribution of PG in response to a pH gradient. The reason for this behavior is not presently clear. This data does not appear to directly support the suggestion that packing constraints stoichiometrically inhibit transport since at 30 mol% PG more PG is transported than the total amount of PG in the vesicles at 5 mol%. A potential constraint on the amount of PG transported could be the generation of a surface charge asymmetry. For example, SUVs have a surface charge which is approximately linearly related to the PG content of the vesicles (below 30 mol% PG) and high ionic strength (Winiski et al., 1986). From the the data presented in Winiski et al. one can approximate that the surface potential for the 10 mol% EPG in EPC at equilibrium would be 25 mV for the IM and 5 mV 142 Table 5. Lipid factors which influence the extent of PG transport. LUVs were prepared in 300 mM sodium EPPS (pH 9.0) and the external buffer was exchanged for 200 mM sodium sulfate, 1 mM sodium EPPS (pH 9.0). The transport at 45°C or 60°C was initiated by adding an aliquot of 100 mM sodium citrate sodium sulphate (pH 4.0) and the transport was stopped by adding cold 100 mM sodium acetate, 100 mM citrate (pH 6.0) at 60 min or 30 min respectively. The PG asymmetry was determined as described in Section 5.2.5 Lipid Composition Temp. (°C) PG Transported (% of OM PG) EPCEPG 9.5:0.5 45 54 EPCEPG 9:1 45 44 EPCEPG 8:2 45 27 EPCEPG 7:3 45 26 EPCEPG 9:1 45 50 EPCP&EPG 8:1:1 45 48 EPC:PS:EPG 7:2:1 45 51 EPGPArEPG 8:1:1 45 32 DOPCDOPG 9:1 45 60 DOPGMOPG 9:1 45 60 DOPGMOPCDOPG 8:1:1 45 66 EPCEPG 9:1 45 58 EPCChol:EPG 8:1:1 60 58 EPCChokEPG 7:2:1 60 60 EPCChol:EPG 6:3:1 60 66 EPC:Chol:EPG 5:4:1 60 60 143 for the OM (Winiski et al., 1986). This is an approximation since the ionic strength of the buffers utilized in this work is slightly higher than that utilized for the work presented in Winiski et al. Inclusion of the acidic phospholipid phosphatidylserine (PS) in the LUVs would be expected to decrease the ratio of the IM:OM surface charge at equilibrium. The results of Table 5 indicate that the inclusion of 20 mol% PS had no apparent influence on the extent of transport of EPG in this system. In addition, the inclusion of divalent cations such as M n 2 + in the internal or external buffer of the LUVs (which would be expected to markedly decrease the surface charge of the vesicles) (Lau et al., 1981)) did not influence the extent of PG transport (data not shown). Consequently, this suggests that the generation of surface charge asymmetry did not influence the extent of PG transport at apparent equilibrium. Inclusion of PA (a charged lipid which would be expected to be transported under these conditions) decreased the extent of transport of PG indicating that it is not the surface charge which is important in inhibiting the transport but rather the actual transport of the molecule itself. In addition to the possible constraints imposed by changes in the packing pressure imposed by PG translocation it is of interest to directly examine whether changes in the size of the apolar moiety influence the extent of PG transport. A comparison of the extent of transport of MOPG with that of DOPG indicate that the presence of at least one altered acyl chain has no influence on the extent of PG asymmetry. This suggests that it is the polar head group rather than the acyl chain which is limiting the extent of PG transport. Finally, it is of interest to directly examine whether cholesterol influences the extent of PG transport in response to pH gradients. Since the inclusion of cholesterol in the vesicles inhibits the rate of PG transport the experiment was performed at elevated temperatures to insure that PG had reached equilibrium (see also Chapter 4). It is clear from the results presented in Table 5 that cholesterol had no influence on the extent of PG redistribution at equilibrium. 144 5.4 Discussion This chapter presents new evidence concerning the transport of phospholipids across model systems. In particular, factors which influence the pH dependent transport of PG across LUVs are examined. There are several points of interest. Transport bf PG can be treated as an exponential decay to an apparent equilibrium. The extent of transport of PG across the LUVs is considerably less than one would expect from the distribution of a weak acid in response to a pH gradient of greater than 2 units. Transport of the PG appears to occur via the uncharged protonated molecule and does not involve the complementary movement of other phospholipids. The result also indicate that the primary energetic barrier to transbilayer transport of PG appears to be the hydrophobic region of the membrane rather than the polar head group region. These points