Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies of the generation and function of phospholipid asymmetry Eastman, Simon J. 1992

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

Item Metadata

Download

Media
831-ubc_1992_spring_eastman_simon_j.pdf [ 1.98MB ]
Metadata
JSON: 831-1.0086810.json
JSON-LD: 831-1.0086810-ld.json
RDF/XML (Pretty): 831-1.0086810-rdf.xml
RDF/JSON: 831-1.0086810-rdf.json
Turtle: 831-1.0086810-turtle.txt
N-Triples: 831-1.0086810-rdf-ntriples.txt
Original Record: 831-1.0086810-source.json
Full Text
831-1.0086810-fulltext.txt
Citation
831-1.0086810.ris

Full Text

STUDIES OF THE GENERATION AND FUNCTION OF PHOSPHOLIPID ASYMMETRY by SIMON J. EASTMAN BSc. Biochemistry, Carleton University, 1986  A THESIS SUBMITTED IN PARTIAL FULFILMENT 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 October, 1991 © Simon Eastman, 1991  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.  (Signature)  Department of  Biochemistry  The University of British Columbia Vancouver, Canada Date  DE.6 (2/88)  November 4. 1991  ABSTRACT It is well established that biological membranes maintain an asymmetric transbilayer distribution of component molecules, including lipids. The mechanisms by which this lipid asymmetry is established and maintained are not well understood. In addition, little is known concerning the biological significance of lipid asymmetry. This thesis employs large unilamellar vesicle (LUV) model membrane systems to examine the ability of transmembrane pH gradients (ApH) to generate lipid asymmetry and investigate the consequences of lipid asymmetry in membrane fusion phenomena. The first area of investigation demonstrates that transmembrane pH gradients can influence the inter-vesicular exchange of stearylamine and oleic acid. Vesicles containing stearylamine are shown to aggregate immediately with vesicles containing phosphatidylserine and disaggregation occurs as stearylamine equilibrates between the two vesicle populations. Despite visible flocculation during the aggregation phase, vesicle integrity is maintained. It is also shown that stearylamine is the only lipid to exchange, fusion does not occur and vesicles are able to maintain a pH gradient. When stearylamine is sequestered to the inner monolayer in response to a transmembrane pH gradient (inside acidic) aggregation is not observed and diffusion of stearylamine to acceptor vesicles is greatly reduced. The ability of ApH-dependent lipid asymmetry to modulate lipid exchange is also demonstrated for fatty acids. Oleic acid can be induced to transfer from one population of vesicles to another by maintaining a basic interior pH in the acceptor vesicles. It is also shown that the same acceptor vesicles can deplete serum albumin of bound fatty acid. The second area of investigation concerns asymmetric transbilayer distributions of dioleoylphosphatidic acid (DOPA) induced by transmembrane pH gradients. A  11  fluorescent assay is developed employing 2-(p-toluidinyl)naphthalene-6-sulfonic acid (TNS) as a probe of lipid asymmetry. The kinetics of DOPA transport are shown to be consistent with the transport of the uncharged (protonated) form. Transport of the neutral species can be rapid, exhibiting half-times for transbilayer transport of approximately 25 s at 45°C. These studies also indicate that the transport of DOPA is associated with a large activation energy (28 Kcal/mol). The third area builds on the ability to generate LUVs with an asymmetric distribution of DOPA and concerns studies on the ability of lipid asymmetry to regulate 2 stimulated fusion of LUV systems. It is shown that for LUVs composed of Ca DOPC:DOPE:PI:DOPA (25:60:5:10 mol/mol) rapid and essentially complete fusion is Ca is added. + observed by fluorescent resonance energy transfer techniques when 2 Alternatively, for LUVs with the same lipid composition but when DOPA has been sequestered to the inner monolayer, due to the presence of a pH gradient (interior basic), little or no fusion is observed upon addition of Ca . It is demonstrated that the extent of 2 + induced fusion correlates with the amount of exterior DOPA. It is also shown that 2 Ca LUVs containing only 2.5 mol% DOPA, but when all the DOPA is in the outer monolayer, can be induced to fuse to the same extent and with the same initial rate as LUVs containing 5 mol% DOPA. These results strongly support a regulatory role for lipid asymmetry in membrane fusion and indicate that the fusogenic tendencies of lipid bilayers are largely determined by the properties of one monolayer.  111  1.6.2 Fusion of Model Membranes 1.6.3 Molecular Mechanism(s) of Membrane Fusion 1.7 Thesis Outline  Chapter 2 Intervesicular Exchange of Lipids Influence of Transmembrane pH Gradients 2.1 Introduction 2.2 Materials and Methods 2.2.1 Lipids and Chemicals 2.2.2 Vesicles 2.2.3 Turbidity Experiments to Monitor Vesicle Aggregation 2.2.4 Stearylamine Exchange Monitored by Ion Exchange Chromatography using DEAE-Sephacel 2.2.5 Oleic Acid Exchange 2.3 Results 2.3.1 Stearylamine Exchange 2.3.2 Effect of a Transmembrane ApH on Stearylamine Exchange 2.3.3 Fatty Acid Exchange Between Membranes 2.3.4 Exchange of Fatty Acids Between Vesicles and BSA 2.4 Discussion  .41 43 47  48 48 51 51 51 52 53 54 55 55 58 59 62 67  Chapter 3 Transbilayer Transport of Phosphatidic Acid in Response to a Transmembrane pH Gradient 71 3.1 Introduction 3.2 Materials and Methods 3.2.1 Lipids and Chemicals 3.2.2 Preparation of Large Unilamellar Vesicles 3.2.3 Induction of Transbilayer Transport of Acidic Phospholipids 3.2.4 Detection of Phosphatidyiglycerol Asymmetry by Periodate Oxidation 3.2.5 Detection of Asymmetry Using TNS 3.2.6 Measurement of the Internal pH of LUVs 3.2.7 Kinetic Analysis of Phosphatidic Acid Transport 3.3 Results 3.3.1 TNS Fluorescence Assay of Asymmetry 3.3.2 Comparison of the TNS Assay to a Chemical Assay 3.3.3 Kinetic Analysis of PA Transport 3.3.4 Influence of pH and Temperature on PA Transport 3.3.5 Transport of DOPA to the Outer Monolayer 3.3.6 Response of Various Phospholipids to a Transmembrane ApH 3.4 Discussion  V  71 73 73 73 73 74 75 75 76 79 79 80 83 85 86 89 93  Chapter 4 Influence of Lipid Asymmetry on Fusion Between Large Unilamellar Vesicles 4.1 Introduction 4.2 Materials and Methods 4.2.1 Lipids and Chemicals 4.2.2 Preparation of Large Unilamellar Vesicles 4.2.3 Detection of Fusion 4.2.4 Induction of DOPA Asymmetry 4.2.5 31 P NMR Studies 4.2.6 Freeze Fracture Electron Microscopy 4.3 Results 4.3.1 Vesicle Composition 4.3.2 Effect of Ca 2 on LUV Fusion 4.3.3 Effect of DOPA Content on LUV Fusion 4.3.4 Effect of DOPA Asymmetry on LUV Fusion 4.3.5 Fusion of LUVs with DOPA Exclusively on the Outer Monolayer. 4.3.6 Polymorphic Phase Preferences 4.3.7 Freeze-Fracture Studies of Vesicle Fusion 4.4Discussion  .96  .  96 97 97 98 98 99 101 101 102 102 103 103 106 108 110 110 114  Chapter 5 Summary  117  References  123  V  TABLE OF CONTENTS  Abstract  ii  Table of Contents  iv  List of Figures  vii  List of Tables  x  Abbreviations  xi  Acknowledgments  .xiii  Dedication  xiv  Chapter 1 Introduction  1  1.1 The Structure and Function of Biological Membranes 1.2 Model Membranes 1.2.1 Monolayers 1.2.2 Planar Bilayers 1.2.3 Liposomes 1.3 Properties of Membrane Lipids 1.3.1 Structure 1.3.2 Gel-Liquid Crystalline Phase Transitions 1.3.3 Acid-Base Properties 1.3.4 Lipid Polymorphism 1.3.4.1 Factors Affecting Lipid Polymorphism 1.3.4.2 The Function of Lipid Polymorphism in Biological Membranes 1.4 Lipid Transport and Exchange 1.4.1 Extracellular Fatty Acid Transport 1.4.2 Intracellular Fatty Acid Transport 1.5 Membrane Asymmetry 1.5.1 Methods of Detecting Lipid Asymmetry 1.5.2 Phospholipid Asymmetry in Biological Membranes 1.5.3 Lipid Asymmetry in Model Membranes 1.5.4 Biological Significance of Lipid Asymmetry 1.6 Membrane Fusion 1.6.1 Methods of Detecting Membrane Fusion 1.6.1.1 Lipid MixingAssays 1.6.1.2 Mixing of Aqueous Contents  iv  1 5 6 6 7 11 11 14 15 18 18 21 21 22 26 26 26 27 33 35 36 37 38 39  1.6.2 Fusion of Model Membranes 1.6.3 Molecular Mechanism(s) of Membrane Fusion 1.7 Thesis Outline  Chapter 2 Intervesicular Exchange of Lipids : Influence of Transmembrane pH Gradients 2.1 Introduction 2.2 Materials and Methods 2.2.1 Lipids and Chemicals 2.2.2 Vesicles 2.2.3 Turbidity Experiments to Monitor Vesicle Aggregation 2.2.4 Stearylamine Exchange Monitored by Ion Exchange Chromatography using DEAE-Sephacel 2.2.5 Oleic Acid Exchange 2.3 Results 2.3.1 Stearylamine Exchange 2.3.2 Effect of a Transmembrane ApH on Stearylamine Exchange 2.3.3 Fatty Acid Exchange Between Membranes 2.3.4 Exchange of Fatty Acids Between Vesicles and BSA 2.4 Discussion  .41 43 47  48 48 51 51 .51 52 53 54 55 55 58 59 62 67  Chapter 3 Transbilayer Transport of Phosphatidic Acid in Response to a Transmembrane pH Gradient 71 3.1 Introduction 3.2 Materials and Methods 3.2.1 Lipids and Chemicals 3.2.2 Preparation of Large Unilamellar Vesicles 3.2.3 Induction of Transbilayer Transport of Acidic Phospholipids 3.2.4 Detection of Phosphatidylglycerol Asymmetry by Periodate Oxidation 3.2.5 Detection of Asymmetry Using TNS 3.2.6 Measurement of the Internal pH of LUVs 3.2.7 Kinetic Analysis of Phosphatidic Acid Transport 3.3 Results 3.3.1 TNS Fluorescence Assay of Asymmetry 3.3.2 Comparison of the TNS Assay to a Chemical Assay 3.3.3 Kinetic Analysis of PA Transport 3.3.4 Influence of pH and Temperature on PA Transport 3.3.5 Transport of DOPA to the Outer Monolayer 3.3.6 Response of Various Phospholipids to a Transmembrane ApH 3.4 Discussion  V  71 73 73 73 73 74 75 75 76 79 79 80 83 85 86 89 93  Chapter 4 Influence of Lipid Asymmetry on Fusion Between Large Unilamellar Vesicles 4.1 Introduction 4.2 Materials and Methods 4.2.1 Lipids and Chemicals 4.2.2 Preparation of Large Unilamellar Vesicles 4.2.3 Detection of Fusion 4.2.4 Induction of DOPA Asymmetry 4.2.5 31 P NMR Studies 4.2.6 Freeze Fracture Electron Microscopy 4.3 Results 4.3.1 Vesicle Composition 4.3.2 Effect of Ca 2 on LUV Fusion 4.3.3 Effect of DOPA Content on LUV Fusion 4.3.4 Effect of DOPA Asymmetry on LUV Fusion 4.3.5 Fusion of LUVs with DOPA Exclusively on the Outer Monolayer. 4.3.6 Polymorphic Phase Preferences 4.3.7 Freeze-Fracture Studies of Vesicle Fusion 4.4Discussion  .96  .  96 97 97 98 98 99 101 101 102 102 103 103 106 .108 110 110 114  Chapter 5 Summary  117  References  123  vi  LIST OF FIGURES Figure 1 Freeze Fracture Electron Micrographs of LUVs Produced by Extrusion  10  Figure 2 The Structure of a Phospholipid and Commonly Occurring Headgroups  13  Figure 3 Gel to Liquid-Crystalline Phase Transition  16  Figure 4 Polymorphic Phase Behavior of Lipids  20  Figure 5 Fatty Acid Transport Into the Interstitial Space  24  Figure 6 Proposed Mechanisms of Fatty Acid Transport Across Cell Membranes  25  Figure 7 Phospholipid Asymmetry in Mammalian Plasma Membranes  28  Figure 8 Fluorescence Assays to Monitor Membrane Fusion  40  Figure 9 Mechanism of Membrane Fusion Procceding Via Intermediates of the Bilayer to Hexagonal H 11 Phase Transition  46  Figure 10 Mechanism of Net Acidic Lipid Transport in Response to a Transmembrane pH Gradient  50  Figure 11 Turbidity Measurements of Vesicle Aggregation  .56  Figure 12 Characterization of Vesicle Elution from DEAE-Sephacel Columns  57  Figure 13 Effect of Transmembrane ApH on Vesicle Aggregation  60  Figure 14 Effect of a Transmembrane ApH on Stearylamine Exchange  61  -  vii  Figure 15 Oleic Acid Exchange in Aggregating Systems  63  Figure 16 Oleic Acid Exchange in Non-Aggregating Systems  65  Figure 17 Exchange of Oleic Acid Between BSA and Vesicles  66  Figure 18 Standard Curve of TNS Fluorescence as a Function of DOPA Concentration in DOPC/DOPA LUVs  81  Figure 19 Influence of a Transmembrane pH Gradient on TNS Fluorescence  82  Figure 20 Comparison of the TNS Assay for Lipid Asymmetry to a Chemical Assay (Periodate Oxidation) for DOPC/DOPG LUVs 84 Figure 21 Transbilayer Transport of DOPA in Response to a Transmembrane pH Gradient (Inside Basic)  87  Figure 22 Influence of the External pH on the Rate of the Transbilayer Transport of DOPA. 88 Figure 23 Temperature Dependence of ApH Driven DOPA Asymmetry  90  Figure 24 Transport of DOPA to the Outer Monolayer in Response to a Transmembrane pH Gradient (Interior Acidic) 91 Figure 25 Effect of a Transmembrane ApH on the Transbilayer Distributions of Various Acidic Phospholipids 92 Figure 26. 2 Concentration on the Fusion of LUVs Containing Effect of Ca l0mol%DOPA  104  Figure 27 Effect of DOPA Concentration on Vesicle Fusion  105  Figure 28 Modulation of Membrane Fusion by Lipid Asymmetry  107  viii  Figure 29 Effect on Fusion of DOPA Transport to the Outer Monolayer  109  Figure 30 Polymorphic Phase Preferences of Non-Fusogenic Vesicles (no DOPA) and Fusogenic Vesicles (10 mol% DOPA) in the Absence and Presence of Excess Ca 2  112  Figure 31 Freeze-Fracture Electron Micrographs of LUVs in the Absence and Presence of Ca 2  113  Figure 32 Possible Structure of Vesicles Exhibiting an Asymmetric Distribution of Acidic Phospholipids 120  ix  LIST OF TABLES Table 1 Gel-Liquid Crystalline Phase Transition of Some Representative Lipids  x  17  ABBREVIATIONS  MLV  Multilamellar vesicle  FATMLV  Frozen and thawed multilamellar vesicle  SUV  Small unilamellar vesicle  LUV  Large unilamellar vesicle  LUVET  Large unilamellar vesicle by extrusuion techniques  CMC  Critical micellar concentration  Lipids FA  Fatty acids  SA  Stearylamine  OA  Oleic acid  PC  Phosphatidylcholine  PA  Phosphatidic acid  PE  Phosphatidylethanol amine  P1  Phosphatidylinositol  PS  Phosphatidylserine  PG  Phosphatidyiglycerol  CL  Cardiolipin  DOPE  Dioleoylphosphatidyl ethanolamine  DOPC  Dioleoylphosphatidylcholine  DPPC  Dipalmitoylphosphatidylcholine  xi  DOPA  Dioleoylphosphatidic acid  DOPS  Dioleoylphosphatidylserine  Rh-PE  N-(lissamine rhodamine B sulfonyl) dioleoylphosphatidylethanolamine  NBD-PE  N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl) dioleoylphosphatidylethanolamine Transmembrane electrochemical potential  ApH  Transmembrane pH gradient  11 Fl  Hexagonal phase  NMR  Nuclear magnetic resonance  ESR  Electron spin resonance  QELS  Quasi-elastic light scattering  HEPES  N-(2-hydroxyethyl)piperazine-N’-2-ethanesulfonic acid  MES  2-(N-morpholino)ethanesulfonic acid  EPPS  N-(2-hydroxyethyl)piperazine-N-3-propanesu1fonic acid  PIPES  Piperazine-N,N’-bis(2-ethanesulfonic acid)  TNS  2-(p-toluidinyl)naphthalene-6-sulfonic acid  ATP  Adenosine 5’-triphosphate  BSA  Bovine serum albumin  EGTA  Ethyleneglycol-bis-(3-aminoethyl ether)-N,N,N’,N’ tetraacetic acid  TX-100  Triton-x-100  xl’  ACKNOWLEDGEMENTS There are a great many people to whom I am indebted for helping me throughout my graduate studies. Firstly I would like to thank all the members of the Cullis lab. There are too many people in the lab to mention everyone, although I probably would not have made it through the past five years without a great deal of help from many of these anonymous people. Pool games and beer at the Grad centre are good remedies for almost any ailment. I would especially like to thank Mick Hope for being my co-supervisor and a constant source of knowledge and Kim Wong who is constantly helping everybody in the lab, usually all at the same time. To all the friends I have made in the department, I thank you for all the great times had in the past five years. I would like to thank my brothers David, Michael and Nigel, my sister Nicola and my mom and dad for giving me the confidence to pursue graduate studies and for all the support they have provided me throughout the years. I would also like to thank the members of my newest family for all their generosity and for making me feel so welcome. Everyone who knows me, knows that Brenda is at least half the reason that I am in a position to finish my graduate studies and I can not thank her enough for everything she has done for me. Finally, I would like to thank Pieter Cullis for all his help both in editing this thesis and for being such an extraordinary supervisor. Pieter not only provides his students with excellent ideas and support throughout their research but provides an atmosphere in which people can enjoy science and become good friends.  xli’  TO MY MOM AND DAD, MY FAMILY AND MY WIFE BRENDA  xiv  CHAPTER 1 INTRODUCTION  1.1 THE STR UCTURE AND FUNCTION OF BIOLOGICAL MEMBRANES  Biological membranes act as highly selective permeability barriers and define the boundaries of a cell or organelle. Specific membrane proteins form pores, channels or transporters to regulate the flow of ions, metabolites and other molecules between compartments, thereby controlling the intracellular environment and regulating intracellular communication. In addition, membranes are involved in communication between cells, express immunogenetic determinants and participate in a host of other functions which include processes such as procaryotic DNA replication, protein biosynthesis, protein secretion, bioenergetics and hormonal responses (Gennis, 1989; Houslay & Stanley, 1982) Biological membranes are comprised mainly of lipids and proteins. The basic lipid bilayer structure of membranes was first proposed in 1925 by Gorter and Grendel. This model was modified by Danielli and Davson in 1935 who postulated that proteins coated the surface of the lipid bilayer. To date, biological membranes are best characterized by the fluid mosaic model (Singer & Nicholson, 1972) in which proteins are either embedded in a liquid crystalline lipid bilayer (integral membrane proteins) or attached to the surface of the membrane by ionic interactions or hydrogen bonding (peripheral proteins). The liquid crystalline nature of the membrane, characteristic of all biological membranes (see Section 1.3.2), is manifested by the rapid lateral diffusion of lipids and many proteins in the plane of the bilayer and the slow transbilayer movement of these molecules. Typically, a phospholipid will have a diffusion coefficient in the  1  order of 10-8 cm /sec in a pure lipid bilayer. This corresponds to a net average velocity 2 of 2 pm/sec. For proteins in biological membranes, the lateral diffusion coefficient can range from approximately 10-10 cm /sec to values of D 2  <  10-12 cm /sec where the 2  protein is considered to be essentially immobile (Gennis, 1989). The rapid lateral diffusion of lipids in the bilayer contrasts with a very slow rate of transbilayer movement which involves the penetration of the polar lipid headgroups into the hydrocarbon interior in order to cross the membrane. The rate of the transbilayer movement of lipids is dependent on many factors such as the nature of the polar headgroup and the possible existence of specific lipid transport proteins (“flippases” or “translocases”) in certain biological membranes (see Sections 1.5.2 and 1.5.3). Thus, in metabolically active membranes such as the cytoplasmic membrane of E. coli (Donohue-Rolfe & Schaechter, 1980) or the endoplasmic reticulum (Bishop & Bell, 1985), the half-time of transbilayer lipid transport has been reported to be on the order of seconds or minutes, whereas the half-time for transbilayer lipid transport in pure phospholipid bilayers is on the order of days to months (Roseman et al., 1975; Low & Zilversmit, 1980). The lipid component provides the permeability barrier of biological membranes and an appropriate environment for the function of membrane proteins. However, there is a great diversity of lipids in biological membranes. The erythrocyte membrane contains more than 100 different molecular species (van Deenen & de Gier, 1974), for example. It is difficult to explain such diversity if the only function of lipids is to form an inert liquid-crystalline bilayer permeability barrier, which could be satisfied by a single lipid species. It has become increasingly evident that membrane lipids provide more than just a backbone structure for the membrane and in fact are involved in many of the functional aspects of membranes. These include signal transduction (Majerus et al.,  2  1987; Berridge, 1987), the production of metabolic energy (Bass, 1988), participation in biosynthetic pathways (Jackson et a!., 1984), regulation of cell growth (Spiegel & Fishman, 1987), receptor recognition (Cheresh et a!., 1987; Hoekstra, 1990), and membrane fusion (Duzgunes, 1985; Burger & Verkleij, 1990) among others. A general feature of biological membranes is the asymmetric transbilayer distribution of the constituents. Proteins maintain an absolute asymmetry in membranes  (Op den Kamp, 1979), such that every copy of a particular protein has the same orientation with respect to the bilayer. This asymmetry is achieved during biosynthesis and is preserved during the lifetime of the membrane. Protein asymmetry provides functional asymmetry for the membrane and results in the vectorial transport of solutes across the membrane (Houslay & Stanley, 1982), for example. Phospholipids also maintain an asymmetric transbilayer distribution in biological membranes, but this asymmetry is not absolute (see Section 1.5). Phospholipids exhibit a preferential distribution across the bilayer, with the choline containing lipids predominating on the outer monolayer while the aminophospholipids predominate on the inner monolayer of plasma membranes (Houslay & Stanley, 1982). In addition to maintaining an asymmetric distribution of their components, biological membranes maintain a transmembrane electrical potential (Aip) which is generated by the vectorial transport of ions across the bilayer without the offsetting movement of another ion. Furthermore, transmembrane ion gradients are present across most membranes. For example, by way of active transport (pumps) plasma membranes maintain low Na/K ratios inside cells while high Na/K ratios exist in the intercellular fluid. Typically the Na+ and K+ gradients across animal cell membranes involve a difference in concentration of 10 to 15 fold, although this value can be much greater in  3  specialized tissues such as the avian salt gland (Finean et al., 1984). Other pumps keep the intracellular [Ca ] low with respect to the extracellular medium with gradients in the 2 range of  to 1O observed across mammalian membranes (Finean et a!., 1984). In  addition, transmembrane pH gradients exist across many biological membranes including lysosomes, mitochondria, chioroplasts and endocytic vesicles (Rottenberg, 1979). The magnitude of the pH gradient across organelles can range from less than 1 unit to greater than 3 units. For example the pH gradient (tXpH  =  1 pH pH ) across rat liver 0 -  mitochondria has been determined to be in the range of 1 pH unit (+ 0.7 (Rottenberg, 1973) to  +  1.4 units (Nicholls, 1974)), whereas the pH gradient across chioroplasts under  high light conditions can be as large as 3.5 units (Rottenberg & Grunwald, 1972). -  The consistent observation of lipid asymmetry in eukaryotic plasma membranes suggests it is crucial to cell viability. The mechanism(s) by which this asymmetry is generated and maintained is a matter of debate, as are the functional consequences. It is possible that transmembrane ion gradients may play a role in these processes. In this thesis, the importance of ion gradients, notably pH gradients, in the generation of lipid asymmetry is investigated employing model membrane liposomal systems. The role of lipid asymmetry in biological processes such as lipid exchange and membrane fusion has also been studied. This chapter provides an overview of the characteristics of lipids and membranes which are relevant to the studies undertaken in this thesis. Since model membranes were exclusively used in this thesis, a brief section is provided to discuss the methods by which various model systems are produced and the characteristics of each system. The basic physical and chemical properties of lipids are discussed in Section 1.3. The physical and chemical properties of lipids are important for understanding how  4  transmembrane ion gradients can affect the transbilayer distribution of lipids and how in turn this transbilayer asymmetric distribution of lipids can affect the properties of the membrane itself. For example, neutral forms of acidic lipids are transported across the bilayer at a far greater rate than charged lipids. Remaining sections introduce background aspects of membrane biochemistry investigated in this thesis. Lipid transport and exchange is discussed as the effect of lipid asymmetry on this process is the subject of Chapter 2. The asymmetric transbilayer distribution of membrane components, specifically lipids, is a major focus of this thesis and is reviewed in Section 1.5. Finally, an overview of membrane fusion is provided in Section 1.6, which serves as an introduction to the studies on the influence of lipid asymmetry on membrane fusion, summarized in Chapter 4.  