Liposomes, Lipid Bilayers and Model Membranes: From Basic Research to Application [1 ed.] 1466507098, 9781466507098

Citation preview

Liposomes, Lipid Bilayers and Model Membranes From Basic Research to Application

Edited by

Georg Pabst Norbert Kucerka Mu-Ping Nieh John Katsaras

Liposomes, Lipid Bilayers and Model Membranes From Basic Research to Application

Liposomes, Lipid Bilayers and Model Membranes From Basic Research to Application

Edited by

Georg Pabst Norbert Kucerka Mu-Ping Nieh John Katsaras

Boca Raton London New York

CRC Press is an imprint of the Taylor & Francis Group, an informa business

CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2014 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Version Date: 20140108 International Standard Book Number-13: 978-1-4665-0711-1 (eBook - PDF) This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com

Contents Preface...............................................................................................................................................ix Editors ...............................................................................................................................................xi Contributors ................................................................................................................................... xiii

PART 1 Basic Research Topics Chapter 1

Soft Matter Physics of Lipid Membrane–Based Assemblies .......................................3 Daniel Harries and Uri Raviv

Chapter 2

Nonlamellar Lipid Aggregates ................................................................................... 31 Charlotte E. Conn and John M. Seddon

Chapter 3

Extractant Molecules as Hosts in Surfactant Monolayers or Bilayers ....................... 49 Olivier Diat, Pierre Bauduin, Thomas Zemb, Jean-Francois Dufrêche, and Daniel Meyer

Chapter 4

Molecular Dynamics of Lipid Bilayers: Standards, Successes, and Works in Progress................................................................................................ 69 Edward Lyman and Sandeep Patel

Chapter 5

New Insights into the Peptide–Membrane Partitioning Equilibrium from In Silico Free Surface-to-Bilayer Peptide Insertion ..........................................99 Jakob P. Ulmschneider

Chapter 6

Basic Aspects and Applications of Lipids and Protein Dynamics ........................... 111 Maikel C. Rheinstädter

Chapter 7

Lipid Diversity and Its Implications for Membrane Organization ........................... 125 Jianjun Pan, Norbert Kučerka, Mu-Ping Nieh, Frederick A. Heberle, Paul Drazba, and John Katsaras

Chapter 8

Liposome-Based Models for Membrane Rafts: Methodology and Applications..... 143 Frederick A. Heberle, Robin S. Petruzielo, Shih Lin Goh, Tatyana M. Konyakhina, David G. Ackerman, Jonathan J. Amazon, and Gerald W. Feigenson

v

vi

Chapter 9

Contents

Nanoscale Membrane Mimetics for Biophysical Studies of Membrane Proteins.... 167 Catherine J. Baker, Ilia G. Denisov, and Stephen G. Sligar

Chapter 10 Microemulsions: Biomimetic Systems for Characterization of Biomembranes and Their Associated Biomolecules ......................................................................... 179 Douglas G. Hayes Chapter 11 Locations of Small Biomolecules in Model Membranes ......................................... 199 Drew Marquardt and Thad A. Harroun Chapter 12 Membrane Medicine ................................................................................................ 217 Georg Pabst and Karl Lohner Chapter 13 Structural Diversity of DNA–Phospholipid Aggregates .......................................... 247 Daniela Uhríková and Petra Pullmannová Chapter 14 An Update on Active Membranes ............................................................................ 271 David Lacoste and Patricia Bassereau

PART 2 Technology Chapter 15 Medical Applications of Lipid Nanoparticles .......................................................... 291 David B. Fenske and Pieter R. Cullis Chapter 16 Polymer-Modiied Liposomes: From Long-Circulating to Multifunctional ............ 317 Aleksandr Piroyan, Alexander Koshkaryev, Robert D. Riehle, and Vladimir P. Torchilin Chapter 17 Drug Formulations Based on Self-Assembled Liquid Crystalline Nanostructures ..................................................................................................... 341 Anan Yaghmur, Jesper Østergaard, Susan Weng Larsen, Henrik Jensen, Claus Larsen, and Michael Rappolt Chapter 18 Tethered Lipid Membranes ...................................................................................... 361 Wolfgang Knoll, Renate L. C. Naumann, and Christoph Nowak Chapter 19 Ion-Transporting Supported and Tethered Lipid Bilayers That Incorporate Biological Membrane Transport Proteins ................................................................ 383 Donald K. Martin, Bruce A. Cornell, Lavinia Liguori, Jean-Luc Lenormand, Jean-Pierre Alcaraz, Gwenaël Scolan, and Philippe Cinquin

vii

Contents

Chapter 20 Role of Liposomes in Textile Dyeing ....................................................................... 401 Meritxell Martí, Alfonso de la Maza, José L. Parra, and Luisa Coderch Chapter 21 Micro- and Nanoparticles for Controlling Microorganisms in Foods ..................... 415 Jochen Weiss, Qixin Zhong, Federico Harte, and P. Michael Davidson

Preface I wonder how much it would take to buy a soap bubble, if there were only one in the world. (Mark Twain, 1835–1910)

Soap bubbles are thin, spherical ilms of soapy water that only survive for a few seconds, depending on their water content and chemical composition. They are commonly used by children for play, but it is also fascinating how soap bubbles have been used to answer physical problems. In some ways, cell membranes are analogous to soap bubbles, including the importance of water in determining their stability. Cell membranes, however, are much more than soap bubbles. Biological membranes are functional, selectively permeable barriers which surround the various cell organelles (e.g., mitochondria, endoplasmic reticulum, Golgi apparatus, etc.), enabling them to maintain their characteristic differences from the cytosol. Their mimetics—for example, liposomes—are used in a number of scientiic and technological applications, some of which are covered in this book. The broad range of such applications can be attributed mostly to Janus-faced properties of amphiphilic molecules that make up these aggregates (lipids, surfactants, polymers), where one side of the molecule prefers to associate with water, while the other associates with oil. These faces can also be “tuned” in a manner whereby aggregates can assume morphologies, ranging from micelles, lamellar and nonlamellar phases, to microemulsions. Importantly, each of these self-assembled morphologies possesses its own unique physical properties, and offers the possibility for scientiic insight and/or technological application. This compilation describes state-of-the-art research and future directions in the dynamic and ever-changing ield of membranes, which over the decades has evolved from studies of physicochemical properties of amphiphiles to their application in industry and medicine. Most recently, sophisticated computing techniques have been used to predict the structural and physical properties of these self-organized assemblies to a degree that can rarely be achieved by experiment. Biomimetic membrane research is driven by diverse interests and needs. For example, in order to optimize given processes or to better understand the function of biological membranes (in order to improve or develop new drugs or sensors), different physical and computational experiments are needed. The science described in this book is presented by leading researchers in their ields, and should serve as a useful reference for both the novice and expert. Although the book is organized into two parts, namely, basic and applied research, this differentiation is not strict, and certain chapters could have easily been assigned to either part of the book. The part on basic research covers a range of topics, beginning with a chapter on soft matter physics of membranes by Harries and Raviv, which is then followed by the chapter by Conn and Seddon, in which they describe nonlamellar phases and their applications. Diat et al. then discuss research involving the extraction of molecules (mainly metal ions) by using the interfacial properties of amphiphilic molecules in an immiscible liquid–liquid (e.g., water–oil). The chapter by Lyman and Patel, as well as that by Ulmschneider, clearly shows the signiicant developments that have taken place over the last decade in molecular dynamics simulations, starting from lipid-only simulations, to atomistic simulations of proteins embedded in bilayers, to peptides partitioning in membranes. Subsequent chapters evolve from theory and computational simulations to some of the most upto-date experimental research. Membrane dynamics, as studied by inelastic neutron scattering, is covered in the chapter by Rheinstädter, while elastic scatterings (neutrons and x-rays) are the topic of discussion by Pan et al. Both works report on recent advances of experimental techniques that enable one to explore the role of lipid diversity in the organization of membranes and the formation of well-deined nanostructures (e.g., lamellae, vesicles, ribbons, and nanodiscs). Lipid diversity leading to lateral heterogeneity in lipid bilayers has received considerable attention, especially in ix

x

Preface

connection to lipid “raft” in living cells. It is widely accepted that rafts play a central role in cellular processes, for example, signal transduction. The aggregate morphology of lipid binary mixtures composed of different chain lengths exhibits a strong dependence on temperature, concentration, and charge. Heberle et al. review a number of current lipid-based models which allow for the study of membrane rafts at a length scale on the order of microns. Baker, Denisov and Sligar introduce the formation of apolipoprotein-associated nanodiscs, which are being used to extract membrane proteins in their native low-curvature environment for biochemical and biophysical studies. Using a different approach, Hayes describes the potential of microemulsions to advance our understanding of model biological systems, including the study of associated biomolecules. The chapters by Marquardt and Harroun, and Pabst and Lohner describe, respectively, experimental techniques for locating molecules within membranes, and how membrane active compounds couple with the membrane to impart function in systems currently of interest to the medical community. Medical needs are also central for the development of novel gene-transfection systems. Lipoplexes, that is, lipid/DNA aggregates, seem to be promising carriers in addressing the transfer of genetic materials. The interplay between electrostatic and elastic interactions, and the structural variety that these aggregates form are described in the chapter by Uhríková and Pullmannová. The irst part of the book concludes with the chapter by Lacoste and Bassereau on lipid vesicles decorated with active proteins. They demonstrate the complex modiications of membrane properties (e.g., elasticity) upon the introduction of proteins, and provide theoretical descriptions. The second part of the book focuses on the technological applications of amphiphiles. This section begins with the chapter by Fenske and Cullis on liposome-based nanoparticles for drug delivery, in which they describe vesicle formulations and loading characteristics with different drugs. They are followed by Piroyan et al., whose chapter shows how liposomes can be modiied with membraneassociated polymers to prolong their in vivo circulation lifetime, and how they can be functionalized for speciic medical purposes. These two chapters on drug delivery are followed by the chapter by Yaghmur et al., in which novel drug-delivery systems made up of cubosomes, hexosomes, and micellar cubosomes are discussed; these assemblies hold promise for increased drug loading. Knoll, Naumann and Nowak, and Martin et al. introduce technology platforms using tethered membranes and reconstituted proteins for bio-sensing applications. They are followed by chapters by Martí et al., describing the use of liposomes in textile dyeing, and by Weiss et al. who describe how lipidic nanoparticles are used by the food industry. The use of lipid-based particles for dyeing materials results in reduced cost and energy, while increasing dye quality in wool. In the case of the food industry, lipidic nanoparticles are used to improve processing times and extend the shelf life of commonly available food products. We are grateful to all the authors who contributed their time and energy to this book. Without them, this opus would not have been possible. The aim of each chapter is to help the readers understand the different membrane platforms used by basic research laboratories and industry. In the end, we hope that the research presented in this book will inspire many researchers to create better bridges between model systems, biology, healthcare, and industrial applications. Georg Pabst Graz, Austria Norbert Kučerka Chalk River, Ontario Mu-Ping Nieh Storrs, Connecticut John Katsaras Oak Ridge, Tennessee

Editors Georg Pabst earned his PhD in physics from Graz University of Technology (Austria) and completed his postdoctoral research at the National Research Council (Chalk River, Ontario, Canada). After returning to Austria, he was a senior research oficer at the Austrian Academy of Sciences and is presently an assistant professor at the University of Graz. His research is focused on the physics of biological membranes with the aim to understand the functional role of membrane lipids in cellular transport and signaling. Norbert Kučerka earned his PhD in biophysics from the Faculty of Mathematics, Physics and Informatics at Comenius University in Bratislava (Slovakia). He completed his postdoctoral research at Carnegie Mellon University (Pittsburgh, Pennsylvania), and as a National Science and Engineering Council fellow, at the Canadian Neutron Beam Centre (Chalk River, Ontario, Canada). In 2008, he joined the National Research Council of Canada, and is presently an associate research oficer. His research is focused on determining the structure of biological model membranes, and unraveling the structure–function relationships. Mu-Ping Nieh earned his PhD from the Department of Chemical Engineering at the University of Massachusetts (Amherst, USA). He completed his a postdoctoral research at the National Institute of Standards and Technology, Pennsylvania State University, and the National Research Council (Chalk River, Ontario, Canada). At the National Research Council he was promoted to assistant research oficer. He is presently an associate professor at the University of Connecticut (Storrs, USA). His research focuses on the mechanisms behind the spontaneous formation in soft materials, including lipids, polymers, and proteins. John Katsaras is a senior scientist and distinguished R&D staff at Oak Ridge National Laboratory (ORNL). He earned his PhD in biophysics from the University of Guelph (Ontario, Canada). Prior to joining ORNL, he was principal research oficer at the National Research Council of Canada. He is internationally recognized for scientiic contributions to the ield of membrane biophysics, to materials of biological and medical relevance, and the application of x-ray and neutron scattering techniques to biologically relevant systems.

xi

Contributors David G. Ackerman Department of Molecular Biology and Genetics Cornell University Ithaca, New York Jean-Pierre Alcaraz TIMC-IMAG/GMCAO Laboratory, CNRS Université Joseph Fourier-Grenoble Grenoble, France Jonathan J. Amazon Department of Molecular Biology and Genetics Cornell University Ithaca, New York Catherine J. Baker Department of Chemistry University of Illinois Urbana, Illinois Patricia Bassereau Institut Curie, Centre de Recherche, CNRS Université Pierre et Marie Curie Paris, France Pierre Bauduin Institut de Chimie Séparative de Marcoule (ICSM) Bagnols sur Céze, France Philippe Cinquin TIMC-IMAG/GMCAO Laboratory, CNRS Université Joseph Fourier-Grenoble Grenoble, France Luisa Coderch Department of Chemical and Surfactants Technology Institute of Advanced Chemistry of Catalonia (IQAC-CSIC) Barcelona, Spain

Charlotte E. Conn CSIRO Materials Science and Engineering Victoria, Australia Bruce A. Cornell Surgical Diagnostics Pty Ltd. St. Leonards, New South Wales, Australia Pieter R. Cullis Department of Biochemistry and Molecular Biology University of British Columbia Vancouver, British Columbia, Canada P. Michael Davidson Department of Food Science and Technology University of Tennessee Knoxville, Tennessee Ilia G. Denisov Department of Biochemistry University of Illinois Urbana, Illinois Olivier Diat Institut de Chimie Séparative de Marcoule (ICSM) Bagnols sur Céze, France Paul Drazba Biology and Soft Matter Division Oak Ridge National Laboratory Oak Ridge, Tennessee and Department of Physics University of Tennessee Knoxville, Tennessee Jean-Francois Dufrêche Institut de Chimie Séparative de Marcoule (ICSM) Bagnols sur Céze, France

xiii

xiv

Contributors

Gerald W. Feigenson Department of Molecular Biology and Genetics Cornell University Ithaca, New York

John Katsaras Biology and Soft Matter Division Oak Ridge National Laboratory and Joint Institute for Neutron Sciences Oak Ridge, Tennessee

David B. Fenske Department of Chemistry University of the Fraser Valley Abbotsford, British Columbia, Canada

and

Shih Lin Goh Department of Molecular Biology and Genetics Cornell University Ithaca, New York

and

Daniel Harries Institute of Chemistry and the Fritz Haber Research Center The Hebrew University Jerusalem, Israel Thad A. Harroun Department of Physics Brock University St. Catharines, Ontario, Canada Federico Harte Department of Food Science and Technology University of Tennessee Knoxville, Tennessee Douglas G. Hayes Department of Biosystems Engineering and Soil Science University of Tennessee Knoxville, Tennessee

Department of Physics University of Tennessee Knoxville, Tennessee Canadian Neutron Beam Centre Chalk River, Ontario, Canada Wolfgang Knoll AIT Austrian Institute of Technology GmbH Vienna, Austria and Centre for Biomimetic Sensor Science Singapore Tatyana M. Konyakhina Department of Molecular Biology and Genetics Cornell University Ithaca, New York Alexander Koshkaryev Department of Pharmaceutical Sciences Northeastern University Boston, Massachusetts Norbert Kučerka Canadian Neutron Beam Centre Chalk River, Ontario, Canada and

Frederick A. Heberle Biology and Soft Matter Division Oak Ridge National Laboratory Oak Ridge, Tennessee

Department of Physical Chemistry of Drugs Comenius University Bratislava, Slovakia

Henrik Jensen Department of Pharmacy University of Copenhagen Copenhagen, Denmark

David Lacoste ESPCI, Laboratoire de Physico-Chimie Théorique, CNRS Paris, France

xv

Contributors

Claus Larsen Department of Pharmacy University of Copenhagen Copenhagen, Denmark Susan Weng Larsen Department of Pharmacy University of Copenhagen Copenhagen, Denmark Jean-Luc Lenormand TIMC-IMAG/TheReX Laboratory, CNRS Université Joseph Fourier-Grenoble Grenoble, France Lavinia Liguori Istituto di Bioisica Consiglio Nazionale delle Ricerche and Fondazione Bruno Kessler Trento, Italy Karl Lohner Institute of Molecular Biosciences University of Graz Graz, Austria

Alfonso de la Maza Department of Chemical and Surfactants Technology Institute of Advanced Chemistry of Catalonia (IQAC-CSIC) Barcelona, Spain Daniel Meyer Institut de Chimie Séparative de Marcoule (ICSM) Bagnols sur Céze, France Renate L. C. Naumann AIT Austrian Institute of Technology GmbH Vienna, Austria Mu-Ping Nieh Department of Chemical, Materials and Biomolecular Engineering University of Connecticut Storrs, Connecticut Christoph Nowak AIT Austrian Institute of Technology GmbH Vienna, Austria and

Edward Lyman Department of Physics and Astronomy and Department of Chemistry and Biochemistry University of Delaware Newark, Delaware

Centre for Biomimetic Sensor Science Singapore Jesper Østergaard Department of Pharmacy University of Copenhagen Copenhagen, Denmark

Drew Marquardt Department of Physics Brock University St. Catharines, Ontario, Canada

Georg Pabst Institute of Molecular Biosciences University of Graz Graz, Austria

Meritxell Martí Department of Chemical and Surfactants Technology Institute of Advanced Chemistry of Catalonia (IQAC-CSIC) Barcelona, Spain

Jianjun Pan Biology and Soft Matter Division Oak Ridge National Laboratory Oak Ridge, Tennessee

Donald K. Martin TIMC-IMAG/GMCAO Laboratory, CNRS Université Joseph Fourier-Grenoble Grenoble, France

José L. Parra Department of Chemical and Surfactants Technology Institute of Advanced Chemistry of Catalonia (IQAC-CSIC) Barcelona, Spain

xvi

Sandeep Patel Department of Chemistry and Biochemistry University of Delaware Newark, Delaware Robin S. Petruzielo Department of Molecular Biology and Genetics Cornell University Ithaca, New York Aleksandr Piroyan Division of Molecular Pharmaceutics University of North Carolina at Chapel Hill Chapel Hill, North Carolina Petra Pullmannová Department of Inorganic and Organic Chemistry Heyrovského, Hradec Králové, Czech Republic Michael Rappolt School of Food Science and Nutrition University of Leeds Leeds, United Kingdom Uri Raviv Institute of Chemistry The Hebrew University Jerusalem, Israel Maikel C. Rheinstädter Department of Physics and Astronomy McMaster University Hamilton, Ontario, Canada and Canadian Neutron Beam Centre Chalk River, Ontario, Canada Robert D. Riehle Department of Pharmaceutical Sciences Northeastern University Boston, Massachusetts Gwenaël Scolan TIMC-IMAG/GMCAO Laboratory, CNRS Université Joseph Fourier-Grenoble Grenoble, France

Contributors

John M. Seddon Department of Chemistry Imperial College London London, United Kingdom Stephen G. Sligar Departments of Biochemistry and Chemistry University of Illinois Urbana, Illinois Vladimir P. Torchilin Department of Pharmaceutical Sciences Northeastern University Boston, Massachusetts Daniela Uhríková Department of Physical Chemistry of Drugs Comenius University Odbojárov, Bratislava, Slovakia Jakob P. Ulmschneider Institute of Natural Sciences Shanghai Jiao Tong University Shanghai, China Jochen Weiss Department of Food Science and Biotechnology University of Hohenheim Stuttgart, Germany Anan Yaghmur Department of Pharmacy University of Copenhagen Copenhagen, Denmark Thomas Zemb Institut de Chimie Séparative de Marcoule (ICSM) Bagnols sur Céze, France Qixin Zhong Department of Food Science and Technology University of Tennessee Knoxville, Tennessee

Part 1 Basic Research Topics

1

Soft Matter Physics of Lipid Membrane–Based Assemblies Daniel Harries and Uri Raviv

CONTENTS 1.1 1.2

Introduction .............................................................................................................................. 3 Lipid Structure, Assembly, and Interactions ............................................................................ 4 1.2.1 Interactions between Membranes .................................................................................5 1.2.2 Measuring Forces between Membranes .......................................................................6 1.2.3 Membrane Elasticity ..................................................................................................... 8 1.2.4 Effect of Temperature ................................................................................................... 8 1.3 Examples of Lipid Membranes and Their Interactions ............................................................9 1.3.1 Entropic Attraction Condenses Like-Charged Membranes ......................................... 9 1.3.2 Neutral and Charged Membranes in Brine................................................................. 13 1.3.2.1 Ion Association with Membranes Dictates Forces between Membranes and Lateral Order ..................................................................... 13 1.3.2.2 Structure of Ions and Zwitterionic Lipids Regulates the Charge of Dipolar Membranes ................................................................................. 14 1.4 Lipids and Guests ................................................................................................................... 17 1.4.1 Sterol Ordering in Lipid Membranes: Possible Implications to Membrane Protein Structure ......................................................................................................... 17 1.4.2 Domain Formation: Regulating the Size and Stabilization of Lipid Raft-Like Domains and Using Calcium Ions as Their Probe ..................................................... 18 1.4.3 Lipid Composition as Modiier of Protein Stable State in the Membrane ................. 22 1.4.4 Proteins as Modiiers of Lipid Structure: A Possible Biological Role........................ 23 1.5 Concluding Remarks ..............................................................................................................24 References ........................................................................................................................................24

1.1

INTRODUCTION

Mention lipids to a biologist and you likely conjure images of biomembranes that form the boundaries between living cells and their exterior and between organelles within the cell. These membranes are complex and heterogeneous complex luids, typically composed of phospholipid bilayer aggregates that contain in addition proteins and other guest molecules. As borders of living cells, these bilayers must be sturdy and selective on the one hand, yet lexible enough to adapt to the surrounding environmental conditions on the other. This dual nature is achieved, in part, by the amphipathic structure of lipids composed of hydrophobic tails and hydrophilic headgroups. The luid character of bilayers allows them to respond to the presence of interacting macromolecules through membrane deformations or variations in local lipid composition. In turn, these macromolecules may respond by changing their stable conformation, or by altered associations and organization. The cross talk between lipids and guests such as proteins is increasingly recognized as an important part of the physiological role of membranes in the cell. 3

4

Liposomes, Lipid Bilayers and Model Membranes

To a physicist, lipids are a compelling example of materials that can spontaneously self-assemble in aqueous solutions into regular aggregates, such as spherical micelles or lipid membranes. These aggregates can be described through their material properties and their interactions with other aggregates. An intricate balance of forces determines these material properties based not only on the molecular identity of the self-assembling lipids but also on their bathing solution. For example, the effect of ions on the structure and function of cells and cell membranes can be substantial: indeed, it is well appreciated that multivalent cations are critical for maintaining cell membrane structural physiological function and integrity. Many membrane-associated processes including transport properties, cell communication, and membrane fusion are inluenced by the properties of ions across charged and dipolar membranes or interfaces. In addition, the interactions between lipid aggregates as well as their material properties are also modiied by the static and dynamic properties of ions at changed interfaces. If questioned on lipids, biotechnologists may attest to their potent ability to serve as vehicles for drug delivery. Lipid vesicles, or liposomes, can sometimes package drugs and release them upon encountering the target cells because of the related nature of synthetic and cell membranes. Finding formulations that will act eficiently and speciically is a signiicant current research undertaking. Being able to unravel the biological role of lipid membranes as mediators of physiological processes on the one hand, and understanding the properties that lead to ideal liposomal carriers on the other requires a proper understanding of the physical properties of lipids. Rather than providing a comprehensive review (excellent overviews of lipid membranes and their interactions are available in numerous references (e.g., Mouritsen 2005, Chernomordik and Kozlov 2008, Tristram-Nagle and Nagle 2004, Nagle and Tristram-Nagle 2000, Marsh 2007, Pabst et  al. 2010, Safran 2003, Israelachvili 2011, Evans and Wennerstrom 1999)), in this chapter, we describe examples of the current understanding of lipid membranes and their interactions, as derived from over half a century of extensive studies. We focus on how different ions and other guest molecules, such as cholesterol, peptides, or proteins, mediate the structure, dynamics, and interactions between and within membranes, and how membrane composition can in turn affect the stability of their guests. The central emerging theme is the interplay of these effects and how their combination leads to important conclusions not only on the biological role of lipids and membranes but also on the way lipid membrane properties are determined by the surrounding and embedded components. We begin by describing the most fundamental interactions that take place between charged or dipolar lipid membranes, and discuss how these are altered by the presence of guest molecules. We then demonstrate how this knowledge can be used to follow and inform on the formation of lateral structures within bilayers, such as raft-like microdomains in lipid mixtures, as well as to assess the mutual effects of guests and lipid on determining the emerging properties of the composite membranes. We inally comment on the ways that some of the remaining questions in this research area may be answered in future studies.

1.2

LIPID STRUCTURE, ASSEMBLY, AND INTERACTIONS

Serving as possibly the most important component of biological membranes, phospholipids typically consist of two fatty acids (saturated or unsaturated) connected to a polar headgroup. In the most abundant phospholipids, glycerol phospholipids, the two fatty acids are bound to carbon atoms of glycerol. The third carbon of glycerol, however, is bound to a phosphate group, which in turn is frequently attached to another small polar (charged or neutral) moiety, such as choline, serine, inositol, or ethanolamine. The lipids can have different chain lengths, degrees of saturation, head group sizes, and charge (Mouritsen 2005). The charge of the headgroups can be negative or net neutral (zwitterionic). Positively charged headgroups, however, do not occur naturally. Each of these structures speciications inluences the properties of the entire self-assembled bilayer. The amphiphilic character of lipids enables their self-assembly in aqueous solutions. Lipids often form a multilamellar structure (the L α or L β phases, where α and β represent the liquid and gel state

5

Soft Matter Physics of Lipid Membrane–Based Assemblies

of the bilayers, respectively), but may also form spherical or tubular vesicles (the L 4 phase), spherical or cylindrical micelles (the L1 phase), inverted micelles, or inverted hexagonal phases (the HII phase) as well as other structures. First proposed by Israelachvili, Mitchell, and Ninham (Israelachvili et al. 1977), the packing parameter, p = VC /( Ah ⋅ lC ), can often well predict the assembled structure based on the shape of a lipid molecule, as described by the volume of the lipid’s hydrophobic tails, VC, the area per headgroup in the aggregate, Ah, and the lipid chain length, lC. Small values of p < 1 correspond to aggregates with “positive” curvature, such as micelles, while p > 1 shows tendencies to “negatively” curved interfaces, such as the HII phases (Israelachvili 2011).

1.2.1 InteractIons between MeMbranes Both charged and neutral membranes can form lamellar phases, which can appear as stacks in onionlike multilamellar vesicles (MLVs) that exhibit regular interlamellar spacing, dw. Alternatively, membranes may appear as separated sheets or unilamellar vesicles at equilibrium. If a stack is formed, the equilibrium distance is determined by a balance of attractive and repulsive forces. Between neutral membranes in lamellar stacks, short-ranged repulsive interactions act between the lamellae, associated with hydration (or solvation) forces, leading to an added free energy f hyd, which is associated with structuring of water molecules at the interface differently than in the bulk, as well as a repulsive entropic interaction resulting from the thermal undulations of bilayers, fund, away from the lat state at nonzero temperature. Balancing these repulsive forces, bilayers experience mutual attractive van der Waals (vdW) interactions, fvdW. In the case of charged membranes, electrostatic interactions, felec, should also be added to the total free energy. These energetic terms are functions of the distance between membranes, as well as of the lipid and solution properties. It has been shown that the interaction free energy per unit area, f, as a function of the water spacing between bilayers, dw, at temperature, T, can often be well approximated by the expression (Andelman 1995, Petrache et al. 2006a) f (dW , T ) = fvdW + fhyd + fund + felec d

=−

2

− W H  1 2 1  k T 1 λh − + λ P e A e + + B  h h 12π  dW2  2π  κ fl (dW + δ )2 (dW + 2δ )2 

− dW λ fl

+ felec .

(1.1)

Here, H is the Hamaker coeficient, δ is the membrane thickness, Ph is the hydration pressure constant, λh is the hydration length (typically of a few Å), κ is the membrane bending rigidity, and Al and λl are parameters associated with membrane luctuations. If the membranes are charged, with a surface charge density σ, or become charged by adsorbing ions, the electrostatic interaction between the membranes, felec, can often be approximated by using the mean ield Poisson–Boltzmann (PB) theory (Andelman 1995). The resulting PB equation can be derived from a simple model that assumes that the free energy of ions in solution is determined by the balance of two forces. First, ion entropy tends to smear out the charge density, to allow the maximum number of states for the ions, and second, the ion electrostatic energy drives oppositely charged groups to attract. In the limit of low temperature, systems with an equal number of positive and negative charges would collapse. Thermal luctuations, however, guarantee that ions do not fully condense or “crystallize” on each other at any nonzero temperature. The result of minimization of this free energy is the nonlinear PB equation, which, for two parallel lat charged surfaces immersed in a 1:1 salt solution and separated at some distance, dw, along the z direction, takes the form 8πeρ∞ ∂ 2ψ  eψ  =− sinh  . 2 εW ∂z  kBT 

(1.2)

6

Liposomes, Lipid Bilayers and Model Membranes

Here, ψ is the electrostatic potential, ρ∞ is the salt ion concentration where ψ = 0, e is the charge of an electron, εw is the dielectric constant of water, and kB is the Boltzmann constant. For surfaces with equal charge density σ, the boundary condition to the PB equation is εW (∂ ψ/∂z)s = 4πσ, where the potential gradient is taken at the membrane–water interface and is pointing into the solution. This equation requires a solution that, for most conigurations, needs to be calculated numerically. For equally charged oppositely faced surfaces, however, useful approximate solutions exist for different regimes. These solutions depend on the bulk salt concentration, described by the Debye screening length λD, and on the surface charge density of the bilayer, expressed by the Gouy–Chapman (GC) length, b. Speciically (Andelman 1995, Steiner et al. 2012)  π  1 1   4ld  1 − πlσd + (πlσd )2  W W  w  8 − dW e λD  felec  πl λ D = k BT − 1 ln(d ) (1.3) W  πlb   λ D coth  dW  − 1   2 λ   πlb2  D 

λ D > dW > b

Gouy − Chapman (GC) region

dW > λ D > b

Intermediate region

b > dW, λ D2 > b ⋅ dW

Ideal gas(IG) region

λ D > dW , λ D2 < b ⋅ dW or λ D < b, λ D < dW

ɺɺ (DH) region’s Debye −Hucke

Here, l = e2/(εwkBT) is the Bjerrum length, describing the distance at which two elementary charges interact with an energy of kBT (~7 Å for water at room temperature), and b is the GC length, given by b = e/(2π•σ•l). Finally, λ D = (8π l ∑ i ρ∞,i zi2 )−1/ 2 is the Debye screening length, where ρ∞,i is the bulk salt concentration of the ith ion and zi is its ionic valence. This naïve PB theory deals with “simple” ions that show only electrostatic interactions, but not, for example, water structuring around them, and does not include nonelectrostatic interactions between ions and interfaces. These additional interactions can often become important for understanding ion–membrane interactions better. Recent extensions to the PB approach have been reviewed (Ben-Yaakov et al. 2011), including the possible association of the counterions with the membrane, and ion-speciic effects (Harries et al. 2006, Dvir et al. 2013). We return to discuss this point in the examples in Section 1.3.2.

1.2.2 MeasurIng Forces between MeMbranes The pressure between membranes is given in terms of the derivative of the free energy with respect to the separation between layers, Π(dW , T ) = −∂ f (dW , T ) /∂dW, so that at equilibrium, a membrane stack will show zero pressure. It is possible to act on the membrane stack and apply an additional external pressure by exerting an osmotic stress using polymers that are excluded from the MLVs (Leneveu et al. 1976). The corresponding distance between layers at some pressure can be precisely measured using small-angle x-ray scattering (SAXS). Together, the information on the osmotic pressure with SAXS can be used to derive pressure–distance curves (or “equations of state”) that can be modeled using expressions such as Equation 1.1. An example of such force curves is shown in Figure 1.1, where the underlying contributions to the total force are shown, too. (Note that the forces are always measured for compression of stacks, rather than for pulling them apart.) In turn, this allows deriving the relevant force constants associated with the different terms in the membrane stack’s free energy. Moreover, SAXS can inform on the undulations of membranes: larger luctuations result in broader Bragg diffractions. It is therefore possible to extract information on membrane luctuations and hence on their elastic properties using analysis of the Bragg diffraction peaks (Roux and Sainya 1988).

7

Soft Matter Physics of Lipid Membrane–Based Assemblies

Polymer solution

Scattered beam

Incident beam

Membrane allows only water to pass

Diffraction pattern

log Πosm (dyn/cm2)

9 8

Hydration

Fluctuation

%PEG –40 –20

7 6

–10 –5

5

–2

4 Πosm = 0

5.0

5.5

6.0

–0 6.5

Interlamellar spacing D (nm)

FIGURE 1.1 The osmotic stress technique relies on excluded polymers to set the osmotic pressure in solution, whereas SAXS is used to measure the spacing between bilayers in a multilamellar stack. Since large enough PEG polymers are completely excluded from the bilayer stack, no additional semipermeable membrane, which allows only water to pass through, is required. An example of pressure–distance curves for DMPC at 30°C (solid circles) and 35°C (solid diamonds) is shown together with theoretical decomposition into vdW (solid curve), hydration (dashed), and luctuation repulsion (dotted). Two regimes are distinguished: at low pressure, the repulsion is dominated by thermal luctuations, whereas at high pressure, the repulsion is dominated by hydration forces. The right vertical axis shows the concentrations of PEG 20,000 g/mol that generated the pressures indicated on the left, as measured by vapor pressure osmometry. (Experimental curves adapted from Petrache, H. I., Harries, D., and Parsegian, V. A. 2005. Macromolecular Symposia, 219, 39–50. With permission.)

Alternative approaches to the osmotic stress technique include the surface forces apparatus (SFA), whereby two, typically rigid, surfaces are brought into proximity and the force required to push them to that distance is accurately measured (Marra and Israelachvili 1985, Israelachvili 2011, Raviv and Klein 2002, Raviv et al. 2003). The information gained from this approach is restricted to membranes that are supported on such surfaces. The elastic constants can also be evaluated using vesicle pipetting and aspiration techniques, where pressure is exerted on a membrane by sucking on a pipette, and the subsequent changes in luctuations and membrane undulations are recorded (Evans and Parsegian 1986, Mouritsen 2005).

8

Liposomes, Lipid Bilayers and Model Membranes

1.2.3 MeMbrane elastIcIty The elastic energy of a bilayer can be described to a irst approximation by an expression that is harmonic in the curvature of a thin sheet. According to the Canham–Helfrich–Evans formalism (Canham 1970, Helfrich 1973, Evans 1974), the free energy of the membrane is 1 Felas = dA κ (C − 2C0 )2 + κK . 2

∫

(1.4)

In this expression, C = C1 + C2 and K = C1C2 are the mean and Gaussian curvatures of the membrane, respectively, and C1 = 1/R1 and C2 = 1/R2 are the principal curvatures of each monolayer, where R1 and R2 are the principal radii of the curvature. C0 = 1/R0 is the spontaneous curvature of the bilayer (or more typically of a monolayer), where R0 is its spontaneous radius; the value of C0 is determined by the packing parameter of the lipid (or lipids) that form the bilayer. When two symmetric lipid monolayers form a bilayer, the bilayer itself has no intrinsic tendency to curve away from the lat interface, and its spontaneous curvature is zero. Owing to the harmonic nature of Equation 1.4, any deviation from the lat coniguration would cause a higher energetic penalty than remaining lat. The membrane, therefore, assumes the lat coniguration and pays the smallest possible “frustration” cost of bending two monolayers that typically each have C0 with some nonzero value. Deviations of the mean curvature from C0 cost elastic energy, determined by the membrane bending modulus, κ. The measured values of κ for phospholipid membranes typically range between 5 and 40 kBT, and it is often found that κ is proportional to the square of the hydrocarbon thickness, lC2 , and inversely proportional to the area per headgroup to some power, p, ranging between 3 and 5 (Sackmann 1994, Lipowsky and Sackmann 1995, Szleifer et al. 1988, 1990). The saddle-splay modulus, κ , measures the energy cost of saddle-like deformations (where the two curvatures C1 and C2 have opposite sign) relative to isotropic, sphere-like deformations (where the two curvatures have the same sign). The elastic constants, κ and κ , are determined by the type of lipid through the head–head and tail–tail interactions. It is expected that these moduli will be strongly dependent on the surfactant chain length but only weakly dependent on the head–head interaction strength. The free energy cost of imposing membrane deformation over a length scale ξκ can be described by a scale-dependent renormalized rigidity κ(ξκ) (Morse 1994, Degennes and Taupin 1982). The membrane is softened by luctuations, and subsequently, κ(ξκ) decreases as κ (ξκ ) = κ − [3kBT / (4π )]ln(ξκ /a ), where a is a microscopic cutoff length (on the order of the molecular size). The length scale at which κ(ξκ) vanishes, ξκ ≅ a exp[ 4πκ / (3kBT )], is known as the persistence length and describes the correlation length beyond which the positional orientations of distant points on a membrane of simple topology become uncorrelated (Safran 1994). As we further discuss in Section 1.3.1, the tendency of layers to luctuate, as well as to remain stacked, sensitively depends on the elastic constants. In Equation 1.1, this is described by the added repulsion between layers expected for softer membranes. Moreover, the additional free energy terms to those presented in Equation 1.4 may be necessary to account for lipid tail compression or extension, as well as for lipid tilt. These extensions become even more important when guest molecules are presented in the bilayer, and are further discussed in Section 1.4. Membrane bending rigidity can be measured using optical microscopy image analysis of luctuating membranes (Sackmann 1994) or x-ray line-shape analysis of correlation peaks, associated with multilamellar structures (Roux and Sainya 1988, Ben-Nun et al. 2010).

1.2.4

eFFect oF teMperature

In general, temperature is a challenging parameter to study because temperature may change several material parameters at once. The Hamaker coeficient depends on temperature, so that

Soft Matter Physics of Lipid Membrane–Based Assemblies

9

2

H (T ) ≅ kBT [(ε w − ε m ) / (ε w + ε m )] , where εm is the membrane dielectric constant (Parsegian 2006). The temperature dependence of the water dielectric constant can be approximated by ε w (T2 ) = ε w T1 (T2 / T1 )−1.36 , where εw (298 K) = 78 (Israelachvili 2011). This suggests that l, b, and λD only weakly depend on temperature. However, felec increases with temperature, according to Equation 1.3. The reason for this increase can be qualitatively understood as follows. Next to a charged surface, the density proile of counterions is determined by the balance between their thermal energy (that increases with temperature) to drive ions away from the interface and to evenly spread in solution, and their electrostatic attraction to the oppositely charged surface (that does not vary with temperature in simple PB theory). When two charged surfaces are brought together, their counterions are trapped between the two surfaces as they must remain in the gap to neutralize them. The repulsion between the surfaces is proportional to the excess counterion concentration in the midplane between the surfaces, with respect to the ion concentration in the bulk. The counterion density proile extends farther away from the surface as temperature increases; hence, the repulsion between the surfaces increases with temperature. In addition, the osmotic pressure is also proportional to the temperature. The undulation term in Equation 1.1 represents the entropy of luctuating membranes and grows quadratically with temperature. Finally, the bending rigidity, κ, and the luctuation decay length, λl, were shown to be temperature dependent (Szekely et al. 2012).

1.3

EXAMPLES OF LIPID MEMBRANES AND THEIR INTERACTIONS

In this section, we discuss recent indings that provide a more comprehensive insight into the self-organization of charged and dipolar membranes in pure water or salt solutions under various conditions.

1.3.1

entropIc attractIon condenses lIke-charged MeMbranes

Charged membranes composed of phospholipids, such as dilauryl(C12:0)-, dioleoyl(C18:1)-, or distearoyl(C18:0)-sn-glycero-3-phospho-l-serine (PS), corresponding to dilauryl(C12:0)-sn-glycero-3phospho-l-serine (DLPS), dioleoyl(C18:1)-sn-glycero-3-phospho-l-serine (DOPS), or distearoyl(C18:0)sn-glycero-3-phospho-l-serine (DSPS), respectively, have a persistence length, ξκ , which is enormously large. Equations 1.1 and 1.3 predict that, in pure water, like-charged lipid bilayers made of those lipids are stiff and should repel one another (Israelachvili 2011, Safran 1994). When the lipid concentrations exceed the critical micelle concentration (ca. 10−10 M), at which bilayers begin to form, charged bilayers are expected to form a single, ideally swollen (space-illing) multilamellar phase with an ideal-swelling multilamellar repeat distance of Dideal = δ/ϕ, where δ is the membrane thickness and ϕ is the lipid volume fraction (Deme et al. 2002a,b, Dubois and Zemb 1991, 1998, 2000, Dubois et al. 1992, Cowley et al. 1978, Lipowsky and Leibler 1986). Recently, we showed (Steiner et  al. 2012) (Figure 1.2) that although charged membranes are suficiently rigid, in order to swell to Dideal, they deviate from the behavior of typical like-charged solids when diluted below a critical concentration, ϕc, of ca. 15 wt%. Unexpectedly, these membranes swell into multilamellar structures with spacing that is up to four times shorter than Dideal (i.e., if the bilayers were illing the entire available space). This process is reversible with respect to changing the lipid concentration (Figure 1.2). These indings hold for a wide range of conditions, including varying the membrane charge density, bending rigidity, monovalent salt concentration, and under conditions of typical biological, living systems (Steiner et al. 2012). The deviations from ideal-swelling behavior are accompanied by deviations from the theoretical (Equations 1.1 and 1.3) pressure–distance curve, measured by the osmotic stress method (Figure 1.3). The repeat lamellar spacing, D, deviates from the theoretical values when D(Π) reached values that correspond to lipid volume fractions (given by ϕ = δ/[D(Π)]) of ca. ϕc or less.

10

Liposomes, Lipid Bilayers and Model Membranes 1.0

D/Dideal

0.8 0.6 Diluting Concentrating Ideal swelling

0.4 0.2 0.0

0.1 0.2 Lipid volume fraction, φ

0.3

FIGURE 1.2 Deviation of charged membranes from ideal-swelling behavior. The multilamellar periodicity, D, measured by x-ray scattering, normalized by the ideal-swelling distance, Dideal = δ/ϕ, is plotted versus the volume fraction, ϕ, of DOPS in pure water. The horizontal dashed line corresponds to ideal-swelling behavior. Solid symbols correspond to samples that were prepared at high ϕ and diluted to lower ϕ. Open symbols correspond to the reverse process, where low ϕ samples were allowed to evaporate and reach higher ϕ values.

Owing to their counterion entropy, the repulsion between charged interfaces increases with temperature (see Section 1.2.3); hence, a smaller deviation from the ideal-swelling behavior is expected at higher temperatures. We found that like-charged membranes condense further with increasing temperature (Figure 1.4). This effect is also shown to be reversible. To understand the effect better, we carefully analyzed the effect of temperature on the structure of lipid bilayers. We found that bilayers thin with temperature (inset in Figure 1.4) and that the area per lipid increases with temperature (Szekely et al. 2011c). The volume of each lipid does not change much in the range of temperatures used. The increase of the area per headgroup, A, compensates for the stronger lateral entropic repulsion between the tails and the charged headgroups at higher temperatures. Although the bilayers thin with temperature (Szekely et  al. 2011c), they do not thin enough to explain the

Osmotic stress, Π (Pa)

106

φ = 0.052 φ = 0.027 φ = 0.014 Theory

105

104

103

102 5.5 10 55 100 Repeat distance, D (nm)

FIGURE 1.3 Osmotic stress, Π, as a function of the multilamellar repeat distance, D, for solutions of DOPS in pure water at different volume fractions, ϕ, as indicated in the igure. Osmotic stress was applied by solutions of PEG (Mw = 20,000 g/mol). The solid line corresponds to the calculated Π using the PB theory (Equations 1.1 and 1.2). The arrows point to the measured values of D when Π = 0.

11

Soft Matter Physics of Lipid Membrane–Based Assemblies

Membrane thickness (nm)

D/Dideal

1.0

0.9

4.1

4.0

3.9

3.8

0

20 40 60 80 Temperature (°C)

φ = 0.052 φ = 0.084 0.8 30

60 Temperature (°C)

FIGURE 1.4 The deviation from the temperature-dependent ideal-swelling distance, Dideal(T) = δ(T)/ϕ, as a function of T for DOPS, ϕ = 0.084 (solid symbols), and ϕ = 0.052 (open symbols). The measured values of the lamellar repeat distance, D(T), were normalized by Dideal(T), using the mean values of the membrane thickness δ(T), shown as the inset and the corresponding ϕ value. The broken horizontal line corresponds to the ideal-swelling behavior as a function of temperature.

observed decrease in the interlamellar spacing. In other words, even after we take into account the decrease of the ideal-swelling distance (Dideal(T) = δ(T)/ϕ) with increasing temperature, the deviation of the repeat lamellar spacing, D, from Dideal still grows with temperature. Why do self-assembled like-charged rigid lipid bilayers not swell ideally? When D < Dideal, a second phase must coexist with the multilamellar phase. Cryogenic transmission electron and light microscopy images showed that samples that deviated from ideal-swelling microphase separated into a multilamellar phase and a disordered lipid phase, containing vesicles and closed-tubular structures (Steiner et al. 2012). We therefore associate the deviations from ideal swelling with the formation of the disordered phase. This disordered phase originates from the self-assembled character of the charged lipid bilayers and their ability (unlike solid interfaces) to rearrange to reduce their total free energy. Traditionally (Lipowsky and Leibler 1986, Safran 1994), there were two limiting cases in the luid multilamellar, L α , phase of membranes, separated by water spacing, dw, with only steric repulsion: 1. If the membrane persistence length ξκ > dw, the sheets tend to be parallel. 2. If ξκ < dw, the sheets are wrinkled and can build up an isotropic disordered phase.

κ of typical phospholipid bilayers (of order 10 kBT) implies that ξκ is exponentially large and the membranes are stiff. It turns out, however, that the stability limits of the L α phase should also take into account the membrane topological Gaussian rigidity, κ. The Helfrich free energy limits (Templer et al. 1998, Marsh 2006, Porte et al. 1989) κ to the range: −2κ < κ < 0. The mean bending modulus, κ, must be positive for a bilayer to form because any deformation from a lat state increases the membrane free energy. The sign of κ is not determined a priori. For κ < 0, spherical deformation (vesicles or the L4 phase) will lower the free energy. When κ > 0, the only way for the surface to lower its free energy is by forming saddle-like deformed surfaces. The energy needed to curve a lat bilayer into a closed vesicle (if C0 = 0) is Eves = 4π (2κ + κ ) (Claessens et al. 2007a). Charged membranes tend to have κ < 0. Eves is highest for uncharged bilayers and decreases with increasing membrane charge density, so that charged lipids can easily form vesicles (Claessens et al. 2004, 2007a,b). These vesicles have been termed entropically stabilized vesicles

12

Liposomes, Lipid Bilayers and Model Membranes

because the driving force for their formation is the entropy gain from the melting of the multilamellar phase into vesicles. With analogy to the bending rigidity, the Gaussian rigidity, κ (ξ ), is also renormalized by the free energy cost for the formation of handles of size ξ, κ (ξ ) = κ − (kBT /π )ln(ξ /a ). A lat membrane is energetically stable only when κ + ≡ κ + κ /2 is positive. When κ+ < 0, the lat state is unstable and many spherical surfaces can form; when κ − < 0, saddle-splay surfaces form. By analogy to the deinition of the persistence length ξκ, one can deine ξ− ~ a exp[12πκ − / (5kBT )] and ξ+ ~ a exp[3πκ + / (kBT )], as the length scales at which the rigidities κ+ and κ− vanish. The multilamellar phase “melts” at a critical spacing dw ~ min(ξκ, ξ+, ξ−), corresponding to the irst value of dw, for which one of the rigidities, κ, κ+, or κ− vanishes (Morse 1994). ξκ is always intermediate between the two other topological persistence lengths. Thus, the instability associated with κ (ξκ) is always preempted by one of the other topological rigidities. New stability limits for the L α phase can now be formulated: when κ ≥ −(10 / 9)κ , the L α phase is expected to melt upon dilution to dw ~ ξ− ≪ ξκ ≪ ξ + into a saddle surface termed the L3 or the sponge phase. When κ ≤ −(10 / 9)κ , the L α phase is expected to melt upon dilution to dw ~ ξ + ≪ ξκ ≪ ξ− into the L 4 phase, consisting of primarily spherical surfaces, as observed in our experiments. When κ + (10 / 9)κ ≪ kBT so that ξ− ~ ξκ ~ξ +, the L α phase melts, when dw ~ ξκ, for which all the rigidities are of the order of thermal energy (Morse 1994, Marsh 2006). By taking into account the topological rigidity, transitions from the L α phase to the L 4 phase are obtained, when the lamellar spacing is much smaller than the membrane persistent length, ξκ. Charged membranes have a Gaussian modulus that is typically large and negative. The Gaussian modulus of our charged membranes promoted, with little energy cost (i.e., Eves ≈ 0), partial melting of the multilamellar phase and the formation of the disordered phase, at much shorter interlamellar separations, compared to the membrane persistence length, ξκ. The disordered phase forms owing to the gain in translational and conigurational entropy of the bilayers in that phase. The lamellar phase depletes the disordered phase that contains spherical or tubular vesicles, whose typical dimensions are much larger than dw (or D). Under these conditions, D ≪ Dideal, because the disordered phase applies an osmotic stress to the lamellar phase. This pressure is an effective entropic attractive contribution to the free energy and accounts for the difference between the theoretical and experimental results in Figure 1.3. At the same time, the repulsive interactions between the charged bilayers in the lamellar phase apply pressure to the entropically stabilized disordered phase. The equilibrium D is set at the spacing in which the pressures of the two phases balance one another and the water chemical potential is equal in both phases. At high volume fractions, Dideal is relatively small and the disordered phase does not form because its effective entropic attraction cannot overcome the internal (mainly electrostatic) pressure of the multilamellar phase at small gaps. As temperature increases, the effective entropic attractive contribution overcomes the increase in the (mainly electrostatic) net repulsion between the charged membranes and the disordered phase further condenses the membranes in the lamellar phase. In other words, the effective entropic attraction reduces the effective temperature of the lamellar phase. The added monovalent salt changes the screening length, λD . When NaCl is added (up to ca. 0.1 M), the behavior is essentially similar to that of pure water although the balance of pressures and water chemical potentials occurs at shorter D values owing to screening of the electrostatic repulsion (Steiner et al. 2012). Higher salt concentrations considerably screen the electrostatic repulsion and the vdW attraction dominates the free energy. In addition, simple PB theory is insuficient to account for the observed D values. Theories that account for the discrete character of the solvent and the inite size of the salt ions can better predict the salt dependence of D (Ben-Yaakov et al. 2009, 2011). At very high salt concentrations (>3 M), three-dimensional (3D) membrane crystals form owing to the high osmotic stress that the salt applies to the membranes (Dvir et al. 2013). We also showed that membranes composed of lipids with dipolar headgroups and saturated lipid tails can become charged if placed in salt solutions containing multivalent cations (e.g., CaCl2 or ZnCl2) above some critical concentration (ca. 1 mM in the case of CaCl2). The cations adsorbed

Soft Matter Physics of Lipid Membrane–Based Assemblies

13

onto the membranes and charge the initially dipolar lipid headgroups. The behavior of these membranes is very similar to the behavior of membranes containing charged lipids (Szekely et al. 2011b).

1.3.2

neutral and charged MeMbranes In brIne

Beyond direct coulombic electrostatic forces, speciic interactions can also occur at membrane interfaces with interaction energies that depend on hydration and the ions polarizability, as well as on the ion interaction with the membrane interface. Below, we review some examples of short-range surface-speciic interactions. These interactions can be described within a modiied PB formalism that adds to the free energy surface terms that are noncoulombic. These added interactions can lead to membrane charging or neutralization, to lower the total free energy (sometimes referred to as “charge regulation,”) and could even lead to phase separation in the membrane plane, in cases where line tension arises between ion-adsorbed and ion-desorbed patches. 1.3.2.1 Ion Association with Membranes Dictates Forces between Membranes and Lateral Order When ions can adsorb on charged or uncharged lipid membranes through forces that go beyond their coulomb interactions, the resulting charged surface cannot be described by a constant charge density. Instead, the amount of charge is regulated according to association and dissociation of surface ionic groups. This effect was realized in the work of Ninham and Parsegian (1971), who introduced the concept of a variable surface charge density that self-adjusts according to other system properties, in what has become known as the charge-regulated boundary condition. This model proved to be very useful and was later applied in many other contexts (Chan et al. 1975, 1976, Prieve and Ruckenstein 1976, von Grunberg 1999). For example, consider a charged membrane with some oppositely charged ions neutralizing it. If we further assume ideal mixing of charged and uncharged lipids surface, the charge density on a charged membrane can be expressed using the Ninham– Parsegian charge regulation condition (Ninham and Parsegian 1971): η (1 − η ) = exp( eψ s kBT ). Here, η is the surface fraction of charges (due to ion adsorption), and ψs is the surface potential. A modiication in the interactions between two bounding surfaces will then be affected by the changes in the ionic proiles and the electrostatic ield. Surface free energies that further include nonelectrostatic interactions between ions may lead to surface phase separations. In fact, such phase separation can couple to the spacing between the charged layers at equilibrium, as we discuss in the following. One extension of the Ninham–Parsegian model considers short-range nonelectrostatic interactions between associated and dissociated charged groups at the surface (Harries et al. 2006). This generalization was motivated by experiments by Zemb and coworkers (Dubois et al. 1998), where a lamellar–lamellar phase transition was observed in certain bilayer-forming lipids and surfactants, such as didodecyldimethylammonium (DDA) halides, for a series of three homologous halide counterions: Cl−, Br −, and I−, as shown in Figure 1.5. A discontinuous transition was found for the interlamellar spacing D as a function of applied osmotic stress, but only when counterions such as bromide were used. In contrast, for chloride, there is no phase transition and the isotherm, Π(dw), follows the PB prediction, while for iodide, the lamellar phase did not disperse at all in water. There is a probable link between the surface activity of ions and the lamellar–lamellar phase transition characterized by this observed discontinuous jump in the interbilayer separation. A possible model that explains these indings has used additional terms to the ideal mixing that includes an ion–lipid adsorption free energy as well as an unfavorable interaction energy between ion-adsorbed and ion-desorbed lipids. The nonelectrostatic ion binding to the surface is modeled by terms linear in the surface adsorbed fraction, whereas the interaction between the bound and dissociated groups is represented by a term quadratic on the surface charge density with constant κs.

14

Liposomes, Lipid Bilayers and Model Membranes 108

Π (Pa)

106

104

102

100

DDABr DDACl 101

102

103

D (Å)

FIGURE 1.5 Effect of ion association on the interlamellar spacing of lipid membrane stacks. Ion association with a membrane bilayer can effectively charge a membrane (if the lipids are net neutral) or neutralize it (if the membrane is charged). If, in addition, the membrane charged lipids and neutralized lipids acquire a free energy penalty for mixing, a lateral phase separation could show up as a jump in the interlamellar spacing in the stack. This could be the reason for the jump seen in Br ions but not in Cl ions in DDA+ membranes. (Reprinted with permission from Harries, D. et al. 2006. Ion induced lamellar–lamellar phase transition in charged surfactant systems. Journal of Chemical Physics, 124(22), 224702. Copyright 2006, American Institute of Physics.)

As previously discussed (Ben-Yaakov et al. 2011), minimizing the complete free energy leads to the regular PB equation inside the bulk solution and to a boundary condition in the form of the Langmuir–Frumkin–Davis adsorption isotherm (Davis 1958). The solution of the PB equation with this boundary condition yields the osmotic pressure as a function of separation dw. The analysis of the solution shows that if the parameter κs is large enough (typically ≈ 10 kBT), an in-plane phase transition can occur, with a coexistence region in the phase diagram between ion-adsorbed and ion-depleted phases. This transition, in turn, is coupled to the bulk phase transition causing a jump in the interlamellar spacing dw as the osmotic pressure is changed. Qualitatively, the behavior of this lamellar system, seen for different ions, can be understood in terms of the speciic counterion interaction with the charged surfactant bilayer, Figure 1.5. The Cl− counterion always dissociates from the bilayer-forming DDA+ surfactant, resulting in PB-like behavior, and a continuous Π(dw) isotherm. For Br−, the dissociation is partial, leading to a irst-order transition and a coexistence between the two lamellar phases of two different values of D spacing. Finally, for the I− counterion, the ion always stays associated with the DDA+ surfactant and there is not suficient repulsive interaction to stabilize the swelling of the stack under any osmotic pressure. Most interestingly, the in-plane transition is coupled to a transition in interlamellar spacings. 1.3.2.2

Structure of Ions and Zwitterionic Lipids Regulates the Charge of Dipolar Membranes Sodium (Na+) and calcium (Ca2+) ions are crucial for the regulation and function of many membrane-associated processes. While these ions are expected to bind or remain close to charged interfaces (Sundler and Papahadjopoulos 1981, Wilschut et al. 1981, Hauser and Shipley 1984, Philipson 1984), there is a great deal of curiosity about what these mobile ions do near or at zwitterionic interfaces, for example, lipids with phosphatidylcholine (PC) headgroups. The surface of membranes with zwitterionic headgroups is electrostatically neutral. In pure water, zwitterionic lipids form a multilamellar stack with an equilibrium interbilayer (water) spacing, dW, determined by the

Soft Matter Physics of Lipid Membrane–Based Assemblies

15

balance between the vdW attraction, the repulsive hydration forces (Leneveu et al. 1976, Lis et al. 1982, Rand and Parsegian 1989), and the repulsive undulation forces (Helfrich 1978, Roux and Sainya 1988, Petrache et al. 1998) between the bilayers (the irst three terms in Equation 1.1 and Figure 1.1). When multivalent cations (e.g., Ca2+) are added, they may bind to the PC phosphate moiety (Yabusaki and Wells 1975, Grasdalen et al. 1977, Binder et al. 2001, Szekely et al. 2011b), positively charge the membranes (Inoko et al. 1975, McLaughlin et al. 1978, Lau et al. 1980, Lis et al. 1980, 1981a,b, Klein et al. 1987, Tatulian 1987, Satoh 1995, Sabín et al. 2007), and increase their interbilayer spacing (Marra and Israelachvili 1985). As the concentration of CaCl2 increases, the bilayer spacing decreases due to screening of the electrostatic repulsion (Figure 1.6) (Inoko et al. 1975, Lis et al. 1981a,b, Tatulian et al. 1991, Szekely et al. 2011b). Nuclear magnetic resonance (NMR) (Westman and Göran Eriksson 1979, Akutsu and Seelig 1981), infrared (IR) linear dichroism (Binder et  al. 2001), and small-angle neutron scattering (SANS) (Uhríková et al. 2008) studies suggest that the binding of calcium ions induces an electric ield, which orients the zwitterionic headgroups so that the phosphate group will face the Ca2+ ions (Aruga et al. 1985, McManus et al. 2003), decreases the area per headgroup (Huster et al. 1999, Böckmann and Grubmüller 2004, Pabst et al. 2007, Uhríková et al. 2008), and increases membrane thickness (Pabst et al. 2007, Uhríková et al. 2008, Szekely et al. 2011c). When the ionic strength is low, the long-ranged attraction between the bound Ca2+ ions and the headgroup dipoles realigns the headgroups, compresses the hydrocarbon chains (Aruga et al. 1985, Kataoka et al. 1985), and stabilizes the lipid gel phase. The packing of the tails tightens with CaCl2 concentration (Shah and Schulman 1967, Graddick et al. 1979, Ganesan et al. 1982, Kataoka et al. 1985, Zidovetzki et al. 1989, Shibata 1990, Binder et al. 2001, Binder and Zschörnig 2002, Böckmann and Grubmüller 2004, Yeap et al. 2008). The PC headgroup has a dipole moment that ranges between 6 and 19 Debye (depending on its orientation (Seelig et al. 1977, Shepherd and Büldt 1978, Hauser et al. 1981)). Saturated lipids adsorb the ions more readily, owing to their smaller area per headgroup that allows them to gain ion–dipole interaction with little entropic cost (Szekely et al. 2011c), associated with the reduction in the area per lipid, induced by the adsorption of the cations (Szekely et al. 2011b). The binding of Ca2+ was suggested to induce headgroup conformational changes (Uhríková et al. 2008) that lead

40

dw (nm)

30

20

10

0

0.01 0.1

1 10 100 1000 CaCl2 (mM)

FIGURE 1.6 Interlamellar spacing, dW, as a function of CaCl2 concentration of 15 mg/mL lipid solutions containing saturated tails—DLPC (squares), DMPC (circles), and DPPC (triangles). When two phases coexisted, solid symbols correspond to the dominant phase.