will be discussed in turn. It is perhaps not surprising that the transport of PG in response to pH gradients results in an exponential decay of the external PG to an equilibrium value. The transport of phospholipids between vesicles via monomer diffusion (Nichols and Pagano, 1981; Nichols and Pagano, 1982; Brown and Thompson, 1987) and the transport of phospholipids across model membrane systems (Homan and Pownall, 1988) is also thought to occur via a first order reaction. In addition, transbilayer flux of non-ionic solutes appears to follow an exponential decay (Fettiplace and Haydon, 1980)." Treating the transport as an exponential decay emphasizes that the transport is thought to occur by simple monomer diffusion and does not require either the flip-flop mechanism suggested by Bretscher (1971) or the mechanism involving the formation of non-bilayer phases (Cullis and de Kruijff, 1979). In both cases, the expectation is that transport of PG would be expected to drive net transport of other lipids. Since no net accumulation of LPC occurred during the transport of PG it is unlikely that these two latter mechanisms are operative. In addition, the non-bilayer preferring lipid DOPE had little influence on the rate of transport of PG (data not shown). Thus, the transbilayer transport of 145 PG across LUVs appears to occur by a mechanism involving monomer diffusion and the kinetics can be described by an exponential decay. However, the transport of PG did not decay to the equilibrium value expected for the transbilayer distribution of a weak acid in response to a ApH of greater than 2 units. The reasons for this are not presently understood. Increasing the PG content of LUVs led to an increase in the absolute amount of PG transported but to a decrease in the transmembrane PG gradient observed at equilibrium. This did not appear to be due to the establishment of a transmembrane surface potential gradient since the inclusion of PS or the presence of divalent cations on either side of the membrane did not influence the extent of PG transport. The limitation on the extent of PG asymmetry observed at high PG content of the vesicles was also observed for PA containing LUVs. This suggests that it is the transport of the polar head group which is involved in limiting the extent of PG asymmetry. Surprisingly, factors which would be expected to influence the acyl chain packing had no influence on the extent of PG transport. For example, cholesterol did not influence the extent of PG transport and MOPG was transported to a similar extent as DOPG. Consequently, the limitation of the PG asymmetry appears to be a function of transport of the polar head group rather than a function of the acyl chain packing. Examining the factors which influence the rate constant for transport of PG will hopefully clarify the mechanism of transport. The pH dependence of PG transport is consistent with a mechanism where transport occurs via the protonated form. The relatively high activation energy for PG transport is somewhat surprising. It is useful to compare the activation energy for PG transport in model membrane systems with that reported for glycerol in EPC MLVs since it appears that the phospholipid headgroup is the primary factor involved in the high activation energy. The activation energy for transport, of glycerol has been reported to be 11 kcal mol"1 (Cohen, 1975a) or 19 kcal mol"1 (de Gier et. al., 1970). Despite the discrepancy between the reported values both studies concluded that the transport of sugars across membranes occurred by a dehydrated intermediate. This conclusion was supported by the correlation between the 146 activation energy of transport for a variety of short chain alcohols and the number of hydrogen bonds that the alcohols could form (de Gier et al., 1970; Cohen, 1975a). Consequently the hydrogen bonding capacity of the polar head group could account for the relatively high activation energy observed for PG transbilayer transport. This would suggest that PA which has less hydrogen bonding capacity would have a slightly lower activation energy for transport whereas PI which has a slightly higher hydrogen bonding capacity may have a higher activation energy. However, there is presently no direct evidence to confirm this hypothesis. The discussion to this point has emphasized that PG transport across model membrane systems in response to transmembrane pH gradients occurs via the protonated form. The results presented suggest that the major energetic barrier to transport of PG across the membrane is provided by the acyl chain component of the bilayer. The relatively high activation energy may be related to the hydrogen bonding capacity of the polar head-group. This suggests that