1.2 MODEL MEMBRANES  Model membranes have been developed to study the properties of pure lipids, lipid mixtures and reconstituted lipid-protein mixtures. Model membrane systems have provided much useful information about the structure and function of biological membranes since the characteristics of individual components of membranes can be determined using these systems. In addition, the environment of the model systems can be easily manipulated in order to determine the effects of specific factors. The complexity of biological membranes often precludes such studies aimed at understanding the physical properties and functional roles of individual components. Three basic types of model membranes exist. These are monolayers, planar bilayers and liposomes. These systems are briefly discussed in the following sections.  5  1.2.1 Monolayers At an air-water interface, phospholipids form an oriented monolayer with the polar portions in contact with the aqueous phase and the hydrocarbon tails extended into the air. Monolayer films can be compressed and the resistance to compression measured. Studies of the compression pressure versus the surface area (occupied by the film) provides information about the molecular packing of lipids and lipid-protein interactions. Possibly the best-known observation derived from monolayers studies is the condensation effect of cholesterol and phospholipid, where the area occupied by a typical membrane phospholipid molecule and a cholesterol molecule in a monolayer is less than the sum of their molecular areas in isolation (Demel & de Kruijff, 1976). Monolayers have also been used to study the physical chemistry of lipid headgroups and the enzymology of soluble proteins that act at the lipid-water interface (Gennis, 1989). These systems have restricted uses and are obviously not useful for studying characteristics of lipid bilayers, such as transbilayer lipid asymmetry.  1.2.2 Planar Bilayers Planar bilayers can be formed by painting a concentrated solution of lipid in an organic solvent across a small orifice in a nonpolar partition between two aqueous compartments. The solvent tends to collect at the perimeter of the orifice, leaving a bilayer film across the center (Fettiplace et al., 1974). Planar lipid bilayers have proven to be excellent systems to study pores, channels and transporters of charged molecules because they allow for electrical measurements across the bilayer through use of electrodes in the buffered compartments. Various proteins can be incorporated into the membranes if they are soluble in the solvent used to dissolve the lipid. Limitations with  6  these systems arise due to the presence of the hydrocarbon solvent which may affect the normal properties of the lipid bilayer being examined. Furthermore, the relatively small amount of bilayer present largely precludes studies on the topology of the membrane itself.  1.2.3 Liposomes A liposome is a lipid bilayer structure which encloses an aqueous volume and can either consist of multiple bilayers in a series of concentric shells, referred to as a multilamellar vesicle (MLV), or a single-walled unilamellar vesicle. MLVs, were first described by Bangham et al. (1965) and can be simply prepared by drying down a solution of lipid in organic solvent to a thin film on the wall of a vessel, followed by the addition of an aqueous buffer and agitation of the system. The vesicles produced are heterogeneous in size (0.5 10 !m) and have a low ratio of trapped aqueous -  volume to lipid  (— 0.5  L/imol) although this ratio can be greatly increased (5 10 -  L/[Lmo1) by subjecting the MLVs to freeze-thaw cycles (Mayer et al., 1985). Furthermore, only a small amount of the total lipid is present in the outer monolayer and thus exposed to the external medium (Hope et al., 1986). MLVs are very useful in studies of the structural properties of lipids such as lipid polymorphism and the factors that modulate these preferences, since the regular arrays of bilayers in MLVs are suited for X-ray studies and the relatively large size of MLVs makes structural and motional analysis by nuclear magnetic resonance (NMR) more straightforward than in smaller systems. Unilamellar vesicles are the most commonly used model membrane systems and are usually classed as small unilamellar vesicles (SUVs), with diameters ranging between  7  25 to 40 nm, and large unilamellar vesicles (LUVs) which typically have diameters of 50  to 200 nm. There are several ways to produce unilamellar vesicles. SUVs are usually produced by the sonication of MLVs to form limit size vesicles (Huang, 1969). These vesicles can also be produced using a French press (Barenholz et a!., 1979). SUV systems are comparatively simple to prepare and and are relatively homogeneous in size for defined lipid compositions. However, these systems have small interior aqueous volumes so that the trapping efficiencies obtained are poor and the small radii of curvature associated with SUVs can perturb the physical properties of the lipids being studied (Cullis et al., 1985). LUVs can be produced from organic solvents (Szoka & Papahadjopoulos, 1980) by injecting a solution of lipids in ethanol or ether into an aqueous medium and removing the organic solvent by heating the solution above the boiling point of the solvent. Alternatively the solvent can be removed by dialysis or gel filtration. Another procedure for the production of vesicles from organic solvents, the reverse phase evaporation procedure (REV), involves dissolving the lipids in an organic solvent, such as ether, and forming an emulsion with the appropriate aqueous buffer. The organic solvent is subsequently removed under vacuum resulting in a thick gel of hydrated lipid which can be diluted and sized by extrusion through polycarbonate filters (Szoka & Papahadjopoulos, 1978). Vesicles produced using reverse phase evaporation procedures can exhibit large trapping efficiences. Difficulties often occur in the preparation of vesicles from organic solvents due to the differing solubilities of lipid species in various solvents. Furthermore, residual solvents can affect the physical characteristics of the lipids.  8  A second approach to the formation of unilamellar vesicles is to dissolve lipids in a detergent and then to remove the detergent by dialysis (Mimms et al., 1981) or gel filtration. Detergents with a low critical micellar concentration (CMC) can often be removed employing hydrophobic beads (Hope et a!., 1986; Gennis, 1989). However, these vesicles are particularily useful in the study of membrane proteins since solubilized proteins can be reconstituted into the vesicles during the dialysis period. These systems have low trapping efficiencies and suffer from similar problems as the REV method, that is differential solubilities of various lipid species in detergents and the presence of residual detergents. A third method of producing LUVs is to repeatedly extrude MLVs at intermediate pressures through polycarbonate filters of defined pore size. These vesicles are sometimes referred to as LUVETs (large unilamellar vesicles by extrusion techniques). This process has many advantages over the previously mentioned methods. First, the process is very rapid, secondly, there are no residual solvents or detergents present, thirdly, a relatively high trapping efficiency is obtained ( 30%) and finally, it is usually possible to produce homogeneously sized vesicle populations. The size of the unilamellar vesicles generated by extrusion can range from approximately 50 200 nm -  depending on the pore size of the polycarbonate filters used. Larger vesicles can be prepared by this procedure but a proportion of the vesicles are multilamellar. The ability to generate homogeneous populations of vesicles by extrusion techniques is illustrated in Figure 1, where freeze-fracture electron micrographs show vesicles produced by the extrusion of egg-PC through polycarbonate filters of various pore sizes ranging from 30400 nm. The LUVs used to undertake the studies described in this thesis were all produced by the extrusion technique.  9  Figure 1 Freeze Fracture Electron Micrographs of LUVs Produced by Extrusion -  Vesicles were prepared by extruding frozen and thawed egg-PC MLVs, at a concentration of 100 mM lipid, 20 times through polycarbonate filters of various pore sizes. (A) 400 nm, (B) 200 nm (C) 100 nm, (D) 50 nm and (E) 30 nm. The bar in panel A represents 150 nm and all panels exhibit the same magnification (Mayer et al., 1986).  10  1.3 PROPERTIES OF MEMBRANE LIPIDS  Lipids have many characteristic properties which include differences in their chemical diversity, their gel-liquid crystalline phase behavior, their polymorphic phase preferences under various conditions and their acid-base characteristics among others. These topics are briefly reviewed in the following sections.  1.3.1 Structure By definition, lipids are water-insoluble biological molecules that are highly soluble in organic solvents such as chloroform. This is a rather general definition and includes an extremely diverse class of molecules. The most abundant class of lipids in most biological membranes are the phospholipids (see Figure 2). These lipids are composed of a hydrophilic head group linked to the sn 3 position of glycerol-3-phosphate via a phosphate ester. Two acyl chains are attached to the sn 1 and sn 2 positions via ester linkages. Figure 2 illustrates the diversity of polar head groups which defines the classes of phospholipid, the most common being the phosphatidyicholines (PC), phosphatidylethanolamines (PE), phosphatidylserines (PS), phosphatidylglycerols (PG), phosphatidic acids (PA), phosphatidylinositols (P1) and the diphosphatidyiglycerols or cardiolipins (CL). These head groups differ with respect to their charge, polarity, size, chemical reactivity and hydrogen bonding capacity and therefore have a large influence on the properties of lipid bilayers. For example, substitution of ethanolamine for choline produces drastic differences in the gel-liquid crystalline phase transition temperature (see Section 1.3.2) and the polymorphic phase preferences of these lipids (see Section 1.3.4).  11  Sphingolipids are another important class of membrane lipids. Ceramide, which is the precursor to all sphingolipids, is composed of a fatty acyl chain linked to the amino group of sphingosine which is an alcohol with a long-chain hydrocarbon tail. The addition of phosphocholine to the C-i hydroxyl of ceramide produces sphingomyelin which is a major component of nervous tissue and many plasma membranes but only a minor constituent of intracellular membranes. Other sphingolipids include cerebrosides, which are formed by the attachment of one or more carbohydrate groups to the C-i hydroxyl of ceramide, and gangliosides, which contain several sugar residues and one or more sialic acid residues attached to the C-i hydroxyl of ceramide. A third important class of membrane lipids are the sterols. Cholesterol, the major lipid in this class, is found almost exclusively in the plasma membranes of mammalian cells, where it can comprise up to 45 mol% of the total lipid. Cholesterol has a large apolar region and a hydroxyl group constituting the polar domain. Cholesterol is buried in the nonpolar hydrocarbon area of bilayers with the hydroxyl function exposed to the aqueous surface. In addition to these lipids, there are a number of other lipid components in membranes which comprise only minor quantities of the total lipid. These include fatty acids, lyso-phospholipids, plasmalogens and diglycerides among others. Although these lipids are only present in small quantities, they can play major roles in membrane function. For example small amounts of diacylgicerols have been shown to destabilise bilayers and induce fusion (Siegel et a!., 1989). Furthermore, phosphorylated derivatives of phosphatidylinositol (P1), a phospholipid present usually only in very small quantities in biological membranes, are important in transmembrane signalling.  12  Figure 2 The Structure of a Phospholipid and Commonly Occurring Headgroups  Phospliatychone  0  2 Ca  / jic  0  Eltianaarnine (PhospatidylehanoIamine)  223  ‘I  —  A  NK3  V  — CH 2 —CH  V  Ser.ne  (P1osphatidyiserine) • C =0  COO  —  —  —  CH(OH)CH CH O 2 H  CH2  GeroT (Dphospha{idylgIycero  ac-OH —  3 CH  Gefo (PhosphaIidygIyceioI)  01  cardohpin)  2 CH H  OH  My0n0sd0 (Phosphatidyhnositol)  OH 3 CH  H  H  13  1.3.2 Gel-Liquid Crystalline Phase Transitions Phospholipids can exist in a frozen “gel” state or a fluid “liquid-crystalline” state depending on the temperature (Figure 3A). The transition of a bilayer from the “gel” phase (crystalline) to the liquid-crystalline phase occurs at a characteristic temperature (Ta) for a specific lipid species. This transition temperature is dependent upon the nature of the lipid headgroup, the acyl chains and the environment in which the lipid is dispersed. Below the transition temperature (T the acyl chains adopt an extended alltrans configuration. As the temperature approaches T, the acyl chains start to develop “kinks” due to the 1200 rotation of C-C bonds to form gauche isomers (see Figure 3B). As the temperature rises above T the average number of gauche isomers per acyl chain increases. The “kinks” in the acyl chains decrease the length of the acyl chains and increase the distance between the individual molecules. Thus, the lipid bilayer decreases in thickness but undergoes lateral expansion during the transition from the gel to liquidcrystalline states. The acyl chains markedly effect the transition temperature (See Table 1). For example, in a given lipid species, the temperature at which the gel to liquidcrystalline phase transition occurs increases with increasing acyl chain length and is higher for saturated acyl chains than for unsaturated acyl chains. The lipid headgroup is also a major determinant of the transition temperature. Charge repulsion between adjacent negatively charged phospholipids can cause lateral expansion of the bilayer, favouring the liquid-crystalline state. Thus, the phase transition temperature for dipalmitoylphosphatidic acid (DPPA) drops from 66°C at pH 6 (1 negative charge) to 43°C at pH 12 where the PA carries two negative charges (Marsh, 1990). Divalent cations can also serve to increase the transition temperature by reducing the charge repulsion between negatively charged lipids. For example, the gel to liquid  14  crystalline phase transition temperature of 1,2-(14:0) PG increases from 26°C to 81°C upon binding of 1 Ca 2 molecule for every 2 PG molecules (Marsh, 1990). The conformation of the phospholipid headgroup can also affect the phase transition by perturbing the packing of the bilayer in the crystalline state. The significance of the ability of lipids to adopt the gel phase is unknown, although it has been suggested that lateral inhomogeneities (ie. areas of gel-state domains within a liquid-crystalline bilayer) may play roles in various physiological processes. On this note it should be mentioned that there is no evidence for gel-state lipid components in eucaryotic membranes at physiological temperatures (Cullis & Hope, 1985).  1.3.3 Acid-Base Properties The polar regions of lipids contain various ionizable groups (phosphate, carboxyl groups and amino groups) which have weak acid or base characteristics. Fatty acids, phosphatidylinositol (P1), phosphatidylserine (PS), phosphatidylglycerol (PG), cardiolipin (CL) and phosphatidic acid (PA) are acidic molecules that carry a net negative charge at physiological pH. The state of ionization of phospholipids and fatty acids in a membrane can in principle be described by an equilibrium characterized by a given dissociation constant. However, the behavior of an ionizable group in solution can differ greatly from its behavior at the surface of a membrane. The charge on the lipid molecules generates an electrostatic surface potential which causes the redistribution of cations and anions at the lipid/water interface. Thus a membrane containing acidic (negatively charged) phospholipids will attract cations to the lipid water interface, increasing the concentration of H+ ions at the interface, for example, resulting in a lower pH at the interface than in  15  Figure 3 Gel to Liquid-Crystalline Phase Transition (A) A phosphatidyicholine bilayer in the gel phase exhibiting acyl chains tilted with respect to the bilayer normal (L) undergoing a phase transition the the liquid-crystalline state (LcJ. (B) A fatty acid chain in the all trans state and the distortion of the acyl chain upon the introduction of a “kink” due to the formation of a gauche isomer (from Houslay & Stanley, 1982).  A  Acy) chain ordered  Acyl chains disordered  Polar headgroup  1 RRR R Crystalline state solid 13 L i  Liquid-crysta)line state fluid L  B  9 9  (B) First-order Kink (2G1) ...GTG..  (A) All trans  16  Table 1 Gel-Liquid Crystalline Phase Transition of Some Representative Lipids Lipid  Transition Temperature (T)I°C  Phosphatidyicholines 1,2-(16:0) 1,2-(18:0) 1-(16:0)-2-(18:0) ) 9 1-(16:0)-2-(18:lcA 1,2-(16:1 cA ) 9 1,2-(18:1 cA ) 9  41.5 54.5 49 20 -36 -19  Phosphatidylethanolamines 1,2-(18:0) ) 9 1,2-(18:lcA  74 -16  Phosphatidyiglycerols 1,2-(18:0) ) 9 1,2-(18:lcA Phosphatidylserines 1,2-(18:0) ) 9 1,2-(18:lcA  Phosphatidic Acid 1,2-(16:0) ) 9 1,2-(18:lcA  Conditions  54.5 75. -18  pH7.0 pHl.0  70. 79. -11 -10.  pH7.0 pH2.0 pH7.0 pH9.5  66 8  Note mean temperatures, Data obtained from Small, 1986; Houslay & Stanley, 1982 and Marsh, 1990. -  17  the bulk solution. This in turn is reflected by apparent pKa values that are higher for membrane associated molecules than those free in bulk solution. For example the PKa for various fatty acids in a lipid bilayer has been reported to be between 7.0 7.5 which is -  much higher than the pK of 4.8 reported for free fatty acids (Tocanne & Tessie, 1990) .The pK of the phosphate groups of the various phospholipids listed above vary from <1 to approximately 4 depending on the ionic conditions and the detection methods used. Phosphatidic acid and cardiolipin have two ionizable groups. The PKa for the two phosphate groups on cardiolipin has been reported to be -1 (Marsh, 1990) while the pK 2 for phosphatidic acid is between 8 and 9 (Tocanne and Tessie, 1990).  1.3.4 Lipid Polymorphism Liquid crystalline lipids can adopt different phases upon hydration depending on many factors such as the lipid species and its environment (Tilcock, 1986). These phases include bilayers (La), micelles and the inverted hexagonal phase (H ). Lipids capable of 11 adopting non-bilayer phases in isolation appear to be common to all biological membranes.  1.3.4.1 Factors Affecting Lipid Polymorphism Individual species of lipids found in membranes can adopt different structures (micellar, bilayer or hexagonal ) 11 depending on the nature of the lipid headgroup, the H size and degree of unsaturation of the hydrocarbon tail, temperature, pH, the degree of hydration, the ionic strength of the medium, the presence of divalent cations and the presence of other lipids or proteins (Cullis et al., 1986). The phase adopted can be predicted by the “shape” properties of the lipid in question. Briefly, lipids with a  18  headgroup that has a larger cross-sectional area than that of the hydrophobic end (“inverted cone” shape) will form micelles, lipids with approximately equal crosssectional areas for the polar and non-polar regions (“cylindrical” shape) will form bilayers and lipids with headgroups with small cross-sectional areas with respect to the hydrophobic end (“cone” shape) will adopt the hexagonal H 11 phase (see Fig. 4A). Any factor that either increases the cross-sectional area of the headgroup or hydrophobic region of the lipid will alter the effective shape of the molecule and can therefore change the phase adopted by the lipid (see Fig. 4B). For example an increase in the degree of unsaturation of the fatty acyl chain of a phospholipid will increase the effective crosssectional area of the hydrophobic tail and drive the structure towards the hexagonal phase. Similarily an increase in temperature increases the effective cross-sectional area of the hydrophobic tails by increasing the the number of gauche isomers in the acyl chains. Protonation of acidic phospholipids such as PS decreases the charge repulsion between molecules thereby reducing the effective cross-sectional area of the headgroups which serves to drive structures towards the hexagonal H 11 phase. Certain divalent cations will act to dehydrate various phospholipid headgroups, thereby reducing the effective area of the headgroup. For example PS forms a tight complex with Ca 2 that results in the formation of cochleate lipid cylinders (Papahadjopoulos et al., 1975). PA + and 2 and CL can also form a non-bilayer phase (hexagonal ) 11 in the presence of Ca H 2 (Papahadjopoulos et al., 1976; Vail & Stollery, 1979). Mg  19  Figure 4 Polymorphic Phase Behavior of Lipids (A) Polymorphic phases formed by lipids upon hydration of the lipid at concentrations above the critical micellar concentration and the corresponding shapes of the lipids. (B) Factors influencing the bilayer to hexagonal H 11 phase transition. A  LIPID  PHASE  MOIECUIAR SHAPE  LTSOPHOSPHOUpiOS  DETERGENTS  V MICELLAR  INVERTED CONE  PNOSENATIDYLCHO(INE SI’IIINGOMYEUN PHOSPHATIDYLSERINE P**OPHATIDYLINOSITOL PHOSPHATIOYLGLYCEROL PH0SPIIATIOIC ACID CAEOIOLIPIN DTGALACTOSVLDtGLYcTRID€  ——  SILAYER  CYLINDRICAL  PNOSPHATIOELETHANOLAUIN (UP4SATI*ATEO) + 2 0 CARDTOLIPIN_C pHospIlArloIc 2 ACIO-C (pH<6.O P1IOSPHATIDIC ACID (pH<3.0I PHOSPHATIDYLSERINE  (pH<4.Oj MONOGALJCTOSYLDIGLYCERIOE HEXAGONAL (H J 11  CONE  LfRt1ELLAR  B  0 a: a: D  a: a: CCC  a: Ca  z  w N  C—  CIT  Ill I—  z  a: —  UT  2: 0  aD o a:  C-)  a: UT  ID IC w  —  v HEXAGONAL 20  --  H  1.3.4.2 The Function ofLipid Polymorphism in Biological Membranes Biological membranes are composed of mixtures of lipids, individual species of which adopt different polymorphic phases. Obviously, the lipids present in a biological membrane must maintain the appropriate characteristics required for cell viability. That is, they must form a fluid bilayer structure capable of providing a suitable permeability barrier to the flow of ions and metabolites while allowing for transmembrane communication and membrane protein function among many other demands. It has been suggested that non-bilayer forming lipids regulate such phenomena as membrane order (Lafleur et al., 1990) the transport of proteins across bilayers (Batenburg & de Kruijff, 1988; de Vrije et al., 1990) and the formation of various membrane structures such as tight junctions (Kachar & Resse, 1982). However, membrane fusion is the main biological phenomenon for which there is extensive evidence to indicate that non-bilayer forming lipids are involved and this is discussed in greater depth in Section 1.6.3  1.4 LIPID TRANSPORTAND EXCHANGE The trafficking of lipids within cells and between cells is important in many biological events including membrane biogenesis and homeostasis and the sorting of proteins by lipid carrier vesicles. It is now well established that the membranes of intracellular organelles maintain unique lipid compositions. Since these organelles usually do not synthesize all of the lipids contained in their membranes, the synthesis, sorting and transport of these lipids is crucial for the maintenance of membrane structure. Within cells, transport is regulated by three general mechanisms. These include vesicular transport, where lipids are transported between membranes by budding and fusing of  21  vesicles, monomer transport, the transport of single molecules by diffusion through the aqueous phase and lateral diffusion, where molecules exchange between two connected organelles (Sleight, 1987). The trafficking of lipids throughout cells is complex and an area of much current research (van Meer, 1989; Pagano, 1990; Voelker, 1990). The process of lipid trafficking would be expected to be influenced by the transbilayer distribution of lipids. For example, lipids located on one monolayer of a cell or organelle membrane would not be available for monomer transport on the opposite side. In the case of vesicular transport the fusion processes could be affected by the transbilayer distribution of lipids (see Section 1.6 and Chapter 4). Lateral diffusion may also be affected by the transbilayer lipid distribution as is the case for tight junctions (see Section 1.5.4). The effects of the transbilayer distribution of fatty acids on their transport (introduced below) is the subject of Chapter 2.  