16

Liposomes, Lipid Bilayers and Model Membranes

to dehydration of the lipid headgroups (Tatulian et al. 1991). The presence of Ca2+ also reduces the vdW Hamaker coeficient (Tatulian et al. 1991). The binding of divalent cations does not affect the short-range hydration repulsion, but makes the repulsive double-layer forces longer ranged (Marra and Israelachvili 1985). The binding of Ca2+ (or other divalent cations) increases the main gel-toliquid phase transition temperature, Tm (Simon et al. 1975, Chapman et al. 1977, Mishima et al. 1984, Aruga et al. 1985, Kataoka et al. 1985, Tatulian 1987). For 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), it increases from 23°C without Ca2+ up to ~29°C in the presence of 1 M CaCl2 (Chapman et al. 1977), and for 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC)—from 41°C without Ca2+ up to ~44°C in the presence of 300 mM CaCl2 (Aruga et al. 1985). This increase is attributed to the tighter lipid chain packing (Ganesan et al. 1982) and the more stabilized membrane gel phase (Shah and Schulman 1967, Chapman et al. 1977). Ca2+ also increases the pretransition temperature (by a few degrees centigrade), in which the hydrocarbon chains straighten (Chapman et al. 1977, Graddick et al. 1979) from a tilted β′ conformation into a β conformation (Tardieu et al. 1973, Rand et al. 1975, Janiak et al. 1976). The gel-to-liquid transition is accompanied by lateral expansion of the headgroup lattice, which loosens the ion–dipole interaction between Ca2+ ions and the surrounding PC headgroups, reducing the surface potential (measured by zeta potential) by cation desorption at the phase transition temperature (Ohki et al. 1981, Tatulian 1987). The PC headgroup binds Ca2+ relatively weakly compared with anionic lipids (McLaughlin et al. 1981, Macdonald and Seelig 1987a,b, Coorssen and Rand 1995, Akashi et al. 1998, Huster et al. 2000, Sinn et al. 2006). When the PC bilayers are pushed together by osmotic stress, the bound Ca2+ desorbs above some critical pressure (Lis et al. 1981a,b). Saturated-tail membranes (such as DLPC, DMPC, and DPPC) bind Ca2+ ions more strongly than unsaturated-tail membranes (Lau et al. 1980, Huster et al. 2000, Szekely et al. 2011c), and gel-phase membranes (frozen acyl chains) bind more divalent cations than liquid-disordered membranes (melted acyl chains) (Lau et al. 1980, Lis et al. 1980, 1981a, Ganesan et al. 1982, Marra and Israelachvili 1985, Klein et al. 1987, Tatulian 1987, Satoh 1995, Yamada et al. 2005, Yeap et al. 2008, Szekely et al. 2011b). The natural lipid EggPC is composed of a variety of zwitterionic lipids with different hydrocarbon chains with different degrees of saturation. The most abundant lipid in EggPC (about 70% of the total phospholipid composition (Tattrie et al. 1968)) is similar to the synthetic 1-palmitoyl-2-oleoyl-snglycero-3-phosphocholine (POPC), containing a saturated and an unsaturated tail. Therefore, EggPC interacts weakly with Ca2+ (Hauser et al. 1975, Szekely et al. 2011b). Ca2+ binds more strongly to DMPC and DPPC membranes than Mg2+ (Lau et al. 1980, Lis et al. 1980, 1981a,b, Mishima et al. 1984, Kataoka et al. 1985, Zidovetzki et al. 1989). The binding of the monovalent cation and anion is signiicantly weaker than that of the divalent cation (Szekely et al. 2011b). The binding of monovalent ions has a similar structural effect on the area per lipid headgroup and bilayer thickness as that of divalent ions do, as suggested by molecular dynamics (MD) simulations (Pandit et al. 2003) and SAXS measurements (Pabst et al. 2007). There is, however, no effect on the headgroup orientation and hydration upon binding (Pandit et al. 2003, Petrache et al. 2006a). The binding of Cl− is weaker than the binding of Na+ because of the different extent of dehydration for the two types of ions (K Na+ = 0.4 ± 0.2 M−1, KCl- = 0.16 M−1 for binding to DPPC membranes (Pandit et al. 2003)). The binding of monovalent ions also increases the main and pretransition temperatures of DPPC and DMPC (Chapman et al. 1977), but these changes are signiicantly smaller than that induced by divalent cations (Cunningham et al. 1986b). Large, polarizable anions interact with PC and confer a negative charge on the lipid, due to anion binding. This effect is apparent mainly for the interaction of lipids with Thiocyanate (SCN−) (Chapman et al. 1977, Cunningham et al. 1986a, Macdonald and Seelig 1988) or, according to the Hofmeister series (Rydall and Macdonald 1992, Sachs and Woolf 2003, Aroti et al. 2007, Leontidis et al. 2007), with ions with increasing polarizability. SAXS measurements showed that the multilamellar phase of DPPC swells due to SCN− or Br − binding (Cunningham et al. 1986a) (or other anions (Aroti et al. 2007)) and 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC) membranes swell with

Soft Matter Physics of Lipid Membrane–Based Assemblies

17

Br − or Cl− (Petrache et al. 2006a). The swelling is stronger for shorter tails and when the tails have double bonds (Petrache et al. 2006a). Br− binds more strongly to DLPC membranes than Cl− does, with an intrinsic binding constant of KBr- = 0.22 M−1, determined by itting osmotic stress data to the equation of state for interbilayer interactions (Petrache et al. 2006b). The measured surface potential, due to Br − binding to DPPC, is reduced as the bilayers are pushed together by external osmotic stress, indicating desorption of Br − ions from the surface (Figure 1.5) (Cunningham and Lis 1989). The absolute surface potential, however, is much smaller than the surface potential of DPPC with Ca2+ or other divalent cations (Szekely et al. 2011b). The addition of monovalent salts screens the vdW attraction between bilayers (Marra and Israelachvili 1985) and the Hamaker coeficient is reduced (Petrache et al. 2006a,b); hence, the swelling of PC membranes monotonically increases with salt concentration (Petrache et al. 2006b, Szekely et al. 2011b). The interactions between ions (or charged molecules) and dipolar membranes can be used for protein or drug encapsulation and delivery by adsorbing charged macromolecules (Huster and Arnold 1998), such as proteins or peptides, to polar interfaces, or by using Ca2+ ions to adsorb deoxyribonucleic acid (DNA) onto zwitterionic surfaces (McManus et al. 2003, Uhríková et al. 2005, Mengistu et al. 2009). DNA binds to DPPC membranes through “calcium ion bridges,” and is “sandwiched” between lipid bilayers: the Ca2+ ions bind and positively charge the lipid bilayer, the positive end of the PC headgroup interacts with the negatively charged phosphate groups of the DNA, and form a DNA–lipid complex (McManus et  al. 2003, Uhríková et  al. 2005). This complex can also form when DNA is added to the unsaturated lipid 1,2-dioleoyl-sn-glycero-3phosphocholine (DOPC) with Ca2+ or Mg2+ (Uhríková et al. 2005, Mengistu et al. 2009). Solidstate NMR measurements showed that DMPC, Ca 2+, and the negatively charged dextran sulfate (DS) form similar structures (Huster and Arnold 1998). In charged membranes, such as in lipids containing phosphatidylserine (PS) headgroups, the charges are covalently bound to the surface and are neutralized by counterions. In zwitterionic membranes, however, the surface charge is created by physically adsorbed ions. This noncovalent binding results in a more lexible regulation of the electrostatic interactions as ions may adsorb or desorb from the surface, for example, under osmotic stress (Figure 1.5) (Lis et  al. 1981a,b). For charged membranes, the effective surface charge density is regulated only by the counterion distribution next to the surfaces. Therefore, the interactions between dipolar membranes can be regulated by controlling the amounts and types of ions.

1.4 LIPIDS AND GUESTS Once guests are introduced to lipid membranes, the medium can become highly complex with multiple interactions determining the inal outcome of membrane structure and stability. Many components of biological and technologically utilized membranes can modulate these membranes’ properties. Examples include cholesterol, highly charged phosphatidylinositol-based lipids, and “stealth lipids” that carry a headgroup tethered to a polyethylene glycol (PEG) polymer. In addition, larger molecules may ind residence in lipid membranes, including peptides and proteins (Moshe et al. 2013). In this section, we demonstrate the intimate relationship that can develop between guest molecules and the host membranes. We relate these indings to the material properties of the lipid and guest, and the forces between them.

1.4.1

sterol orderIng In lIpId MeMbranes: possIble IMplIcatIons to MeMbrane proteIn structure

No matter how you deine cholesterol, a guest or host in lipid membranes, it is so essential to the proper function of mammalian cell membranes that even small errors in cholesterol synthesis can

18

Liposomes, Lipid Bilayers and Model Membranes

have serious implications. Changes in the physiological balance of cholesterol have been linked to various pathologies ranging from coronary heart disease to genetic metabolic disorders that involve the biochemical pathways of cholesterol synthesis. But while it has long been recognized as necessary to proper cellular membrane function, cholesterol has been getting bad rap for decades, mainly due to its involvement in cardiovascular diseases. A striking example for cholesterol’s vital role is the Smith–Lemli–Opitz syndrome (SLOS) caused by an inborn deicient activity of 3β-hydroxysterol D7-reductase (DHCR7), the enzyme responsible for the inal step in cholesterol synthesis from 7-dehydrocholesterol (7DHC) to cholesterol. This DHCR7 deiciency is responsible for the accumulation of 7DHC and reduced levels of cholesterol in patients with SLOS, leading to multiple congenital malformations (Gondre-Lewis et al. 2006, Tulenko et al. 2006, Porter and Herman 2011). Puzzlingly, these dire consequences are all due to just one double bond present in 7DHC but not in cholesterol (Petrache et al. 2005, Paila et al. 2008, Shrivastava et al. 2008). To try and resolve this puzzle, a combination of MD simulations with SAXS experiments has been used to compare mixed sterol with DMPC (a fully saturated PC lipid) membranes over a wide range of sterol compositions for the two types of sterols: cholesterol and its immediate metabolic precursor 7DHC. Whereas most membrane properties are only slightly affected by the replacement of one sterol by the other, the rigidity of membranes containing cholesterol is signiicantly larger than that of membranes that contain 7DHC over a large range of sterol concentrations (Rog et al. 2008, 2009, Petrache et al. 2005, Gondre-Lewis et al. 2006, Khelashvili et al. 2010, 2011). This change in the material properties seems to be strongly linked to the way that cholesterol is situated within the lipid membrane. Speciically, cholesterol seems to stand perpendicular to the membrane interfacial plane signiicantly more strongly than 7DHC, which in turn seems to be more “loppy” and luctuates more strongly. Could this be the basis for the mechanism leading to SLOS pathologies? In trying to answer this question, we have been hypothesizing that membranes that contain 7DHC affect proteins that reside in membranes differently from cholesterol-containing membranes (Khelashvili et al. 2010, 2011, Khelashvili and Harries 2013). This could lead to structural (conformational) differences in membrane proteins that can alter their biological function. First evidences for these effects come from computer simulations of membrane peptide dynorphin A (1–17) (or dynA) that is an endogenous ligand for the κ-opioid G protein-coupled receptor (GPCR) Figure 1.7. Cholesterol creates favorable conditions for dynA to be positioned near the critical residues of the target GPCR, thereby facilitating the entry of dynA into the transmembrane (TM) bundle of the κ-opioid receptor. Interestingly, while dynA is well positioned in cholesterol containing the membrane to perform its function, we have found that dynA is positioned 2–3 Å deeper in 7DHC membranes (see Figure 1.7). This serves as an indication that cholesterol in membranes can affect biological function through changes in the membrane material properties (Epand 2006), possibly explaining at least some of the phenotypes of cholesterol-related metabolic diseases. It is tempting to speculate that the aligning ield exerted by cholesterol (Khelashvili and Harries 2013, Khelashvili et al. 2013), compared with the much weaker ield experienced in the presence of 7DHC, can alter the environment for peptide folding so strongly that the peptide is unable to fold in the cholesteroldepleted membrane.

1.4.2

doMaIn ForMatIon: regulatIng the sIze and stabIlIzatIon oF lIpId raFt-lIke doMaIns and usIng calcIuM Ions as theIr probe

Lipid domains in biological membranes, often called rafts, have a typical diameter ranging between 10 and 30 nm (Pralle et al. 2000, de Almeida et al. 2005). Unless stabilized by proteins, the rafts are transient (Edidin 2001) and form owing to the lateral segregation of different lipids, driven by their intermolecular interactions. The composition (Lipowsky 2002) and thickness (Binder et  al. 2003, Tokumasu et al. 2003, Kuzmin et al. 2005, Akimov et al. 2007) of the domains differ from that of the membrane matrix. Model membrane studies provide insight into the possible behavior of lipid rafts in

Soft Matter Physics of Lipid Membrane–Based Assemblies

19

FIGURE 1.7 (See color insert.) Peptides fold in membranes that contain cholesterol, but fail to do so in membranes containing 7DHC. The left panel shows MD simulations of cholesterol (in blue) containing membranes with folded dynA peptide (in gold) inserted, whereas the right panel shows a peptide that fails to achieve the same fold when 7DHC replaces cholesterol. For clarity, lipid molecules are omitted in images. (Reprinted from Chemistry and Physics of Lipids, 169, Khelashvili, G. and Harries, D., How sterol tilt regulates properties and organization of lipid membranes and membrane insertions, 113–123, Copyright (2013), with permission from Elsevier.)

cells (London 2005). Binary mixtures containing an unsaturated lipid with a low melting temperature, Tm, and a saturated lipid with a high Tm, usually do not form nanoscale liquid domains (Veatch and Keller 2003). They may form, however, domains of a solid-like gel phase (i.e., solid ordered or Lβ phase), which coexist with domains of a luid phase (Lee 1977) (i.e., liquid disordered or Lα phase). The line tension at the interface between the lipids results from membrane curvature and chainpacking mismatch (Lentz et al. 1976, Elliott et al. 2005), which separates these domains. To form nanoscale rafts, the interfacial line tension should be close to zero (Kuzmin et al. 2005). Without stabilizing agents, nanoscopic and microscopic domains are unstable and eventually coalesce into phaseseparated macroscopic domains to reduce the interfacial line tension (Lentz et al. 1976, Elliott et al. 2005, Kuzmin et al. 2005, Tokumasu et al. 2003, Yethiraj and Weisshaar 2007). The saturated lipid DPPC and the unsaturated lipid DOPC phase separate (Veatch and Keller 2003, Szekely et al. 2011a). Theory predicts (Brewster et al. 2009) that a line-active component in cell membranes can stabilize domains of a particular size over biologically useful timescales. POPC is a hybrid lipid, containing saturated and unsaturated tails, and may act as a line-active component at saturated/ unsaturated interfaces. POPC can therefore preferentially go to the interface between the two coexisting bulk phases in a DOPC/DPPC mixture and can lower the interfacial line tension, without affecting the properties of the lipid molecules within the domains. This line-tension reduction is expected because the saturated–unsaturated tail orientation of POPC at the interface is energetically favorable, and hence suficient to overcome the entropy loss in bringing the hybrid lipid molecules to the interface from a state of mixing (Brewster et al. 2009). The predicted domain sizes under these conditions range from tens to hundreds of nanometers (Brewster and Safran 2010). As explained in Section 1.3.2.2, Ca2+ ions preferentially adsorb onto dipolar membranes composed of saturated zwitterionic (PC) lipids such as DPPC (Figure 1.6) (Szekely et al. 2011b). The adsorbed ions charge the initially neutral membranes and their lamellar repeat distance, D, increases

20

Liposomes, Lipid Bilayers and Model Membranes

from ca. 6 nm to a signiicantly larger distance, determined by the concentration of the added salt (Szekely et al. 2011b). The Ca2+ afinity of the unsaturated lipid DOPC or the hybrid lipid POPC is low and their D barely changes when CaCl2 is added. The unsaturated and hybrid lipids have larger areas per headgroup and their headgroups can freely rotate about their own location (Szekely et al. 2011b). The adsorption of Ca2+ ions limits this free rotation, costs entropy, and is therefore less energetically favorable than adsorption onto saturated lipids. Using the selectivity of Ca2+ binding, we investigated how the concentration of the hybrid lipid POPC controls the size of domains in binary mixtures of the segregating lipids DOPC and DPPC (Szekely et al. 2011a). We used the different binding afinities (Szekely et al. 2011b) of Ca2+ ions to the three lipids to indirectly probe the structure of the different lipidic populations in binary (DPPC/ DOPC or DPPC/POPC) and ternary lipid mixtures (DPPC/DOPC/POPC), as a function of their composition. In the binary mixtures, the lipid phase separated into a DPPC-rich phase (Figure 1.8) that adsorbed Ca2+ ions and its separation increased with decreasing volume fraction of DPPC, χDPPC. The second phase is DPPC-poor and essentially did not adsorb Ca2+ ions. In binary mixtures, the DPPC-rich phase showed an increase of the lamellar repeat distance D with decreasing χDPPC (Figure 1.8), whereas D of the ternary DPPC-rich phase (Figure 1.9) decreased with decreasing χDPPC. This observation can be explained by a ternary interaction between the three lipids, which cannot occur in any of the binary mixtures. In the ternary mixtures, the three lipids mix over a wide range of concentrations due to the effect of the hybrid lipid that reduces the line tension between the saturated (DPPC) and unsaturated (DOPC) lipids. As the hybrid and unsaturated lipids went into the ternary DPPC-rich phase and did not adsorb Ca2+ ions, the membrane charge density decreased; hence, D decreased with increasing χPOPC (Figure 1.9). Using this approach, our indings suggest that POPC stabilizes and regulates the size of lipid raft-like domains in vitro and can even dissolve the domains and ideally mix the three lipids when its concentration is suficiently high (Figure 1.10). This study (Szekely et al. 2011a) provides experimental support for the hypothesis that line-active cosurfactant molecules reduce the interfacial line tension between lipid domains, in vitro. It is possible that a similar mechanism operates in vivo.

45 40 35

D (nm)

30 25 20 15 10 5

0.0

0.2

0.4 0.6 χDOPC

0.8

1.0

FIGURE 1.8 Repeat lamellar spacing, D, as a function of DOPC mole fraction, χDOPC, of binary DOPC– DPPC mixtures, in 5 mM CaCl2 solutions. The inal total lipid concentration was 15 mg/mL. The lipid mixture microphase separates into two coexisting lamellar phases: a lamellar phase dominated by DOPC (squares) and a lamellar phase dominated by the DPPC–calcium complex (triangles).

21

Soft Matter Physics of Lipid Membrane–Based Assemblies 45 40 35

D (nm)

30 25 20 15 10 5

0.0

0.2

0.4 0.6 χDOPC

0.8

1.0

FIGURE 1.9 The lamellar repeat distance, D, as a function of POPC volume fraction, χPOPC, of ternary mixtures, in 5 mM CaCl2 solutions. The DOPC:DPPC molar ratio was ixed at 1:4 and the amount of POPC was adjusted as needed. The inal total lipid concentration, in all the mixtures, was 15 mg/mL.

FIGURE 1.10 (See color insert.) Illustration of the arrangement of the lipids in the binary and ternary mixtures. The unsaturated lipid, DOPC, is in blue, the saturated lipid, DPPC, is in red, and the hybrid lipid, POPC, is colored half red and half blue. In the binary mixtures, the lipids phase separates. POPC can sit at the interface between the DPPC-rich domain and the DOPC-rich domain. When POPC is added, the domain size decreases (middle cartoon). When the molar ratio between the saturated, unsaturated, and the hybrid lipid crosses 1:1:1 (left cartoon), the three lipids can completely mix. The relative size of the different lipids and the way they are incorporated in the membrane are based on our interpretation of our experimental indings. (Reprinted with permission from Szekely, O. et al. 2011a. Regulating the size and stabilization of lipid raft-like domains and using calcium ions as their probe. Langmuir, 27(24), 14767–14775. Copyright 2011, American Chemical Society.)

22

Liposomes, Lipid Bilayers and Model Membranes

This model can be improved by studying the effect of cholesterol that is an important component of cellular membranes.

1.4.3

lIpId coMposItIon as ModIFIer oF proteIn stable state In the MeMbrane

Lipid membranes generally carry stress. As discussed in Section 1.2.4, bringing together two monolayers to form a lat bilayer is a necessary compromise for both monolayers’ tendency to curve toward or away from each other, as shown in Figure 1.11. Interestingly, guest molecules can augment or alleviate this stress (Khelashvili et al. 2008, 2009, Sperotto et al. 2006, May 2000, May and BenShaul 1999, Nyholm et al. 2007, Lundbaek et al. 2010). If the guest (say protein) can change its state, the state that alleviates more stress will be selected over the other see Figure 1.11. And, in doing so, the lipid environment can shift equilibrium states of favorable or unfavorable guest states. A similar mechanism is realized in the idea of “hydrophobic mismatch,” whereby proteins that match in their tail’s tendencies to stretch or compress will best accommodate the proteins. Deviations from this state will encourage proteins to change their conformations or to associate so as to alleviate some of the hydrophobic mismatch by burying it in protein–protein interactions, Figure 1.11. These shifts in membrane stress can be described using the lateral stress proiles of lipids. It has been shown that the work required to go from one protein state to another seems to match the changes that can be exerted by the lipid stress (Marsh 2007, Bezrukov 2000, Lundbaek et al. 2010). Striking examples of these effects have been shown in single-molecule measurements of proteins that can insert into membranes. The immediate consequences of this lipid stress are witnessed as changes in conformational equilibrium between different functional forms of integral proteins and peptides, membrane-induced interactions between proteins, and partitioning of proteins between different membranes and between the bulk and the membrane. Moreover, by changing the energetics of the spontaneous formation of nonlamellar local structures, lipid packing stress inluences membrane stability and fusion (Chernomordik and Kozlov 2008). As an example of these types of measurements, we mention two model channels: alamethicin and gramicidin (Lundbaek et al. 2010, Bezrukov 2000). Two strategies were used to introduce elastic stress to these reconstituted channels in model membranes. First, alamethicin or gramicidin channels were reconstituted into bilayer lipid membranes of changing lipid composition to vary the elastic stress of lipid packing. Second, bilayers

c0 > 0

c0 = 0

c0 < 0

FIGURE 1.11 Lipid monolayers typically show nonlamellar tendencies even when they form bilay-

ers, leading to membrane “frustration.” This frustration can be alleviated or exasperated by the presence of membrane guests. If the guest better satisies some of the nonlamellar tendencies of each monolayer when it is in one of its molecular states (as in the example of dimerized versus monomeric gramicidin peptides shown in the bottom panel), the overall equilibrium between the different guest states will tend toward that preferred state. This commensurate state can be achieved through smaller chain compression (straight double arrow) or better matching lipid curvatures at the interface of the two oppositely faced monolayers (bent double arrow).

23

Soft Matter Physics of Lipid Membrane–Based Assemblies

were formed only from one lipid species, PS, and, while monitoring single gramicidin or alamethicin channels, the elastic stress was varied by changing the pH of the bathing solution. It was found that an increase in elastic stress in the hydrocarbon tail region decreased the gramicidin channel lifetime and increased the duration of the alamethicin single-channel “burst.” Thus, manipulations that suppressed gramicidin channels promoted alamethicin channels by favoring larger alamethicin aggregates. The mechanism of gramicidin channel suppression by negative curvature stress, as well as the effects of hydrophobic mismatch, is well understood and documented. However, there is no consensus on the effect of lipid on alamethicin, and several models have been proposed to explain this feature.

1.4.4

proteIns as ModIFIers oF lIpId structure: a possIble bIologIcal role

So far, we have discussed evidence that the lipid environment can signiicantly alter the structure and arrangement of membrane proteins (or peptides). Interestingly, the converse effect is also an important consequence of lipid–protein interactions. Proteins undergoing conformational changes and structural rearrangements such as dimerization and aggregation can alter the structure and dynamics of lipid membranes. A few examples for the effect of proteins on the membrane state have emerged from MD simulations (Shan et al. 2012, Mondal et al. 2011a,b). An illustrative example concerns the 5-HT2A receptor for the neurotransmitter serotonin (5-HT), which is a GPCR, and is targeted by an extensive and diverse collection of external stimuli, Figure 1.12. Interestingly, the serotonin 5-HT2A receptor can elicit similar stimuli in the cell upon binding of structurally diverse ligands, but, for the binding of quite similar ligands, it shows dramatically different responses. MD simulations of molecular models of the serotonin 5-HT2A receptor in complex with pharmacologically distinct ligands show the dynamic rearrangements of the receptor molecule to be different for these ligands, and the nature and extents of the rearrangements relect the known pharmacological properties of the ligands. The different rearrangements of the receptor molecule were shown to produce different alterations of the surrounding membrane, a remodeling of the environment that can have differential ligand-determined effects on receptor function and association in the cell’s membrane, Figure 1.12. A major component of lipid rearrangement is due to the hydrophobic mismatch discussed in Section 1.4.3. The differential reorganization of the receptor lipid environment is relected in two ways. First, the cholesterol in the membrane shows involvement in the activation of the 5-HT2A receptor. Second, different extents and patterns of membrane deformations are seen for different receptor states, Figure 1.12. These indings

4 1 2 5-HT 7 6

4 3 5

4 2 KET

1

1 2 LSD 7 6

3

7

5

6

3 5

(C2-C2 distance, Å) 36

38

40

42

44

46

48

50

FIGURE 1.12 Snapshots of three different states of the 5-HT2A receptor bound to three different ligands. The shaded ield corresponds to membrane thickness and is calculated to be markedly different in the different receptor states.

24

Liposomes, Lipid Bilayers and Model Membranes

likely carry functional consequences and have been used to predict a mechanism of ligand-speciic GPCR oligomerization based on this lipid rearrangement. Speciically, calculations show that the quantiied residual exposure at speciic TM segments (that is not alleviated by bilayer adaptation) predicts favorable contact interfaces in oligomeric arrays. Taken together, these indings suggest that distinct ligand-induced conformations of GPCRs may elicit different functional responses through differential effects on the membrane environment.

1.5

CONCLUDING REMARKS

In this chapter, we have used a series of examples to demonstrate how lipid material properties are modiied by their surrounding solution, and how, in turn, these properties inluence the intermolecular forces between membranes, their structure, and the states of guests within these membranes. The interplay between membranes, solution conditions, and macromolecules embedded within them is an important regulatory mechanism that is apparently used by the biological membrane more generally than previously thought. It is also a consideration that biotechnologists are using more often in designing effective drug formulations, as our understanding of lipid membrane material properties deepens (Zucker et al. 2012). It will be interesting to apply the new concepts described in this chapter, to ind new ways to harness lipid material properties in designing new lipid-based nanoencapsulation vehicles.

REFERENCES Akashi, K.-I. et al. 1998. Formation of giant liposomes promoted by divalent cations: Critical role of electrostatic repulsion. Biophysical Journal, 74(6), 2973–2982. Akimov, S. A. et al. 2007. Lateral tension increases the line tension between two domains in a lipid bilayer membrane. Physical Review E, 75(1), 011919. Akutsu, H. and Seelig, J. 1981. Interaction of metal ions with phosphatidylcholine bilayer membranes. Biochemistry, 20(26), 7366–7373. Andelman, D., ed., 1995. Electrostatic properties of membranes: The Poisson–Boltzmann theory. Handbook of Biological Physics, Chapter 12, North-Holland, Amsterdam: Elsevier, 603–642. Aroti, A. et  al. 2007. Effects of monovalent anions of the Hofmeister series on DPPC lipid bilayers part I: Swelling and in-plane equations of state. Biophysical Journal, 93(5), 1580–1590. Aruga, S., Kataoka, R. and Mitaku, S. 1985. Interaction between Ca2+ and dipalmitoylphosphatidylcholine membranes: I. Transition anomalies of ultrasonic properties. Biophysical Chemistry, 21(3–4), 265–275. Ben-Nun, T. et  al. 2010. X plus: A comprehensive computationally accelerated structure analysis tool for solution x-ray scattering from supramolecular self-assemblies. Journal of Applied Crystallography, 43, 1522–1531. Ben-Yaakov, D. et al. 2009. Ions in mixed dielectric solvents: Density proiles and osmotic pressure between charged interfaces. Journal of Physical Chemistry B, 113, 6001–6011. Ben-Yaakov, D. et al. 2011. Ion-speciic hydration effects: Extending the Poisson–Boltzmann theory. Current Opinion in Colloid & Interface Science, 16, 542–550. Bezrukov, S. M. 2000. Functional consequences of lipid packing stress. Current Opinion in Colloid & Interface Science, 5(3–4), 237–243. Binder, H. et al. 2001. Interaction of Zn2+ with phospholipid membranes. Biophysical Chemistry, 90(1), 57–74. Binder, H. and Zschörnig, O. 2002. The effect of metal cations on the phase behavior and hydration characteristics of phospholipid membranes. Chemistry and Physics of Lipids, 115(1–2), 39–61. Binder, W. H., Barragan, V. and Menger, F. M. 2003. Domains and rafts in lipid membranes. Angewandte Chemie International Edition, 42(47), 5802–5827. Böckmann, R. A. and Grubmüller, H. 2004. Multistep binding of divalent cations to phospholipid bilayers: A molecular dynamics study. Angewandte Chemie International Edition, 43(8), 1021–1024. Brewster, R., Pincus, P. A. and Safran, S. A. 2009. Hybrid lipids as a biological surface-active component. Biophysical Journal, 97(4), 1087–1094. Brewster, R. and Safran, S. A. 2010. Line active hybrid lipids determine domain size in phase separation of saturated and unsaturated lipids. Biophysical Journal, 98(6), L21–L23.

Soft Matter Physics of Lipid Membrane–Based Assemblies

25

Canham, P. B. 1970. Minimum energy of bending as a possible explanation of biconcave shape of human red blood cell. Journal of Theoritical Biology, 26, 61–81. Chan, D., Healy, T. and White, R. L. 1976. Electrical double-layer interactions under regulation by surface ionization equilibria-dissimilar amphoteric surfaces. Journal of the Chemical Society, Faraday Transactions, 1, 2844. Chan, D. et al. 1975. Regulation of surface potential at amphoteric surfaces during particle–particle interaction. Journal of the Chemical Society, Faraday Transactions, 1, 1046. Chapman, D. et al. 1977. Lipid phase transitions in model biomembranes: The effect of ions on phosphatidylcholine bilayers. Biochimica et Biophysica Acta (BBA)—Biomembranes, 464(2), 260–275. Chernomordik, L. V. and Kozlov, M. M. 2008. Mechanics of membrane fusion. Nature Structural & Molecular Biology, 15(7), 675–683. Claessens, M. et  al. 2007a. Entropic stabilization and equilibrium size of lipid vesicles. Langmuir, 23(11), 6315–6320. Claessens, M. et al. 2007b. Bending rigidity of mixed phospholipid bilayers and the equilibrium radius of corresponding vesicles. Physical Review E, 76(1), 011903. Claessens, M. et  al. 2004. Charged lipid vesicles: Effects of salts on bending rigidity, stability, and size. Biophysical Journal, 87(6), 3882–3893. Coorssen, J. R. and Rand, R. P. 1995. Structural effects of neutral lipids on divalent cation-induced interactions of phosphatidylserine-containing bilayers. Biophysical Journal, 68(3), 1009–1018. Cowley, A. C. et  al. 1978. Measurement of repulsive forces between charged phospholipid bilayers. Biochemistry, 17(15), 3163–3168. Cunningham, B. A. and Lis, L. J. 1989. Interactive forces between phosphatidylcholine bilayers in monovalent salt solutions. Journal of Colloid and Interface Science, 128(1), 15–25. Cunningham, B. A., Lis, L. J. and Quinn, P. J. 1986a. The inluence of monovalent anions on dipalmitoylphosphatidylcholine bilayer phase transitions: A time resolved x-ray diffraction study. Molecular Crystals and Liquid Crystals, 141, 361–367. Cunningham, B. A. et  al. 1986b. The inluence of ion species on phosphatidylcholine bilayer structure and packing. Chemistry and Physics of Lipids, 39(1–2), 135–143. Davis, J. T. 1958. Adsorption of long-chain ions. I. Proceedings of the Royal Society A, 245, 417–428. de Almeida, R. F. M. et al. 2005. Lipid rafts have different sizes depending on membrane composition: A timeresolved luorescence resonance energy transfer study. Journal of Molecular Biology, 346(4), 1109–1120. Degennes, P. G. and Taupin, C. 1982. Micro-emulsions and the lexibility of oil–water interfaces. Journal of Physical Chemistry, 86(13), 2294–2304. Deme, B. et al. 2002a. Giant collective luctuations of charged membranes at the lamellar-to-vesicle unbinding transition. 1. Characterization of a new lipid morphology by SANS, SAXS, and electron microscopy. Langmuir, 18(4), 997–1004. Deme, B., Dubois, M. and Zemb, T. 2002b. Giant collective luctuations of charged membranes at the lamellarto-vesicle unbinding transition. 2. Equation of state in the absence of salt. Langmuir, 18(4), 1005–1013. Dubois, M. and Zemb, T. 1991. Osmotic-pressure of diluted charged bilayers. Abstracts of Papers of the American Chemical Society, 201, 44-Coll. Dubois, M. and Zemb, T. 1998. Molecular force measurement between bilayers using a controlled osmoticpressure x-ray scattering technique. Journal De Physique IV, 8(P5), 55–62. Dubois, M. and Zemb, T. 2000. Swelling limits for bilayer microstructures: The implosion of lamellar structure versus disordered lamellae. Current Opinion in Colloid & Interface Science, 5(1–2), 27–37. Dubois, M. et  al. 1992. Osmotic-pressure and salt exclusion in electrostatically swollen lamellar phases. Journal of Chemical Physics, 96(3), 2278–2286. Dubois, M. et al. 1998. Equation of state of a charged bilayer system: Measure of the entropy of the lamellar– lamellar transition in DDABr. Journal of Chemical Physics, 108(18), 7855–7869. Dvir, T. et al. 2013. Charged membranes under coninement induced by polymer-, salt-, or ionic liquid solutions. Soft Matter, 9(44), 10640–10649. Edidin, M. 2001. Membrane cholesterol, protein phosphorylation, and lipid rafts. Science STKE, 2001(67), PE1. Elliott, R. et al. 2005. Phase separation of saturated and mono-unsaturated lipids as determined from a microscopic model. The Journal of Chemical Physics, 122(4), 044904. Epand, R. M. 2006. Cholesterol and the interaction of proteins with membrane domains. Progress in Lipid Research, 45(4), 279–294. Evans, D. P. and Wennerstrom, H., 1999. The Colloidal Domain. 2nd ed. Wiley-VCH: New York.

26

Liposomes, Lipid Bilayers and Model Membranes

Evans, E. A. 1974. Bending resistance and chemically-induced moments in membrane bilayers. Biophysical Journal, 13, 923–931. Evans, E. A. and Parsegian, V. A. 1986. Thermal–mechanical luctuations enhance repulsion between biomolecular layers. Proceedings of the National Academy of Sciences of the United States of America, 83(19), 7132–7136. Ganesan, M. G., Schwinke, D. L. and Weiner, N. 1982. Effect of Ca2+ on thermotropic properties of saturated phosphatidylcholine liposomes. Biochimica et Biophysica Acta (BBA)—Biomembranes, 686(2), 245–248. Gondre-Lewis, M. C. et al. 2006. Abnormal sterols in cholesterol-deiciency diseases cause secretory granule malformation and decreased membrane curvature. Journal of Cell Science, 119(9), 1876–1885. Graddick, W. F. et al. 1979. Order–disorder and the pretransition in dipalmitoyl phosphatidylcholine multilayers. Biochemical and Biophysical Research Communications, 88(3), 907–912. Grasdalen, H. et al. 1977. Surface potential effects on metal ion binding to phosphatidylcholine membranes. 31P NMR study of lanthanide and calcium ion binding to egg-yolk lecithin vesicles. Biochimica et Biophysica Acta (BBA)—Biomembranes, 469(2), 151–162. Harries, D. et al. 2006. Ion induced lamellar–lamellar phase transition in charged surfactant systems. Journal of Chemical Physics, 124(22), 224702. Hauser, H. et al. 1981. Preferred conformation and molecular packing of phosphatidylethanolamine and phosphatidylcholine. Biochimica et Biophysica Acta (BBA)—Reviews on Biomembranes, 650(1), 21–51. Hauser, H. et al. 1975. Ion-binding to phospholipids. European Journal of Biochemistry, 58(1), 133–144. Hauser, H. and Shipley, G. G. 1984. Interactions of divalent cations with phosphatidylserine bilayer membranes. Biochemistry, 23(1), 34–41. Helfrich, W. 1973. Elastic properties of lipid bilayers—Theory and possible experiments. Zeitschrift Fur Naturforschung C—A Journal of Biosciences, C 28(11-1), 693–703. Helfrich, W. 1978. Steric interaction of luid membranes in multilayer systems. Zeitschrift Fur Naturforschung Section A—A Journal of Physical Sciences, 33(3), 305–315. Huster, D. and Arnold, K. 1998. Ca2+-mediated interaction between dextran sulfate and dimyristoyl-sn-glycero3-phosphocholine surfaces studied by 2H nuclear magnetic resonance. Biophysical Journal, 75(2), 909–916. Huster, D., Arnold, K. and Gawrisch, K. 2000. Strength of Ca2+ binding to retinal lipid membranes: Consequences for lipid organization. Biophysical Journal, 78(6), 3011–3018. Huster, D. et al. 1999. Investigation of phospholipid area compression induced by calcium-mediated dextran sulfate interaction. Biophysical Journal, 77(2), 879–887. Inoko, Y. et  al. 1975. Effects of cations on dipalmitoyl phosphatidylcholine–cholesterol–water systems. Biochimica et Biophysica Acta, 413(1), 24–32. Israelachvili, J. N. 2011. Intermolecular and Surface Forces. 3rd ed. Academic Press: Amsterdam. Israelachvili, J. N., Mitchell, D. J. and Ninham, B. W. 1977. Theory of self-assembly of lipid bilayers and vesicles. Biochimica et Biophysica Acta, 470(2), 185–201. Janiak, M. J., Small, D. M. and Shipley, G. G. 1976. Nature of the thermal pretransition of synthetic phospholipids: Dimyristoyl- and dipalmitoyllecithin. Biochemistry, 15(21), 4575–4580. Kataoka, R. et  al. 1985. Interaction between Ca2+ and dipalmitoylphosphatidylcholine membranes: II. Fluorescence anisotropy study. Biophysical Chemistry, 21(3–4), 277–284. Khelashvili, G. and Harries, D. 2013. How sterol tilt regulates properties and organization of lipid membranes and membrane insertions. Chemistry and Physics of Lipids, 169, 113–123. Khelashvili, G., Weinstein, H. and Harries, D. 2008. Protein diffusion on charged membranes: A dynamic mean-ield model describes time evolution and lipid reorganization. Biophysical Journal, 94, 2580–2597. Khelashvili, G., Harries, D. and Weinstein, H. 2009. Modeling membrane deformations and lipid demixing upon protein–membrane interaction: The BAR dimer adsorption. Biophysical Journal, 97(6), 1626–1635. Khelashvili, G., Pabst, G. and Harries, D. 2010. Cholesterol orientation and tilt modulus in DMPC bilayers. Journal of Physical Chemistry B, 114(22), 7524–7534. Khelashvili, G., Kollmitzer, B., Heftberger, P. Pabst, G. and Harries, D. 2013. Calculating the bending modulus for multicomponent lipid membranes in different thermodynamic phases. Journal of Chemical Theory and Computation, 9(9), 3866–3871. Khelashvili, G. et  al. 2011. Impact of sterol tilt on membrane bending rigidity in cholesterol and 7DHCcontaining DMPC membranes. Soft Matter, 7(21), 10299–10312. Klein, J. W. et al. 1987. Phospholipid dependence of calcium ion effects on electrophoretic mobilities of liposomes. Chemistry and Physics of Lipids, 43(1), 13–23.

Soft Matter Physics of Lipid Membrane–Based Assemblies

27

Kuzmin, P. I. et al. 2005. Line tension and interaction energies of membrane rafts calculated from lipid splay and tilt. Biophysical Journal, 88(2), 1120–1133. Lau, A. L. Y. et al. 1980. The adsorption of alkaline earth cations to phosphatidyl choline bilayer membranes: A unique effect of calcium. Bioelectrochemistry: Ions, Surfaces, Membranes. Washington, DC: American Chemical Society, 49–56. Lee, A. G. 1977. Lipid phase transitions and phase diagrams II. Mixtures involving lipids. Biochimica et Biophysica Acta (BBA)—Reviews on Biomembranes, 472(3–4), 285–344. Leneveu, D. M., Rand, R. P. and Parsegian, V. A. 1976. Measurement of forces between lecithin bilayers. Nature, 259(5544), 601–603. Lentz, B. R., Barenholz, Y. and Thompson, T. E. 1976. Fluorescence depolarization studies of phase transitions and luidity in phospholipid bilayers. 2. Two-component phosphatidylcholine liposomes. Biochemistry, 15(20), 4529–4537. Leontidis, E. et al. 2007. Effects of monovalent anions of the Hofmeister series on DPPC lipid bilayers part II: Modeling the perpendicular and lateral equation-of-state. Biophysical Journal, 93(5), 1591–1607. Lipowsky, R. 2002. Domains and rafts in membranes—Hidden dimensions of selforganization. Journal of Biological Physics, 28(2), 195–210. Lipowsky, R. and Leibler, S. 1986. Unbinding transitions of interacting membranes. Physical Review Letters, 56(23), 2541–2544. Lipowsky, R. and Sackmann, E., eds., 1995. Structure and Dynamics of Membranes: Generic and Speciic Interactions. Amsterdam: Elsevier. Lis, L. J. et al. 1981a. Adsorption of divalent cations to a variety of phosphatidylcholine bilayers. Biochemistry, 20(7), 1771–1777. Lis, L. J., Parsegian, V. A. and Rand, R. P. 1981b. Binding of divalent cations to dipalmitoylphosphatidylcholine bilayers and its effect on bilayer interaction. Biochemistry, 20(7), 1761–1770. Lis, L. J. et al. 1982. Interactions between neutral phospholipid bilayer membranes. Biophysical Journal, 37(3), 657–665. Lis, L. J., Rand, R. P. and Parsegian, V. A. 1980. Measurement of the adsorption of Ca2+ and Mg2+ to phosphatidyl choline bilayers. Bioelectrochemistry: Ions, Surfaces, Membranes. Washington, DC: American Chemical Society, 41–47. London, E. 2005. How principles of domain formation in model membranes may explain ambiguities concerning lipid raft formation in cells. Biochimica et Biophysica Acta (BBA)—Molecular Cell Research, 1746(3), 203–220. Lundbaek, J. A. et al. 2010. Lipid bilayer regulation of membrane protein function: Gramicidin channels as molecular force probes. Journal of the Royal Society Interface, 7(44), 373–395. Macdonald, P. M. and Seelig, J. 1987a. Calcium binding to mixed cardiolipin–phosphatidylcholine bilayers as studied by deuterium nuclear magnetic resonance. Biochemistry, 26(19), 6292–6298. Macdonald, P. M. and Seelig, J. 1987b. Calcium binding to mixed phosphatidylglycerol–phosphatidylcholine bilayers as studied by deuterium nuclear magnetic resonance. Biochemistry, 26(5), 1231–1240. Macdonald, P. M. and Seelig, J. 1988. Anion binding to neutral and positively charged lipid membranes. Biochemistry, 27(18), 6769–6775. Marra, J. and Israelachvili, J. 1985. Direct measurements of forces between phosphatidylcholine and phosphatidylethanolamine bilayers in aqueous electrolyte solutions. Biochemistry, 24(17), 4608–4618. Marsh, D. 2006. Elastic curvature constants of lipid monolayers and bilayers. Chemistry and Physics of Lipids, 144(2), 146–159. Marsh, D. 2007. Lateral pressure proile, spontaneous curvature frustration, and the incorporation and conformation of proteins in membranes. Biophysical Journal, 93(11), 3884–3899. May, S. 2000. Theories on structural perturbations of lipid bilayers. Current Opinion in Colloid and Interface Science, 5(3–4), 244–249. May, S. and Ben-Shaul, A. 1999. Molecular theory of lipid–protein interaction and the l-alpha-H-II transition. Biophysical Journal, 76(2), 751–767. McLaughlin, A., Grathwohl, C. and McLaughlin, S. 1978. The adsorption of divalent cations to phosphatidylcholine bilayer membranes. Biochimica et Biophysica Acta (BBA)—Biomembranes, 513(3), 338–357. McLaughlin, S. et al. 1981. Adsorption of divalent cations to bilayer membranes containing phosphatidylserine. The Journal of General Physiology, 77(4), 445–473. McManus, J. J., Radler, J. O. and Dawson, K. A. 2003. Does calcium turn a zwitterionic lipid cationic? The Journal of Physical Chemistry B, 107(36), 9869–9875.

28

Liposomes, Lipid Bilayers and Model Membranes

Mengistu, D. H., Bohinc, K. and May, S. 2009. Binding of DNA to zwitterionic lipid layers mediated by divalent cations. The Journal of Physical Chemistry B, 113(36), 12277–12282. Mishima, K., Satoh, K. and Ogihara, T. 1984. The effects of pH and ions on myelin igure formation in phospholipid–water system. Chemical Physics Letters, 106(6), 513–516. Mondal, S. et al. 2011a. Quantitative modeling of membrane deformations by multihelical membrane proteins: Application to G-protein coupled receptors. Biophysical Journal, 101(9), 2092–2101. Mondal, S. et al. 2011b. Quantitative analysis of membrane deformation by multi-helical transmembrane proteins. Biophysical Journal, 100(3), 492–492. Morse, D. C. 1994. Topological instabilities and phase-behavior of luid membranes. Physical Review E, 50(4), R2423–R2426. Moshe, L. et al. 2013. Modulating the structure and interactions of lipid–peptide complexes by varying membrane composition and solution conditions. Soft Matter, 9(29), 7117–7126. Mouritsen, O. G. 2005. Life—As a Matter of Fat. Springer: Berlin. Nagle, J. F. and Tristram-Nagle, S. 2000. Structure of lipid bilayers. Biochimica et Biophysica Acta—Reviews on Biomembranes, 1469(3), 159–195. Ninham, B. W. and Parsegian, V. A. 1971. Electrostatic potential between surfaces bearing ionizable groups in ionic equilibrium with physiologic saline solution. Journal of Theoretical Biology, 31(3), 405–428. Nyholm, T. K. M., Ozdirekcan, S. and Killian, J. A. 2007. How protein transmembrane segments sense the lipid environment. Biochemistry, 46(6), 1457–1465. Ohki, K. et al. 1981. Physical properties of phosphatidylcholine–phosphatidylinositol liposomes in relation to a calcium effect. Biochimica et Biophysica Acta (BBA)—Biomembranes, 644(2), 165–174. Pabst, G. et al. 2007. Rigidiication of neutral lipid bilayers in the presence of salts. Biophysical Journal, 93(8), 2688–2696. Pabst, G. et al. 2010. Applications of neutron and x-ray scattering to the study of biologically relevant model membranes. Chemistry and Physics of Lipids, 163(6), 460–479. Paila, Y. D. et al. 2008. Signaling by the human serotonin(1A) receptor is impaired in cellular model of Smith– Lemli–Opitz syndrome. Biochimica et Biophysica Acta—Biomembranes, 1778(6), 1508–1516. Pandit, S. A., Bostick, D. and Berkowitz, M. L. 2003. Molecular dynamics simulation of a dipalmitoylphosphatidylcholine bilayer with NaCl. Biophysical Journal, 84(6), 3743–3750. Parsegian, V. A. 2006. Van der Waals Forces A Handbook for Biologists, Chemists, Engineers and Physicists. New York: Cambridge University Press. Petrache, H. I., Harris, D. and Parsegian, V. A. 2005. Alteration of lipid membrane elasticity by cholesterol and its metabolic precursors. Macromolecular Symposia, 219, 39–50. Petrache, H. I. et al. 1998. Interbilayer interactions from high-resolution x-ray scattering. Physical Review E, 57(6), 7014–7024. Petrache, H. I. et  al. 2006a. Swelling of phospholipids by monovalent salt. Journal of Lipid Research, 47(2), 302–309. Petrache, H. I. et al. 2006b. Salt screening and speciic ion adsorption determine neutral-lipid membrane interactions. Proceedings of the National Academy of Sciences of the United States of America, 103(21), 7982–7987. Philipson, K. D. 1984. Interaction of charged amphiphiles with Na+–Ca2+ exchange in cardiac sarcolemmal vesicles. Journal of Biological Chemistry, 259(22), 13999–14002. Porte, G. et al. 1989. l-Alpha to l3—A topology driven transition in phases of ininite luid membranes. Journal De Physique, 50(11), 1335–1347. Porter, F. D. and Herman, G. E. 2011. Malformation syndromes caused by disorders of cholesterol synthesis. Journal of Lipid Research, 52(1), 6–34. Pralle, A. et al. 2000. Sphingolipid–cholesterol rafts diffuse as small entities in the plasma membrane of mammalian cells. The Journal of Cell Biology, 148(5), 997–1007. Prieve, D. C. and Ruckenstein, E. 1976. The surface potential and double-layer interaction force between surfaces characterized by multiple ionizable groups. Journal of Theoretical Biology, 56, 205. Rand, R. P., Chapman, D. and Larsson, K. 1975. Tilted hydrocarbon chains of dipalmitoyl lecithin become perpendicular to the bilayer before melting. Biophysical Journal, 15(11), 1117–1124. Rand, R. P. and Parsegian, V. A. 1989. Hydration forces between phospholipid bilayers. Biochimica et Biophysica Acta (BBA)—Reviews on Biomembranes, 988(3), 351–376. Raviv, U. et al. 2003. Lubrication by charged polymers. Nature, 425(6954), 163–165. Raviv, U. and Klein, J. 2002. Fluidity of bound hydration layers. Science, 297(5586), 1540–1543. Rog, T. et al. 2009. Ordering effects of cholesterol and its analogues. Biochimica et Biophysica Acta— Biomembranes, 1788(1), 97–121.

Soft Matter Physics of Lipid Membrane–Based Assemblies

29

Rog, T. et al. 2008. Comparison of cholesterol and its direct precursors along the biosynthetic pathway: Effects of cholesterol, desmosterol and 7-dehydrocholesterol on saturated and unsaturated lipid bilayers. Journal of Chemical Physics, 129(15), 154508. Roux, D. and Sainya, C. R. 1988. A synchrotron x-ray study of competing undulation and electrostatic interlayer interactions in luid multimembrane lyotropic phases. Journal De Physique, 49(2), 307–318. Rydall, J. R. and Macdonald, P. M. 1992. Investigation of anion binding to neutral lipid membranes using deuterium NMR. Biochemistry, 31(4), 1092–1099. Sabín, J. et al. 2007. Fractal aggregates induced by liposome–liposome interaction in the presence of Ca2+. European Physical Journal E: Soft Matter and Biological Physics, 24(2), 201–210. Sachs, J. N. and Woolf, T. B. 2003. Understanding the Hofmeister effect in interactions between chaotropic anions and lipid bilayers: Molecular dynamics simulations. Journal of the American Chemical Society, 125(29), 8742–8743. Sackmann, E. 1994. Membrane bending energy concept of vesicle- and cell-shapes and shape-transitions. FEBS Letters, 346, 3–16. Safran, S. A. 1994. Statistical Thermodynamics of Surfaces, Interfaces and Membranes. Reading, MA: Addison-Wesley. Safran, S. A. 2003. Statistical Thermodynamics of Surfaces, Interfaces, and Membranes. Westview Press: Colorado. Satoh, K. 1995. Determination of binding constants of Ca2+, Na+, and Cl− ions to liposomal membranes of dipalmitoylphosphatidylcholine at gel phase by particle electrophoresis. Biochimica et Biophysica Acta (BBA)—Biomembranes, 1239(2), 239–248. Seelig, J., Gally, H.-U. and Wohlgemuth, R. 1977. Orientation and lexibility of the choline head group in phosphatidylcholine bilayers. Biochimica et Biophysica Acta (BBA)—Biomembranes, 467(2), 109–119. Shah, D. O. and Schulman, J. H. 1967. Inluence of calcium, cholesterol, and unsaturation on lecithin monolayers. Journal of Lipid Research, 8(3), 215–226. Shan, J. F. et  al. 2012. Ligand-dependent conformations and dynamics of the serotonin 5-HT2A receptor determine its activation and membrane-driven oligomerization properties. PLoS Computational Biology, 8(4), e1002473. Shepherd, J. C. W. and Büldt, G. 1978. Zwitterionic dipoles as a dielectric probe for investigating head group mobility in phospholipid membranes. Biochimica et Biophysica Acta (BBA)—Biomembranes, 514(1), 83–94. Shibata, T. 1990. Pulse NMR study of the interaction of calcium ion with dipalmitoylphosphatidylcholine lamellae. Chemistry and Physics of Lipids, 53(1), 47–52. Shrivastava, S. et  al. 2008. Differential effects of cholesterol and its immediate biosynthetic precursors on membrane organization. Biochemistry, 47(20), 5668–5677. Simon, S. A. et al. 1975. A calorimetric and monolayer investigation of the inluence of ions on the thermodynamic properties of phosphatidylcholine. Biochimica et Biophysica Acta (BBA)—Biomembranes, 375(3), 317–326. Sinn, C. G., Antonietti, M. and Dimova, R. 2006. Binding of calcium to phosphatidylcholine–phosphatidylserine membranes. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 282, 410–419. Sperotto, M. M., May, S. and Baumgaertner, A. 2006. Modelling of proteins in membranes. Chemistry and Physics of Lipids, 141(1–2), 2–29. Steiner, A. et al. 2012. Entropic attraction condenses like-charged interfaces composed of self-assembled molecules. Langmuir, 28(5), 2604–2613. Sundler, R. and Papahadjopoulos, D. 1981. Control of membrane fusion by phospholid head groups I. Phosphatidate/phosphatidylinositol speciicity. Biochimica et Biophysica Acta (BBA)—Biomembranes, 649(3), 743–750. Szekely, O. et al. 2011a. Regulating the size and stabilization of lipid raft-like domains and using calcium ions as their probe. Langmuir, 27(24), 14767–14775. Szekely, O. et al. 2011b. The structure of ions and zwitterionic lipids regulates the charge of dipolar membranes. Langmuir, 27(12), 7419–7438. Szekely, P. et al. 2011c. Effect of temperature on the structure of charged membranes. Journal of Physical Chemistry B, 115(49), 14501–14506. Szekely, P. et al. 2012. Effect of temperature on the interactions between dipolar membranes. Journal of Physical Chemistry B, 116(11), 3519–3524. Szleifer, I. et al. 1990. Molecular theory of curvature elasticity in surfactant ilms. Journal of Chemical Physics, 92(11), 6800–6817. Szleifer, I. et al. 1988. Curvature elasticity of pure and mixed surfactant ilms. Physical Review Letters, 60(19), 1966–1969.

30

Liposomes, Lipid Bilayers and Model Membranes

Tardieu, A., Luzzati, V. and Reman, F. C. 1973. Structure and polymorphism of hydrocarbon chains of lipids— Study of lecithin–water phases. Journal of Molecular Biology, 75(4), 711–733. Tattrie, N. H., Bennett, J. R. and Cyr, R. 1968. Maximum and minimum values for lecithin classes from various biological sources. Canadian Journal of Biochemistry, 46(8), 819–824. Tatulian, S. A. 1987. Binding of alkaline–earth metal cations and some anions to phosphatidylcholine liposomes. European Journal of Biochemistry, 170(1/2), 413–420. Tatulian, S. A. et al. 1991. A neutron diffraction study of the inluence of ions on phospholipid membrane interactions. Biochimica et Biophysica Acta (BBA)—Biomembranes, 1070(1), 143–151. Templer, R. H., Khoo, B. J. and Seddon, J. M. 1998. Gaussian curvature modulus of an amphiphilic monolayer. Langmuir, 14(26), 7427–7434. Tokumasu, F. et al. 2003. Nanoscopic lipid domain dynamics revealed by atomic force microscopy. Biophysical Journal, 84(4), 2609–2618. Tristram-Nagle, S. and Nagle, J. F. 2004. Lipid bilayers: Thermodynamics, structure, luctuations, and interactions. Chemistry and Physics of Lipids, 127(1), 3–14. Tulenko, T. N. et  al. 2006. A membrane defect in the pathogenesis of the Smith–Lemli–Opitz syndrome. Journal of Lipid Research, 47(1), 134–143. Uhríková, D. et al. 2005. The structure of DNA–DOPC aggregates formed in presence of calcium and magnesium ions: A small-angle synchrotron x-ray diffraction study. Biochimica et Biophysica Acta (BBA)— Biomembranes, 1713(1), 15–28. Uhríková, D. et al. 2008. Structural changes in dipalmitoylphosphatidylcholine bilayer promoted by Ca2+ ions: A small-angle neutron scattering study. Chemistry and Physics of Lipids, 155(2), 80–89. Veatch, S. L. and Keller, S. L. 2003. Separation of liquid phases in giant vesicles of ternary mixtures of phospholipids and cholesterol. Biophysical Journal, 85(5), 3074–3083. von Grunberg, H. H. 1999. Chemical charge regulation and charge renormalization in concentrated colloidal suspensions. Journal of Colloid and Interface Science, 219, 339. Westman, J. and Göran Eriksson, L. E. 1979. The interaction of various lanthanide ions and some anions with phosphatidylcholine vesicle membranes: A 31P NMR study of the surface potential effects. Biochimica et Biophysica Acta (BBA)—Biomembranes, 557(1), 62–78. Wilschut, J., Duezguenes, N. and Papahadjopoulos, D. 1981. Calcium/magnesium speciicity in membrane fusion: Kinetics of aggregation and fusion of phosphatidylserine vesicles and the role of bilayer curvature. Biochemistry, 20(11), 3126–3133. Yabusaki, K. K. and Wells, M. A. 1975. Binding of calcium to phosphatidylcholines as determined by proton magnetic resonance and infrared spectroscopy. Biochemistry, 14(1), 162–166. Yamada, N. L. et al. 2005. SAXS, SANS and NSE studies on “unbound state” in DPPC/water/CaCl2 system. Journal of the Physical Society of Japan, 74(10), 2853–2859. Yeap, P. K. et al. 2008. Effect of calcium ions on the density of lecithin and its effective molecular volume in lecithin–water dispersions. Chemistry and Physics of Lipids, 151(1), 1–9. Yethiraj, A. and Weisshaar, J. C. 2007. Why are lipid rafts not observed in vivo? Biophysical Journal, 93(9), 3113–3119. Zidovetzki, R., Atiya, A. W. and De Boeck, H. 1989. Effect of divalent cations on the structure of dipalmitoylphosphatidylcholine and phosphatidylcholine phosphatidylglycerol bilayers—An 2H-NMR study. Membrane Biochemistry, 8(3), 177–186. Zucker, D. et  al. 2012. Characterization of PEGylated nanoliposomes co-remotely loaded with topotecan and vincristine: Relating structure and pharmacokinetics to therapeutic eficacy. Journal of Controlled Release, 160(2), 281–289.