the rate determining step for transport of PG across the bilayer is penetration of the polar head group into the hydrocarbon rather than diffusion across the hydrocarbon. This is consistent with the previous studies of permeation of short chain sugars across model membrane systems (de Gier et al., 1970; Cohen, 1975a). However, the actual mechanism by which molecules penetrate the hydrocarbon is not presently understood. Two model have been discussed (for example see Diamond and Wright, 1969; Deamer and Bramhall, 1986) which stipulate either that non-ionic solutes simply dissolve in the hydrocarbon and diffuse across membranes (solubility-diffusion model) or that the molecules interact with a hydrated defect in the hydrocarbon which facilitates passage through the hydrocarbon. There is no definitive evidence presented here to distinguish between the two models and it should be emphasized that they are not mutually exclusive. However, it is difficult to reconcile the influence of the lipid factors on PG transport with the mechanism involving transport via a hydrated defect. For example, decreasing the length of the acyl chain by 2 carbons or increasing the unsaturation of the hydrocarbon significantly increased the rate of transport though there is no reason to suggest that the presence of defects would be markedly influenced. 147 SUMMARY This thesis investigates the influence of transmembrane ion gradients on the net flux across model membrane systems of several weak acids and bases of biological interest. In order to conduct these investigations an understanding of the relationship between transmembrane electrical potentials (A*) and transmembrane pH gradients (ApH) is required. It is demonstrated that an imposed A * or an imposed ApH will drive the net transport of protons (or their equivalents) across LUVs. However, transport of protons does not reach the expected electro-chemical equilibrium over the time course investigated (8 hr). In addition, A * dependent transport is decreased in the presence of external Na + ions. These observations are discussed in light of four mathematical models which describe the transport of protons across model membrane systems by transient hydrogen bonded water chains (tHBC) (Nagle, 1987). The evidence supports a model in which the rate limiting step for transport of protons is recombination of a transient hydrogen bonded water chain in one leaflet carrying an excess hydrogen with a transient hydrogen bonded water chain in the other leaflet (Deamer and Nichols, 1983). This qualitatively accounts for the observation that proton flux in response to a pH gradient of 1 unit appears to be equal to transport in response to an electrical potential of 150 mV. It has been previously demonstrated that lipophilic amines will be accumulated into LUVs in response to transmembrane electrical potentials (inside negative) (Mayer et al., 1985a; Bally et al., 1988) and transmembrane pH gradients (Nichols and Deamer, 1976; Mayer et al., 1988; Bally et al., 1988). These observations are consistent with a mechanism whereby lipophilic amines can cross membranes as both the charged and uncharged form and suggest that the amines may be able to act as protonionophores. However, recently this hypothesis was questioned by Madden and co-workers (1988) who were unable to demonstrate significant redistribution of several lipophilic amines into proteoliposomes which exhibit electrical potentials (inside negative). In addition the amines demonstrated no uncoupler like action in 148 these proteoliposomes. In Chapter 3 a model membrane system exhibiting a stable A * and no ApH is used to investigate the mechanism of transport of several lipophilic amines of biological interest. Transport at neutral pH of the ^-blocker timolol, the local anaesthetic dibucaine, the antischizophrenic agent chlorpromazine and the antineoplastic agent doxorubicin is demonstrated to occur primarily via the neutral unprotonated form rather the charged form. In addition, it is demonstrated that the amines do not act as proton ionophores except in the presence of the lipophilic anion tetraphenylboron. The ability of transmembrane pH gradients to drive transport of lipophilic weak bases (Chapter 3) and acids (Hope and Cullis, 1987) into LUVs suggests that transmembrane pH gradients would drive net transport of acidic phospholipids across model membrane systems. The results presented in Chapter 4 demonstrate that a transmembrane pH gradient (inside basic) will drive the net transbilayer transport of phosphatidylglycerol (PG) and phosphatidic acid (PA) but not phosphatidylserine (PS) into the inner monolayer of LUVs. This implies that the presence of a net charge on the polar head group of a phospholipid