1.4.1 Extracellular Fatty Acid Transport Fatty acids provide the main energy source for most mammalian tissues and are essential components of the structural lipids of cell membranes (Bass, 1988). Since fatty acids are hydrophobic, special mechanisms are required to deliver fatty acids through the plasma to the tissue where they are employed. However, the mechanism(s) by which fatty acids are released from the specific carrier molecules and cross cell membranes en route to their site of utilization is still a matter of debate. Short chain fatty acids (2-10 carbon atoms long) are absorbed directly through the villi of the intestinal mucosa, diffuse across the mucosal cytosol and enter the venous blood while remaining in the free fatty acid form. Long-chain fatty acids (12 or more carbon atoms), on the other hand, are solubilized by bile salts and form micelles. These  22  micelles are transported to the brush borders or microvilli of the intestinal epithelial cells where the fatty acids dissociate from the micelle and diffuse across the epithelial cell membrane. Inside the epithelial cell, the fatty acids are esterified to form triacylglycerols which are integrated with free cholesterol, cholesterol esters, phospholipids and proteins to form chylomicrons. The chylomicrons are transported through the epithelial cell membrane into the lacteals of the lymphatic system, which delivers them into the general circulation via the thoracic duct (Rawn, 1989). Serum albumins are also responsible for the transporting a large proportion of fatty acids through the circulation. At the surface of endothelial cells, at the fatty acids are hydrolyzed from triglycerides contained in lipoproteins by lipoprotein lipases and pass from the capillary to the interstitial space through the endothelial cells of the capillary wall. Alternatively, as shown in Fig. 5, albumin-fatty acid complexes may pass the endothelium through clefts or directly through the capillary endothelial cell by way of plasmalemmal vesicles (Paulussen & Veerkamp, 1990). Upon reaching the target cell the fatty acids must dissociate from the carrier molecule and cross the plasma membrane. The mechanism by which the fatty acid crosses the membrane is not clear. There is evidence to support passive diffusion of the fatty acid, where the rate of diffusion is limited by the equilibrium distribution of fatty acids between albumin and the plasma membrane (Noy et al., 1986). Other data supports the presence of fatty acid transport systems. These include systems where fatty acids either passively diffuse across the membrane or are transported by membrane proteins after the albumin binds to a membrane associated receptor (Stremmel, 1988). It has also been suggested that fatty acids diffuse from albumin into the membrane where they are transported by plasma membrane fatty acid binding proteins or translocators (see Fig. 6).  23  Figure 5 Fatty Acid Transport Into the Interstitial Space Transport of fatty acids (FA) through the capillary endothelial cells into the extracellular fluid followed by transport into the target cell. The fatty acid can cross the endothelial cell either in the free fatty acid form or bound to serum albumin (see Section 1.4.1). PMFABP (Plasma membrane fatty acid binding protein).  :tracellular Fluid  PMFABP A Albumin Receptor  24  Figure 6 Proposed Mechanisms of Fatty Acid Transport Across Cell Membranes Models of cellular uptake of fatty acids (FA). (A) Passive diffusion through the plasma membrane. (B) passive diffusion through membrane after albumin (ALB) binds to a specific receptor that facilitates removal of the fatty acid. (C) An extrinsic fatty acid binding protein seives to help dissociate FA from serum albumin followed by FA diffusion across the membrane. (D) Intrinsic fatty acid translocator. (E) Intrinsic fatty acid translocator with an external albumin binding site and an intracellular FABP receptor (see Section 1.4.1).  A  B  D  C  E  FA  25  1.4.2 Intracellular Fatty Acid Transport Once delivered to the cytoplasm, fatty acids may be utilized in biosynthetic pathways, in the production of metabolic energy through n-oxidation or they can be re esterified to triacyiglycerols for energy storage. In the cytoplasm, the fatty acids are bound to fatty acid binding proteins (FABPs) which probably participate in the intracellular transport and storage of fatty acids much like albumin does extracellularly (Spener et al., 1989). FABPs have been proposed to promote cellular uptake of fatty acids, protect cellular structures from the detergent effects of fatty acids, modulate enzyme activities and to target fatty acids to specific metabolic pathways (Bass, 1988; Spener et al., 1989; Spener & Mukherjea, 1990; Paulussen & Veerkamp, 1990).  1.5 MEMBRANE ASYMMETRY  As previously mentioned (Section 1.1), biological membranes maintain an asymmetric transbilayer distribution of their components. For lipids, this asymmetry is not absolute and must be maintained during the lifetime of the cell or organelle. In this section, some of the mechanisms used to establish and maintain lipid asymmetry both in biological membranes and model membrane systems are briefly reviewed and some of the implications that lipid asymmetry has for the functions of biological membranes are discussed.  1.5.1 Methods of Detecting Lipid Asymmetry  Most techniques used to detect lipid asymmetry specifically modify one side of the lipid bilayer of a sealed membrane. In order for reliable results to be obtained, several criteria must be met. Firstly, the membrane to be probed must be highly purified.  26  This is often not the case when studying organelles. Secondly, the isolated vesicles must have a unique sidedness (ie. either all right side out or all inside out). Thirdly, the membrane must be impermeable to the probe so that only one leaflet of the bilayer is examined. Fourthly, the probe must not alter the permeability of the membrane and finally the reagent (enzyme, chemical probe, exchange protein etc.) must be able to detect all exposed membrane components (Op den Kamp, 1979; Houslay and Stanley, 1982). The techniques used to determine lipid asymmetry include immunological techniques, modification with chemical reagents, enzymatic degradation of lipids, fluorescent probes, lipid exchange proteins and physical techniques such as NMR and ESR among others (Op den Kamp, 1979; Etemadi, 1980; Houslay & Stanley, 1982; Schroeder, 1985; Gennis, 1989).  1.5.2 Phospholipid Asymmetry in Biological Membranes Bretscher (1972) first reported that phospholipids were asymmetrically distributed across erythrocyte plasma membranes. For normal human erythrocytes, it has been determined that all the phosphatidylserine is located on the inner monolayer along with 80% of the total phosphatidylethanolamine. The outer monolayer consists mainly of the choline containing lipids phosphatidyicholine and sphingomyelin, and glycolipids (see Figure 7). In the twenty years since this finding, a large number of membranes have been reported to show asymmetric distributions of phospholipids. Some examples include bacterial membranes (Barsukov et a!., 1977; Donohue-Rolfe & Schaechter, 1980; Goldflne et a!., 1982), intestinal (Barsukov et al., 1986) and renal brush- border membranes (Venien & Le Grimellec, 1988), beef heart mitochondrial inner membranes  27  Figure 7 Phospholipid Asymmetry in Mammalian Plasma Membranes (A) Human erythrocyte membrane, (B) rat liver blood sinusoidal membrane, (C) rat liver continuous plasma membrane, (D) pig platelet plasma membrane and (E) VSV envelope derived from hamster kidney BHK-21 cells. From Cullis & Hope, 1985  Outer Monor  go  A  100  80 70 60 50 40 31)  PC  -SF  8  Ii  C  sp  PC  I-I  1  PC  PC  I  lji  PC P1 P6÷ +Sp  •  10  ÷  100  s  80 70 60 50 40 30 20 10  lips  0 10 20  20  30 40  40 50 60 70 80 90 1(t)  E  0  II  II —I  I  j  50 60  go  80 100  Inner monolayer  28  (Krebs et a!., 1979), cardiac sarcolemma (Post et a!., 1988), viral membranes (Rothman et al., 1976; Schafer et al., 1974), platelets (Sanchez-Yague & Llanillo, 1986) and plasma membranes from many other sources as well as many subcellular organelle membranes. Initially, this phospholipid asymmetry was believed to be a static phenomenon, similar to protein asymmetry, where the asymmetry was established during biosynthesis and maintained due to an extremely slow rate of transbilayer movement. However, evidence of the rapid transport of phospholipids in various biological systems (Zilversmit & Hughes, 1977; de Kruijff et al., 1979; van den Besselaar et al., 1979; Hutson & Higgins, 1982) seemed to argue against a purely static mechanism of maintaining asymmetry. Furthermore, reports in 1984 by Seigneuret and Devaux indicating that spin labelled analogues of aminophospholipids were specifically transported by an ATP dependent mechanism in human erythrocytes provided evidence that asymmetry was actively maintained. Bishop and Bell (1985) reported a similar finding for phosphatidylcholine analogues in rat liver microsomes, where the transport of phosphatidylcholine was found to be rapid and saturable and could be inhibited by structural analogues of PC or by proteases. These studies indicate that lipids can be transported across membranes by specific protein translocases. Different mechanisms are probably involved since the transport of PC and its analogues is independent of metabolic energy and has been classified as facilitated diffusion, while aminophospholipids are transported by an ATP-dependent active transport system (Zachowski & Devaux, 1990; Devaux, 1991). Further evidence for the presence of aminophospholipid translocases has been provided by Schroit et a!., (1987), who reported rapid movement of fluorescent analogues of PS and PE across erythrocyte membranes. PS translocation has also been inferred  29  from morphological changes in erythrocytes after the incorporation of PS into the outer monolayer. The transport of PS was shown to be dependent on the concentration of 2 ions and ATP and also on the state of protein sulfhydryl groups (Daleka & Huestis, Mg 1985). Connor and Schroit (1987) have shown that a photoactivatible analogue of PS preferentially bound to a 30 kDa protein in the erythrocyte membrane. Latter work indicated pyridyldithioethylamine (PDA), an inhibitor of the PS translocase, bound to a 30 31 kDa protein and co-migrated with band 7 from the erythrocyte membrane -  (Connor & Schroit, 1988). Other researchers have reported that the aminophospholipid Mg + ATPase (Damiana et al., translocase may be a 115-130 kDa vanadate-sensitive 2 1987). Zachowski et al. (1986), showed that PE and PS compete for the same translocase with PS having a higher affinity for the translocase. More recent studies have provided evidence for translocases in various other membranes such as human platelets (Sune et al., 1987) and in chromaffin granules (Zachowski et al., 1989). Isolation and reconstitution would be the ideal method of positively identifying translocases but this could be very difficult as it has been suggested that the translocase is present in a very low copy number in erythrocytes (< 1000 copies/cell) (Devaux, 1991). A translocase has been reconstituted into liposomes from total protein extracts of rat liver microsomes but this has not served to purify or identify the translocase (Backer & Dawidowicz, 1987). Although there is now a large body of evidence to support the existence of phospholipid translocases, other factors may also be involved in the development and maintenance of phospholipid asymmetry. For example, it has been suggested that the cytoskeleton may play a role in maintaining phosphatidylserine asymmetry (Bevers et al., 1987; Verhallen et al., 1987; Comfurius et al., 1989). This relationship has been inferred from experiments which indicate that the breakdown of the cytoskeleton during platelet  30  activation is accompanied by the appearance of PS on the outer monolayer of platelets (Verhallen et al., 1987) and experiments which indicate a reversible specific binding of PS to the cytoskeleton (Comfurius et al., 1989). Williamson et al., 1987 have also stated that a kinetic analysis of the data collected by Devaux and co-workers on PS transport across erythrocyte membranes indicates that the translocase is actually a bidirectional flip-flop enzyme and could not in itself maintain aminophospholipid asymmetry. However, other research has provided evidence to suggest that the cytoskeleton is not required to maintain PS asymmetry. For example, Middelkoop et al. (1989) have shown that treatment of erythrocytes with diamide to crosslink cytoskeletal proteins did not affect the phospholipid asymmetry but the asymmetry was lost if ATP was depleted. Also, studies with vesicles formed from heat treated erythrocytes show that the vesicles are capable of maintaining aminophospholipid asymmetry as long as ATP is present at sufficient concentrations. This indicates that cytoskeletal proteins are not required for asymmetry since the vesicles contain only about 25% of the spectrin associated with erythrocytes and are devoid of the other cytoskeletal proteins (Calvez et al, 1988). In contradiction to these findings, recent research by Connor & Schroit (1990) argues that an endofacial protein (probably a cytoskeletal component) is required to maintain PS asymmetry. This argument is based on several observations. Firstly, diamide did not react with the sulfhydryl group of the 32 kDa protein which they have tentatively identified as the PS transporter, but did inhibit PS transport. If the 32 kDa protein they have identified as the PS transporter is the only protein involved in PS translocation then diamide should not affect PS transport if it does not react with this protein. Secondly, oxidative cross-linking and inhibition of PS transport by diamide was reversible by intracellular glutathione but PDA inhibition (PDA binds the sulfhydryl of  31  the 32 kDa protein) was not. This suggests that the inhibition of PS transport is a result of these chemicals reacting with two different proteins. Finally, it was claimed that oxidation of either the PDA sensitive site or the diamide sensitive site after the PS was transported to the inner monolayer, did not promote the movement of PS from the inner to the outer monolayer, suggesting that the PS was stabilized on the inner monolayer. It is stated that this finding is inconsistent with other data indicating that an ATP-dependent outward movement of PS occurs in RBCs and that cytoskeletal proteins are not involved in the transport of exogenously added PS analogues. Thus, the role of the cytoskeleton in maintaining phospholipid asymmetry in RBC and other plasma membranes is obviously still a matter of debate. Other factors may also be involved in the development and maintenance of phospholipid asymmetry. For example the addition of gramicidin to erythrocytes greatly increases the rate of transbilayer movement of phosphatidyicholine. However, formylated gramicidin does not affect the rate of PC flip-flop (Classen et al., 1987). This has been correlated to the ability of gramicidin to induce non-bilayer (H 1 phases in membranes and the inability of the formylated gramicidin to induce this phase (Killian & de Kruijff, 1986). It is also possible that membrane potentials play a role in developing lipid asymmetry. Donohue-Rolfe and Schaechter (1980) have shown that the transport of phosphatidylethanolamine across the cytoplasmic membrane of E. coli is extremely rapid. This transport is not affected by depletion of ATP or the inhibition of protein or lipid synthesis, but is markedly reduced if the proton-motive force is depleted. The complexity of biological membranes makes determining the role of various mechanisms to induce and maintain lipid asymmetry a difficult task.  32  In addition to the complications of different mechanisms acting to induce lipid transport in biological membranes, the discovery that phospholipid transport in biological membranes is rapid has rendered the results of some techniques of determining lipid asymmetry questionable. If the rate of transport of the lipids is fast with respect to the method used to determine asymmetry, then the results of the assay could be invalid. Also, techniques which perturb the membrane could result in the rapid redistribution of lipids (Etemadi, 1980; Devaux, 1991). This dynamic aspect of lipid asymmetry along with the difficulties of preparing pure membrane fractions from eucaryotic cells could explain the often different extents of asymmetry reported by various laboratories for the same membranes. Although the rapid redistribution of lipids under some conditions may lead to difficulties in quantifying the exact extent of lipid asymmetry, the research to date is unambiguous in indicating that lipid asymmetry is a general feature of most or all biological membranes.  1.5.3 Lipid Asymmetry in Model Membranes Model systems have been used to study both the mechanisms by which asymmetry is generated and maintained, and the functions of lipid asymmetry in biological membranes. Liposomes exhibiting lipid asymmetry have been produced by many procedures including enzymatic or chemical modifications of external lipids, for example phospholipase digestion (de Kruijff & Baken, 1978; Low & Zilversmit 1980), treatment of vesicles containing phosphatidylserine with PS-decarboxylase (Schroit, 1986) and the conversion of phosphatidylethanolamine to its amidine derivative using isethionyl acetimidate (Roseman et al., 1975). Lipids have also been incorporated into the exterior monolayer of vesicles by spontaneous transfer of various lipids such as  33  trisialogangliosides (Feigner et ai., 1981) and phospholipids with fluorescent molecules attached to one of the acyl chains (Pagano et al., 1981) or through the use of phospholipid transfer proteins (Low & Zilversmit, 1980; de Kruijff & Wirtz, 1977). Small unilamellar vesicles (SUVs) prepared by the sonication of mixtures of phospholipids have also been used to study lipid asymmetry (Barsukov et al., 1980; Bramhall, 1986; Berden et al., 1975; Lentz et al., 1980). In these systems the spontaneous asymmetric distribution of  lipids occurs upon sonication, presumably due to the small radius of curvature of the vesicle which favors the packing of specific headgroups at the inner monolayer. Phospholipid transport in these systems has been shown to be very slow, usually with a half-time on the order of days (Low & Zilversmit, 1980; Ganong & Bell, 1984). Perturbing the bilayer with enzymes (de Kruijff & Baken, 1978), or due to the presence of residual detergent (Kramer et al., 1981) increases the rate of lipid transport but not to levels observed in biological membranes. Divalent cations (Lentz et al., 1982), the incorporation of proteins into membranes (de Kruijff et al., 1978; Gerritsen et al., 1980) and incorporation of lipids into the external bilayer (Barsukov et al., 1980; de Kruijff & Wirtz, 1977) also produce relatively fast rates of transport. In these cases half-times for transport are usually on the order of hours instead of days. Although the rate of transport of phospholipids in these model membranes is often less than that observed for biological membranes, especially membranes that synthesize lipids, model membranes have indicated the importance of various factors in lipid transport. For example the rate of transbilayer lipid redistribution in vesicles with reconstituted integral membrane proteins is similar to the rate observed in erythrocytes indicating that a base rate of lipid flip-flop in certain membranes may simply be due to packing defects around membrane proteins. Other studies have shown that differences in  34  the lateral packing due to differences in the acyl chains may play a role in developing asymmetry (Gabriel & Roberts, 1987) and the rate of lipid flip-flop (de Kruijff & Wirtz, 1977). The observation that divalent cations can increase the rate of lipid transport of certain lipids, for example PG (Lentz et al., 1982) indicates that phospholipids can undergo rapid flip-flop under the appropriate ionic conditions. These results combined with evidence exhibiting the accumulation of Ca 2 in vesicles containing PA (Nayar et al., 1984; Serham et al., 1981a, 1981b) implicates phospholipids as possible ionophores. Model systems have also been used to show the importance of the phospholipid headgroup structure in the transbilayer diffusion of lipids (Homan & Pownall, 1988) and work by Ganong and Bell (1984), employing phosphatidylthioglycerol and dioleoylthioglycerol, has shown that the phospholipid headgroup is the major barrier to transbilayer lipid transport. In addition, future experiments involving the reconstitution of membrane proteins may eventually lead to the isolation and identification of lipid translocases.  1.5.4 Biological Significance of Lipid Asymmetry The fact that all the biological membranes studied to date have shown an asymmetric distribution of their phospholipids and that the transbilayer transport of certain lipids appears to be mediated by proteins, strongly indicates that the asymmetry of lipids is important for cell function. Indeed, many functions have been proposed for the role of lipid asymmetry. Dragsten et al. (1981) proposed that the ability of membrane components to pass through tight junctions between the apical and basolateral surfaces of epithelial cells depends upon their being located on the inner monolayer of the membrane. Thus lipids and other molecules which are capable of being transported to  35  the inner monolayer of the membrane pass freely through tight junctions while proteins and other non transportable molecules remain segregated on their respective surfaces. It has been suggested that lipid asymmetry may be responsible for the maintenance of the discoid shape of erythrocytes (Seigneuret & Devaux, 1984). Research has shown that erythrocytes which have lost their asymmetry are recognized and phagocytosed by macrophages more readily than asymmetric erythrocytes (McEvoy et a!., 1986). Since aged erythrocytes show an increased exposure of phosphatidylethanolamine and phosphatidylserine on their outer monolayer (Shulka & Hanahan, 1982), it is possible that the loss of asymmetry is responsible for the clearance of aged erythrocytes by splenic macrophages (Tanaka & Schroit, 1983; Schroit et al., 1985). The role of platelets in blood coagulation has been correlated to the exposure of phosphatidylserine on the outer monolayer of the platelets upon activation (Zwaal, 1978; Bevers et a!., 1987). Other research (see below) has indicated that lipid asymmetry may also play a role in membrane fusion events (Cullis & Hope, 1988; Tullius et al., 1989; Santini et a!., 1990) especially exocytosis (Nayar et al., 1982; Zachowski et al., 1989). Finally, it has been speculated that specific lipids may be necessary for the enzymatic activity of certain proteins (Houslay & Stanley, 1982).  1.6 MEMBRANE FUSION Membrane fusion is a central process in many biological events including endo and exocytosis, fertilization, lipid and protein trafficking and muscle biogenesis. In biological systems, membrane fusion is a tightly regulated event and although fusion has been extensively studied, the molecular mechanisms involved in controlling fusion are still poorly understood. Studies suggest that biological membrane fusion is controlled by  36  many factors including the presence of various ions, specific proteins and the lipid composition of the membrane. There are many examples of the importance of proteins in the regulation of membrane fusion in vivo. Exocytosis and fusion of enveloped virus to cell membranes are two examples. The role of proteins is quite different in these two processes. In exocytosis a group of calcium binding proteins, collectively referred to as annexins, appear to play a major role in regulating fusion by aggregating the appropriate membranes. However, it has been suggested that the actual fusion event is initiated by other factors that destabilize the lipid bilayer, such as fatty acids and diacylglycerol (Zaks & Creutz 1990). In contrast, it has been proposed that the fusion of enveloped viruses to target membranes is controlled entirely by the viral glycoproteins. That is, the glycoprotein is responsible for recognition and binding to the target membrane and the initiation of fusion by destabilizing the lipid bilayer (Hoekstra & Kok, 1989; Hoekstra, 1990). Regardless of the mechanism by which the fusing membranes are destabilized, the final fusion event must involve the lipid bilayers. Model membranes are therefore an ideal system for studying membrane fusion. A brief discussion on the fusion of model membranes and some of the methods used to assay fusion events, follows.  1.6.1 Methods of Detecting membrane Fusion Techniques that have been used to detect membrane fusion include differential scanning calorimetry (DSC), nuclear magnetic resonance (NMR) and electron spin resonance (ESR) spectroscopy, electron microscopy, gel filtration chromatography, turbidity measurements and fluorescence techniques (Wilschut & Hoekstra, 1986).  37  Fluorescence techniques have proved to be among the most useful procedures for studying fusion and a few techniques will be briefly discussed.  