2

Nonlamellar Lipid Aggregates Charlotte E. Conn and John M. Seddon

CONTENTS 2.1 Introduction ............................................................................................................................ 31 2.2 Amphiphile Self-Assembly and Lipid Packing Parameter ..................................................... 32 2.3 Interfacial Curvature .............................................................................................................. 33 2.4 Generic Lyotropic Phase Diagram..........................................................................................34 2.5 Factors Controlling the Preferred Interfacial Curvature ........................................................ 35 2.6 Curvature Elastic Energy........................................................................................................ 36 2.7 2D Hexagonal Phases, Type I and Type II ............................................................................. 38 2.8 Chain Packing Frustration ...................................................................................................... 38 2.9 3D Bicontinuous Cubic Phases, Type I and Type II ............................................................... 39 2.10 Sponge Phase .......................................................................................................................... 39 2.11 3D Ordered Micellar Phases, Type I and Type II ...................................................................40 2.12 Intermediate Phases ................................................................................................................ 42 2.13 Control of Phase Behavior ...................................................................................................... 42 2.14 Phase Transitions between Nonlamellar Phases ..................................................................... 43 2.15 Dispersed Nanoparticles of Nonlamellar Lipid Aggregates...................................................44 2.16 Examples of Applications of Nonlamellar Phases..................................................................44 Acknowledgments............................................................................................................................46 References ........................................................................................................................................46

2.1 INTRODUCTION Amphiphilic molecules such as phospholipids, glycolipids, and monoacylglycerols (Figure 2.1) can spontaneously self-assemble into a wide range of ordered lyotropic liquid-crystalline mesophases (Shearman et  al., 2006, 2010). The structural ordering within such phases can extend from the nanoscale (2–3 nm), up to much larger values (>100 nm), although locally the molecules are liquidlike.* Such lyotropic phases, in addition to their great relevance to understanding biological membrane structures, also have great potential in applications such as nanoencapsulation, nanoreactors, drug delivery, gene transfer, and delivery of Si-RNA to cells (Fong and Drummond, 2012, Leal et al., 2010). Apart from the bilayer lamellar L α phase, all other luid lyotropic mesophases are based upon ordered arrangements of curved interfaces, separating the amphiphiles from the water regions. These ordered curved phases may be divided into three main structural classes: 2D packings of cylindrical interfaces (discontinuous phases), 3D packings of saddle surfaces (bicontinuous phases), and 3D packings of spherical/ellipsoidal interfaces (discontinuous phases). Furthermore, each of these three classes occurs in two variants, either oil-in-water (type I), where the interfaces have net mean curvature toward the lipid hydrocarbon chain regions, or water-in-oil (inverse, type II), where the interfaces curve toward the water regions and away from the hydrocarbon chains (Figure 2.2). Type I curved mesophases generally break up into disordered micellar solutions upon high dilution in water. Type II mesophases, on the other hand, are usually stable in the presence of an excess *

Or soft solids in the case of the low-temperature lamellar gel phases.

31

32

Liposomes, Lipid Bilayers and Model Membranes O

(a)

O

OH OH

OH (b)

OH OH

(c)

O

O P O –O O

O O

N+

O

(d)

O O

O

O O

O– O P O

NH+3

FIGURE 2.1 Examples of amphiphilic lipids: (a) monoolein; (b) phytantriol; (c) DOPC; (d) DOPE.

FIGURE 2.2 By convention, type I lyotropic mesophases have positive interfacial mean curvature, whereas type II (inverse) phases have negative interfacial mean curvature. (From Shearman, G. C. et al. 2006. J. Phys. Condens. Mat., 18, S1105–S1124.)

water phase, and are therefore much more suitable for any applications where stable ordered selfassembled structures are required.

2.2

AMPHIPHILE SELF-ASSEMBLY AND LIPID PACKING PARAMETER

The hydrophobic effect causes amphiphiles in water to aggregate into micelles or ordered mesophases above a certain concentration known as the critical micelle concentration, or cmc. The value of the cmc depends strongly on the length and number of hydrocarbon chains, and varies from subnanomolar for long double-chained phospholipids, to submolar for short, single-chained amphiphiles. A useful way to think about amphiphile self-assembly is to associate a given amphiphile, under speciied conditions of temperature, pressure, hydration, and so on, with a “shape” that will determine which type of aggregate structure will be preferred. The simplest treatment assumes that the amphiphile, of hydrocarbon chain(s) volume v, has an optimum headgroup area ao, and that the

33

Nonlamellar Lipid Aggregates

FIGURE 2.3 The preferred “shape” of an amphiphile is dictated by its packing parameter P, which can have values of P < 1 (left), P = 1 (center), or P > 1 (right). (From Shearman, G. C. et al. 2006. J. Phys. Condens. Mat., 18, S1105–S1124.)

chain packing energy is constant and independent of size and shape, so long as no part of the interior is more than the maximum length lc of the hydrocarbon chain(s) from the polar–nonpolar interface. This then deines an amphiphile “packing parameter” (see Figure 2.3) P=

v ao lc

A combination of entropy of mixing and geometric considerations then predicts the following aggregate structures with increasing packing parameter: P < 1/3: spherical micelles 1/3 < P 1: inverse phases

2.3

INTERFACIAL CURVATURE

At any point P on a surface, we locate the two principal curvatures c1 and c2, which are at right angles to each other (Figure 2.4). They combine to give the mean and the Gaussian curvature at that point: 1 (c + c2 ) 2 1 K = c1c2

H =

The sign of H is arbitrary: it is commonly deined to be positive for curvature of the interface toward the hydrocarbon chains (i.e., as in a normal micelle). Thus, the inverse phases will have negative H. The sign of the Gaussian curvature K characterizes the nature of the surface: • c1 and c2 have the same sign: K > 0; elliptic surface (e.g., spheres). • Either c1 or c2 is zero: K = 0; parabolic surface (e.g., cylinders or lat sheets). • c1 and c2 have opposite sign: K < 0; hyperbolic surface (e.g., saddle surfaces). When c1 = −c2 at all points, we have a minimal surface (with H = 0 at all points on the surface, and K varying smoothly from zero at lat points, to being most negative at saddle points).

34

Liposomes, Lipid Bilayers and Model Membranes

FIGURE 2.4 The principal curvatures at a point on a surface are given by c1 = 1/R1 and c2 = 1/R2. (From Shearman, G. C. et al. 2006. J. Phys. Condens. Mat., 18, S1105–S1124.)

Such minimal surfaces are intimately involved in the formation of sponge-like lyotropic phases (e.g., bicontinuous cubic phases). Note that bilayers and cylinders have different values of H, but both have zero K; inverse spherical micelles have the opposite sign of H from normal micelles, but both have positive values of K. For a curved interface, the cross-sectional area depends on the distance d moved perpendicular to the interface. The result from geometry is a(d ) = ao [1 + 2 Hd + Kd 2 ] where we take the positive d-direction to be moving from the chain end toward the polar headgroup. Thus, a luid bilayer draped onto a minimal surface will have maximal cross-sectional area per molecule at the bilayer mid-plane, and minimal area at the polar headgroups. Note that the headgroup area will vary smoothly along the polar–nonpolar interface, due to the variation in the underlying Gaussian curvature at the bilayer mid-plane.

2.4

GENERIC LYOTROPIC PHASE DIAGRAM

It is useful to arrange the various lyotropic mesophases according to their average interfacial mean curvature H (see Figure 2.5). The lat luid lamellar phase (H = 0) occupies the central region of this “phase diagram.” The interfacial tension γ is balanced by the headgroup lateral repulsions πh, and also by the lateral repulsions between the chains πch. Calculations and computer simulations show that πch is maximal in the middle of the chain region, and falls off toward both the interface and the methyl endgroups (bilayer midplane). Thus, the local pressure π(z) varies continuously across the lipid bilayer, giving rise to a lateral pressure proile. π(z) is positive in the headgroup region, strongly negative at the water/hydrocarbon chain interface, and positive in the chain region (see Figure 2.6). Note that the local pressure within a bilayer can be very high, ≥500 atm, which is large enough to induce conformational changes of any proteins embedded in the lipid membrane. For a luid bilayer to be at mechanical equilibrium, the zeroth moment of the lateral pressure proile must be zero:

∫ π (z)dz = 0

35

Nonlamellar Lipid Aggregates

LII

III

HII

QII RbII MhII BcII

Lα

QI

HI

II LI

RbI MhI BcI

Mean interfacial curvature

FIGURE 2.5 A schematic lyotropic “phase diagram” showing the natural sequence of phases in terms of the average interfacial mean curvature. (Adapted from Seddon, J. M. and Templer, R. H. 1993. Cubic phases of self-assembled amphiphilic aggregates. Philos. Trans. R. Soc. London, Ser. A, 344, 377–401 with permission from The Royal Society, copyright (1993).) z Headgroup pressure

Interfacial pressure π (z) Chain pressure

FIGURE 2.6 Typical lateral stress proile across a luid lipid bilayer. (Adapted from Shearman, G. C. et al. 2006. J. Phys. Condens. Mat., 18, S1105–S1124.)

and the area per molecule will adjust itself to its optimal value ao in order to bring this about. Thus, increasing the temperature will tend to increase ao; conversely, increasing the hydrostatic pressure will tend to reduce ao. However, increasing the temperature will also build up a strain in the bilayer, due to the increasing desire for splay caused by the increasing conformational disorder of the chains, which tries to force each monolayer to curve toward the water region (for P > 1). In practice, experimental phase diagrams must still be determined experimentally. The temperature–composition phase diagrams for two common lipids, monoolein and phytantriol, are given in Figure 2.7. Note that not all phases are observed in a given experimental phase diagram, and the sequence as a function of water content may not be as expected (see Section 2.9).

2.5

FACTORS CONTROLLING THE PREFERRED INTERFACIAL CURVATURE

The most direct way to induce spontaneous curvature in a bilayer is to engineer an asymmetric lipid composition between the two monolayers making up the bilayer. Most cell membranes do in fact have such asymmetric compositions, maintained by cellular processes, although the underlying reasons and implications of this are still very poorly understood. However, asymmetry will be rapidly lost if the spontaneous lip-lop of lipids from one monolayer across to the other occurs at an appreciable rate. For bicontinuous phases such as the inverse bicontinuous cubic phases, on the other hand, indeinite swelling in water could in principle occur without incurring signiicant packing penalty from the chain region. However, calculations suggest that thermal luctuations should destroy such phases for lattice parameters larger than approximately 30 nm (Bruinsma, 1992), although this upper limit depends on the bilayer bending modulus, and might therefore be overcome by appropriate choice of lipid.

36

Liposomes, Lipid Bilayers and Model Membranes

(a)

(b) 120 Pn3m

80

Lα

Lα + Ia3d

Pn3m + Ia3d

Lc + Pn3m Lc + water

0 Lc

L2 + H2O

60

Pn3m + water

Ia3d

40

HII + water

Temperature, °C

HII

FI

Temperature, °C

70

FI + water

40

10 20 30 40 Composition, % (w/w) water

50

20

HII + H2O

Q224 + H2O Q230

30

Lc + ice

HII

L2

50

Lα 5

Q224

10 15 20 25 30 Composition, % (w/w) water

35

FIGURE 2.7 Experimental phase diagrams for (a) monoolein and (b) phytantriol. (Reprinted from Biomaterials, 21, Qiu, H. and Caffrey, M., The phase diagram of the monoolein/water system: Metastability and equilibrium aspects, 223–234, Copyright (2000), with permission from Elsevier; Reprinted with permission from Barauskas, J. and Landh, T. 2003. Phase behavior of the phytantriol/water system. Langmuir, 19, 9562–9565. Copyright 2003, American Chemical Society.)

2.6 CURVATURE ELASTIC ENERGY It costs elastic energy to bend a lipid membrane away from its preferred shape. The curvature elastic energy per unit area of a lipid bilayer is given (Helfrich, 1973) by gc = 2κ ( H − H o )2 + κ G K where Ho is the spontaneous mean curvature, that is, the value when the membrane is relaxed; H and K are the mean and Gaussian curvatures, respectively; κ is the mean curvature (bending) modulus; and κG is the Gaussian curvature modulus. The total curvature energy of a lipid membrane is Gc = gc dA, where the integral extends over the whole membrane surface. κ measures the energy required to bend the membrane away from its spontaneous curvature Ho, and is always positive. For a bilayer, κ is twice the value for a lipid monolayer: κ b = 2κ m. Most common lipids have values of κ m in the region of 4 × 10−20 J (approximately 10 kT), and thus κb = 8 × 10−20 J (approximately 20 kT). Single-chain lipids have values some ivefold smaller, which means that thermally induced luctuations will be much more pronounced. The product κ bHo is proportional to the irst moment of the lateral pressure proile across the bilayer (Helfrich, 1981):

∫

∫

κ b H o = π ( z )zdz Because κ b is positive, this means that for a symmetric composition bilayer, H ob must be zero, but will be nonzero for an asymmetric bilayer. Note that most cell membranes are strongly asymmetric in their lipid composition, and will therefore have nonzero values of H ob . For a monolayer, H om can be either positive, zero, or negative, depending on how the lateral pressures are distributed across the layer. Thus, the monolayer may wish to bend, even when the bilayer does not. This means that a lat lipid bilayer will usually contain a curvature frustration energy equal to (from the Helfrich equation) ∆gc = 4κ m ( H om )2

37

Nonlamellar Lipid Aggregates

which is the energy required to force the two monolayers into a planar shape. The monolayer Gaussian modulus κ Gm is given (Helfrich, 1981) by the second moment of the lateral pressure proile, where the origin is set at the neutral surface, where the area does not change upon bending, which is a distance t from the ends of the chains: dm − t

κ Gm = −

∫

π ( z )z 2 dz

−t

The neutral surface tends to lie just below the oil–water interface, and κG is expected normally to be negative for a monolayer, with a value of approximately −0.8κ m (Templer et al., 1998). The value of κG for a bilayer is not simply twice the value for a monolayer, but is given (Porte et al., 1989; Helfrich and Rennschuh, 1990) by the more complex expression: κ Gb = 2(κ Gm − 2κ m H om t ) where t is the distance from the center of the bilayer to the neutral surface.* Thus, κ Gb can be positive even when κ Gm is negative, if the second term is suficiently positive, which can occur for H om negative. This tends to occur when the interfacial tension γ is balanced primarily by the chain lateral pressure πch rather than the headgroup lateral pressure πh (i.e., for small, weakly hydrated headgroups). A positive κG will try to force the membrane to develop negative Gaussian curvature, in order to lower the Helfrich curvature energy. Thus, κG controls the topological complexity of the membrane. A remarkable result from differential geometry is the Gauss–Bonnet theorem:

∫ KdA = 4π (1 − g) where the genus g describes the topology of any closed interface: Sphere: g = 0 Torus: g = 1 Double torus: g = 2 Thus

∫ κ

G

KdA = κ G 4π (1 − g )

and the integrated Gaussian contribution to the curvature elastic energy is constant for any closed shape of membrane, unless a topological change occurs. If a membrane has a positive κG, it can lower its curvature elastic energy by increasing its genus g, which introduces regions of negative Gaussian curvature in the surface. Such a change requires membrane fusion events to occur. For example, if a fusion channel between two bilayers is formed, with the mean curvature remaining close to zero, the change in curvature elastic energy is G fusion = −4πκ G If many fusion channels form between membranes, they will arrange themselves into 3D lattices of channels: this underlies the formation of the inverse bicontinuous cubic phases. We can describe *

This equation has sometimes been incorrectly quoted with a factor of 4 rather than 2 within the brackets.

38

Liposomes, Lipid Bilayers and Model Membranes

these structures as consisting of a single luid lipid bilayer draped onto different periodic minimal surfaces, G, D, and P, for the Ia3d, Pn3m, and Im3m cubic phases, respectively (see Section 2.9).

2.7 2D HEXAGONAL PHASES, TYPE I AND TYPE II The tendency for monolayer curvature increases with increasing temperature, as increasing conformational disorder is created in the hydrocarbon chains, and can lead to a phase transition from a lamellar bilayer phase, to an inverse hexagonal HII phase (Seddon, 1990a), where there is a negative interfacial curvature of the lipid monolayers (see Figure 2.8). This tends to relax the monolayer curvature frustration, but at the expense of introducing a packing frustration arising from the need for the chains to ill all of the hydrophobic volume, including the triangular “voids” at the center points of the hydrocarbon chain region. Modeling the chains as harmonic springs shows that this chain packing energy term becomes highly signiicant, and generally prevents the monolayers from being able to adopt their preferred curvature H om , unless alkanes are added to relieve the packing frustration.

2.8 CHAIN PACKING FRUSTRATION For cylindrical structures such as the inverse hexagonal HII phase, only a limited swelling can occur in water, due to packing constraints in the hydrocarbon chain region. To a limited extent, swelling can be enhanced by the addition of nonpolar solutes such as alkanes, which can relax such constraints. However, there is a risk that such solutes will destroy the ordered lattice. What will happen if the preferred monolayer mean curvature Ho has a value more negative than can be accommodated by the inverse hexagonal HII phase? The answer is that the long inverse cylinders break up into short or even spherical inverse micelles, which can then pack onto cubic or other 3D lattices. The chain packing frustration for a simple sc, bcc, or fcc packing of inverse spheres is very large (the “void” volume of an fcc phase would be 26% of the unit cell, compared to 9% for an HII phase), and such phases have not so far been observed experimentally. However, more complex structures have been observed (see Section 2.11), most commonly an inverse micellar cubic phase of spacegroup Fd3 m, which consists of two types of quasispherical

Inverse bicontinuous cubic phase (QII)

Inverse hexagonal phase (HII)

Inverse discontinuous micellar cubic phase (Fd3m)

Inverse micellar

D

Diamond (QII )

G

Gyroid (QII)

P

Primitive (QII)

Increasing curvature

FIGURE 2.8 (See color insert.) Schematic of some inverse nonlamellar mesophases that can be adopted by lipid systems. Phases are shown with curvature becoming more negative from left to right. (Reprinted from Int. J. Pharm., 395, Mulet, X. et al., High throughput preparation and characterisation of amphiphilic nanostructured nanoparticulate drug delivery vehicles, 290–297, Copyright (2010), with permission from Elsevier.)

Nonlamellar Lipid Aggregates

39

inverse micelle, of different diameters. Curiously, the “void” volume is 29%, worse than for an fcc packing of uniform spheres. However, it turns out that the variation in distance from the interface to the center points of the hydrophobic regions is much reduced, leading to a smaller chain-stretching free energy cost.

2.9

3D BICONTINUOUS CUBIC PHASES, TYPE I AND TYPE II

As for hexagonal phases, bicontinuous cubic phases may be normal (oil-in-water) or inverse (water-inoil). The vast majority of cubic phases observed for lipid systems are inverse bicontinuous cubic phases, with normal cubic phases commonly observed for block copolymer and surfactant systems. The inverse bicontinuous cubic phases consist of a single, continuous bilayer; the bilayer midplane is considered to be coincident with an ininite periodic minimal surface and subdivides space into two interpenetrating, but not connected, water networks. Three types of inverse cubic phase have been identiied in lipid systems (Figure 2.8). These are based on the Schwarz diamond (D), Schwarz primitive (P), and the Schoen gyroid (G) minimal surfaces, and are denoted as the QIID, QIIP, and QIIG phases, and have crystallographic spacegroups Pn3m, Im3m, and Ia3d, respectively (Luzzati et al., 1968). As a range of other minimal surfaces exist, it has been suggested that the adoption of cubic phases based only on the G, D, and P minimal surfaces relects a low level of Gaussian curvature inhomogeneity /2 (and hence a low level of packing frustration) for these surfaces (Schwarz and Gompper, 2000). Bicontinuous cubic phases have been observed for many different types of amphiphiles, including monoacylglycerides, glycolipids, urea and urea-like amphiphiles, and monoethanolamides (Kullkarni et al., 2011; Fong and Drummond, 2012). Cubic-like structures, known as cubic membranes, have also been observed by transmission electron microscopy in many eukaryote cell types under both physiological and pathological conditions, most frequently in the endoplasmic reticulum, but also in the plasma membrane, the nuclear envelope, and the Golgi complex (Almsherqi et al., 2009; Landh, 1995; Hyde et al., 1997). With increasing hydration, bicontinuous cubic phases typically adopt the phase sequence Q IIG –Q IID –Q IIP, since the order of compactness of the underlying minimal surfaces at the same values of curvature is G > D > P. While most lipids do not adopt all three cubic phases, the sequence appears to be universally followed within lipid–water phase diagrams. Usually, only the Q IID and Q IIP phases are observed at high water contents, although recently a Q IIG phase, which appears to be thermodynamically stable, has been observed for the synthetic branched chain glycopyranoside 2-hexyl-dexyl-β -d-glucopyranoside (β -Glc-OC10 C6) under excess water conditions (Brooks et al., 2011). If we consider the hypothetical phase sequence with increasing interfacial mean curvature (Figure 2.5), inverse cubic phases are expected to lie to lower water content than lamellar phases. However, inverse bicontinuous cubic phases are typically more hydrated than the corresponding lamellar phase. In this case, it has been suggested (Shearman et al., 2006) that water does not act to tune the curvature of the interface (which would tend to inhibit inverse phase formation). Rather, the water acts as inert “iller” for the aqueous channels of the cubic phase, permitting the already fully hydrated interface to adopt its preferred curvature.

2.10 SPONGE PHASE The sponge (L3) phase may be visualized as a kind of “melted” or translationally disordered bicontinuous cubic phase, with locally bicontinuous, highly lexible, curved lipid bilayers. It has been mainly observed for the lipid monoolein in the presence of additives and solvents, including polyethylene glycol and 2-methyl-2,4-pentanediol (Engstrom et al., 1998; Efrat et al., 2007; Imberg et al., 2003; Landh, 1994). Addition of the lipid diglycerol monooleate to monoolein has also been shown to induce sponge phase formation. Sponge phases may function as structural intermediates during the lamellar to cubic phase transition (Conn et al., 2006; Mulet et al., 2009).

40

Liposomes, Lipid Bilayers and Model Membranes

The sponge phase region in a phase diagram typically occurs over a narrow water content range, and is bordered by two-phase regions, which include a lamellar phase to lower water content and another luid phase to higher water content (Engstrom et al., 1998). Sponge phases were initially characterized as having two to three times larger aqueous pores compared to the bicontinuous cubic phase formed by the same lipid. However, recently, weakly hydrated sponge phases have been observed for pharmaceutical-grade glycerol monooleate (Angelova et al., 2011), and for its mixtures with monoolein (Angelov et al., 2011).

2.11 3D ORDERED MICELLAR PHASES, TYPE I AND TYPE II Ordered micellar phases (cubic or 3D hexagonal) consist of an ordered 3D packing of normal (type I) and inverse (type II) micellar aggregates. They are optically isotropic and highly viscous. The ive currently known type I ordered micellar phases are listed in Table 2.1. Type I ordered micellar phases have been mainly observed for saturated lysophospholipids, which have inherently high solubility in water, and gangliosides, which have a large effective headgroup size driving positive curvature (Shearman et al., 2010). The Pm3n phase (Figure 2.9a) is the most common type I micellar cubic phase and consists of a 3D packing of two quasispherical and six slightly asymmetric, disk-shaped micelles per unit cell. The less common discontinuous Fm3m and Im3m cubic phases (Figure 2.9b and c) consist of quasispherical micelles arranged onto fcc and bcc lattices, respectively. The Fd3m phase is, by far, the most commonly observed inverse micellar cubic phase. One unit cell consists of two different types of quasispherical micelle: eight larger hexakaidecahedra and 16 smaller dodecahedra (Figure 2.9d). A minimum of two different lipid components is usually

TABLE 2.1 Summary of the Known Nonlamellar Lipid Mesophases Phase

Symbols

Possible Spacegroups

Fluid isotropic

L1 , L2

Ordered micellar

II, III

Pm3n, Fm3m, Im3m, Fd3m, P63/mmc

Hexagonal

HI, HII

p6mm

Bicontinuous cubic

QI, QII

Ia3d, Pn3m, Im3m

Sponge

L3

Ribbon

RbI, RbII

c2mm, p2gg, cmm

Mesh

MhI, MhII

Bicontinuous intermediate

BcI, BcII

0 (random mesh), I4mm (tetragonal mesh), R3m (rhombohedral mesh) No spacegroups currently identiied

Brief Description Disordered luid of normal or inverse micelles Spherical or nonspherical, normal or inverse micelles packed onto cubic or 3D hexagonal lattices Long cylindrical micellar or inverse micellar aggregates packed onto a 2D lattice Lipid bilayer draped over a periodic minimal surface subdividing space into two water channels or vice versa A “melted cubic” or translationally disordered, locally bicontinuous phase Long lat micellar “ribbons” packed onto 2D lattices of oblique, rectangular, or hexagonal symmetry Based on the structure of the lamellar phase, but with the bilayer punctured by water-illed pores Based on the structure of the bicontinuous cubic phases, stretched along highsymmetry directions

Note: The subscript I or II refers to normal (type I) and inverse (type II) ordered phases, respectively. For disordered phases (micellar, inverse micellar, and sponge), the accepted nomenclature is L1, L2, and L3.

Nonlamellar Lipid Aggregates

41

FIGURE 2.9 The packing of the ordered micellar phases: (a) Pm3n, (b) Fm3m, (c) Im3m, (d) Fd3m, (e) P63/mmc. The irst three are type I structures. (With permission from Sakya, P. et al. 1997. Micellar cubic phases and their structural relationships: The nonionic surfactant system C12EO12/water. Langmuir, 13, 3706–3714.) The latter two are shown as type II, inverse micellar phases. (With permission from Shearman, G. C. et al. 2010. Ordered micellar and inverse micellar lyotropic phases. Liq. Cryst., 37, 679–694, Copyright (2010), Taylor & Francis.)

required to form the Fd3m phase, with partial segregation of the two lipid components into each micelle. It has been observed in several hydrated binary lipid mixtures, including monoolein/ oleic acid (Mariani et al., 1988), unsaturated diacylglycerol/phosphatidylcholine mixtures such as DOPC:DOG (Seddon, 1990b), and PC/fatty alcohol mixtures (Seddon and Templer, 1995). However, an inverse Fd3m phase has also been observed over a wide water content and temperature range for a number of purely binary glycolipid/water mixtures (Seddon et al., 1996). In this case, different conformations of the sugar ring may permit adoption of micelles of different curvature by a single lipid. The Fd3m phase may be biologically relevant, for example, functioning as a local watertight “patch” in a punctured membrane following lipolytic enzyme attack (Nieva et al., 1995). A 3D hexagonally packed inverse micellar phase of symmetry P63/mmc (Figure 2.9e) has been recently discovered in a dioleoyl phosphatidylcholine/dioleoyl glycerol system doped with a small amount of cholesterol (Shearman et al., 2009). The phase consists of an hcp packing of identical quasispherical inverse micelles. Its formation was a surprise because the chain packing frustration in this phase should be as large as that of an fcc packing of inverse spheres, which has never been observed. It is possible that the cholesterol is able to partially relieve the chain packing frustration in the hcp phase because of its relatively weak pinning to the polar headgroup region of the phase.

42

Liposomes, Lipid Bilayers and Model Membranes

A cubic phase of spacegroup P4332 has been identiied within a ternary monoolein/cytochrome c/water system (Mariani et al., 1988). The putative structure is based on the Ia3d bicontinuous cubic phase structure, but with one of the water channels replaced by a network of protein-encapsulated, quasispherical inverse micelles. A poorly ordered discontinuous micellar cubic phase with primitive symmetry, known as the cubic liquid (QL) phase, has been recently reported (Efrat et  al., 2007) in the monoolein– ethanol–water system. This phase is transparent, nonbirefringent, and characterized by having a low viscosity.

2.12 INTERMEDIATE PHASES Intermediate phases are divided into three main classes: ribbon, mesh, and noncubic bicontinuous phases (Holmes, 1998). As for bicontinuous cubic phases, they are typically observed in phase diagrams intermediate between the lamellar and hexagonal phases. They are differentiated from the bicontinuous cubic phases by being anisotropic (and therefore birefringent) and less viscous. All intermediate phases possess nonuniform interfacial curvature resulting from a competition between the preferred interfacial curvature and the molecular length. They are more common for systems with longer or more rigid alkyl chains. Ribbon and mesh phases have been observed in a variety of surfactant systems, including sodium dodecyl sulfate/water, but, to date, have not been observed for lipid systems. The proposed system of nomenclature for the intermediate phases denotes the ribbon, noncubic bicontinuous, and mesh phases as Rb, Bc, and Mh, respectively, with the subscript I or II relecting a normal or inverse phase (Holmes, 1998) The ribbon, mesh, and noncubic bicontinuous phases may be considered as distorted hexagonal, lamellar, and bicontinuous cubic phases. Ribbon phases (RbI, RbII) consist of long, lat “ribbons” of elliptical cross section packed onto 2D lattices of oblique, primitive rectangular, centered rectangular, or hexagonal symmetry. They are more commonly observed for type I than for type II systems. The stabilization of ribbon phases frequently requires a mixed surfactant system, with a variation in surfactant composition between the lat and the curved regions of the ribbon. However, an inverse ribbon phase has been recently observed for a single branchedchain polyoxyethylene surfactant, with the ribbon phase region stabilized by pressure (Shearman et al., 2011). Mesh phases are based on the structure of the lamellar phase, but with the bilayers punctured by water-illed defects. The simplest mesh phase to visualize is the random mesh MhI(0) where the defects are not correlated between the bilayers. Mesh phases with correlated defects have been observed with crystallographic spacegroups I4mm (tetragonal mesh) and R3m (rhombohedral mesh). Finally, bicontinuous intermediate phases (BcI, BcII) are based on the structure of the isotropic bicontinuous cubic phases, but with anisotropy introduced into the system by stretching in one of the three high-symmetry directions of the phase. Some controversy still exists over the existence of the bicontinuous intermediate phases, as their identiication is frequently ambiguous due to insuficient structural information (Schroder-Turk et al., 2011). The additional bending energy cost associated with tetragonal or rhombohedral periodic minimal surfaces compared to cubic minimal surfaces may make their formation unlikely.

2.13 CONTROL OF PHASE BEHAVIOR As described in Section 2.2, the chemical structure (and hence geometry) of the lipid itself determines the phase behavior. In practice, it is dificult to predict the phase adopted by a speciic lipid. A recent review of the literature (Fong and Drummond, 2012) has elucidated a set of molecular design guidelines to promote inverse phase formation. Environmental variables, including temperature, pressure, hydration, and type of solvent, also affect phase behavior. Their effect may be rationalized qualitatively by a consideration of curvature

Nonlamellar Lipid Aggregates

43

elastic energy. An increase in thermal disorder at increased temperatures will induce an increase in chain splay and hence monolayer Ho. This is consistent with the formation of inverse curved phases toward higher-temperature regions of phase diagrams. Conversely, on application of elevated pressures, denser chain packing results in a decrease in Ho and the formation of inverse-curved phases becomes less favorable. Nonlamellar phase formation has recently been demonstrated in alternative solvents, including protic ionic liquids, suitable for nonaqueous applications (Greaves and Drummond, 2008). Finally, incorporated additives, both within the lipid region (lipids, cholesterol, hydrophobic and amphiphilic drugs, peptides, and proteins) and within the solvent (polar solutes, ions, polymers such as polyethylene glycol, and hydrophilic drugs, peptides, and proteins) have been shown to affect the lipid phase behavior.

2.14

PHASE TRANSITIONS BETWEEN NONLAMELLAR PHASES

Transitions between nonlamellar phases, for example, cubic–hexagonal or intercubic transitions are luid–luid transitions involving a change of symmetry and/or topology. Such transitions may be effected by means of a change in a thermodynamic variable such as temperature or pressure. They may also be induced isothermally and isobarically by changes in water content or pH, or by the addition of an additive such as a salt or polymer. The thermodynamic driving force for the transition is proportional to the difference in chemical potential energy between the two phases. The enthalpy (and hence entropy) change is typically small for a luid–luid transition. Understanding the dynamics of lyotropic phase transitions is of practical interest for materials processing of amphiphilic materials (particularly for drug delivery and crystallization applications) and of relevance to fundamental biophysical processes such as endo- and exocytosis, fat digestion, and membrane budding. The intercubic transitions are perhaps the best-studied transitions between nonlamellar phases. All three bicontinuous cubic phases are interrelated by a mathematical transformation known as the Bonnet transformation, which isometrically maps the underlying minimal surfaces onto each other, so that all angles, distances, and areas on the surface are preserved. The Bonnet transformation predicts the ratio of lattice parameters of two cubic phases coexisting with excess water under equilibrium conditions. This has the value a(G)/a(D) = 1.576 and a(P)/a(D) = 1.279, and these ratios have been observed experimentally, including during a cubic–cubic phase transition. The Bonnet transformation is, however, considered to be physically unreasonable as the actual mechanism of transformation, because it requires self-intersection of surfaces during the transition. Alternative theoretical models for the cubic–cubic transition mechanism have been described both in terms of the “skeletal graphs,” which describe the water channels (Sadoc and Charvolin, 1989), and the minimal surfaces upon which the bilayer is draped (Fogden and Hyde, 1999). A recently developed model can be used to estimate the “cooperative unit” (the number of molecules that undergo the transition cooperatively, in a synchronized two-state manner) and the average mean curvature of the intermediate transition state based on experimental rate data (Squires et al., 2009). Experimental data on nonlamellar phase transitions are limited, mainly due to dificulties in obtaining time-resolved data on the timescale of the transition. However, the development of new synchrotron facilities and techniques has facilitated the monitoring of fast transitions such as intercubic transitions in real time (Brooks et al., 2011). For both temperature- and pressure-jump experiments, it has been observed that the transition rate depends on the difference between the inal temperature (or pressure) and the temperature (or pressure) at the phase boundary. Structural intermediates, including other cubic phases, have been detected during cubic–cubic transitions, but not in all cases. The transition from a cubic to a hexagonal phase involves a much greater change in topology than an intercubic transition and has been rather less studied, both experimentally and theoretically. The transition is generally slower than intercubic ones, and limited by the transport of water in or out of the cubic phase.

44

Liposomes, Lipid Bilayers and Model Membranes (a)

(c)

100 nm

(e)

100 nm 100 nm

(d)

(b)

50 nm

(f)

50 nm

20 nm

(g)

100 nm

(h)

50 nm

FIGURE 2.10 Representative cryo-TEM micrographs of nanoparticles of nonlamellar mesophases. (a–d): Cubosomes of GMO/F127/water (weight ratio: 1.88/0.12/98.0) viewed along the [001] (a and b) and [111] (c and d) directions. (e and f): Spongeosomes of DGMO/GDO/P80/water (weight ratio: 2.13/2.13/0.74/95.0). (g and h): Hexosomes of DGMO/GDO/F127/water (weight ratio: 2.25/2.25/0.5/95.0). Fourier’s transforms of magniied areas are shown in panels b, d, f, and h. (Reprinted with permission from Barauskas, J., Johnsson, M., and Tiberg, F. 2005. Self-assembled lipid superstructures: Beyond vesicles and liposomes. Nano Lett., 5, 1615–1619. Copyright 2005, American Chemical Society.)

2.15

DISPERSED NANOPARTICLES OF NONLAMELLAR LIPID AGGREGATES

Owing to their stability against dilution in excess water, type II nonlamellar lipid phases can be dispersed into submicron particles (Yaghmur and Glatter, 2009). The most common (Barauskas et al., 2005) are dispersions of the bicontinuous cubic and hexagonal phases, known as cubosomes (Figure 2.10a through d) and hexosomes (Figure 2.10g and h), respectively, although dispersed particles of sponge mesophase (spongeosomes) (Figure 2.10e and f) and micellar cubic phase (micellosomes) have been reported. Cubosomes and hexosomes are typically 200–300 nm in diameter and retain the symmetry of the bulk parent phase. The term “nanoparticles” is commonly employed to describe these particles, although some strict deinitions of this term limit it to particle sizes OH − > F − > HCOO − > PO34− > CH 3 COO − > Cl − > Br − > I − > NO3− > ClO 4− > SCN − And for cations: Ba 2 + > Ca 2 + > Mg2 + > Li + > Na + > K + > Rb + ≈ Cs+ > NH 4+ with Cl− and Na+ marking a borderline case for the anions and cations, respectively. The ions classiied on the left side of this borderline are considered as reducing the protein solubility and is called salting-out, kosmotropes, or water-structure-makers, while on the right side the ions are called salting-in, chaotropes, or water-structure-breakers. In the former case, the tendency of these ions to stay well-hydrated in solution favors the attractive interaction between the hydrophobic surface patches of proteins, which leads to the folding of the proteins and their precipitation. In the latter case, the ions on the right side increase the protein solubility, by interacting with the hydrophobic patches of the proteins, preventing them from folding and stabilizing them thus in solution. Their inluence on colloids and soft matter has been studied extensively; even inversions of the series have been observed for some systems (Leontidis 2002; Lo Nostro et al., 2004; Pinna et al., 2005; Schwierz et al., 2010; Vlachy et al., 2009). In this context, ion pairing was described by Collins to be a critical parameter inluencing the micellar properties (Collins, 1997). It was suggested that kosmotropes form ion pairs better with other kosmotropes and chaotropes pair with chaotropes, by preferring counterions with similar free energy of hydration. This ion-pairing tendency will contribute to the free energy of the system, as ion pairs are less hydrated than dissociated ions. Indeed, it was shown by Hedin that the micellar surface perturbs the hydration shell of bromide more than those of chloride ions (Hedin and Furo, 1999; Hedin et al., 2000). Ninham introduced another approach by claiming that better polarizable chaotropic ions have stronger dispersion interactions with interfaces than their kosmotropic counterparts (Bostrom et al., 2002). Furthermore, Leontidis stated that dispersion forces might not be the only reason for the Hofmeister series and that there are indications of anions acting through direct interactions with surfaces (Leontidis, 2002). In last decades, the role of interfacial hydration on aggregation parameters like size and shape has been discussed. In this context, we talk about “interfacial” water (in contrast to the “bulk” water) as an integral part of the aggregation structure. Therefore, the aggregate structure is sensitive to the balance of short-range interactions within the interfacial region between the hydrophobic effect and the free energy of hydration of counterions and head groups. Hence, sphere-to-rod transitions of amphiphiles depend on the amphiphile as well as on the type and concentration of the counterion. This transition occurs as a consequence of the dehydration of the interfacial region, which favors the formation of head group–counterion pairs as evidenced from chemical trapping experiments (Geng and Romsted, 2005; Romsted, 2007). More recently, the surfactant head groups were also considered as chaotropic or kosmotropic moieties, which formed the basis of interesting

Extractant Molecules as Hosts in Surfactant Monolayers or Bilayers

55

papers (Vlachy et al., 2008, 2009) with the aim to establish a Hofmeister series for surfactant head groups in order to explain and predict their interaction with a given type of counterion. For example, alkyl sulfate head groups were classiied as chaotropes while alkyl carboxylates were considered as kosmotropes. Up to now the inluence of anions onto the self-assembly of amphiphilic molecules was of central interest because of their more pronounced effects (Abezgauz et  al., 2010; Aswal and Goyal, 2002). The bulk cations, often ixed as Na+ or K+, which are also present due to the dissociation of the salt were not often considered in the studies, but they should not be fully neglected (Bhat et al., 2008). Clearly, ions at interfaces can inluence their molecular structure and despite all the attention solvent extraction has received, however, the molecular structure of the L–L interface remains quite elusive and, as a result, the mechanism of complex (ion + extractant molecules + counterions) formation and transfer that inluences the kinetics of extraction is not well understood. Although this chapter focuses on solvent extraction, other types of reactions at liquid–liquid interfaces with a great fundamental interest are considered because they also play a key role in many important chemical and biological systems (Volkov, 2001), such as phase-transfer catalysis (Fiamegos and Stalikas, 2005), micellar catalysis (Zhang et al., 2009), biomembrane activity, and enzymatic reactions on fat lipases (Reis et al., 2009). Systems in which such reactions take place are dificult to study because their global behavior is the result of reactant diffusion in bulk and interfacial mechanisms. To study speciically liquid interfaces, different experimental systems were developed: systems with a unique and controlled interface between two immiscible liquids such as interface of pendant drops (Jeribi et al., 2002), interface developed for relectance measurements (Zarbakhsh et al., 2009), interface created into a microchannel (Watanabe, 2009), nanoscopic interface supported by nanopipette (Li et al., 2009), and systems with interfaces stabilized by amphiphilic molecules such as vesicles (Mircheva et al., 2008), micelles, and microemulsions (Tondre et al., 2001). To tackle this highly challenging problem, it is important to understand the physical and chemical behavior of an extractant molecule in bulk but also at the water/oil interface at equilibrium and during an extraction process (“active interface”). In the following parts of this chapter, a few studies on the amphiphilic properties of extractant molecules in interaction with ions at the water/oil interfaces will be included. Since extractant molecules are weakly surface active, their interfacial behavior can be analyzed by incorporating them within a thermodynamically stable phase characterized by a high speciic oil:water interface like microemulsions, lyotropic lamellar phases, or in more classical micellar systems.

3.5 EXTRACTANT WITHIN A MONOLAYER Considering a four-component system with water or salt solution, oil, hydrophilic surfactant, and hydrophobic cosurfactant, a tetrahedron representation (Kahlweit et al., 1991; Reimer et al., 2003) of the phase diagram of such a system can be illustrated as in Figure 3.2. By adding cosurfactant stepwise to a surfactant–water–oil mixture, two phases are observed at irst. A surfactant-rich phase, identiied as an oil-in-water microemulsion, is then in equilibrium with an excess oil phase (Winsor I, 2). By increasing the cosurfactant content, the well-known three-phase body occurs. A microemulsion coexists then with excess water and oil (Winsor III, 3). At higher cosurfactant content, the system is driven into a two-phase region where a water-in-oil microemulsion and excess water phase coexists (Winsor II, 2). At suficiently high surfactant concentrations, a one-phase microemulsion can be obtained (Winsor IV, 1). By ixing the water-to-oil volume ratio at 1:1, the schematic phase boundaries in the resulting two-dimensional section of the tetrahedron have the shape of a ish. The head of the ish represents the 2–3–2 phase transitions, whereas the tail situated at high surfactant concentrations corresponds to the 2–1–2 phase transitions. Principally, the

56

Liposomes, Lipid Bilayers and Model Membranes Cosurfactant

– 2 1

3 Water

Surfactant 2

Oil

FIGURE 3.2 Schematic of the tetrahedron phase diagram of a quaternary water/surfactant/cosurfactant/oil system. A cut at the constant water/oil volume fraction γ = 0.5 shows phase boundaries describing the shape of a ish.

form of the ish head and the position of the critical point (transition 2–2) are of interest, since it is directly related to the solubilization power of the amphiphile. For classical ternary systems composed of oil, water, and polyethoxylated surfactants, temperature is used to tune the curvature of the surfactant ilm in order to determine the ish cuts (Solans and Kunieda, 1997) but in the pseudofour-phase diagram, the cosurfactant is used to tune the curvature of the surfactant ilm from direct to reverse systems. The aim of the recent C. Bauer work was to study the cosurfactant behavior of a highly hydrophobic extractant molecule, here TBP, in comparison to classical n-alcohol cosurfactants. In the nuclear industry, for instance, TBP is the benchmark extractant used in the so-called PUREX process (plutonium and uranium reining by extraction), where uranium and plutonium are selectively extracted from an acidic aqueous solution (Rao and Kolarik, 1996). TBP, like many other extractants in the solvent extraction process, is hydrophobic, nearly insoluble in water, and has an amphiphilic structure composed of a complexing polar part and one or more alkyl chain(s) as a nonpolar part. Hence, extractants have all the properties required to play the role of a cosurfactant (Zana, 1995). Incorporating extractants in microemulsions might permit to artiicially functionalize the interfacial ilm by a complexing group. An effective strategy was to mix this hydrophobic molecule with a highly hydrophilic surfactant C8G1 and to realize ish cuts with water and dodecane by varying the nature and the concentration of the cosurfactant: n-pentanol, n-hexanol, n-octanol, or TBP in order to compare them, as shown in Figure 3.3. Neither the surfactant nor the cosurfactant presents pronounced temperature dependence (Ruiz, 2008) as the cosurfactant-to-surfactant ratio is varied. The full “ish” diagrams are presented by Bauer (2011). For the n-alcohol series, by increasing the chain length, the ish is shifted both to (i) lower surfactant concentrations and (ii) lower cosurfactant-to-surfactant ratios. This can be interpreted respectively as (i) an increase in the solubilization eficiency (Solans and Kunieda, 1997), that is, less surfactant is needed to cosolubilize water and dodecane, and (ii) an increase in the hydrophobicity of the cosurfactant, that is, less cosurfactant is required to reverse the system from direct to reverse microemulsions. Such an evolution as a function of the alcohol chain length is usually observed in microemulsions (Solans and Kunieda, 1997). Also note that the head of the ish becomes thinner as the length of the n-alcohol increases. This is usually interpreted by an increase in the interfacial ilm rigidity (Andelman et al., 1987; Degennes and Taupin, 1982; Kegel and Lekkerkerker, 1993), which becomes too stiff to accommodate water and oil, that is, to form a soft bicontinuous structure usually observed in

Extractant Molecules as Hosts in Surfactant Monolayers or Bilayers

57

3

γ

2

1

0 0.05

0.10

0.15 WSURF

0.20

0.25

FIGURE 3.3 (See color insert.) Fish cuts for the different cosurfactants, pentanol (squares), hexanol (triangles), octanol (rhombi), TBP (arrows), and TBP with addition of Nd(NO3)3 (0.3 M) (stars) as a function of the cosurfactant/surfactant molar ratio on the y axis.

Winsor III microemulsions. When rigidity diverges, phase diagrams become temperature insensitive and phase limits are then imposed by local packing constraints (Zemb, 2009). In the case of TBP, the ish diagram is observed and is comparable to the one obtained with n-pentanol but only very slightly shifted to higher surfactant concentrations. Hence, TBP is identiied as an eficient cosurfactant to form microemulsions and is just a little less eficient in terms of solubilization power compared to n-pentanol. If a mole representation of the ish diagram is used, for example, by plotting γ (Figure 3.3), the ish determined for TBP is shifted far below the ishes corresponding to the n-alcohol series. Therefore, TBP is far more eficient in mole compared to n-alcohols, for example, three times more eficient in mole than n-pentanol, in changing the microemulsion curvature. Since TBP has three saturated hydrocarbon chains, this shows that it forms a mixed ilm with C8G1. The inluence of the presence of salt on the ish cut was also studied for TBP by using Nd(NO3)3 0.3 M instead of pure water, as shown in Figure 3.3. Neodymium salt was chosen because it is supposed to have the greatest effect due to (i) its three charges and (ii) its strong interaction with TBP forming a complex with Nd3+ (Sawada et al., 2008). A slight shift in the solubilization power to lower surfactant concentration is observed in the presence of Nd3+. Moreover, the phase transition appears slightly at higher cosurfactant-to-surfactant ratios. This is likely related to a salting-in effect, subsequent to ion adsorption at the TBP-functionalized interface. This leads to an increase of the hydrophilicity of organic molecules, resulting in increasing the ilm curvature toward water. Nevertheless, the addition of salt on the studied system affects only slightly the phase transitions. This is not surprising because salt addition is well known to have large effects on the Winsor-type system based on ionic surfactants and only slight effects on nonionic ones (Yamaguchi et  al., 1999). Hence, this observation is here conirmed with TBP, a nonionic extractant, as cosurfactant, and with a nonionic surfactant. Using TBP in the presence of Nd3+ does not affect much the phase transitions. A similar study was carried out recently by T. Hellweg et al. investigating the catalytic activity of a conined enzyme on a hydrophobic toxic nerve agent within a sugar-based surfactant microemulsion. This study reveals a high stability of the microemulsion in static and dynamic states during the hydrolysis with no signiicant effect on the monolayer structure and curvature (Wellert et al., 2011).

58

Liposomes, Lipid Bilayers and Model Membranes

However, there exist some examples where a cation can have a dramatic effect on interfaces and this is the case of the well-known Ca-induced membrane fusion mechanism. Ca2+ interacts preferentially with carboxylate or phosphate groups. The inherent mechanism is a partial dehydration of the polar heads of the amphiphilic species that can be counterbalanced by a reorganization of the assembly when fast molecular diffusion is permitted (Jackson and Chapman, 2008; Nir et al., 1979; Papahadjopoulos et al., 1990; Portis et al., 1979).

3.6

EXTRACTANT WITHIN A BILAYER

Lyotropic lamellar phases, periodical stacking of oily layers (made by the surfactant tails) diluted by water are also very interesting systems with multiple interfaces. In thermodynamic equilibrium, these phases are structurally and mechanically responsive to weak effects, such as the Hofmeister ion effect (Aroti et  al., 2007; Leontidis et  al., 2007; Petrache et  al., 2006) or the partitioning of host molecules within bilayers (Dhez et al., 2000; Ladbrooke et al., 1968; Mills et al., 2008; Tamai et al., 2008). In a recent work of Banc (Banc et al., 2010, 2011), the insertion of a lipophilic ioncomplexing agent, N1,N3-dimethyl-N1,N3-dibutyl-2-tetradecyl malonamide (DMDBTDMA), in a lamellar phase was stabilized by a nonionic surfactant: pentaethyleneglycol dodecyl ether (C12E5) was described and characterized. DMDBTDMA is like the DMDOHEMA, an extractant molecule used in the nuclear industry to separate minor actinides from high-level radioactive liquid wastes by liquid–liquid extraction processes (Nash and Lumetta, 2010) (DIAMEX process) or in more conventional hydrometallurgy for the rare earth element recycling (Estler et  al., 2003; Tian and Hughes, 1994). The structural parameters of a lamellar system enabled to describe the partition of the extractant molecules between the core and the interface of bilayers. This simple approach was based on the evolution of interfacial area with the addition of extractants, which was measurable due to the multitude of interfaces in lamellar phases, as for microemulsion systems described earlier (Tchakalova et al., 2008a,b). To summarize, the effect of ion complexation (the neodynium in this case) with the extractant molecules was analyzed in order to follow the change in the amphiphilic property of the extractant. First, the system was characterized using SAXS and transmission electron microscopy (TEM) coupled with freeze fracture to follow structural parameters, and polarized attenuated total relectance– Fourier transform infrared (ATR–FTIR) spectroscopy to measure the orientation of dipoles and to quantify the ion complexation phenomenon. Figure 3.4a shows the evolution of bilayer thickness (δHC) with the percentage of Nd(NO3)3 in the aqueous solution from the dilution law for the C12E5/DMDBTDMA (80/20 molar) as well as similar evolution for the reference system without the extractant molecules. An effect is clearly identiied that the extractant molecules is more or less equally partitioned between the surface and inside (surrounded by the aliphatic chain of the nonionic surfactants) of the bilayers (Figure 3.4b) (Banc et al., 2010), regardless of the extractant concentration within the lamellar phase. However, when a complexing cation is added, this partitioning is varying with an increase of the extractant molecules at the interface and then surrounded by the polar head of the surfactants. Then, a thermodynamic approach of liquid–liquid extraction was developed on the basis of the experimental data by considering the ion extraction process as the sum of three equilibria: 1. Positioning of free extractant polar head areas at the interfaces 2. Complexation of cations with extractants at the interfaces 3. Burying of bonded extractants within the bilayers Hence, by determining the extractant distribution within the bilayers and taken into account using the spectroscopic technique as shown in Figure 3.5a, the relative proportion of complexed and noncomplexed extractant (Figure 3.5b), the free energy contribution of each step deined earlier

59

Extractant Molecules as Hosts in Surfactant Monolayers or Bilayers (a)

20

C12E5 C12E5/DMDBTDMA

(b) 28 κ=0

26

19

24 δ HC (Å)

δ HC (Å)

18 17

22 20 18

16

+Nd3+

κ=1

16 15 0

20

40 60 % Nd(NO3)3

80

14

100

0

5

10

15

20

25

30

35

% DMDBTDMA

FIGURE 3.4 (a) Evolution of bilayer thickness (δHC) with the percentage of Nd(NO3)3 in the aqueous solution. Empty dots display the reference system (C12E5) whereas full dots display the mixed system C12E5/ DMDBTDMA (80/20 molar). (b) Evolution of bilayer thickness as function of the percentage of DMDBTDMA in the C12E5/DMDBTDMA mixture. The dotted line (κ = 0) displays the case for which extractants are fully embedded within bilayers, whereas the dashed line (κ = 1) represents the case for which extractants fully participate in the interfacial area as the surfactant molecules. Full dots are experimental points obtained for a system hydrated with 1 M LiNO3. The arrow indicates the decrease of the bilayers thickness by exchanging lithium with neodymium cations at constant DMDBTDMA percentage. (From Banc, A. et al. 2011. Journal of Physical Chemistry B 115(6):1376–1384. With permission.)

was estimated and compared (see Figure 3.6). This approach elucidates the possible mechanism of extraction by DMDBTDMA in this model system. As observed for membrane fusion, ion effect seems to have irst a “pulling” effect on the extractant or ligand, so then the aliphatic chain can contact with water with a local dehydration. Then, depending on the physical properties of the interfacial assembly, this effect induces a luctuation

Intensity (a.u.)

4×104 3×104 2×104 1×104

0% 10% 20% 30% 40% 50% 60% 70% 80% 100%

0 1700 1680 1660 1640 1620 1600 1580 1560 1540 Wavenumber (cm–1)

(b) 0.35 0.3 Fraction of bonded C — —O

(a) 5×104

0.25 0.2 0.15 0.1 0.05 0

0

20

40 60 % Nd(NO3)3

80

100

FIGURE 3.5 (a) Evolution of the carbonyl stretching vibration band with the percentage of neodynium salts in the aqueous solution. (b) Fraction of bonded carbonyl versus the percentage of neodymium. (From Banc, A. et al. 2011. Journal of Physical Chemistry B 115(6):1376–1384. With permission.)

60

Liposomes, Lipid Bilayers and Model Membranes

Energy

Free extractant in bulk

Free extractant at interface

ΔG1

ΔG2

Bonded extractant in bulk

ΔGTot

ΔG3 Bonded extractant at interface Extraction pathway

FIGURE 3.6 Energy diagram of the extraction reaction divided into four elementary equilibrium: (1) Free extractant within the bilayers, (2) free extractants with their polar head at the bilayers interfaces, (3) extractants complexed with cations at interfaces, and (4) complexed extractant within the bilayers. (From Banc, A. et al. 2011. Journal of Physical Chemistry B 115(6):1376–1384. With permission.)

in the packing or cluster formation and local densiication. On the other hand, if the complexation (ion pairing) is optimal, a “pushing” effect with a transfer of the neutral species back to the organic phase can be observed. Similar studies were carried out by mixing a hydrophilic cationic surfactant, trimethyltetradecylammonium hydroxide (TTAOH) with an anionic and the well-known acidic organophosphorus extractant for rare earths, the di(2-ethylhexyl) phosphate (HDEHP) (Yuan et al., 2008). The mixed catanionic system is an excellent combination since HDEHP is a monoacid with a pKa of 3.6 and cannot be dissolved alone in pure water. However, in the presence of TTAOH, HDEHP is soluble in water due to an acid–base reaction. A rich phase diagram was obtained varying the concentration ratio in water. Multilamellar vesicles and cylinders and tubes joining with vesicles were observed using cryo-TEM. However, in this work, neither the information on geometric packing effects that should be important nor the charge density of the headgroups, which should tune speciic interactions between headgroups and affect the long-range repulsive force that controls the swelling of the phases, is shown (Silva et al., 2010). These reports show the complexity of the amphiphilic character of this type of molecules that nevertheless have to be understood for developing kinetic models for ion extraction or separation at a water/oil interface, free or structured. Ultimately, such µEs and lamellar liquid crystal phases can be considered as model systems to indirectly study ion adsorption and complexation at the water/oil interface using an extractant molecule. These interfaces do not exactly mimic the active interface of the emulsion that is involved in the L–L extraction process since it is covered with the surfactant. However, these model phases permit to stabilize the ligand close to the interface in order to study their amphiphilic properties, which is indispensable to understand the kinetics of extraction.