will prevent significant transbilayer transport. Work presented in Chapter 5 further investigates the mechanism of transport of PG. Transport of PG is demonstrated to follow a first order exponential decay to an apparent equilibrium PG redistribution. The observed equilibrium is not that expected for the redistribution of a weak acid in response to a transmembrane pH gradient of 2 units. The extent of redistribution is influenced by the PG content of the vesicles and the presence of PA but not by the presence of other acidic phospholipids, the presence of cholesterol or changes in the acyl chain of the PG. The pH dependence of the kinetics of transport is consistent with a model in which the neutral protonated form is transported. In addition, transport of PG has an. activation energy of 31 kcal mol"1. The kinetics of transport are decreased by increasing the chain length of the host lipid, increasing the saturation, the presence of cholesterol or decreasing the surface charge of the vesicles. The results are consistent with a model in which the neutral form of phospholipids which are weak acids are transported across model membrane systems in response 149 to transmembrane pH gradients. The major energetic barrier to transport of the PG across the model membrane systems is the acyl chain component of the membrane. 150 References Addanki, S., Cahill, F.D. and Sotos, J.F. (1968) J. Biol. Chem. 243, 2337. Andersen, O.S. and Fuchs, M. (1975) Biophys. J. 15, 795-830. Anderson, G.W. and Orci, L. (1988) / . Cell Biol. 106, 539-543. Azzi, A. and Scarpa, A. (1967) Biochem. Biophys. Acta 135, 1087-1088. Backer, J.M. and Dawidowicz, E.A. (1987) Nature, 327, 341-343. Bally, M.B., Hope, M.J., van Echteld, C.J.A. and Cullis, P.R. (1985) Biochem. Biophys. Acta 812, 66-76. Bally, M.B., Mayer, L.D., Loughrey, H., Redelmeier, T.E., Madden, T.D., Wong, K . M . Hope M.J. and Cullis, P.R (1988) Chem. Phys. Lipids, 47, 97-107. Bangham, A.D., Standish, M.M. and Watkins, J.D. (1965) J. Mol. Biol. 13, 238-244. Bangham, A.D. and Hill, M.W. (1986) Chem. Phys. Lipids 40, 189-206. Barchfeld, G.L. and Deamer, D.W. (1985) Biochem. Biophys. Acta 819, 161-169. Bar-on, Z. and Degani, H. (1985) Biochim. Biophys. Acta 813, 207-212. Barrenholz, Y., Amselem, S., and Lichtenberg, D. (1986) FEBS Lett., 99, 210-214. Barry, P.H. and Diamond, J.M. (1984) Physiol. Rev. 64, 763-872. Barsukov, L.T., Victorov, A.V. Vasilenko, LA. Evstigneeva, R.P and Bergelson, L.D. (1980) Biochem. Biophys. Acta 598, 153-168. Berden, J.A. Barker, R.W. and Radda, G.K. (1975) Biochem. Biophys. Acta 375, 186-208. Biegel, C M . and Gould, J.M. (1982) Biochemistry 29, 3474. Bishop, W.R. and Bell, R.M. (1985) Cell 42, 51-60. Bittman, R. and Blau, L. (1972) Biochemistry 2, 4831-4839. Blok, M.C., de Gier, J. and van Deenen, L .L .M. (1974) Biochim. Biophys. Acta 367, 202-209. Blok, M.C., van der Neut-Kok, E.C.M., van Deenen, L .L .M. and de Gier, J. (1975) Biochim. Biophys. Acta 406, 187-196. Boggs, J.M., (1987) Biochem. Biophys. Acta 906, 353-404. Bottcher, C.J.F., vam Gent, C M . and Pries, C. (1961) Anal. Chim. Acta 24, 203-204. Bresseleers, G.J.M., Goderis, H.L. and Tobback, P.P. (1984) Biochem. Biophys. Acta 772, 374-382. 151 Bretscher, M.S. (1973) Science 181, 622-629. Brown, R.E. and Thompson, T.E. (1987) Biochemistry 26, 5454-5460. Cafiso, D.S. and Hubbell, W.L. (1978a) Biochemistry 17, 187-195. Cafiso, D.S. and Hubbell, W.L. (1978b) Biochemistry 17, 3871-3788. Cafiso, D.S. and Hubbell, W.L. (1982) Biophys. J. 39, 263-272. Cafiso , D.S. and Hubbell, W.L. (1983) Biophys. J. 44, 49-57. Carruthers, A. and Melchior, D.L. (1983) Biochemistry 22, 5797-5807. Cass, A. and Finkelstein, A. (1967) J. Gen. Physiol 50, 1765-1770. Castaing, M. , Morel, F. and Lehn, J.M. (1986) J. Memb.Biol. 89, 251-267. Cevc, G.,,Watts, A. and Marsh, D. (1981) Biochemistry 20, 4955-4965. Cevc, G. (1987) Biochemistry 26, 6305-6310. Chap, H.J., Zwaal, R.F.A., Van Deenen, L .L .M. (1977) Biochem. Biophys. Acta 247, 146-164. Chapman, D. (1975) Q. Rev. Biop. 8, 185-235. Classen, J., Haest, C.W.M., Tournois, H. and Deuticke, B. (1987) Biochemistry 26, 6604-6612. Clement, N.R. and Gould, J.M. (1981) Biochemistry 20, 1534-1538. Cohen, B.E. (1975a) J. Memb. Biol. 20, 205-234. Cohen, B.E. (1975b) J. Memb. Biol. 20, 235-268. Comfurius, P. and Zwaal, R.F.A. (1977) Biochem. Biophys. Acta 488, 36-42. Connor, J. and Scroit, AQ.J. (1988) Biochemistry 27, 848-85L Cools, A.A. and Janssen, L.H.M. (1986) Exper. 42, 954-956. Conrad, M.P., and Strauss, H.L. (1985) Biophys. J. 48, 117-124. Croft, A.R. (1967) / . Biol. Chem. 242, 3352-3359. Cullis; P.R. and de Kruijff, B. (1978) Biochim. Biophys. Acta 507, 207-218. Cullis, P.R. and de Kruijff, B. (1979) Biochim. Biophys. Acta 559, 399-420. Cullis, P.R., de Kruijff, B., Hope, M.J., Verkleij, A.J.,* Nayar, R., Farren, S.B., Tilcock, C.P.S., Madden, T.D. and Bally, M.B. (1983) in Membrane Fluidity Biology, Vol 1 (Aloia, R.C. ed), Academic Press, N.Y., pp 40-83. 