1.6.1.1 Lipid Mixing Assays Several lipid mixing assays have been developed to monitor fusion. The most commonly used method is based on resonance energy transfer (RET) between two fluorescently labelled lipids, often derivatives of phosphatidylethanolamine, N-(lissamine rhodamine B sulfonyl) phosphatidylethanolamine (N-Rh-PE) and N-(7-nitro-2,1,3benzoxadiazol-4-yl) phosphatidylethanolamine (N-NBD-PE) (Struck et al., 1981). Rhodamine is an efficient quencher of NBD fluorescence because its excitation maximum is close to the emission maximum of NBD. In this assay the fluorescent probes are incorporated into liposomes in equal concentrations (typically  <  1% of the  total lipid). The labelled vesicles are mixed with an excess of unlabelled target membrane and NBD fluorescence is constantly monitored. Fusion is detected as an increase in NBD fluorescence as the fluorescently tagged lipids diffuse away from each other upon mixing with the unlabelled lipids. Alternatively, the two lipids can be incorporated into separate populations of vesicles and fusion can be monitored as a decrease in NBD fluorescence due to energy transfer between the fluorescent probes upon lipid mixing. The RET assay has proven to be very useful as the increase in NBD fluorescence is proportional to the amount of fusion under specific conditions. However, since the fluorophores must be incorporated into vesicles, at least one of the membranes must be artificial. An assay for fusion that does allow monitoring of intact biological membranes relies on the fluorescent dequenching of octadecyirhodamine B chloride (R ) (Hoekstra 18  38  et al, 1984). This fluorophore spontaneously incorporates into lipid bilayers after its addition in an ethanolic solution. At high enough concentrations in the intact membrane the R 18 effectively quenches its own fluorescence. Upon dilution into unlabelled membranes due to fusion, the efficiency of quenching is reduced and an increase in fluorescence is observed (see Fig. 8).  1.6.1.2 Mixing ofAqueous Contents Many assays have been used to detect the mixing of aqueous contents upon fusion. The two most commonly used fluorescent assays are the terbium (Th j, 3 dipicolinic acid (DPA) assay (Wilschut et al., 1980) and the 1-amino-naphthalene-3,6,8trisulfonic acid / N,N’-p-xylenebis-(pyridinium bromide) (ANTS/DPX) assay (Ellens et al., 1985). In both cases the respective probes are encapsulated in two separate populations Tb + IDPA assay, the Tb + is encapsulated with a low concentration 3 of vesicles. For the 3 of citrate to prevent it from binding to negatively charged lipids. If vesicle fusion occurs a Tb(DPA) 3 complex forms which is highly fluorescent. EDTA is present in the external medium to prevent the formation of the fluorescent complex due to leakage from the vesicles. This assay is not useful in acidic conditions (pH  <  4.0) because the  protonated form of DPA is membrane permeable. The ANTS/DPX assay, in which the decrease in fluorescence of ANTS upon chelation by DPX is measured, can be used in acidic conditions. Leakage of contents to the external medium does not affect ANTS fluorescence because the quenching of fluorescence by DPX is highly concentration dependent. These assays are restricted to systems where the internal contents can be exchanged.  39  Figure 8 Fluorescence Assays to Monitor Membrane Fusion (A) Mixing of internal contents (Th /DPA). (B) lipid mixing assays using resonance 3 energy transfer (RET) and (C) relief of fluorescence self-quenching of . 18 R  A  C  0  40  1.6.2 Fusion of Model Membranes Model systems ranging in complexity from single cells and permeabilized cells to cell-free systems composed of isolated intracellular membranes to pure lipid bilayers have provided the most information about the molecular mechanism of membrane fusion (Burger & Verkleij, 1990). The most useful systems have been the simplest model membranes, that is systems composed of pure lipids. These systems allow for the precise control of the environment of the membrane and therefore individual parameters can be tested for their effect on fusion. For example, calcium has been shown to induce fusion of vesicles composed of acidic phospholipids (Papahadjopoulos et al., 1974, 1976, 1977; Wilschut et al., 1980, 1981). Since Ca 2 is required for many biological fusion events, the possibility arises that the interaction of Ca 2 with acidic phospholipids may mediate membrane fusion. + required to cause fusion between vesicles composed 2 However, the concentration of Ca purely of acidic phospholipids is usually several orders of magnitude larger than that observed in the intracellular milieu (Duzgunes, 1985). This does not necessarily rule out the role of acidic lipids in intracellular membrane fusion events since areas may exist with transiently high local Ca 2 concentrations. In addition, as previously mentioned (Section 1.6) certain proteins (annexins among others) act to aggregate fusing + 2 membranes. These proteins have been shown to greatly reduce the concentration of Ca required to induce the fusion of acidic phospholipid vesicles (Papahadjopoulos et al., 1990; Zaks & Creutz, 1990). More likely Ca 2 could mediate fusion through other Ca has been + processes acting synergistically to destabilize the bilayer. Furthermore, 2 suggested to induce a conformational change in a fusion protein to make it fusogenic  41  (Papahadjopoulos et a!., 1990). Alternatively, the presence of Ca 2 could activate phospholipases to produce free fatty acids or diacyiglycerols, both of which have been shown to destabilize bilayer structures and cause membrane fusion (Zaks & Creutz, 1990). Other enzymatic reactions such as phosphorylation of proteins by protein kinases could also play a role (Burger & Verkleij, 1990).  It is also possible that other ions such  as Zn 2 or Mg 2 act synergistically with the Ca 2 to reduce the concentration of Ca 2 required to induce fusion (Deleers et al., 1986). It is likely that many of these factors are responsible for the initiation and regulation of membrane fusion. Other components, besides the presence of divalent cations, involved in the induction and regulation of membrane fusion may be specific proteins which can either aid in fusion by aggregating membranes or may be directly involved in destabilizing the lipid bilayer through hydrophobic interactions (Martinez-Bazenet et al., 1988; Meers et a!., 1988(a); Maezawa et al., 1989; Arvinte et al., 1986). Peptides acting in a similar manner to proteins (Walter et al., 1986; Suenga et a!., 1989), free fatty acids or diacylglycerols which can induce the formation of non-bilayer phase structures (Cullis & Hope, 1978; Meers et a!., 1988(b); Siegel et a!., 1989) lipid composition (Duzgunes, 1985), lipid polymorphic properties (see Section 1.6.3) (Cullis & Hope, 1988; Ellens et al., 1989) and other factors may also be important in the induction and regulation of fusion. Since model systems can address questions that are not easily examined in complex biological systems they have lead to several proposed mechanisms for membrane fusion.  42  1.6.3 Molecular Mechanism(s) of Membrane Fusion  In recent years several mechanisms for membrane fusion which share common features have been proposed. These include the close apposition of membrane bilayers, the formation of an intermediate structure in which the outer monolayers of the vesicles coalesce (membrane destabilization), the formation of a pore between the two vesicles and the mixing of the internal contents as the pore expands. The differences in the mechanisms of fusion lie in the nature of the intermediate structures and the parameters which cause close membrane apposition and bilayer destabilization. For example Papahadjopoulos et al., (1990) have presented three pathways for the molecular rearrangements of lipids during membrane fusion. In all three pathways, the initial step involves two intact membranes making contact with each other within a short distance with partial or complete dehydration at the point of contact. In the next stage, a “contactinduced defect” is postulated, resulting from the local dehydration and other intermembrane interactions. The resulting structures are unstable since parts of the acyl chains become exposed to the aqueous phase. The second step is considered to be the first committed step to fusion and has been termed the “hydrophobic contact intermediat&’. In this step the lipids of the outer monolayers can mix. The next stage involves the formation of a “curved bilayer annulus”, which is said to be a short-lived high-curvature intermediate forming a water filled channel between the two membranes. Finally, the aqueous channel enlarges, thus removing the areas of high-curvature and producing a larger vesicle. Difficulties with this model include the fact that the dehydration step, which can be induced by Ca 2 in PS containing systems, is not generally available for other lipid systems.  43  The most likely intermediate structures in membrane fusion are the intermediary structures postulated in the formation of non-bilayer lipid phases (hexagonal H 11 and cubic phase) (see Figure 9). The role of inverted micellar intermediates (IMIs) in membrane fusion was first proposed in the late 1970’s (Cullis et al., 1978; Verkleij et al., 1979; Cullis & Hope, 1979) and this theory has been enhanced by more recent kinetic analyses (Siegel 1986). Siegel’s kinetic analysis shows that inverted micelles have a very short lifetime (103 sec range). This value agrees well with the rate of biological fusion events (also of the i0 sec range) and is compatible with a local-point contact fusion event, which morphological evidence indicates is the method by which biological membranes fuse (Chandler, 1980, 1984; Heuser et al., 1979; Knoll et al., 1988). The next step in membrane fusion involves the formation of interlamellar attachment sites (ILAs) where an aqueous channel is formed between the two membranes (see Fig. 9). Alternatively aggregation of IMIs between apposed bilayers can lead to the formation of the H 11 phase which involves leakage of vesicle contents (Bentz et al., 1985). The involvement of these intermediary structures in membrane fusion is supported by a large body of evidence. For example, Hope & Cullis (1981) showed that the ability of molecules to induce the fusion of erythrocyte phospholipids correlated with their ability to induce non-bilayer phases and that chemiclly related molecules that were non-fusogenic did not induce the formation of non-bilayer phases. In addition, almost any biological membrane studied contains a significant amount of lipid that in isolation will adopt the hexagonal phase under physiological conditions. (Burger & Verkleij, 1990) and other lipids that will adopt this phase under various other conditions, such as the addition of divalent cations or cis-unsaturated fatty acids.  44  Other studies of membrane fusion in various systems are also consistent with a model of fusion involving IMIs. These include studies which show the presence of particles in influenza induced fusion events (Knoll et al., 1988) and exocytosis (Schimdt et a!., 1983) by fast-freeze freeze-fracture, the enhancement of envelope virus fusion with model membranes by the addition of H 11 forming lipid (Van Meer et a!., 1985; White et al., 1983) and the indication of fusion pore formation during exocytosis by capacitance measurements (Zimmerberg, 1987, Brenckenridge & Walmers, 1987). Finally as stated earlier, factors which induce the formation of the H 11 phase, such as fusion proteins, diacylglycerol and cis-unsaturated fatty acids also tend to induce fusion (Hope & Cullis, 1981; Siegel et al., 1989).  45  Figure 9 Mechanism of Membrane Fusion Procceding Via Intermediates of the Bilayer to Hexagonal H 11 Phase Transition Model of membrane fusion proceeding via an inverted micellar intermediate (IMI). Formation of the hexagonal Hi phase occurs via aggregation of IMIs (see Section 1.6.3). From Ellens et a!., 1989.  gij  3 [  H  a  1  i  I  IMI  H phase pre curio rs (leakage)  1  Isotropic or inverted Cubic Phases  H 1 1  46  1.7 THESIS OUTLINE The mechanisms by which lipid asymmetry is generated and maintained in biological membranes are still poorly understood. The functions of lipid asymmetry in biological membranes are also ill defined. Previous research has indicated that ion gradients, specifically pH gradients, can induce lipid asymmetry in model membrane LUV systems. The purpose of this thesis is to study the mechanism by which pH gradients induce this asymmetry for phosphatidic acid and other acidic lipids. Further, the role of lipid asymmetry in such processes as lipid exchange and membrane fusion are examined. The influence of transmembrane pH gradients on the exchange of simple lipids (stearylamine and fatty acids) between LUVs, and between LUVS and bovine serum albumin is discussed in Chapter 2. In Chapter 3, the mechanism by which transbilayer pH gradients induce the asymmetric distribution of dioleoylphosphatidic acid (DOPA) is examined. Additionally, a new fluorescent assay to detect the asymmetric distribution of acidic phospholipid is described. Finally, in Chapter 4, the ability of lipid asymmetry to modulate membrane fusion between LUVs is examined.  47  CHAPTER 2 INTERVESICULAR EXCHANGE OF LIPIDS: INFLUENCE OF TRANSMEMBRANE PH GRADIENTS  2.1 INTRODUCTION  Lipid transport and exchange are important physiological processes. For example, the diffusion of fatty acids between micelles, fatty acid binding proteins and cell membranes represents fundamental steps in the catabolism and biosynthesis of lipids. In addition, fatty acids are the major source of metabolic energy of most mammalian tissues. Due to the low solubility of fatty acids in aqueous media and the considerable daily flux of fatty acids from adipose triglyceride stores into the plasma (— 700 mmol/day in the average human) complex mechanisms have evolved in order to transport fatty acids to their site of utilization (Bass, 1988). These include the transport of fatty acids through the plasma in the form of relatively inert triglycerides in lipoproteins or bound to fatty acid binding proteins (FABPs), of which serum albumin is the main example in plasma. Although, there is more than one mechanism for the delivery of fatty acids to target membranes (see Chapter 1, section 1.4.1), much evidence exists that fatty acid exchange between membranes, or between membranes and binding proteins, occurs via monomer diffusion from one surface to the other. The rate limiting step in this process is the release of monomers from the donor surface to the aqueous medium (Ferrel et al., 1985; Roseman & Thompson, 1980; Nicholls & Pagano, 1982; Doody et al., 1980; Noy et al., 1986). Thus the transbilayer distribution of an exchangeable lipid in a lipid bilayer should have a profound effect on the rate at which the lipid is exchanged between two  48  surfaces as lipids sequestered on the inner monolayer are not available to dissociate into the external aqueous environment. Previous studies have shown that transmembrane pH gradients can induce the asymmetric transbilayer distribution of simple lipids with acid or base characteristics, such as stearylamine and fatty acids (Hope & Cullis, 1987). A model of the mechanism by which lipids are transported across lipid bilayers in response to transmembrane pH gradients is depicted in Figure 10. The ability of a transmembrane ApH to regulate the transbilayer distribution of lipids should affect the exchange characteristics of these simple lipids. This chapter examines the effects of transmembrane pH gradients on the exchange of simple lipids between model membranes and between membranes and bovine serum albumin (BSA). The results are discussed with respect to the mechanism and modulation of lipid flow both in vitro and in vivo.  49  Figure 10 Mechanism of Net Acidic Lipid Transport in Response to a Transmembrane pH Gradient If the exterior pH is low a certain proportion of the acidic lipids are protonated (neutral) and freely permeable across the bilayer. However, upon reaching the more basic interior environment, the lipids are deprotonated and negatively charged. The charged form of the lipid is relatively impermeable to the bilayer such that the lipids effectively become trapped on the inner monolayer.  Inside  Outside  •OOC  tcoo  I  +  1 W  COOH  HOOC  II 0 [H], [R-C001  IR-COOiilHii  0 [R-COOH]  [R-COOH]  At equilibrium 0 [R-COOH]  1 [R-COOHJ  Thus (R-COO] 0 [R—C00]  [HJI  50  2.2 MATERIALS AND METHODS  2.2.1 Lipids and Chemicals Dioleoylphosphatidylcholine (DOPC) and N-(lissamine rhodamine B sulfonyl) dioleoylphosphatidylethanolamine (Rh-PE), N-(7-nitro-2, 1,3-benzoxadiazol-4-yl) dioleoylphosphatidylethanolamine (NBD-PE) and brain phosphatidylserine (PS) were obtained from Avanti Polar Lipids (Birmingham, AL). Stearylamine (SA) and oleic acid (OA) were purchased from Sigma (St. Louis) H} dipalmitoylphosphatidycholine (DPPC) was purchased from NEN. [ 3 while the [ C] 1 4 Oleic acid was obtained from ICN and the 4 [‘ C ] stearylamine was obtained from Dr. J. Wilschut. Lipid compositions of vesicles are expressed as molar ratios.  2.2.2 Vesicles Large unilamellar vesicles were prepared by extrusion through polycarbonate filters (Hope et a!., 1985) using the Lipex Extruder obtained from Lipex Biomembranes Inc., Vancouver, Canada. Lipid mixtures were dried down from chloroform under a stream of nitrogen gas and residual chloroform was removed under vacuum for one hour. The appropriate buffer composed of either 150 mM sodium citrate, 10 mM HEPES, 5 S0 (pH 7.4) or 150 mM sodium citrate, 5 mM 4 2 K mM 4 S0 (pH 4.0) was added to the 2 K lipid film and vortexed to prepare a liposomal suspension of approx. 10 p.mol/ml total phospholipid. The liposomes were freeze-thawed five times using liquid nitrogen-warm water cycles in order to increase the trapped volume of the vesicles and to promote equilibrium solute distributions. These freeze and thawed multilamellar vesicles  51  (FATMLVs) (Mayer et a!., 1985) were then extruded ten times through two stacked 0.1 urn polycarbonate filters (Nuclepore). After preparation, all vesicles were passed down Sephadex G-50 columns equilibrated with 10 mM , 4 S 2 Na O 10 mM HEPES, 5 mM 4 S 2 K 0 (pH 7.4). Exchange of the initial hydrating buffer for the low ionic strength buffer was necessary to enable binding of charged vesicles to the ion exchange columns.  2.2.3 Turbidity Experiments to Monitor Vesicle Aggregation Vesicles (DOPC/PS (8:2) mol/mol or DOPC/SA (9.5:0.5) mol/mol) were prepared as described above by hydrating the lipid films in buffer containing 150 mM citrate (pH 7.4). After exchanging the external buffer employing G-50 columns the turbidity of both sets of vesicles (at a concentration of 1 umol/ml total phospholipid) was monitored at 550 nm using a Shimadzu UV-160 spectrophotorneter. Vesicles containing PS were then added into a cuvette to a final concentration of 0.5 p.mol/ml. After monitoring the turbidity of the solution for approx. lOs,  stearylamine-containing vesicles (0.5 umol/ml final concentration) were added and mixed. The change in absorbance was then measured over the next 2 mm. To examine the effects of transmembrane pH gradients (ApH) on the aggregation of LUVs, vesicles composed of DOPCIPS(8:2) or DOPC/SA (9.5:0.5) were prepared as described above except the vesicles containing SA were hydrated in a buffer containing 150 mM citrate (pH 4.0). The turbidity of both sets of vesicles were measured separately. PS-containing vesicles were added to the cuvette to a final concentration of 0.5 p.mol/ml. After approx. 20 s stearylamine-containing vesicles were added and mixed.  Approx. 40 s later valinomycin and nigericin (0.1 and 0.01 umol, respectively) were added to the cuvette with mixing.  52  2.2.4 Stearylamine Exchange Monitored by Ion Exchange Chromatography using DEAE-Sephacel Vesicles composed of DOPC/PSINBD-PE (8:2:0.1) containing trace quantities of H]DPPC or DOPC/[ 3 [ C]SAIRh-PE (9.5:0.5:0. 1) were prepared in 150 mM citrate, 10 14 mM HEPES, 5 mM 4 S0 buffer (pH 7.4) and passed down Sephadex G-50 columns 2 K equilibrated with 10 mM 4 SO 10 mM HEPES, 5 mM 4 2 Na , S0 buffer (pH 7.4). Then, 2 K 1 tmol of each set of vesicles were passed down DEAE-Sephacel (Pharmacia) columns equilibrated in external buffer (10 mM 4 SO 10 mM HEPES, 5 mM 4 2 Na , S0 (pH 7.4)). 2 K After two fractions (approx. 3.5 ml per fraction) were collected, the eluting buffer was changed (500 mM NaC1, 10 mM HEPES (pH 7.4)) and four more fractions were collected. After establishing elution profiles for the two vesicle populations, 1 &mo1 of each population of vesicles were added together, mixed, and allowed to incubate for 5 mm at room temperature (approx. 20°C). The mixture of vesicles was then applied to a DEAE-Sephacel column and eluted under the same conditions as described above. Radioactive decays were monitored using a Packard Tricarb 2000 CA liquid scintillation counter. The presence of fluorescent lipid was assayed employing an SLM Aminco SPF SOOC fluorometer. To study the effects of a transmembrane pH gradient on stearylamine exchange, brain phosphatidylserine-containing vesicles ([ H]DOPCIPS or DOPC!PS (8:2)) and 3 stearylamine-containing vesicles (DOPC/[’ C]SA or 4 4 H]DOPC/[’ 3 [ C ]SA (9.5:0.5)) were prepared as described above with the exception of the buffers. The PS-containing vesicles were hydrated at pH 7.4 while the SA-containing vesicles were hydrated at pH 4.0. Vesicles (1 tmo1 of each population) were mixed ([ H]DOPC/PS + DOPCI[’ 3 C]SA 4 or DOPC/PS + 14 H]DOPC/[ in the pH 7.4 buffer and incubated either in the 3 [ C]SA)  53  presence or absence of the ionophores valinomycin and nigericin (0.1 and 0.01 tmol, respectively). After a 5 mm incubation at room temperature, the vesicle populations were separated on DEAE columns and the fractions were assayed for 3 H and 14 C.  2.2.5 Oleic Acid Exchange Oleic acid containing vesicles (DOPC/PS/[’ C]OA (8:2:0.5)), DOPC vesicles and 4 DOPC vesicles containing stearylamine (DOPC/SA (9.5:0.5)) were hydrated at pH 10.0 (100 mM 4 SO 50 mM 3 2 Na , B0 5 mM 4 H , S0 or pH 7.0 (150 mM sodium citrate, 10 2 K ) mM HEPES, 5 mM 4 S0 as required. After extrusion, vesicle populations were 2 K ) passed down G-50 columns equilibrated in pH 7.0 external buffer (20 mM 4 SO 10 2 Na , mM HEPES, 5 mM ) 4 S 2 K 0 . Vesicle populations (1 !.tmol of each) were mixed either with or without the presence of ionophores, and after a 5 mm incubation at room temperature (— 21°C) the populations were separated on DEAE-Sephacel columns as described above. Fractions from the column were then counted employing a dual label program on a Packard 2000 CA liquid scintillation counter. For experiments monitoring the extraction of fatty acids from BSA, [‘ C]fatty 4 acid were mixed with BSA at a ratio of 2:1 (F.A./BSA). Fatty acid loaded BSA was added to 1 mL of 20 mM DOPC vesicles to a final concentration of 5 mg/mL BSA. 200 L of the mixture was applied to a 20 x 1.5 cm Sepharose CL-4B column and eluted with  the external pH 7.0 buffer. The fractions were measured for 3 H and 14 C counts. All results presented in this thesis are average values obtained from a minimum of three experiments, except where noted. In this Chapter, Figures 11 and 13 present results from a single experiment which are typical of multiple trials.  54  2.3 RESULTS  2.3.1 Stearytamine Exchange The aggregation-disaggregation behavior of vesicles consisting of DOPC and 20 mol% PS and donor vesicles composed of DOPC and 5 mol% stearylamine is shown in Fig. 11. The turbidity profiles of vesicles prepared with a transmembrane ApH  =  0 (pH  7.4 inside and outside) demonstrate that separately PS and stearylamine-containing vesicles are only slightly turbid, giving an absorbance of approx. 0.06 at 550 nm. However, when mixed the turbidity of the solution immediately increases then slowly decreases until it reaches the original turbidity of the separate vesicles. This process can also be followed visually. It is logical to suggest that this aggregation-disaggregation phenomenon occurs due to an initial electrostatic attraction giving rise to aggregation of the oppositely charged vesicles during which time exchange of the stearylamine occurs, followed by disaggregation of the vesicles after the surface charges of acceptor and donor vesicles have equilibrated. This interpretation is supported by the data shown in Fig 12. Figs. 12A and 12B show the elution profiles from DEAE-Sephacel columns of H]DOPC/PS (8:2) and DOPC/[ 3 [ C]SA (9.5:0.5) vesicles, respectively. In Fig. 12C the 14 two vesicle populations were mixed and incubated for 5 mm before being applied to the DEAE column. All the tritium counts are associated with the PS-containing vesicles, indicating that there has been no exchange of [ H]DPPC. The C] 3 14 stearylamine counts [ on the other hand are approximately equally distributed between the two vesicle populations. These results clearly suggest that only stearylamine is undergoing exchange in this vesicle system. It was also of interest to incorporate the fluorescent energy transfer probes rhodamine-PE and NBD-PE into the stearylamine and PS-containing  55  Figure 11 Turbidity Measurements of Vesicle Aggregation Turbidity measurements at 550 nm of (A) (- -) DOPC,PS (8:2) vésicles (1 iimol/ml), (B) (— —) DOPC/stearylainine (9.5:0.5) vesicles (1 imol/ml) and (C) ( ) a mixture of 0.5 tmol/ml of the two vesicle populations. The results for (A) and (B) overlap. The.arrow indicates the time at which the SA-containing vesicles were added. -  1  10.8  .  