3.7 PARTITIONING The preceding examples were referring to lipophilic extractant but sometimes. However on the extractant molecule conformation, interesting distribution between aqueous and organic media can be observed. This is the case of a biological and very eficient surfactant, surfactin that can complex metal cation. This is a natural cyclic lipopeptide (Figure 3.7a) produced

Extractant Molecules as Hosts in Surfactant Monolayers or Bilayers

61

(b)

(a) O

O

HO

N NH H HN H

O

O O

H N

O

O

NH

H N

H N

O OH

O O O

FIGURE 3.7 (See color insert.) (a) Structure of the lipoheptapeptide surfactin (main isomer) produced by Bacillus subtilis. (b) Schematic representation of coexisting surfactin reverse spherical aggregates in alkanerich phase in equilibrium with direct elongated micelles in aqueous phase.

by Bacillus subtilis strains (Arima et  al., 1968). It is known that surfactin interacts strongly with phospholipid monolayers and destabilizes phospholipid bilayers by a deep insertion of the peptide moiety. These effects are promoted by its cyclic form and by hydrophobic interactions between fatty chains and between lateral hydrophobic groups (Brasseur et  al., 2007; Eeman et al., 2006; Heerklotz and Seelig, 2001; Heerklotz et al., 2004; Kell et al., 2007; Shen et al., 2010a,b). Moreover, phospholipid vesicles are easily converted into globular mixed micelles in the presence of surfactin, indicating its strong detergency (Boettcher et  al., 2010; Kell et  al., 2007; Liu et al., 2010). The lipopeptide also readily adsorbs at soft interfaces and forms stable monomolecular ilms, adopting a ball-like structure like an amphipathic nanoparticle rather than a classical surfactant (Shen et al., 2009). Molecular modeling has shown that the peptidic backbone is very lexible and that surfactin tends to self-associate, forming clusters at the water/ hexane interface (Nicolas, 2003). Surfactin molecule contains two pH-sensitive carboxylate groups and thus solubilizes in aqueous solutions when becoming suficiently ionized, that is, at pH values above the pKa of 5.8 (Nicolas, 2003). Above the CMC, surfactin can self-assemble in small spherical micelles in water with low aggregation numbers (Shen et al., 2009). The small size is mainly attributed to strong electrostatic repulsions between charged head groups, imposing high interface curvature. In addition, the large surfactant head group imposes high packing constraints, leading to small aggregation numbers, as observed for other amphiphiles with bulky head groups (Auzely-Velty et  al., 2000). In micelles, surfactin molecules arrange in a core–shell structure. Negatively charged carboxylates are the hydrophilic polar head groups, whereas aliphatic chains (fatty chains and leucines lateral groups) constitute the micellar hydrophobic core (Shen et al., 2009). Compared to classical surfactants, surfactin presents an additional original property. Indeed, environmental conditions can affect the conformation of the peptidic sequence (Knoblich et  al., 1995; Osman et al., 1998). This modiies the interactions between polar head groups, and also indirectly the order between aliphatic chains. As a consequence, the peptide conformation inluences the self-assembly properties. For example, upon increasing the pH and ionic strength, globular micelles undergo a transition to larger aggregates such as ellipsoidal elongated micelles (Han et al., 2008; Ishigami et al., 1995). The transition is not sharp and two kinds of micelles may coexist, showing that surfactin is not a classical lexible surfactant (Figure 3.7b). The importance of peptide conformation has been evidenced regarding the aggregation and arrangement of the molecules (Vass et al., 2001). The conformational lexibility might be responsible for enabling the coexistence of bilayers and micelles in the same sample of surfactin, indicating that the shape parameter strongly depends on the conformation of the peptidic part of the molecule.

62

Liposomes, Lipid Bilayers and Model Membranes

Among various biological potential applications (Mills et al., 2008; Seydlova and Svobodova, 2008a,b; Singh and Cameotra, 2004), surfactin is also a good candidate for environmental applications like water and soil remediation owing to an eficient removal power of hydrocarbons (Lai et al., 2009) or heavy metals (Mulligan, 2005; Mulligan et al., 1999). Surfactin not only presents a rich and complex self-assembling pattern but also has an afinity for cations and oils. Therefore, it has recently been shown that surfactin could be a model to explore the role of self-organization in biphasic liquid–liquid extraction of metal cations (Dejugnat et al., 2011). The mixture of water/oil/ alcohol/surfactin, which is composed of complex luids, contains either direct or reverse aggregates in equilibrium in both subphases aqueous and organic, respectively (Figure 3.7b). As mentioned earlier, surfactin is an eficient agent for trace metal recovery. Indeed, aspartic and glutamic acid residues in surfactin are involved in the chelation of divalent metal cations (Hosono and Suzuki, 1983; Zou et al., 2010) and experiments using the lipoheptapeptide in liquid– liquid extraction-based metal separation, for example, in hydrometallurgy and in the ield of nuclear spent fuel recycling, have been carried out. We have therefore considered three model metal cations (Dejugnat et al., 2011): Cu2+, which is known to interact with peptides (in the classical biuret reaction, or involved in β-amyloid neurotoxicity (Bortolato et al., 1997)), Nd3+ as a nonradioactive model of trivalent americium and curium, and Fe3+ as a model of Pu4+ (John et al., 2001; Rozga and Bal, 2010). Competitive extraction experiments have been then conducted using equimolar amounts of surfactin and different metal cations that were added successively and by using different complementary techniques (scattering and spectroscopies) no extraction of Nd3+ was observed, whereas Fe3+ was extracted almost completely. An intermediate situation was observed with Cu2+ for which the extraction remains low (about 12%) but not negligible. It was shown that clearly the extraction eficiency does not depend simply on a cation’s charge but complexation mechanisms with surfactin were involved. Regarding the size of the aggregates, there is no modiication when the extraction is low (case of Nd3+ and Cu2+), whereas an increase in size of reverse micelles is observed in the case of strongly extracted Fe3+. The swelling observed in the case of iron could be related to the nucleation of reverse micelles induced by extracted analytes (water, ions) (Eicke, 1980; Eicke and Christen, 1978b). It appears that all the partially extracted Cu2+ is immediately released to the aqueous phase after being easily displaced by Fe3+. From these simple experiments, it is possible to establish a classiication of the cations, with respect to the strength of their interaction with surfactin. However, this exciting molecule may be used as a model in fundamental research due to its advantage of the “chameleon” amphiphilic behavior.

3.8 SUMMARY The analysis of phenomena that happen at L–L interfaces with extractant molecules and the fundamental studies on this speciic domain is dificult to perform because these interfaces are inherently buried, thin, and hard to probe, independently from the bulks. Moreover, owing to their nanometric extension in the normal direction and small quantity, the signal collected from these interfaces is relatively weak. This explains why rather few experimental reports directly probed speciically the interfacial domain and most of the research focused on indirect measurements of the L–L interfacial phenomena. Nevertheless, we have shown that many concepts used in surfactant science can be applied for extractant molecules to understand their amphiphilic behavior. Some thermodynamic model can be developed based on molecular geometric constraints in the extractant ilm to predict the formation of reverse micelles and the water solubility within the extractant reverse micelles. Electrostatic effects and ion polarizability also have to be evaluated to understand salt solubility in reverse extractant micelles or through a water/oil interface as well as the ion competition taking place in ion separation processes. Similar problems are encountered in interfacial catalysis or in ion transport for several cellular functions. Furthermore, the interface between water and oil-containing extractant molecules is far from being lat, and is inherently dynamic (Baaden et  al., 2001a,b).

Extractant Molecules as Hosts in Surfactant Monolayers or Bilayers

63

Beyond the studies on “static” planar and curved interfaces discussed in this chapter, the issue with liquid–liquid interfaces “in action” has to be investigated, that is, to follow the kinetics and spectroscopic signature of the multistep extraction processes. The results should serve as a basis for further understanding the extraction mechanism, to improve the eficiency and kinetics of existing processes for further development. The methodology, for example, “prediction” and interpretation of second harmonic generation signature (Martin-Gassin et al., 2011) will also allow us to study such liquid interfaces. Experiments as well as MD investigations have to focus on the distribution of solutes and solvent molecules with more attention devoted to the precise orientation of these species near the interface, in conjunction with the ion dehydration and transfer process and related spectroscopic signature.

REFERENCES Abecassis, B., Testard, F., Zemb, T., Berthon, L., and Madic, C. 2003. Effect of n-octanol on the structure at the supramolecular scale of concentrated dimethyldioctylhexylethoxymalonamide extractant solutions. Langmuir 19(17):6638–6644. Abezgauz, L., Kuperkar, K., Hassan, P.A., Ramon, O., Bahadur, P., and Danino, D. 2010. Effect of Hofmeister anions on micellization and micellar growth of the surfactant cetylpyridinium chloride. Journal of Colloid and Interface Science 342(1):83–92. Andelman, D., Cates, M.E., Roux, D., and Safran, S.A. 1987. Structure and phase-equilibria of microemulsions. Journal of Chemical Physics 87(12):7229–7241. Antonio, M.R., Chiarizia, R., Gannaz, B., Berthon, L., Zorz, N., Hill, C., and Cote, G. 2008. Aggregation in solvent extraction systems containing a malonamide, a dialkylphosphoric acid and their mixtures. Separation Science and Technology 43(9–10):2572–2605. Arima, K., Kakinuma, A., and Tamura, G. 1968. Surfactin a crystalline peptidelipid surfactant produced by Bacillus subtilis—Isolation characterization and its inhibition of ibrin clot formation. Biochemical and Biophysical Research Communications 31(3):488–494. Aroti, A., Leontidis, E., Dubois, M., and Zemb, T. 2007. Effects of monovalent anions of the Hofmeister series on DPPC lipid Bilayers part I: Swelling and in-plane equations of state. Biophysical Journal 93(5):1580–1590. Aswal, V.K. and Goyal, P.S. 2002. Role of counterion distribution on the structure of micelles in aqueous salt solutions: Small-angle neutron scattering study. Chemical Physics Letters 357(5–6):491–497. Auzely-Velty, R., Djedaini-Pilard, F., Desert, S., Perly, B., and Zemb, T. 2000. Micellization of hydrophobically modiied cyclodextrins. 1. Micellar structure. Langmuir 16(8):3727–3734. Baaden, M., Berny, F., and Wipff, G. 2001a. The chloroform TBP aqueous nitric acid interfacial system: A molecular dynamics investigation. Journal of Molecular Liquids 90(1–3):1–9. Baaden, M., Burgard, M., and Wipff, G. 2001b. TBP at the water-oil interface: The effect of TBP concentration and water acidity investigated by molecular dynamics simulations. Journal of Physical Chemistry B 105(45):11131–11141. Bachmann, R.T., Wiemken, D., Tengkiat, A.B., and Wilichowski, M. 2010. Feasibility study on the recovery of hexavalent chromium from a simulated electroplating efluent using Alamine 336 and reined palm oil. Separation and Puriication Technology 75(3):303–309. Banc, A., Bauduin, P., Desbat, B., Ly, I., and Diat, O. 2011. Ion extraction mechanism studied in a lyotropic lamellar phase. Journal of Physical Chemistry B 115(6):1376–1384. Banc, A., Bauduin, P., and Diat, O. 2010. Tracking an ion complexing agent within bilayers. Chemical Physics Letters 494(4–6):301–305. Barnard, K.R., Turner, N.L., and Shiers, D.W. 2010. LIX (R) 63 stability in the presence of Versatic 10 under proposed commercial extract and strip conditions, Part III: Effect of manganese and cobalt loading on oxime stability at 30 degrees C. Hydrometallurgy 104(2):268–277. Bauer, C. 2011. Metal ion extractant in microemulsions: Where solvent extraction and surfactant science meet. Chimie et Physicochimie des Matériaux: (Univ. Montpellier II, Montpellier). Bauer, C., Bauduin, P., Diat, O., and Zemb, T. 2011. Liquid interface functionalized by an ion extractant: The case of Winsor III microemulsions. Langmuir 27(5):1653–1661. Bauer, C., Bauduin, P., Dufreche, J.F., Zemb, T., and Diat, O. 2012. Liquid/liquid metal extraction: Phase diagram topology resulting from molecular interactions between extractant, ion, oil and water. European Physical Journal—Special Topics 213(1):225–241.

64

Liposomes, Lipid Bilayers and Model Membranes

Bhat, M.A., Dar, A.A., Amin, A., and Rather, G.M. 2008. Co- and counterion effect on the micellization characteristics of dodecylpyridinium chloride. Journal of Dispersion Science and Technology 29(4):514–520. Boettcher, C., Kell, H., Holzwarth, J.F., and Vater, J. 2010. Flexible loops of thread-like micelles are formed upon interaction of L-alpha-dimyristoyl-phosphatidylcholine with the biosurfactant surfactin as revealed by cryo-electron tomography. Biophysical Chemistry 149(1–2):22–27. Bortolato, M., Besson, F., and Roux, B. 1997. Inhibition of alkaline phosphatase by surfactin, a natural chelating lipopeptide from Bacillus subtilis. Biotechnology Letters 19(5):433–435. Bostrom, M., Williams, D.R.M., and Ninham, B.W. 2002. Ion speciicity of micelles explained by ionic dispersion forces. Langmuir 18(16):6010–6014. Brasseur, R., Braun, N., El Kirat, K., Deleu, M., Mingeot-Leclercq, M.P., and Dufrene, Y.F. 2007. The biologically important surfactin lipopeptide induces nanoripples in supported lipid bilayers. Langmuir 23(19):9769–9772. Brennan, L. and Owende, P. 2010. Biofuels from microalgae—A review of technologies for production, processing, and extractions of biofuels and co-products. Renewable and Sustainable Energy Reviews 14(2):557–577. Camel, V. 2002. Extraction techniques. Analytical and Bioanalytical Chemistry 372(1):39–40. Christen, H. and Eicke, H.F. 1974. Monte-Carlo model of micelle formation. Journal of Physical Chemistry 78(14):1423–1427. Collins, K.D. 1997. Charge density-dependent strength of hydration and biological structure. Biophysical Journal 72(1):65–76. Cote, G. 1998. Extraction liquide-liquide – Déinition du procédé – Réactifs industriels. Technique de l’Ingénieur J2762. Cotton, S. 2006. Lanthanide and Actinide Chemistry (Wiley, John and Sons Inc.), (West Sussex: England), DOI: 10.1002/0470010088. Danesi, P.R. and Chiarizia, R. 1980. The kinetics of metal solvent-extraction. CRC Critical Reviews in Analytical Chemistry 10(1):1–126. Degennes, P.G. and Taupin, C. 1982. Micro-emulsions and the lexibility of oil-water interfaces. Journal of Physical Chemistry 86(13):2294–2304. Dejugnat, C., Diat, O., and Zemb, T. 2011. Surfactin self-assembles into direct and reverse aggregates in equilibrium and performs selective metal cation extraction. ChemPhysChem 12(11):2138–2144. Dhez, O., Konig, S., Roux, D., Nallet, F., and Diat, O. 2000. Lamellar-to-nematic phase transition in a lipidsurfactant mixture. European Physical Journal E 3(4):377–388. Dourdain, S., Hofmeister, I., Pecheur, O., Dufr̂che, J.-F., Turgis, R., Leydier, A., Jestin, J., Testard, F., PelletRostaing, S., and Zemb, T. 2012. Synergism by coassembly at the origin of ion selectivity in liquid− liquid extraction. Langmuir 28:11319−11328. Eeman, M., Berquand, A., Dufrene, Y.F., Paquot, M., Dufour, S., and Deleu, M. 2006. Penetration of surfactin into phospholipid monolayers: Nanoscale interfacial organization. Langmuir 22(26):11337–11345. Eicke, H.-F. 1980. Surfactants in nonpolar solvents, aggregation and micellization. Micelles, 87, 85–145. Eicke, H.F. 1975. Structure and properties of surfactant associates in apolar solvents due to hydrophile-lipophile balance. Chimia 29(4):176–177. Eicke, H.F. and Christen, H. 1974a. Nucleation process of micelle formation in apolar solvents. Journal of Colloid and Interface Science 48(2):281–290. Eicke, H.F. and Christen, H. 1974b. Stability of micelles in apolar media. Journal of Colloid and Interface Science 46(3):417–436. Eicke, H.F. and Christen, H. 1978a. Is water critical to formation of micelles in apolar media. Helvetica Chimica Acta 61(6):2258–2263. Eicke, H.F. and Denss, A. 1978b. Deinition of a micelle revisited. Journal of Colloid and Interface Science 64(2):386–388. Estler, F., Eickerling, G., Herdtweck, E., and Anwander, R. 2003. Organo-rare-earth complexes supported by chelating diamide ligands. Organometallics 22(6):1212–1222. Fiamegos, Y.C. and Stalikas, C.D. 2005. Phase-transfer catalysis in analytical chemistry. Analytica Chimica Acta 550(1–2):1–12. Geng, Y. and Romsted, L.S. 2005. Origin of the sphere-to-rod transition in micellar solutions: Speciic ion hydration matters. Abstracts of Papers of the American Chemical Society 230:U2956–U2957. Grifin, W.C. 1949. Classiication of surface-active agents by. Journal of the Society of Cosmetic Chemists 1:311. Gupta, C.K. and Nagaiyar, K. 2004. Extractive Metallurgy of Rare Earths (CRC Press: Boca Raton, London, New York, Washington, D.C.). Guveli, D.E. 1984. The association of some novel cationic surfactants in benzene. Journal of Colloid and Interface Science 101(2):344–355.

Extractant Molecules as Hosts in Surfactant Monolayers or Bilayers

65

Han, Y.C., Huang, X., Cao, M.W., and Wang, Y.L. 2008. Micellization of surfactin and its effect on the aggregate conformation of amyloid beta(1–40). Journal of Physical Chemistry B 112(47):15195–15201. Hedin, N. and Furo, I. 1999. Fast diffusion of Br− ions on a micellar surface. Journal of Physical Chemistry B 103(44):9640–9644. Hedin, N., Furo, I., and Eriksson, P.O. 2000. Fast diffusion of the Cl- ion in the headgroup region of an oppositely charged micelle. A Cl-35 NMR spin relaxation study. Journal of Physical Chemistry B 104(35):8544–8547. Heerklotz, H. and Seelig, J. 2001. Detergent-like action of the antibiotic peptide surfactin on lipid membranes. Biophysical Journal 81(3):1547–1554. Heerklotz, H., Wieprecht, T., and Seelig, J. 2004. Membrane perturbation by the lipopeptide surfactin and detergents as studied by deuterium. Journal of Physical Chemistry B 108(15):4909–4915. Helfrich, W. 1978. Steric interaction of luid membranes in multilayer systems. Zeitschrift Fur Naturforschung Section A—A Journal of Physical Sciences 33(3):305–315. Hosono, K. and Suzuki, H. 1983. Acylpeptides, the inhibitors of cyclic adenosine-3′,5′-monophosphate phosphodiesterase. 3. Inhibition of cyclic-amp phosphodiesterase. Journal of Antibiotics 36(6):679–683. Hou, M.J. and Shah, D.O. 1987. Effects of the molecular-structure of the interface and continuous phase of solubilization of water in water oil microemulsions. Langmuir 3(6):1086–1096. Ishigami, Y., Osman, M., Nakahara, H., Sano, Y., Ishiguro, R., and Matsumoto, M. 1995. Signiicance of beta-sheet formation for micellization and surface-adsorption of surfactin. Colloids and Surfaces B— Biointerfaces 4(6):341–348. Israelachvili, J.N., Mitchell, D.J., and Ninham, B.W. 1976. Theory of self-assembly of hydrocarbon amphiphiles into micelles and bilayers. Journal of the Chemical Society—Faraday Transactions II 72:1525–1568. Jackson, M.B. and Chapman, E.R. 2008. The fusion pores of Ca2+-triggered exocytosis. Nature Structural and Molecular Biology 15(7):684–689. Jeribi, M., Almir-Assad, B., Langevin, D., Henaut, I., and Argillier, J.F. 2002. Adsorption kinetics of asphaltenes at liquid interfaces. Journal of Colloid and Interface Science 256(2):268–272. John, S.G., Ruggiero, C.E., Hersman, L.E., Tung, C.S., and Neu, M.P. 2001. Siderophore mediated plutonium accumulation by Microbacterium lavescens (JG-9). Environmental Science and Technology 35(14):2942–2948. Kahlweit, M., Strey, R., and Busse, G. 1991. Effect of alcohols on the phase-behavior of microemulsions. Journal of Physical Chemistry 95(13):5344–5352. Kalmykov, S.N. and Denecke, M.A. 2011. Actinide Nanoparticle Research (Springer-Verlag, Berlin, Heidelberg). Kegel, W.K. and Lekkerkerker, H.N.W. 1993. Phase-behavior of an ionic microemulsion system as a function of the cosurfactant chain-length. Colloids and Surfaces A—Physicochemical and Engineering Aspects 76:241–248. Kell, H., Holzwarth, J.F., Boettcher, C., Heenan, R.K., and Vater, J. 2007. Physicochemical studies of the interaction of the lipoheptapeptide surfactin with lipid bilayers of L-alpha-dimyristoyl phosphatidylcholine. Biophysical Chemistry 128(2–3):114–124. Kentish, S.E. and Stevens, G.W. 2001. Innovations in separations technology for the recycling and re-use of liquid waste streams. Chemical Engineering Journal 84(2):149–159. Kitahara, A. 1965. Solubility behavior of polyoxyethylene nonylphenol ethers in cyclohexane and effect of water by a light-scattering method. Journal of Physical Chemistry 69(8):2788–2791. Knoblich, A., Matsumoto, M., Ishiguro, R., Murata, K., Fujiyoshi, Y., Ishigami, Y., and Osman, M. 1995. Electron cryo-microscopic studies on micellar shape and size of surfactin, an anionic lipopeptide. Colloids and Surfaces B—Biointerfaces 5(1–2):43–48. Kubota, S., Ozaki, S., Onishi, J., Kano, K., and Shirai, O. 2009. Selectivity on ion transport across bilayer lipid membranes in the presence of gramicidin A. Analytical Sciences 25(2):189–193. Kumar, J.R., Kim, J.S., Lee, J.Y., and Yoon, H.S. 2010. Solvent extraction of uranium(VI) and separation of vanadium(V) from sulfate solutions using Alamine 336. Journal of Radioanalytical and Nuclear Chemistry 285(2):301–308. Kunz, W., Henle, J., and Ninham, B.W. 2004. ‘Zur Lehre von der Wirkung der Salze’ (about the science of the effect of salts): Franz Hofmeister’s historical papers. Current Opinion in Colloid and Interface Science 9(1–2):19–37. Kunz, W., Testard, F., and Zemb, T. 2009. Correspondence between curvature, packing parameter, and hydrophilic-lipophilic deviation scales around the phase-inversion temperature. Langmuir 25(1):112–115. Ladbrooke, B.D., Williams, R.M., and Chapman, D. 1968. Studies on lecithin-cholesterol-water interactions by differential scanning calorimetry and x-ray diffraction. Biochimica et Biophysica Acta 150(3):333–340.

66

Liposomes, Lipid Bilayers and Model Membranes

Lai, C.C., Huang, Y.C., Wei, Y.H., and Chang, J.S. 2009. Biosurfactant-enhanced removal of total petroleum hydrocarbons from contaminated soil. Journal of Hazardous Materials 167(1–3):609–614. Lamas, J.P., Sanchez-Prado, L., Garcia-Jares, C., and Llompart, M. 2010. Determination of fragrance allergens in indoor air by active sampling followed by ultrasound-assisted solvent extraction and gas chromatography-mass spectrometry. Journal of Chromatography A 1217(12):1882–1890. Lemaire, B., Bothorel, P., and Roux, D. 1983. Micellar interactions in water-in-oil microemulsions.1. Calculated interaction potential. Journal of Physical Chemistry 87(6):1023–1028. Leontidis, E. 2002. Hofmeister anion effects on surfactant self-assembly and the formation of mesoporous solids. Current Opinion in Colloid and Interface Science 7(1–2):81–91. Leontidis, E., Aroti, A., Belloni, L., Dubois, M., and Zemb, T. 2007. Effects of monovalent anions of the Hofmeister series on DPPC lipid bilayers Part II: Modeling the perpendicular and lateral equation-ofstate. Biophysical Journal 93(5):1591–1607. Li, Q., Xie, S.B., Liang, Z.W., Meng, X., Liu, S.J., Girault, H.H., and Shao, Y.H. 2009. Fast ion-transfer processes at nanoscopic liquid/liquid interfaces. Angewandte Chemie—International Edition 48(43):8010–8013. Liu, J., Zou, A.H., and Mu, B.Z. 2010. Toluidine blue: Aggregation properties and distribution behavior in surfactin micelle solution. Colloids and Surfaces B—Biointerfaces 75(2):496–500. Lo Nostro, P., Lo Nostro, A., Ninham, B.W., Pesavento, G., Fratoni, L., and Baglioni, P. 2004. Hofmeister speciic ion effects in two biological systems. Current Opinion in Colloid and Interface Science 9(1–2):97–101. Martin-Gassin, G., Gassin, P.M., Couston, L., Diat, O., Benichou, E., and Brevet, P.F. 2011. Second harmonic generation monitoring of nitric acid extraction by a monoamide at the water-dodecane interface. Physical Chemistry Chemical Physics 13(43):19580–19586. Meridiano, Y., Berthon, L., Crozes, X., Sorel, C., Dannus, P., Antonio, M.R., Chiarizia, R., and Zemb, T. 2009. Aggregation in organic solutions of malonamides: Consequences for water extraction. Solvent Extraction and Ion Exchange 27(5–6):607–637. Mills, T.T., Tristram-Nagle, S., Heberle, F.A., Morales, N.F., Zhao, J., Wu, J., Toombes, G.E.S., Nagle, J.F., and Feigenson, G.W. 2008. Liquid-liquid domains in bilayers detected by wide angle x-ray scattering. Biophysical Journal 95(2):682–690. Mircheva, K., Minkov, I., Ivanova, T., Panaiotov, I., Proust, J.E., and Verger, R. 2008. Comparative study of lipolysis by PLA(2) of DOPC substrates organized as monolayers, bilayer vesicles and nanocapsules. Colloids and Surfaces B—Biointerfaces 67(1):107–114. Mitchell, D.J. and Ninham, B.W. 1981. Micelles, vesicles and micro-emulsions. Journal of the Chemical Society—Faraday Transactions II 77:601–629. Mulligan, C.N. 2005. Environmental applications for biosurfactants. Environmental Pollution 133(2):183–198. Mulligan, C.N., Yong, R.N., Gibbs, B.F., James, S., and Bennett, H.P.J. 1999. Metal removal from contaminated soil and sediments by the biosurfactant surfactin. Environmental Science and Technology 33(21):3812–3820. Nash, K.L. and Lumetta, G.J. 2010. Advanced Separation Techniques for Nuclear Fuel Reprocessing and Radioactive Waste Treatment (Woodhead Puublishing Series in Energy n°2, Cambridge, United Kingdom). Nicolas, J.P. 2003. Molecular dynamics simulation of surfactin molecules at the water-hexane interface. Biophysical Journal 85(3):1377–1391. Nir, S., Bentz, J., and Portis, A.R. 1979. Promotion and inhibition of aggregation and fusion of acidic phospholipid vesicles by cations. Biophysical Journal 25(2):A261–A261. Osman, M., Hoiland, H., Holmsen, H., and Ishigami, Y. 1998. Tuning micelles of a bioactive heptapeptide biosurfactant via extrinsically induced conformational transition of surfactin assembly. Journal of Peptide Science 4(7):449–458. Osseo-Asare, K. 1991. Aggregation, reversed micelles and microemulsions in liquid-liquid extraction: The tri-n-butyl phosphate-diluent-water-electrolytre system. Advances in Colloid and Interface Science 37:123–173. Papahadjopoulos, D., Nir, S., and Duzgunes, N. 1990. Molecular mechanisms of calcium-induced membranefusion. Journal of Bioenergetics and Biomembranes 22(2):157–179. Petrache, H.I., Zemb, T., Belloni, L., and Parsegian, V.A. 2006. Salt screening and speciic ion adsorption determine neutral-lipid membrane interactions. Proceedings of the National Academy of Sciences of the United States of America 103(21):7982–7987. Pinna, M.C., Bauduin, P., Touraud, D., Monduzzi, M., Ninham, B.W., and Kunz, W. 2005. Hofmeister effects in biology: Effect of choline addition on the salt-induced super activity of horseradish peroxidase and its implication for salt resistance of plants. Journal of Physical Chemistry B 109(34):16511–16514.

Extractant Molecules as Hosts in Surfactant Monolayers or Bilayers

67

Portis, A., Newton, C., Pangborn, W., and Papahadjopoulos, D. 1979. Studies on the mechanism of membranefusion—Evidence for an inter-membrane Ca2+-phospholipid complex, synergism with Mg2+, and inhibition by spectrin. Biochemistry 18(5):780–790. Rao, P.R.V. and Kolarik, Z. 1996. A review of third phase formation in extraction of actinides by neutral organophosphorus extractants. Solvent Extraction and Ion Exchange 14(6):955–993. Reimer, J., Soderman, O., Sottmann, T., Kluge, K., and Strey, R. 2003. Microstructure of alkyl glucoside microemulsions: Control of curvature by interfacial composition. Langmuir 19(26):10692–10702. Reis, P., Holmberg, K., Watzke, H., Leser, M.E., and Miller, R. 2009. Lipases at interfaces: A review. Advances in Colloid and Interface Science 147–48:237–250. Romsted, L.S. 2007. Do amphiphile aggregate morphologies and interfacial compositions depend primarily on interfacial hydration and ion-speciic interactions? The evidence from chemical trapping. Langmuir 23(2):414–424. Rozga, M. and Bal, W. 2010. The Cu(II)/A beta/human serum albumin model of control mechanism for copperrelated amyloid neurotoxicity. Chemical Research in Toxicology 23(2):298–308. Ruiz, C.C. 2008. Sugar-Based Surfactant: Fundamentals and Applications. In Surfactant Sciences Series, vol 143 (CRC Press: Boca Raton, FL). Rydberg, J., Cox, M., Musikas, C., and Choppin, G.R. 2004. Solvent Extraction Principles and Practice (Marcel Dekker, New York). Salager, J.L., Marquez, N., Graciaa, A., and Lachaise, J. 2000. Partitioning of ethoxylated octylphenol surfactants in microemulsion-oil-water systems: Inluence of temperature and relation between partitioning coeficient and physicochemical formulation. Langmuir 16(13):5534–5539. Sawada, K., Hirabayashi, D., and Enokida, Y. 2008. Fundamental studies on extraction of actinides from spent fuels using liqueied gases—Conversion of copper into nitrate with NO2 and extraction of Nd(III) nitrate by CO2 with TBP. Progress in Nuclear Energy 50(2–6):483–486. Schwierz, N., Horinek, D., and Netz, R.R. 2010. Reversed anionic hofmeister series: The interplay of surface charge and surface polarity. Langmuir 26(10):7370–7379. Seaborg, G.T. 1980. Our heritage of elements. Metallurgical Transactions B 11B:5–19. Seydlova, G. and Svobodova, J. 2008a. Development of membrane lipids in the surfactin producer Bacillus subtilis. Folia Microbiologica 53(4):303–307. Seydlova, G. and Svobodova, J. 2008b. Review of surfactin chemical properties and the potential biomedical applications. Central European Journal of Medicine 3(2):123–133. Shen, H.H., Thomas, R.K., Chen, C.Y., Darton, R.C., Baker, S.C., and Penfold, J. 2009. Aggregation of the naturally occurring lipopeptide, surfactin, at interfaces and in solution: An unusual type of surfactant? Langmuir 25(7):4211–4218. Shen, H.H., Thomas, R.K., Penfold, J., and Fragneto, G. 2010a. Destruction and solubilization of supported phospholipid bilayers on silica by the biosurfactant surfactin. Langmuir 26(10):7334–7342. Shen, H.H., Thomas, R.K., and Taylor, P. 2010b. The location of the biosurfactant surfactin in phospholipid bilayers supported on silica using neutron relectometry. Langmuir 26(1):320–327. Shen, H.T. and Forssberg, E. 2003. An overview of recovery of metals from slags. Waste Management 23(10): 933–949. Shirai, O., Ozaki, S., Onishi, J., Kozai, N., Ohnuki, T., and Kano, K. 2009. Ion transport across a bilayer lipid membrane in the presence of hydrophobic ions. Chemistry Letters 38(11):1038–1039. Silva, B.F.B., Marques, E.F., Olsson, U., and Pons, R. 2010. Headgroup effects on the unusual lamellarlamellar coexistence and vesicle-to-micelle transition of salt-free catanionic amphiphiles. Langmuir 26(5):3058–3066. Singh, P. and Cameotra, S.S. 2004. Potential applications of microbial surfactants in biomedical sciences. Trends in Biotechnology 22(3):142–146. Solans, C. and Kunieda, H. 1997. How to prepare microemulsions: Temperature-insensitive microemulsions. In: Industrial Applications of Microemulsions, eds. C. Solans, H. Kunieda (Marcel Dekker, New York), Vol (Surfactant Science Series) 66, pp. 21–46. Steytler, D.C., Jenta, T.R., Robinson, B.H., Eastoe, J., and Heenan, R.K. 1996. Structure of reversed micelles formed by metal salts of bis(ethylhexyl) phosphoric acid. Langmuir 12(6):1483–1489. Tamai, N., Uemura, M., Goto, M., Matsuki, H., and Kaneshina, S. 2008. Lateral phase separation in cholesterol/ diheptadecanoylphosphatidylcholine binary bilayer membrane. Colloids and Surfaces B—Biointerfaces 65(2):213–219. Tanford, C. 1973. The Hydrophobic Effect: Formation of Micelles and Biological Membranes (John Wiley and Sons Inc., New York).

68

Liposomes, Lipid Bilayers and Model Membranes

Tchakalova, V., Testard, F., Wong, K., Parker, A., Benczedi, D., and Zemb, T. 2008a. Solubilization and interfacial curvature in microemulsions I. Interfacial expansion and co-extraction of oil. Colloids and Surfaces A—Physicochemical and Engineering Aspects 331(1–2):31–39. Tchakalova, V., Testard, F., Wong, K., Parker, A., Benczedi, D., and Zemb, T. 2008b. Solubilization and interfacial curvature in microemulsions II. Surfactant eficiency and PIT. Colloids and Surfaces A— Physicochemical and Engineering Aspects 331(1–2):40–47. Testard, F., Berthon, L., and Zemb, T. 2007. Liquid-liquid extraction: An adsorption isotherm at divided interface? Comptes Rendus Chimie 10(10–11):1034–1041. Tian, Q.Z. and Hughes, M.A. 1994. Synthesis and characterization of diamide extractants for the extraction of neodymium. Hydrometallurgy 36(1):79–94. Tondre, C., Hebrant, M., and Watarai, H. 2001. Rate of interfacial reactions compared to bulk reactions in liquid-liquid and micellar processes: An attempt to clarify a confusing situation. Journal of Colloid and Interface Science 243(1):1–10. Troxler, L., Baaden, M., Bohmer, V., and Wipff, G. 2000. Complexation of M(3+) lanthanide cations by calix[4] arene-CMPO ligands: A molecular dynamics study in methanol solution and at a water/chloroform interface. Supramolecular Chemistry 12(1):27–51. Valenzuela, F., Fonseca, C., Basualto, C., Correa, O., Tapia, C., and Sapag, J. 2005. Removal of copper ions from a waste mine water by a liquid emulsion membrane method. Minerals Engineering 18(1):33–40. Vass, E., Besson, F., Majer, Z., Volpon, L., and Hollosi, M. 2001. Ca2+-induced changes of surfactin conformation: A FTIR and circular dichroism study. Biochemical and Biophysical Research Communications 282(1):361–367. Vitolo, S., Petarca, L., and Bresci, B. 1999. Treatment of olive oil industry wastes. Bioresource Technology 67(2):129–137. Vlachy, N., Drechsler, M., Verbavatz, J.M., Touraud, D., and Kunz, W. 2008. Role of the surfactant headgroup on the counterion speciicity in the micelle-to-vesicle transition through salt addition. Journal of Colloid and Interface Science 319(2):542–548. Vlachy, N., Jagoda-Cwiklik, B., Vacha, R., Touraud, D., Jungwirth, P., and Kunz, W. 2009. Hofmeister series and speciic interactions of charged headgroups with aqueous ions. Advances in Colloid and Interface Science 146(1–2):42–47. Volkov, A.G. 2001. Liquid Interfaces in Chemical, Biological, and Pharmaceutical Applications (Marcel Dekker, New York). Watanabe, M. 2009. Formation of a water-xylene interface in a microchannel without sidewalls. Analytical Chemistry 81(19):8213–8218. Wellert, S., Tiersch, B., Koetz, J., Richardt, A., Lapp, A., Holderer, O., Gab, J. et al. 2011. The DFPase from Loligo vulgaris in sugar surfactant-based bicontinuous microemulsions: Structure, dynamics, and enzyme activity. European Biophysics Journal with Biophysics Letters 40(6):761–774. Yamaguchi, Y., Aoki, R., Azemar, N., Solans, C., and Kunieda, H. 1999. Phase behavior of cationic microemulsions near the tricritical point. Langmuir 15(22):7438–7445. Yuan, Z.W., Yin, Z.L., Sun, S.X., and Hao, J.C. 2008. Densely stacked multilamellar and oligovesicular vesicles, bilayer cylinders, and tubes joining with vesicles of a salt-free catanionic extractant and surfactant system. Journal of Physical Chemistry B 112(5):1414–1419. Zana, R. 1995. Aqueous surfactant-alcohol systems —A review. Advances in Colloid and Interface Science 57:1–64. Zarbakhsh, A., Webster, J.R.P., and Wojciechowski, K. 2009. Neutron relectivity study of alkylated azacrown ether at the air-liquid and the liquid-liquid interfaces. Langmuir 25(19):11569–11575. Zemb, T. 2009. Flexibility, persistence length and bicontinuous microstructures in microemulsions. Comptes Rendus Chimie 12(1–2):218–224. Zemb, T., Nave, S., Mandin, C., Martinet, L., Berthon, L., Testard, F., and Madic, C. 2004. Supramolecular organisation of tri-n-butyl phosphate in organic diluent on approaching third phase transition. Physical Chemistry Chemical Physics 6(4):799–808. Zhang, J., Meng, X.G., Zeng, X.C., and Yu, X.Q. 2009. Metallomicellar supramolecular systems and their applications in catalytic reactions. Coordination Chemistry Reviews 253(17–18):2166–2177. Zhang, Y.J. and Cremer, P.S. 2006. Interactions between macromolecules and ions: The Hofmeister series. Current Opinion in Chemical Biology 10(6):658–663. Zhu, Z.W. and Cheng, C.Y. 2011. Solvent extraction technology for the separation and puriication of niobium and tantalum: A review. Hydrometallurgy 107(1–2):1–12. Zou, A.H., Liu, J., Garamus, V.M., Zheng, K., Willumeit, R., and Mu, B.Z. 2010. Interaction between the natural lipopeptide [Glu(1), Asp(5)] surfactin-C15 and hemoglobin in aqueous solution. Biomacromolecules 11(3):593–599.

4

Molecular Dynamics of Lipid Bilayers Standards, Successes, and Works in Progress Edward Lyman and Sandeep Patel

CONTENTS 4.1 4.2

Introduction ............................................................................................................................ 69 A Selective History of Lipid Simulation................................................................................. 70 4.2.1 Early Days................................................................................................................... 70 4.2.2 New Millennium ......................................................................................................... 71 4.3 Membrane Simulations: The Present Day .............................................................................. 71 4.3.1 So You Want to Simulate a Membrane? ..................................................................... 72 4.3.2 Timescales and Lengthscales ..................................................................................... 72 4.3.3 Library of Lipid Models ............................................................................................. 74 4.3.3.1 Sterols .......................................................................................................... 74 4.3.3.2 Others ........................................................................................................... 75 4.3.4 State of the Art (and What Is to Come) ...................................................................... 76 4.3.4.1 Diffusion and Finite-Size Effects ................................................................ 76 4.3.4.2 Undulations and Curvature .......................................................................... 77 4.3.4.3 Lipid Mixtures and Immiscibility ............................................................... 78 4.3.4.4 Polarizable Force Fields ............................................................................... 79 4.3.4.5 Transfer of Charged and Polar Side Chains ................................................. 81 4.3.4.6 Helix Tilt ...................................................................................................... 81 4.3.4.7 G-Protein-Coupled Receptors...................................................................... 82 4.3.4.8 Membrane Curvature and Amphipathic Peptides ....................................... 83 4.3.5 Coarse-Grained Modeling .......................................................................................... 83 4.4 Summary ................................................................................................................................ 85 References ........................................................................................................................................ 85

4.1 INTRODUCTION Since its inception, the ield of lipid bilayer simulation has contributed a great deal to our understanding of the structure and dynamics of lipid bilayers, and by extension cell membranes. In many ways, the line between simulation and experiment has been blurry for many years, with experimental and simulation data often acting as two halves of a whole, rather than separate disciplines. The situation is somewhat analogous to the case of protein structure determination, where models and calculations are quite often essential to the determination of structures from NMR (nuclear magnetic resonance) or diffraction data. Bilayer simulations differ signiicantly from protein simulations in several key ways that we discuss in the next paragraph. We then 69

70

Liposomes, Lipid Bilayers and Model Membranes

review the historical development and current state of lipid bilayer simulations, with an emphasis on what is and is not possible with the current technology. Our goal is to provide a reader who is not an expert in molecular simulation with an understanding of how and where molecular simulation might be useful in one’s own research study. The chapter is by no means comprehensive, as we will limit the discussion to molecular dynamics of particle-based models of lipids and to bilayer geometries. For discussions on continuum approaches, the reader is referred to several recent review articles [1–3]. Other complementary publications include a book on multiscale coarse graining [4], and a host of review articles on lipid bilayer simulations dating back many years [5–23]. What is different about simulations of lipid bilayers compared to simulations of proteins or nucleic acids? A bilayer (if we, for the moment, limit ourselves to single-component bilayers) is composed of many copies of the same molecule, which presents a number of advantages. For instance, the average area per lipid is a sensible quantity, measureable by experiment and simulation alike. (Although trickier to measure in a simulation than it appears at irst glance, as discussed below.) Analogous quantities are dificult to identify for simulations of proteins, as the estimation of such observables by computer simulation is often confounded by the slow conformational dynamics of the protein. Indeed, the availability of quantities that are directly measureable both by simulation and experiment has contributed enormously to our understanding of lipid bilayers—experimental and computational study of bilayers is a two-way street. We will see that this theme emerged early in the history of bilayer simulations, and continues to this day. After all, the physics of bilayers and membranes contains interesting phenomena over an enormous range of length and timescales, from the isomerizations of the hydrocarbon chains to collective undulations. Thus, as advances in hardware and software push forward the accessible length- and timescales, new phenomena become accessible to computation. This theme runs through the history of lipid simulations, to which we turn next.

4.2 A SELECTIVE HISTORY OF LIPID SIMULATION 4.2.1

early days

The earliest all-atom (or nearly all-atom) simulations of hydrated lipid bilayers were published in the late 1980s and early 1990s. In 1988, Egberts and Berendsen published a simulation of a single-chain lipid bilayer, which included explicit water molecules, sodium ions, and partial atomic charges [24]. In 1991, Berkowitz and Raghavan published simulations of a more biological lipid (dilauroyl phosphatidylethanolamine) containing two hydrocarbon tails [25]. Again employing explicit water, these authors determined that the orientational motion of water at the interface was decoupled from the motion of the headgroups. This was a key result, conirming that the orientational polarization of the interfacial water was not a good order parameter for describing the hydration force between apposed lipid surfaces.* Starting in the early- to mid-1990s, with advances in computational power and lipid models, it became possible to simulate lipid bilayers in the luid (L α) phase for several hundred picoseconds. An early lesson that emerged from this study was the sensitivity of the interfacial structure to the treatment of long-range electrostatics [26,27] and the inclusion of a suficient hydration layer [28], lessons that must be heeded to this day [29,30]. By 1993, it was possible to simulate a hydrated lipid bilayer in the luid phase for hundreds of picoseconds [31–38]. With the timescale for isomerization of the hydrocarbon chains now accessible, Venable et al. showed that the observed fast decay of orientational correlations of CH bond vectors is due to the isomerization dynamics of the hydrocarbon chains, allowing comparison of simulation data with the carbon–deuterium order *

At the end of their paper, these authors caution that “. . . the data should not be taken literally . . . since the potential functions in the simulations are known only approximately, and the run times are still relatively short.” Words that, in some cases, still ring true today.

Molecular Dynamics of Lipid Bilayers

71

parameters obtained from the quadropolar spectra [37]. Also during this period, several authors recognized the value of a careful comparison between simulation data and experimental structural data as obtained by diffraction of lamellar stacks [39,40]. This approach admits quantitative validation of the simulation model parameters by comparison of several observables, including average atomic positions, area per lipid, and carbon–deuterium order parameters. These early comparisons set the stage for many years—it is now considered an essential validation of a new model to compute these same quantities.

4.2.2

new MIllennIuM

By the late 1990s and early 2000, it had become possible to simulate bilayers for longer timescales (tens of nanoseconds) and longer lengthscales (1000 lipids). The extension in timescale enabled the structural results of earlier simulations to be checked, and enabled computation of some dynamical observables for the irst time. Although, in this period, the number of publications becomes too large to comprehensively discuss, we choose to focus on a few key points which we feel represent critical progress. In 1995, Essman and Berkowitz computed rotational and lateral diffusion coeficients, noting the dificulties that attend the determination of a well-deined rotational axis for a molecule with many internal degrees of freedom [41]. This was an important advance, as it extended previous work investigating the orientational dynamics of CH bond vectors [37], but now on timescales that admitted observation of slower contributions arising from rotations about the lipid long axis. Indeed, disentangling the various contributions to relaxation as observed by NMR is a major contribution of bilayer simulations [7,42]. The quality of agreement between simulated and measured carbon–deuterium order parameters has since become a standard metric for the evaluation of lipid forceields [43–47]. Another case in which simulation offered a key insight into NMR relaxation is provided by the work of Feller, Huster, and Gawrisch, who used simulations to rationalize the complex relaxation phenomena observed by nuclear Overhauser enhancement spectroscopy [48]. A seminal paper during this period demonstrated that scattering density proiles could be computed directly from all-atom simulations, providing a stringent test of the models [43]. The analysis yielded speciic recommendations regarding where the models needed improvement, and has partly motivated the latest generation of improved models. Like the NMR order parameters and the area per lipid, the analysis of scattering density proiles has since become a standard benchmark for lipid models. Also during this period, the extension of lengthscales into the 1000 lipid regime enabled the computational investigation of membrane undulations and area compressibility [49,50]. An important outcome of this work was the recognition that a bilayer simulation should obtain the correct area under conditions of zero surface tension, a goal which challenged the force-ield development community for the following decade, but for which recent progress is promising [46]. Also around this time the irst simulations of cholesterol-containing bilayers were published [51–55]; we will see below that modeling multicomponent bilayers is an area of very active research, as it presents signiicant challenges to simulation.

4.3 MEMBRANE SIMULATIONS: THE PRESENT DAY Over the past decade, researchers have leveraged improvements in lipid models and computational power to study more complex problems, incorporating multicomponent membranes and membrane proteins. A huge number of papers have been published during the last decade, an exhaustive discussion is both impossible and probably not very useful. Instead, we will single out a few problems that we feel highlight areas of special importance. The selection is, of course, biased toward our own interests. The discussion is organized by asking what is needed to study a particular problem in membrane biophysics.

72

Liposomes, Lipid Bilayers and Model Membranes

4.3.1

so you want to sIMulate a MeMbrane?

Let us say that you have an interesting problem in membrane biophysics/chemistry, and that you would like to use molecular simulation as a tool to help unravel it. There are three questions that you have to ask before even starting; we list them next. We will use these questions to organize the rest of the chapter, answering them in the order that they are posed here. In answering these questions, we will provide an overview of the current state of the art in molecular simulation of membranes. 1. Are the timescale and lengthscale of my problem tractable in the computer? All-atom simulations of bilayers on conventional hardware are limited to timescales of roughly 1 µs and lengthscales of 20 nm. Modestly longer length and timescales may be achieved through the use of massively parallel computing resources and/or specialized hardware; much longer length and timescales may be achieved at the expense of chemical detail by employing coarse-grained models. 2. Are there well-validated models available for all of the pieces of my problem? The parameters used to model the chemistry of lipids, proteins, nucleic acids, and sugars are under constant development. Some are more reliable than others, some do not even yet exist. As we will see, even “simple” lipids are a challenge to parameterize. Embarking on projects that hinge on untested parameters is not to be taken lightly. 3. Is the important biophysics/chemistry captured by the current state of the art? This is clearly a broad question, and yet at the same time is problem speciic. It also clearly stands on the shoulders of the previous two questions. By reviewing the current state of the art, as well as obvious outstanding challenges, we hope that the reader will come away with a sense of what is and is not currently possible.

4.3.2

tIMescales and lengthscales

It is critical to understand the limitations of scales that apply to all-atom simulation of lipid bilayers. These limitations derive from details of both hardware and software, and are an essential consideration when determining the feasibility of a simulation project. In all widely used MD software, the integration of Newton’s equation of motion (and related stochastic differential equations) in the computer is accomplished by approximating the positions and velocities of all the atoms as time is advanced in discrete steps [56]. The accuracy to which the positions and velocities are approximated is determined by the choice of integration algorithm, the precision of the arithmetic, and the length of the discrete timestep. For a given algorithm, the maximum timestep length is determined by the fastest motion in the system, which for a biomolecular system is the vibration of covalently bound hydrogen. A good rule of thumb is that for a stable (approximately energy conserving) integration, the timestep must be no more than one-tenth this timescale, or about 1 fs. All commonly used MD software implements some methods to constrain these hydrogen bonds to their equilibrium lengths, which yields good structural and dynamic properties and allows safely doubling the timestep to 2 fs. The total duration of a simulation is then simply the product of the timestep and the number of integration steps that are computed. The computation of a single timestep demands evaluation of the force acting on each atom, and therefore the computational cost of a single timestep depends on the model that is chosen and the size of the system, with more accurate models and larger systems being more computationally expensive. The intermediate to long-range pairwise forces are the most expensive portion of the calculation, since the number of bonded force calculations scales linearly in the number of atoms. The long range pairwise forces typically contain two contributions: Electrostatic forces, and a 12–6 Lennard–Jones (L–J) potential to model dispersion and exchange. To eliminate boundary effects, the common practice is to employ periodic boundary conditions, choosing a system size large enough to minimize periodic artifacts. Naively, the cost of a single timestep therefore scales

Molecular Dynamics of Lipid Bilayers

73

as N2, where N is the number of atoms in the system. In practice, the computational complexity scales like N ln N. This is achieved by two algorithmic tricks. First, a cutoff is imposed on the L–J interactions, typically on the order of 10 Å. How the L–J potential is modiied at and beyond the cutoff is important—empirical potentials are parameterized for a particular cutoff scheme, and changing the cutoff procedure amounts to changing the forceield. The electrostatic interactions require a more sophisticated treatment. Since the Coulomb potential decays much more slowly than the L–J potential, simply cutting off the electrostatic interactions yields serious artifacts [26,28–30]. Direct computation of the periodic Coulomb interactions yields a different problem—the sum does not converge. The sum is rendered convergent by the Ewald method, in which the charges are screened by a ictitious cloud of compensating charge, the effects of which are then subtracted after the fact [57]. Modern MD software does not compute the full (N2) Ewald sum, but rather approximates the interactions by computing the long-range contributions on a discrete mesh by fast Fourier transform—the “Smooth Particle Mesh Ewald” method [58]. The parameters of the PME calculation are the size of the mesh and the order of the interpolation scheme used to compute forces in between mesh locations. Note that using a coarser mesh and lower order interpolation scheme reduces the cost of the calculation at the expense of accuracy, with potentially drastic results, including violation of the equipartition theorem and drastic errors in transport coeficients [29,30]. Apart from PME, of note is the recent publication of an alternate approach to approximating long-range electrostatics, the Gaussian split Ewald sum [59]. Which algorithm is optimal for a given problem will depend on details of the problem (especially system size), the hardware (parallelization and memory), and the level of required accuracy in the computed forces. For the interested reader, recent reviews can be found in Refs. [60,61]. Given a model and a system, the real (wallclock) time needed to compute a single timestep depends on the computational resource. The evaluation of the forces may be parallelized, reducing the wallclock time needed for each integration step by dividing the force calculation up among multiple processors. This scheme only works up to a point, since eventually the time spent computing forces is less than the time spent communicating the results of the separate calculations. Exactly where that point is depends on the computational resource and the code. However, a recent breakthrough is the introduction of “neutral territory methods,” which reduce communication overhead and enable signiicantly increased parallelism [62]. MD codes implementing neutral territory methods can achieve signiicantly higher throughput relative to codes from just a few years ago, a recent publication demonstrates simulation rates of 45 ns/day for a 65,000 atom bilayer/protein system on a cluster of 150 Intel Xeon E5630 processors communicating via Mellanox Ininiband at 40 GB/s [63]. Simulation rates up to 150 ns/day have been reported for similar sized systems by scaling up to 2048 processors [64]; however, achieving this performance demands utilizing few compute cores per ininiband port (on the order of 8), an expensive and therefore uncommon cluster architecture. Realistically, therefore, most groups can expect to simulate a membrane protein system for at most a few µs contiguously on commodity hardware. Other hardware and architectures promise potentially signiicant improvements in the near future, with many efforts to implement MD algorithms on graphics processing units making dramatic recent progress [65–68]. Still more dramatic simulation rates may be achieved on the Anton machine, a special purpose supercomputer built for molecular dynamics simulation [69]; simulation rates of tens of microseconds/day have been demonstrated [69]. Such hardware is, unfortunately, a very inite resource. The lengthscale of bilayer simulations is also resource-limited. Roughly speaking, simulations are parallelized by dividing the simulation volume into boxes, each of which is assigned to a different processor. For a ixed communication overhead, one can always simulate a larger system (more boxes) by adding more processors. The development of petascale computing resources has therefore pushed the size-limit envelope of particle-based simulations, beyond 1 million atoms. Leveraging such a resource for a bilayer system would translate into a 50 × 50 nm patch of bilayer. All-atom bilayer simulations on such a lengthscale would seem to be of limited value however, since the time needed to relax membrane height luctuations scales like the dimension in the bilayer plane to the

74

Liposomes, Lipid Bilayers and Model Membranes

fourth power [70]. Additionally, the relaxation of concentration luctuations in a multicomponent bilayer is governed by diffusion, and so the mixing timescale increases at least linearly with the area of the bilayer. (Both of these issues are discussed in more detail below.)

4.3.3

lIbrary oF lIpId Models

Early simulations of lipid bilayers focused on a handful of lipids selected on the basis of chemical simplicity and availability of experimental data for validation. There was an expectation that chemical simplicity is synonymous with simplicity of parameterization. Typical of this period were phospholipids, especially phosphatidyl choline and ethanolamine, with saturated hydrocarbon chains. Despite their apparent chemical simplicity, two or three generations of empirical force ield [45,46] have been required to achieve a satisfactory level of agreement with experimental observables for these lipids, which speaks to the subtlety of the intermolecular interactions that collectively determine the bulk properties of the bilayer. The current standard for quantitative comparison with experimental data is set by the recently published Charmm36 parameter set, which addresses several long-standing deiciencies of lipid models [46]. One can therefore have conidence in published models for these relatively “simple” phospholipids, with glycerol backbone, hydrocarbon chains of varying lengths, and up to one (centrally located) unsaturation. Several headgroups have been parameterized and validated, including phosphatidyl ethanolamine, -serine, -glycerol, and -choline. Other parameters have been developed based on other forceield development methodologies. These include the “Berger” lipids [44], which are a hybrid of the Gromos and OPLS forceields, and lipids parameterized to be compatible with the generalized Amber forceield [71,72]. The choice of forceield is motivated by the demands of the application—while Charmm36 is the best choice for most protein-free bilayer simulations, certain aspects of peptide or nucleic acid chemistry may well be better represented by one of the other forceields. Generally, it is a bad idea to mix parameters from different forceields, as such combinations have in most cases not been carefully tested. However, we emphatically stress that, when it comes to forceield parameters, details matter. As an example, it was shown that the description of dihedral rotations in the vicinity of a double bond in a hydrocarbon tail has a dramatic impact on the observed membrane structural properties [73]. The design of any simulation project must bear in mind all of these details. For the interested reader, a more thorough discussion of lipid forceield development can be found in Ref. [74] and elsewhere in this volume [75]. While lipid bilayers are interesting soft matter systems in their own right, the development of accurate and predictive phospholipid models is also an essential step along the way toward modeling cell membranes. It is both a proof of principle, demonstrating that classical ixed-charge models capture many critical properties of membranes at a quantitative level, and an important building block, since cell membranes contain signiicant fractions of phospholipid. On the other hand, the lipid fraction of cell membranes is spectacularly more complex, composed of a dizzying array of lipids and sterols. Motivated by this fact, many recent efforts have focused on expanding the library of membrane components available for biomolecular simulation. Here, we report on these efforts to provide the reader a panoramic view of the state of the ield, but we stress that—despite many years of effort by dedicated researchers—this work is in progress. One cannot assume that simply because several papers have been published employing a model of lipid X that the parameters for lipid X are well validated. After all, it has taken two or three decades to develop good parameters for “simple phospholipids.” This is in no way meant to detract from published work with preliminary models, as the process of lipid parameter development is essential although an incremental one, but as a word of caution to newcomers to the ield. 4.3.3.1 Sterols Cholesterol is central to the properties of cell membranes; accordingly, bilayer simulations incorporating cholesterol were published as early as 1999 [54]. A great deal of work has focused on

Molecular Dynamics of Lipid Bilayers

75

understanding precisely how cholesterol modulates the observable properties of bilayers, especially the “condensing” effect of cholesterol—titrating cholesterol into a pure saturated chain, phospholipid bilayer reduces the average area per lipid—and the closely related increase in chain order that is observed in cholesterol-containing bilayers. (For an excellent review of simulation work in this area, see Ref. [76].) While one can ind evidence to support a number of competing models for the organization of cholesterol [77–79]—superlattice, umbrella model, and so on—one interesting general conclusion to emerge is that the average tilt of the sterol ring is intimately related to the cholesterol ordering/condensing effects. Remarkably, it has been observed that deletion of the methyl groups on the β face of cholesterol reduces its ordering and condensation effects, apparently due to a decreased ability to align with the membrane normal [80]. This is surprising, since it was already known that the smoothness of the α face promotes tighter packing of saturated chains, and thus is at least partly responsible for the condensing effect [81,82]. Simulations have also demonstrated that the position of unsaturated bonds plays a crucial role in modulating the interaction of cholesterol and unsaturated lipids, with a monounsaturation at the center of the chain (where they are typically found) having the largest effect [73]. More recently, attention has focused on cholesterol incorporated into bilayers with a signiicant mol fraction of polyunsaturated fatty acids (PUFAs). Motivated by neutron scattering data which show evidence for cholesterol lying parallel to the membrane surface and at the bilayer core [83], such conigurations were then observed by CG simulation, along with an enhanced rate of cholesterol lip-lop [84]. So far, the results have not been reproduced by all-atom simulations. It should be mentioned, however, that the preference of cholesterol for saturated chains was already demonstrated some years before by simulation [85], and that the empirical force ields for both cholesterol [86] PUFAs are under active development [87]. Other sterol derivatives, including ergosterol, have been the subject of simulations for several years [76,88–90]. This work promises to continue an important line of inquiry, allowing comparison of the molecular mechanisms of these related but quite distinct sterols, in order to uncover the molecular interactions that determine the remarkable chemistry of cholesterol. 4.3.3.2 Others Here, we touch briely on other lipid species that are the subject of ongoing parameterization efforts. We stress that the ease and quality of parameterization is driven in part by the availability of good quality, unambiguous experimental data. For some of the phospholipids just discussed, pure bilayers may be prepared and studied by a number of techniques with relative ease. For many other lipid species this is not the case—they may not form bilayers, or they may resist traditional experimental techniques. Note also that bilayer mixtures challenge parameter validation, since the data are confounded by the possibility of nonideal mixing (more to come on this topic later), which would require disambiguating the contributions to the experimental observable from each phase. Apart from the surprising results regarding cholesterol just discussed, PUFAs are of course an essential component of the membranes of many cells. Given their special chemistry and dramatic effect on the bulk material properties of bilayers, we can expect that if they are present in a membrane, they are likely to play an important role. With an eye toward understanding their special role, a model for a PUFA with 1-stearoyl-2-docosahexanoyl (DHA) tails was published over a decade ago [91]. This model was then used to provide a membrane environment for rhodopsin [92], for which it is known that the membrane environment—in particular, ω -3 PUFAs—shifts the meta-I/ meta-II equilibrium [93]. The initial work demonstrated that rhodopsin has a marked preference for solvation by DHA; further work demonstrated signiicant structural changes in the protein [94]. A series of articles considered generic effects on protein function [95], the inluence of the number of unsaturations [96], the properties of cis versus trans unsaturation [97,98], and the preference of cholesterol for the saturated chain over the unsaturated chain [85]. PUFA models continue to be the subject of parameterization efforts [87], in recognition of the fact that subtle differences occur as a function of position along the chain.