152 Cullis, P.R. and Hope, M.J. (1985) in Biochemistry of Lipids and Membranes (Vance, D.E. and Vance, G.E., ed) Benjamin/Cummings Publishing Co., Menlo Park, CA, pp. 25-70 . Dabadie, P., Bendriss, P., Erny, P. and Mazat, J.P. (1987) FEBS 226, 77-82. Daleke, D.L. and Huestis, W.H. (1985) Biochemistry 24, 5406-5416. Dalmark, M. and Storm, H.H. (1981) / . Gen. Physiol. 78, 349-364. Danielli, J.F. (1982) in Membranes and Transport, Vol. 1 Plenum Press pp 3-14. Davis, J. (1983) Biochem. Biophys. Acta 737, 117-171. Dawidowicz, E.A., (1987) in Current topics in Membranes and Transport, Vol. 29, Academic Press Inc. pp 175-202. de Gier, J., Mandersloot and van Deenen, L .L .M. (1968) Biochem. Biophys. Acta 150, 666-675. de Gier, J., Mandersloot, J.G., Hupkes, J.V., McElhaney, R.N. and Van Beek, W.P. (1970) Biochem. Biophys. Acta 233, 610-618. de Kruijff, B. and Wirtz, K.W.A. (1977) Biochem. Biophys. Acta 468, 318-326. de Kruijff, B. and Baken, P. (1978) Biochem. Biophys. Acta 507, 38-47. de Kruijff, B. and van Zoelen, E.J.J. (1978) Biochem. Biophys. Acta 511, 105-115. de Kruijff, B. van Zoelen, E.J.J, and van Deenen, L .L .M. (1978) Biochem. Biophys. Acta 509, 537-542. Deamer, D,W., Prince, R.C. and Croft, A.R. (1972) Biochim. Biophys. Acta 274, 323-335. Deamer, D.W. and Bangham, A.D. (1976) Biochem. Biophys. Acta 443, 629-634. Deamer, D.W. (1982) in Intracellular pH: Its Measurement, Regulation and Utilization in Cellular Functions, Alan R. Liss Ins., New York pp. 173-187. Deamer, D.W. and Nichols, J.W. (1983) Proc. Natl. Acad. Sci. 80, 165-168. Deamer, D.W. and Bramhall, J.(1986) Chem. Phys. Lipids 40, 167-188. Deamer, D.W. (1987) / . Bioenerg. Biom. 19, 457-468. Demel, R.A., and de Kruijff, B. (1976) Biochem. Biophys. Acta 457, 109-132. Demura, M. , Kamo, N., and Kobatake, Y. (1985a) Biochim. Biophys. Acta 812, 377-386. Demura, M. , Kamo, N., and Kobatake, Y. (1985b) Biochim. Biophys. Acta 820, 60-74. Demura, M. Kamo, N. and Kobatake, Y. (1987) Biochim. Biophys. Acta 903, 303-308. Devaux, P.F. and Seigneuret, M. (1985) Biochem. Biophys. Acta 822, 63-125. 153 Denkins, Y.M. and Schroit, A.J. (1986) Biochem. Biophys. Acta 862, 343-351. Diamond, J.M. and Wright, E.M. (1969) Ann. Rev. Phys. 31, 581-646. Dilger, J.P. and McLaughlin (1979) Science 206, 1196-1198. Dilger, J,P„ Fisher, L.R. and Haydon, D.A. (1982) Chem. Phys. Lipids 30, 159-176. Edidin, M. (1987) in Current topics in Membranes and Transport, Vol. 29, Academic Press Inc. pp 91-127. Eilers, M. and Schatz, G. (1988) Cell 52, 481-483. Elamri, K. and Blume, A. (1983) Biochim. Biophys. Acta 111, 22-30. Eksborg, S. (1977) J. Pharm. Sci. 67, 782-785. Etemadi, A.H. (1980a) Biochem. Biophys. Acta 604,347-422. Etemadi, A.H. (1980b) Biochem. Biophys. Acta 604, 423-475. Feigner, P.L., Freire, E. Barenholz, Y. and Thompson, T.E. (1981) Biochemistry 20, 2168-2172. Ferrel, J.E. Lee, K.J. , Jr. and Huestis, W.H. (1985) Biochemistry 24, 2849-2857. Fettiplace, R., Gordon, L.G.M., Hladky, S.B., Requena, J., Zingshen, H.P. and Haydon, D.A. (1974) in Methods in Membrane Biology, Vol. 4 (Korn, E.D. ed.) New York, pp 1-75. Fettiplace, R., (1978) Biochem. Biophys. Acta 513, 1-10. Fettiplace, R. and Haydon, D.A. (1980) Phys Rev. 60, 510-550. Finkelstein, A. and Cass, A. (1968) / . Gen. Physiol. 52, 145s-172s. Finkelstein, A. (1976) / . Gen. Physiol 68, 127-135. Finean, J.B. and Hutchinson, A.L. (1988) Chem. Phys. Lipids 46, 63-71. Fioler, J.W.T., Bakker, E.P. and van Dam, K. (1974) Biochim. Biophys. Acta 368, 432-445. Fisk, C H . and Subborow, Y. (1925) J. Biol. Chem. 66, 375. Flewelling, R.F. and Hubbell, W.L. (1986a) Biophys. J. 49, 531-540. Flewelling, R.F. and Hubbell, W.L. (1986b) Biophys. J. 49, 541-552. Frezatti, W.A., Toselli, W.R. and Screier, S. (1986) Biochem. Biophys. Acta 860, 531-538. Gaenssien, R.E. and McCarty, R.E. (1971) Arch Biochim. Biophys. 147, 55-56. 154 Ganong, B.R. and Bell, R.M. (1984) Biochemistry 23, 4977-4983. Garcia, M.L., Kitada, M. , Eisenstein, H.C. and Krulwich, T.A. (1984) Biochim. Biophys. Acta 766, 109-115. Garlid, K.D. and Nakashima, R.A. (1983) J. Biol. Chem. 258, 7974-7980. Gerritsen, W.J., De Kruijff, B. Verkleij, A.J., De Gier, J. and van Deenen, L .L .M. (1980a) Biochem. Biophys. Acta 598, 554-560. Gerritsen, W.J., Henricks, P.A.J., De Kruijff, B. and van Deenen, L .L .M. (1980b) Biochem. Biophys. Acta 600, 607-619. Goldfine, H., Johnston, N.C. and Bishop, D.G. (1982) Biochem. Biophys. Res. Comm. 108, 1502-1507. Goldman, D.E., (1944) / . Gen. Physiol. 