0.6  Ui ()  z  0.4  0 (I)  0.2 0.0 60 80 TIM[ (seconds)  56  140  Figure 12 Characterization of Vesicle Elution from DEAE Sephacel Columns Elution profiles from DEAE-Sephacel columns of (A) [ H]DOPC/PS (8:2) vesicles, (B) 3 C]stearylamine (9.5:0.5) vesicles, (C) a mixture of vesicles in (A) and (B), and 4 DOPC/[’ (D) a mixture of DOPC/PS/NBD-PE (8:2:0.1) and DOPC/stearylamine/Rh-PE (9.5:0.5:0.1) vesicles. (.) 3 H disintegrations, (D) 14 C disintegrations. In panel (D), (.) indicates rhodamine fluorescence (ex. 560, em. 590) and (o) indicates NBD fluorescence (ex. 480, em. 530). The arrow indicates the point at which elution with high ionic strength buffer was started.  j  FRACTION  57  vesicles, respectively, in order to ascertain whether there was evidence of lipid mixing and fusion during aggregation. The results, shown in Fig. 12D, show that these probes do not exchange between the acceptor and donor vesicles during the aggregation stage. Moreover, there was no detection of fluorescence energy transfer during the aggregation-disaggregation reaction.  2.3.2 Effect of a Transmembrane ApH on Stearylamine Exchange Having established that stearylamine rapidly equilibrates between the two vesicle populations in the absence of a transmembrane pH gradient, vesicles of DOPC/SA (9.5:0.5) were prepared with an internal pH of 4.0. Given the external pH of 7.4, a pH  gradient of -3.4 units (calculated as inside pH minus outside pH) is then formed across the bilayer. Previous work has shown that such a gradient induces stearylamine to accumulate at the inner monolayer of the vesicle resulting in a reduction of the outer monolayer concentration of stearylamine by up to three orders of magnitude (Hope & Cullis, 1987). Fig. 13 demonstrates the change in the outer surface charge associated with stearylamine-containing vesicles in the presence of a pH gradient (acidic inside). When DOPC/PS (8:2) and DOPC/stearylamine (9.5:0.5) vesicles are mixed there is no aggregation, consistent with an absence of a positive surface charge on the DOPC/stearylamine vesicles. However, dissipating the gradient using a combination of nigericin, a proton ionophore, and valinomycin, a K+ ionophore, in the presence of K+ ions allows stearylamine to equilibrate across the bilayer. A positive surface charge is then restored and aggregation with the negatively charged PS-containing vesicles is observed as an increase in turbidity. Subsequently, inter-vesicle exchange of stearylamine takes place,  58  and the vesicles disaggregate. This mechanism is confirmed in Fig. 14 which demonstrates that in the presence of a pH gradient (ApH = -3.4), when stearylamine is located at the inner monolayer of the vesicle, exchange of stearylamine between vesicles does not occur. Specifically, [ H]DOPC/PS LUVs were incubated with 3 DOPC/[ C]stearylami 14 ne LUVs (pH 4.0 inside) and after a 5 mm incubation, separation on an ion exchange column shows that 97% of the stearylamine is still associated with the DOPC vesicles. However, as illustrated in Fig. 14B, in the presence of ionophores stearylamine equilibrates with the PS-containing vesicles. It should be noted that exchange of phospholipid label between the vesicle systems was never observed.  2.3.3 Fatty Acid Exchange Between Membranes The results presented above deal with the transfer of stearylamine between vesicles. Vesicles exhibiting a positive ApH (basic inside) sequester fatty acids to the inner monolayer (Hope & Cullis, 1987), and consequently fatty acid exchange between membranes should also be influenced by transmembrane proton gradients. In order to monitor the exchange of fatty acid between two vesicle populations and test this hypothesis we studied the following systems. In the first DOPC/SA (9.5:0.5) vesicles were incubated with DOPC/PS/[ C]OA(8:2:0.5) LUVs. Vesicles containing oleic acid 14 (OA) were prepared with a pH gradient that was basic inside (see Section 2.2.5) yielding a ApH  =  3.0. As we have described above (see Figs. 11 and 13), mixing positively  charged stearylamine-containing vesicles and negatively charged PS-containing vesicles results in immediate aggregation followed by a slower disaggregation due to the equilibration of stearylamine between the vesicle populations. The same phenomenon was observed for DOPCISA and DOPCIPS/[ C]OA vesicle systems. However, as 14  59  Figure 13 Effect of Transmembrane ApH on Vesicle Aggregation Turbidity measurements at 550 nm of a mixture of DOPC/PS (8:2) and DOPC/stearylamine (9.5:0.5) vesicles with a transmembrane pH gradient (ApH = -3.4) DOPCIPS vesicles were added at t = O;’L addition of DOPC/stearylamine vesicles; addition of ionophores (nigericin and valinomycin) to collapse the pH gradient.  4,  E 0  to  u) C,)  TIME (secs..)  60  Figure 14 Effect of a Transmembrane ApH on Stearylamine Exchange Elution profiles of [ H]DOPCIPS (8:2) vesicles and DOPC/[ 3 C]stearylamine vesicles. 14 (A) DOPC/[ Cjstearylamine vesicles have a transmembrane pH gradient (ApH = -3.4) 14 acidic interior. (B) The same vesicle mixture plus the ionophores, nigericin and valinomycin, used to collapse the transmembrane pH gradient. (.) , H disintergrations, 3 (D) 14 C disintergrations.  U  U  0  1  2-  3 FRACTION  61  4  5  6  shown in Fig. 15A, the exchange of oleic acid is greatly reduced by the presence of a positive ApH. It is worthwhile pointing out that despite vesicle aggregation and stearylamine exchange the bulk of the oleic acid remains associated with the inner monolayer of the DOPC/PS/OA vesicles. This suggests that vesicle aggregation and lipid exchange does not significantly enhance the proton permeability of the PS-containing vesicles, otherwise the ApH would collapse enabling oleic acid to exchange. When the ApH is deliberately dissipated employing the ionophores valinomycin and nigericin in the presence of K, C] 14 oleic acid is observed to elute from the DEAE column in both [ vesicle fractions indicating equilibration of the fatty acid between DOPC/SA and DOPC/PS/OA vesicles (Fig 15B). The second approach involved monitoring the transfer of fatty acid from one population of vesicles to another in non-aggregating systems as shown in Fig. 16. LUVs composed of [ H]DOPC were incubated with vesicles of DOPC/PS![’ 3 C]OA 8:2:0.5). 4 The results of Fig. 16A show that in the absence of a ApH oleic acid rapidly equilibrates between the two vesicle populations. However, when the DOPC LUVs exhibit a ApH of 3.0 units (inside basic) oleic acid preferentially moves into the DOPC vesicles (Fig. 16B). At equimolar concentrations of donor and acceptor vesicles, 85% of the oleic acid transfers to the DOPC vesicles. At higher ratios of acceptor to donor vesicles more than 90% of the oleic acid transfers to the DOPC vesicles.  2.3.4 Exchange of Fatty Acids Between Vesicles and BSA The above experiments clearly demonstrate an ability of transmembrane pH gradients to modulate lipid flow between membranes. However, a large proportion of free fatty acid in vivo is delivered to peripheral tissues bound to serum albumin. In the  62  Figure 15 Oleic Acid Exchange in Aggregating Systems Elution profiles from DEAE-Sephacel columns of mixtures of DOPC/SA (9.5:0.5) and C]OA (8:2:0.5) vesicles. (A) The oleic acid-containing vesicles had a basic 4 DOPCIPS/[’ transmembrane pH gradient (ApH = 3.0). (B) The vesicles in the presence of ionophores to collapse the pH gradient. (0) 14 C disintergrations. The arrow indicates the point at which elution with high ionic strength buffer was started.  100  0.  1  0  0  1  2  3  4  Fraction  63  5  6  7  light of our observations on the modulation of fatty acid flow between membranes by transmembrane pH gradients, it was of interest to see whether albumin could be depleted of fatty acid when incubated with a population of vesicles that exhibit a positive ApH (inside basic). Fig. 17 presents the data from such an experiment. When incubated with vesicles composed of DOPC (ApH  =  0) fatty acid equilibrates between protein and  membrane as shown in Fig. 17A. However, in the presence of a positive ApH, net flux of fatty acid occurs in the direction of the vesicles, significantly depleting albumin of lipid (see Fig. 17B).  64  Figure 16 Oleic Acid Exchange in Non-Aggregating Systems Elution profiles from DEAE-Sephacel columns of [ H]DOPC, DOPC/PS/[’ 3 C]OA 4 vesicle mixtures. (A) Neither vesicle population has a transmembrane pH gradient. (B) The [ H]DOPC vesicles have a basic pH gradient (ApH = 3.0). (.) H disitergrations; 3 (D) 1 C disitergrations.  100  a  1  a  0  1  2  3  4  FRACTION  65  5  6  7  Figure 17 Exchange of Oleic Acid Between BSA and Vesicles Extraction of C]OA 14 from fatty acid loaded bovine serum albumin (BSA) by DOPC [ vesicles. Separation of vesicles and BSA on Sepharose CL-4B columns. (A) [ H]DOPC 3 vesicles without a transmembrane pH gradient (ApH = 0.0). (B) [ HjDOPC vesicles with 3 a basic transmembrane pH gradient (ApH = 3.0). (.) 3 H disitergrations, (o) l C 4 disitergrations. BSA elutes between fractions 18-28.  100 A ApHO.0  80  60  40  20 n  1  0.  0  5  10  15  20 25 FRACTION  66  30  35  40  DISCUSSION  The results presented here illustrate the remarkable exchange abilities of stearylamine and oleic acid and the sensitivity of these exchange characteristics to transmembrane pH gradients across vesicle membranes. Here the mechanisms modulating these exchange processes and their implications for lipid exchange in vivo are discussed. The aggregation and subsequent disassociation processes observed for positively charged (DOPC/SA) and negatively charged (DOPC/PS) vesicles provides a graphic illustration of rapid intervesicular exchange of stearylamine. There are three points of interest. First, with regard to the exchange mechanism, whereas the initial attraction between positively and negatively charged vesicles is sufficiently intense to produce visible flocculation under our experimental conditions, neither membrane fusion nor exchange of diacylphospholipids was observed. The lack of phospholipid exchange indicates that intervesicular mixing between external monolayers, commonly observed in fusing systems (Wilschut & Hoekstra, 1986) does not occur and that the bilayers of these aggregated vesicles remain intact. This is also supported by the observation that the proton permeability barrier for acceptor vesicles is maintained during the aggregation disaggregation process involving exchange of stearylamine (Fig. 15). Thus, exchange of stearylamine may be most logically suggested to proceed via intermediary partition of SA into the aqueous phase separating the vesicles. In the case of oleic acid, it is widely accepted that non-mediated lipid exchange occurs via this mechanism, involving desorption of monomers from the bilayer into the aqueous solution, and subsequent  67  diffusion to an acceptor site (Ferrel et al., 1985; Roseman & Thompson, 1980; Nicholls & Pagano, 1982; Doody et al., 1980). This is also supported by the result presented here for non-aggregating systems (Fig. 16) where oleic acid is observed to rapidly equilibrate between the two vesicle populations. The second point concerns the non-exchangeable characteristics of the fluorescent lipids rhodamine-PE and NBD-PE. As indicated in Fig. 12, even under conditions of vesicle aggregation and stearylamine exchange, there is no evidence of lipid mixing as reported by these probe molecules. This provides an additional confirmation of indications that these fluorescent lipids do not readily exchange between aggregated systems (Wilschut & Hoekstra, 1986; Struck et al., 1981), supporting their use as nonexchangeable markers in studies of membrane fusion. The third point concerns the modulation of stearylamine and oleic acid diffusion by transmembrane pH gradients. These phenomena are relatively straight-forward to understand in the light of previous observations (Hope & Cullis, 1987) that lipophilic weak bases, such as stearylamine, will partition to the inner monolayer in vesicles exhibiting a transmembrane pH gradient (inside acidic) (see Fig. 10). Alternatively, lipophilic weak acids such as oleic acid will partition to the inner monolayer when the interior is basic (Hope & Cullis, 1987). Thus the absence of aggregation and lipid exchange when appropriate pH gradients are employed is a graphic consequence of the ApH-dependent lipid asymmetry. Modulation of lipid exchange by transmembrane pH gradients may also occur in vivo. For example, the results presented here show that fatty acid can be induced to transfer from one population of vesicles to another by simply maintaining a basic interior pH in the acceptor vesicles. Such vesicles can also induce fatty acid depletion of bovine  68  serum albumin. Although serum and cytoplasmic pH values are strictly maintained within narrow limits (Deamer, 1982), in organelles such as the lysosome, which maintain an acidic interior (Hamilton & Cistola, 1986; Hope & Cullis, 1987), fatty acid would be expected to mainly reside in the outer (cytoplasmic) monolayer. In addition the adsorption of short chain fatty acids in the digestive tract has been associated with the presence of pH gradients formed by various mechanisms including proton pumps which create acidic microclimates at the epithelial cell surface. Phospholipid asymmetry, common to many plasma membranes  (Op den Kamp, 1979) may also play a role.  Phosphatidylserine, for example, is negatively charged and frequently located in the cytoplasmic monolayer of cell membranes, resulting in a negative surface potential (Cullis & Hope, 1985). This potential will repel anions from and attract cations to the lipid/water interface and will result in a significantly lower pH at the cytoplasmic membrane surface when compared to the exterior bulk pH. Such a gradient could be important in enhancing the flow of fatty acids out of adipocytes, for example. The possibility of fatty acid flow from lysosomes into the cytoplasm because of the acidic pH gradient across the lysosomal membrane has been discussed previously (Hamilton & Cistola, 1986; Hope & Cullis, 1987). Mitochondria, on the other hand, develop a basic internal pH during oxidative phosphorylation (Nichols, 1982). Given the results presented here this might be expected to culminate in an accumulation of free fatty acids by mitochondria. However, cells have evolved a complex system by which fatty acids are first activated and subsequently converted to carnitine derivatives which are then transported into the mitochondrial matrix. The data described here suggest that intracellular free fatty acid may be deliberately kept at a very low concentration to prevent these lipids accumulating within organelles, such as the mitochondria, in an  69  unregulated manner. The potential harm of this type of accumulation is illustrated by the ability of low concentrations of lysosomotropic detergents to kill cells (Miller et al., 1983). These molecules are aliphatic amines which accumulate in lysosomes in response to the pH gradient, disrupt membrane integrity and cause the release of lysosomal enzymes into the cytoplasm.  70  CHAPTER 3 TRANSBILAYER TRANSPORT OF PHOSPHATIDIC ACID IN RESPONSE TO A TRANSMEMBRANE PH GRADIENT  3.1 INTRODUCTION It is now generally accepted that biological membranes exhibit asymmetric transbilayer distributions of their components. For certain lipids (glycolipids & sphingolipids) and proteins, this asymmetry is established during synthesis and maintained throughout the lifetime of the molecule. However, the rate of transbilayer transport of glycerolipids in most biological membranes precludes a purely static mechanism for the maintenance of asymmetry (see Op den Kamp, 1979; Etemadi, 1980; Devaux, 1991; Chapter 1 Section 1.5). Although a large body of evidence now exists to indicate the existence of specific proteins to translocate phospholipids across various membranes (Seigneuret & Devaux, 1984; Bishop & Bell, 1985; Zachowski et al., 1986; Backer & Dawidowicz, 1987; Connor & Schroit, 1988), such proteins have yet to be isolated or positively identified. In addition, translocators for many species of lipids such as phosphatidic acid have not as yet been reported. It is probable that other mechanisms exist in conjunction with specific translocators in order to generate and maintain the asymmetry of various phospholipids. For example, Zachowski et al. (1985) have suggested that the asymmetry of the aminophospholipids may provide the driving force for PC asymmetry in erythrocytes. Furthermore, it has been suggested that non-bilayer phase lipids, the presence of integral membrane proteins and ion gradients may play roles in the induction and maintenance of lipid asymmetry.  71  In order to investigate the mechanisms by which lipid asymmetry is generated and maintained, model membrane systems are useful. A major emphasis of this study concerns the influence of ion gradients, especially pH gradients, on the transbilayer distribution of lipids in large unilamellar vesicles (LUVs). Previous studies have shown that the transbilayer distribution of simple lipids with weak acid or base characteristics, such as fatty acids and stearylamine (Hope & Cullis, 1987; Eastman et a!., 1989; Chapter 2), and certain acidic phospholipids (phosphatidyiglycerol & phosphatidic acid) (Hope et al., 1989; Redelmeier et al., 1990), can be modulated by the presence of transmembrane pH gradients. The kinetics of ApH driven transport of phosphatidyiglycerol has been examined in detail (Redelmeier et al., 1990). The results indicated that PG is transported in the neutral (protonated) form which could exhibit half-times for transbilayer movement on the order of seconds. However, the assays used to study the transbilayer distribution of PG were either specific for PG (periodate oxidation) or involved the costly synthesis of 13 C labelled phospholipids. A quantitative analysis of the transport of phosphatidic acid was precluded by the lack of an appropriate assay for PA asymmetry. This chapter describes the use of a new fluorescent assay employing the probe 2-(ptoluidinyl)naphthalene-6-sulfonic acid (TNS) to detect PA asymmetry. TNS is a fluorescent probe previously used to report on the surface potential of membranes (McLaughlin & Haray, 1976; Eisenberg et al., 1979; Searle & Barber, 1979). This assay is shown to be useful as a general assay to detect the transbilayer distribution of acidic lipids. Employing TNS, the kinetics of ApH driven DOPA transport were examined, showing that DOPA is transported via the neutral form with an activation energy similar to that observed for PG.  72  3.2 MATERIALS AND METHODS  3.2.1 Lipids and Chemicals. All phospholipids were obtained from Avanti Polar Lipids (Peiham, AL) and were used without further purification. These included dioleoylphosphatidylcholine (DOPC), dioleoylphosphatidic acid (DOPA), dioleoylphophatidylserine (DOPS), bovine liver phosphatidylinositol (P1), and bovine heart cardiolipin (CL). All lipids were in the Na salt form. TNS [2-(p-toluidinyl)naphthalene-6-sulfonic acid] and all buffers were obtained from Sigma Chemical Co. (St. Louis, MO.).  3.2.2 Preparation of Large Unilamellar Vesicles. Large unilamellar vesicles (LUVs) of the desired lipid compositions were prepared in the appropriate buffers by extrusion techniques as described in section 2.2.2.  3.2.3 Induction of Transbilayer Transport of Acidic Phospholipids. Vesicles were prepared in 300 mM EPPS, pH 9.0, and passed down Sephadex G 25 columns (Pharmacia) equilibrated in 150 mM / 4 S 2 Na 1 O mM EPPS, pH 9.0. At this point, phosphate assays (Fiske & Subbarow, 1925) were performed on the vesicle preparations, and the vesicles were diluted to a concentration of 10.5 mM total phospholipid. Vesicles were then placed in test tubes and preheated to an appropriate temperature. At time t  =  0, an equal volume of a preheated 100 mM citrate buffer, pH  4.0, was added to each tube to obtain the desired ApH. The tubes were then incubated for appropriate times, and PA transport was quenched by placing 200 itL of the sample into test tubes containing 500 iL of ice-cold 100 mM ammonium acetate/100 mM sodium  73  citrate, pH 6.0, and the sample was then stored on ice until assayed for asymmetry. The transbilayer movement of PA or PG under these conditions is negligible as the pH gradient is reduced by the presence of acetate and transport is extremely slow at 0°C (see Results). For “zero time” time points, 100 tL samples were removed from the test tubes before the citrate (pH 4.0) was added and placed into the ice-cold acetate and citrate mixture (500 iL). To measure PA movement to the outer monolayer, the same procedure was followed except that the vesicles were hydrated in 300 mM citrate, pH 4.0. Untrapped buffer was exchanged for 150 mM I 4 S 2 Na O mM citrate, pH 4.0, employing column 1 chromatography, and the external pH was subsequently adjusted by the addition of 100 mM EPPS, pH 9.0.  3.2.4 Detection of Phosphatidyiglycerol Asymmetry by Periodate Oxidation The transmembrane asymmetry of PG was assayed by two methods. The first method involved periodate oxidation as previously described (Lentz et al., 1980; Hope et al., 1989). Briefly, the PG on the external monolayer of the vesicles was oxidized by the addition of 100 ii.L of freshly prepared 100 mM sodium periodate to each sample. To assay the total amount of PG in the vesicles, 50 1 iL of 200 mM sodium cholate was added to the sample before the addition of the periodate. The oxidation of the PG was quenched after 11 mm by the addition of 100 iL of 1 M sodium arsenite in 1 N 4 S0 2 H . The formaldehyde resulting from the oxidation of the glycerol was detected by the Hantzsh reaction (Nash, 1953).  74  3.2.5 Detection of Asymmetry Using TNS. Following the establishment of PG or PA asymmetry, two samples (200 i1 L) were removed and placed into test tubes. To each of these samples 3 mL of 3 i.M TNS  inS  mM ammonium acetate /5 mM HEPES (pH 7.0) was added and the samples were mixed. The fluorescence of the samples was then measured employing either an SLM Aminco SPF 500C or a Perkin Elmer IS-SO fluorometer using an excitation wavelength of 321 nm and an emission of 445 nm. To determine asymmetry, two sets of vesicles were utilized. The first contained only PC while the second contained a well-defined mixture of PC and the acidic lipid being assayed. The acidic lipid made up a maximum of 5% of the total lipid since it was found that the TNS fluorescence varied linearly with acid lipid content over the range 06% PA or PG (see Figure 19). Due to this linearity, the percentage of acidic lipid in the outer monolayer could be calculated according to x  =  -f(PC)]} X 0 {[f-f(PC)]/[f , where f 0  is the TNS fluorescence for the sample for which asymmetry is assayed, f(PC) is the fluorescence of the sample containing no acidic phospholipid, f 0 is the fluorescence associated with the sample prior to induction of asymmetry, and X 0 is the mole percent of acidic phospholipid in the LUVs.  3.2.6 Measurement of the Internal pH of LUVs. The internal pH was monitored employing entrapped pyranine according to the method of Rossignol et al. (1982). This first required construction of a standard curve, utilizing the LUVs with an external buffer of 150 mM 4 SO / 1 mM EPPS (pH 9.0) 2 Na containing 1 mM pyranine and diluted to a concentration of 10 mM total lipid. To this dispersion was added an equal volume of a buffer with a pH in the range of 5.0 9.0. -  75  This buffer was capable of buffering over the range pH 5.0 9.0, containing 150 mM -  SO 20 mM MES, 20 mM PIPES, 20 mM HEPES, and 20 mM EPPS. In order to 2 Na , 4 ensure that the internal pH was the same as the external pH, 1 tM nigericin and 1 tg/mL gramicidin was also present. The fluorescence was then monitored at 45°C) by employing excitation wavelengths of 405 and 463 nm (emission wavelength 511 nm), and the ratio of Ip to 1405 vs pH was employed to produce the pH titration curve. This was utilized to obtain the internal pH of vesicles with a transmembrane pH gradient (external pH 4.0) by monitoring 1463/1405 as a function of incubation time at 45°C.  3.2.7 Kinetic Analysis of Phosphatidic Acid Transport. The analysis of the kinetics of DOPA transport follows a model for the transport of acidic phospholipids across lipid bilayers in response to a transmembrane pH gradient previously described in Redelmeier et al., 1990. This model assumes that only the neutral (protonated) form of the acidic phospholipid is able to move across the membrane. Thus the net inward flux (net) of phosphatidic acid is expected to be a function of the concentration gradient of the neutral species across the membrane, the membrane area and the permeability coefficient.  