76

Liposomes, Lipid Bilayers and Model Membranes

The ceramide sphingomyelin is a crucial component of the outer cell membrane in many cell types, believed to play a role in the lateral organization of the outer plasma membrane [99]. Published work has focused on sphingomyelin–cholesterol interactions, since these interactions are thought to drive the lateral structure of mixtures, including the outer cell membrane [100,101]. This is an important step, but we stress that the sphingomyelin models have not been subjected to rigorous validation like many of the lipids just discussed. Present in small amounts in native membranes (~1 mol%) but crucial in many signaling pathways, the inositols have been the subject of parameterization efforts recently [102,103]. Inositols present a dual challenge to parameterization: The inositol ring and all its variants, and a polyunsaturated tail. Once again, inositol models are available in the published literature, but have not yet been subjected to the same level of validation as their simpler cousins. Some published works that focus on interactions between the headgroup and peripheral membrane proteins [104,105] avoid the complication of the polyunsaturated tail by grafting the inositol headgroup onto oleoyl chains, creating a synthetic lipid that is sometimes employed in experiments as well, and avoiding the challenge of the polyunsaturated tail. It is expected, however, that the polyunsaturated tail is an important aspect of the PiP lipids, and it is therefore recommended that a model incorporating this aspect be used.

4.3.4

state oF the art (and what Is to coMe)

Having discussed length and timescale limitations and lipid models, we next turn our attention to aspects of membrane biophysics and chemistry that present interesting challenges to simulation. We irst consider bilayer only simulations, and then survey membrane protein simulations. 4.3.4.1 Diffusion and Finite-Size Effects A key advantage of molecular dynamics is that it, in principle, provides access to both thermodynamic and time-dependent quantities. In the latter category, the transport properties of the bilayer environment are of fundamental importance. They govern the spatio-temporal dynamics of the membrane, and therefore play a role in every process that involves the membrane. Accordingly, measurement of the diffusion coeficient has been the subject of simulation work for many years [41,106–109]. Many publications report diffusion coeficients, which are of course measureable by any number of experimental techniques. Generally, “agreement” is found between the simulated coeficients and their experimentally measured counterparts, though agreement is typically taken to mean roughly half an order of magnitude. But how are diffusion coeficients actually measured in simulation? And what might the limitations be? The mean-squared displacement (MSD) is the second moment of the time-dependent probability distribution of lipid displacements: r 2 (t ) =

∫ r P ( r , t ) dr 2

(4.1)

P(r,t) is the probability to ind a lipid at position r at time t given that it was located at r = 0 at time t = 0 [110]. In a simulation context, this ensemble average could be estimated by averaging this quantity over all the lipids in the system. In practice, it is common to improve the statistics by also averaging over time—for a given value of t, the trajectory of a particular lipid provides several pairs of points for which this quantity can be time-averaged, when t is less than the total trajectory length T. If the simulation is ergodic, then this is a theoretically sound practice. The mean-squared displacement determined in this way will then have at least two regimes: A ballistic regime at short times, and a diffusive regime at long times. The diffusion coeficient is then proportional to the slope of the MSD as a function of time during the diffusive regime.

Molecular Dynamics of Lipid Bilayers

77

Note, however, that the diffusion coeficient is strongly dependent on the system size for small systems, since the motion of neighboring lipids is correlated [106,109]. To escape its local “cage” formed by neighboring lipids, a concerted rearrangement is needed, which extends to next-nearest lipid neighbors—for small systems (on the order of 72 lipids per lealet), these correlations may span the entire system. An alternate approach therefore is to adopt a “jump” model. In the jump model, the motion of a lipid is considered to be a continuous time random walk with a variable step size [110], with the steps or jumps corresponding to the concerted rearrangements just mentioned. Diffusion coeficients determined in this way are also in agreement with experimental values [109], but also reveal a subtler inite size effect that persists to longer lengthscales. When evaluating the transport of lipids by either method, it is important to consider also the center of mass motion of the bilayer as a whole. Given a perfect integration scheme, this is a nonissue, since momentum would be perfectly conserved. In practice, however, there are no perfect integrators, as discussed in Section 4.3.2. The extent to which spurious center of mass motion contaminates a calculation of the MSD will depend on both the integration scheme and the system. However, it is common practice to remove the overall center of mass motion, with some authors arguing that this should be done for each bilayer separately [108]. Above, our discussion of lipid transport was based on the assumption that it is well described by a normal diffusion model. There is, however, increasing evidence that the dynamics of lipids is considerably more complex. Starting roughly 5 years ago, Falck et al. presented evidence for collective lows of lipids on the timescale of tens of nanoseconds [106]. They conclude that, rather than moving as “jumps,” lipid motion is better described as a collective or correlated phenomenon, with groups of lipids moving together. More recently, there is growing interest in the possibility for more exotic, “subdiffusive” transport, in which the MSD grows sublinearly in time. Flenner et al. [111] have shown that, in a pure bilayer, there is indeed a region of sublinear scaling that extends from the ballistic regime to a crossover to normal diffusion at about the 100 ns timescale. These authors provide a convincing explanation in terms of a mode-coupling approach, which indicates that there is no clear separation of timescales below the 100 ns timescale. Very recently, Jeon et al. [112] presented data for mixed bilayers containing cholesterol, and found evidence for cholesterol-dependent subdiffusion. At this point, things are far from settled, and there remains much work to be done. On the basis of the recently discussed results, it is clear, however, that one should be cautious when determining diffusion coeficients from the slope of an MSD plot. 4.3.4.2 Undulations and Curvature The bending rigidity of a bilayer is a key material property. Measureable by experiment and sensitive to composition, it offers an attractive avenue for validation of lipid models and for drawing connections between microscopic and bulk properties. The Canham–Helfrich-Hamiltonian [113,114] describes how the cost of a deformation depends on the material constants of the bilayer: The bending rigidity Kc, the spontaneous curvature c0, and the Gaussian curvature modulus Kg:

H =

∫

2 K  1 1 1   dS  c  + − 2c0  + K g  R1 R2    2  R1 R2 

(4.2)

The integral runs over the surface S, which is parameterized here by the local radii of curvature R1 and R2. To second order in the deviation h(r) from a lat membrane, the cost of deforming a (periodic) bilayer contains a term Kc[∇2 h(r)]2/2. Thus, Kc may be estimated from simulations by measuring the undulation spectrum of a lat bilayer [2,49,115,116]. In practice, however, simulations are restricted in lengthscale, cutting off the undulations and limiting the range over which they may be itted, and making the estimation of Kc by this method challenging for all-atom simulations. (Petascale resources enable longer lengthscales, but are still limited in timescale. Since the

78

Liposomes, Lipid Bilayers and Model Membranes

timescale over which membrane undulations relax scales like the 4th power of the wavelength, the microsecond bound on simulation timescales imposes an effective simulation lengthscale for membrane simulations, if undulations are of importance.) Note, however, that the height luctuation analysis is improved by direct Fourier analysis of the exact lipid positions, rather than analysis of a smoothed discretized surface that approximates the lipid positions [117]. Recently, however, an entirely distinct approach has been suggested by Watson et al. Their approach recognizes that Kc is also related to luctuations in the lipid orientations, when expressed in the correct gauge [118]. As these authors show, this insight enables good estimates for Kc to be obtained from relatively modestly sized simulations. Apart from measuring undulations, it is sometimes desirable to directly measure the curvature of a simulated lipid bilayer. For example, the area per lipid, a fundamental and apparently simple observable, depends on the actual surface area of the bilayer, and is therefore sensitive to the undulations of the membrane [119,120]. Other quantities (like estimation of carbon deuterium order parameters) may depend on the deinition of a local membrane normal, which also requires calculation of the local surface. This is not a trivial detail—different approaches will yield different answers. A precise, quantitative estimation of such quantities from simulation therefore demands consideration of these issues. Indeed, the lamellar stacks for which such quantities are measured experimentally are themselves not lat. Experimental measurement of area per lipid therefore is increasingly incorporating molecular dynamics simulation data as an integral part of the analysis [120–122]. The preceding paragraphs consider relatively modest curvature. The various membranes of the cell can however be very highly curved, with radii of curvature as small as 5 nm at the necks of budding vesicles. In the cell, such highly curved membranes are generated by specialized proteins, which we discuss below. Here, we focus on issues intrinsic to such highly curved bilayers. Let us say that you would like to simulate a highly curved bilayer. You might, for example, be interested in how the bilayer/water interface changes with curvature, as it has been demonstrated that some interfacial membrane proteins sense curvature by identifying lipid packing defects [123– 125]. A moment of relection reveals that building such a bilayer coniguration is challenging—it must be of small enough size to be tractable, and the balance of forces along the lateral direction must yield a stable coniguration, perhaps with additional stabilization provided at the interface by a curvature generating protein. In recent work, Cui et al. [126] recycled a curved coniguration that was generated by a curvature generating protein, removing the protein and stabilizing the curvature by imposing a ixed simulation volume. This approach is obviously not easily reproduced. Very recently, Sodt and Pastor have developed a much more generally useful approach in which a monolayer is curved into the desired radius, and then the interior is illed with hexane. By judicious choice of the quantity of hexane, the curvature forces are balanced, yielding a stable bilayer [127]. 4.3.4.3 Lipid Mixtures and Immiscibility Pure, bilayer-forming lipids already possess an interesting phase diagram, with several phases and phase transitions [128]. Biological membranes, on the other hand, are quite far from a pure system, comprising a dizzying variety of lipid species and proteins. It may be, however, that there are “simple” mixtures of lipids containing just a few components that display some of the key features that are essential to understand biologically realistic membranes. Given the complexity of pure lipid phase diagrams, it is not surprising that even relatively simple mixtures of lipids possess quite complex phase diagrams. In the following paragraphs, we focus on the regions in which the bilayer is luid, of interest to the study of biologically relevant mixtures, and consider the implications for simulation. The existence of coexisting liquid phases in simple mixtures of lipids has been known for many years [129], but is still the most remarkable feature of such systems, rich with fascinating consequences. As a function of composition and temperature, ternary systems containing cholesterol, a high melting temperature lipid, and a low melting temperature lipid can form two distinct liquid phases, which we here call “liquid ordered (Lo)” and “liquid disordered (L d).” Careful recent

Molecular Dynamics of Lipid Bilayers

79

experiments have mapped out the phase diagrams of several such systems [130–133], demonstrating a region of two-phase coexistence and a miscibility transition. Keller and coworkers have gone a step further, carefully measuring a line of miscibility critical points, demonstrating exponents consistent with the 2D Ising universality class [134,135]. Though much debate focuses on the relevance of the phase behavior of simple mixtures to live cell membranes, these phenomena are of interest in their own right, with key questions unanswered. For example, a deep understanding of how the macroscopic phase diagram emerges from the microscopic, molecular scale interactions is lacking. Molecular dynamics simulation would seem ideally suited for this problem, as one can perform many “controls” that are not available experimentally. However, the study of immiscible phases by conventional molecular dynamics is still beyond reach. The reason is simply that the mixing timescale is much longer than the microsecond timescale achievable on commodity resources. Over a microsecond, a lipid explores an area on the order of 1 nm2. Therefore, a (e.g., 100 nm2) simulation of a ternary mixture that is initiated from a random initial coniguration does not have time to phase separate. Note that the diffusive timescale represents a lower bound on the necessary simulation timescale—the dynamics of composition luctuations in the critical region are likely to be much slower, as the correlated motions typical of the critical region result in the well-known phenomenon of critical slowing down [136]. In the authors’ view, this is a critical limitation of lipid bilayer simulations. The same basic problem also applies to simulations of membrane proteins, in cases where the protein is expected to prefer to interact with certain lipid species. Several authors have developed advanced sampling methods to overcome this limitation. The irst to recognize the need for advanced sampling methods for mixing degrees of freedom were Scott and coworkers, who developed a hybrid Monte Carlo/ MD approach over a decade ago [137]. More recently, Tajkorshid and coworkers have developed an exciting approach based on replacing the membrane interior with a mimetic, in this way decoupling the headgroups and the tails [138]. But perhaps the most promising method so far is a statistically rigorous grand-canonical Monte Carlo approach developed by Kindt and coworkers [139–142]. Their method is suited to problems in which the compositional degrees of freedom are structurally similar, as it relies on “mutation” type Monte Carlo moves in which one lipid is converted into another. All of these methods are exciting developments, but there remains signiicant and important work to be done. Progress in this area is essential to the continued development of lipid simulations toward more realistic membrane models. 4.3.4.4 Polarizable Force Fields Molecular dynamics simulations of the kind discussed in this review necessarily rely on an approximation to the underlying quantum mechanical interactions. In the models discussed in Section 4.3.3, the electrostatic interaction is modeled by Coulomb interactions between static charges assigned to individual sites of molecules. Thus, there is no mechanism for capturing a dynamic electronic response to changes in the local chemical environment. In strongly anisotropic environments (e.g., interfacial systems like the membrane surface, the protein–solvent interface, or ion conduction from bulk solution through integral membrane protein channels), one can argue that at the atomistic level, a ixed-charge representation may not faithfully model the underlying physics. Currently, several approaches for irst-generation polarizable force ields for biological molecules (proteins, DNA/RNA, ions) are actively being pursued. In spirit, all approaches attempt to model the induced dipole (in the linear response limit) of atoms or molecules, µinduced = α E, within some systematic formalism. These include point dipole polarizable models [143–145], shell (Drude) models [146–149], charge equilibration (electronegativity equalization, chemical potential equalization) models [150–161], luctuating dipole [162], Charge on Spring (COS) models [163,164], and the Sum of Interactions Between Fragments ab initio computed (SIBFA) [165–168], models based on molecular polarizability as introduced by Thole, and recent atomic multipole methods such as the AMOEBA [169–171] force ield (implemented within the TINKER [172] modeling package). The point dipole

80

Liposomes, Lipid Bilayers and Model Membranes

polarizable models generally ascribe atomic polarizabilities to various sites within the molecular construct along with ixed charges, and self-consistently evaluate the induced dipole moments arising from the local electric ield generated by the nearby charge density and coniguration. Shell (Drude) models introduce inite-mass charged sites generally coupled to more polar nuclei via a harmonic spring with a force constant which intimately determines the local polarizability (the “atomic” polarizability in such models goes as the inverse of the force constant as one would intuitively presume); effectively, one models, locally, a system with two charges whose separation oscillates with time. Thole models introduce intra- and inter-molecular dipole–dipole interactions by assigning atomic polarizabilities determined from empirical its to experimental molecular polarizability tensors [173]. Thole introduced a damping function in order to attenuate the intra-molecular dipole–dipole interactions at short distances. The AMOEBA force ield introduces permanent atomic monopole, dipole, and quadrupole moments, and explicitly treats polarization by allowing mutual induction of dipoles at prescribed sites with contributions from both permanent multipoles and induced dipoles; in this respect, an iterative approach to self-consistency is required in order to determine the instantaneous moments. Permanent multipoles interact through a multipole interaction matrix. Atomic polarizabilities are it to reproduce experimental molecular polarizabilities in the spirit of Thole. These values are used in conjunction with a damping scheme introduced to attenuate the intra-molecular dipole–dipole interactions in order to avoid polarization catastrophes at small separations. Though each approach demonstrates unique advantages and shortcomings, it currently remains to be seen if any one approach is superior to the others, much as the current situation with ixed-charge empirical force ields for biomacromolecular modeling. We refer the reader to recent reviews on polarizable force ields for biomacromolecular simulations [174]. Though much work has been accomplished in the last decade, building upon the initial foundations based on small-molecule systems, much remains to be done with respect to the routine realization of application of polarizable force ields in large-scale molecular simulations. Membrane and lipid bilayer systems present natural systems for applications of polarizable force ields due to the inherent, strongly anisotropic environments encountered, particularly in processes associated with transfer of a wide variety of molecular species through the anisotropy. We discuss briely recent applications of polarizable force ields to such systems. 4.3.4.4.1 Membrane Monolayer Potential The membrane dipole potential plays an important role in the movement of molecular and ionic species across the water–lipid interface. A recent study using Drude’s oscillator models of the DPPC monolayer explored the surface dipole potential of a water–lipid monolayer system [175]. The authors demonstrated the improvement in the prediction of the relative interfacial potential, ΔV = Vmonolayer–air − Vwater–air, over ixed-charge nonpolarizable force-ield representation. The authors suggest that, compared to the bilayer dipole potential, the monolayer potential is a less ambiguous measurement for comparing force-ield predictions to experiment. The estimated relative dipole potential was 0.35 V in agreement with the range of experimental values from 0.30 to 0.45 V. In separate work [176], a charge equilibration model was also used to explore the same quantity, the authors using their model to predict a relative interfacial potential of 0.64 V, slightly higher than the experimental range, but lower than the value of 0.8 V predicted by the nonpolarizable standard CHARMM27 force ield. 4.3.4.4.2 Potassium Permeation Free Energetics in Gramicidin A Owing to its narrow pore, lined with polar carbonyl groups, the Gramicidin A (gA) channel has attracted attention as a model system for testing of modern force ields [177]. Experimentally consistent potentials of mean force through this simple channel were dificult to estimate via MD simulations and free energy methods. There was conjecture regarding the suitability of nonpolarizable force ields for such systems. In response, a study to compute the one-dimensional equilibrium potential of mean force for potassium permeation through gA using a charge equilibration force ield was discussed [178].

Molecular Dynamics of Lipid Bilayers

81

The resulting potential of mean force showed a dramatic decrease in the central barrier to ion permeation, being almost one-half that predicted by standard force ields. The global minimum at 9.5 Å agrees with solid state 15N NMR chemical shift anisotropy experiments of Tian et al. [179]. Moreover, the site at 7.5 Å is seen to be of low free energy (almost commensurate in stability to the global binding site) but separated by a signiicant free energy barrier of 5 kcal/mol. This further coincides with the NMR measurements [179] suggesting an internal binding site of signiicantly reduced signal relative to the external binding sites (conjectured to be of equal free energetics). 4.3.4.4.3 Protein- and Peptide-Bilayer Systems In this section, we discuss the applications of molecular dynamics to protein/bilayer systems, again emphasizing what is and is not possible with current technology. Membrane proteins—especially integral membrane proteins—challenge experimental efforts aimed at structure and functional characterization. This presents an opportunity for simulations, provided the researcher is aware of the limitations and calibrates her expectations accordingly. The review is by no means exhaustive—the goal is rather to give the reader a sense of the limitations of membrane protein simulation, and to touch on a few areas that are presently at the forefront of the simulation ield. The reader is referred to other recent review articles for a broader view [9,12–15,19,20]. 4.3.4.5 Transfer of Charged and Polar Side Chains Recent years have witnessed great interest in understanding the molecular origins of the presence of charged and polar amino acid residues in ostensibly hydrophobic lipid bilayer environments [180– 192]. The motivation for such a microscopic understanding stems from the broad range of biophysical processes predicated on the interactions between such protein residues and hydrophobic lipid chains. These processes range from voltage gating in select ion channels [193–195], permeation of cationic residue enriched cell-penetrating peptides for transporting cargo across the cellular membrane [183,196–199], and the action of antimicrobial peptides upon interaction with native cellular membranes. Understanding these protein–lipid interactions has sought recourse in hydrophobicity scales quantifying relative partitioning propensities of different amino acid side chains from aqueous to bilayer-like environments [200]. Elaborating upon ideas of partitioning of functional chemical groups between hydrophilic and hydrophobic environments, recent work has broadened the palette of hydrophobicity scales attempting to address the relative free energetics of partitioning; this has been possible due to novel experiments on well-characterized integral-membrane protein systems [188] as well as elucidation of structural aspects of the machinery implicated in the synthesis and insertion of membrane proteins upon synthesis in the ribosome [201]. Further factors possibly contributing to interactions of charged and polar species in lipid bilayers include bilayer thickness, nonadditivity of interactions between nonbilayer components [200,202], speciicity of protein sequence to speciic bilayer composition [203–206] interactions of a particular amino side chain in the bilayer with lipid head groups and water, and lipid deformation (coupled with the ease of deformability of the lipid) [183,207,208]. More recently, there appears to be a convergence of molecular modeling-based predictions of relative free energetics of different amino acid side chains as part of a macromolecular assembly and experiment [181]. Nevertheless, despite the numerous advances in computational methods addressing partitioning into membranes and lipid bilayers, it appears that the current challenges reside in determining unambiguous scales for comparing computational and experimental results. For a discussion of this issue, refer to [186] and references therein. 4.3.4.6 Helix Tilt Computational study of the tilt and orientation of isolated transmembrane (TM) helices has attracted considerable interest over the past several years. This relatively simple problem admits the controlled investigation of the role of hydrophobic mismatch in determining the conformation of TM helices. Of course, the preferred orientation of a helical segment in a given bilayer environment is also an important aspect of the folding of membrane proteins.

82

Liposomes, Lipid Bilayers and Model Membranes

The interest in helix tilt within the computational community was initially stimulated in part by experimental investigation of the WALP peptides [209]. With a sequence of adjustable length and hydrophobicity, WALP peptides afford control over the key biophysical parameters that govern interaction with the lipid bilayer. X-ray diffraction of WALP peptides in oriented bilayer stacks showed that, in contrast with gramicidin, the WALP peptides did not alter the thickness of the membrane to match the hydrophobic length of the peptide. Initial computational work based on an implicit membrane model followed the folding and insertion of WALP peptides [210]. Later work using an explicit representation of the bilayer environment demonstrated the importance of the conigurational entropy in stabilizing tilted conigurations [211]. More recently, exhaustive umbrella sampling simulations of a large number of peptide sequences and bilayers have demonstrated that the peptide tilts to match the bilayer, while the orientation is determined by the chemistry of the lanking lipid anchors [212]. In total, these publications show that sampling helix orientational degrees of freedom is tractable in the computer. Building on this work, an important emerging area is the rationalization of solid-state NMR data for TM helices with an ensemble of conigurations sampled by molecular dynamics [213]. 4.3.4.7 G-Protein-Coupled Receptors Comprising the target of roughly half of all drugs on the market and perhaps 2% of our protein coding genome, the G-protein-coupled receptors (GPCRs) are a rich and worthy target for simulation [214,215]. On the one hand, simulations offer a tantalizing complement to structural data, promising a window onto the mechanism of activation [13,216]. Yet on the other hand, simulations are beholden to structural data—a simulation has to start somewhere, and that is usually at a crystal structure. The GPCR simulation literature has therefore followed the GPCR structure literature. For a decade, molecular dynamics simulations were mostly limited to rhodopsin, the high-resolution structure of which was irst published in 2000 [217]. (Discussion of homology models of GPCRs is beyond the scope of this chapter, for a review see Ref. [218].) Modeling rhodopsin presents special challenges, including palmitoylation sites, pH-dependent activation [219], a covalently bound ligand with delocalized electronic structure that challenges conventional MD force ields [220], and sensitivity to the unusual membrane of the rod outer segment membrane. Early efforts reported unbiased simulations in the 10–100 ns timescale range, and were directed at the activation mechanism [221–223]. More recent efforts have implemented experimentally derived restraints [224] or very long (>1 µs) simulations in an effort to observe activation [225,226], suggesting that rhodopsin becomes internally hydrated upon activation [227]. Despite these insights, it is fair to say that no one has yet observed the activation cycle of rhodopsin by unbiased simulation—dark state, MetaI, MetaII. However, simulation has made signiicant contributions to demonstrate the importance of the membrane environment for rhodopsin function, especially cholesterol and polyunsaturated fatty acids, both major components of the rod outer segment membrane [92,94,228,229]. This is seminal work which sets a high standard for membrane protein simulation—many simulations of membrane proteins are published, but few bother to acknowledge the key role of the membrane. (These studies inspired a similar investigation of cholesterol interactions with the A2A adenosine receptor, discussed below.) More recently, a slew of new crystal structures have energized the ield, including structures of dark state squid rhodopsin [230] and the MetaII state [231]. Squid rhodopsin has a more rapid turnover than bovine, motivating simulations with an eye toward activation [227]. The MetaII structures will no doubt stimulate renewed interest, though to the author’s knowledge nothing is yet published. Recent successes crystallizing other GPCRs [232–240] have spawned a wave of simulations as well [216]. One of the irst observations, distilled from µs simulations of the β2 adrenergic receptor [241–243] and the A2A adenosine receptor [63,244–246], is that other GPCRs are much more conformationally variable than rhodopsin. In all cases published so far, the “ionic lock” motif is found to be conformationally heterogeneous, less a “lock” than a propensity. Other recent work has focused on the ligand, with ligand binding to β2AR observed via ultra-long timescale simulations on a special purpose computer [247], and ligand unbinding observed for A2A on commodity

Molecular Dynamics of Lipid Bilayers

83

resources [63]. A recent publication identiied three speciic binding sites for cholesterol on A2A [244]; one of these sites is corroborated by a recent crystal structure [248]. Recent work combining experimental data and simulations of a cannabinoid receptor has shown a novel access mechanism for ligand binding, via the TM region [249]. 4.3.4.8 Membrane Curvature and Amphipathic Peptides Many cellular processes require the generation and maintenance of highly curved membrane surfaces [250]. These surfaces are not formed spontaneously, but rather by the action of specialized curvature generating proteins. The Bin/Aphiphysin/Rvs family of proteins are involved in forming some of the most highly curved membranes, including the necks of budding vesicles and the T-tubule network in muscle tissue [251,252]. The mechanism by which N-BAR proteins (the N stands for an N-terminal amphipathic helix) generate curvature has been studied with simulation techniques for several years, starting with the seminal work of Blood et al. [253] in which curvature generation by a single N-BAR protein (amphiphysin) was observed by all-atom simulation. This result was exciting, as it was not clear initially that local curvature would be generated on simulation timescales. A key aspect of the simulation was that the membrane along the long axis of the BAR domain was more than twice the length of the BAR domain. As discussed earlier, the inite lengthscale of the simulation cell cuts off the undulations of the membrane, effectively stiffening the membrane against bending. By increasing the simulation lengthscale, Blood et al. were able to access longer wavelength, softer modes of undulation. Building on this work, a number of all-atom simulation studies reported on the mechanism of curvature generation by amphiphysin [254,255] and endophilin [256]. A number of recent papers also report coarse-grained and multiscale simulations of membrane remodeling [257–260]. The area of research emerging from this work are the hybrid methods that combine electron microscopy data and molecular simulation [261,262]. Molecular simulation (coarse-grained simulation in particular) provides a powerful complement to experimental data, when samples are not ordered enough for high-resolution reconstruction from experimental data alone. A closely related area concerns curvature sensing by amphipathic motifs. A number of proteins that are known to upconcentrate at curved membrane interfaces contain amphipathic motifs, including the N-BAR domains, epsin’s N-terminal homology domain, and α-synuclein. A general mechanism based on lipid packing defects has been proposed [124]; observation of amphipathic motif concentration on single liposomes of known curvature has demonstrated a Langmuir-like binding isotherm for the interaction [123,125]. Inspired by the experimental work, Cui et al. [126] used molecular simulation of a curved bilayer to directly connect the size distribution of packing defects with membrane curvature. Using metadynamics to accelerate the sampling of the folding of the amphipathic helix of endophilin, they demonstrated that endophilin’s N-terminal helix will only fold stably at a positively curved membrane interface. At such an interface, hydrophobic packing defects large enough to accommodate a bulky hydrophobic side chain are observed with suficient frequency to nucleate folding. The work discussed here represents the tip of the iceberg. Interest in the interplay of membrane curvature, curvature sensing motifs, and curvature generation is growing at a rapid pace. Such problems are ripe for applications combining molecular simulation and experimental data in new and innovative ways, as these problems tend to confound conventional structure-based approaches.

4.3.5

coarse-graIned ModelIng

In this section, the length- and timescale limitations discussed in Section 4.3.2 are used to motivate a brief discussion of “coarse-grained” modeling, an important and rapidly developing area of research that aims to overcome these limitations. In a “coarse-grained” model, one considers a model with many fewer “effective” degrees of freedom than in an all-atom representation. Detail is traded in exchange for computational expediency. Such a trade is often dictated by the problem. For example, if the goal of a simulation is to capture some long wavelength material properties of

84

Liposomes, Lipid Bilayers and Model Membranes

a bilayer (bending modulus, say), then it is neither necessary nor desirable to maintain an all-atom level of complexity. But where does one get a coarse-grained model? What we now call coarse-grained models of bilayers have their roots in the soft condensed matter physics literature [263–267]. Motivated by the success of the Helfrich treatment of a bilayer as a two-dimensional sheet with the peculiar properties of luid bilayers, a generation of physicists turned their attention to developing models that resolve individual lipids. These models represent a lipid as a few effective interaction sites (typically three or less), with the interactions between sites determined in an empirical way, by demanding that the long wavelength properties of interest are well represented. This generation of models demonstrated that capturing a few key, generic features of lipids—amphiphilic nature, aspect ratio, size—is suficient to model some of the long wavelength physics—bending modulus, area compressibility, and spontaneous curvature. Lipid models with a similar level of resolution are still used in many applications, such as coarse-grained models of N-BAR domain remodeling [257,259,262], lipoprotein nanodisc assembly [268], and membrane deformation by virus particles [269,270]. After all, the model resolution is to a large extent dictated by the problem at hand. If you want to simulate a 500 nm diameter vesicle, roughly 300,000 lipids are required, which demands a model with few sites per lipid. The next generation of coarse-grained models sought to achieve a higher degree of chemical speciicity [263,271–273]. Additional interaction sites were introduced to model the hydrocarbon chains and headgroup regions, often with multiple sites to distinguish different chemical regions in each. At this level, it becomes possible to capture details such as headgroup chemistry and unsaturated bonds. Of course, the increased chemical detail demands a more careful parameterization of the interactions. Several different solutions to this problem have been developed, which may broadly be characterized as “bottom-up” and “top-down.” In the spirit of the earlier generation of models discussed above, top-down approaches adjust inter-site interactions to reproduce some experimental observable. By far the most successful and widely used top-down approach is the MARTINI model, introduced by Marrink and coworkers [274,275]. (For a review, see the chapter in Ref. [4].) MARTINI assumes Lennard–Jones interactions between sites—a convenient choice for implementation in existing software—and then adjusts the well depth and radius to reproduce measured partition coeficients. The effect of water is modeled with a bath of Lennard–Jones solvent particles, each with a size of about four waters, to be consistent with the roughly 4-to-1 heavy atom-to-CG site resolution of MARTINI. MARTINI has proven to be remarkably effective at capturing material properties of membranes and their dependence on features of the lipids, such as length of acyl chains, and composition, such as cholesterol content. The spontaneous formation of cholesterol and saturated hydrocarbon-rich domains has been observed with MARTINI [276], as well as many other phenomena involving partitioning of lipids and proteins in heterogeneous bilayers [277–279]. These are noteworthy results, since a preferential interaction between domain-forming species is not built in “by hand,” but rather emerges from the parameterization based on solvation propensities. Other recent applications of MARTINI include fusion [280,281] and polyunsaturated fatty acid interactions with cholesterol [282]. The MARTINI model has also been extended to peptides, but is under active development—coarsegrained modeling of proteins is an extremely challenging problem in its own right [283]. For a more complete overview, see Ref. [284]. “Bottom-up” approaches to CG model parameterization seek to determine the CG interactions directly from all-atom simulation data [271,273,285–287]. Bottom-up approaches are based on the recognition that the CG model is rigorously related to the underlying all-atom model by a projection—the conigurational distribution of coarse-grained coordinates is determined by the Boltzmann-averaging over all-atom conigurations:

P(R) =

∫ drδ (R − Φ(r))exp(− βH (r)) Z (β )

(4.3)

Molecular Dynamics of Lipid Bilayers

85

where R is a coniguration of the CG model, r is a coniguration of the all-atom model, β is the inverse temperature, H(r) is the Hamiltonian of the all-atom system, and Z(β) is the all-atom partition function. Φ(r) projects an all-atom coniguration onto a CG one, the delta function therefore ensures that only all-atom conigurations consistent with R contribute to the average. Equation 4.3 offers a straightforward recipe for determining the CG potential of mean force that is consistent with the underlying all-atom model, by the Boltzmann-averaging over all-atom conigurations that are consistent with a given CG coniguration. Unfortunately, direct evaluation of Equation 4.3 is numerically impractical, as it presupposes good sampling of coniguration space—precisely the problem that coarse-grained models are intended to solve. Several solutions have been developed to circumvent the problem of direct evaluation of Equation 4.3. The earliest approaches used the structure of the liquid state as observed in all-atom simulation as the target data. In the irst such application known to the present authors, Shelley et al. [273] showed that radial distribution functions could be well reproduced by iterative adjustment of Lennard–Jones parameters based on the Boltzmann-inverted PMF [273]. At nearly the same time, Müller-Plathe and coworkers used a simplex approach to it a CG force ield to all-atom radial distribution target data, showing that arbitrarily complex representations of two-body CG potentials could be it [287]. More recently, a different approach to determining effective interaction potentials from RDFs, the iterative Monte Carlo method, has been applied to bilayer systems [285]. A different approach, called force-matching, takes as the target data the forces observed in allatom simulation. First applied to CG biomolecular systems by Izvekov and Voth [271], force-matching was originally based on earlier work that derived effective, empirical force ields from ab initio data [288–290] by least-squares itting of CG forces to their all-atom counterparts. Force-matching by least squares is called “Multiscale coarse-graining” (MS-CG). Initially seen as a practical solution, it was later shown that for an isotropic, homogeneous system the MS-CG equations are in fact equivalent to the generalized Yvon–Born–Green equations [291]. The pair potential derived by MS-CG therefore incorporates two- and three-body interactions observed at the all-atom level, providing a rigorous connection between scales [292–294]. To summarize, there are several approaches that offer a route to accessing longer length and time scales at the expense of chemical detail. Deciding on an approach will depend on the details of the problem at hand, and demands a thorough understanding of the limitations of each approach.

4.4

SUMMARY

Molecular simulation of lipids and membranes is an area of exciting and rapid discovery. Driven by advances in hardware and software, as well as the continued development of models and force ields, the next decade promises to continue the tradition of careful science that has helped to shape our understanding of the biophysics of bilayers. However, given that molecular simulation of lipid bilayers is very much a work in progress, we caution the uninitiated to carefully consider the three questions that organize the review before embarking on a simulation project. (1) Are the timescale and lengthscale of my problem tractable in the computer? (2) Are there well-validated models available for all of the pieces of my problem? (3) Is the important biophysics/chemistry captured by the current state of the art? After all, any simulation will produce numbers, but it is up to you to ensure that they are meaningful data.

REFERENCES 1. Fan J, Sammalkorpi M, Haataja M. Formation and regulation of lipid microdomains in cell membranes: Theory, modeling, and speculation. FEBS Letters. 2010;584(9):1678–84. 2. Brown FLH. Elastic modeling of biomembranes. Annual Reviews of Physical Chemistry. 2008;59:685–712. 3. Ayton GS, Voth GA. Systematic multiscale simulation of membrane protein systems. Current Opinion in Structural Biology. 2009;19(2):138–44.

86

Liposomes, Lipid Bilayers and Model Membranes

4. Voth GA, editor. Coarse-Graining of Condensed Phase and Biomolecular Systems. Boca Raton, FL: CRC Press; 2009. 5. Pastor RW. Molecular dynamics and Monte Carlo simulations of lipid bilayers. Current Opinion in Structural Biology. 1994;4(4):486–92. 6. Feller SE. Molecular dynamics simulations of lipid bilayers. Current Opinion in Colloid and Interface Science. 2000;5:217–23. 7. Pastor RW, Venable RM. Molecular and stochastic dynamics simulations of lipid membranes. In: van Gunsteren WF, Weiner PK, Wilkinson AJ, editors. Computer Simulation of Biomolecular Systems. Leiden: ESCOM; 1993. pp. 443–63. 8. Berkowitz ML, Kindt JT. Molecular Detailed Simulations of Lipid Bilayers. Reviews in Computational Chemistry. Hoboken, NJ: John Wiley & Sons, Inc.; 2010. pp. 253–86. 9. Faraldo-Gómez JD, Forrest LR. Modeling and simulation of ion-coupled and ATP-driven membrane proteins. Current Opinion in Structural Biology. 2011;21(2):173–9. 10. Feig M, Brooks Iii CL. Recent advances in the development and application of implicit solvent models in biomolecule simulations. Current Opinion in Structural Biology. 2004;14(2):217–24. 11. Forrest LR, Sansom MSP. Membrane simulations: Bigger and better? Current Opinion in Structural Biology. 2000;10(2):174–81. 12. Gumbart J, Wang Y, Aksimentiev A, Tajkhorshid E, Schulten K. Molecular dynamics simulations of proteins in lipid bilayers. Current Opinion in Structural Biology. 2005;15(4):423–31. 13. Johnston JM, Filizola M. Showcasing modern molecular dynamics simulations of membrane proteins through G protein-coupled receptors. Current Opinion in Structural Biology. 2011;21(4):552–8. 14. Khalili-Araghi F, Gumbart J, Wen P-C, Sotomayor M, Tajkhorshid E, Schulten K. Molecular dynamics simulations of membrane channels and transporters. Current Opinion in Structural Biology. 2009; 19(2):128–37. 15. Lindahl E, Sansom MSP. Membrane proteins: Molecular dynamics simulations. Current Opinion in Structural Biology. 2008;18(4):425–31. 16. Lipowsky R. The morphology of lipid membranes. Current Opinion in Structural Biology. 1995;5(4):531–40. 17. Merz Jr KM. Molecular dynamics simulations of lipid bilayers. Current Opinion in Structural Biology. 1997;7(4):511–7. 18. Mouritsen OG, Jørgensen K. Small-scale lipid-membrane structure: Simulation versus experiment. Current Opinion in Structural Biology. 1997;7(4):518–27. 19. Risselada HJ, Grubm̈ller H. How SNARE molecules mediate membrane fusion: Recent insights from molecular simulations. Current Opinion in Structural Biology. 2012;22(2):187–96. 20. Sansom MSP. Models and simulations of ion channels and related membrane proteins. Current Opinion in Structural Biology. 1998;8(2):237–44. 21. Scott HL. Modeling the lipid component of membranes. Current Opinion in Structural Biology. 2002;12(4):495–502. 22. Sherwood P, Brooks BR, Sansom MSP. Multiscale methods for macromolecular simulations. Current Opinion in Structural Biology. 2008;18(5):630–40. 23. Shinoda W, DeVane R, Klein ML. Computer simulation studies of self-assembling macromolecules. Current Opinion in Structural Biology. 2012;22(2):175–86. 24. Egberts E, Berendsen HJC. Molecular dynamics simulation of a smectic liquid crystal with atomic detail. The Journal of Chemical Physics. 1988;89(6):3718–32. 25. Berkowitz ML, Raghavan K. Computer simulation of a water/membrane interface. Langmuir. 1991;7(6):1042–4. 26. Alper HE, Bassolino D, Stouch TR. Computer simulation of a phospholipid monolayer–water system: The inluence of long range forces on water structure and dynamics. The Journal of Chemical Physics. 1993;98(12):9798–807. 27. Alper HE, Bassolino-Klimas D, Stouch TR. The limiting behavior of water hydrating a phospholipid monolayer: A computer simulation study. The Journal of Chemical Physics. 1993;99(7):5547–59. 28. Marrink SJ, Berkowitz M, Berendsen HJC. Molecular dynamics simulation of a membrane/water interface: The ordering of water and its relation to the hydration force. Langmuir. 1993;9(11):3122–31. 29. Patra M, Karttunen M, Hyvonen MT, Falck E, Lindqvist P, Vattulainen I. Molecular dynamics simulations of lipid Bilayers: Major artifacts due to truncating electrostatic interactions. Biophysical Journal. 2003;84(6):3636–45. 30. Feller SE, Pastor RW, Rojnuckarin A, Bogusz S, Brooks BR. Effect of electrostatic force truncation on interfacial and transport properties of water. The Journal of Physical Chemistry. 1996;100(42):17011–20.

Molecular Dynamics of Lipid Bilayers

87

31. Damodaran KV, Merz KM, Gaber BP. Structure and dynamics of the dilauroylphosphatidylethanolamine lipid bilayer. Biochemistry. 1992;31(33):7656–64. 32. Egberts E, Marrink SJ, Berendsen HJC. Molecular dynamics simulation of a phospholipid membrane. European Biophysics Journal 1994;22(6):423–36. 33. Heller H, Schaefer M, Schulten K. Molecular dynamics simulation of a bilayer of 200 lipids in the gel and in the liquid crystal phase. The Journal of Physical Chemistry. 1993;97(31):8343–60. 34. Robinson AJ, Richards WG, Thomas PJ, Hann MM. Head group and chain behavior in biological membranes: A molecular dynamics computer simulation. Biophysical Journal. 1994;67(6):2345–54. 35. Shinoda W, Fukada T, Okazaki S, Okada I. Molecular dynamics simulation of the dipalmitoylphosphatidylcholine (DPPC) lipid bilayer in the luid phase using the Noś-Parrinello-Rahman NPT ensemble. Chemical Physics Letters. 1995;232(3):308–12. 36. Tu K, Tobias DJ, Klein ML. Constant pressure and temperature molecular dynamics simulation of a fully hydrated liquid crystal phase dipalmitoylphosphatidylcholine bilayer. Biophysical Journal. 1995;69(6):2558–62. 37. Venable RM, Zhang Y, Hardy BJ, Pastor RW. Molecular dynamics simulations of a lipid bilayer and of hexadecane: An investigation of membrane luidity. Science. 1993;262(5131):223–6. 38. Zhou F, Schulten K. Molecular dynamics study of a membrane–water interface. The Journal of Physical Chemistry. 1995;99(7):2194–207. 39. Petrache HI, Feller SE, Nagle JF. Determination of component volumes of lipid bilayers from simulations. Biophysical Journal. 1997;72(5):2237–42. 40. Tu K, Tobias DJ, Blasie JK, Klein ML. Molecular dynamics investigation of the structure of a fully hydrated gel-phase dipalmitoylphosphatidylcholine bilayer. Biophysical Journal. 1996;70(2):595–608. 41. Essmann U, Berkowitz ML. Dynamical properties of phospholipid bilayers from computer simulation. Biophysical Journal. 1999;76(4):2081–9. 42. Pastor RW. Techniques and applications of Langevin dynamics simulations. In: Luckhurst GR, Veracini CA, editors. The Molecular Dynamics of Liquid Crystals. Netherlands: Kluwer Academic Publishers; 1994. pp. 85–138. 43. Benz RW, Castro-Román F, Tobias DJ, White SH. Experimental validation of molecular dynamics simulations of lipid bilayers: A new approach. Biophysical Journal. 2005;88(2):805–17. 44. Berger O, Edholm O, Jähnig F. Molecular dynamics simulations of a luid bilayer of dipalmitoylphosphatidylcholine at full hydration, constant pressure, and constant temperature. Biophysical Journal. 1997;72(5):2002–13. 45. Feller SE, MacKerell AD. An improved empirical potential energy function for molecular simulations of phospholipids. Journal of Physical Chemistry B. 2000;104:7510–5. 46. Klauda JB, Venable RM, Freites JA, O’Connor JW, Tobias DJ, Mondragon-Ramirez C, Vorobyov I, MacKerell AD, Pastor RW. Update of the CHARMM all-atom additive force ield for lipids: Validation on six lipid types. Journal of Physical Chemistry B. 2010;114(23):7830–43. 47. Smondyrev AM, Berkowitz ML. Molecular dynamics study of Sn-1 and Sn-2 chain conformations in dipalmitoylphosphatidylcholine membranes. The Journal of Chemical Physics. 1999;110(8):3981–5. 48. Feller SE, Huster D, Gawrisch K. Interpretation of NOESY cross-relaxation rates from molecular dynamics simulation of a lipid bilayer. Journal of the American Chemical Society. 1999;121(38):8963–4. 49. Lindahl E, Edholm O. Mesoscopic undulations and thickness luctuations in lipid Bilayers from molecular dynamics simulations. Biophysical Journal. 2000;79(1):426–33. 50. Feller SE, Pastor RW. Constant surface tension simulations of lipid bilayers: The sensitivity of surface areas and compressibilities. The Journal of Chemical Physics. 1999;111(3):1281–7. 51. Åman K, Lindahl E, Edholm O, Håkansson P, Westlund P-O. Structure and dynamics of interfacial water in an L± phase lipid bilayer from molecular dynamics simulations. Biophysical Journal. 2003;84(1):102–15. 52. Hofsäß C, Lindahl E, Edholm O. Molecular dynamics simulations of phospholipid bilayers with cholesterol. Biophysical Journal. 2003;84(4):2192–206. 53. Pasenkiewicz-Gierula M, RÛg T, Kitamura K, Kusumi A. Cholesterol effects on the phosphatidylcholine bilayer polar region: A molecular simulation study. Biophysical Journal. 2000;78(3):1376–89. 54. Smondyrev AM, Berkowitz ML. Structure of dipalmitoylphosphatidylcholine/cholesterol bilayer at low and high cholesterol concentrations: Molecular dynamics simulation. Biophysical Journal. 1999;77(4):2075–89. 55. Smondyrev AM, Berkowitz ML. Molecular dynamics simulation of dipalmitoylphosphatidylcholine membrane with cholesterol sulfate. Biophysical Journal. 2000;78(4):1672–80. 56. Frenkel D, Smit B. Understanding Molecular Simulation: 2nd ed. San Diego: Academic Press; 2002.

88

Liposomes, Lipid Bilayers and Model Membranes

57. Ewald P. Die Berechnung optischer und elektrostatischer Gitterpotentiale. Annals of Physics. 1921;369:253–87. 58. Essmann U, Perera L, Berkowitz ML, Darden T, Lee H, Pedersen LG. A smooth particle mesh Ewald method. The Journal of Chemical Physics. 1995;103(19):8577–93. 59. Shan Y, Klepeis JL, Eastwood MP, Dror RO, Shaw DE. Gaussian split Ewald: A fast Ewald mesh method for molecular simulation. The Journal of Chemical Physics. 2005;122(5):054101–13. 60. Karttunen M, Rottler Jr, Vattulainen I, Sagui C. Chapter 2 electrostatics in biomolecular simulations: Where are we now and where are we heading? In: Scott EF, editor. Current Topics in Membranes. Waltham, MA: Academic Press; 2008. pp. 49–89. 61. Tobias DJ. Electrostatics calculations: Recent methodological advances and applications to membranes. Current Opinion in Structural Biology. 2001;11(2):253–61. 62. Bowers KJ, Chow E, Huafeng X, Dror RO, Eastwood MP, Gregersen BA, Klepeis JL, Kolossvary I, Moraes MA, Sacerdoti FD, Salmon JK, Yibing S, Shaw DE, editors. Scalable Algorithms for Molecular Dynamics Simulations on Commodity Clusters. SC 2006 Conference, Proceedings of the ACM/IEEE; 2006 11–17 Nov. 2006. 63. Lee JY, Lyman E. Agonist dynamics and conformational selection during microsecond simulations of the A2A adenosine receptor. Biophysical Journal. 2012;102(9):2114–20. 64. Chow E, Rendleman CA, Bowers KJ, Dror RO, Hughes DA, Gullingsrud J, Sacerdoti FD, Shaw DE. Desmond perfromance on a cluster of multicore processors: D. E. Shaw Research Technical Report DESRES/TR–2008. 65. Bauer BA, Davis JE, Taufer M, Patel S. Molecular dynamics simulations of aqueous ions at the liquid– vapor interface accelerated using graphics processors. Journal of Computational Chemistry. 2011;32(3): 375–85. 66. Ganesan N, Bauer BA, Lucas TR, Patel S, Taufer M. Structural, dynamic, and electrostatic properties of fully hydrated DMPC bilayers from molecular dynamics simulations accelerated with graphical processing units (GPUs). Journal of Computational Chemistry. 2011;32(14):2958–73. 67. Götz AW, Williamson MJ, Xu D, Poole D, Le Grand S, Walker RC. Routine microsecond molecular dynamics simulations with AMBER on GPUs. 1. Generalized Born. Journal of Chemical Theory and Computation. 2012;8(5):1542–55. 68. Stone JE, Hardy DJ, Uimtsev IS, Schulten K. GPU-accelerated molecular modeling coming of age. Journal of Molecular Graphics and Modelling. 2010;29(2):116–25. 69. Shaw DE, Deneroff MM, Dror RO, Kuskin JS, Larson RH, Salmon JK, Young C et al. Anton, a special-purpose machine for molecular dynamics simulation. Proceedings of the 34th Annual International Symposium on Computer Architecture; San Diego, California, USA. 1250664: ACM; 2007. pp. 1–12. 70. Seifert U. Conigurations of luid membranes and vesicles. Advances in Physics. 1997;46:13–137. 71. Dickson CJ, Rosso L, Betz RM, Walker RC, Gould IR. GAFFlipid: A general amber force ield for the accurate molecular dynamics simulation of phospholipid. Soft Matter. 2012;8(37):9617–27. 72. Jójárt B, Martinek TA. Performance of the general amber force ield in modeling aqueous POPC membrane bilayers. Journal of Computational Chemistry. 2007;28(12):2051–8. 73. Martinez-Seara H, Róg T, Pasenkiewicz-Gierula M, Vattulainen I, Karttunen M, Reigada R. Effect of double bond position on lipid bilayer properties: Insight through atomistic simulations. The Journal of Physical Chemistry B. 2007;111(38):11162–8. 74. Klauda JB, Venable RM, MacKerell AD, Pastor RW. Considerations for lipid force ield development. In: Feller SE, editor. Current Topics in Membranes: Computational Modeling of Membrane Bilayers. Amsterdam: Elsevier; 2008. pp. 1–48. 75. Ulmschneider JP. New insights into the peptide-membrane partitioning equilibrium from in silico free surface-to-bilayer peptide insertion. In: Pabst G, Kucˇerka N, Nieh M-P, Katsaras J, editors. Liposomes, Lipid Bilayers and Model Membranes. Boca Raton, FL: CRC Press; 2013. pp. 99–110. 76. Róg T, Pasenkiewicz-Gierula M, Vattulainen I, Karttunen M. Ordering effects of cholesterol and its analogues. Biochimica et Biophysica Acta (BBA)—Biomembranes. 2009;1788(1):97–121. 77. Chiu SW, Jakobsson E, Mashl RJ, Scott HL. Cholesterol-induced modiications in lipid bilayers: A simulation study. Biophysical Journal. 2002;83(4):1842–53. 78. Pandit SA, Khelashvili G, Jakobsson E, Grama A, Scott HL. Lateral organization in lipid–cholesterol mixed bilayers. Biophysical Journal. 2007;92(2):440–7. 79. Falck E, Patra M, Karttunen M, Hyv̈nen MT, Vattulainen I. Lessons of slicing membranes: Interplay of packing, free area, and lateral diffusion in phospholipid/cholesterol bilayers. Biophysical Journal. 2004;87(2):1076–91.

Molecular Dynamics of Lipid Bilayers

89

80. Róg T, Pasenkiewicz-Gierula M, Vattulainen I, Karttunen M. What happens if cholesterol is made smoother: Importance of methyl substituents in cholesterol ring structure on phosphatidylcholine sterol interaction. Biophysical Journal. 2007;92(10):3346–57. 81. Róg T, Pasenkiewicz-Gierula M. Cholesterol effects on the phosphatidylcholine bilayer nonpolar region: A molecular simulation study. Biophysical Journal. 2001;81(4):2190–202. 82. Róg T, Pasenkiewicz-Gierula M. Non-polar interactions between cholesterol and phospholipids: A molecular dynamics simulation study. Biophysical Chemistry. 2004;107(2):151–64. 83. Harroun TA, Katsaras J, Wassall SR. Cholesterol hydroxyl group is found to reside in the center of a polyunsaturated lipid membrane. Biochemistry. 2006;45(4):1227–33. 84. Marrink SJ, de Vries AH, Harroun TA, Katsaras J, Wassall SR. Cholesterol shows preference for the interior of polyunsaturated lipid membranes. Journal of the American Chemical Society. 2007;130(1):10–1. 85. Pitman MC, Suits F, MacKerell AD, Feller SE. Molecular-level organization of saturated and polyunsaturated fatty acids in a phosphatidylcholine bilayer containing cholesterol. Biochemistry. 2004;43(49):15318–28. 86. Lim JB, Rogaski B, Klauda JB. Update of the cholesterol force ield parameters in CHARMM. The Journal of Physical Chemistry B. 2011;116(1):203–10. 87. Klauda JB, Monje V, Kim T, Im W. Improving the CHARMM force ield for polyunsaturated fatty acid chains. The Journal of Physical Chemistry B. 2012;116(31):9424–31. 88. Czub J, Baginski M. Comparative molecular dynamics study of lipid membranes containing cholesterol and ergosterol. Biophysical Journal. 2006;90(7):2368–82. 89. Smondyrev AM, Berkowitz ML. Molecular dynamics simulation of the structure of dimyristoylphosphatidylcholine bilayers with cholesterol, ergosterol, and lanosterol. Biophysical Journal. 2001;80(4):1649–58. 90. Perlmutter JD, Sachs JN. Inhibiting lateral domain formation in lipid bilayers: Simulations of alternative steroid headgroup chemistries. Journal of the American Chemical Society. 2009;131(45):16362–3. 91. Feller SE, Gawrisch K, MacKerell AD. Polyunsaturated fatty acids in lipid bilayers: Intrinsic and environmental contributions to their unique physical properties. Journal of the American Chemical Society. 2001;124(2):318–26. 92. Feller SE, Gawrisch K, Woolf TB. Rhodopsin exhibits a preference for solvation by polyunsaturated docosohexaenoic acid. Journal of the American Chemical Society. 2003;125(15):4434–5. 93. Brown MF. Inluence of non-lamellar forming lipids on rhodopsin. Current Topics Membr. 1997;44:285–356. 94. Pitman MC, Grossield A, Suits F, Feller SE. Role of cholesterol and polyunsaturated chains in lipid– protein interactions: Molecular dynamics simulation of rhodopsin in a realistic membrane environment. Journal of the American Chemical Society. 2005;127(13):4576–7. 95. Carrillo-Tripp M, Feller SE. Evidence for a mechanism by which ω-3 polyunsaturated lipids may affect membrane protein function. Biochemistry. 2005;44(30):10164–9. 96. Eldho NV, Feller SE, Tristram-Nagle S, Polozov IV, Gawrisch K. Polyunsaturated docosahexaenoic versus docosapentaenoic acid: Differences in lipid matrix properties from the loss of one double bond. Journal of the American Chemical Society. 2003;125(21):6409–21. 97. Roach C, Feller SE, Ward JA, Shaikh SR, Zerouga M, Stillwell W. Comparison of cis and trans fatty acid containing phosphatidylcholines on membrane properties. Biochemistry. 2004;43(20):6344–51. 98. Soni SP, Ward JA, Sen SE, Feller SE, Wassall SR. Effect of trans unsaturation on molecular organization in a phospholipid membrane. Biochemistry. 2009;48(46):11097–107. 99. Lingwood D, Simons K. Lipid rafts as a membrane-organizing principle. Science. 2010;327(5961):46–50. 100. Pandit SA, Vasudevan S, Chiu SW, Jay Mashl R, Jakobsson E, Scott HL. Sphingomyelin-cholesterol domains in phospholipid membranes: Atomistic simulation. Biophysical Journal. 2004;87(2): 1092–100. 101. Róg T, Pasenkiewicz-Gierula M. Cholesterol-sphingomyelin interactions: A molecular dynamics simulation study. Biophysical Journal. 2006;91(10):3756–67. 102. Lupyan D, Mezei M, Logothetis DE, Osman R. A molecular dynamics investigation of lipid bilayer perturbation by PIP2. Biophysical Journal. 2010;98(2):240–7. 103. Li Z, Venable RM, Rogers LA, Murray D, Pastor RW. Molecular dynamics simulations of PIP2 and PIP3 in lipid bilayers: Determination of ring orientation, and the effects of surface roughness on a PoissonBoltzmann description. Biophysical Journal. 2009;97(1):155–63. 104. Lai C-L, Jao CC, Lyman E, Gallop JL, Peter BJ, McMahon HT, Langen R, Voth GA. Membrane binding and self-association of the epsin N-terminal homology domain. Journal of Molecular Biology. 2012;423(5):800–17.

90

Liposomes, Lipid Bilayers and Model Membranes

105. Lai C-L, Landgraf KE, Voth GA, Falke JJ. Membrane docking geometry and target lipid stoichiometry of membrane-bound PKCŒ± C2 domain: A combined molecular dynamics and experimental study. Journal of Molecular Biology. 2010;402(2):301–10. 106. Falck E, Ŕg T, Karttunen M, Vattulainen I. Lateral diffusion in lipid membranes through collective lows. Journal of the American Chemical Society. 2007;130(1):44–5. 107. Lindahl E, Edholm O. Molecular dynamics simulation of NMR relaxation rates and slow dynamics in lipid bilayers. The Journal of Chemical Physics. 2001;115(10):4938–50. 108. Patra M, Karttunen M, Hyv̈nen MT, Falck E, Vattulainen I. Lipid bilayers driven to a wrong lane in molecular dynamics simulations by subtle changes in long-range electrostatic interactions. The Journal of Physical Chemistry B. 2004;108(14):4485–94. 109. Klauda JB, Brooks BR, Pastor RW. Dynamical motions of lipids and a inite size effect in simulations of bilayers. The Journal of Chemical Physics. 2006;125(14):144710–8. 110. Van Kampen N. Stochastic Processes in Physics and Chemistry, 3rd ed. Oxford: Elsevier; 2007. 111. Flenner E, Das J, Rheinstädter MC, Kosztin I. Subdiffusion and lateral diffusion coeficient of lipid atoms and molecules in phospholipid bilayers. Physical Review E. 2009;79(1):011907. 112. Jeon J-H, Martinez-Seara H, Javanainen M, Metzler R. Anomalous diffusion of phospholipids and cholesterols in a lipid bilayer and its origins. Physical Review Letters. 2012;109:188103. 113. Canham PB. The minimum bending energy as a possible explanation of the biconcave shape of the red blood cell. Journal of Theoretical Biology. 1970;26:61–81. 114. Helfrich W. Eleastic properties of lipid bilayers: Theory and possible experiments. Z Nat. 1973;28:693–703. 115. Goetz RD, Gompper G, Lipowsky R. Mobility and elasticity of self-assembled membranes. Physical Review Letters. 1999;82(1):221–4. 116. Marrink SJ, Mark AE. Effect of undulations on surface tension in simulated bilayers. The Journal of Physical Chemistry B. 2001;105(26):6122–7. 117. Brandt EG, Braun AR, Sachs JN, Nagle JF, Edholm O. Interpretation of luctuation spectra in lipid bilayer simulations. Biophysical Journal. 2011;100(9):2104–11. 118. Watson MC, Brandt EG, Welch PM, Brown FLH. Determining biomembrane bending rigidities from simulations of modest size. Physical Review Letters. 2012;109(2):028102. 119. Braun AR, Brandt EG, Edholm O, Nagle JF, Sachs JN. Determination of electron density proiles and area from simulations of undulating membranes. Biophysical Journal. 2011;100(9):2112–20. 120. Braun AR, Sachs JN. Extracting experimental observables from molecular dynamics simulations of membranes. Annual Reports in Computational Chemistry. 2011;7:125–50. 121. Kucerka N, Nagle JF, Sachs JN, Feller SE, Pencer J, Jackson A, Katsaras J. Lipid bilayer structure determined by the simultaneous analysis of neutron and x-ray scattering data. Biophysical Journal. 2008;95(5):2356–67. 122. Kucerka N, Perlmutter JD, Pan J, Tristram-Nagle S, Katsaras J, Sachs JN. The effect of cholesterol on short- and long-chain monounsaturated lipid bilayers as determined by molecular dynamics simulations and x-ray scattering. Biophysical Journal. 2008;95(6):2792–805. 123. Bhatia VK, Madsen KL, Bolinger P-Y, Kunding A, Hedegard P, Gether U, Stamou D. Amphipathic motifs in BAR domains are essential for membrane curvature sensing. The EMBO Journal. 2009;28(21):3303–14. 124. Drin G, Casella J-F, Gautier R, Boehmer T, Schwartz TU, Antonny B. A general amphipathic [alpha]helical motif for sensing membrane curvature. Nature Structural and Molecular Biology. 2007;14(2): 138–46. 125. Hatzakis NS, Bhatia VK, Larsen J, Madsen KL, Bolinger P-Y, Kunding AH, Castillo J, Gether U, Hedegard P, Stamou D. How curved membranes recruit amphipathic helices and protein anchoring motifs. Nature Chemical Biology. 2009;5(11):835–41. 126. Cui H, Lyman E, Voth GA. Mechanism of membrane curvature sensing by amphipathic helix containing proteins. Biophys Journal 2011;100(5):1271–9. 127. Sodt AJ, Pastor RW. The tension of a curved surface from simulation. The Journal of Chemical Physics. 2012;137(23):234101–12. 128. Mabrey S, Sturtevant JM. Investigation of phase transitions of lipids and lipid mixtures by sensitivity differential scanning calorimetry. Proceedings of the National Academy of Sciences of the United States of America. 1976;73(11):3862–6. 129. Jain MK, White III HB. Long range order in biomembranes. Advances in Lipid Research. 1977;15:1–60. 130. Feigenson GW, Buboltz JT. Ternary phase diagram of dipalmitoyl-PC/dilauroyl-PC/cholesterol: Nanoscopic domain formation driven by cholesterol. Biophysical Journal. 2001;80(6):2775–88. 131. Veatch SL, Keller SL. Seeing spots: Complex phase behavior in simple membranes. Biochimica Biophysica Acta. 2005;1746:172–85.