27, 37-60. Goldman et al., (1978) Biochem. Biophys. Acta 512, 254-269. Gruner, S.M., Lenk, R.P., Janoff, A S and Ostro, M.J. (1985) Biochemistry 24, 2833-2842. Gruner, S. (1987) in Liposomes From Biophysics to Therapeutics (Ostro, M.J. ed) Marcel Dekker pp 1-38. Gutnecht, J. and Walter, A. (1981a) Biochem. Biophys. Acta 641, 183-188. Gutnecht, J. and Walter, A. (1981b) Biochem. Biophys. Acta 649, 149-154. Gutnecht, J. and Walter, A. (1982) Biochem. Biophys. Acta 685, 233-240. Gutnecht, J., (1984) / . Memb. Biol. 82, 105-112. Gutnecht, J., (1987a) / . Bioenerg. Biom. 19, 427-442. Gutnecht, J., (1987b) Proc. Natl. Acad Sci. 84, 6443-6446. Gutnecht, J. (1987c) Biochem. Biophys. Acta 898, 97-108. Hansen and Elkins (1971) Chem. Rev. 71, 526-616. Hartsel, S. C. and Cafiso, D.S. (1986) Biochemistry 25, 8214-8219. Hatefi, Y. (1985) Ann. Rev. Biochem. 54, 1015-1069. Hauser, H. , Oldani, D., and Philips, M.C. (1973) Biochemistry 2, 4507-4517. Hauser, H. (1987) Chem. Phys. Lipids 43, 283-299. Haydon, D.A. and Hladky, S.B. (1972) Quart. Rev. Biop. 5, 187-282. 155 Helenius, A., Mellman, I., Wall, D. and Hubbard, A. (1983) Trends Biochem. Sci. 8, 245-250. Herbette, L .G. , Katz, A .M. and Sturtevant, J. (1983) Mol. Pharmacol. 24,259-269. Herbette, L. Blaisie, J.K. Defoor, P. Fleischer, S., Bick, R.J., van Winkle, W.B., Tate, C A . and Entman, M L . (1984) Biochem. Biophys. Acta 234, 235-242. Herbette, L .G. , Chester, D.W. and Rhodes, D.G. (1986) Biophys. J. 49, 91-94. Higgins, J.A. and Piggot, C A . (1982) Biochem. Biophys. Acta 693, 151-158. Homan, R. and Pownall, H.J. (1988) Biochem. Biophys. Acta 938, 155-166. Hope, M.J., Bally, M.B., Webb, G., and Cullis, P.R. (1985) Biochim. Biophys. Acta 812, 55-65. Hope, M.J., Bally, M.B., Mayer, L.D. Janoff, A.S. and Cullis, P.R. (1986) Chem. Phys. Lipids 40, 89-108. Hope, M.J. and Cullis, P.R. (1987) / . Biol. Chem. 262, 4360-4366. Hope, M.J., Redelmeier, T.E., Wong, K.F. Rodrigueza, W. and Cullis, P.R. (1988) Biochemistry in press. Houslay, M.D. and Stanley, K .K. (1982) Dynamics of Biological Membranes, John Wiley and Sons, Toronto. Huang, C H . (1969) Biochemistry 8, 344-352. Huang, C H . (1977) Lipids 12, 348-356. Huang, C H . and Mason, J.T. (1982) in Membranes and Transport, Vol. 1 (Martonosi, A.N. ed.) Plenum Press pp. 15-24. Hui, S., (1987) in Current topics in Membranes and Transport, Vol. 29, Academic Press Inc. pp 30-70. Jacobson, K. and Papahadjopoulos, D. (1975) Biochemistry 14, 152-161. Jacobson, K. and Papahadjopoulos, D. (1976) Biophys. J. 16, 549-560. Jain, M.K. (1980) in Introduction to Biological Membranes, (Jain, M.K. and Wagner, R.C. ed) John Wiley and Sons, Toronto pp 117-142. Jain, M.K. (1983) in Membrane Fluidity in Biology, Vol 1 (Aloia, R.C. ed) Academic Press, New York pp 2-39. Johnson, S.M. (1973) Biochem. Biophys. Acta 307, 27-41. Johnson, M.L. and Frasier, S.G. (1985) Methods Enzymol. 117, 301-343. 156 Johnston, N.C. and Goldfine, H. (1985) Biochem. Biophys. Acta 813, 10-18. Jung, D.W. and Brierly, G.P. (1979) Arch. Biochim. Biophys. 193, 76-87. Kamo, N., Muratsuga, M. , Horigoh, R. and Kobatake, Y. (1979) J. Memb. Biol. 49, 105-121. Kasianowicz, J., Benz, R. and McLaughlin, S. (1987) J. Memb. Biol. 95, 73-89. Kates, M. and Sastry, P.S. (1969) Meth. Enzym. 14, 197-211. Kawashima, Y. and Bell, R,M. (1987) J.Biol. Chem. 262, 16495-16502/ Kawato, S., Kinosito, K., and Ikegani, A. (1978) Biochemistry 14, 3927-3933. Kell, D.B., and Morris, J.G. (1980) / . Biochem. Biophys. Methods 3, 143. Kirby, C. and Gregoriadis, G. (1984) in Liposome Technology Vol 1, (ed. Gregoriadis, G.) CRC Press pp 19-27. Kitagawa, T., Inoue, K. and Nojima, J. (1976) / . Biochem. 79, 1123-1133. Kleinfeld, A.M. , (1987) in Current topics in Membranes and Transport, Vol. 29, Academic Press Inc. pp 1-27. Konishi, T., Murakami, N., Hatano, Y. and Nakazato, K. (1986) Biochem. Biophys. Acta 862, 278-284. Kramer, R.M., Hasselbach, H. and Semenza, G. (1981) Biochem. Biophys. Acta 643, 233-242. Kreba, J.J.R., Hauser, H. and Carafoli, E. (1979) / . Biol. Chem. 254, 5308-5316. Kremer, J.M.H., Esker, M.W.J., Pathmamanohovan, C. and Wiersema, P.H. (1977) Biochemistry 16,3932-3935. Krishnamoorthy, G. and Hinkle, P.C. (1984) Biochemistry 23, 1640-1645. Kumar, A. and Gupta, C M . (1984) Biochem. Biophys. Acta 769, 419-420. Ladbroke, B.D., Williams, R.M. and Chapman, D. (1968) Biochem. Biophys. Acta 150, 333-340. Lange's Handbook of Chemistry (1979) (ed J.A. Dean) McGraw-Hill Book Co., Toronto, pp 5.40 -5.42. Langley, K . E . and Kennedy, E.P. (1979) Proc. Natl. Acad. Sci. 76, 6245-6249. Lau, A., McLaughlin, A. and McLaughlin, S. (1981) Biochem. Biophys. Acta 645, 279-292. Lee, H.C. and Forte, J.G. (1980) Biochem. Biophys. Acta 601, 152-166. 