dN(A) t 0t / dt net = 0  =  P Am 1 [M4] 0 ([M] )  (1)  t0t is the total number of PA molecules in the outer monolayer, [AH] is the 0 where N(A)  (surface) concentration of the neutral form of the acid, P is the permeability coefficient, Am is the area of the membrane and the subscripts o and i represent the outer and inner monolayer respectively.  76  From the law of the conservation of mass  N(A) t 0 0t I Am  =  0 + N(AH) (N(A) ) / Am 0  =  0 [Aj  +  0 [MI]  (2)  and from the acid dissociation constant  3 K  =  0 0 [H] [Aj 0 I [AH]  (3)  Rearranging equation 3 and substituting into equation 2 for [A-] gives  N(A) t 0 0t I Am  =  (1  +  3/0 K [H] [AH] ) 0  (4)  Substituting equation 4 into equation 1 gives  (dN(A) t 0 0t I dt) I Am  (1  +  0 I dt 3 I [H] K ) d[AH] 0  If it is assumed that K>> [H] , [H] 0 1  <<  0 and that the concentration of PA in the [H]  inner and outer monolayers are equal at time t  (dN(A) t 0 0t I dt) I Am  =  (5)  =  3I0 (K [H] 0 ) d[AH] I dt  Therefore:  77  0, then  P [AH] 0  (6)  Idt=-P pL[AHLI;=-kvJq 0 d[AH] 0 where  (8)  An analytical solution to 7 gives:  (9)  [AH(t)b=[AH(0)Le+C  Since [AH(t)J , Is proportional to [A(t)J(, 4  , = [k(0)L et + C 1 [Att)J  (10)  If ft Is assumed that the exponential decay Is to some equilibrium value [A’(eq)J:  ,=C 1 [A(eq)J  (11)  Substituting equation 11 Into equation 10 glves  ) I [k(0)], = e 0 ([k(t)L [k(eq)]  (12)  hi ([k(t)J , [k(eq)]) I [k(0)J, = -kt 1  (13)  -  -  78  Thus a plot of in ([A(t)] [A(eq)] I [A(O)J ) 0 0 0 vs t should yield a straight line -  with slope k and units of V . 1  3.3 RESULTS  3.3.1 TNS fluorescence Assay of Asymmetry. Asymmetry of acidic phospholipids in LUVs composed of PC/acidic lipid mixtures can be detected by ion-exchange chromatographic techniques (Hope & Cullis, 1987; Hope et al., 1989). Unfortunately, this technique does not provide quantitative measures of asymmetry. In the case of PG, 13 C NMR studies on 13 C-labelled varieties (Hope et a!., 1989) or chemical assays specific for PG (Redelmeier et al., 1990) provide more quantitative information. In order to facilitate asymmetry studies on PA and other acidic phospholipids a more general and flexible assay for the presence of such lipids in the outer monolayer of LUV systems was required. An obvious approach is to monitor the surface potential of the outer monolayer, which will reflect the presence of negatively charged phospholipids. Studies to investigate the utility of TNS, a probe of membrane surface potential introduced by McLaughlin and co-workers (McLaughlin & Haray, 1976; Eisenberg et a!., 1979) as a probe of asymmetry were therefore conducted. In this regard, it should be noted that TNS is a fluorescent lipophilic weak acid (pKa = 4) which exhibits enhanced fluorescence when associated with a lipid biiayer. Thus, under the assay conditions employed here, the presence of acidic lipids in the outer monolayer of the LUVs will result in decreased partitioning of the negatively charge probe into the bilayer and correspondingly reduced fluorescence intensity. This effect is illustrated in Figure 18 for 100 nm DOPC/DOPA LUVs containing 0 10 mol% DOPA. It may be -  79  noted that the decrease in fluorescence intensity with PA content is linear over the range  o  -  6 mol% DOPA, and thus most asymmetry experiments were restricted to this range  for ease of analysis. Additional control experiments to establish the utility of the TNS assay were required, however. This is due to the weak acid characteristics of TNS, which would suggest that it could be accumulated into LUVs exhibiting a basic interior due to permeation of the neutral form. Such accumulation would be expected to result in enhanced fluorescence intensity arising from increased partitioning of the probe into the interior lipid monolayer due to the small aqueous to lipid volume ratio in the LUV interior. Behavior corresponding to this effect is shown in Figure 19 for 100 nm DOPC LUVs with a interior pH of 9.0 (300 mM EPPS) and an exterior pH of 4.0. However, it is also shown in Figure 19 that dissipating the transmembrane pH gradient by raising the exterior pH and adding 100 mM ammonium acetate to the exterior medium eliminated such effects.  3.3.2 Comparison of the TNS Assay to a Chemical Assay In order to further validate the TNS asymmetry assay for acidic phospholipids, a direct comparison with the periodate assay for PG asymmetry was performed in DOPC/DOPG (95:5, mol/mol) LUV systems. As shown in Figure 20, a very similar rate and extent of PG asymmetry were reported by both procedures.  80  Figure 18 Standard Curve of TNS Fluorescence as a Function of DOPA concentration in DOPC/DOPA LUVs DOPCIDOPA LUVs (100 nm) were prepared from lipid dispersions containing 0 10 mol% DOPA employing a protocol similar to that employed for inducing asymmetry (see Section 3.2.3), with the exceptions that the heating step was omitted and the ammonium acetate buffer (pH 6.0) was added prior to the citrate buffer (pH 4.0) to avoid generating a pH gradient. Fluorescence is expressed as a ratio of the fluorescence measured for the DOPA containing vesicles with respect to that observed for pure DOPC LUVs. Error bars indicate standard deviations from three sets of triplicate results. -  1.0  0.8 C, C 0 0  0.6  1  0 li  0.4  0.2 0  2  4  6  % Phosphatidic Acid  81  8  .  10  Figure 19 Influence of a Transmembrane pH Gradient on TNS Fluorescence Influence of a transmembrane pH gradient (interior basic) on the fluorescent response of TNS in DOPC LUVs. Vesicles (100 nm diameter) were prepared in EPPS buffer (pH 9.0), and the exterior buffer was exchanged for the 4 SO buffer (see materials and 2 Na methods). A transmembrane pH gradient was then generated by addition of 100 [it of the citrate buffer (pH 4.0) to 100 L of the vesicle solution (10.5 mM phospholipid). The pH gradient was then either quenched or maintained by addition of either 500 jL of the ammonium acetate buffer (pH 6.0) or 500 iL more of the citrate buffer (pH 4.0). The TNS response was then monitored following addition of 10.5 mL of a 3 pM TNS solution. (A) indicates the response for the unquenched system at 45°C; (A) the quenched system at 45°C; (o) the unquenched system at 22°C; (.) the quenched system at 22°C. The fluorescence is expressed as a ratio of that observed at t = 0 to that obtained at time t. Error bars indicate standard deviations from three sets of triplicate results.  V C., C V C, C,, V I  0 D Li  0  20  40  60  Time  (miri)  82  80  3.3.3 Kinetic Analysis of PA Transport. Initial experiments on PA transport in response to ApH were designed to monitor the time course of ApH-induced PA asymmetry as assayed by TNS and to test the applicability of the kinetic analysis employed elsewhere [see Section 3.2.7 and Redelmeier et a!. (1990)]. As shown in Figure 21A, the presence of ApH, interior basic 0 (pH  =  4.0, pH 1  =  9.0), in 100 nm DOPC!DOPA (95:5 mol/mol) LUVs results in the  depletion of DOPA in the outer monolayer to approximately 5% of the original content after a 30 mm incubation at 45°C. This corresponds to a DOPA content in the inner monolayer of 9.3 mol% and an exterior DOPA content of 0.26 mol%, which is nearing the detection limits of the TNS assay. It is interesting to note that the maximum ApH induced PA asymmetry detected ([PA] 1 I [PA] 0 = 39) is considerably greater than that detected for DOPG under similar conditions, where a maximum inside:outside ratio of 3 was obtained (Redelmeier et al., 1990). The kinetic analysis employed here assumes that [PAH] >.>.[PAH], where PAH 0 represents the neutral form of PA. It is therefore important that the pH 1 remains sufficiently high to satisfy this condition. Employing entrapped pyranine, as described in the last part of this section, the pH 1 at 30 mm was measured to be 7.3, indicating that 0 / [PAH] [PAH] 1  —  50, which satisfies this demand. It may be noted that the decrease in  1 from pH 9 to pH 7.3 during PA transport cannot be accounted for by the import of pH associated protons. It may be suggested that packing problems resulting from the import of PA may result in shape changes and partial lysis, resulting in release of buffer. However, freeze-fracture studies revealed no difference in shape induced by the lipid asymmetry (results not shown), and the lack of lysis was indicated by the lack of release  83  Figure 20 Comparison of the TNS Assay for Lipid Asymmetry to a Chemical Assay (Periodate Oxidation) for DOPC/DOPG LUVs DOPG asymmetry assayed by TNS (.) and periodate assay (o) procedures. Vesicles (100 nm diameter) containing 5 mol% DOPG (in DOPC) were prepared as indicated under materials and methods to exhibit a transmembrane pH gradient (pH 1 = 9.0; pH 0 = 4.0) and subsequently incubated at 45°C for the times indicated. For details of the TNS and periodate assay procedures, see Sections 3.2.4 and 3.2.5. Error bars indicate standard deviations from three sets of duplicate results.  L  50N  40  10  1’oo31o5’o6o Time (mm)  84  of entrapped radiolabelled citrate during PA import. It is possible that the reduction in pH arose in part due to leakage of internal Na+ ions, which would allow the inward movement of protons. As indicated under Materials and Methods, assuming that only the neutral (protonated) form of PA is transported, a plot of in {[[PA(t)] 0 0 [PA(eq)J / [PA(0)] ] } vs 0 -  time should yield a straight line with slope k, where the half-time (t p for transbilayer 1 movement of the PA is given by 2 ,, = 0.693 k 1 t . As shown in Figure 21B, a very good 4 linear fit employing the data of Figure 21A could be achieved employing k and [PA 0 as variables. This analysis results in a rate constant of 1.67 x 10-i mm (eq)] , 4 corresponding to a half-time for transbilayer movement (t ) of DOPA of 4.1 mm at 112 45°C.  3.3.4 Influence of pH and Temperature on PA Transport. The kinetic analysis of PA transport employed here assumes that the neutral (protonated) form of PA is the permeating species. Indeed, it is generally accepted that weak acids and bases permeate through membranes in the neutral form [see, for example, Gutnecht and Walter (1981b)J. However, it has been reported that weak bases can cross lipid bilayers in the charged form (McLaughlin, 1975) and that fatty acids can act as proton ionophores (Gutnecht & Walter, 1981a) which implies they can move across the membrane in both the charged and uncharged form. As previously shown for PG (Redelmeier et a!., 1990), determination of the rate constant k as a function of exterior pH provides a method for unambiguously determining whether the neutral form is the primary permeating species, as k should vary linearly with exterior proton concentration via the relation k = 0 [W] kn/Ka if this is the case. As shown in Figure 22A, the rate of  85  PA transport was found to be strongly dependent on the exterior pH, increasing by nearly an order of magnitude for every unit pH 0 is lowered. A plot of log k vs log [H] 0 (Figure 22C) reveals a straight line with a slope 0.9 +1- 0.05, strongly indicating that PA is  permeating the membrane in the neutral (protonated) form. The temperature also strongly influenced the rate of DOPA transport. An analysis of transport rates over the range 25 45°C revealed that the transport rate increased -  nearly 5-fold for every 10CC increase in temperature as shown in Fig. 23A. An Arrhenius plot of the rate constants (Fig 23C) indicated an activation energy for DOPA transport of 28 kcal/mol or 117 id/mo!.  3.3.5 Transport of DOPA to the Outer Monolayer The results to this point demonstrate transport of DOPA from the outer monolayer to the inner monolayer of LUVs with a basic interior. Conversely, it would be expected that in LUVs with an acidic interior, PA should move from the inner to the outer monolayer. This behavior is illustrated in Figure 24 where it is shown that the percentage of DOPA in the outer monolayer increases from 50% of the total DOPA to more than 85% over a 1 h time course for DOPCIDOPA (95:5, mol/mol) LUVs with an initial interior pH of 4.0 and an exterior pH of 9.0.  86  Figure 21 Transbilayer Transport of DOPA in Response to a Transmembrane pH Gradient (Inside Basic) (A) Influence of a pH gradient (interior basic) on the transbilayer distrbution of DOPA in DOPC/DOPA (95:5 mol/mol) LUVs. Vesicles were prepared as indicated under materials and methods (pH 0 = 4.0) and incubated for the times indicated prior 1 = 9.0; pH to quenching transport. The amount of DOPA remaining in the outer monolayer was assayed employing TNS as described in Section 3.2.5. (B) Best fit to these data employing the kinetic analysis descrbed in materials and methods. From the slope of this plot, the rate constant k can be determined to be 1.67 X 10-1 mm . 4  L >. C 0 C 0  V  0 C  a 10  Time (miri)  15  Time (miri)  87  Figure 22 Influence of the External p11 on the Rate of the Transbilayer Transport of DOPA Vesicles, DOPC/DOPA (95:5 mol/mol) were prepared in 300 mM NaEPPs, pH 9.0 buffer S1 2 Na I O mM EPPS pH 9.0 (see and the exterior buffer was exchanged for 150 mM 4 Section 3.2.3). The vesicles were then introduced into citrate solutions with varying pH: 0 = 3.5. After 0 =4.0; (a) pH 0 = 4.5; (A) pH 0 = 5.0; (A) pH (o) pH 0 = 5.5; (.) pH the amount of quenched and incubation at 40°C for the indicated times, transport was Best fit to this TNS. (B) DOPA remaining in the outer monolayer assayed employing vs the log Plot of log k data employing the kinetic analysis in materials and methods. (C) of the external proton concentration.  >-  a C 0  a C  a K  Tim. (mtn)  Tim. (mm)  0 log [H+1  88  3.3.6 Response of Various Phospholipids to a Transmembrane ApH As indicated above, the TNS assay for asymmetry should be generally applicable to determine the transbilayer distributions of a variety of acidic lipids in addition to PA and PG. We have therefore employed this assay to determine possible ApH-induced asymmetry in DOPC systems containing 3 mol% DOPS, 3 mol% bovine heart cardiolipin and 5 mol% bovine liver P1. As shown in Figure 25A, no ApH-induced asymmetry could be detected at 45°C for these phospholipids under conditions similar to those for which DOPA and egg PG exhibit considerable transbilayer movement. This is in agreement with preliminary observations described elsewhere (Hope et al., 1989). As shown in Figure 25B, this inability to induce asymmetry for P1, PS and cardiolipin does not arise from depletion of the transmembrane pH gradient, as ApH values in excess of 2 units are maintained over the experimental time course.  89  Figure 23 TeLnperature Dependence of ApH Driven DOPA Asymmetry Vesicles, DOPCIDOPA (95:5 mol/mol), were prepared in 300 mM NaEPPS as previously described. The vesicles and the citrate buffer (pH 4.0) were pre-incubated to the appropriate temperature before establishing the transmembrane DpH. After incubation for the times indicated, transport was terminated and the amount of DOPA remaining in the outer monolayer was measured employing TNS (see Section 3.2.5). (o) 25°C; (.) 30°C; (A) 35°C; (A) 40°C; (ci) 45°C. (B) Best fit of data to the kinetic analysis described in Section 3.2.7. (C) An Arrhenius plot of the rate constants derived from B.  0 0  .<  o -3.0  C  flmi (miti) C  —6.0 —65 —7.0 —7_S C  —6.0 —6.3 —LO —15  —  3.00  3.10  3.70  10 / 3 r  3.30  ) 1 (IC  90  340  3.30  Figure 24 Transport of DOPA to the Outer Monolayer in Response to a Transmembrane pH Gradient (Interior Acidic) Vesicles, DOPCIDOPA (97.5:2.5 mol/mol), were prepared in 300 mM citrate (pH 4.0) and the external buffer was exchanged for 150 mM Na SOdl mM citrate (pH 4.0). At 2 zero time, outward DOPA transport was initiated by the addition of 100 mM EPPS (pH 9.0). Transport was quenched after incubation for the indicated times at 40°C by addition of 100 mM ammonium acetate/100 mM citrate buffer (pH 6.0) precooled to 0°C. DOPA transport was assayed employing the TNS protocol.  95 1.  >-  55../  45.  o  1 10  I 20  30  Time (mm)  91  40  50  60  Figure 25 Effect of a Transmembrane ApH on the Transbilayer Distributions of Various Acidic Phospholipids (A) Transbilayer distributions of acidic phospholipid following incubation at 45°C in the presence of a transmembrane pH gradient (pH 1 = 9.0, pH 0 = 4.0). The LUVs were prepared, pH gradients were established, and the amount of acidic lipid in the outer monolayer was assayed as indicated in Section 3.2.5. The lipid compositions corresponding to various symbols are as follows: (a) DOPCIDOPA (95:5); (A) DOPC/EPG (95:5); (A) DOPCIDOPS (97:3); (.) DOPC/bovine heart cardiolipin (97:3); (o) DOPC/bovine brain P1(95:5). (B) Residual pH gradients detected employing pyranine as a probe of internal pH. Vesicles were prepared as described earlier except that 1 mM pyranine was added to the internal buffer. The internal pH of the vesicles was monitored (see Section 3.2.6) by measuring the fluorescence of samples using excitation wavelengths of 405 and 463 nm and an emission wavelength of 511 nm. C  >..  a 0 C 0 I C  0 C  0. 0 0. 0  a ci  -a ci  Time (mn) B L5  x  0  a C  1. C  C  6S  10  20  I 30  I 40  50  Time  (mm)  92  60  70  I 80  90  3.4 DISCUSSION  The results of this Chapter establish TNS as a useful probe for determining transbilayer distributions of acidic lipids in LUV systems and provide new information on the mechanism and kinetics of the transbilayer movement of PA. With regard to the TNS assay, the obvious advantages are convenience and generality. Tedious syntheses to achieve 13 C-labelled or spin-labelled varieties of acidic phospholipids are avoided, the behavior of the acidic phospholipid itself (rather than a labelled variety) is detected, and the assay is relatively rapid. A potential disadvantage of the TNS assay is that the asymmetry to be assayed must be relatively stable. This is clearly not a problem for the DOPA asymmetries assayed here. The half-time for transbilayer movement of DOPA at 20°C and pH 6.0 can be estimated from the results of this study as 12.6 days (see below). Such stable asymmetries are not available for other lipids which are weak acids, such as fatty acids, which exhibit much faster transbilayer diffusion rates (Hope & Cullis, 1987). With regard to the mechanism of DOPA transport in response to transmembrane pH gradients, the results of this investigation strongly support a first-order process involving permeation of the neutral (protonated) form. The linear dependence of the rate [H+] provides particularly compelling constant on the exterior proton concentration 0 evidence in this respect. This behavior is consistent with that previously documented for egg phosphatidylglycerol (EPG) and DOPG (Redelmeier et al., 1990) and the generally accepted view that weak acids permeate through membranes in the neutral form (Gutnecht & Walter, 1981b). Within this formalism, the rate constant k can be written as k  =  [H+lokn/Ka where k is the rate constant for transbilayer movement of the neutral  form and Ka is the weak acid dissociation constant. Given that k  93  =  1.67 x 10-1 min - at 1  45°C for DOPA in LUVs with an exterior pH of 4.0, we obtain k 1,  corresponding to k  -  =  (1.67 x 3 10 ) Ka min  t (t 1.67 min 112 = 25 s), assuming a Ka for DOPA of i-  (Tocanne & Teissie, 1990). This is somewhat smaller than, but comparable to, the rate constant for the neutral form of PG under similar conditions (K = 6 min ; Redelmeier et t al., 1990). A more precise comparison is difficult to achieve given the variability in Ka values reported for PG and PA (Tocanne & Teissie, 1990). The high activation energy (28 kcal/mol) observed for DOPA transport is similar to that observed for EPG (31 kcal/mol) and likely reflects requirements related to dehydration of the phospholipid headgroup, as discussed elsewhere for PG (Redelmeier et al., 1990). In this regard, the similarity between PA and PG activation energies clearly establishes the (protonated) phosphate group as the dominant impediment to transbilayer diffusion. It should also be noted that the combination of a high activation energy and the linear dependence on the proton concentration imparts an exquisite sensitivity of the rate constant for transbilayer movement of DOPA (and PG) to the experimental temperature and pH. Given k = 1.67 x 10-1 mind at 45°C, pH 4.0, a generalized rate constant for DOPA can be written as k(T,pH)  =  1.67 x 3 i0 -pHexp [44.3(1 318jT)] min 1 -  where T is temperature (in Kelvin). Thus, at pH 4 and 600C, the rate constant can be calculated to be 1.23 mind (t 112 = 34 s), whereas at pH 7.0 and 20°C k 112 (t  =  =  3.8 x 10-6 min 1  127 days). This obviously allows the preparation of LUVs exhibiting stable  asymmetric transmembrane lipid distributions following a brief incubation at low pH and/or high temperature to induce the asymmetry. Aside from the fact that this provides convenient conditions for assaying asymmetry as indicated above, such systems are of potential utility in their own right. Two areas of interest concern the influence of lipid  94  asymmetry on membrane fusion processes and the influence of lipid asymmetry on the transbilayer movement of other lipids. The inability to induce asymmetry for other acidic phospholipids (bovine brain P1, bovine heart cardiolipin, and DOPS) is of interest. In the case of DOPS, this can be attributed to the zwitterionic nature even when both acidic functions are protonated. The lack of response of P1 and cardiolipin is surprising, and likely reflects the influence of the bulky polar inositol group and/or low phosphate pKa values. In summary, this investigation establishes TNS as a useful probe of asymmetric transbilayer distributions of acidic phospholipids in LUV systems. Application of this assay to monitor the ApH-dependent transport of DOPA in LUV systems indicates that DOPA traverses the membrane in the neutral form which exhibits transbilayer redistribution times which can be on the order of seconds. Finally, the sensitivity of the rate constant for transport of pH and temperature allows the generation of systems with relatively stable asymmetric distributions of PA.  95  CHAPTER 4 INFLUENCE OF LIPID ASYMMETRY ON FUSION BETWEEN LARGE UNILAMELLAR VESICLES  4.1 INTRODUCTION The asymmetric transbilayer distributions of lipids commonly observed in biological membranes (see Section 1.5 and Chapter 3) may be expected to play a role in regulating membrane fusion in vivo. Model membranes composed of unsaturated phosphatidylethanolamine (PE) and phosphatidylserine (PS), approximating the inner monolayer composition of the erythrocyte membrane, for example, fuse readily in the presence of physiological stimuli such as Ca 2 (Hope et al., 1983). Alternatively, vesicles composed of phosphatidylcholine (PC) and sphingomyelin, the outer monolayer composition, are resistant to fusion. It may therefore be expected that membranes whose external monolayers contain fusogenic lipids such as PE and PS will fuse more readily than membranes with identical lipid compositions but where the fusogenic lipids are localized to the inner monolayer under conditions where the fusion stimulus is in contact with the outer monolayer. These speculations are supported by several observations. For example, Sessions & Horowitz (1981, 1983), have shown that the lipid composition of the external leaflet of the plasma membrane of myoblasts, which undergo fusion to form myotubes, contain more phosphatidylethanolamine and phosphatidylserine than the outer monolayer of the erythrocyte. Further, the concentrations of these lipids in the outer monolayer increases prior to fusion (Santini et al., 1990). It may also be noted that erythrocytes which have lost lipid asymmetry fuse more readily than erythrocytes exhibiting asymmetric lipid distributions with PE and PS in the inner monolayer (Tullius  96  et al., 1989). Alternatively, in most cells, the inner monolayer of the plasma membrane contains more PS and PE than the outer monlayer. Since fusion events are nescessary for exocytosis to occur, it would be expected that the inner monolayer of these cells should be capable of undergoing fusion under appropriate conditions. However, cell to cell fusion occurs only in very specific circumstances (some of which are mentioned above) which would be consistent with a lipid composition of the the outer monolayer of the majority of cell types which is relatively resistant to fusion. The regulatory role of lipid asymmetry in fusion has proven difficult to investigate, due primarily to the lack of an appropriate model system. However, recent work has shown that lipid asymmetry can be generated in large unilamellar vesicles (LUVs) by imposing transmembrane pH gradients (Hope et al., 1989; Redelmeier et al., 1990; Eastman et al., 1991, Chapter 3). Here we utilize this phenomenon to investigate the role of lipid asymmetry in the regulation of Ca 2 induced membrane fusion. It is shown that lipid asymmetry can profoundly regulate fusion phenomena between LUV systems and that the composition of the outer monolayer plays a dominant role in determining the rate and extent of fusion in the system studied here.  4.2 MATERIALS AND METHODS  4.2.1 Lipids and Chemicals Dioleoylphosphatidylcholine (DOPC), dioleoylphosphatidylethanolamine (DOPE), dioleoylphosphatidic acid (DOPA), bovine liver phosphatidylinositol (P1), N-(7nitro-2, 1,3-benzoxadiazol-4-yl) phosphatidylethanolamine (N-NBD-PE), and N (lissamine rhodamine B sulfonyl) phosphatidylethanolamine (N-Rh-PE) were obtained  97  from Avanti Polar Lipids (Peiham, AL). TNS [2-(p-toluidinyl) naphthalene-6-sulfonic acid] and all buffers were obtained from Sigma Chemical Co. (St. Louis, MO).  