Molecular Dynamics of Lipid Bilayers

91

132. Veatch SL, Polozov IV, Gawrisch K, Keller SL. Liquid domains in vesicles investigated by NMR and luorescence microscopy. Biophysics Journal. 2004;86(5):2910–22. 133. Mills TT, Tristram-Nagle S, Heberle FA, Morales NF, Zhao J, Wu J, Toombes GES, Nagle JF, Feigenson GW. Liquid–liquid domains in bilayers detected by wide angle x-ray scattering. Biophysical Journal. 2008;95(2):682–90. 134. Veatch SL, Soubias O, Keller SL, Gawrisch K. Critical luctuations in domain-forming lipid mixtures. Proceedings of the National Academy of Sciences of the United States of America. 2007;104(45):17650–5. 135. Honerkamp-Smith AR, Cicuta P, Collins MD, Veatch SL, den Nijs M, Schick M, Keller SL. Line tensions, correlation lengths, and critical exponents in lipid membranes near critical points. Biophysical Journal. 2008;95(1):236–46. 136. Hohenberg PC, Halperin BI. Theory of dynamic critical phenomena. Reviews of Modern Physics. 1977;49(3):435–79. 137. Chiu SW, Jakobsson E, Subramaniam S, Scott HL. Combined Monte Carlo and molecular dynamics simulation of fully hydrated dioleyl and palmitoyl-oleyl phosphatidylcholine lipid bilayers. Biophysical Journal. 1999;77(5):2462–9. 138. Ohkubo YZ, Pogorelov TV, Arcario MJ, Christensen GA, Tajkhorshid E. Accelerating membrane insertion of peripheral proteins with a novel membrane mimetic model. Biophysical Journal. 2012;102(9):2130–9. 139. de Joannis J, Coppock PS, Yin F, Mori M, Zamorano A, Kindt JT. Atomistic simulation of cholesterol effects on miscibility of saturated and unsaturated phospholipids: Implications for liquid-ordered/liquiddisordered phase coexistence. Journal of the American Chemical Society. 2011;133(10):3625–34. 140. Coppock PS, Kindt JT. Determination of phase transition temperatures for atomistic models of lipids from temperature-dependent stripe domain growth kinetics. The Journal of Physical Chemistry B. 2010;114(35):11468–73. 141. Kindt JT. Atomistic simulation of mixed-lipid bilayers: Mixed methods for mixed membranes. Molecular Simulation. 2011;37:516–24. 142. Yin F, Kindt JT. Hydrophobic mismatch and lipid sorting near OmpA in mixed bilayers: Atomistic and coarse-grained simulations. Biophysical Journal. 2012;102(10):2279–87. 143. Xie W, Pu J, MacKerell Jr. AD, Gao J. Development of a polarizable intermolecular potential function (PIPF) for liquid amides and alkanes. Journal of Chemical Theory and Computation. 2007;3(6):1878–89. 144. Gao J, Habibollazadeh D, Shao L. A polarizable intermolecular potential function for simulation of liquid alcohols. Journal of Physical Chemistry. 1995;99:16460–7. 145. Gao J, Xia X. A priori evaluation of aqueous polarization effects through Monte Carlo QM-MM simulations. Science. 1992;258:631. 146. Vorobyov IV, Anisimov VM, Alexander D. MacKerell J. Polarizable empirical force ield for alkanes based on the classical Drude oscillator model. Journal of Physical Chemistry B. 2005;109:18988–99. 147. Vorobyov IV, Anisimov VM, Greene S, Venable RM, Moser A, Pastor RW, MacKerell AD. Additive and classical drude polarizable force ields for linear and cyclic ethers. Journal of Chemical Theory and Computation. 2007;3:1120. 148. Anisimov VM, Lamoureux G, Vorobyov IV, Huang N, Roux B, Alexander D. MacKerell J. Determination of electrostatic parameters for a polarizable force ield based on the classical Drude oscillator. Joural of Chemical Theory and Computation. 2005;1:153–68. 149. Anisimov VM, Vorobyov IV, Roux B, Alexander D. MacKerell J. Polarizable empirical force ield for the primary and secondary alcohol series based on the classical Drude model. Journal of Chemical Theory and Computation. 2007;3:1927–46. 150. Patel S, Brooks CL, III. CHARMM luctuating charge force ield for proteins: I. Parameterization and application to bulk organic liquid simulations. Journal of Computational Chemistry. 2004;25:1–15. 151. Patel S, Brooks CL, III. A nonadditive methanol force ield: Bulk liquid and liquid–vapor interfacial properties via molecular dynamics simulations using a luctuating charge model. Journal of Chemical Physics. 2005;122:024508. 152. Patel S, Brooks CL, III. Structure, thermodynamics, and liquid–vapor equilibrium of ethanol from molecular dynamics simulations using nonadditive interactions. Journal of Chemical Physics. 2005;123:164502. 153. Patel S, Brooks CL, III. Revisiting the hexane–water interface via molecular dynamics simulations using non-additive alkane-water potentials. Journal of Chemical Physics. 2006;124:204706. 154. Patel S, MacKerell AD, Jr., Brooks CL, III. CHARMM Fluctuating charge force ield for proteins: II Protein/solvent properties from molecular dynamics simulations using a non-additive electrostatic model. Journal of Computational Chemistry. 2004;25:1504–14.

92

Liposomes, Lipid Bilayers and Model Membranes

155. Patel S, Brooks CL, III. Fluctuating charge force ields: Recent developments and applications from small molecules to macromolecular biological systems. Molecular Simulation. 2006;36:231–49. 156. Rick SW. Simulations of ice and liquid water over a range of temperatures using the luctuating charge model. Journal of Chemical Physics. 2001;114(5):2276. 157. Rick SW, Berne BJ. Dynamical luctuating charge force ields: The aqueous solvation of amides. Journal of the American Chemical Society. 1996;118:672–9. 158. Rick SW, Berne BJ. Free energy of the hydrophobic interaction from molecular dynamics simulations: The effects of solute and solvent polarizability. Journal of Physical Chemistry B. 1997;101: 10488–93. 159. Rick SW, Stuart SJ. Potentials and algorithms for incorporating polarizability in computer simulations. In: Lipkowitz KB, Boyd DB, editors. Reviews in Computational Chemistry. Hoboken, NJ: John Wiley and Sons, Inc.; 2002. p. 89. 160. Rick SW, Stuart SJ, Bader JS, Berne BJ. Fluctuating charge force ields for aqueous solutions. Journal of Molecular Liquids. 1995;65/66:31–40. 161. Rick SW, Stuart SJ, Berne BJ. Dynamical luctuating charge force ields: Application to liquid water. Journal of Chemical Physics. 1994;101(7):6141–56. 162. Kaminski GA, Stern HA, Berne BJ, Friesner RA. Development of an accurate and robust polarizable molecular mechanics force ield from ab initio quantum chemistry. Journal of Physical Chemistry A. 2004;108:621–7. 163. Yu H, Gunsteren WFv. Accounting for polarization in molecular simulation. Computer Physics Communications. 2005;172:69–85. 164. Yu H, Geerke DP, Liu H, Gunsteren WFv. Molecular dynamics simulations of liquid methanol and methanol-water mixtures with polarizable models. Journal of Computational Chemistry. 2006;27: 1494–504. 165. Piquemal J-P, Chelli R, Procacci P, Gresh N. Key role of polarization anisotropy of water in modeling classical polarizability. Journal of Physical Chemistry A. 2007;111(33):8170–6. 166. Ledecq M, Lebon F, Durant F, Giessner-Prettre C, Marquez A, Gresh N. Modeling of copper(II) complexes with the SIBFA polarizable molecular mechanics procedure. Application to a new class of HIV-1 protease inhibitors. Journal of Physical Chemistry B. 2003;107(38):10640–52. 167. Gresh N. Energetics of Zn2+ binding to a series of biologically relevant ligands: A molecular mechanics investigation grounded on ab initio self-consistent ield supermolecule computations. Journal of Computational Chemistry. 1995;16:856–82. 168. Gresh N, Garmer DR. Comparative binding energetics of Mg2+, Ca2+, Zn2+, and Cd2+ to biologically relevant ligands: Combined ab initio SCF supermolecule and molecular mechanics investigation. Journal of Computational Chemistry. 1996;17:1481–95. 169. Jiao D, Golubkov PA, Darden TA, Ren P. Calculation of protein-ligand binding free energy by using a polarizable potential. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(17):6290. 170. Ren P, Ponder JW. Consistent treatment of inter- and intramolecular polarization in molecular mechanics calculations. Journal of Computational Chemistry. 2002;23(16):1497–506. 171. Ren P, Ponder JW. Polarizable atomic multipole water model for molecular mechanics simulation. Journal of Physical Chemistry B. 2003;107:5933–47. 172. Ponder JW. TINKER, Version 4.2. 4.2 ed: School of Medicine, Washington University; 2002. 173. Thole B. Molecular polarizabilities calculated with a modiied dipole interaction. Chemical Physics. 1981;59:341. 174. Antila HS, Salonen E. Polarizable force ields. In: Monticelli L, Salonen E, editors. Biomolecular Simulations: Methods and Protocols, Methods in Molecular Biology. New York: Springer Science + Business Media; 2012. 175. Harder E, MacKerell AD, Roux B. Many-body polarization effects and the membrane dipole potential. Journal of the American Chemical Society. 2009;131(8):2760–1. 176. Lucas TR, Bauer BA, Davis JE, Patel S. Molecular dynamics simulation of hydrated DPPC monolayers using charge equilibration force ields. Journal of Computational Chemistry. 2011;33(2):141–52. 177. Allen TW, Bastug T, Kuyucak S, Chung S-H. Gramicidin A channel as a test ground for molecular dynamics force ields. Biophysical Journal. 2003;84:2159–68. 178. Patel S, Davis J, Bauer B. Exploring ion permeation energetics in gramicidin A using polarizable charge equilibration force ields. Journal of American Chemical Society. 2009;131(39):13890–1. 179. Tian F, Cross TA. Cation transport: An example of structural based selectivity. Journal of Molecular Biology. 1999;285(5):1993–2003.

Molecular Dynamics of Lipid Bilayers

93

180. Dorairaj S, Allen TW. On the thermodynamic stability of a charged arginine sidechan in a transmembrane helix. Proceedings of the National Academy of Sciences of the United States of America. 2007;104:4943–8. 181. Fleming PJ, Freites JA, Moon CP, Tobias DJ, Fleming KG. Outer membrane phospholipase A in phospholipid bilayers: A model system for concerted computational and experimental investigations of amino acid side chain partitioning into lipid bilayers. Biochimica et Biophysica Acta (BBA) Biomembranes. 2012;1818(2):126–34. 182. Gumbart J, Roux B. Determination of membrane-insertion free energies by molecular dynamics simulations. Biophysics Journal 2012;102:795–801. 183. Hristova K, Wimley WC. A look at arginine in membranes. Journal of Membrane Biolgy. 2011;239:253–9. 184. Li L, Vorobyov I, Allen TW. Potential of mean force and pKa proile calculation for the lipid membraneexposed arginine side chain. Journal of Physical Chemistry B. 2008;112:9574–87. 185. Li L, Vorobyov I, Dorairaj S, Allen TW. Charged protein side chain movement in lipid bilayers explored with free energy simulation. Current Topics in Membrane Research. 2008;60:505–459. 186. MacCallum JL, Bennett WFD, Tieleman DP. Partitioning of amino acid side chains into lipid bilayers: Results from computer simulations and comparison to experiment. Journal of General Physiology. 2007;129:371–7. 187. MacCallum JL, Bennett WFD, Tieleman DP. Distribution of amino acids in a lipid bilayer from computer simulations. Biophysics Journal. 2008;94:3393–404. 188. Moon CP, Fleming KG. Side-chain hydrophobicity scale derived from transmembrane protein folding into lipid bilayers. Proceedings of the National Academy of Sciences of the United States of America. 2011;108:10174–7. 189. Schow EV, Freites JA, Cheng P, Bernsel A, von Heijne G, White SH, Tobias DJ. Arginine in membranes: The connection between molecular dynamics simulations and translocon-mediated insertion experiments. Journal of Membrane Biology. 2011;239:35–48. 190. Tieleman DP, MacCallum JL, Ash WL, Kandt C, Xu Z, Monticelli LM. Membrane protein simulations with a united atom lipid and all atom protein model: Side chain transfer free energies and model proteins. Journal of Physics Condensed Matter. 2006;18:S1221–34. 191. Vorobyov I, Li L, Allen TW. Assessing atomistic and coarse-grained force ields for protein–lipid interactions: The formidable challenge of an ionizable side chain in a membrane. Journal of Physical Chemistry B. 2008;112:9588–602. 192. Yoo J, Cui Q. Does arginine remain proteonated in the lipid membrane? Insights from microscopic pKa calculations. Biophysics Journal. 2008;94:L61–L3. 193. Jensen MO, Jogini V, Borhani DW, Lefler AE, Dror RO, Shaw DE. Mechanism of voltage gating in potassium channels. Science. 2012;336:229–33. 194. Jiang Y, Lee A, Chen J, Cadene M, Chait BT, MacKinnon R. The open pore conformation of potassium channels. Nature. 2002;417:523–6. 195. Jiang Y, Lee A, Chen J, Ruta V, Cadene M, Chait BT, MacKinnon R. X-ray structure of a voltage-dependent K+ channel. Nature. 2003;423:33–41. 196. Futaki S. Oligorarginine vectors for intracelular delivery: Design and cellular-uptake mechanisms. Biopolymers. 2006;84:241–9. 197. Herce HD, Garcia AE. Cell penetrating peptides: How do they do it? Journal of Biology and Physics. 2007;33:345–56. 198. Herce HD, Garcia AE, Litt J, Kane RS, Martin P, Enrique N, Rebolledo A, Milesi V. Arginine-rich peptides destabilize the plasma membrane, consistent with a pore formation translocation mechanism of cell-penetrating peptides. Biophysics Journal. 2009;97:1917–25. 199. Vives E, Schmidt J, Pelegrin A. Cell-penetrating and cell-targeting peptides in drug delivery. Biochimica et Biophysica Acta (BBA)—Reviews on Cancer. 2008;1786(2):126–38. 200. MacCallum JL, Bennett WFD, Tieleman DP. Transfer of arginine into lipid bilayers is nonadditive. Biophysics Journal. 2011;101:110–7. 201. Hessa T, White SH, von Heijne G. Membrane insertion of a potassium channel voltage sensor. Science. 2005;307:1427. 202. Vorobyov I, Bekker B, Allen TW. Electrostatics of deformable lipid membranes. Biophysics Journal. 2010;98:2904–13. 203. Bogdanov M, Heacock PN, Dowhan W. A polytopic membrane protein displays a reversible topology dependent on membrane lipid composition. EMBO Journal. 2002;21:2107–16. 204. Bogdanov M, Xie J, Dowhan W. Lipid–protein interactions drive membrane protein topogenesis in accordance with the positive inside rule. Journal of Biological Chemistry. 2009;284(15):9637–41.

94

Liposomes, Lipid Bilayers and Model Membranes

205. Bogdanov M, Xie J, Heacock PN, Dowhan W. To lip or not to lip: Lipid–protein charge interactions are a determinant of the inal membrane protein topology. Journal of Cell Biology. 2008;182:925–35. 206. Zhang W, Bogdanov M, Pi J, Pittard AJ, Dowhan W. Reversible topological organization within a polytopic membrane protein is governed by a change in membrane phospholipid composition. Journal of Biology and Chemistry. 2003;278:50128–35. 207. Marks J, Placone J, Hristova K, Wimley WC. Spontaneous membrane-translocating peptides by orthogonal high-throughput screening. Journal of American Chemical Society. 2011;133:8995–9004. 208. Wimley WC, Hristova K. Antimicrobial peptides: Successes, challenges, and unanswered questions. Journal of Membrane Biology. 2011;239:27–34. 209. Weiss TM, van der Wel PCA, Killian JA, Koeppe RE, Huang HW. Hydrophobic mismatch between helices and lipid bilayers. Biophysical Journal. 2003;84(1):379–85. 210. Im W, Brooks CL. Interfacial folding and membrane insertion of designed peptides studied by molecular dynamics simulations. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(19):6771–76. 211. Lee J, Im W. Transmembrane helix tilting: Insights from calculating the potential of mean force. Physical Review Letters. 2008;100(1):018103. 212. Kim T, Im W. Revisiting hydrophobic mismatch with free energy simulation studies of transmembrane helix tilt and rotation. Biophysical Journal. 2010;99(1):175–83. 213. Jo S, Im W. Transmembrane helix orientation and dynamics: Insights from ensemble dynamics with solid-state NMR observables. Biophysical Journal. 2011;100(12):2913–21. 214. Congreve M, Marshall F. The impact of GPCR structures on pharmacology and structure-based drug design. British Journal of Pharmacology. 2010;159(5):986–96. 215. Fredholm BB, Ijzerman AP, Jacobson KA, Linden J, Müller CE. International Union of Basic and Clinical Pharmacology. LXXXI. Nomenclature and classiication of adenosine receptors—An update. Pharmacological Reviews. 2011;63(1):1–34. 216. Grossield A. Recent progress in the study of G protein-coupled receptors with molecular dynamics computer simulations. Biochimica et Biophysica Acta (BBA)—Biomembranes. 2011;1808(7):1868–78. 217. Palczewski K, Kumasaka T, Hori T, Behnke CA, Motoshima H, Fox BA, Trong IL et al. Crystal structure of rhodopsin: A G protein-coupled receptor. Science. 2000;289(5480):739–45. 218. Mobarec JC, Sanchez R, Filizola M. Modern homology modeling of g-protein coupled receptors: Which structural template to use? Journal of Medicinal Chemistry. 2009;52(16):5207–16. 219. Mahalingam M, Martínez-Mayorga K, Brown MF, Vogel R. Two protonation switches control rhodopsin activation in membranes. Proceedings of the National Academy of Sciences. 2008;105(46):17795–800. 220. Tajkhorshid E, Baudry Jrm, Schulten K, Suhai Sn. Molecular dynamics study of the nature and origin of retinal’s twisted structure in bacteriorhodopsin. Biophysical Journal. 2000;78(2):683–93. 221. Crozier PS, Stevens MJ, Forrest LR, Woolf TB. Molecular dynamics simulation of dark-adapted rhodopsin in an explicit membrane bilayer: Coupling between local retinal and larger scale conformational change. Journal of Molecular Biology. 2003;333(3):493–514. 222. Lemaître V, Yeagle P, Watts A. Molecular dynamics simulations of retinal in rhodopsin: From the darkadapted state towards lumirhodopsin. Biochemistry. 2005;44(38):12667–80. 223. Saam J, Tajkhorshid E, Hayashi S, Schulten K. Molecular Dynamics investigation of primary photoinduced events in the activation of rhodopsin. Biophysical Journal. 2002;83(6):3097–112. 224. Hornak V, Ahuja S, Eilers M, Goncalves JA, Sheves M, Reeves PJ, Smith SO. Light activation of rhodopsin: Insights from molecular dynamics simulations guided by solid-state NMR distance restraints. Journal of Molecular Biology. 2010;396(3):510–27. 225. Grossield A, Pitman MC, Feller SE, Soubias O, Gawrisch K. Internal hydration increases during activation of the G-protein-coupled receptor rhodopsin. Journal of Molecular Biology. 2008;381(2):478–86. 226. Martínez-Mayorga K, Pitman MC, Grossield A, Feller SE, Brown MF. Retinal counterion switch mechanism in vision evaluated by molecular simulations. Journal of the American Chemical Society. 2006;128(51):16502–3. 227. Jardón-Valadez E, Bondar A-N, Tobias DJ. Coupling of retinal, protein, and water dynamics in squid rhodopsin. Biophysics Journal. 2010;99(7):2200–7. 228. Grossield A, Feller SE, Pitman MC. A role for direct interactions in the modulation of rhodopsin by ω-3 polyunsaturated lipids. Proceedings of the National Academy of Sciences of the United States of America. 2006;103(13):4888–93. 229. Khelashvili G, Grossield A, Feller SE, Pitman MC, Weinstein H. Structural and dynamic effects of cholesterol at preferred sites of interaction with rhodopsin identiied from microsecond length molecular dynamics simulations. Proteins: Structure, Function, and Bioinformatics. 2009;76(2):403–17.

Molecular Dynamics of Lipid Bilayers

95

230. Murakami M, Kouyama T. Crystal structure of squid rhodopsin. Nature. 2008;453(7193):363–7. 231. Choe H-W, Kim YJ, Park JH, Morizumi T, Pai EF, Krausz N, Hofmann KP, Scheerer P, Ernst OP. Crystal structure of metarhodopsin II. Nature. 2011;471(7340):651–5. 232. Jaakola V-P, Grifith MT, Hanson MA, Cherezov V, Chien EYT, Lane JR, Ijzerman AP, Stevens RC. The 2.6 angstrom crystal structure of a human A2A adenosine receptor bound to an antagonist. Science. 2008;322(5905):1211–7. 233. Cherezov V, Rosenbaum DM, Hanson MA, Rasmussen SGF, Thian FS, Kobilka TS, Choi H-J et al. Highresolution crystal structure of an engineered human β2-adrenergic G protein coupled receptor. Science. 2007;318(5854):1258–65. 234. Chien EYT, Liu W, Zhao Q, Katritch V, Won Han G, Hanson MA, Shi L et al. Structure of the human dopamine D3 receptor in complex with a D2/D3 selective antagonist. Science. 2010;330(6007):1091–5. 235. Hanson MA, Cherezov V, Grifith MT, Roth CB, Jaakola V-P, Chien EYT, Velasquez J, Kuhn P, Stevens RC. A speciic cholesterol binding site is established by the 2.8 ang structure of the human β2 adrenergic receptor. Structure. 2008;16(6):897–905. 236. Lebon G, Warne T, Edwards PC, Bennett K, Langmead CJ, Leslie AGW, Tate CG. Agonist-bound adenosine A2A receptor structures reveal common features of GPCR activation. Nature. 2011;474(7352): 521–5. 237. Rasmussen SGF, Choi H-J, Rosenbaum DM, Kobilka TS, Thian FS, Edwards PC, Burghammer M et al. Crystal structure of the human β2 adrenergic G-protein-coupled receptor. Nature. 2007;450(7168): 383–7. 238. Warne T, Serrano-Vega MJ, Baker JG, Moukhametzianov R, Edwards PC, Henderson R, Leslie AGW, Tate CG, Schertler GFX. Structure of a β1-adrenergic G-protein-coupled receptor. Nature. 2008;454(7203):486–91. 239. Wu B, Chien EYT, Mol CD, Fenalti G, Liu W, Katritch V, Abagyan R et al. Structures of the CXCR4 chemokine GPCR with small-molecule and cyclic peptide antagonists. Science. 2010;330(6007):1066–71. 240. Xu F, Wu H, Katritch V, Han GW, Jacobson KA, Gao Z-G, Cherezov V, Stevens RC. Structure of an agonist-bound human A2A adenosine receptor. Science. 2011;332(6027):322–7. 241. Dror RO, Arlow DH, Borhani DW, Jensen M, Piana S, Shaw DE. Identiication of two distinct inactive conformations of the β2-adrenergic receptor reconciles structural and biochemical observations. Proceedings of the National Academy of Sciences of the United States of America. 2009;106(12):4689–94. 242. Dror RO, Arlow DH, Maragakis P, Mildorf TJ, Pan AC, Xu H, Borhani DW, Shaw DE. Activation mechanism of the β2-adrenergic receptor. Proceedings of the National Academy of Sciences of the United States of America. 2011;108(46):18684–9. 243. Romo TD, Grossield A, Pitman MC. Concerted interconversion between ionic lock substrates of the β2 adrenergic receptor revealed by microsecond timescale molecular dynamics. Biophysics Journal. 2010;98(1):76–84. 244. Lee JY, Lyman E. Predictions for cholesterol interaction sites on the A2A adenosine receptor. Journal of the American Chemical Society. 2012;134(40):16512–5. 245. Lyman E, Higgs C, Kim B, Lupyan D, Shelley JC, Farid R, Voth GA. A role for a speciic cholesterol interaction in stabilizing the inactive, apo coniguration of the human A2A adenosine receptor. Structure. 2009;17:1660–8. 246. Rodrĺguez D, Piñeiro Á, Gutiérrez-de-Terán H. Molecular dynamics simulations reveal insights into key structural elements of adenosine receptors. Biochemistry. 2011;50(19):4194–208. 247. Dror RO, Pan AC, Arlow DH, Borhani DW, Maragakis P, Shan Y, Xu H, Shaw DE. Pathway and mechanism of drug binding to G-protein-coupled receptors. Proceedings of the National Academy of Sciences of the United States of America. 2011;108(32):13118–23. 248. Liu W, Chun E, Thompson AA, Chubukov P, Xu F, Katritch V, Han GW et al. Structural basis for allosteric regulation of GPCRs by sodium ions. Science. 2012;337(6091):232–6. 249. Hurst DP, Grossield A, Lynch DL, Feller S, Romo TD, Gawrisch K, Pitman MC, Reggio PH. A lipid pathway for ligand binding is necessary for a cannabinoid G protein-coupled receptor. Journal of Biological Chemistry. 2010;285(23):17954–64. 250. McMahon HT, Gallop JL. Membrane curvature and mechanisms of dynamic cell membrane remodelling. Nature. 2005;438(7068):590–6. 251. Frost A, Unger VM, De Camilli P. The BAR domain superfamily: Membrane-molding macromolecules. Cell. 2009;137(2):191–6. 252. Mim C, Unger VM. Membrane curvature and its generation by BAR proteins. Trends in Biochemical Sciences. 2012;37(12):526–33.

96

Liposomes, Lipid Bilayers and Model Membranes

253. Blood PD, Voth GA. Direct observation of Bin/amphiphysin/Rvs (BAR) domain-induced membrane curvature by means of molecular dynamics simulations. Proceedings of the National Academy of Sciences. 2006;103(41):15068–72. 254. Blood PD, Swenson RD, Voth GA. Factors inluencing local membrane curvature induction by N-BAR domains as revealed by molecular dynamics simulations. Biophysical Journal. 2008;95(4):1866–76. 255. Lyman E, Cui H, Voth GA. Water under the BAR. Biophysical Journal. 2010;99(6):1783–90. 256. Cui H, Ayton GS, Voth GA. Membrane binding by the endophilin N-BAR Domain. Biophysical Journal. 2009;97(10):2746–53. 257. Arkhipov A, Yin Y, Schulten K. Four-scale description of membrane sculpting by BAR domains. Biophysical Journal. 2008;95(6):2806–21. 258. Arkhipov A, Yin Y, Schulten K. Membrane-bending mechanism of amphiphysin N-BAR domains. Biophysical Journal. 2009;97(10):2727–35. 259. Ayton GA, Lyman E, Voth GA. Hierarchical coarse-graining strategy for protein-membrane systems to access mesoscopic scales. Faraday Discuss. 2010;144:347–57. 260. Ayton GS, Lyman E, Krishna V, Swenson RD, Mim C, Unger VM, Voth GA. New insights into BAR domain-induced membrane remodeling. Biophysical Journal. 2009;97(6):1616–25. 261. Mim C, Cui H, Gawronski-Salerno JA, Frost A, Lyman E, Voth GA, Unger VM. Structural basis of membrane bending by the N-BAR protein endophilin. Cell. 2012;149:137–45. 262. Lyman E, Cui H, Voth GA. Reconstructing protein remodeled membranes in molecular detail from mesoscopic models. Physical Chemistry, Chemical Physics. 2011;13:10430. 263. Smit B, Esselink K, Hilbers PAJ, Van Os NM, Rupert LAM, Szleifer I. Computer simulations of surfactant self-assembly. Langmuir. 1993;9(1):9–11. 264. Goetz R, Lipowsky R. Computer simulations of bilayer membranes: Self-assembly and interfacial tension. The Journal of Chemical Physics. 1998;108(17):7397–409. 265. Stevens MJ, Hoh JH, Woolf TB. Insights into the molecular mechanism of membrane fusion from simulation: Evidence for the association of splayed tails. Physical Review Letters. 2003;91(18):188102. 266. Cooke IR, Deserno M. Solvent free model for self-assembling luid bilayer membranes: Stabilization of the luid phase based on broad attractive tail potentials. Journal of Chemical Physics. 2005;123:224710. 267. Cooke IR, Kremer K, Deserno M. Tunable generic model for luid bilayer membranes. Physical Review E. 2005;72(1):011506. 268. Shih AY, Arkhipov A, Freddolino PL, Schulten K. Coarse grained protein, lipid model with application to lipoprotein particles. The Journal of Physical Chemistry B. 2006;110(8):3674–84. 269. Ayton GS, Voth GA. Multiscale computer simulation of the immature HIV-1 Virion. Biophysical Journal. 2010;99(9):2757–65. 270. Reynwar BJ, Illya G, Harmandaris VA, Muller MM, Kremer K, Deserno M. Aggregation and vesiculation of membrane proteins by curvature-mediated interactions. Nature. 2007;447(7143):461–4. 271. Izvekov S, Voth GA. A multiscale coarse-graining method for biomolecular systems. The Journal of Physical Chemistry B. 2005;109(7):2469–73. 272. Marrink SJ, Mark AE. Molecular dynamics simulation of the formation, structure, and dynamics of small phospholipid vesicles. Journal of the American Chemical Society. 2003;125(49):15233–42. 273. Shelley JC, Shelley MY, Reeder RC, Bandyopadhyay S, Klein ML. A coarse grain model for phospholipid simulations. The Journal of Physical Chemistry B. 2001;105(19):4464–70. 274. Marrink SJ, de Vries AH, Mark AE. Coarse grained model for semiquantitative lipid simulations. The Journal of Physical Chemistry B. 2003;108(2):750–60. 275. Marrink SJ, Risselada HJ, Yeimov S, Tieleman DP, de Vries AH. The MARTINI force ield: A coarse grained model for biomolecular simulations. The Journal of Physical Chemistry B. 2007;111(27):7812–24. 276. Risselada HJ, Marrink SJ. The molecular face of lipid rafts in model membranes. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(45):17367–72. 277. Domański J, Marrink SJ, Schäfer LV. Transmembrane helices can induce domain formation in crowded model membranes. Biochimica et Biophysica Acta (BBA)—Biomembranes. 2012;1818(4):984–94. 278. Schäfer LV, de Jong DH, Holt A, Rzepiela AJ, de Vries AH, Poolman B, Killian JA, Marrink SJ. Lipid packing drives the segregation of transmembrane helices into disordered lipid domains in model membranes. Proceedings of the National Academy of Sciences of the United States of America. 2011;108(4):1343–8. 279. Schäfer LV, Marrink SJ. Partitioning of lipids at domain boundaries in model membranes. Biophysical Journal. 2010;99(12):L91-L3. 280. Fuhrmans M, Marrink SJ. Molecular view of the role of fusion peptides in promoting positive membrane curvature. Journal of the American Chemical Society. 2011;134(3):1543–52.

Molecular Dynamics of Lipid Bilayers

97

281. Risselada HJ, Marelli G, Fuhrmans M, Smirnova YG, Grubm̈ller H, Marrink SJ, M̈ller M. Linetension controlled mechanism for inluenza fusion. PLoS ONE. 2012;7(6):e38302. 282. Kucerka N, Marquardt D, Harroun TA, Nieh M-P, Wassall SR, de Jong DH, SchaÃàfer LV, Marrink SJ, Katsaras J. Cholesterol in bilayers with PUFA chains: Doping with DMPC or POPC results in sterol reorientation and membrane-domain formation. Biochemistry. 2010;49(35):7485–93. 283. Monticelli L, Kandasamy SK, Periole X, Larson RG, Tieleman DP, Marrink S-J. The MARTINI Coarse-grained force ield: Extension to proteins. Journal of Chemical Theory and Computation. 2008;4(5):819–34. 284. Marrink SJ, Fuhrmans M, Risselada HJ, Periole X, editors. The MARTINI Force Field. Boca Raton, FL: CRC Press; 2008. 285. Murtola T, Falck E, Patra M, Karttunen M, Vattulainen I. Coarse-grained model for phospholipid/cholesterol bilayer. The Journal of Chemical Physics. 2004;121(18):9156–65. 286. Wang Z-J, Deserno M. A systematically coarse-grained solvent-free model for quantitative phospholipid bilayer simulations. The Journal of Physical Chemistry B. 2010;114(34):11207–20. 287. Meyer H, Biermann O, Faller R, Reith D, Muller-Plathe F. Coarse graining of nonbonded inter-particle potentials using automatic simplex optimization to it structural properties. The Journal of Chemical Physics. 2000;113(15):6264–75. 288. Hone TD, Izvekov S, Voth GA. Fast centroid molecular dynamics: A force-matching approach for the predetermination of the effective centroid forces. The Journal of Chemical Physics. 2005;122(5):054105–7. 289. Izvekov S, Parrinello M, Burnham CJ, Voth GA. Effective force ields for condensed phase systems from ab initio molecular dynamics simulation: A new method for force-matching. The Journal of Chemical Physics. 2004;120(23):10896–913. 290. Ercolessi F, Adams JB. Interatomic potentials from irst-principles calculations: The force-matching method. EPL (Europhysics Letters). 1994;26(8):583. 291. Noid WG, Chu J-W, Ayton GS, Voth GA. Multiscale coarse-graining and structural correlations: Connections to liquid-state theory. The Journal of Physical Chemistry B. 2007;111(16):4116–27. 292. Krishna V, Noid WG, Voth GA. The multiscale coarse-graining method. IV. Transferring coarse-grained potentials between temperatures. The Journal of Chemical Physics. 2009;131(2):024103–12. 293. Noid WG, Chu J-W, Ayton GS, Krishna V, Izvekov S, Voth GA, Das A, Andersen HC. The multiscale coarse-graining method. I. A rigorous bridge between atomistic and coarse-grained models. The Journal of Chemical Physics. 2008;128(24):244114–11. 294. Noid WG, Liu P, Wang Y, Chu J-W, Ayton GS, Izvekov S, Andersen HC, Voth GA. The multiscale coarsegraining method. II. Numerical implementation for coarse-grained molecular models. The Journal of Chemical Physics. 2008;128(24):244115–20.

5

New Insights into the Peptide–Membrane Partitioning Equilibrium from In Silico Free Surfaceto-Bilayer Peptide Insertion Jakob P. Ulmschneider

CONTENTS 5.1 Introduction ............................................................................................................................99 5.2 Determining Partitioning Properties via MD Simulation .................................................... 101 5.3 Length-Dependent Polyleucine Partitioning ........................................................................ 101 5.4 Insertion Free Energy and the Partitioning Equilibrium...................................................... 104 5.5 Role of Lipid Shape on the Peptide Partitioning Equilibrium .............................................. 107 5.6 Outlook ................................................................................................................................. 108 References ...................................................................................................................................... 108

5.1 INTRODUCTION Determination of the insertion energetics of transmembrane (TM) α-helices into membranes has proved dificult. The chief challenge is to overcome the tendency of nonpolar helices to aggregate and precipitate out of aqueous solution.1,2 So far this has been unsuccessful. Numerous alternative experimental and computational approaches have been presented over the last decades to obtain closely related transfer properties.3 Several approaches in particular have provided estimates of the energetics of protein insertion and stability (Figure 5.1). The irst is based on recent in vitro experiments using the Sec61 translocon (Figure 5.1a). Cells have conquered aggregation by means of the translocon machinery, consisting primarily of the SecY complex of membrane proteins in bacteria and archaea and the highly homologous Sec61 complex in eukaryotes. The SecY/Sec61 translocons receive nascent membrane chains directly from the ribosome and guide their insertion into the membrane cotranslationally. All available evidence suggests that the TM segments partition between the translocon complex and the lipid bilayer following physicochemical principles.4−6 The code, in the form of a biological hydrophobicity scale, is highly correlated with physical hydrophobicity scales determined, for example, from measurements of the partitioning of amino acids between water and n-octanol.7 Another recent set of experiments has used the folding and refolding capability of outer membrane phospholipase A (OmpLA) as a scaffold to determine the relative transfer free energies of amino acids into lipid bilayers (Figure 5.1b).8 Although the broad energetics are similar, quantitatively these two scales have proved dificult to reconcile with other hydrophobicity scales as well as to generalize for arbitrary sequences. The key dificulty lies in interpreting the partitioning properties in the absence of detailed structural 99

100

Liposomes, Lipid Bilayers and Model Membranes

(b) (a)

Peptide chain

ΔG0W, I Enter

Ribosome

OmpLa

ΔGapp

Translocon

ΔΔGOmpLA

Exit (c) Free energy

W

W

(d) ΔG‡ S

W

WU ΔGS→TM

ΔGW→S

Membrane normal TM

TM

S

SU

S

ΔGS→TM

FIGURE 5.1 (See color insert.) Schematic depiction of the partitioning processes studied to quantify the interaction of peptide sequences with lipid bilayers. (a) The partitioning equilibrium probed by the transloconmediated insertion experiments is presumably between translocon pore and bilayer (ΔGapp), while the entry of the peptide into the translocon (“Enter”), and subsequent secretion (“Exit”) are thought to be nonequilibrium processes. (b) The outer membrane phospholipase A (OmpLA) can fold reversibly into synthetic liposomes in vitro. Host–guest experiments using OmpLA as a scaffold yield relative water-to-bilayer transfer free energies (ΔΔG OmpLA) of amino acid side chains on the protein surface. (c) Direct partitioning simulations of freely inserting, suficiently hydrophobic peptides reveal the equilibrium between surface adsorbed (S) and transmembrane inserted (TM) states. No soluble state (W) exists for these peptides, which unfold in water and precipitate out of solution. (d) Partitioning simulations of polar and charged sequences reveal the equilibrium between water soluble and interfacially adsorbed orientations. Peptide aggregation is not considered in this scheme.

knowledge of the proteins or peptides studied and their dynamic interactions with the lipid bilayer environment. In particular, the translocon assay utilizes a complex cellular machinery, whose mechanics and thermodynamic processes are only poorly understood at present,9 and the OmpLA assay lacks the protein backbone contribution. In addition, the noninserted reference state is not known for either experiment. Therefore, currently there is no reliable data set that allows prediction of transfer energetics for arbitrary peptides and provides an atomic detail understanding of the reasons behind the partitioning properties. As discussed by Schow et al.,9 a quantitative comparison is necessary for completing our understanding of the translocon-to-bilayer partitioning process, and for connecting membrane protein stability to translocon-guided membrane protein assembly. To circumvent the experimental

New Insights into the Peptide–Membrane Partitioning Equilibrium

101

challenges of partitioning TM segments across lipid membranes, a recent computational approach is to use molecular dynamics (MD) simulations carried out in the microsecond time regime.10−13 Such µs-scale equilibrium MD simulations can measure and quantify the transfer properties of monomeric peptides into lipid bilayers (Figure 5.1c,d): sequences are allowed to transition spontaneously into and out of the lipid bilayer membrane, thus providing the true thermodynamic partitioning equilibrium. The strength of this partitioning approach is that all states populated at equilibrium are directly detected, and the free energy between them is obtained from their relative occupancies. These simulations allow to compare direct peptide partitioning with translocon-to-bilayer partitioning.

5.2 DETERMINING PARTITIONING PROPERTIES VIA MD SIMULATION Microsecond-length partitioning simulations can now be run routinely in 1–2 months on a single modern Intel/AMD CPUs. The typical simulation setup consists of all-atom MD simulations carried out in palmitoyloleoyl-phosphatidylcholine (POPC) lipid bilayers, with a length of 1–2 µs per run. Simulations are generally performed at elevated temperatures (T ≥ 80°C), as this greatly speeds up sampling.12 This is possible since the thermodynamic properties of the systems appear independent of temperature, even for very hot systems (>200°C), as will be elaborated further below. This approach directly reveals all states populated at equilibrium. For hydrophobic peptides, only surface bound and TM inserted helices are observed. The free energy of insertion is directly obtained from the relative occupancy of these states. The irst sequences to be studied in this way have been a set of polyleucine peptide (Ln) constructs that have previously been investigated experimentally.14 These polyleucine constructs have the overall sequence: (i) unlanked acetyl-(L)n-amide constructs (Ln), and (ii) lanked acetyl-GGPG-(L)n-GPGG-amide peptides (GLn), with n = 5–12. The GLn sequences were used in the translocon assay, with the GGPG lanks serving as helix breakers, insulating the polyleucine “guest” segments from the host sequence.14 Changing the length of the peptide leads to the partitioning equilibrium being shifted from TM inserted to noninserted. In the translocon experiments, the insertion free energy as a function of peptide length n can be itted to a simple linear function ∆G(n) = n ⋅ ∆GLeu + ∆G0 , indicative of a two-state equilibrium model. The free partitioning of hydrophobic peptides into lipid bilayers is schematically illustrated in Figure 5.1c. The two principal states are a surface-bound helix (S) and a transmembrane inserted helix (TM). Water-solvated states (W) are much higher in free energy and not populated at equilibrium. This is consistent with experiments that show these peptides precipitate out of solution.1,2 Thus, as demonstrated further below, the partitioning for these peptides takes place between S and TM states, rather than between water and TM. For comparison, Figure 5.1a depicts the process assumed to be probed by the translocon-mediated insertion experiments.4−6 Recent simulations as well as experiments indicate this likely represents an equilibrium partitioning process of peptides between translocon channel and bilayer, rather than between water-soluble and TM states (see discussion below).3,9,15−18 In this model, the entry of the peptide into the translocon (‘Enter’) and the translocation (‘Exit’) is energetically driven and nonequilibrium.

5.3

LENGTH-DEPENDENT POLYLEUCINE PARTITIONING

The typical peptide insertion pathway observed in the simulations is shown in Figure 5.2. All peptides were initially unfolded and placed into bulk water about 15 Å from the bilayer surface (W). Rapid adsorption (U) at t = 5–10 ns, consistent with insoluble hydrophobic peptides, is followed by interfacial folding into a surface-bound helix (S). Subsequently, the peptides insert and adopt a TM helix orientation. This pathway agrees well with the thermodynamic models of White19 and Engelman,20 in which folding precedes insertion. The adsorption process was irreversible in all simulations, with no subsequent expulsion of the peptide into the water phase (W), or unfolded

102

Liposomes, Lipid Bilayers and Model Membranes L10

25

W (0 ns)

20 U (2 ns)

zCM (Å)

15

U S (40 ns)

10 5

S

0 TM (65 ns)

–5 0

25

50 Helicity (%)

75

100

TM

FIGURE 5.2 Illustration of a fast folding and adsorption process from the initial water solvated unfolded state (W). The insertion depth zCM is plotted versus the peptide helicity for the pathway taken by L10 at 80°C. Interfacial adsorption from the initial state in water occurs in ~2 ns (U). The peptide then folds (S) and subsequently inserts (TM). Only the S and TM states are observed for the remainder (1–2 µs) of the simulations.

conformations (U) observed after ~50 ns. Thus, the simulations directly conirm the theoretical scheme in Figure 5.1c. Only the α-helical S and TM states remained for the following 1–2 µs. These can be distinguished by their characteristic center-of-mass position along the membrane normal (zCM) and the helix tilt angle (θ). The TM state is deeply buried in the center of the bilayer and aligned along the membrane normal (zCM ≈ 0 Å, θ ≈ 10°), whereas the S helix is parallel to the membrane surface (zCM ≈12 Å, θ ≈ 90°). The density proile of the S-state (Figure 5.3a) reveals the peptide to be deeply buried near the edge of the acyl chains, just below the glycerol/carbonyl groups (Figure 5.3a). The S position substantially deforms the monolayer surface containing the peptide, with up to ~2.5 Å local thinning through the disturbance of the lipids (Figure 5.3b). However, water is not pulled into the bilayer in signiicant numbers by the peptide, as can be seen by the water density curve that is essentially identical to that of the opposing interface. For short peptides (n ≤ 6), S conigurations dominate, while longer polyleucine segments (n ≥ 10) mainly insert to form TM helices. Peptides of intermediate lengths (n = 7–9) display an equilibrium alternating between S and TM conigurations. Depending on the temperature, there is a signiicant change in the peptide insertion and expulsion rates. This is illustrated in Figure 5.4 for GL8. As the temperature is raised from 175°C to 217°C, the peptide transits much more frequently between the TM and S states (Figure 5.4a). Average insertion and expulsion rates, kin and kout, can be computed from these simulations. The resulting kinetics are summarized (for L7, L8, and GL8) as an Arrhenius plot in Figure 5.4b. In all cases, a it of ln k versus 1/T results in a straight line, indicating a irst-order, single-barrier process. From the slope of the it, the activation enthalpies for both insertion and expulsion ΔH‡ can be estimated via ln k = –ΔH‡/RT + const. The barriers for both L7 and L8 are quite weak, with ΔH‡ ~ 5–8 kcal/mol and transition times of up to ~0.5 µs at 30°C. However, the situation is very different for GL8, with vastly increased barriers of ΔH‡ ~ 20–24 kcal/mol. It is possible to obtain the insertion and expulsion rates at 30°C by extrapolating the Arrhenius plot. This gives τ ≈ 9–226 ms, roughly ~106 times slower than for L8, which is beyond the timescales that can currently be reached in simulations. The cause of the slower rate for GL8 is the barrier for translocating one GPGG lanking tetrapeptide that strongly resists burial in the membrane interior. As the lanking sequence tends to be unfolded, this would expose unpaired backbone hydrogen bonds to the hydrophobic membrane interior, which is energetically highly disfavored. These results however show that it is now possible to directly study peptide partitioning via increasing the temperature. How reliable is this approach? Key to its success is the lack of unfolding

New Insights into the Peptide–Membrane Partitioning Equilibrium

Membrane normal (Å)

(a)

Water

103

Total

15 Peptide 0

CH2 G/C PO4

–15

0

0.5 1 Density (g/mL) Deformation (Å)

(b) 10 y-axis (Å)

POPC

0

0 –1 –2 –3

–10

–10

0 10 x-axis (Å)

FIGURE 5.3 (See color insert.) Peptide location and bilayer deformation of the interfacial surface-bound (S) state of Ln and Gln peptides: (a) the density cross-section proile of the bilayer shows that in the S state the peptide (here L7) is buried below the water interface. A representative conformer is shown to scale. The leucine sidechains (green) are chiely in contact with the acyl tails (CH2), and there is only a small overlap with the phospho-choline headgroups and carbonyl-glycerol (C/G) groups. Other peptides behave exactly similar. (b) The peptide induced distortion to the bilayer at equilibrium can be visualized by plotting the time-averaged phosphate position from the bilayer center. This shows local thinning by 5–10% for L7 as the lipid headgroups bend over the peptide to cover the termini (phosphate is represented as an orange sphere).

observed for hydrophobic sequences in lipid membranes. As shown in Figure 5.5 for L8, L12, and GL12, there is no signiicant loss of helicity ( 0.99). All curves display two-state behavior, with a transition to TM inserted conigurations for longer peptides. Figure 5.6b shows that ΔGn decreases perfectly linearly with n in both simulation and experiment. However, the

105

New Insights into the Peptide–Membrane Partitioning Equilibrium (b)

40

GL12 L8 CD: GL12

50

0 –40 180

(c) 1.5

Density (g/mL)

100

L12

T (°C) 45 65 85

80

1.0

200 220 240 Wavelength (nm)

260

20

60

0 220

Total

T (°C) 30 80 120

CH2

H 2O

P/C

0.5 G/C 0

100 140 180 Temperature (°C)

Helicity (%)

(θ) (m° cm2 dmol–1)

(a)

–20

CH3 0

–20 20 Membrane normal (Å)

0

20

FIGURE 5.5 (See color insert.) Thermostability of the peptide and bilayer: (a) circular dichroism spectra of the secondary structure of GL12 in POPC vesicles (peptide/lipid ratio = 1/100) over a temperature range of 45–85°C. The spectra indicate predominantly helical conformers and display low sensitivity to temperature. (b) The thermostability of the peptides observed in the MD simulations is comparable (shown are L8, L12, and GL12; all other peptides behave similar). Shorter peptides are marginally less helical due to terminal fraying. (c) The effect of heating on the lipid bilayer can be visualized by plotting the equilibrium trans-bilayer density proiles in the presence of peptide (here L8) for temperatures in the range 30–120°C (dark to light colors). Comparison of the principal structural groups (CH3 = methyl, CH2 = acyl tails, P/C = phosphocholine headgroups, G/C = carbonyl-glycerol linker, H2O = water) shows temperature-induced broadening of the Gaussians and a slight decrease in the total density. However, the trans-bilayer density proile and location of the principal structural groups do not change signiicantly.

offset and slope vary slightly, relecting a shift of the MD insertion probability curve toward shorter peptides by ~2.4 leucine residues, corresponding to a ΔΔG = ΔG translocon – ΔGdirect = 1.9 ± 0.1 kcal/mol offset between the experimental and computational insertion free energies. This offset is a constant for all peptides. What is apparent from Figure 5.6 is the insensitivity of the insertion curves on temperature. Further 2 µs investigations for L7, L8 (30–160°C), and GL8 (147–217°C) conirmed this intriguing behavior, and the resulting values of pTM and ΔG S→TM are shown in Figure 5.7. The observed thermodynamic behavior of the partitioning process best its a model in which ΔG(T)S→TM ≈ const. and pTM = [1 + exp(– βΔG S→TM)]−1, as illustrated in Figure 5.7. Using ΔG S→TM = ΔHS→TM – TΔSS→TM and assuming only a small explicit dependence of ΔH and ΔS on T, this implies that ΔSS→TM ≈ 0. Why do entropic effects play only a minor role? First, the peptides do not unfold, and thus there are no entropic contributions from conformational changes. Second, the entropic penalty ΔSimmobilize on immobilizing the peptide inside the membrane24 arising from the restriction of the rigid-body rotational (e.g., tilting) and translational motions (e.g., diffusion along the membrane normal) is identical for the S and TM states: for example, the center of mass and tilt angle luctuations of GL8 are ±1.9 Å/ ± 10.1° for the TM, and ±2.1 Å/ ± 11.1° for the S state. While a model where ΔSS→TM ≈ 0

106

Liposomes, Lipid Bilayers and Model Membranes (a)

(c) 100 TM Ln 30°C Ln 80°C Ln 120°C Ln 160°C GLn 217°C Experiment

Translocon Ribosome 4 (b)

6

12

25 0 14 4

(d)

S

8 10 Leucines (#)

50

PTM (%)

75

ΔG* (exp.) ΔG** (pred.) 2 ΔG (fit exp.) ΔG (fit comp.)

TM

0 –2

ΔG (kca/mol)

H Peptide chain

–4 4

6

8 10 Leucines (#)

12

14

FIGURE 5.6 (See color insert.) Bilayer insertion eficiency and transfer free energy as a function of peptide length n. (a) The experimental values are for translocon-mediated insertion into dog pancreas rough microsomes of GGPG-(L)n -GPGG constructs embedded into the leader peptidase carrier sequence.14 (b) The computed values are for spontaneous partitioning of Ln peptides into POPC lipid bilayers at 30–160°C, and for GGPG-(L)n -GPGG at 217°C. (c) Both measurements display perfect two-state Boltzmann behavior (R2 > 0.99), with a transition in the native state from surface bound to TM inserted upon lengthening of the peptide. (d) This is relected in the free energy of insertion ΔG(n) as a function of peptide length n (insertion for negative ΔG—shaded). The straight lines indicate the two-state Boltzmann it, while the data points show the computed (red, green) and experimental (blue) values for the individual peptides (*measured ΔG,5 peptide IDs: 43 & 380–383; *predicted ΔG, http://dgpred.cbr.su.se/).

1 (kcal/mol)

50

TM

pTM (%)

100

0

∆GS→

L7 L8 GL8

0

–1

–2 20

100 Temperature (°C)

180

20

100 Temperature (°C)

180

FIGURE 5.7 Temperature dependence of the insertion propensities pTM and the transfer free energies ΔG S→TM. Overall insertion propensities for L7, L8, and GL8 as a function of temperature. Error bars are derived from block averaging (10 blocks). (D) The corresponding free energies of insertion ΔG S→TM appear to show no systematic variation with temperature.

New Insights into the Peptide–Membrane Partitioning Equilibrium

107

its the data well, assuming ΔHS→TM ≈ const. neglects changes in heat capacity. It is likely that small nonvanishing entropic contributions exist, and both ΔH(T) and ΔS(T) are temperature dependent. The results show that the lanking sequences used in the translocon do not affect the insertion propensity, with almost identical results for (L)n and GGPG-(L)n-GPGG. A possible explanation arises from the structural properties of the lanks, which were speciically designed to remain unfolded through the use of Gly and Pro residues.4 The high cost of partitioning unmatched polar backbone groups into the hydrophobic core prohibits insertion. Consequently, the lanking sequences remain in the polar lipid headgroup region throughout the simulations, either with both lanks at the same interface (S state), or one on each side of the bilayer (TM state). The net contribution to Ln GLn the insertion propensities is ∆∆GSflanks → TM ≈ 0 ( ∆GS→ TM ≈ ∆GS→ TM ). Interestingly, similar observations were made in the translocon experiments: adding additional glycines to the GPGG lanks had an almost negligible effect, with free energy shifts of ΔΔG = +0.18 kcal/mol for GPGGG to −0.17 kcal/mol for GPGGGGGG. Substituting all Gly residues with Asn led to only a small increase flanks ≈ 0. Flanking sequences are not inserted into the bilayer of ΔΔG = +0.5 kcal/mol. Hence, ∆Gexp core, and thus do not contribute to ΔG S→TM. Flanking sequences have, however, a dramatic effect on the insertion barriers. Barriers are ΔH‡S↔TM = 20–24 kcal/mol per GPGG lank or ~5–6 kcal/mol per lanking residue. This compares well with previous estimates for the transfer of an unmatched hydrogen bonding pair into alkane of ~6.4 kcal/mol.22 Without lanks, the peptide partitioning barrier is signiicantly reduced to ΔH‡S↔TM = 5.5–8.1 kcal/mol. What is the reason for the shift between translocon mediated and free partitioning seen in Figure 5.6? More or less 2 fewer leucines are required for the peptides to insert on their own than predicted by the translocon scale. The explanation is that two different partitioning processes are involved. In contrast to partitioning between bilayer interface and hydrocarbon core, the translocon crystal structure suggests that hydrophobic peptides are released laterally from the protein conducting channel into the membrane.25 This means that the experiments measure the partitioning of peptides between translocon channel and bilayer, rather than between water-soluble and TM conigurations (Figure 5.1b).3,9,16−18 Experimental support for this interpretation comes from a recent mutagenesis study by Junne et al. in which an increase in polarity of the Sec61 translocon protein conducting channel was found to reduce drastically the minimal peptide hydrophobicity required for membrane insertion.15 This is consistent with a translocon-to-bilayer equilibrium being at the heart of the translocon-mediated insertion probability. Recently, Gumbart et al. have performed simulations that seem to conirm this view:26 The transfer free energies from the translocon to the membrane were found to be signiicantly smaller than those of water-to-bilayer transfer, and more in line with the translocon experiments. Thus, the close correlation of our results indicates that the surface-bound helical state of the peptides is located in a region of similar hydrophobicity to that of the internal translocon pore. This suggests that the equilibria of spontaneous partitioning and translocon-mediated insertion are likely independent, with no thermodynamic cycle connecting the two insertion paths, as recently discussed.9 Both differ highly from the much larger free energy changes involved in nonequilibrium water-to-bilayer partitioning.

5.5 ROLE OF LIPID SHAPE ON THE PEPTIDE PARTITIONING EQUILIBRIUM For a given sequence, the partitioning equilibrium is a fundamental property of the lipid bilayer composition, and expected to vary with lipid shape, mixture, and charge. Figure 5.8 reveals the shift of the insertion free energy as a function of the hydrophobic width of the bilayer (different lipid acyl lengths), and the level of lipid tail saturation. ΔG is very similar for POPC and DPPC, indicating that the acyl chain saturation plays only a minor role. However, the partitioning equilibrium is shifted by −0.9 kcal/mol toward the TM state in a DMPC bilayer, demonstrating that bilayer thickness greatly affects the insertion free energy. This effect is caused by the energetic cost of membrane deformation, which is lower in thinner membranes as the hydrophobic mismatch is reduced.14,27 However, as the thickness of the POPC bilayer is comparable to that of the ER membrane,14,28,29 this is unlikely to be the cause of the shift between experimental and simulated insertion.

108

Liposomes, Lipid Bilayers and Model Membranes (b) 0.5

10

POPC ∆G (kcal/mol)

Bilayer deformation (Å)

(a)

5

0

5

10 # Leucines

15

0

DPPC

–0.5 –1 –1.5

DMPC 10 12 14 Hydrophobic mismatch (Å)

FIGURE 5.8 The role of bilayer thickness and deformation on shifting the partitioning equilibrium. (a) Peptide-induced bilayer deformation for GLn (n = 6 − 16) sequences in their TM-inserted orientation. The deformation rises with increased negative hydrophobic mismatch. (Inset: GL8, the thick line indicating the average position of the phosphate groups). (b) The partitioning equilibrium is shifted toward the TM orientation upon decreasing the bilayer thickness due to a reduced hydrophobic mismatch.