157 Lentz, B.R. and Litman, B.J. (1978) Biochemistry 17, 5537-5543. Lentz, B.R., Alford, D.R. and Dombrose, F.A. (1980) Biochemistry 19, 2555-2559. Lentz, B.R., Madden, S. and Alford, D.R. (1982) Biochemistry 21, 6799-6807. Lecuyer, H. and Dervichian, D.G. (1969) J. Mol. Biol 45, 39-57. Lieb, W.R. and Stein, W..D. (1971) in Current topics in membranes and transport Vol. 2 (ed Bonner F. and Kleinzeller, A.) pp 1-39. Litman, B.J. (1974) Biochemistry 12, 2545-2554. Luznat, M. and Galla, H.J. (1986) Biochim. Biophys. Acta856, 274-282. Macey, R.I. and Farmer, R.E.L. (1970) Biochem. Biophys. Acta 211, 104-106. Madden, T.D. (1986) Chem. Phys. Lipids 40, 207-226. Madden, T.D., Redelmeier, T.E. and Cullis, P.R. (1988) manuscript in preparation. Mandersloot J.G., Reman, F.C., van Deenen, L .L .M. and de Gier, J. (1975) Biochem. Biophys. Acta 382, 22-26. Massari, S. and Pozzan, T. (1976) Experientia 32, 868-869. Massari, S., Pascolini, D. and Gradenigo, G. (1978) Biochemistry 17, 4465-4469. Mayer, L.D., Bally, M.B., Hope, M.J. and Cullis, P.R., (1985a) /. Biol. Chem. 260, 802-808. Mayer, L.D., Bally, M.B., Hope, M.J. and Cullis^ P.R. (1985b) Biochim. Biophys. Acta 816, 294-302. Mayer, L.D., Hope, M.J., Cullis, P.R. and Janoff, A.S.( 1985c) . Biochim. Biophys. Acta 817, 193-196. Mayer, L.D., Bally, M.B. and Cullis, P.R., (1986a) Biochim. Biophys. Acta 857, 123-126. Mayer, L.D., Bally, M.B., Hope, M.J. and Cullis, P.R., (1986b) Biochim. Biophys. Acta 858, 161-168. Mayer, L.D., Bally, M.B. and Cullis, P.R., (1986c) Chem. Phys. Lip. 40, 333-345. Mayer, L.D., Wong, K. , Menon, K., Chong, C , Harrigan, R.H. and Cullis, P.R. (1988) Biochemistry 27, 2053-2060. McElhaney, R.N. (1986) Biochem. Cell Biol. 64, 58-65. McLaughlin, A (1982) Biochemistry 21, 4879-4885. McLaughlin, S. (1975) in Molecular Mechanisms of Anaesthesia Vol. 1 (Fink, B., Ed) Raven Press, New York, pp. 193-221. 158 McLaughlin, S. (1977; Curr. Top. Membr. Trans. 9, 71-144 McLaughlin, S.G.A. and Dilger, J.P. (1980) Pharm. Rev. 60, 825-863. Mellman, I., Fuchs, R., and Helenius, A. (1986) Annu. Rev. Biochem.. 55, 663-700. Miller, I.R. (1987) Biophys. J. 52, 497-500. Mitchell, P. (1961) Nature 191, 144-148. Mueller, P. Rudin, D.O., Tien, H.T. and Wescott, W.C. (1962) Nature 194, 979-980. Nagle, J.F. (1987) J. Bioenerg. Biom. 19, 413-426. Nash, T. (1953) Biochemical J. 55, 416-421. Nayar, R., Hope, M.J. and Cullis, P.R. (1982) Biochemistry 21, 4583-4589. Nayar, R., Mayer, L.D. Hope, M.J. and Cullis, P.R. (1984) Biochem. Biophys. Acta 777, 343-346. Nelson, J.G. (1967) J. Lipid Res. 8, 374-379. Neumcke, B. and Lauger, P. (1969) Biophys. J. 9, 1160-1170. Newton and Kluza (1978) Drug Intel!. Clin. Pharm: 12, 546-54. Nichols, J.W. and Deamer, D.W. (1976) Biochem. Biophys. Acta 455, 269-271. Nichols, J.W. and Deamer, D.W. (1980) Proc. Natl. Acad. Sci. 77, 2038. Nichols, J.W. Hill, M.W. Bangham, A.D. (1980) Biochim. Biophys. Acta 596, 393-403. Nichols, J.W. and Pagano, R.E. (1981) Biochemistry 20, 2783-2789. Nordlund, J.R. Smidt, C F . Dicken, S. and Thompson, T.E. (1981) Biochemistry 20, 3237-3241. Olson, F., Hunt, C.A., Szoka,. E . C , Vail, W.J., and Papahadjopoulos, D. (1979) Biochem. Biophys. Acta 557, 9-23. Op den Kamp, J.A.F. (1979) Annu. Rev. Biochem. 48, 47-71. Orbach, E. and Finkelstein, A. (1980) / . Gen. Phys. 75, 427-436. O'Shea, P.S., Thelen, S., Petrone, G. and Azzi, A. (1984) FEBS 172, 103-108. Pagano, R.E. and Longmuir, K.J. (1985) J. Biol. Chem. 260, 1909-1916. Papahadjopoulos, D., Jacobson, K., Nir, S., and Isac T. (1973) Biochem. Biophys. Acta 311, 330-348. Papahadjopoulos, D. and Watkins, J.C. (1967) Biochem. Biophys. Acta 135, 639-652. 159 Parsegian, V.A. (1969) Nature 221, 844-846. Perkins, W.R., and Cafiso, D.S. (1986) Biochemistry 25, 2270-2276. Perkins, W.R., and Cafiso, D.S. (1987a) J. Memb. Biol. 96, 165-173. Perkins, W.R., and Cafiso, D.S. (1987b) J. Bioenrg. Biom. 19, 443-455. Philipson, K.D. and Ward, P.R. (1985) J. Biol. Chem 260, 9666-9671. Pletcher, C.H., Resnick, R.M., Wei, G.J., Bloomfield, V.A. and Nelsestuen, G.L. (1980) J. Biol. Chem. 255, 7433-7438. Poznansky, M.S., Tong, P.C., White, J.M., Milgram, J.M. and Solomon, A.K.(1976) J, Gen. Physiol 67, 45-66. Pushkin, J.S. (1977) J. Memb. Biol. 35,39-55. Putney, J.W., Jr., Weiss, S.J., van de Walle, C M . and Haddas, R.A. (1980) Nature 284, 345-347. Ralston, E. Blumenthal, R., Weinstein, J.N. Sharrow, S.O. and Henkart, P. (1980) Biochem. Biophys. Acta 597, 543-551. Redelmeier, T.E. and Cullis, P.R. (1988) Biochemistry manuscript in preparation. Redelmeier, T.E., Mayer, M.B.,Cullis, P.R. and Bally, M.B. (1988) Biochemistry submitted. Ritchie, R.J. and Gibson, J. (1987) Arch. Biochem. Bioph. 