4.2.2 Preparation of Large Unilamellar Vesicles Vesicles of the appropriate lipid composition were prepared in the appropriate buffers by extrusion procedures described in Section 2.2.2. During the freeze-thaw cycles, the temperature of the water used to thaw the vesicles was maintained below 20°C in order to prevent the vesicles undergoing a phase transition to the hexagonal 1 H 1 phase due to the presence of DOPE.  4.2.3 Detection of Fusion Vesicle fusion was monitored using resonance energy transfer (RET) as described by Struck et al (1981) (see Fig. 8B). Briefly, unlabelled vesicles of the appropriate lipid composition were mixed with similar vesicles containing 0.7 mol% of each of NBD-PE and Rh-PE, in a 3:1 ratio. The vesicles were prepared in 300 mM HEPES pH 7.5 and were subsequently run down a Sephadex G-25 column pre-equilibrated with 300 mM sucrose, 1 mM HEPES, pH 7.5. All buffers were adjusted to the appropriate pH with arginine free base. The vesicles were then diluted to a concentration of 10 mM total phospholipid. A small aliquot of the vesicles (25 L) was added to a cuvette containing 1.9 mL of 100 mM sucrose, 50 mM MES, pH 5.5. Fusion was initiated by the injection of 80 pL of a 200 mM CaC1 2 solution to obtain a final Ca 2 concentration of 8 mM. The increase in NBD-PE fluorescence due to fluorescence dequenching as the fluorophores exchanged with lipids on the unlabelled vesicles was monitored using a Perkin Elmer LS 50 spectrofluorimeter equipped with a thermostated, stirred cuvette holder. An excitation  98  wavelength of 465 nm and an emission wavelength of 535 nm were employed and a cutoff filter for light less than 530 nm was used to reduce the effects of light scattering. Zero fluorescence was set at the residual fluorescence of the NBD-PE in the labelled vesicles while 100% fluorescence corresponded to complete mixing of the lipids. Using a 3:1 ratio of unlabelled to labelled vesicles, complete mixing yields a fluorescence intensity of 75% of the infinitely diluted probe. The fluorescence intensity for the infinitely diluted probe was obtained by the addition of Triton X-100 to a final concentration of 0.1 mM and correcting for its effects on the quantum yield of NBD fluorescence by measuring the decrease in fluorescence of vesicles containing 0.7 mol% NBD-PE after the addition of TX-100 to a 0.1 mM concentration. + concentration on the fusion of the LUVs a 2 In order to determine the effect of Ca sufficient amount of a stock solution of 200 mM CaC1 2 was added to a cuvette containing 100 mM sucrose, 50 mM MES, pH 5.5 to give the desired Ca 2 concentration in a final volume of 2 mL. A 50 L aliquot of the LUVs containing 10 mol% PA (5 mM total phospholipid), was injected into the cuvette and the increase in NBD fluorescence due to fusion was monitored as described above.  4.2.4 Induction of DOPA Asymmetry LUVs containing 10 mol% DOPA were prepared in 300 mM HEPES, pH 7.5 and passed down a Sephadex G-25 column equilibrated with 300 mM sucrose, 1 mM HEPES, pH 7.5. In order to initiate the transport of phosphatidic acid to the inner monolayer, these vesicles (10 mM total phospholipid) were mixed with 100 mM sucrose, 20 mM citrate, pH 4.0 (with arginine). The ability of Ca 2 to induce fusion in these systems was  99  monitored at various time intervals ranging from t  =  0 to t  =  6 h using the RET fusion  assay. The transport of DOPA to the inner monolayer was monitored using TNS as described in Section 3.2.5 with slight modifications. Briefly, unlabelled vesicles containing 10 mol% PA were prepared in 300 mM HEPES, pH 7.5 and passed down a Sephadex G-25 column equilibrated in 300 mM sucrose, 1 mM HEPES, pH 7.5. In order to induce DOPA transport, the external pH of the vesicles was reduced by diluting an equal volume of vesicles with a buffer containing 100 mM sucrose, 20 mM citrate, pH 4.0. To stop PA transport, 100 p.L of the vesicles were added to 500 tL of ice cold 100 mM citrate, 100 mM ammonium acetate, pH 6.0 and the vesicles were stored on ice. Asymmetry was assessed by adding 2 mL of 3 1 iM TNS in 5 mM HEPES, 5 mM acetate, pH 7.0 and reading the fluorescence using excitation and emission wavelengths of 321 and 445 nm respectively. This value was normalized with respect to vesicles containing no PA (100% fluorescence). The external concentration of PA over time was calculated by comparing the fluorescence of the samples to a standard curve prepared using vesicles containing 0 to 10 mol% PA. For the transport of DOPA to the outer monolayer, LUVs were prepared in 300 mM citrate, pH 4.0 (with arginine) and passed down Sephadex G-25 columns equlibrated in 300 mM sucrose, 1 mM citrate,. pH 4.0. Transport of DOPA to the outer monolayer was initiated by the addition of an equal volume of 100 mM sucrose, 20 mM HEPES, pH 7.5. The ability of Ca 2 to induce vesicle fusion was monitored using the RET assay.  100  P NMR Studies 4.2.5 31 Frozen and thawed multilamellar vesicles (FATMLVs) were prepared in a 100 mM sucrose, 50 mM MES pH 5.5 buffer employing five freeze-thaw cycles. The proton decoupled 81.0 MHz 31 P NMR spectra were then recorded employing a Bruker MSL 200 spectrometer under the following conditions. The free induction decay (FID) was collected using a pulse width of 3.7 [Is (90°) with an interpulse delay of 1.0 is. After 1000 scans, the FID was Fourier transformed employing a line broadening function of 25 2 was added to the liposomes to bring the Ca Hz. Sufficient 1 M CaC1 2 concentration to 100 mM, a value in large excess of the DOPA concentration (100 mM total lipid, 10 mM DOPA). The MLVs were freeze-thawed two additional times to equilibrate the Ca 2 and spectra were then re-recorded.  4.2.6 Freeze Fracture Electron Microscopy Vesicles used for freeze fracture studies were prepared as for fusion assays. Vesicles (100 mM total lipid) containing either no DOPA or 10 mol% DOPA were prepared in 300 mM HEPES and passed down Sephadex G-25 columns equlibrated in 300 mM sucrose, 1 mM HEPES , pH 7.5. The vesicles were then mixed with an equal volume of 100 mM sucrose, 20 mM citrate pH 4.0. and incubated for times t  =  0 or t =4  h at pH 4.0 and then passed down Sephadex G-50 columns equilibrated in 100 mM sucrose, 50 mM MES, pH 5.5. To induce fusion, a small aliquot of 200 mM CaC1 2 was added to the vesicles to + concentration to 8 mM. Vesicles were incubated with 2 2 Ca for times t + bring the Ca  =  0, 5, 10, 15, and 30 sec before fusion was stopped. Fusion was stopped by adding 100 [LL of the vesicle solution to a tube containing 25 pL of 200 mM EGTA. Glycerol was then  101  added to the samples (-10 mM phospholipid) to a final concentration of 25%. Samples were then frozen in Freon cooled in liquid nitrogen and fractured. Platinum/carbon replicas of the samples were prepared as previously described (Fisher & Branton, 1974). The graphs presented in this Chapter were produced from single experiments but are representative of multiple trials.  4.3 RESULTS  4.3.1 Vesicle Composition Previous studies have shown that vesicles composed of phosphatidic acid or mixtures of PE and PA fuse readily in the presence of Ca 2 (Hope et al., 1983). The + 2 incorporation of PC into the membrane increases their stability and prevents Ca induced fusion at high PC contents (Duzgunes, 1985). In the studies performed here, various mixtures of DOPC,DOPE,PI and DOPA were tested for their ability to fuse in the +. P1 was incorporated to prevent aggregation of vesicles containing little 2 presence of Ca or no DOPA. It was also chosen because earlier studies have shown that it is not transported to the inner monolayer of vesicles under conditions where PA is transported (Eastman et al., 1991) and that it is a non-fusogenic lipid in isolation (Sundler and Papahadjopoulos, 1981). PC and PE are zwitterionic lipids that are also not transported in response to transmembrane pH gradients. It was found that LUVs formed from mixtures of these lipids in the ratios DOPC:DOPE:PI / 25:60:5 did not fuse appreciably , but did when DOPA was present. Further, this system did not 2 in the presence of Ca fuse in acidic environments which was important in order to allow the generation of lipid asymmetry by the transmembrane ApH.  102  4.3.2 Effect of (2+ on LUV Fusion The effect of Ca 2 on the fusion of symmetric LUVs containing 10 mol% PA (DOPC:DOPE:PI:DOPA /25:60:5:10) is shown in Figure 26A. The extent and initial + up to about 8 mM 2 rate of fusion are influenced greatly by the concentration of Ca Ca + 2 , where nearly 100% fusion occurred within 30 sec. Higher concentrations of Ca + 2 had little further effect on the extent of fusion and led to rapid fusion rates which were inconvenient to measure. The initial rate of vesicle fusion for vesicles containing 10 mol% DOPA as a function of Ca 2 concentration is shown in Figure 26B.  4.3.3 Effect of DOPA Content on LUV Fusion Vesicles were prepared with DOPA concentrations ranging from 0 to 10 mol% (in + 2 1% increments). These vesicles were tested for fusion in the presence of 8 mM Ca using the resonance energy transfer (RET) assay. Fluorescence traces of the LUVs containing various concentrations of DOPA are shown in Figure 27. The vesicles containing no DOPA fuse to only a small extent, whereas the addition of as little as 1% DOPA greatly enhances fusion, The extent of fusion continues to increase with increasing contents of PA and approaches 100% fusion at concentrations of DOPA greater than 7 mol% of the total phospholipid.  103  Figure 26. + Concentration on the Fusion of LUVs 2 Effect of Ca Containing 10 mol% DOPA + concentration on the fusion of vesicles 2 Fluorescence traces of the effect of Ca composed of DOPC:DOPE:PI:DOPA (25:60:5:10) (A). Vesicles were prepared as described in Section 4.2.2 and the ability of various concentrations of Ca 2 to induce fusion between the vesicles was monitored using the resonance energy + concentration 2 transfer (RET) assay. The initial rate of vesicle fusion at each Ca is shown in (B).  0  U  Time (sec)  10• C  B  •  / / 78  ] (mM) 2 fCa  104  Figure 27 Effect of DOPA Concentration on Vesicle Fusion Fluorescence traces indicating the effect of DOPA concentration on fusion of vesicles composed of DOPC:DOPE:PI. Vesicles, both labelled and unlabelled, were prepared containing various amounts of phosphatidic acid (0-10 mol%) as described in Section 4.2.2. The rate and extent of fusion of these vesicle populations in the presence of 8 mM Ca 2 was monitored employing the RET assay.  100 90 80 70 60 D  50 40 30 20 10 0 0  10  20  30  50  40  Time (sec)  105  60  4.3.4 Effect of DOPA Asymmetry on LUV Fusion Previous research has shown that the transmembrane pH gradients can induce the asymmetric transbilayer distribution of DOPA in LUVs (Hope et al., 1989; Eastman et al, 1991; Chapter 3). In particular, in LUVs with a basic interior, DOPA will move to the inner monolayer to the extent that greater than 95% of the total DOPA is present in this monolayer. This asymmetry can be readily assayed employing TNS (see Section 4.2.4) and the ability of a transmembrane ApH (pH 1  =  7.5 / pH 0  =  4.0) to induce the inward  movement of DOPA is shown in Figure 28A. It is important to note that the other lipids in this system (DOPE,DOPC and P1) are not transported to the inner monolayer of LUVs under these conditions (Eastman et al., 1991; Chapter 3). The transport of DOPA to the inner monolayer is very rapid with essentially all the external DOPA being transported to the inner monolayer after a 4 h incubation at room temperature (—21°C). 2 to induce fusion of LUVs containing 10 mol% DOPA was The ability of Ca monitored at various time intervals during the induction of DOPA asymmetry (Fig. 28B). Immediately after establishing the transmembrane ApH (t  =  0) the vesicles fused at the  same rate and to the same extent as vesicles containing 10 mol% DOPA symmetrically distributed between the two monolayers (ie. no ApH). However, the rate and extent of fusion decreased markedly as the external DOPA concentration decreased, indicating that the fusion of the LUVs was regulated by the amount of DOPA in the outer monolayer. These results are consistent with the time course of DOPA transport (Fig. 28A) and the concentration dependance of fusion on the DOPA content of the LUVs (Fig. 28).  106  Figure 28 Modulation of Membrane Fusion by Lipid Asymmetry Detection of the transbilayer transport of DOPA in response to a transmembrane ApH employing TNS (see methods) (A) and the effect of this transbilayer asymmetry on the fusion of DOPC:DOPE:PI:DOPA (25:60:5:10) vesicles. Vesicles were prepared with an internal pH of 7.5 as described in Section 4.2.4. A transmembrane ApH was established across the vesicles and the ability of the vesicles to fuse was monitored using the RET assay at various times after the induction of the ipH ranging from t = 0 to t = 4 h (B).  L >‘  U 0 C 0 1  0  A  9 8 7 6 5  C  4 0  a  3 2  0  1’  0  30  60  90  150  120  Time (miri)  C 0 C U  30  40  Time (sec)  107  180  210  240  4.3.5 Fusion of LUVs with DOPA Exclusively on the Outer Monolayer The ability of DOPA transbilayer asymmetry to inhibit fusion of DOPA containing LUVs is clearly demonstrated in Fig. 28. These results show that the characteristics of one monolayer, the monolayer exposed to the fusion inducing agent, can regulate the fusion characteristics of the bilayer as a whole. In order to study this ability further, LUVs were prepared such that all of the DOPA from the inner monolayer was transported to the external monolayer. The fusion characteristics of these vesicles were then compared to vesicles containing symmetrically distributed DOPA at the same external DOPA concentration. This was performed by preparing LUVs containing 2.5 mol% DOPA, inducing the asymmetric distribution of the DOPA in response to a transmembrane pH gradient (inside acidic) to achieve  5 mol% DOPA in the external  monolayer and comparing the rate and extent of fusion of these LUVs to vesicles containing 5 mol% DOPA symmetrically distributed between the two monolayers. The results of this experiment are shown in Figure 29. After all the DOPA was transported from the inner monolayer to the outer monolayer (final DOPA concentration on outer monolayer —4.9 mol%) the vesicles fused at the same rate and to the same extent as vesicles containing 5 mol% DOPA. In contrast, control vesicles containing 2.5 mol% DOPA symmetrically distributed between the leaflets, fused at a slower rate and to a lesser extent than the same vesicles with the DOPA asymmetrically located on the outer monolayer.  108  Figure 29 Effect on Fusion of DOPA Transport to the Outer Monolayer Fluorescence traces showing the extent of vesicle fusion for vesicles containing 2.5 mol% DOPA at times t = 0 (— _)and t = 4 h ( )after inducing transport of the DOPA to the outer monolayer (see Section 4.2.4) and of vesicles containing 5 mol% DOPA ( ).  80 70 50 C 0  50 40  U  30 20 10 0 0  10  20  30  40  Time (sec)  109  50  60  4.3.6 Polymorphic Phase Preferences Intermediates in bilayer to hexagonal 1 H 1 phase transitions, such as inverted micelles and interlamellar attachment sites are the most likely intermediates in the membrane fusion process (Cullis & Hope, 1978; Siegel et al., 1989; Ellens et a!., 1989). p NMR experiments were performed to determine the structural preferences of lipids 31 comprising the outer monolayer of fusing and non-fusing LUVs in the absence and presence of excess Ca . As shown in Figure 30, MLVs composed of DOPC:DOPE:PI 2 (25:60:5) exhibit bilayer spectra both in the presence and absence of Ca . However, 2 MLVs containing 10% DOPA exhibit bilayer structures in the absence of Ca 2 but + (see Figure 30). This 2 undergo a transition to the 1 H 1 phase in the presence of excess Ca is indicated by the transition from a 31 P NMR lineshape with a low field shoulder and high field peak to a spectra which exhibits reversed asymmetry and was a factor of two narrower (Cullis & de Kruijff, 1979).  4.3.7 Freeze-Fracture Studies of Vesicle Fusion Freeze-fracture studies were also carried out, in order to visualize the fusogenic behavior of vesicles exhibiting symmetric and asymmetric transbilayer lipid distributions As shown in Figure 31, LUVs which do not contain DOPA do not fuse appreciably +. This is in agreement with the RET assay 2 either in the absence or presence of Ca results (see Figure 28). On the other hand, LUVs containing 10 mol% DOPA symmetrically distributed on both monolayers fuse very quickly in the presence of 8 mM Ca + 2 . After 5 s many large structures are visible and almost all the vesicles appear to have undergone some degree of fusion. After 30 s all the vesicles have fused into huge  110  complexes. These results correspond very well with the fluorescence data which shows a very fast initial rate of fusion for the 10% DOPA vesicles and nearly 100% mixing of lipids. In contrast, LUVs containing 10 mol% DOPA, but where the DOPA has been sequestered to the inner monolayer in response to a pH gradient (interior basic) show very little fusion in the presence of 8 mM Ca 2 even after 30 s.  111  Figure 30 Polymorphic Phase Preferences of Non-Fusogenic Vesicles (no DOPA) and Fusogenic Vesicles (10 mol% DOPA) in the Absence and Presence of Excess Ca 2 31 NMR spectra of freeze-thawed MLVs (FATMLVs) composed of P DOPC:DOPE:PI (25:60:5) in the absence (A) and presence (B) of excess Ca 2 or DOPC:DOPE:PI:DOPA (25:60:5:10) in the absence (C) and presence (D) of +. Spectra were recorded for the MLVs in the absence of Ca 2 excess Ca + on a 2 Bruker MSL 200 NMR spectrometer. Following the addition of excess Ca 2 the MLVs were exposed to 2 additional freeze-thaw cycles to equilibrate the Ca 2 and the spectra were re-recorded.  A  B  C  D  PPM  -25  2l5oPI —25 PPM  112  Figure 31 Freeze-Fracture Electron Micrographs of LUVs in the Absence and Presence of Ca Freeze fracture electron microgaphs of various vesicle systems. DOPC:DOPE:PI (25:60:5) in the absence of Ca 2 (A) and 30 seconds after the addition of 8 mM 2 (C) and at 2 (B). DOPC:DOPE:PI:DOPA (25:60:5:10) in the absence of Ca Ca various times after the addition of 8 mM Ca 2 [5 sec (D) and 30 sec (E)]. DOPC:DOPE:PI:DOPA (25:60:5:10) vesicles which have been exposed to a transmembrane pH gradient for 4 hours in order to induce PA asymmetry. In the 2 [5 sec (0) and absence of (2+ (F) and at various times after the addition of Ca 30 sec (H)]. Fusion reactions were terminated during the time course by the addition of excess EGTA (see Section 4.2.6).  113  DISCUSSION The results of this investigation clearly demonstrates the potential regulatory role of lipid asymmetry in membrane fusion processes. This is shown by the strong correlation between the amount of DOPA in the outer monolayer of the LUVs employed here and fusion detected by the resonance energy transfer technique and freeze-fracture procedures. Three aspects of this work which are of interest concern the influence of the inner monolayer on the fusion process, the mechanism of fusion and the relation to regulation of fusion processes in vivo. Each of these points is discussed in turn. The results presented here show that as the amount of DOPA in the outer monolayer of DOPC:DOPE:PI:DOPA (25:60:5:10) LUVs decreases due to net transport of DOPA to the inner monolayer, the fusion rate decreases correspondingly (Fig. 28). This is perhaps not surprising given the strong influence of DOPA content on fusion (Fig. 27), however it clearly establishes the importance of the outer monolayer lipid composition as a determinant of fusion. A related question concerns the role of the inner monolayer, and whether the composition of the inner monolayer influences the propensity of the LUV as a whole to fuse. The results of Figure 29 where LUVs containing only 2.5% DOPA, but where all the DOPA is located in the outer monolayer, fuse at effectively the same rate as symmetric LUVs containing 5 % DOPA suggest that it does not. The lack of influence of the composition of the inner monolayer, not initially involved in membrane fusion, on the rate and extent of fusion argues for the validity of lipid asymmetry as a regulatory agent in fusion in vivo and has interesting implications for generating more controlled fusion processes. In particular, fusion between model membrane LUV systems in vitro is, in contrast to membrane fusion in vivo, generally a  114  leaky process. This is at least in part because the stimuli employed to initiate fusion in vitro commonly promote hexagonal (H ) phase in the lipid mixture (Burger & Verkleij, 11 1990; Hope & Cullis, 1981). End-stage formation of hexagonal structure is not compatible with maintenance of a permeability barrier. However, systems exhibiting asymmetric transbilayer distributions of lipid clearly have the potential to be self regulating and possibly to exhibit leak tight fusion. For example, in asymmetric LUV systems exhibiting a fusogenic outer monolayer but where the overall lipid mixture will not support fusion, fusion will presumably proceed until the fusogenic potential of the outer monolayer is dissipated by mixing with inner monolayer lipids. It will be of particular interest to determine the leakiness of fusion events in such systems. Regarding the mechanism of fusion, the results presented here are consistent with a fusion process which relies on the ability of component lipids to adopt transitory non bilayer structures. This is indicated by the fact that MLVs composed of DOPC:DOPE:PI (25:60:5) did not adopt H 11 phase in the presence of excess Ca 2 and LUVs with the same  lipid composition did not fuse appreciably in the presence of Ca , whereas the addition 2 of 10 % DOPA resulted in the ability of excess Ca 2 to induce the H 11 phase and stimulated fusion between corresponding LUVs. As has been well discussed elsewhere (Cullis & Hope, 1978; Siegel et al., 1989; Ellens et al., 1989) the actual intermediary in fusion is unlikely to be the hexagonal phase per Se. More logical structures include the inverted micelle and interlamellar attachment sites, which are likely intermediaries in the 11 transition process (see Siegel et al., 1989). bilayer-to-H The regulatory role that lipid asymmetry could play in vivo is clearly of major interest. Two types of regulation are clearly possible a passive form where the -  previously generated static asymmetry determines whether fusion can proceed and a  115  more active regulation where the local generation of asymmetry promotes or inhibits fusion. An example of passive regulation could be the stable transbilayer asymmetry observed for plasma membranes. The presence of phosphatidylcholine and or sphingomyelin in the outer monolayer inhibits fusion with extracellular entities except under exceptional conditions. Alternatively, the inner monolayer composed predominantly of phosphatidylethanolamine and phosphatidylserine would be expected to fuse more readily with internal organelles or secretory vesicles in response to local Ca concentrations. + stimuli such as increased 2 The possibility that lipid asymmetry may actively, locally regulate fusion is more speculative and clearly requires that lipids can be quickly mobilized from one side of the bilayer to the other as appropriate, or can be rapidly generated on demand. The relatively slow rates of transbilayer lipid movement in plasma membranes suggests that transbilayer mobilization would not be a feasible regulatory process in vivo. However, in membranes such as the endoplasmic reticulum membrane, where transbilayer flip-flop rates are rapid (Zilversmit & Hughes, 1977; Bishop & Bell, 1985) it is possible that fusion could be regulated by such a mechanism. Alternatively, of course, the local enzymatic generation of fusogenic lipids such as diglycerides or phosphatidic acid could also result in a similar end. In summary, the results of this investigation clearly demonstrate the potential regulatory role of lipid asymmetry in membrane fusion phenomena. Further, the properties of the one monolayer appears to determine the fusogenic tendencies of the bilayer as a whole.  