5.6 OUTLOOK Obtaining partitioning properties via direct partitioning simulations constitutes a simple, eficient, and general tool to determine single molecule partitioning properties as well as transfer kinetics of peptides into lipid bilayers. One of the exciting future applications of this simulation method will be to determine the exact code for the peptide partitioning free energy as a function of membrane topology, such as the charge state, lipid mixture ratios, and lipid shape. It will also greatly aid the investigations of the insertion mechanism of membrane active peptides (e.g., antimicrobials, cellpenetrating, and fusion peptides), and de novo membrane protein structure prediction via ab initio folding–partitioning and assembly simulations.

REFERENCES 1. Ladokhin, A. S., White, S. H. Interfacial folding and membrane insertion of a designed helical peptide. Biochemistry 2004, 43(19), 5782–5791. 2. Wimley, W. C., White, S. H. Designing transmembrane alpha-helices that insert spontaneously. Biochemistry 2000, 39(15), 4432–4442. 3. White, S. H. Membrane protein insertion: The biology-physics nexus. J. Gen. Physiol. 2007, 129(5), 363–369. 4. Hessa, T., Kim, H., Bihlmaier, K., Lundin, C., Boekel, J., Andersson, H., Nilsson, I., White, S. H., von Heijne, G. Recognition of transmembrane helices by the endoplasmic reticulum translocon. Nature 2005, 433(7024), 377–381. 5. Hessa, T., Meindl-Beinker, N. M., Bernsel, A., Kim, H., Sato, Y., Lerch-Bader, M., Nilsson, I., White, S.  H., von Heijne, G. Molecular code for transmembrane-helix recognition by the Sec61 translocon. Nature 2007, 450(7172), 1026–1030. 6. Hessa, T., White, S. H., von Heijne, G. Membrane insertion of a potassium-channel voltage sensor. Science 2005, 307(5714), 1427. 7. Wimley, W. C., Creamer, T. P., White, S. H. Solvation energies of amino acid side chains and backbone in a family of host-guest pentapeptides. Biochemistry 1996, 35(16), 5109–5124. 8. Moon, C. P., Fleming, K. G. Side-chain hydrophobicity scale derived from transmembrane protein folding into lipid bilayers. Proc. Natl. Acad. Sci. USA 2011, 108(25), 10174–10177. 9. Schow, E., Freites, J., Cheng, P., Bernsel, A., von Heijne, G., White, S., Tobias, D. Arginine in membranes: The connection between molecular dynamics simulations and translocon-mediated insertion experiments. J. Mem. Biol. 2011, 239(1), 35–48. 10. Ulmschneider, J., Andersson, M., Ulmschneider, M. Determining peptide partitioning properties via computer simulation. J. Mem. Biol. 2011, 239(1), 15–26.

New Insights into the Peptide–Membrane Partitioning Equilibrium

109

11. Ulmschneider, J. P., Smith, J. C., White, S. H., Ulmschneider, M. B. In silico partitioning and transmembrane insertion of hydrophobic peptides under equilibrium conditions. J. Am. Chem. Soc. 2011, 133(39), 15487–15495. 12. Ulmschneider, M. B., Doux, J. P. F., Killian, J. A., Smith, J., Ulmschneider, J. P. Mechanism and kinetics of peptide partitioning into membranes. J. Am. Chem. Soc. 2010, 132, 3452–3460. 13. Ulmschneider, M. B., Smith, J. C., Ulmschneider, J. P. Peptide partitioning properties from direct insertion studies. Biophys. J. 2010, 98, L60–L62. 14. Jaud, S., Fernandez-Vidal, M., Nilsson, I., Meindl-Beinker, N. M., Hubner, N. C., Tobias, D. J., von Heijne, G., White, S. H. Insertion of short transmembrane helices by the Sec61 translocon. Proc. Natl. Acad. Sci. USA 2009, 106(28), 11588–11593. 15. Junne, T., Kocik, L., Spiess, M. The hydrophobic core of the Sec61 translocon deines the hydrophobicity threshold for membrane integration. Mol. Biol. Cell 2010, 21(10), 1662–1670. 16. Von Heijne, G. Formation of transmembrane helices in vivo—Is hydrophobicity all that matters? J. Gen. Physiol. 2007, 129(5), 353–356. 17. White, S. H., von Heijne, G. Do protein-lipid interactions determine the recognition of transmembrane helices at the ER translocon? Biochem. Soc. Trans. 2005, 33(Pt 5), 1012–1015. 18. White, S. H., von Heijne, G. Transmembrane helices before, during, and after insertion. Curr. Opin. Struct. Biol. 2005, 15(4), 378–386. 19. White, S. H., Wimley, W. C. Membrane protein folding and stability: Physical principles. Annu. Rev. Biophys. Biomol. Struc. 1999, 28, 319–365. 20. Engelman, D. M., Chen, Y., Chin, C.-N., Curran, A. R., Dixon, A. M., Dupuy, A. D., Lee, A. S., Lehnert, U., Matthews, E. E., Reshetnyak, Y., K., Senes, A., Popot, J.-L. Membrane protein folding: Beyond the two stage model. FEBS Lett. 2003, 555, 122–125. 21. Ladokhin, A. S., White, S. H. Folding of amphipathic α-helices on membranes: Energetics of helix formation by melittin. J. Mol. Biol. 1999, 285, 1363–1369. 22. Ben-Tal, N., Sitkoff, D., Topol, I. A., Yang, A.-S., Burt, S. K., Honig, B. Free energy of amide hydrogen bond formation in vacuum, in water, and in liquid alkane solution. J. Phys. Chem. B 1997, 101(3), 450–457. 23. Pabst, G., Katsaras, J., Raghunathan, V. A. Enhancement of steric repulsion with temperature in oriented lipid multilayers. Phys. Rev. Lett. 2002, 88(12), 128101. 24. Ben-Tal, N., Ben-Shaul, A., Nicholls, A., Honig, B. Free-energy determinants of α-helix insertion into lipid bilayers. Biophys. J. 1996, 70(4), 1803–1812. 25. Van den Berg, B., Clemons, W. M., Jr., Collinson, I., Modis, Y., Hartmann, E., Harrison, S. C., Rapoport, T. A. X-ray structure of a protein-conducting channel. Nature 2004, 427, 36–44. 26. Gumbart, J., Chipot, C., Schulten, K. Free-energy cost for translocon-assisted insertion of membrane proteins. Proc. Natl. Acad. Sci. USA 2011, 108(9), 3596–3601. 27. Krishnakumar, S. S., London, E. Effect of sequence hydrophobicity and bilayer width upon the minimum length required for the formation of transmembrane helices in membranes. J. Mol. Biol. 2007, 374(3), 671–687. 28. Mitra, K., Ubarretxena-Belandia, I., Taguchi, T., Warren, G., Engelman, D. M. Modulation of the bilayer thickness of exocytic pathway membranes by membrane proteins rather than cholesterol. Proc. Natl. Acad. Sci. USA 2004, 101(12), 4083–4088. 29. Gawrisch, K., Gaede, H., Mihailescu, M., White, S. Hydration of POPC bilayers studied by 1H-PFGMAS-NOESY and neutron diffraction. Eur. Biophys. J. 2007, 36(4), 281–291.

6

Basic Aspects and Applications of Lipids and Protein Dynamics Maikel C. Rheinstädter

CONTENTS 6.1 Introduction .......................................................................................................................... 111 6.2 Membrane Diffusion Studied in Single Supported Bilayers ................................................ 113 6.3 Determination of Membrane Elastic Properties by Inelastic Scattering Techniques ........... 115 6.4 Protein–Protein Interaction in Purple Membranes............................................................... 118 6.5 Concluding Remarks ............................................................................................................ 121 References ...................................................................................................................................... 122

6.1 INTRODUCTION The understanding of dynamics and functioning of biological membranes is one of the greatest challenges in modern biology and biophysics. Few experimental techniques can access dynamical properties in biological materials on the nanometer scale, and resolve the dynamics of proteins, lipid, and hydration water molecules, and the interaction between them. In this context, inelastic neutron scattering turned out to be a very powerful tool to study dynamics and interactions in biomolecular materials up to relevant nanosecond timescales and down to the nanometer length scale. This chapter reviews and discusses inelastic neutron scattering experiments to study lipid diffusion, membrane elasticity, and protein–protein interactions of transmembrane proteins. High-tech life sciences include the emerging biotechnology and biomedical device industries, functional foods, and nutraceuticals; however, they also include the development of new biomaterials and pharmaceuticals. Even though biological membranes were studied for decades, very few biologically relevant processes were revealed on a molecular level. The reason is the combination of very small nanometer length scales and very fast dynamics of pico- and nanoseconds, which pose particular experimental challenges. Dynamical properties are often less well understood, but are important for many fundamental biomaterial properties such as elastic properties and interaction forces. They also determine or strongly affect certain functional aspects, such as diffusion and transport through a membrane, and are essential for protein function. The proper functioning of membrane proteins depends on membrane composition and physical properties, such as its elastic properties and hydrophobic thickness. Dynamic processes in complex biological membranes also involve interactions between the membrane’s different constituents, such as lipids, cholesterol and proteins (Rheinstädter et al. 2008, 2009). Neutron and x-ray scattering can be used as a microscope to study the structure and dynamics in these systems as they give access to the relevant length and timescales. Biological materials can be thought of as “multiscale” materials, due to the fact that relevant dynamics takes place over extended length and timescales (Frauenfelder et al. 1991, Fenimore et al. 2004, Bayerl 2000, Rheinstädter et al. 2006b). To address this multiscale behavior experimentally, different techniques must be applied. Figure 6.1 depicts the length and timescales accessible by high-speed atomic force microscopy (AFM), inelastic neutron scattering, inelastic x-ray scattering, dynamic light scattering (DLS), Brillouin and Raman’s scattering, and dielectric spectroscopy. 111

112

Liposomes, Lipid Bilayers and Model Membranes Scattering vector (Å–1) 10–2

Raman

Brillouin

DLS

Macroscopic

104

10–1

100

101

102 104 103 102 Inelastic x-ray ng scattering i 101 r te cat s 100 n tro eu 10–1 n ic t s la 10–2 Ine 10–3 10–4 10–5 10–6 10–7 10–8 10–9 Mesoscopic High speed Macroscopic 10–10 atomic force 10–11 microscopy

103

102 101 Length scale (Å)

100

Energy (meV)

10–3 10–2 10–1 100 101 102 103 104 105 106 107 108 109 1010 1011 1012

10–3

Dielectric spectroscopy

Time (ps)

10–4

10–1

FIGURE 6.1 (See color insert.) Accessible length and timescales, and corresponding energy and momentum transfer, for some spectroscopic techniques covering microscopic to macroscopic dynamics. Light scattering techniques include Raman, Brillouin, and DLS. Inelastic x-ray and neutron scattering access dynamics on nanometer length scales. Dielectric spectroscopy probes the length scale of an elementary electric dipole, which can be estimated by the bond length of a C–O bond to about 140 pm. The area marked by the dashed box is the dynamical range accessible by computer simulations. High-speed AFM is an emerging technique, which allows imaging in real space. (Adapted from Armstrong, C.L., Sandqvist, E., and Rheinstädter, M.C. 2011a. Protein Pept. Lett. 18:344–353.)

The relevant length scale for dielectric spectroscopy is in the order of an elementary molecular electric dipole, which can be estimated by the bond length of a C–O bond to about 140 pm, and frequencies from kilohertz to gigahertz can be measured. Since the wavelength of the probe is usually around λ ~ 510 nm or λ ~ 632 nm, light scattering techniques are limited to small momentum transfers of about 10−4 –10−3 Å−1, corresponding to a length scale of about 100 nm. Inelastic neutron and x-ray scattering access length scales from smaller than angstrom to more than 100 nm and timescales from picoseconds to almost 1 µs. High-speed AFM has combined a high spatial resolution of about 5 Å with a time resolution of milliseconds (Voïtchovsky et al. 2009). In recent years, molecular dynamics (MD) simulations have become an invaluable tool in developing models for molecular structure and dynamics in membranes and proteins. Because of the ever-increasing computing power and optimized algorithms, large complex systems (i.e., many hundreds of molecules) and long simulation times can now be addressed (e.g., Smith 1991, Hayward and Smith 2002, Tarek et al. 2001, Tarek and Tobias 2002, Wood et al. 2007, Meinhold et al. 2007, Marrink et al. 2007). The dashed rectangle in Figure 6.1 marks the dynamic range currently accessed by computer simulations—the elementary timescale for simulations is in the order of femtoseconds. In contrast to other spectroscopic techniques, inelastic neutron scattering results in wave vectorresolved access to molecular dynamics. For example, excitation frequencies and relaxation rates are measured at different internal length scales of the system. A typical dynamic scattering experiment measures (Q, ħω) pairs, resulting in a frequency along with a corresponding length scale, and possibly, a corresponding direction, for example, parallel or perpendicular to a protein’s axis. This additional information is of paramount importance when it comes to relating dynamical information

Basic Aspects and Applications of Lipids and Protein Dynamics

113

to structure. In short, the suite of inelastic instruments used to study soft and biologically relevant materials consists of time-of-light, backscattering, triple-axis, and spin-echo spectrometers (Teixeira et al. 2008, Rheinstädter et al. 2006b,c). The ield is currently boosted by interdisciplinary interest in life sciences. New developments and improvements in scattering instrumentation, sample preparation and environments, and, eventually, the increasingly powerful sources give quantitative access to molecular properties of the bilayers on lateral length scales between micrometers down to a few angstroms. Examples for the application and beneit of inelastic scattering will be illustrated below by detailing lateral diffusion in membranes (Armstrong et al. 2010), by the determination of membrane material properties (Rheinstädter et al. 2006a), and by the determination of protein–protein interaction between transmembrane proteins (Rheinstädter et al. 2009). All studies were conducted utilizing reconstituted lipid or native membranes oriented on solid supports. As such, lateral dynamics in the plane of the membranes can be directly accessed by aligning the membranes and measuring the in-plane component of the momentum transfer, q‖.

6.2 MEMBRANE DIFFUSION STUDIED IN SINGLE SUPPORTED BILAYERS Despite their potential relevance to bioengineering applications, such as biosensors and surface functionalization (Sackmann 1996, Tanaka and Sackmann 2005), dynamic neutron scattering experiments in single membranes have been limited in the past, as the minimal amount of sample material in a single bilayer results in a low scattering signal. Recent developments in neutron scattering instruments, and increasingly powerful neutron sources, now make it possible to observe dynamics in single bilayers. This enables studies of biologically relevant dynamics at interfaces and surfaces; a prerequisite for the development of smart, functional bioinspired surface coatings. Among other techniques, quasielastic neutron scattering (QENS) is an important tool to study nanoscale dynamics and diffusion on a nanometer length scale. The following example documents results obtained from the corresponding spectra measured on the neutron backscattering spectrometer BASIS at the Spallation Neutron Source (SNS, Oak Ridge, TN). Single bilayers of dimyristoylphosphatidylcholine (DMPC) were prepared on silicon wafers (Armstrong et al. 2010). To increase the scattering signal, 100 wafers (resulting in 200 single bilayers) were mounted horizontally into the spectrometer, such that the scattering was sensitive to the lateral diffusion of the lipid molecules, as shown in Figure 6.2a. Figure 6.2b and c depict exemplary neutron spectra at in-plane q‖ values of q‖ = 0.9 and 1.3 Å−1, respectively. The total scattering consists of a narrow central component whose width is due to instrumental resolution, and a quasi-elastic broadening between about 1 and 24 µeV due to the lateral diffusion of the lipid molecules. The quasi-elastic broadening follows a Lorentzian function. There are different models used to describe diffusion (Bée 1988). For a particle diffusing via random Brownian motion, the displacement of the particle can be characterized as a function of time (Armstrong et al. 2011b) σd =

(6.1)

2Dt ,

where D is the translational diffusion coeficient of the system. With this characteristic length scale, one can deine a self-time-dependent pair-correlation function for incoherent scattering Fd (r , t ) = (4Dt )3 2 e − r

2

( 4 Dt )

,

(6.2)

which is a solution of Fick’s law ∂Fd (r , t ) = D∇ 2 Fd (r , t ). ∂t

(6.3)

114

Liposomes, Lipid Bilayers and Model Membranes 0.6 Å–1

(a)

2.0 Å–1

(c) q|| = 0.9 Å–1

Neutron counts (arb. units)

Neutron counts (arb. units)

(b)

102

101 –80

(d) 30

–40

0 40 w (meV)

80

q|| = 1.3 Å–1

102

101 –80

–40

0 40 w (meV)

80

Diffusion constant = 60 × 10–12 m2/s

FWHM

25 20 15 10 5 0

0.5

1

1.5 2 q||2 = (Å–2)

2.5

3

3.5

FIGURE 6.2 (See color insert.) (a) Single solid-supported bilayer aligned with the plane of the membrane in the scattering plane of the neutron spectrometer. (b) and (c) Spectra for q‖ = 0.9 and 1.3 Å−1. The data were itted (black) with the instrumental resolution (green), a Lorentzian (yellow), and an additional broad Lorentzian (red). (d) FWHM of the Lorentzian quasi-elastic broadening component as a function of q‖2. When excluding data points around q‖ = 1 Å−1, data can be it using a linear function. Diffusion was found to be enhanced around q‖ = 1 Å−1, as indicated by the peak. The inset shows a sketch of the lipid packing that might give rise to an enhanced nearest-neighbor diffusion. (Adapted from Armstrong, C.L. et al. 2010. Soft Matter, 6:5864–5867.)

Basic Aspects and Applications of Lipids and Protein Dynamics

115

This results in an intermediate scattering function that decays exponentially 2

(6.4)

I d (q|| , t ) = e − q|| Dt , and can be Fourier transformed to give a Lorentzian incoherent scattering function Sd (q|| , ℏω ) =

 q||2 D 1  .  2 2 2 πℏ  (q|| D) + (ℏω ) 

(6.5)

For a continuous, Brownian lateral-diffusion process, the quasi-elastic broadening is, therefore, described by a Lorentzian peak shape, with a full width at half maximum (FWHM) that shows a q‖2 dependence FWHM L (q|| ) = 2ℏDq||2 ,

(6.6)

where D is the diffusion constant. Figure 6.2d depicts the FWHM of the Lorentzian, following deconvolution with the instrumental resolution, as a function of q‖2. The data were it by a straight line, making it possible to determine the diffusion constant using Equation 6.6 with a diffusion constant of D = 60 × 10−12 m2/s. However, the data deviated from the straight line around q‖ = 1 Å−1, giving rise to faster dynamics. Diffusion seems to be enhanced at the nearest-neighbor distance corresponding to the lipid molecules of about 6 Å (2π/1 Å−1). This change of character and enhanced nearest-neighbor diffusion is most likely due to the highly ordered luid phase of the lipids caused by the coninement of the defect-free bilayer on the substrate (Armstrong et al. 2010). This experiment, in particular, demonstrated the feasibility of inelastic neutron scattering experiments in single membranes. Complex membranes with different contents of membrane-embedded proteins, as well as cholesterol, can be prepared. By tuning bilayer properties, such as elasticity, the impact of membrane properties on protein function may be elucidated. While inelastic neutron scattering experiments used to be strongly limited by the availability of sample material, the sample used for the above-described study contained 30 wt.%) than pure long-chain lipid solutions (Ram and Prestegard 1988, Sanders and Prestegard 1990), thus greatly enhancing the concentration of membrane proteins. Moreover, since a biological membrane’s underlying structure is a lipid bilayer, bicellar mixtures lend themselves as a better membrane mimic for membrane-associated proteins than commonly used detergent-based substrates. One of the most attractive features of bicellar mixtures, especially for nuclear magnetic resonance (NMR) studies, is their ability to align in the presence of a strong magnetic ield. In fact, magnetically aligned lipid mixtures have been reported as early as the late 1970s (Sanders and Schwonek 1992, Forrest and Reeves 1979, 1981). It is well known that the alignment of bicellar mixtures is strongly dependent on temperature (T). If T is lower than the melting transition temperature, TM, of the longchain lipid, the mixture is in a morphology that is not capable of being aligned in the presence of an external magnetic ield. However, as T approaches, or is even slightly greater than TM, the system’s viscosity increases and it aligns in a manner that its bilayer normal (NB) is perpendicular to the external magnetic ield (M) (i.e., NB⊥M) (Forrest and Reeves 1979, 1981, Sanders and Prestegard 1990, Ram and Prestegard 1988, Vold and Prosser 1996)—lipid molecules are diamagnetic. While such an alignment restricts NB in one plane, the membranes form a powder with respect to the magnetic ield, as shown in Figure 7.5a. However, Prosser et al. found that by doping small amounts of lanthanide (paramagnetic) ions, such as Tm3+, Er3+, Yb3+, or Eu3+ into bicellar mixtures, they were able to alter the orientation of NB from ⊥M to || M, as shown in Figure 7.5b (Prosser et al. 1998, 1996). As a result, the freedom of NB is reduced to one orientation, making the system a more utile “goniometer” for NB

(a)

(b) M

NB

NB

M

NB

FIGURE 7.5 Bilayers align their normals, NB either (a) perpendicular (NB⊥M) to or (b) parallel (NB||M) with the external magnetic ield, M depending on whether or not they are doped with paramagnetic lanthanide ions—lanthanide ions result in the NB||M scenario.

132

Liposomes, Lipid Bilayers and Model Membranes

structural studies of membrane-associated proteins. It has been reported that the temperature range in which these lipid mixtures are magnetically alignable depends on the charge density of the membrane, that is, mixtures with a higher molar ratio of charged lipids or paramagnetic ions (Nieh et al. 2002). Compared with solid substrate aligned bilayers, magnetically aligned membranes provide a more biomimetic membrane environment for integral proteins. However, the requisite strong magnetic ields can only be realistically implemented using only a few characterization techniques (e.g., NMR and SANS). (It should be mentioned that a 0.9 T ixed ield device was developed by Harroun et al. for use with standard optical microscopes (Harroun et al. 2006a).) Nieh et al. also examined a lanthanide-free charged bicellar mixture and found that the presence of an oscillating shear low induced good alignment in membranes, with their NB aligning parallel to the direction of the shear (Nieh et al. 2003). Importantly, alignment persisted for a period of hours after shear low ceased, thus greatly simplifying the apparatus needed for inducing alignment. In doing so, this method of alignment may enable a number of other physical techniques to interrogate these interesting and versatile systems. In bicellar mixtures, membrane alignability is closely associated with aggregate morphology. Over the years, the so-called alignable morphology has evolved from disk-like micelles (Vold and Prosser 1996, Sanders and Schwonek 1992, Forrest and Reeves 1979, 1981, Sanders and Prestegard 1990, Ram and Prestegard 1988), to bilayered ribbons (Nieh et al. 2004, Soong et al. 2010, Harroun et al. 2005, van Dam et al. 2004), to perforated lamellae in the case of charged systems (Katsaras et al. 2005, Nieh et al. 2001)—these results were derived from a combination of cryogenic transmission electron microscopy (van Dam et  al. 2004), NMR (Gaemers and Bax 2001), and SANS studies (Harroun et al. 2005, van Dam et al. 2004, Katsaras et al. 2005). The currently accepted structural diagrams for both zwitterionic and charged bicellar mixtures, derived from SANS data, are shown in Figure 7.6 (Nieh et al. 2005). Generally speaking, and with the exception of zwitterionic systems at low Clp where multilamellar vesicles (MLVs) are observed (Figure 7.6a), bilayered micelles are found at low temperatures. In a zwitterionic system (Figure 7.6a), MLVs and ribbons, or nonswellable lamellae composed of ribbons (Nieh et al. 2001, 2005), are found at low and high Clp samples, respectively (Gaemers and Bax 2001), while further increases in temperature result in the presence of MLVs at all Clp. In charged bicellar mixtures (Figure 7.6b), as T increases, unilamellar vesicles (ULVs) and perforated lamellae that are capable of taking up water are found in low and high Clp samples, respectively (Gaemers and Bax 2001). These morphological transitions are closely related to the location of the short-chain lipid in the aggregate morphology. Presumably, segregation of the short-chain lipid from the long-chain lipid (as a result of the immiscibility between liquid

(a)

(b)

MLV

Ribbon-meshed nonswelling lamella

T

ULV Swelling perforated lamella

T Bilayered ribbon

Nanodisc

Nanodisc Clp

Clp

FIGURE 7.6 (See color insert.) Structural diagrams of (a) zwitterionic and (b) charged bicellar mixtures. The range of magnetically alignable structures is indicated by the red regions. The dashed lines represent not well deined boundaries, and the gray region in (b) consists of either bilayered micelles or ULVs, depending on the system’s charge density.

133

Lipid Diversity and Its Implications for Membrane Organization

disordered and gel phases) favors structures with a larger total circumference, as the short-chain lipid, which has a larger spontaneous curvature, is able to stabilize the high curvature edge. Besides aligning membrane proteins, bicellar mixtures have also been used as substrates to crystallize membrane-associated proteins. For example, Fahama and Bowie successfully used a bicellar mixture to crystallize bacteriorhodopsin extracted from Halobacterium salinarum. They took advantage of the inherent low-viscosity solution formed by bilayered micelles at low T, which allowed the application of general screening methods (Faham and Bowie 2002). For further details regarding this method of crystallizing proteins, the reader is referred to the review by Ujwal and Bowie (2011), while other applications of bicellar mixtures, as studied by NMR and other spectroscopic techniques, are summarized in Table 7.1.

TABLE 7.1 Composition of Bicellar Mixtures and Their Applications Long Chain/Short Chain DMPC/DHPC

Applications Structural determination of membrane-associated molecules by NMR Aligning water-soluble proteins Protein crystallization Separation and sensor devices Application to skin

DMPC/CHAPS

Structural determination of membrane-associated molecules

DMPC/CHAPSO

Structural determination of membrane-associated molecules

DPPC/DHPC

DLPC/DHPC DLPC/CHAPSO POPC/DHPC DPC/SDS DMLPC/DHPC DIOMPC/DIOHPC SM/DHPC TBBPC/DHPC

Protein crystallization Structural determination of membrane-associated molecules Protein/drug carrier Application to skin Carbon nanotube assembly Structural determination of membrane-associated molecules

Structural characterization of bicellar mixtures

References Andersson and Mäler (2002), Zandomeneghi et al. (2003), Marcotte and Auger (2005), Prosser et al. (2006) Ottiger and Bax (1998a, b, 1999), Martin-Pastor and Bush (2000) Faham and Bowie (2002), Ujwal and Bowie (2011) Mills and Holland (2004), Pappas and Holland (2008), Luo et al. (2010) Barbosa-Barros et al. (2008a), Rodríguez et al. (2010, 2011) Booth et al. (1997), Sugiyama et al. (1999), Kim et al. (2001), Renthal and Velasquez (2002), Andersson et al. (2007), McKibbin et al. (2007), Gayen and Mukhopadhyay (2008), Krishnamani et al. (2012) Sanders and Prestegard (1990, 1991), Aubin et al. (1993), Salvatore et al. (1996), Chen and Gouaux (1999), Kawaguchi et al. (2003), Wang et al. (2012) Faham et al. (2005) Lind et al. (2008) Nieh et al. (2006), Rubio et al. (2011) Barbosa-Barros et al. (2008b, 2009), Rodríguez et al. (2010, 2011) Wallace and Sansom (2009) Lind et al. (2008) Wang et al. (1998) Chou et al. (2004), Wang et al. (2004) Baek et al. (2011) Aussenac et al. (2005) Evanics and Prosser (2005) Yamaguchi et al. (2012) Loudet et al. (2010)

134

7.7

Liposomes, Lipid Bilayers and Model Membranes

LIPID BILAYER DETERMINES ANTIMICROBIAL PEPTIDE ORGANIZATION

Antimicrobial peptides (AMPs) are a class of small molecules which are capable of disrupting biological membranes through a number of different mechanisms. However, their ability to differentiate between eukaryotic and prokaryotic membranes makes them promising therapeutic agents against certain pathogens importantly, without inducing drug resistance—an often occurring problem with drugs targeting a speciic protein or gene. Alamethicin (Alm) is a small AMP that spontaneously aggregates to form a membrane-spanning bundle (Figure 7.7). To compensate for the hydrophobic mismatch between the bilayer’s hydrophobic core and the protein’s embedded hydrophobic domain, the membrane is deformed—this is because the peptide is stiffer than the bilayer. The energy cost associated with such deformation depends on the membrane’s thickness, its bending and area stretch moduli, and its intrinsic curvature. Pan et al. studied how the Alm bundle structure behaves in two lipid bilayers, namely, di-18:1 PC and di-22:1 PC (Pan et al. 2009b). These bilayers have similar physical properties, except that the di-22:1 PC bilayer is about 7 Å thicker than di-18:1 PC. It was found that Alm forms a hexametric bundle in di-18:1 PC, while a nonamer structure was discovered in di-22:1 PC. The smaller Alm bundle in di-18:1 PC was the result of hydrophobic thickness matching between di-18:1 PC bilayers and Alm—as mentioned, the hydrophobic thickness of di-22:1 PC bilayers is 7 Å larger. This notion is consistent with the well-known functional cutoff effect (Balgavý and Devínsky 1996) observed, for example, in Ca2+ -transporting ATPase incorporated in lipid bilayers (Karlovská et al. 2006). The proper function of a membrane protein in a biological membrane, thus depends on the structural and dynamical properties of the underlying lipid matrix. The close interplay between the lipid matrix and associated AMPs has also emerged from other studies. Sani et al. reported that lipid composition is an important regulator in controlling maculatin 1.1’s conformation and orientation (Sani et al. 2012). In the case of zwitterionic PC lipid bilayers, the peptide’s helical content—a good indicator of the peptide’s interaction potential with lipid bilayers—was found to depend on lipid hydrocarbon chain length and degree of unsaturation, while in anionic lipid bilayers, maculatin 1.1 interacted strongly and oriented orthogonal to the bilayer normal. In another study involving a PC/PG mixture and the cationic AMP, aurein, Cheng et al. found that AMP–membrane interactions were affected by the amount of charged PG lipid present and the hydrophobic thickness of the lipid bilayer (Cheng et al. 2009). More recently, MD simulations of gramicidin A in different lipid bilayers have shown a radial dependence of lipid bilayer perturbation, induced by the addition of gramicidin A (Kim et al. 2012). From these studies, one can surmise that AMP organization and function are to a great extent regulated by the host lipid bilayer through a variety of chemical and physical interactions. In-depth studies of AMP interactions with model membranes are paving the way in deciphering how AMPs

FIGURE 7.7 The Alm hexameric structure in a lipid bilayer. (a) Side view and (b) top view.

Lipid Diversity and Its Implications for Membrane Organization

135

interact with the different lipid species that make up biological membranes. We are of the belief that such knowledge will prove to be invaluable when designing and developing more effective peptidebased antibiotics.

7.8

ION-SPECIFIC EFFECTS IN BACTERIAL MEMBRANES

In addition to lipid–peptide (and lipid–protein) interactions, the signiicance of the aqueous phase for the proper function of biological membranes cannot be overestimated. Biological membranes are surrounded by an electrolytic liquid containing Na+, K+, Ca2+, Mg2+, and Cl− ions. Their interactions with cell membranes are understood to inluence, for example, the gating of ion channels, membrane fusion, and membrane luidity, to name a few. Over the years, there have been copious amounts of biophysical reports demonstrating that ions affect the physical properties of lipid bilayers (see, e.g., Pabst et al. 2010, for a recent review). The effect of Ca2+ cations was recently reported in bacterial mimetic membranes composed of lipopolysaccharides (LPSs) (Kučerka et al. 2008b). LPSs are the major lipid component making up the outermost lealet of the asymmetric outer membrane (OM) of Gram-negative bacteria (Wilkinson 1996, Nikaido 2003). It contributes to the OM’s structural integrity and also protects the bacteria from a variety of toxic molecules, such as certain antibiotics (e.g., penicillin), digestive enzymes (e.g., lysozyme), detergents, heavy metals, bile salts, and some dyes. The passage of nucleotides, disaccharides, amino acids, vitamins, and iron for nutritional growth are usually transported through the OM by porin proteins, but it is LPS that provides the bacteria with its remarkable selectively permeable membrane that is resistant to a variety of deleterious agents. In particular, Pseudomonas aeruginosa is well noted for its recalcitrance to conventional antibiotic therapy, partly as a result of its unique surface chemistry (Rocchetta et al. 1999). For this reason, and also due to the ubiquity of P. aeruginosa and its impact upon health as both an opportunistic and nosocomial pathogen, this organism represents an attractive candidate for medical and pharmacological studies. Although LPS molecules are structurally diverse, they share a common architecture composed of three basic units. The irst is a lipid A moiety that anchors the LPS molecule into the hydrophobic domain of the OM. It consists of two phosphorylated glucosamine units that are typically acylated with four to six fatty acids and is considered to be responsible for most of the toxicity associated with LPS. Second, the LPS’ core oligosaccharide is made up of 8–12 monosaccharide units, and is connected to lipid A by a 2-keto-3-deoxyoctonoic acid (Kdo). Finally, the third part is formed by repetitive monosaccharide subunits (i.e., O-side chain), which are responsible for much of the bacterium’s immunospeciicity (Caroff and Karibian 2003). However, recent experiments revealed a determining effect of counterions involved in the system. Small-angle neutron diffraction (SAND) data showed that water penetrates Ca2+ –LPS bilayers to a lesser extent than Na+ - and Mg2+ –LPS bilayers (Kučerka et al. 2008b). While Ca2+ cations make LPS bilayers more compact and less permeable to water, a signiicant amount of water penetrates deep into Mg2+ –LPS and Na+ –LPS bilayers, including the bilayer’s hydrophobic core (Figure 7.8). It is believed that such increased levels of hydration could be associated with enhanced biological activity in these bacterial membranes. As such, a more accurate determination of the membrane’s structure may allow for a better understanding of membrane function. For example, differences in a bilayer’s permeability to water could have implications with regard to how small molecules permeate through the OM of Gram-negative bacteria, aiding in the development of more effective antibiotics.

7.9 CONCLUSIONS Cell membranes possess a ubiquitous bilayer architecture that is vital for proper biological function. Lipids make up the underlying membrane scaffold enabling proteins to carry out their various

136

Liposomes, Lipid Bilayers and Model Membranes 7

ρ (×10–6 Å–2)

6

LPS + Na+ LPS + Mg2+

5 4 3 2 1 0

LPS + Ca2+ –40

–20

0 z (Å)

20

40

FIGURE 7.8 (See color insert.) 1D NSLD proiles obtained from the Fourier reconstruction of diffraction data from oriented LPS bilayers hydrated in 100% D2O. The solid red and green lines correspond to Na+ - and Mg2+ –LPS bilayers, respectively, whereas the solid blue line corresponds to Ca2+ –LPS bilayers. The schematic shows the structural model of Na+ - and Mg2+ –LPS bilayers on the left, and Ca2+ –LPS on the right. The dashed lines demarcate the limits of water penetration.

functions. Membrane heterogeneity (e.g., lateral compartmentalization and lealet asymmetry) that is essential for cell signaling and traficking, can be achieved by dynamically recruiting/expelling speciic lipids into/from functional compartments. Thus, to better understand the underpinning driving force of lipid homeostasis and to enhance our predictability of more complex biomembrane events, we presented a few examples illustrating how lipid diversity affects membrane organization. These ranged from pure lipid bilayer systems to ones containing other biomolecules. Speciically, we presented a transverse lipid bilayer model that deconstructs a disordered luid lipid bilayer into a mathematically manageable number of moieties. The resultant SDP model enabled us to precisely determine the structural properties of bilayers, including the much soughtafter area per lipid. From such physical studies, biologically relevant insights can be obtained. For example, net-charged bilayers highlighted the importance of electrostatic interactions in governing lipid lateral packing. Importantly, as naturally occurring lipids have different headgroup moieties with a different net charge (e.g., neutral PE and PC, monoanionic PG and PS, and dianionic cardiolipin), distinct localized structures, and therefore functions, can be elicited by compartmentalizing compositional diverse lipids with differing headgroups. We then went on to describe the close interplay between lipid matrix and a well-known membrane modulator, namely, cholesterol. The prevailing concept of cholesterol’s rigidifying effect on membrane lexibility (i.e., bending modulus) was challenged when different degrees of hydrocarbon chain unsaturation were considered. For example, lipid bilayers with dimonounsaturated chains were found to be equally stiff in the presence or absence of 30 mol% cholesterol. In an extreme case where lipid hydrocarbon chains possess multiple unsaturated bonds (i.e., PUFA), cholesterol was found to segregate in the middle of the lipid bilayer. However, the nominal upright orientation of cholesterol was retrieved by doping with less unsaturated lipids. This clearly implies that the composition of a lipid’s hydrocarbon chains is important in modulating the orientation of biomolecules. A unique lipid mixture (i.e., micelle) was included in this chapter to illustrate how simple lipid mixtures can assume a number of different morphologies, simply altering temperature or lipid concentration. The micellar system has served as an important platform for a broad spectrum of biochemical and biophysical studies by providing a biomimetic membrane environment for membrane peptides and proteins. The active role played by the lipid matrix in regulating peptide/protein organizations was illustrated by reporting on the Alm bundle size in two lipid bilayers. It was found that by altering lipid bilayer thickness, Alm assumes a different membrane-spanning bundle structure. As each lipid has its own unique structure (e.g., thickness), different membrane thicknesses required for proper

Lipid Diversity and Its Implications for Membrane Organization

137

membrane function can be achieved by varying lipid composition. Finally, we showed that not only the membrane, but also ions in the aqueous medium surrounding the biomembrane play an important role in modulating membrane structure and function.

REFERENCES Abraham, T. et al. 2008. Monolayer ilm behavior of lipopolysaccharide from Pseudomonas aeruginosa at the air–water interface. Biomacromolecules, 9(10), 2799–2804. Andersson, A. et al. 2007. The membrane-induced structure of melittin is correlated with the luidity of the lipids. Biochimica et Biophysica Acta—Biomembranes, 1768(1), 115–121. Andersson, A. and Mäler, L. 2002. NMR solution structure and dynamics of motilin in isotropic phospholipid bicellar solution. Journal of Biomolecular NMR, 24(2), 103–112. Aubin, Y. et al. 1993. Structure and dynamics of the sialic acid moiety of GM3-ganglioside at the surface of a magnetically oriented membrane. Biochemistry, 32(49), 13405–13413. Aussenac, F., Lavigne, B., and Dufourc, E. J. 2005. Toward bicelle stability with ether-linked phospholipids: Temperature, composition, and hydration diagrams by 2H and 31P solid-state NMR. Langmuir, 21(16), 7129–7135. Aveldano, M. I. 1989. Dipolyunsaturated species of retina phospholipids and their fatty-acids. Biomembranes and Nutrition, 195, 87–96. Baek, S. B. et al. 2011. An NMR study on the conformation of substance P in acidic bicelles. Bulletin of the Korean Chemical Society, 32(10), 3702–3706. Balgavý, P. and Devínsky, F. 1996. Cut-off effects in biological activities of surfactants. Advances in Colloid and Interface Science, 66, 23–63. Barbosa-Barros, L. et  al. 2008a. Effect of bicellar systems on skin properties. International Journal of Pharmaceutics, 352(1–2), 263–272. Barbosa-Barros, L. et al. 2008b. Penetration and growth of DPPC/DHPC bicelles inside the stratum corneum of the skin. Langmuir, 24(11), 5700–5706. Barbosa-Barros, L. et al. 2009. Lipid nanostructures: Self-assembly and effect on skin properties. Molecular Pharmaceutics, 6(4), 1237–1245. Baritaki, S. et al. 2007. Reversal of tumor resistance to apoptotic stimuli by alteration of membrane luidity: Therapeutic implications. Advances in Cancer Research, 98, 149–190. Booth, P. J. et al. 1997. Evidence that bilayer bending rigidity affects membrane protein folding. Biochemistry, 36(1), 197–203. Brown, D. A. and London, E. 2000. Structure and function of sphingolipid- and cholesterol-rich membrane rafts. Journal of Biological Chemistry, 275(23), 17221–17224. Caroff, M. and Karibian, D. 2003. Structure of bacterial lipopolysaccharides. Carbohydrate Research, 338(23), 2431–2447. Cevc, G. 1990. Membrane electrostatics. Biochimica et Biophysica Acta, 1031(3), 311–382. Chen, G. Q. and Gouaux, E. 1999. Probing the folding and unfolding of wild-type and mutant forms of bacteriorhodopsin in micellar solutions: Evaluation of reversible unfolding conditions. Biochemistry, 38(46), 15380–15387. Cheng, J. T. J. et al. 2009. Effect of membrane composition on antimicrobial peptides aurein 2.2 and 2.3 from Australian southern bell frogs. Biophysical Journal, 96(2), 552–565. Cheng, Z. Y. and Li, Y. Z. 2007. What is responsible for the initiating chemistry of iron-mediated lipid peroxidation: An update. Chemical Reviews, 107(5), 2165–2165. Chou, J. J., Baber, J. L., and Bax, A. 2004. Characterization of phospholipid mixed micelles by translational diffusion. Journal of Biomolecular NMR, 29(3), 299–308. Danielli, J. F. 1975. The bilayer hypothesis of membrane structure. In: Weissmann, G. and Clairborne, R. eds. Cell Membranes: Biochemistry, Cell Biology, and Pathology. New York: Hospital Practice, 3–11. Danielli, J. F. and Davson, H. 1935. A contribution to the theory of permeability of thin ilms. Journal of Cellular and Comparative Physiology, 5(4), 495–508. Diller, A. et al. 2009. Bicelles: A natural molecular goniometer for structural, dynamical and topological studies of molecules in membranes. Biochimie, 91(6), 744–751. Eckert, G. P., Wood, W. G., and Muller, W. E. 2010. Lipid membranes and beta-amyloid: A harmful connection. Current Protein and Peptide Science, 11(5), 319–325. Evanics, F. and Prosser, R. S. 2005. Discriminating binding and positioning of amphiphiles to lipid bilayers by 1H NMR. Analytica Chimica Acta, 534(1), 21–29.

138

Liposomes, Lipid Bilayers and Model Membranes

Faham, S. et  al. 2005. Crystallization of bacteriorhodopsin from bicelle formulations at room temperature. Protein Science, 14(3), 836–840. Faham, S. and Bowie, J. U. 2002. Bicelle crystallization: A new method for crystallizing membrane proteins yields a monomeric bacteriorhodopsin structure. Journal of Molecular Biology, 316(1), 1–6. Fan, Z. and Makielski, J. C. 1997. Anionic phospholipids activate ATP-sensitive potassium channels. Journal of Biological Chemistry, 272(9), 5388–5395. Forrest, B. J. and Reeves, L. W. 1979. Studies in membrane processes X: A deuterium magnetic resonance study of dipalmitoyl lecithin and palmitic acid guests in magnetically-oriented hexadecyltrimethyl-ammonium bromide liquid crystalline system. Chemistry and Physics of Lipids, 24(2), 183–192. Forrest, B. J. and Reeves, L. W. 1981. New lyotropic liquid crystals composed of inite nonspherical micelles. Chemical Reviews, 81(1), 1–14. Gaemers, S. and Bax, A. 2001. Morphology of three lyotropic liquid crystalline biological NMR media studied by translational diffusion anisotropy. Journal of the American Chemical Society, 123(49), 12343–12352. Gayen, A. and Mukhopadhyay, C. 2008. Evidence for effect of GM1 on opioid peptide conformation: NMR study on leucine enkephalin in ganglioside-containing isotropic phospholipid bicelles. Langmuir, 24(10), 5422–5432. Gorter, E. and Grendel, F. 1925. On bimolecular layers of lipoids on the chromocytes of the blood. Journal of Experimental Medicine, 41(4), 439–443. Gracia, R. S. et  al. 2010. Effect of cholesterol on the rigidity of saturated and unsaturated membranes: Fluctuation and electrodeformation analysis of giant vesicles. Soft Matter, 6(7), 1472–1482. Halskau, Ø., Muga, A., and Martínez, A. 2009. Linking new paradigms in protein chemistry to reversible membrane protein interactions. Current Protein and Peptide Science, 10(4), 339–359. Harroun, T. A. et al. 2006a. 0.9 T static magnetic ield and temperature-controlled specimen environment for use with general-purpose optical microscopes. Review of Scientiic Instruments, 77(1), 014102. Harroun, T. A., Katsaras, J., and Wassall, S. R. 2006b. Cholesterol hydroxyl group is found to reside in the center of a polyunsaturated lipid membrane. Biochemistry, 45(4), 1227–1233. Harroun, T. A., Katsaras, J., and Wassall, S. R. 2008. Cholesterol is found to reside in the center of a polyunsaturated lipid membrane. Biochemistry, 47(27), 7090–7096. Harroun, T. A. et al. 2005. Comprehensive examination of mesophases formed by DMPC and DHPC mixtures. Langmuir, 21(12), 5356–5361. Heberle, F. A. et al. 2012. Model-based approaches for the determination of lipid bilayer structure from smallangle neutron and x-ray scattering data. European Biophysics Journal, 41, 875–890. Heipieper, H. J. et al. 2007. Solvent-tolerant bacteria for biotransformations in two-phase fermentation systems. Applied Microbiology and Biotechnology, 74(5), 961–973. Hendler, R. W. 1971. Biological membrane ultrastructure. Physiological Reviews, 51(1), 66–97. Henriksen, J. et al. 2006. Universal behavior of membranes with sterols. Biophysical Journal, 90(5), 1639–1649. Iqbal, U. et al. 2011. Small unilamellar vesicles: A platform technology for molecular imaging of brain tumors. Nanotechnology, 22, 195102(1)–195102(15). Karlovská, J. et al. 2006. Inluence of N-dodecyl-N,N-dimethylamine N-oxide on the activity of sarcoplasmic reticulum Ca(2+)-transporting ATPase reconstituted into diacylphosphatidylcholine vesicles: Effects of bilayer physical parameters. Biophysical Chemistry, 119(1), 69–77. Katsaras, J. 1997. Highly aligned lipid membrane systems in the physiologically relevant “excess water” condition. Biophysical Journal, 73(6), 2924–2929. Katsaras, J. 1998. Adsorbed to a rigid substrate, dimyristoylphosphatidylcholine multibilayers attain full hydration in all mesophases. Biophysical Journal, 75(5), 2157–2162. Katsaras, J. et  al. 2005. Bicellar lipid mixtures as used in biochemical and biophysical studies. Naturwissenschaften, 92(8), 355–366. Kawaguchi, K., Kimura, K., and Asakura, T. 2003. Direct observations of high resolution 1H NMR in liquid phase for peptides bound to bicelles. Kobunshi Ronbunshu, 60(4), 199–202. Kim, J. M. et al. 2001. Structure and function in bacteriorhodopsin: The role of the interhelical loops in the folding and stability of bacteriorhodopsin. Journal of Molecular Biology, 308(2), 409–422. Kim, L. et al. 2012. Inluence of hydrophobic mismatch on structures and dynamics of gramicidin A and lipid bilayers. Biophysical Journal, 102(7), 1551–1560. Krishnamani, V. et al. 2012. Secondary and tertiary structure of bacteriorhodopsin in the SDS denatured state. Biochemistry, 51(6), 1051–1060. Kučerka, N. et al. 2012. Scattering density proile model of POPG bilayers as determined by molecular dynamics simulations and small-angle neutron and x-ray scattering experiments. Journal of Physical Chemistry B, 116(1), 232–239.

Lipid Diversity and Its Implications for Membrane Organization

139

Kučerka, N. et al. 2010. Cholesterol in bilayers with PUFA chains: Doping with DMPC or POPC results in sterol reorientation and membrane-domain formation. Biochemistry, 49(35), 7485–7493. Kučerka, N. et al. 2009. The functional signiicance of lipid diversity: Orientation of cholesterol in bilayers is determined by lipid species. Journal of the American Chemical Society, 131(45), 16358–16359. Kučerka, N. et al. 2008a. Lipid bilayer structure determined by the simultaneous analysis of neutron and x-ray scattering data. Biophysical Journal, 95(5), 2356–2367. Kučerka, N. et al. 2008b. Effect of cations on the structure of bilayers formed by lipopolysaccharides isolated from Pseudomonas aeruginosa PAO1. Journal of Physical Chemistry B, 112(27), 8057–8062. Lenard, J. and Singer, S. J. 1966. Protein conformation in cell membrane preparations as studied by optical rotatory dispersion and circular dichroism. Proceedings of the National Academy of Sciences of the United States of America, 56(6), 1828–1835. Leonard, A. et al. 2001. Location of cholesterol in DMPC membranes. A comparative study by neutron diffraction and molecular mechanics simulation. Langmuir, 17(6), 2019–2030. Lind, J., Nordin, J., and Mäler, L. 2008. Lipid dynamics in fast-tumbling bicelles with varying bilayer thickness: Effect of model transmembrane peptides. Biochimica et Biophysica Acta—Biomembranes, 1778(11), 2526–2534. Lingwood, D. and Simons, K. 2010. Lipid rafts as a membrane-organizing principle. Science, 327(5961), 46–50. Loudet, C. et al. 2010. Biphenyl phosphatidylcholine: A promoter of liposome deformation and bicelle collective orientation by magnetic ields. Progress in Lipid Research, 49(3), 289–297. Luo, R. J., Archer-Hartmann, S. A., and Holland, L. A. 2010. Transformable capillary electrophoresis for oligosaccharide separations using phospholipid additives. Analytical Chemistry, 82(4), 1228–1233. Maccarrone, M. et al. 2011. Cannabinoid receptor signalling in neurodegenerative diseases: A potential role for membrane luidity disturbance. British Journal of Pharmacology, 163(7), 1379–1390. Marcotte, I. and Auger, M. 2005. Bicelles as model membranes for solid- and solution-state NMR studies of membrane peptides and proteins. Concepts in Magnetic Resonance Part A: Bridging Education and Research, 24(1), 17–37. Marguet, D. et  al. 2006. Dynamics in the plasma membrane: How to combine luidity and order. EMBO Journal, 25(15), 3446–3457. Marrink, S. J. et al. 2008. Cholesterol shows preference for the interior of polyunsaturated lipid. Journal of the American Chemical Society, 130(1), 10–11. Martin-Pastor, M. and Bush, C. A. 2000. Conformational studies of human milk oligosaccharides using H-1-C13 one-bond NMR residual dipolar couplings. Biochemistry, 39(16), 4674–4683. Matsumori, N. and Murata, M. 2010. 3D structures of membrane-associated small molecules as determined in isotropic bicelles. Natural Product Reports, 27(10), 1480–1492. McClung, C. R. and Davis, S. J. 2010. Ambient thermometers in plants: From physiological outputs towards mechanisms of thermal sensing. Current Biology, 20(24), R1086–R1092. McKibbin, C. et  al. 2007. Opsin stability and folding: Modulation by phospholipid bicelles. Journal of Molecular Biology, 374(5), 1319–1332. Mclaughlin, S. 1989. The electrostatic properties of membranes. Annual Review of Biophysics and Biophysical Chemistry, 18, 113–136. McMullen, T. P. W. and McElhaney, R. N. 1996. Physical studies of cholesterol–phospholipid interactions. Current Opinion in Colloid and Interface Science, 1(1), 83–90. Mills, J. O. and Holland, L. A. 2004. Membrane-mediated capillary electrophoresis: Interaction of cationic peptides with bicelles. Electrophoresis, 25(9), 1237–1242. Newton, A. C. 1993. Interaction of proteins with lipid headgroups—Lessons from protein-kinase-C. Annual Review of Biophysics and Biomolecular Structure, 22, 1–25. Nieh, M.-P. et al. 2001. SANS study of the structural phases of magnetically alignable lanthanide-doped phospholipid mixtures. Langmuir, 17(9), 2629–2638. Nieh, M.-P. et al. 2002. SANS study on the effect of lanthanide ions and charged lipids on the morphology of phospholipid mixtures. Biophysical Journal, 82(5), 2487–2498. Nieh, M.-P. et al. 2006. Spontaneously forming ellipsoidal phospholipid unilamellar vesicles and their interactions with helical domains of saposin C. Langmuir, 22(26), 11028–11033. Nieh, M.-P. et al. 2004. Magnetically alignable phase of phospholipid bicelle mixtures is a chiral nematic made up of wormlike micelles. Langmuir, 20(19), 7893–7897. Nieh, M.-P. et  al. 2005. Spontaneously formed unilamellar vesicles with path-dependent size distribution. Langmuir, 21(15), 6656–6661.

140

Liposomes, Lipid Bilayers and Model Membranes

Nieh, M.-P. et al. 2003. Highly aligned lamellar lipid domains induced by macroscopic coninement. Langmuir, 19(17), 6936–6941. Nikaido, H. 2003. Molecular basis of bacterial outer membrane permeability revisited. Microbiology and Molecular Biology Reviews, 67(4), 593–656. Olofsson, A. et al. 2007. Negatively charged phospholipid membranes induce amyloid formation of medin via an alpha-helical intermediate. Journal of Molecular Biology, 374(1), 186–194. Ottiger, M. and Bax, A. 1998a. Characterization of magnetically oriented phospholipid micelles for measurement of dipolar couplings in macromolecules. Journal of Biomolecular NMR, 12(3), 361–372. Ottiger, M. and Bax, A. 1998b. Determination of relative N–H(N) N–C′, C(α)–C′, and C(α)–H(α) effective bond lengths in a protein by NMR in a dilute liquid crystalline phase. Journal of the American Chemical Society, 120(47), 12334–12341. Ottiger, M. and Bax, A. 1999. How tetrahedral are methyl groups in proteins? A liquid crystal NMR study. Journal of the American Chemical Society, 121(19), 4690–4695. Pabst, G. et  al. 2007. Entropy-driven softening of luid lipid bilayers by alamethicin. Langmuir, 23(23), 11705–11711. Pabst, G. et al. 2010. Applications of neutron and x-ray scattering to the study of biologically relevant model membranes. Chemistry and Physics of Lipids, 163(6), 460–479. Pan, J. et al. 2012. Molecular structures of luid phase phosphatidylglycerol bilayers as determined by small angle neutron and x-ray scattering. Biochimica et Biophysica Acta—Biomembranes, 1818(9), 2135–2148. Pan, J. J. et  al. 2008. Cholesterol perturbs lipid bilayers nonuniversally. Physical Review Letters, 100(19), 198103. Pan, J. J. et al. 2009a. Alamethicin in lipid bilayers: Combined use of x-ray scattering and MD simulations. Biochimica et Biophysica Acta—Biomembranes, 1788(6), 1387–1397. Pan, J. J., Tristram-Nagle, S., and Nagle, J. F. 2009b. Alamethicin aggregation in lipid membranes. Journal of Membrane Biology, 231(1), 11–27. Pan, J. J., Tristram-Nagle, S., and Nagle, J. F. 2009c. Effect of cholesterol on structural and mechanical properties of membranes depends on lipid chain saturation. Physical Review E, 80(2), 021931. Pappas, T. J. and Holland, L. A. 2008. Fluid steering in a microluidic chip by means of thermally responsive phospholipids. Sensors and Actuators B—Chemical, 128(2), 427–434. Phillips, R. et al. 2009. Emerging roles for lipids in shaping membrane–protein function. Nature, 459(7245), 379–385. Prosser, R. S. et al. 2006. Current applications of bicelles in NMR studies of membrane-associated amphiphiles and proteins. Biochemistry, 45(28), 8453–8465. Prosser, R. S. et  al. 1996. Magnetically aligned membrane model systems with positive order parameter: Switching the sign of Szz with paramagnetic ions. Journal of the American Chemical Society, 118(1), 269–270. Prosser, R. S., Hwang, J. S., and Vold, R. R. 1998. Magnetically aligned phospholipid bilayers with positive ordering: A new model membrane system. Biophysical Journal, 74(5), 2405–2418. Quehenberger, O. et al. 2010. Lipidomics reveals a remarkable diversity of lipids in human plasma. Journal of Lipid Research, 51(11), 3299–3305. Ram, P. and Prestegard, J. H. 1988. Magnetic ield induced ordering of bile/salt phospholipid micelles: New media for NMR structural investigations. Biochimica et Biophysica Acta, 940(2), 289–294. Renthal, R. and Velasquez, D. 2002. Self-association of helical peptides in a lipid environment. Journal of Protein Chemistry, 21(4), 255–264. Robertson, J. D. 1957. New observations on the ultrastructure of the membranes of frog peripheral nerve ibers. Journal of Biophysical and Biochemical Cytology, 3(6), 1043–1047. Robertson, J. D. 1959. The ultrastructure of cell membranes and their derivatives. Biochemical Society Symposia, 16, 3–43. Rocchetta, H. L., Burrows, L. L., and Lam, J. S. 1999. Genetics of O-antigen biosynthesis in Pseudomonas aeruginosa. Microbiology and Molecular Biology Reviews, 63(3), 523–553. Rodríguez, G. et  al. 2011. Bicellar systems as modiiers of skin lipid structure. Colloids and Surfaces B— Biointerfaces, 84(2), 390–394. Rodríguez, G. et al. 2010. Application of bicellar systems on skin: Diffusion and molecular organization effects. Langmuir, 26(13), 10578–10584. Rubio, L. et al. 2011. Structural effects of lufenamic acid in DPPC/DHPC bicellar systems. Soft Matter, 7(18), 8488–8497. Ruelland, E. and Zachowski, A. 2010. How plants sense temperature. Environmental and Experimental Botany, 69(3), 225–232.