258, 332-341. Rooney, E.K., East, J.M., McWhirter, A . C and Lee, A.G. (1983) Biochem. Biophys. Acta 728, 159-170. Rossignol, M. , Thomas, P., and Grignon, C. (1982) Biochim. Biophys. Acta 684, 195-199. Rothman, J.E. and Kennedy, E.P. (1977) Proc. Natl. Acad. Sci. 74, 1821-1825. Rottenberg, H., Grunwald, T. and Avron, M. (1972a) Eur. J. Biochem. 25, 54-63. Rottenberg, H.and Grunwald, T. (1972b) Eur. J. Biochem. 25, (71-74). Rottenberg, H. (1973) J. Memb. Biol. 11, 117. Rottenberg, H. (1975) Bioenergetics 7, 61-74. Rottenberg, H. (1979) m Methods in Enzymology Vol. LV, pp. 547-569. Schneider, E. , Haest, C.W.M., Plasa., G. and Deuticke, B. (1986a) Biochem. Biophys. Acta 855, 325-336. Schneider, E. , Haest, C.W.M. and Deuticke, B. (1986b) FEBS 198, 311-314. 160 Schroit, A.J., Madsen, J. and Ruoho, A.E. (1987) Biochemistry 26, 1812-1819. Schuldiner, S. and Kaback, (1975) Biochemistry 14, 5451-5458. Schuldiner, S. Fishkes, H. and Kanner, B.T. (1978) Proc. Natl. Acad. Sci. 75, 3713-3716. Seelig, J. (1977) Biochem. Biophys. Acta 10, 353-418. Seelig, J. (1978) Biochem. Biophys. Acta 515, 105-140. Seigneuret, M. and Devaux, P.F. (1984) Proc. Natl. Acad. Sci. 81, 3751-3755. Senisterra, G.A., Disalvo, E.A. and Gagliardino, J.J. (1988) Biochem. Biophys. Acta 941, 264-270. Serhan, C., Anderson, P., Goodman, E. , Dunham, P. and Weissmann, G. (1980) J.Biol. Chem. 256,2736-2741. Shaw, J.M. and Thompson, T.E. (1982) Biochemistry 21, 920-927. Sheetz, M.P. and Chan, S,I, (1972) Biochemistry 11, 4573-4581. Silvius, J.R. (1982) in Lipid-Protein Interactions, (Jost, P.C. and Griffith, O.H. eds.) John Wiley and Sons, pp. 239-280. Simon, S., Mcintosh, T.J. and LaTorre, R. (1982) Science 216, 65-67. Simons, K. and Virta, H. (1987) EMBO J. 6, 2241-2247. Singer, S.J. and Nicholson, G.L. (1972) Science 175, 720-731. Sleight, R.G. and Pagano, R.E. (1984) / . Cell Biol. 99, 742-751. Smaal, E.B., Schreuder, C , van Baal, J.B,, Tijburg, P.N.M., Mandersloot, J.G., de Kruijff, B. and de Gier, J. (1987) Biochem. Biophys. Acta 897, 191-196. Small, D.M. (1985) in The Physical Chemistry of Lipids from Alkanes to Phospholipids, Plenum Press, New York. Sone, N., Yoshida, M. , Hirata, H., Okamoto, H. and Kagawa, K. (1976) /. Memb. Biol. 30, 121-134. Steels, P.S. and Boulpaep, E.L. (1987) J. Memb. Biol. 100, 165-182. Sune, A., Bette-Bobillo, P., Bienvenue, A., Fellman, P., P. and Devaux, P.F. (1987) Biochemistry 26, 2972-2978. Szoka, F. and Papahadjopoulos, D. (1980) Ann. Rev. Biophys. Bioeng. 9, 467-508. Szoka, F. and Papahadjopoulos, D. (1978) Proc. Natl. Acad. Sci. 79, 4194-4198. Tanford, C. (1980) in The Hydrophobic Effect: Formation of Micelles and Biological Membranes, John Wiley and Sons, New York. 161 Thompson, T.E. and Tillack, T.W. (1985) Annu. Rev. Biophys. Biophys. Chem. 14, 631-686. Tilcock, C.P.S. (1986) Chem. Phys. Lipids 40, 109-126. Tilcock, C.P.S. and Cullis, P.R. (1987) Annals N.Y. Acad. Sci. 492, 88-102. Tilley, L . , Cribier, S., Roelofsen, B., Op den Kamp, J.A.F. and van Deenen, L .L .M. (1986) FEBS Lett. 194, 21-27. Tournois, H. Leunissen-Bijvelt, J., Haest, C.W.M., de Gier, J. and de Kruijff, B. (1987) Biochemistry 26, 6613-6621. Toyoshima, Y., and Thompson, T.E. (1975) Biochemistry 14, 1525-1531. Tsui, F . C , Ojicius, D.M. and Hubbell, W.L. (1986) Biophys. J. 49, 459-468. van den Besselaar, A.M.H.P., de Kruijjf, B., van den Bosch, H. and van Deenen, L .L .M. (1978) Biochem. Biophys. Acta 510, 242-255. van Deenen, L .L .M. and de Gier, J. (1974) in The Red Blood Cell (D. Surgenor ed ) Academic Press pp. 147-213. van Deenen, L .L .M. (1981) FEBS Lett. 123, 3-15. van Duijn, G. , Luiken, J., Verkleij, A.J. and de Kruijff, B. (1986) Biochem. Biophys. Acta 863, 193-204. van Meer, G. and Simons, K. (1987) J. Cell Biol. 105, 1623-1635. Walter, A. and Gutknecht, J. (1984) J. Memb. Biol 77, 255-264. Walter, A. and Gutknecht, J. (1986) / . Memb. Biol 90, 207-217. Watts, A., Harlos, K. , Maschke, W. and Marsh, D. (1978) Biochem. Biophys. Acta 510, 63-74. Westman, J., Boulanger, Y., Ehrenberg, A. and Smith, I.C.P. (1982) Biochem. Biophys. Acta 685, 315-328. Wickner, W. (1983) Annu Rev. Biochem. 52, Winiski, A.P., McLaughlin, A ; C , McDaniel, R.V., Eisenberg, M. and McLaughlin, S. (1986) Biochemistry 25, 8206-8214. Yoshikawa, K. , and Terada, H. (1981) / . Am. Chem. Soc. 103, 7788-7790. Zachowski, A., Fellman, P. and Devaux, P.F. (1985) Biochem. Biophys. Acta 815, 510-514. Zachowski, A., Favre, E. , Cribier, S., Herve, P. and Devaux, P.F. (1986) Biochemistry 25, 2285-2590. 162 Zachowski, A., Herrmann, A., Paraf, A., and Devaux, P.F. (1987) Biochem. Biophys. Acta 897, 197-200. Zilversmit, D.B. and Hughes, M. /E . (1977) Biochem. Biophys. Acta 469, 99-110. Zwaal, R.F.A. (1978) Biochem. Biophys. Acta 515, 163-205. 163 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0098265/manifest

Comment

Related Items