116  CHAPTER 5 SUMMARY The asymmetric nature of biological membranes is now well established with respect to both the constituents of the membrane and the environments on either side of the membrane. Concerning lipids, the mechanism(s) by which asymmetry is generated and maintained in biological membranes are still a matter of contention. One possibility is that transmembrane ion gradients influence the transbilayer distribution of lipids in membranes. In this thesis, the ability of transmembrane pH gradients (ApH) to modulate lipid asymmetry in model membrane liposomal systems has been addressed. Furthermore, the functional significance of lipid asymmetry has been examined with respect to lipid exchange and membrane fusion. Previous studies have shown that transmembrane pH gradients can influence the transbilayer distribution of simple lipids with weak acid or weak base characteristics, such as stearylamine and fatty acids (Hope & Cullis, 1987). In Chapter 2, the ability of transmembrane pH gradients to modulate the exchange of these simple lipids between membranes is examined. It is shown that when these lipids are symmetrically distributed across the bilayer, rapid exchange can occur between donor and acceptor membranes. This is demonstrated for both aggregating and non-aggregating systems. However, if these lipids are sequestered to the inner monolayer in response to a transmembrane ApH, then little or no exchange occurs between the donor and acceptor membranes. Although fatty acids are usually bound to fatty acid binding proteins (FABPs) in the circulation and within cells, transmembrane pH gradients may modulate the exchange of free fatty acids across various biological membranes. For example, in the intestine, transmembrane pH gradients generated by the outward secretion of H ions by the epithelia, results in the  117  protonation of the fatty acids and the absorption of the fatty acids by the epithelial cells (Bugaut, 1986). Additionally, organelles with acidic interiors, such as lysosomes, would be expected to accumulate fatty acids in the outer monolayer where the acids could be removed by fatty acid binding proteins (FABP5) and processed by the cell. The ability of transmembrane pH gradients to modulate the flow of fatty acids between membranes and bovine serum albumin (BSA) could be significant with respect to the uptake of fatty acids by cells. For example, it has been speculated that one mechanism by which fatty acids cross cell membranes is by passive diffusion after dissociating from serum albumin. The results presented in this thesis indicate that transmembrane pH gradients (inside basic) would favor the transfer of fatty acids from the albumin to the acceptor membrane. Furthermore, this observation suggests that the concentration of free fatty acids is purposely maintained at very low levels within cells (through binding to FABPs) in order to prevent the accumulation of these acids in the membranes of organelles with basic interiors, such as mitochondria. In addition to the ability of transmembrane pH gradients to modulate the transbilayer distribution of simple lipids, more recent studies have shown that pH gradients can induce the transbilayer distribution of certain acidic phospholipids, such as PG and PA (Hope et al., 1989; Redelmeier et al., 1990). However, the lack of convenient assays to detect the asymmetry of various acidic phospholipids in model membranes precluded studies into the kinetics and mechanism of lipids other than PG. In Chapter 2, a fluorescence assay which employs TNS to detect the transbilayer distribution of acidic phospholipids is described. Using this assay the kinetics of DOPA transport in response to a transmembrane ApH is examined. DOPA is shown to be transported in the neutral form with an activation energy of 28 Kcal/mol.  118  The ability to generate lipid asymmetry in model membranes allowed studies of the effect of lipid asymmetry on membrane fusion, as discused in Chapter 4. It is shown that vesicles (DOPC:DOPE:PI:DOPA (25:60:5:10)), with a symmetric distribution of lipids fuse rapidly in the presence of Ca . However, if the DOPA is transported from 2 + 2 the outer monolayer to the inner membrane then the vesicles become resistant to Ca induced fusion. This is consistent with results indicating that similar vesicles prepared without DOPA fuse only to a small extent upon addition of Ca . Further studies 2 indicate that vesicles containing 2.5 mol% DOPA asymmetrically distributed on the outer monolayer, fuse to the same extent and at the same initial rate as vesicles containing 5 mol% DOPA symmetrically distributed between the membranes. These results indicate that the asymmetric distribution of lipids can have a profound regulatory effect on the fusion characteristics of lipid bilayers and that one leaflet of the bilayer can control the fusion characteristics of the entire bilayer. The experiments presented in this chapter also support the involvement of non-bilayer structures in fusion processes. There are a number of interesting problems which arise from these studies. The most obvious one concerns the imbalance between the inner and outer monolayers of the LUVs after the inward movement of lipid such as DOPA. In systems containing 10 mol% DOPA, for example, translocation of DOPA to the inner monolayer results in an inner monolayer which contains 20% more lipid than the outer monolayer. This is a topological impossibility. However, as indicated elsewhere (Hope et al., 1990) there is no evidence for a compensatory movement of other lipids (ie. PC) to the outer monolayer. It is possible that alternative structures bud off from the inner monolayer as illustrated in Fig. 32. Presently there is no data to support this hypothesis. Freeze fracture electron micrographs show no evidence of changes to the vesicle structure.  119  Figure 32 Possible Structure of Vesicles Exhibiting an Asymmetric Distribution of Acidic Phospholipids In order to accommodate the extra lipid on the inner monolayer, invaginations could perhaps be formed similar to the cristae observed in the mitichondrial inner membrane but only involving one monolayer.  fflffi  120  However, it is possible that the structure of the vesicles is affected by the cryo-protectant (glycerol) or the structure may undergo changes during the freezing process. Future experiments using cryo-transmission electron microscopy will hopefully be useful in detecting any morphological changes caused during the induction of asymmetry. Cryo transmission electron microscopy should not affect the morphology of the vesicles since no cryo-protectants are used and the freezing process is so rapid that there is no time for rearrangements of lipid structures (Burger & Verkleij, 1990). A second interesting problem concerns the ability of lipid asymmetry to regulate non-leaky membrane fusion. The results of Chapter 4 suggest that LUVs composed of an appropriate lipid composition, could be prepared such that the outer monolayer was fusogenic under conditions where the fusogenic lipid (eg. DOPA) was located exclusively on the outer monolayer, but which were non-fusogenic when the lipid was symmetrically distributed. Such LUVs would be expected to undergo only a limited number of fusion events as it would be expected that the lipids would redistribute (flipflop) during the fusion events producing vesicles with a non-fusogenic outer monolayer lipid composition. It would be of interest to see if such a regulated fusion event proceeded without leakage of vesicle contents. Such a system would more closely resemble the highly regulated, non-leaky fusion events observed in biological systems. LUVs exhibiting asymmetric distributions of lipids and pH sensitive fusion characteristics could also be useful in the delivery of drugs to specific cells. For example, a lipid which induced fusion in the neutral form, but did not affect the lipid bilayer when it was in a charged form, could be sequestered to the inner monolayer of vesicles in response to a transmembrane pH gradient. On the inner monolayer the lipid would be charged and non-fusogenic. However, as the pH gradient across the vesicles  121  dissipated, the fusogenic lipid would move to the outer monolayer. If the exterior pH was such that a proportion of the lipid was neutralized, then the LUV would become fusogenic. If these vesicle systems were targeted to a specific cell, it is plausible that fusion between the vesicle and cell could occur.  122  References  Arvinte, T., Hildenbrand, K., Wahi, P. & Nicolau, C. (1986) Proc. Nat!. Acad. Sci. USA 83, 962-966 Backer, J.M. & Dawidowicz, E.A. (1987) Nature 327, 341-343 Bangham, A.D., Standish, M.M & Watkins, J.D. (1965) J. Mol. Biol. 13, 238-244 Barenho!z, Y., Amselem, S. & Lichtenberg,D. (1979) FEBS Lett. 99, 210-214 Barsukov, L.I., Bergelson, L.D., Spiess, M., Hauser, H. & Semenza, G. (1986) Biochim. Biophys. Acta 862, 87-99 Barsukov, L.I., Kulikov, V.1., & Bergelson, L.D. (1977) Biokhimiya 42, 1539-1555 Barsukov, L.I., Victorov, A.V., Vasilenko, l.A., Evstigneeva, R.P. & Bergelson, L.D. (1980) 598, 153-168 Bass, N.M. (1988) International Rev, of Cytology 111, 143-184 Batenburg, A.M. & de Kruijff, B. (1988) Bioscience Rep. 8, 299-307 Bentz, J., Ellens, H., Lai, M.-Z. & Szoka, F.C. (1985) Proc. Nat!. Acad. Sci. USA 82, 5742-5745 Berden, J.A., Barker, R.W. & Radda, G.K. (1975) Biochim. Biophys. Acta 375, 186-208 Berridge, M.J. (1987) Ann. Rev. Biochem 56, 159-193 Beyers, E.M., Verhallen, P.F.J., Linskens, W.M.A., Comfurius, P. & Zwaal, R.F.A. (1987) Biochim. Biophys. Acta 903, 197-205 Bishop, R.W. & Be!!, R.M. (1985) Ce!! 42, 51-60 Bramhall, J. (1986) Biochemistry 25, 3479-3486 Breckenridge, L.J. & Walmers, W. (1987) Nature 328, 814-817 Bretscher, M.S. (1972) Science 181, 622-629 Bugaut, M. (1987) Comp. Biochem. Physio!. 86B, 439-472 Burger, K.N.J. & Verk!eij, A.J. (1990) Experientia 46, 631-644  123  Calvez, J.-Y., Zachowski, A., Herrmann, A., Morrot, G. & Devaux, P.F. (1988) Biochemistry 27, 5666-5670 Chandler, D.E. & Heuser, J.E. (1980) J. Cell Biol. 86, 666-674 Chandler, D.E. (1984) J. Cell Sci. 72, 23-36 Cheresh, D.A., Pytela, R., Pierschbacher, M.D., Klier, F.G., Ruoslahti, E. & Reisfeld. R.A., (1987) J. Cell Biol. 105, 1163-1173 Classen, J., Haest, C.W.M., Tournois, H. & Deuticke, B. (1987) Biochemistry 26, 66046612 Comfurius, P., Bevers, E.M., & Zwaal, R.F.A. (1989) Biochim. Biophys. Acta 983, 212216 Connor, J. & Schroit, A.J. (1987) Biochemistry 26, 5099-5 105 Connor, J. & Schroit, A.J. (1988) Biochemistry 27, 845-851 Connor, J. & Schroit, A.J. (1990) Biochemistry 29, 37-43 Cullis, P.R. & de Kruijff, B. (1979) Biochim. Biophys. Acta 559, 399-420 Cullis, P.R. & Hope, M.J. (1978) Nature 271, 672-674 Cullis, P.R. & Hope, M.J. (1985) in Biochemistry ofLipids and Membranes (Ed. Vance, D.E & Vance, J.E.) The Benjamin/Cummings Publishing Co., Inc., Menlo Park, California pp. 25-72 Cullis, P.R. & Hope, M.J. (1988) in Molecular Mechanisms ofMembrane Fusion (Ohki, S., Doyle, D., Flanagan, T., Hui, S.W. & Mayhew, E. Ed.) Plenum Publishing Co. pp. 37-51 Cullis, P.R., Hope, M.J. & Tilcock C.P.S. (1986) Chem. Phys. Lipids 40, 127-144 Cullis, P.R., Hope, M.J., de Kruijff, B., Verkleij, A.J. & Tilcock, C.P.S. (1985) in Phospholipids and Cellular Regulation (Kuo, J.F. Ed.) CRC Press, Boca Raton, Florida 1-59 Daleke, D.L. & Huestis, W.H. (1985) Biochemistry 24, 5406-5416 Damiani, E., Margreth, A., Farlan, S., Dahms,A., Arnn, J. & Sabbadini, R. (1987) J. Cell Biol. 104, 461-472 Danielli, J.F. & Davson, H. (1935) J. Cell. Comp. Physiol. 5, 495-508 de Kruijff, B. & Baken, P. (1978) Biochim. Biophys. Acta 507, 38-47  124  de Kruijff, B. & Wirtz, K.W.A. (1977) Biochim. Biophys. Acta 468, 318-326 de Kruijff, B., van den Besselaar, A.M.H.P., van den Bosch, H. & van Deenen, L.L.M. (1979) Biochim. Biophys. Acta 555, 181-192 de Kruijff, B., van Zoelen, E.J.J. & van Deenen, L.L.M. (1978) 509, 537-542 de Vrije, G.J., Batenburg, A.M., Killian, J.A. & de Kruijff, B. (1990) Mo!. Microbiology 4, 143-150 Deamer, D.W. (1982) in Intracellular pH: Its Measurement, Regulation and Utilization in Cellular Functions (Lisas, A. ed.) Elsevier, New York Deleers, M., Servais, J.-P. & Wulfert, E. (1986) Biochemical and Biophysica! Res. Comm. 137, 101-107 Deme!, R.A. & de Kruijff, B. (1976) Biochim. Biophys. Acta 457, 109-132 Devaux, P.F. (1991) Biochemistry 30, 1163-1173 Donohue-Ro!fe, A.M. & Schaechter, M. (1980) Proc. Nat!. Acad. Sci. USA 77, 18671871 Doody, M.C., Powna!l, H.J., Kao, Y.J. & Smith, L.C. (1980) Biochemistry 19, 108-116 Dragsten, P.R., Blumenthal, R. & Handler, J.S. (1981) Nature 294, 718-722 Duzgunes, N. (1985) Subce!!ular Biochem. 11, 195-286 Eastman, S.J., Hope, MJ. & Cu!!is, P.R. (1991) Biochemistry 30, 1740-1745 Eastman, S.J., Wilschut, J., Cul!is, P.R. & Hope, M.J. (1989) Biochim. Biophys. Acta 981, 178-184 Eisenberg, M., Gresa!fi, T., Riccio, T. & McLaughlin, S. (1979) Biochemistry 18, 52135223 Ellens, H., Bentz, J. & Szoka, F.C. (1985) Biochemistry 24, 3099-3106 Ellens, H., Siegel, D.P., Alford, D., Yeag!e, P.L., Boni, L., Lis, L.J., Quinn, P.J. & Bentz, J. (1989) Biochemistry 28, 3692-3703 Epand, R.M. (1990) Biochem. Cell Bio!. 68, 17-23 Etemadi, A.-H. (1980) Biochim. Biophys. Acta 604, 423-475 Feigner, P.L., Freire, E., Barenholz, Y. & Thompson, T.E. (1981) Biochemistry 20, 21682172  125  Ferrell, J.E., Lee, K.J. & Huestis, W.H. (1985) Biochemistry 24, 2857-2864 Fettiplace, R., Gordon, I.G.H., Hladky, S.B., Requens, 3., Zingshen, H.B. & Haydon, D.A. (1974) in Methods in Membrane Biology (Korn, E.D. Ed.) 4, 1-75 Finean, J.B., Coleman, R. & Michell, R.H. (1984) in Membranes and Their Cellular Functions, Blackwell Scientific Publications, Oxford Fisher, K. & Branton, D. (1974) Methods Enzymol. 32, 35-44 Fiske, C. & Subbarow, Y. (1925) J. Biol. Chem. 64, 375-400 Gabriel, N.E. & Roberts, M.F. (1987) Biochemistry 26, 2432-2440 Ganong, B.R. & Bell, R.M. (1984) Biochemistry 23, 4977-4983 Gennis, R.B. (1989) in Biomembranes Molecular Structure and Function, SpringerVerlag, New York Gerritsen, W.J., Henricks, P.A.J., de Kruijff, B. & van Deenen, L.L.M. (1980) Biochim. Biophys. Acta 600, 607-619 Goldfine, H., Johnston, N.C. & Bishop, D.G. (1982) Biochem. Biophys. res. Comm. 108, 1502-1507 Gorter, E. & Grendal, F. (1925) J. Exp. Med. 41, 439-443 Gutnecht J. & Walter, A. (1981a) Biochim. Biophys. Acta 641, 183-188 Gutnecht J. & Walter, A. (1981b) Biochim. Biophys. Acta 649, 149-154 Hamilton, J.A. & Cistola, D.P. (1986) Proc. Natl. Acad. Sci. USA 83, 82-86 Heuser, J.E., Reese, T.S., Dennis, M.J., Jan, Y., Jan, L. & Evans, L. (1979) J. Cell Biol. 81, 275-300 Hoekstra, D. & Kok, J.W. (1989) Bioscience Rep. 9, 273-305 Hoekstra, D. (1990) J. Bioenergetics and Biomembranes 22, 121-155 Hoekstra, D., de Boer, T., Klappe, K. & Wilschut. J. (1984) Biochemistry 23, 5675-568 1 Homan, R. & Pownall, H.J. (1988) Biochim. Biophys. Acta 938, 155-166 Hope, M.J. & Cullis, P.R. (1979) FEBS Lett. 107, 323-326 Hope, M.J. & Cullis, P.R. (1981) Biochim. Biophys. Acta 640, 82-90  126  Hope, M.J. & Cullis, P.R. (1981) Biochim. Biophys. Acta 640, 82-90 Hope, M.J. & Cullis, P.R. (1987) J. Biol. Chem. 262, 4360-4366 Hope, M.J., Bally, M.B., Mayer, L.D., Janoff, A.S. & Cullis ,P.R. (1986) Chem. Phys. Lipids 40, 89-107 Hope, M.J., Bally, M.B., Webb, G. & Cullis, P.R. (1985) Biochim. Biophys. Acta 812, 55-65 Hope, M.J., Redelmeier, T.E., Wong, K.F., Rodngueza, W. & Cullis, P.R. (1989) Biochemistry 28, 4181-4187 Houslay, M.D. & Stanley K.K. (1982) Dynamics ofBiological Membranes, John Wiley & Sons, Toronto Huang, C.H. (1969) Biochemistry 8, 344-352 Hutson, J.L. & Higgins, J.A. (1982) Biochim. Biophys. Acta 687, 247-256 Jackson B.J., Bohin, J.-P. & Kennedy, E.P. (1984) J. Bacteriology 160, 976-98 1 Kachar, B. & Reese, T.S. (1982) Nature 296, 464-467 Killian, J.A. & de Kruijff, B. (1986) Chem. Phys. Lipids 40, 259-284 Knoll, G., Burger, K.N.J., Bron, R. van Meer, G. & Verkleij, A.J. (1988) J. Cell Biol. 107, 2511-2521 Kramer, R.M., Hasselbach, H.-J. & Semenza, G. (1981) Biochim. Biophys. Acta 643, 233-242 Krebs, J.J.R., Hauser, H. & Carafoli, E. (1979) J. Biol. Chem. 254, 5308-5316 Lafleur, M., Cullis, P.R., Fine, B. & Bloom, M. (1990) Biochemistry 29, 8325-8332 Lentz, B.R., Alford, D.R. & Dombrose, F.A. (1980) Biochemistry 19, 2555-2559 Lentz, B.R., Madden, S. & Alford, D.R. (1982) Biochemistry 21, 6799-6807 Low, M.G. & Zilversmit, D.B. (1980) Biochim. Biophys. Acta 596, 223-234 Maezawa, S., Yoshimura, T., Hong, K., Duzgunes, N. & Papahadjopoulos, D. (1989) Biochemistry 28,1422-1428 Majerus, P.W., Wilson, D.B., Connolly, T.M., Bross, T.E. & Neufeld, E.J. (1985) TIBS 10, 168-171  127  Marsh, D. (1990) in CRC Handbook ofLipid Bilayers CRC Press, Boca Raton, Florida Martinez-Bazinet, C., Audigier-Petit, C., Frot-Coutaz, J., Got, R., Nicolau, C. & Letoublon, R. (1988) Biochim. Biophys. Acta 943, 35-42 Mayer, L.D., Hope, M.J., Cullis, P.R. & Janoff, A.S. (1985) Biochim. Biophys. Acta 817, 193-196 Mayer, L.D., Hope, M.J. & Cullis, P.R. (1986) Biochim. Biophys. Acta 858, 161-168 McEvoy, L., Williamson, P. & Schlegel, R.A. (1986) Proc. Nati. Acad. Sci. USA 83, 3311-3315 McLaughlin, S. & Haray, H. (1976) Biochemistry 15, 1941-1948 McLaughlin, S. (1975) in Molecular Mechanisms ofAnaesthesia (Fink B. Ed.) Raven Press, New York pp 193-221 Meers, P., Bentz, J., Alford, D., Nir, S., Papahadjopoulos, D. Hong, K. (1988a) Biochemistry 27, 4430-4439 Meers, P., Hong, K. & Papahadjopoulos, D. (1988b) 27, 6784-6794 Middelkoop, E., Van der Hoek, E.E., Bevers, E.M., Comfurius, P., Slotboom, A.J., Op den Kamp, J.A.F., Zwaal, R.F.A. & Roelofsen, B. (1989) Biochim. Biophys. Acta 981, 151-160  Miller, D.K., Griffiths, E., Lenard, J. & Firestone, R.A. (1983) J. Cell Biol. 97, 18411851 Mimms, L.T., Zampighi, G., Nozaki, Y. Tanford, C. & Reynolds, J.A. (1981) Biochemistry 20, 833-840 Nash,T. (1953) Biochem. J. 55, 416-421 Nayar, R., Mayer, L.D., Hope, M.J. & Cullis, P.R. (1984) Biochim. Biophys. Acta 777, 343-346 Nayar, R., Schmid, S.L., Hope, M.J. & Cullis, P.R. (1982) Biochim. Biophys. Acta 688, 169-176 Nicholls, D.G. (1974) Eur. J. Biochem. 50, 305-315 Nicholls, D.G. (1982) in Bioenergetics:An Introduction to the Chemiosmotic Theory Academic Press, New York Nicholls, J.W. & Pagano, R.E. (1982) Biochemistry 21, 1720-1726  128  Noy, N., Donnelly, T.M. & Zakim, D. (1986) Biochemistry 25, 2013-2021 Op den Kamp, J.A.F. (1979) Ann. Rev. Biochem 48, 423-475  Pagano, R.E. (1990) Current Opinion in Cell Biol. 2, 652-663 Pagano, R.E., Martin, O.C., Schroit, A.J. & Struck, D.K. (1981) Biochemistry 20, 49204927 Papahadjopoulos, D., Nir, S. & Duzgunes, N. (1990) J. Bioenergetics Biomembranes 22, 157-179 Papahadjopoulos, D., Poste, G., Schaeffer, B.E. & Vail, W.J. (1974) Biochim. Biophys. Acta 352, 10-28 Papahadjopoulos, D., Vail, W.J., Jacobson, K. & Poste, 0. (1975) Biochim. Biophys. Acta 394, 483-49 1 Papahadjopoulos, D., Vail, W.J., Newton, C., Nir, S., Jacobson, K., Poste, 0. & Lazro, R. (1977) Biochim. Biophys. Acta 465, 579-598 Papahadjopoulos, D., Vail, W.J., Pangborn, W.A. & Poste, G. (1976) Biochim. Biophys. Acta 448, 265-283 Paulussen, R.J.A. & Veerkamp, J.H. (1990) Subcellular Biochemistry 16, 175-226 Post, J.A., Langer, G.A., Op den Kamp, J.A.F. & Verkleij, A.J. (1988) Biochim. Biophys. Acta 943, 256-266 Quinn, P.J. (1990) Biochemical Soc. Trans. 18, 133-136 Rawn, D.J. (1989) in Biochemistry, Neil Patterson Publishers, Burlington pp 209-456 Redelmeier, T.E., Hope, M.J., & Cullis, P.R. (1990) Biochemistry 29, 3046-3053 Rooney, E.K., East, J.M., McWhirter, A.C. & Lee, A.G. (1983) Biochim. Biophys. Acta 728, 159-170 Roseman, M., Litman, B.J. & Thompson (1975) Biochemistry 14, 4826-4830 Roseman, M.A. & Thompson, T.E. (1980) 19, 439-444 Rossignol, M., Thomas, P. & Grignon, C. (1982) Biochim. Biophys. Acta 684, 195-199 Rothman, J.E., Tsai, D.K., Dawidowicz, A. & Lenard, J. (1976) Biochemistry 15, 23612370 Rottenberg, H. & Grunwald,T. (1972) Eur. J. Biochem. 25, 71-74  129  Rottenberg, H. (1973) J. Membr. Biol 11, 117-137 Rottenberg, H. (1979) in Methods in Enzymology LV, 547-569 Sanchez-Yague, J. & Lianillo, M. (1986) Biochim. Biophys. Acta 856, 193-201 Santini, M.T., Indovina, P.L., Cantafora, A. & Blotta, I. (1990) Biochim. Biophys. Acta 1023, 298-304 Schafer, R., Hinnen, R. & Franklin, R.M. (19764) Eur. J. Biochem 50, 15-27 Schmidt, W., Patzak, A., Lingg, G., Winkler, H. & Plattner, H. (1983) Eur. J. Biochem. 32, 31-37 Schroeder, F. (1985) Subcellular Biochem. 11, 51-101 Schroit, A.J., Madsen, J. & Ruoho, A.E. (1987) Biochemistry 26, 1812-1819 Schroit, A.J., Madsen, J. & Tanaka, Y. (1985) J. Biol. Chem. 260, 5131-5138 Searle, G.F. & Barber, J. (1979) Biochim. Biophys. Acta 545, 508-518 Seelig, J. (1977) Biochim. Biophys. Acta 10, 353-418 Seigneuret, M. & Devaux, P.F. (1984) Proc. Natl. Acad. Sci. USA 81, 3751-3755 Serham, C., Anderson, P. Goodman, E., Dunham, P. & Weissman, G. (1981a) J. Biol. Chem. 256, 2736-2741 Serham, C., Fridovich, J., Goetzl, E.J., Dunham, P.B. & Weissman, G. (1981b) J. Biol. Chem. 257, 4746-4752 Sessions, A. & Horowitz, A.F. (1981) FEBS Lett. 134, 75-78 Sessions, A. & Horowitz, A.F. (1983) Biochim. Biophys. Acta 728, 103-111 Shulka, S.D. & Hanahan, D.J. (1982) Arch Biochem. Biophys. 241, 335-341 Siegel, D.P. (1986) Biophys. J. 49, 1171-1183 Siegel, D.P. Banschbach, J., Alford, D., Ellens, H., Lis, L.J., Quinn, P.J., Yeagle, P.L. & Bentz, J. (1989) Biochemistry 28, 3703-3709 Singer, SJ. & Nicolson, G.L. (1972) Science 175, 720-731 Sleight, R.G. (1987) Ann. Rev. Physiol. 49, 193-208  130  Small, D.M. (1986) in The Physical Chemistry ofLipidsfromAlkanes to Phospholipids Plenum Press, New York Spencer, F. & Mukherjea, M. (1990) Subcellu!ar Biochem. 16, 1-19 Spener, F., Borchers, T. & Mukherjea, M. (1989) FEBS Letters 244, 1-5 Spiegel, S. & Fishman P.H. (1987) Proc. Nati. Acad. Sci. USA 84, 141-145 Stremmel, W. (1988) Path. Res. Pract. 183, 524-528 Struck, D.K., Hoekstra, D. & Pagano, R.E. (1981) Biochemistry 20, 4093-4099 Suenga, M., Lee, S., Park, N. G., Aoyagi, H., Tetsuo, K., Umeda, A. & Amako, K. Biochim. Biophys. Acta 981, 143-150 Sundler, R. & Papahadjopoulos, D. (1981) Biochim. Biophys. Acta 649, 743-750 Sune, A., Bette-Bobillo, P., Bienvenue, A., Fellmann, P. & Devaux, P.F. (1987) Biochemistry 26, 2972-2978 Szoka, F. & Papahadjopoulos, D. (1978) Proc. Nat!. Acad. Sci. USA 79, 4194-4198 Szoka, F. & Papahadjopoulos, D. (1980) Ann. Rev. Bioeng. 9, 467-508 Tanaka, Y. & Schroit, AJ. (1983) J. Biol. Chem. 258, 11335-11343 Tilcock, C.P.S. (1986) Chem. Phys. Lipids 40, 109-125 Tocanne, J.-F. & Teissie, J. (1990) Biochim. Biophys. Acta 1031, 111-142 Tullius, E.K., Williamson, P. & Sch!egel, R.A. (1989) Bioscience Rep. 9, 623-633 Vail, W.J., & Stollery, J.G. (1979) Biochim. Biophys. Acta 551, 74-84 van Deenen, L.L.M. & de Gier, J. (1974) in The Red Blood Cell (Surgenor, D. Ed.) Academic Press, New York pp. 147-213 van den Besselaar, A.M.H.P., de Kruijft B., van den Bosch, H., van Deenen, L.L.M. (1979) Biochim. Biophys. Acta 555, 193-199 van Meer, G. (1989) Ann. Rev. Cell Biol. 5, 247-275 van Meer, G., Davoust, J. & Simons, K. (1985) Biochemistry 24, 3593-3602 Venien, C. & LeGrimellec, C. (1988) Biochim. Biophys. Acta 942, 159-168  131  Verhallen, P.F.J., Bevers, E.M., omfurius, P. & Zwaal, R.F.A. (1987) Biochim. Biophys. Acta 903, 206-217 Verkleij, A.J., Mombers, C., Gerritsen, WJ., Leunissen-Bijvelt, J. & Cullis, P.R. (1979) Biochim. Biophys. Acta 555, 358-362 Voelker, D.R. (1990) Experientia 46, 569-578 Walter, A., Steer, C.L. & Blumenthal, R. (1986) Biochim. Biophys. Acta 861, 319-330 White, J., Kielian, M. & Helenius, A. (1983) Quart. Rev. Biophys. 16, 151-195 Williamson, P., Algarin, L., Bateman, 3., Choe, H.-R. & Schiegel, R.A. (1985) J. Cellular Phys. 123, 209-214 Williamson, P., Antia, R. & Schiegel, R.A. (1987) FEBS Lett. 219, 316-320 Wilschut, J. & Hoekstra, D. (1986) Chem. Phys. Lipids 40, 145-166 Wilschut, J. Duzgunes, N., Fraley, R. & Papahadjopoulos, D. (1980) Biochemistry 19, 6011-6021 Wilschut, 3., Duzgunes, N. & Papahadjopoulos, D. (1981) Biochemistry 20, 3126-3133 Zachowski, A. & Devaux, P.F. (1990) Experientia 46, 643-656 Zachowski, A., Favre, E., Cribier, S., Herve, P. & Devaux, P.F. (1986) Biochemistry 25, 2585-2590 Zachowski, A., Feliman, P. & Devaux, P.F. (1985) Biochim. Biophys. Acta 815, 510-514 Zachowski, A., Henry, 3.-P. & Devaux, P.F. (1989) Nature 340, 75-76 Zaks, W.J. & Creutz, C.E. (1990) 3. Bioenergetics Biomembranes 22, 97-120 Zilersmit, D.B. & Hughes, M.E. (1977) Biochim. Biophys. Acta 469, 99-110 Zimmerberg, J. (1987) Bioscience Rep. 7, 251-268 Zwaal, R.F.A. (1978) Biochim. Biophys. Acta 515, 163-205  132  

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.831.1-0086810/manifest

Comment

Related Items