Lipid Diversity and Its Implications for Membrane Organization

141

Saidi, Y., Finka, A., and Goloubinoff, P. 2011. Heat perception and signalling in plants: A tortuous path to thermotolerance. New Phytologist, 190(3), 556–565. Salvatore, B. A., Ghose, R., and Prestegard, J. H. 1996. NMR studies of a 13C,15N-labeled GM4-lactam glycolipid at an oriented model–membrane interface. Journal of the American Chemical Society, 118(17), 4001–4008. Sanders, C. R. and Landis, G. C. 1995. Reconstitution of membrane-proteins into lipid-rich bilayered mixed micelles for NMR-studies. Biochemistry, 34(12), 4030–4040. Sanders, C. R. and Prestegard, J. H. 1990. Magnetically orientable phospholipid bilayers containing small amounts of a bile salt analog, CHAPSO. Biophysical Journal, 58(2), 447–460. Sanders, C. R. and Prestegard, J. H. 1991. Orientation and dynamics of β-dodecyl glucopyranoside in phospholipid-bilayers by oriented sample NMR and order matrix analysis. Journal of the American Chemical Society, 113(6), 1987–1996. Sanders, C. R. and Schwonek, J. P. 1992. Characterization of magnetically orientable bilayers in mixtures of dihexanoylphosphatidylcholine and dimyristoylphosphatidylcholine by solid-state NMR. Biochemistry, 31(37), 8898–8905. Sani, M. A., Whitwell, T. C., and Separovic, F. 2012. Lipid composition regulates the conformation and insertion of the antimicrobial peptide maculatin 1.1. Biochimica et Biophysica Acta—Biomembranes, 1818(2), 205–211. Schmidt, D., Jiang, Q. X., and MacKinnon, R. 2006. Phospholipids and the origin of cationic gating charges in voltage sensors. Nature, 444(7120), 775–779. Seddon, A. M., Curnow, P., and Booth, P. J. 2004. Membrane proteins, lipids and detergents: Not just a soap opera. Biochimica et Biophysica Acta—Biomembranes, 1666(1–2), 105–117. Shevchenko, A. and Simons, K. 2010. Lipidomics: Coming to grips with lipid diversity. Nature Reviews Molecular Cell Biology, 11(8), 593–598. Silvius, J. R. 2003. Role of cholesterol in lipid raft formation: Lessons from lipid model systems. Biochimica et Biophysica Acta—Biomembranes, 1610(2), 174–183. Singer, S. J. and Nicolson, G. L. 1972. The luid mosaic model of the structure of cell membranes. Science, 175(4023), 720–731. Smith, L. L. 1991. Another cholesterol hypothesis—Cholesterol as antioxidant. Free Radical Biology and Medicine, 11(1), 47–61. Soni, S. P. et al. 2008. Docosahexaenoic acid enhances segregation of lipids between raft and nonraft domains: H-2-NMR study. Biophysical Journal, 95(1), 203–214. Soong, R. et al. 2010. Bicellar mixtures containing pluronic F68: Morphology and lateral diffusion from combined SANS and PFG NMR studies. Langmuir, 26(4), 2630–2638. Sorre, B. et  al. 2009. Curvature-driven lipid sorting needs proximity to a demixing point and is aided by proteins. Proceedings of the National Academy of Sciences of the United States of America, 106(14), 5622–5626. Stillwell, W. and Wassall, S. R. 2003. Docosahexaenoic acid: Membrane properties of a unique fatty acid. Chemistry and Physics of Lipids, 126(1), 1–27. Sugiyama, Y., Fujii, K., and Mukohata, Y. 1999. The effect of carboxyl group modiication on the chromophore regeneration of archaeopsin-1 and bacterioopsin. Journal of Biochemistry, 125(6), 1144–1150. Torres, S., Pandey, A., and Castro, G. R. 2011. Organic solvent adaptation of Gram positive bacteria: Applications and biotechnological potentials. Biotechnology Advances, 29(4), 442–452. Tristram-Nagle, S. and Nagle, J. F. 2007. HIV-1 fusion peptide decreases bending energy and promotes curved fusion intermediates. Biophysical Journal, 93(6), 2048–2055. Ujwal, R. and Bowie, J. U. 2011. Crystallizing membrane proteins using lipidic bicelles. Methods, 55(4), 337–341. Upchurch, R. G. 2008. Fatty acid unsaturation, mobilization, and regulation in the response of plants to stress. Biotechnology Letters, 30(6), 967–977. van Dam, L., Karlsson, G., and Edwards, K. 2004. Direct observation and characterization of DMPC/DHPC aggregates under conditions relevant for biological solution NMR. Biochimica et Biophysica Acta— Biomembranes, 1664(2), 241–256. van Klompenburg, W. et  al. 1997. Anionic phospholipids are determinants of membrane protein topology. EMBO Journal, 16(14), 4261–4266. van Meer, G., Voelker, D. R., and Feigenson, G. W. 2008. Membrane lipids: Where they are and how they behave. Nature Reviews Molecular Cell Biology, 9(2), 112–124. Vist, M. R. and Davis, J. H. 1990. Phase-equilibria of cholesterol dipalmitoylphosphatidylcholine mixtures— H-2 nuclear magnetic-resonance and differential scanning calorimetry. Biochemistry, 29(2), 451–464.

142

Liposomes, Lipid Bilayers and Model Membranes

Vold, R. R. and Prosser, R. S. 1996. Magnetically oriented phospholipid bilayered micelles for structural studies of polypeptides. Does the ideal bicelle exist? Journal of Magnetic Resonance Series B, 113(3), 267–271. Wallace, E. J. and Sansom, M. S. P. 2009. Carbon nanotube self-assembly with lipids and detergent: A molecular dynamics study. Nanotechnology, 20(4), 045101. Wang, H. et al. 1998. A liquid crystalline medium for measuring residual dipolar couplings over a wide range of temperatures. Journal of Biomolecular NMR, 12(3), 443–446. Wang, H., Elferich, J., and Gouaux, E. 2012. Structures of LeuT in bicelles deine conformation and substrate binding in a membrane-like context. Nature Structural and Molecular Biology, 19(2), 212–219. Wang, J. F., Schnell, J. R., and Chou, J. J. 2004. Amantadine partition and localization in phospholipid membrane: A solution NMR study. Biochemical and Biophysical Research Communications, 324(1), 212–217. Warschawski, D. E. et al. 2011. Choosing membrane mimetics for NMR structural studies of transmembrane proteins. Biochimica et Biophysica Acta—Biomembranes, 1808(8), 1957–1974. Wassall, S. R. and Stillwell, W. 2008. Docosahexaenoic acid domains: The ultimate non-raft membrane domain. Chemistry and Physics of Lipids, 153(1), 57–63. Wilkinson, S. G. 1996. Bacterial lipopolysaccharides—Themes and variations. Progress in Lipid Research, 35(3), 283–343. Yamaguchi, T. et al. 2012. NMR-based conformational analysis of sphingomyelin in bicelles. Bioorganic and Medicinal Chemistry, 20(1), 270–278. Zandomeneghi, G. et al. 2003. Switched-angle spinning applied to bicelles containing phospholipid-associated peptides. Journal of Biomolecular NMR, 25(2), 125–132. Zhou, Z. et al. 2010. Changes in freezing tolerance in hybrid poplar caused by up- and down-regulation of PtFAD2 gene expression. Transgenic Research, 19(4), 647–654.

8

Liposome-Based Models for Membrane Rafts Methodology and Applications Frederick A. Heberle, Robin S. Petruzielo, Shih Lin Goh, Tatyana M. Konyakhina, David G. Ackerman, Jonathan J. Amazon, and Gerald W. Feigenson

CONTENTS 8.1 8.2

Introduction .......................................................................................................................... 143 Experimental Tools for Studying Domains in Model Membranes....................................... 145 8.2.1 Methods to Interrogate Phase Behavior of Multicomponent Model Membranes .... 146 8.2.1.1 Nanoscopic Regime (1–10 nm) .................................................................. 146 8.2.1.2 Mesoscopic Regime (10–200 nm).............................................................. 147 8.2.1.3 “Macroscopic” Regime (>200 nm) ............................................................ 148 8.2.1.4 Assessing the Inluence of Probe Molecules on Bilayer Properties .......... 148 8.2.1.5 Probe-Free Methods .................................................................................. 149 8.2.2 Methods to Determine Tielines in Ternary Mixtures .............................................. 149 8.2.3 Methods to Determine Domain Size ........................................................................ 150 8.2.3.1 Use of FRET to Determine Domain Size .................................................. 151 8.2.3.2 Use of SANS to Determine Domain Size .................................................. 152 8.3 Phase Diagrams of Biomimetic Mixtures ............................................................................ 154 8.3.1 Type II Ternary Mixtures: Micron-Sized Liquid Domains ...................................... 154 8.3.2 Type I Ternary Mixtures: Nanometer-Sized Liquid Domains ................................. 155 8.3.3 Type II M Behavior: Modulated Phases in a Four-Component Mixture .................... 156 8.4 Toward a Better Raft Model: Insights from the Type II M Domain Size Transition.............. 158 8.4.1 Modulated Phase Patterns Result from Competing Interactions.............................. 158 8.4.2 Domain Size Depends on Phase Thickness Mismatch in the Nanoscopic Regime .....159 8.5 Moving Forward ................................................................................................................... 160 References ...................................................................................................................................... 161

8.1

INTRODUCTION

The plasma membrane (PM) is a quasi-two-dimensional barrier surrounding the cell, providing a controlled environment for the function of its organelles. When the landmark luid mosaic model was proposed in 1972, membrane lipids were pictured as a passive and luid “sea” of nondescript molecules, the uniform matrix in which the more interesting transmembrane and peripheral proteins diffuse and interact (Singer and Nicolson 1972). Over the next four decades, this picture has undergone radical transformations that continue to this day. It is now clear that lateral heterogeneities in the PM arising from nonrandom mixing of lipids and proteins are important in many cellular processes, though the mechanisms governing their formation and dissolution are still not completely understood. These self-organized domains could compartmentalize the bilayer, controlling 143

144

Liposomes, Lipid Bilayers and Model Membranes

diffusion and local concentrations of membrane-associated proteins that sort preferentially between different compartments. Lipid rafts (Simons and Ikonen 1997) (the moniker bestowed by Kai Simons, who formalized the concept in the late 1990s) have been proposed to play a role in many cellular processes, such as intracellular traficking (Parton and Richards 2003, Rajendran and Simons 2005), viral assembly and exit (Ono 2010), lipid/protein sorting (Cao et al. 2012), and early events in signal transduction. The signiicance of membrane heterogeneity in governing protein functions is illustrated by an example from the ield of immune cell signaling (Holowka et al. 2005). When FcεRI receptors are crosslinked by IgE, they form a stabler patch on the PM outer lealet, which recruits Lyn kinases to initiate phosphorylation of the receptors on the inner lealet. Phosphatases are excluded from this patch. Furthermore, cholesterol depletion disrupts the co-localization of crosslinked receptors and Lyn, and subsequently, causes a loss of Lyn-mediated phosphorylation (Sheets et  al. 1999). This indicates that specialized compartments, compositionally distinct from the rest of the bilayer, form in the PM based on favorable associations between particular lipids and proteins. Similar observations are reported in T-cell signaling, where the regulation of adaptor protein phosphorylation events by kinases is dependent on their co-localization in speciic regions within the PM upon stimulation of T-cell receptors (Fuller and Zhang 2009). Disruption of proper protein localization results in the disturbance of the entire signaling cascade. As with many transformative ideas, the initial burst of excitement surrounding lipid rafts was followed by a period of controversy and doubt (Munro 2003, McMullen et al. 2004, Shaw 2006, Leslie 2011). Rafts in resting cells are small and possibly transient, and cannot be visualized with standard microscopy. Seeing is believing, and by the mid-2000s, the lack of visual evidence for rafts only added fuel to an emerging ire: early biochemical assays for rafts involving detergent extraction were revealed to be prone to signiicant artifacts (Heerklotz 2002, Hancock 2006, Brown 2006). These growing pains underscored the enormous challenges faced by biologists and biochemists studying the complex membranes of living cells. Biological membranes are vastly complex chemical soups that come in a wide variety of lavors, with hundreds of lipid and protein components. By deinition, models must generalize, and the luid mosaic model downplayed lipid diversity in order to emphasize a more complicated picture of protein diffusion and interaction within the plane of the membrane. However, even as the luid mosaic model was gaining acceptance among biologists, biophysicists utilizing a relatively new innovation, the liposome, were uncovering remarkable behaviors in membranes stripped of their proteins, and indeed of nearly all their chemical complexity. In this reductionist approach, a system is whittled down to just a few representative components, such that the global behavior can be understood in terms of individual molecular interactions. Crucially, free-loating liposomes are structurally similar to biological membranes, circumventing unwanted interactions with solid bilayer supports that add to the system’s complexity. Liposomes can be made eficiently, cheaply, reproducibly, and (through their continued technological development) in a staggering range of sizes (tens of nanometers to over 100 microns) suitable for different types of experiments. Starting in the 1960s, calorimetric (Ladbrooke and Chapman 1969, de Kruyff et  al. 1973, 1974, Klopfenstein et  al. 1974, Demel et al. 1977) and spectroscopic (Marsh and Smith 1973, Wu and McConnell 1975) research on submicron-sized liposomes established that “simple” lipid mixtures can display an astonishingly rich set of behaviors, including phase coexistence driven by the immiscibility of unlike lipid chains. A signiicant breakthrough occurred with the development of improved techniques for producing giant (10–100 micron diameter) unilamellar vesicles (GUVs). In the early 2000s, luorescence micrographs of GUVs mimicking the composition of mammalian PM, and stained with lipophilic dyes, revealed stable liquid-phase domains with micrometer dimensions, an observation that dovetailed with the evolving raft hypothesis. Signiicantly, the liquid domains were only observed in mixtures that contained, in addition to two lipid species with a large difference in chain melting transition temperature (TM), a substantial amount of cholesterol (by mole fraction, the most abundant component of animal cell PMs). Was it possible that rafts in cells were a manifestation of

145

Liposome-Based Models for Membrane Rafts

the liquid-ordered/liquid-disordered phase coexistence found in model membranes? The striking images that emerged from these studies had remarkable power: they were easily interpretable, providing a decisive determination of phase behavior, and often allowing for the identiication of the coexisting phases. The pioneering work of Korlach et al. (1999) and Dietrich et al. (2001) triggered an avalanche of research to catalog the phase behavior of three-component lipid mixtures, and to understand the fundamental mechanisms responsible for liquid immiscibility. In this chapter, we discuss methods that have been developed to examine phase behavior in liposome-based model membranes. We then summarize relevant phase diagrams, focusing on the major organizing principle that grew out of these studies: that the phase behavior of many threecomponent models for the mammalian outer lealet is broadly similar, but can differ in a very important aspect depending on the structure of the low-melting lipid. Finally, we discuss the newest wrinkle on domain studies in model membranes: the observation of spatially modulated phase patterns in four-component mixtures. Throughout the text, we reserve the term “raft” for PM heterogeneities, and instead use “domain” when referring to liquid-disordered (Ld) and liquidordered (Lo) phases in model membranes.

8.2

EXPERIMENTAL TOOLS FOR STUDYING DOMAINS IN MODEL MEMBRANES

In this section, we summarize experimental methods that have been developed to study lateral heterogeneities in model membranes. We focus on techniques that can be applied to free-loating liposomes. Many other useful techniques for studying phase behavior require a supported bilayer, including atomic force microscopy (AFM) (Lin et al. 2007) and superresolution imaging (Kuo and Hochstrasser 2011). A crucial irst step is to establish the phase behavior and tielines of the system of interest, as any further characterization of the size and properties of domains for a particular phase-separated composition requires knowledge of the compositions and relative amounts of the coexisting phases. While many methods have been developed for studying phase behavior, they are not all equally useful for every phase region. For example, the size scale accessible to a technique is important: a particular class of three-component lipid mixtures gives rise to phase domains with nanometer dimensions, and many techniques are not capable of detecting phase separation in these mixtures. The irst part of our discussion is organized around this principle, dividing the spatial detection regimes into nano-, meso-, and macroscopic. For reference throughout this section, a schematic representation of the size scales accessible to various experimental techniques is shown in Figure 8.1. CFM 2H

NMR SANS FRET

ESR, FS 1 nm

10 nm

100 nm

1 μm

FIGURE 8.1 (See color insert.) Spatial sensitivity of techniques used to study phase behavior of liposomes. Gray shading indicates sensitivity to phase coexistence when phase domains are of the size indicated by the horizontal axis. Red shading indicates that domain sizes can be determined from the data in favorable cases as described in the text. CFM, confocal luorescence microscopy; 2H NMR, deuterium nuclear magnetic resonance; SANS, small-angle neutron scattering; FRET, Förster resonance energy transfer; ESR, electron spin resonance; FS, luorescence spectroscopy, including anisotropy, quantum yield, and lifetime.

146

Liposomes, Lipid Bilayers and Model Membranes

A second distinction often made in the literature is that of direct versus indirect determination of phase coexistence. Direct methods include visual observation of coexisting phases by microscopy, or the appearance of two-component nuclear magnetic resonance (NMR) or electron spin resonance (ESR) spectra. (We note here that NMR and ESR report on the local environment of nuclear or electron spin, rather than a true phase property such as enthalpy. Two-component spectra can be observed for single-phase micelles and microemulsions, which complicates the widely accepted notion that these techniques provide unambiguous evidence for coexisting phases.) With indirect methods, phase coexistence is inferred from the variation of a signal (e.g., lipid lateral diffusion coeficient, luorescence anisotropy or FRET). If the signal is phase speciic, discontinuities in the composition or temperature dependence are generally observed at phase boundaries, and in favorable cases, the variation within the coexistence region can be mathematically modeled as a partition-weighted sum of the signal in the coexisting phases, provided the tielines are known. After the phase diagram is established, further characterization is often desired, such as determining the size of submicron phase domains. We briely review techniques that are useful for this purpose, emphasizing two such techniques employed in our own work, namely, FRET and smallangle neutron scattering (SANS).

8.2.1

Methods to Interrogate phase behavIor oF MultIcoMponent Model MeMbranes

8.2.1.1 Nanoscopic Regime (1–10 nm) A bilayer probe’s signal is often highly sensitive to its immediate (i.e., nearest-neighbor) lipid environment. For example, differences in hydrocarbon chain order and packing can inluence probe dynamics, resulting in signiicant differences in probe spectra in different bilayer phases. Singleprobe spectroscopic techniques for studying phase coexistence take advantage of such differences, and rely on the ability to distinguish subpopulations of the probe in different chemical environments. Analysis involves decomposing a signal into components arising from the pure phases. In the most favorable cases, observation of a multicomponent spectrum provides evidence for phase coexistence. This situation requires the rate at which lipids exchange between the two phases to be considerably slower than the characteristic timescale of the measurement. Lipid diffusion rates DT in luid phases are typically 1–10 µ2/s (Filippov et al. 2003). In a magnetic resonance experiment, the inverse of the frequency difference between phase-speciic splittings sets a timescale τ, from which the 2D diffusion equation r 2 = 4DTτ can be used to estimate a minimum domain size below which an arithmetic average of the component spectra (rather than distinct superposition) is observed. For ESR, the high-ield components of the Ld and Lo hyperine splittings differ by ~4 G (Collado et al. 2005), corresponding to a frequency difference of ~7 × 107 rad/s (Marsh 2009), or a minimum detectable “domain size” on the order of a lipid diameter. Therefore, ESR is in principle capable of detecting lipid clustering on any size scale, including coexistence of small phase domains. Detection of coexisting phases (i.e., two-component ESR spectra) has been reported for Ld + Lβ (Chiang et  al. 2005), Lo + Lβ (Collado et  al. 2005), and Ld + Lo (Collado et  al. 2005, Smith and Freed 2009, Ionova et al. 2012). ESR has also been used as an indirect probe of phase coexistence. In this mode of analysis, a parameter derived directly from the spectrum (e.g., outer hyperine splitting) or from modeling the spectrum (e.g., rotational diffusion rates or chain order parameters derived from spectral simulation) is plotted as a function of composition and analyzed for discontinuities. Indirect analysis of ESR spectra has been used to examine luid heterogeneities in binary (Sankaram and Thompson 1990) and ternary (Chiang et al. 2004, Heberle et al. 2010) mixtures. Similarly, 1H pulsed ield gradient NMR provides measurements of lipid diffusion from which phase coexistence can be inferred. When domains are smaller than ~1 micron, the observed DT is an average of that in the coexisting phases, and varies linearly with phase fraction along a tieline (Filippov et al. 2003). Fluorescence lifetimes are typically on the order of nanoseconds, such that the spatial coniguration of lipids in a luid bilayer is static on the timescale of absorption and emission of a photon via

Liposome-Based Models for Membrane Rafts

147

luorescence. Fluorescence spectroscopy is therefore sensitive to clustering on all size scales, and many variants have been developed as indirect probes of phase coexistence. Anisotropy, lifetime, and intensity have been used independently and often in combination (Mateo et al. 1995, de Almeida et al. 2003, 2007, Zhao et al. 2007a, Halling et al. 2008, Nyholm et al. 2011). Fluorescence quenching methods that rely on preferential segregation of a quenching molecule between phase domains have also been used to infer phase coexistence (Bjorkqvist et al. 2005, Bakht and London 2007, Bakht et al. 2007). As mentioned previously, ESR and luorescence spectroscopy can detect lipid clustering at all size scales, ranging from compositional luctuations inherent in nearly all multicomponent mixtures to irst-order phase separation of large domains. A signiicant caveat is that these extremes might not be distinguishable. Highly nonideal mixing (i.e., lipid interaction energies that, while nonzero, are not large enough to cause irst-order phase separation) can produce clusters of tens to nearly 100 lipids (Heberle and Feigenson 2011), corresponding to a domain radius of ~2–4 nm. Domains of this size consist of only a few shells of lipid and may not be subject to the constraints of irst-order phase separation implied by phase boundaries and tielines. Furthermore, clusters of like molecules within a one-phase mixture of different components do not take on phase properties distinct from the rest of the mixture. Of course, as composition is varied continuously within a single-phase region, the physical properties of the bilayer (e.g., interfacial polarity and chain order) may exhibit considerable variation. Consequently, it may be impossible with any reasonable degree of certainty to ascribe a smooth variation in a luorescence signal to probe partitioning between coexisting domains, even when data can be it to such a model. In light of the previous discussion, an important variant of the quenching method that is relatively insensitive to small lipid clusters has found considerable use in bilayer phase studies. Förster resonance energy transfer (FRET) is a luorescence method in which an excited-state donor luorophore is quenched nonradiatively by an acceptor luorophore. FRET is quantiied by measuring either a loss of donor luorescence or a gain in acceptor luorescence. Unlike the previously mentioned quenching methods, which proceed via molecular contact between the luorescent species and quencher, FRET involves a through-space interaction between excitation and emission transition dipoles of donor and acceptor, respectively. The eficiency of donor quenching depends predictably on the distance between donor and acceptor or, for freely diffusing lipid probes in a membrane, the distribution of donor– acceptor distances (Wolber and Hudson 1979, Fung and Stryer 1978). This distribution is strongly affected by partitioning of donor and acceptor between the coexisting phase domains (Buboltz et al. 2007a, Buboltz 2007), as well as the size and morphology of such domains (Towles et al. 2007, Towles and Dan 2007). As a general rule, FRET cannot reliably detect phase domains smaller than the Förster distance (R0) of the donor/acceptor pair (Towles et al. 2007), which by deinition is the distance at which energy transfer between a static donor–acceptor pair is 50% eficient. Probe pairs covering a range of R0 from 2 to 6 nm are commonly used, and this range provides a relatively sensitive method of placing limits on domain size (discussed in Section 8.2.3). FRET has found extensive use as a probe for phase coexistence in ternary mixtures (Feigenson and Buboltz 2001, Silvius 2003, Brown et al. 2007b, Buboltz et al. 2007b, Heberle et al. 2010, Pathak and London 2011, Petruzielo et al. 2013). 8.2.1.2 Mesoscopic Regime (10–200 nm) 2H NMR provides information on the average order and orientation of C–D bonds in lipid hydrocarbon chains. In the context of membrane phase behavior, it can also provide evidence of coexisting ordered and disordered phases (Veatch 2007). The terminal methyl splittings of Ld and Lo phases have a frequency difference of ~4 kHz, and 2H NMR can therefore provide evidence for coexisting phases (two-component spectra) for domains much larger than 100 nm. For domains much smaller than 100 nm, an arithmetic average of the spectra from the coexisting phases is observed, whereas for intermediate domain sizes, spectral broadening occurs. Veatch et al. used the terminal methyl splitting in multilamellar vesicles (MLVs) to observe phase separation in DPPC/DOPC/Chol, inding a single splitting above the miscibility transition temperature TM (determined by luorescence microscopy), and three well-resolved splittings (one for Ld phase, and two for Lo phase due to

148

Liposomes, Lipid Bilayers and Model Membranes

the nonequivalence of the sn-1 and sn-2 methyls) at lower temperatures, indicating domain sizes >>100 nm (Veatch et  al. 2004). At temperatures just below TM, spectral broadening was apparent, which was interpreted as evidence for domain sizes on the order of 80 nm. Possibilities for the discrepant results from microscopy and NMR were discussed, including perturbation of phase behavior by luorescence probes, and geometric constraints imposed by the MLVs used for NMR measurements. In a later publication, the spectral broadening was reinterpreted in terms of compositional luctuations (2D Ising critical behavior) on length scales of 200 nm) Fluorescence imaging of GUVs (using epiluorescence, confocal, or two-photon microscopy) provides the most decisive determination of phase coexistence, and has been used extensively in model membrane studies of phase coexistence (Korlach et al. 1999, Dietrich et al. 2001, Scherfeld et al. 2003, Veatch and Keller 2003, 2005a, Baumgart et al. 2003, Hammond et al. 2005, Fidorra et al. 2006, Zhao et al. 2007a, b, Juhasz et al. 2009). Variations on this technique in which contrast is provided by luorescence lifetimes (rather than intensity) have been explored (de Almeida et al. 2007, Haluska et al. 2008, Stockl et al. 2008, Ariola et al. 2009). Many excellent reviews have discussed GUV preparation (Dimova et al. 2006, Morales-Penningston et al. 2010, Walde et al. 2010) and the general application of luorescence microscopy to the study of membrane phase behavior (Veatch and Keller 2005b, Feigenson 2007, 2009, Bagatolli and Kumar 2009). As with any technique, there are limitations and potential artifacts associated with luorescence imaging of GUVs. First, domains smaller than the optical resolution limit (~200 nm) are undetectable, and it is now known that many biologically important model systems exhibit suboptical domains. This includes not only the Ld + Lo regions of some ternary mixtures but also some gelcontaining regions, where domains may exist as thin stripes with at least one suboptical dimension. A well-known example is the failure, in nearly every case, to observe three-phase Ld + Lo + Lβ coexistence. A second major concern is the ever-present tendency to use high concentrations of luorescent dyes and intense illumination to improve image quality. These conditions are now known to promote artifactual phase separation and domain enlargement via lipid breakdown (Ayuyan and Cohen 2006, Zhao et al. 2007b). Feigenson and coworkers have developed protocols for minimizing light-induced domain artifacts, including dramatically reducing the amount of probe, and minimizing the illumination intensity (e.g., by using gated shutters and searching for vesicles with less intense brightield illumination) (Morales-Penningston et al. 2010). Finally, the possibility of a heterogeneous compositional distribution within the GUV preparation has been discussed (Dimova et al. 2006, Baykal-Caglar et al. 2012). 8.2.1.4 Assessing the Inluence of Probe Molecules on Bilayer Properties All the methods discussed to this point utilize a probe molecule. The effect (if any) of such probes on phase behavior has long been a vexing issue. As mentioned in the previous section, in extreme cases, probes can cause artifactual demixing of lipids and enlargement of phase domains. It has also been reported that luorescence probes can alter the miscibility transition temperature even at low concentrations (Veatch et al. 2007a). In recent years, molecular dynamics (MD) simulations have been employed to assess the extent of local probe-induced perturbations at the low probe concentrations used in experiments (Repakova et al. 2006, Holtta-Vuori et al. 2008, Loura et al. 2008, Gullapalli et al. 2008, Muddana et al. 2011, Skaug et al. 2011). As reviewed in Loura and Ramalho (2009), MD studies of luorescent probes

Liposome-Based Models for Membrane Rafts

149

in bilayers have shown that while overall perturbations to bilayer properties are minor in small simulation sizes of 100–200 lipids, they can be more signiicant near the probe. For example, commonly used luorescent DiI molecules can signiicantly alter the order and orientation of nearby lipids, and can disturb lipids directly across from the probe in the opposing lealet (Ackerman et al. 2013). Importantly, however, it was found that these large effects are conined to the irst lipid shell surrounding the probe, and die off within a few nanometers distance from the probe. Furthermore, even in systems containing several hundred lipids, changes in ensemble averages were marginal compared to a probe-free bilayer. It was therefore concluded that DiI probes can effectively be used to study the properties of small nanoscopic domains containing a few hundred lipids without signiicantly altering the bilayer behavior. 8.2.1.5 Probe-Free Methods Without a doubt, probe-based techniques have driven the vast majority of research on phase behavior of model membranes. Probe-free studies—where signals arise from phase properties rather than spectroscopic signals of individual molecules—are rare, but the few that have been undertaken provide important checks on the idelity of probe containing bilayers to the intrinsic phase behavior of the system. Differential scanning calorimetry (DSC) is a classic probe-free technique, used extensively to determine binary phase diagrams for PC mixtures and for high-TM lipid/Chol mixtures. These binary diagrams are frequently used as constraints in the construction of ternary phase diagrams (Pokorny et al. 2006, Nyholm et al. 2011, Petruzielo et al. 2013). However, transition endotherms are dramatically attenuated with the addition of cholesterol. The enthalpy of the Ld + Lo transition is so low, and/or the phase transition is so broad, that DSC has not been convincingly used to study Ld + Lo phase coexistence in ternary mixtures (Pokorny et al. 2006). X-ray solution scattering has been used to study lipid-phase behavior. Lateral phase separation can give rise to multiple lamellar repeat distances (D-spacings) in unoriented MLV samples in the small-angle scattering regime, if phase domains are aligned across multiple adjacent bilayers in the multilamellar stack (Chen et al. 2007, Yuan et al. 2009, Boulgaropoulos et al. 2012, Tayebi et al. 2012). In the wide-angle regime—which probes chain–chain correlation spacings (d-spacings)— multiple d-spacings arise from coexisting phases with different chain packing parameters within a single bilayer (Boulgaropoulos et al. 2012). A weakness of these methods is that the absence of multiple peaks does not imply the absence of phase separation; for example, coexisting phases with similar D- or d-spacings may not be resolved. An alternative approach is the use of oriented bilayer stacks, which provide off-axis scattering that is obscured in unoriented samples. The angular distribution of off-axis scattered intensity is related to chain orientational order, which is substantially different for Ld and Lo phases. Mills and coworkers utilized this fact to examine Ld + Lo coexistence in ternary mixtures, by determining if two chain tilt distributions were required to model scattering data for different compositions in DPPC/DOPC/Chol (Mills et al. 2008). Their indings were consistent with those obtained from probe-based (luorescence microscopy and NMR) techniques (Veatch and Keller 2003, Veatch et al. 2007b).

8.2.2

Methods to deterMIne tIelInes In ternary MIxtures

Quantitative determination of tielines in ternary mixtures has proven challenging because the tieline orientation (its angle with respect to horizontal in a three-component diagram) is not constrained. Spectral subtraction of multicomponent spectra can give a quantitative determination of both tieline orientation and endpoints. This method has been successfully applied to ESR (Chiang et al. 2005, Smith and Freed 2009, Ionova et al. 2012) and 2H NMR (Veatch et al. 2006, 2007b, Juhasz et al. 2009) data. Typically, spectra are acquired for many different compositions throughout the phase coexistence region and along the boundaries of the single-phase regions. Additional information is frequently used to provide constraints. For example, Smith and Freed constrained the locations of phase boundaries using microscopy data and then used spectral subtraction to determine tieline

150

Liposomes, Lipid Bilayers and Model Membranes

orientations in the Ld + Lo region of bSM/DOPC/Chol (Smith and Freed 2009). Conversely, Ionova et al. constrained the tieline orientation by searching for isosbesticity among sets of ESR spectra lying along straight-line trajectories, and then used spectral subtraction to determine tieline endpoints (boundaries) for PSM/POPC/Chol (Ionova et al. 2012). Owing to spatial detection limitations, 2H NMR is not useful for tieline determination when phase domains are smaller than 100 nm, while ESR can in principle be applied to nanoscopic domains. Petruzielo et al. used high-compositional-resolution FRET surfaces to determine the orientation of an Ld + Lo tieline in bSM/DOPC/Chol and bSM/POPC/Chol. As discussed in Section 8.2.1.1, FRET eficiency is strongly sensitive to the distribution of donor–acceptor distances in a bilayer, and hence to the partition coeficients of the donor and acceptor luorophores in phase coexistence regions. If both donor and acceptor prefer the same phase, FRET eficiency is enhanced throughout the phase coexistence region (“region of enhanced eficiency,” REE), but shows a global peak at a composition on the tieline with strongest partitioning (Heberle et al. 2010). Similarly, if probes prefer different phases (“region of reduced eficiency,” RRE), a global minimum in FRET eficiency is observed at a composition on the tieline with strongest probe partitioning. It is reasonable to assume that luorescent lipid probes will have the strongest partitioning between phases that are most dissimilar (i.e., the longest tieline in the phase coexistence region). In a three-probe experiment comprising one RRE and one REE energy transfer pair, the locations of the global REE peak and RRE valley, joined by a straight line, give the orientation of the longest tieline. The tieline orientation obtained from FRET surfaces for bSM/POPC/Chol (Petruzielo et al. 2013) was in good agreement with that determined from ESR spectral subtraction in PSM/POPC/Chol (Ionova et al. 2012), and in qualitative agreement with Ld + Lo tielines of related systems (Veatch et al. 2007b, Juhasz et al. 2009). Uppamoochikkal and coworkers devised an elegant small-angle x-ray scattering method for determining the orientation (though not endpoints) of tielines (Uppamoochikkal et  al. 2010). Though oriented bilayer stacks hydrated from the vapor phase were used in the study, the analysis can in principle be applied to aqueous MLV suspensions, and will be mentioned here for completeness. The method has the advantage of being completely probe-free, and is based on the observation in many phase-separated mixtures of a three-dimensional phase separation, whereby phase domains are aligned in adjacent bilayers in the multilamellar stack (discussed in Section 8.2.1.5). If the lamellar repeat distances D of the two phases are different, two sets of Bragg relections are observed. By varying the relative humidity of the sample cell, the thickness of the water layer between lamellae changes (Kučerka et al. 2005). (To modify the technique for aqueous MLVs, bilayer hydration can be varied by applying osmotic stress with a large neutral polymer such as polyethylene glycol, see Pabst et al. 2009.) A plot of the D1 versus D2 spacings at different hydration levels generates a curve that appears to be a sensitive “ingerprint” for a particular tieline (Uppamoochikkal et al. 2010). In other words, because two samples lying on the same tieline are composed of the same coexisting phases (differing only in the relative fractions), these samples have the same D1/D2 hydration curve. Tieline orientations determined for the Ld + Lo region of DPPC/DOPC/Chol were in reasonable qualitative agreement with 2H NMR results, though greater disagreement was observed at lower cholesterol concentrations, particularly for the location of the three-phase triangle and the orientation of Ld + Lβ tielines. A disadvantage of x-ray diffraction is that lamellar repeat distances may by chance be identical for the coexisting phases. Furthermore, phase domains do not always form aligned superstructures in multilamellar bilayers due to packing frustration, in which case only a single D-spacing is observed. Finally, the method requires precise control over hydration, which requires either a specially designed sample chamber for oriented samples hydrated from the vapor phase (Katsaras 1998), or else preparation of many separate MLV samples with increasing osmolyte concentration.

8.2.3

Methods to deterMIne doMaIn sIze

Accurate phase diagrams including tielines open the door to studies of the differential structural and mechanical properties of coexisting phases. Such studies are still in their infancy, but hold the

Liposome-Based Models for Membrane Rafts

151

promise of unveiling the fundamental mechanisms controlling raft formation and maturation in cells. Among the most important questions are those concerning the size and lifetime of membrane domains, and how these properties are inluenced by membrane composition. Indeed, since the irst observations of rafts, there has been considerable interest in obtaining precise measurements of domain sizes in the nanoscopic regime. We now discuss FRET and SANS methodologies that have been developed for this purpose. 8.2.3.1 Use of FRET to Determine Domain Size Various models have been proposed to extract domain sizes from energy transfer measurements. Nevertheless, experimental data remain rare, a fact that underscores the general dificulty of accounting for the many (and sometimes dificult to measure) parameters affecting FRET eficiency in phaseseparated membranes. In the most rigorous models, these include probe partition coeficients, Förster distances, quantum yields, lifetimes, and transverse bilayer locations, as well as average bilayer molecular areas. Many of these parameters can have different values in the coexisting phases. Monte Carlo (MC) simulations have proven to be a valuable approach: FRET is effectively a geometric problem, and the calculation of FRET eficiency for a static coniguration of donors and acceptors is straightforward. A model needs only to specify the size and morphology of phase domains, after which probe molecules are randomly placed in the bilayer in accordance with their experimental concentrations and partition coeficients, and FRET eficiency is calculated (Towles et al. 2007, Kiskowski and Kenworthy 2007). Frazier and coworkers compared experimental FRET eficiency for bSM/POPC/ Chol 0.35/0.3/0.35 to MC data simulated for a discrete lattice, in which pairwise lipid interaction energies were varied (Frazier et al. 2007). The authors found the best agreement for simulations where phase domains had linear dimensions of a few hundred nanometers. Hof and coworkers developed an off-lattice MC method (Sachl et al. 2011) and measured domain sizes in SM/DOPC/Chol bilayers where the ganglioside GM1 (a minor component) was crosslinked with cholera toxin B subunit, inding that domains varied from 5 to 24 nm in radius depending on SM concentration (Stel et al. 2012). A basic inding of all domain size-dependent FRET models, and veriied by MC simulations, is that small domains effectively reduce the apparent partitioning strength of the acceptor between the coexisting phases. Loura et al. developed a combined analytical and simulation approach that exploited this principle to estimate domain sizes. In the irst step, probe partition coeficients are obtained from domain size-independent methods (e.g., intensity, lifetime, or anisotropy) (Loura et al. 2001). Time-resolved donor decay data in the presence and absence of acceptor is then analyzed globally to recover (among other parameters in the FRET model) the apparent acceptor partition coeficient K A. To the extent that K A recovered from the model does not match the domain size-independent values, an “educated guess” is made as to the size of domains. MC simulations are then used to ine-tune the estimate. de Almeida et al. applied this methodology to PSM/POPC/ Chol (de Almeida et al. 2005), inding that domain size depends on composition within the Ld + Lo region: small (presumably Lo) domains ( 0.003 Å−1. Enhanced scattering was indeed observed upon lowering the temperature through the miscibility transition. However, the scattering curves for 20°C and 25°C could not be successfully

154

Liposomes, Lipid Bilayers and Model Membranes

it using an analytical model for a single, round domain (Anghel et al. 2007). Rather, the multidomain MC model was used, whereby data were it as a weighted sum of simulated curves with different numbers of domains and interpreted accordingly. It was reported that 95% of the ULVs of 1/1/1 composition contain ~30 domains of radius 8 nm, with the remaining 5% of vesicles containing either 1 or 8 domains of 40 or 14 nm radius. Masui et al. studied the same lipid system using sonicated SUVs of ~20 nm diameter, and similarly found that SANS data could not be adequately it with the analytical single domain model (Masui et al. 2008). Instead, a weighted sum of single domain and two domain MC curves provided a good it to the experimental data.

8.3

PHASE DIAGRAMS OF BIOMIMETIC MIXTURES

Throughout the irst decade of the 2000s, considerable effort was put toward solving the phase diagrams of three-component mixtures mimicking the PM composition, primarily using experimental methods reviewed in the previous section. In this section, we will summarize the general indings of those studies. (For the interested reader, detailed information can be found in an extensive compendium of ternary phase diagrams compiled by Marsh 2009.) We organize our discussion around the important observation that for mixtures consisting of a low-TM lipid, a high-TM lipid, and cholesterol, either one or two regions of micron-sized, coexisting domains are observed in microscopy experiments: In every case, gel/luid coexistence is observed at low cholesterol concentrations, but only in some cases is luid/luid coexistence observed at intermediate cholesterol concentrations. However, despite these macroscopic differences, all such mixtures seem to have the same coexistence regions when probed with nanoscopic techniques. Feigenson has classiied these systems as exhibiting Type I or II behavior, to indicate the number of macroscopic coexistence regions (Feigenson 2007).

8.3.1

type II ternary MIxtures: MIcron-sIzed lIquId doMaIns

Type II mixtures have so far been observed only for two particular low-TM lipids that are not abundant in mammalian cell membranes. Speciically, these are lipids with unsaturations in each chain (DOPC), or highly branched chains (DPhPC)—both these lipids have very low TM relative to the biologically abundant low-TM lipids such as POPC and SOPC. Type II phase diagrams are observed for mixtures including DPPC/DOPC/Chol (Veatch et al. 2007b, Buboltz et al. 2007b, Davis et al. 2009, Juhasz et al. 2009); DPPC/DPhPC/Chol (Veatch et al. 2006); DSPC/DOPC/Chol (Zhao et al. 2007a, Heberle et al. 2010); PSM/DOPC/Chol (Veatch and Keller 2005a, Nyholm et al. 2011); SSM/ DOPC/Chol (Farkas and Webb 2010); bSM/DOPC/Chol (Smith and Freed 2009, Petruzielo et al. 2013); and eSM/DOPG/Chol (Vequi-Suplicy et al. 2010). The general characteristics of a Type II phase diagram are exempliied by the room-temperature phase diagram of DSPC/DOPC/Chol (Zhao et al. 2007a) shown in Figure 8.4. At low cholesterol concentrations (χChol < ~0.1), a macroscopic Ld + Lβ(Lβ′) region is observed in GUV images (light shaded region). Phase domains exhibit irregular, straight-line, and branched features, consistent with the presence of a solid phase. At χChol = 0 (i.e., the binary DSPC/DOPC mixture), wide-angle x-ray diffraction data indicates a tilted gel phase (Lβ′) in coexistence with Ld. Experiments in related systems (Mills et al. 2009) show that the addition of cholesterol eliminates the saturated PC chain tilt in a continuous fashion (i.e., without separation of coexisting tilted and untilted phases). The luidus (left-hand) boundary is identical to within error for microscopy and spectroscopy data, and indicates a nearly constant DSPC concentration χDSPC = 0.1 in the luid phase for χChol < 0.3. In contrast, microscopy, spectroscopy, and x-ray data are discrepant with regard to the solidus (righthand) boundary: GUV images are not interpretable for χDSPC > ~0.8, while nanoscopic methods reveal heterogeneity to at least χDSPC = 0.9. Together, these data indicate that at least part of the gel/ luid region is characterized by domains with at least one nanoscopic ( 25%, large round domains were observed. Subsequent experiments revealed that the window for modulated phases depends on composition within the Ld + Lo region (Goh et al. 2013), as discussed in the next section. Building on the terminology introduced in Feigenson (2007), we will refer to a multicomponent mixture with a domain morphology transition as a “Type II M ” system (Figure 8.4). Recent work has found smooth variation of phase boundaries within the four-component space of DSPC/DOPC/POPC/Chol (Konyakhina et  al. 2013). The four-component diagram provides a compelling argument against the view that Ld + Lo nanodomains are a single-phase region. If the macroscopic Type I Ld + Lβ (two phase) region terminates in a critical point, with a “structured single-phase region” above it at higher cholesterol fraction, then as ρ increases, this single-phase region would meet the macroscopic three-phase region found in DOPC-rich compositions, in violation of the phase rule. The solved phase diagram for DSPC/DOPC/POPC/Chol enables precise study of phase-separated samples and their coexisting compositions. In the inal section of this chapter, we discuss signiicant indings from our studies of this Type II M mixture.

158

8.4

Liposomes, Lipid Bilayers and Model Membranes

TOWARD A BETTER RAFT MODEL: INSIGHTS FROM THE TYPE IIM DOMAIN SIZE TRANSITION

The unexpected complexity observed in the four-component DSPC/DOPC/POPC/Chol mixture may have signiicance for rafts in biological systems. The existence of distinct regimes of nanoscopic, macroscopic, and spatially modulated domains might be an important organizing principle in biological membranes. Furthermore, the domain size transition is driven by lipid composition rather than temperature, suggesting a mechanism for homeothermic cells to control domain properties, including size and morphology. We now discuss our recent indings from this Type II M mixture.

8.4.1

Modulated phase patterns result FroM coMpetIng InteractIons

Following the initial observations of modulated phase patterns in DSPC/DOPC/POPC/Chol (Konyakhina et al. 2011), Goh et al. undertook a more detailed investigation of the modulated phase regime, examining samples prepared along two different tielines within the Ld + Lo region (Goh et al. 2013). There are several indings of note from this study. First, the phase fractions obtained from images follow the lever rule, supporting the description of the modulated phase regime as a continuous extension of the macroscopic Ld + Lo region of DSPC/DOPC/Chol. In vesicles exhibiting modulated phase patterns, the Ld phase was always the continuous (percolating) phase regardless of location on the tieline (i.e., a percolation threshold was not observed). At ρ values just large enough to produce round, macroscopic domains, a percolation threshold occurred approximately halfway along the tielines, where phase mole fractions are roughly equal. The transition from uniform GUVs to the appearance of patterned domains was abrupt, occurring within a range of ~5% ρ, which is comparable to the compositional uncertainty inherent in the GUV sample preparation. This observation could have implications for cell membranes, as it suggests a means for cells to abruptly change the size and connectivity of membrane domains through slight changes in local membrane compositions. Interestingly, the compositional window over which modulated phase patterns were observed was found to depend on location within the Ld + Lo region, as shown in Figure 8.5. An example is provided by comparing results for two compositions with large fractions of Lo phase, but located on different tielines. For a tieline located just above the three-phase region, modulated phases were observed over the range 15 < ρ < 35% (Figure 8.5, top panel, red curve). At a higher cholesterol tieline, where the compositions of the coexisting phases are closer and phase properties likely to be more similar, a much wider modulated phase window of 30 < ρ < 65% was observed (Figure 8.5, bottom panel, red curve). For these samples with large fractions of Lo phase, nearly 100% of vesicles displayed modulated phase patterns within the ρ window. As the fraction of Lo phase was reduced (i.e., movement along the tieline toward the Ld phase boundary), the fraction of vesicles showing modulated phases decreased, and the modulated phase window narrowed. The modulated phase patterns observed in GUVs were reproduced in MC simulations based on a competing interactions model. In this model, multiple ields (order parameters) couple to oppose the formation of a single domain, and so can lead to multiple smaller domains or modulated phase domains at equilibrium. A line tension at the domain boundary derives from a local composition ield speciied by the 2D Ising model. Effectively, line tension favors the formation of large, round domains, while a repulsive ield tends to disperse domains. In the original experimental work (Konyakhina et al. 2011), the repulsive ield was considered to be a dipole–dipole repulsion modeled as a Landau–Ginzburg energy functional. Such an interaction is effective only over a distance of a few nanometers, and is unlikely to be responsible for micron-scale periodic patterns. In a subsequent theory and simulation paper, Amazon et al. (2013) considered the effect of bilayer curvature as a repulsive force, modeled with a Helfrich energy functional. In this model, the local composition and curvature ields couple through composition-dependent bilayer material properties that dictate the energetics of bending and stretching in the membrane. In particular, differences

Liposome-Based Models for Membrane Rafts

159

in the bending rigidity of the Ld and Lo phases effectively reduce the energetic cost of domain interface, and can stabilize smaller domains and modulated phases, including striped and honeycomb patterns. The simulated phase patterns closely matched images observed by Konyakhina et al. (2011) and Goh et al. (2013), and reproduced key experimental indings, including the observation that Ld phase is continuous whenever modulated patterns are present. We now discuss several noteworthy indings from this study. Modulated phases were found to be thermodynamically stable by comparing the energy of a single-round domain morphology to that of a modulated-phase morphology. Importantly, this inding indicates that modulated phases are not simply kinetic artifacts. The canonical parameter set that reproduced modulated phase patterns included a line tension of ~0.01 pN and bending moduli of ~10 × 10−19 J (250 k BT) and 100 × 10−19 J (2500 k BT) for the Ld and Lo phases, respectively. The line tension value is of the same order of magnitude as values obtained by analyzing luctuating domain boundaries in GUVs (Esposito et al. 2007, Honerkamp-Smith et al. 2008). However, the bending moduli are at least one order of magnitude larger than experimental values, which range from 10 to 200 k BT (Kuzmin et al. 2005, Semrau et al. 2008). Instead of such a high value for the bending moduli, a second competing interaction not accounted for in the model may combine with the bending energies to compete with line tension. Electrostatic (dipolar) repulsion might be the as yet unidentiied interaction (Amazon et al. 2013). The geometry of phase patterns was closely related to the relative and absolute values of the model parameters. In particular, the difference in bending rigidity between the Ld and Lo phases controlled the width of the Ld lines and the size of the Lo domains. Finally, the model results in a strong prediction that the ratio of vesicle diameter to line tension is a key parameter controlling domain morphology. Taken together, the experimental and simulation results suggest that domain size and morphology are controlled by a mechanism based on line tension. Figure 8.5 shows a schematic interpretation of experimentally observed modulated phase windows. Line tension is presumed to increase monotonically with increasing ρ, here modeled as a linear dependence (black lines). The simulations predict a critical line tension window, within which modulated phases are observed (gray box). For a steep dependence of line tension on ρ, the modulated phase window should be relatively narrow (Figure 8.5, top panel, red box), and occur at lower ρ values. For shallower line tension dependence, the ρ window is expected to widen and shift to larger values of ρ (Figure 8.5, bottom panel, red box).

8.4.2

doMaIn sIze depends on phase thIckness MIsMatch In the nanoscopIc regIMe

While competing interactions between line tension and curvature energies seem to provide a satisfactory explanation for micron-sized domain patterns, the question remains whether similar interactions exist at the nanoscale. To begin to address this question, Heberle et al. investigated lateral and transverse bilayer structure along a ρ trajectory using SANS (Heberle et al. 2013). To examine domain formation, lateral contrast matching conditions were used as described in Section 8.2.3.1. All samples contained 22 mol% Chol, as well as DSPC and low-TM lipid in a 1:1 ratio. To match the SLD of the headgroup and acyl chain regions and eliminate radial contrast (Figure 8.2), 65% of the DSPC was replaced with DSPC-d70. At this composition, the average bilayer SLD matches that of water with 34% D2O. SANS intensity was measured for ρ values from 0% to 100%, over a temperature range of 10–50°C. Using the DSPC/DOPC/Chol (Zhao et  al. 2007a) and DSPC/ POPC/Chol (Heberle et al. 2010) phase diagrams to constrain the compositions and volume fractions of the coexisting phases, SANS curves at 20°C were modeled as spherical shells with a random arrangement of N monodisperse domains, as described in Section 8.2.3.1. This simple model provided a good it to the data, and domain size was found to increase from 68 to 162 Å as ρ varied from 0% to 35%. Interestingly, the scattering curve for the ρ = 100% sample (known to exhibit macroscopic phase separation in GUVs) was not adequately it with a single domain, but could be satisfactorily it with a linear combination of N = 1 and N = 4 curves. This result indicates the presence of a few

160

Domain radius (Å)

Liposomes, Lipid Bilayers and Model Membranes

150

100

50

6.5

7.0

7.5

8.0

Thickness mismatch (Å)

FIGURE 8.6 SANS measurements of domain size and bilayer thickness mismatch are positively correlated. Domain size and thickness mismatch between the Ld and Lo phases were determined by SANS for fourcomponent mixtures of DSPC/DOPC/POPC/Chol.

large domains at equilibrium, consistent with incomplete domain coalescence often observed in luorescence micrographs of GUVs. It is also likely that domain size polydispersity is present at lower values of ρ, though the good quality of the its to a monodisperse model suggests that the size distribution may be relatively narrow. Measurements of bilayer thickness were made at compositions corresponding to the coexisting Ld and Lo phases along the ρ trajectory. The Lo phase thickness was found to be essentially independent of ρ, but the Ld thickness decreased from 38.4 to 35.1 Å as ρ increased from 0% to 100%. The thickness difference between coexisting Ld and Lo phases therefore increased with increasing ρ, such that the domain size correlated positively with the thickness mismatch, as shown in Figure 8.6. Theoretical work predicts a quadratic dependence of line tension on thickness mismatch (Kuzmin et al. 2005). Presumably, increased line tension drives the coalescence of small domains into fewer, larger domains to reduce the total domain perimeter.

8.5

MOVING FORWARD

The study of seemingly simple lipid mixtures continues to generate interesting and unexpected results. In this chapter, we have demonstrated that as few as four lipid components are suficient to produce a remarkable range of phase behavior, domain morphology, and connectivity, with ordersof-magnitude variation in domain sizes. The foundation for understanding these observations is the thermodynamic phase diagram, which allows for systematic investigation of the structural and mechanical properties of coexisting phases. Starting from this foundation, we are now beginning to understand the role of competing order parameters in inluencing domain size and shape. The interplay between experiment and simulation is crucial for understanding the molecular-level details of these interactions. For example, future progress will beneit greatly from the improvement of coarse-grained MD force ields, which currently cannot reproduce phase separation for CG models of the most studied raft-mimicking mixture, DPPC/DOPC/Chol (Davis et al. 2013). Despite four decades of active research, many mysteries of cell membrane organization have yet to be solved. Without exception, biological membranes exhibit transbilayer asymmetry, and it is not known how the phase behavior of the outer and inner lealets is coupled (if at all). Progress in this area has been limited owing to the great technical challenges involved in the preparation of asymmetric liposomes, though recent pioneering work from London and colleagues has paved the way for such studies (Cheng et al. 2009, Cheng and London 2011). Furthermore, the relative inluence of lipid interactions versus protein interactions on bilayer phase behavior has yet to be convincingly

Liposome-Based Models for Membrane Rafts

161

established. Fortunately, the baseline knowledge required to properly interpret such studies—the multicomponent lipid-phase diagram—has dramatically matured in just the past decade.

REFERENCES Ackerman, D., Heberle, F. A., and Feigenson, G. W. 2013. Limited perturbation of a DPPC bilayer by luorescent lipid probes: A molecular dynamics study. Journal of Physical Chemistry B, 117, 4844–4852. Amazon, J. J., Goh, S. L., and Feigenson, G. W. 2013. Competition between line tension and curvature stabilizes modulated phase patterns on the surface of giant unilamellar vesicles: A simulation study. Physical Review E, 87(2), 022708. Anghel, V. N. P. et al. 2007. Scattering from laterally heterogeneous vesicles. II. The form factor. Journal of Applied Crystallography, 40, 513–525. Ariola, F. S. et al. 2009. Membrane luidity and lipid order in ternary giant unilamellar vesicles using a new bodipy-cholesterol derivative. Biophysical Journal, 96(7), 2696–2708. Ayuyan, A. G. and Cohen, F. S. 2006. Lipid peroxides promote large rafts: Effects of excitation of probes in luorescence microscopy and electrochemical reactions during vesicle formation. Biophysical Journal, 91(6), 2172–2183. Bagatolli, L. and Kumar, P. B. S. 2009. Phase behavior of multicomponent membranes: Experimental and computational techniques. Soft Matter, 5(17), 3234–3248. Bakht, O. and London, E. 2007. Detecting ordered domain formation (lipid rafts) in model membranes using Tempo. Methods in Molecular Biology, 398, 29–40. Bakht, O., Pathak, P., and London, E. 2007. Effect of the structure of lipids favoring disordered domain formation on the stability of cholesterol-containing ordered domains (lipid rafts): Identiication of multiple raft-stabilization mechanisms. Biophysical Journal, 93(12), 4307–4318. Baumgart, T., Hess, S. T., and Webb, W. W. 2003. Imaging coexisting luid domains in biomembrane models coupling curvature and line tension. Nature, 425(6960), 821–824. Baykal-Caglar, E. et al. 2012. Preparation of giant unilamellar vesicles from damp lipid ilm for better lipid compositional uniformity. Biochimica et Biophysica Acta, 1818(11), 2598–2604. Bjorkqvist, Y. J. et al. 2005. Domain formation and stability in complex lipid bilayers as reported by cholestatrienol. Biophysical Journal, 88(6), 4054–4063. Boulgaropoulos, B. et  al. 2012. Lipid sorting by ceramide and the consequences for membrane proteins. Biophysical Journal, 102(9), 2031–2038. Brown, A. C., Towles, K. B., and Wrenn, S. P. 2007a. Measuring raft size as a function of membrane composition in PC-based systems: Part 1—Binary systems. Langmuir, 23(22), 11180–11187. Brown, A. C., Towles, K. B., and Wrenn, S. P. 2007b. Measuring raft size as a function of membrane composition in PC-based systems: Part II—Ternary systems. Langmuir, 23(22), 11188–11196. Brown, D. A. 2006. Lipid rafts, detergent-resistant membranes, and raft targeting signals. Physiology, 21, 430–439. Buboltz, J. T. 2007. Steady-state probe-partitioning luorescence resonance energy transfer: A simple and robust tool for the study of membrane phase behavior. Physical Review E, 76(2), 021903. Buboltz, J. T. et al. 2007a. Stern-Volmer modeling of steady-state Forster energy transfer between dilute, freely diffusing membrane-bound luorophores. Journal of Chemical Physics, 127(21), 215101. Buboltz, J. T. et al. 2007b. High-resolution mapping of phase behavior in a ternary lipid mixture: Do lipid-raft phase boundaries depend on the sample preparation procedure? Langmuir, 23(24), 11968–11971. Cao, X., Surma, M. A., and Simons, K. 2012. Polarized sorting and traficking in epithelial cells. Cell Research, 22(5), 793–805. Chen, L., Yu, Z., and Quinn, P. J. 2007. The partition of cholesterol between ordered and luid bilayers of phosphatidylcholine: A synchrotron x-ray diffraction study. Biochimica et Biophysica Acta, 1768(11), 2873–2881. Cheng, H. T. and London, E. 2011. Preparation and properties of asymmetric large unilamellar vesicles: Interlealet coupling in asymmetric vesicles is dependent on temperature but not curvature. Biophysical Journal, 100(11), 2671–2678. Cheng, H. T., Megha, and London, E. 2009. Preparation and properties of asymmetric vesicles that mimic cell membranes: Effect upon lipid raft formation and transmembrane helix orientation. Journal of Biological Chemistry, 284(10), 6079–6092. Chiang, Y. W. et al. 2004. Dynamic molecular structure of DPPC-DLPC-cholesterol ternary lipid system by spin-label electron spin resonance. Biophysical Journal, 87(4), 2483–2496.

162

Liposomes, Lipid Bilayers and Model Membranes

Chiang, Y. W. et al. 2005. New method for determining tie-lines in coexisting membrane phases using spinlabel ESR. Biochimica et Biophysica Acta, 1668(1), 99–105. Collado, M. I. et al. 2005. Domain formation in sphingomyelin/cholesterol mixed membranes studied by spinlabel electron spin resonance spectroscopy. Biochemistry, 44(12), 4911–4918. Davis, R., Kumar, P. B. S., Sperotto, M. M., and Laradji, M. 2013. Predictions of phase separation in threecomponent lipid membranes by the MARTINI force ield. The Journal of Physical Chemistry B, 117(15), 4072–4080. Davis, J. H., Clair, J. J., and Juhasz, J. 2009. Phase equilibria in DOPC/DPPC-d62/cholesterol mixtures. Biophysical Journal, 96(2), 521–539. de Almeida, R. F. et al. 2005. Lipid rafts have different sizes depending on membrane composition: A timeresolved luorescence resonance energy transfer study. Journal of Molecular Biology, 346(4), 1109–1120. de Almeida, R. F. M. et al. 2007. Complexity of lipid domains and rafts in giant unilamellar vesicles revealed by combining imaging and microscopic and macroscopic time-resolved luorescence. Biophysical Journal, 93(2), 539–553. de Almeida, R. F. M., Fedorov, A., and Prieto, M. 2003. Sphingomyelin/phosphatidylcholine/cholesterol phase diagram: Boundaries and composition of lipid rafts. Biophysical Journal, 85(4), 2406–2416. de Kruyff, B. et al. 1973. The effect of the polar headgroup on the lipid-cholesterol interaction: A monolayer and differential scanning calorimetry study. Biochimica et Biophysica Acta, 307(1), 1–19. de Kruyff, B. et al. 1974. Non-random distribution of cholesterol in phosphatidylcholine bilayers. Biochimica et Biophysica Acta, 356(1), 1–7. Demel, R. A. et  al. 1977. Preferential interaction of cholesterol with different classes of phospholipids. Biochimica et Biophysica Acta, 465(1), 1–10. Dietrich, C. et al. 2001. Lipid rafts reconstituted in model membranes. Biophysical Journal, 80(3), 1417–1428. Dimova, R. et  al. 2006. A practical guide to giant vesicles. Probing the membrane nanoregime via optical microscopy. Journal of Physics: Condensed Matter, 18(28), S1151–S1176. Esposito, C. et al. 2007. Flicker spectroscopy of thermal lipid bilayer domain boundary luctuations. Biophysical Journal, 93(9), 3169–3181. Farkas, E. R. and Webb, W. W. 2010. Precise and millidegree stable temperature control for luorescence imaging: Application to phase transitions in lipid membranes. Review of Scientiic Instruments, 81(9), 093704. Feigenson, G. W. 2007. Phase boundaries and biological membranes. Annual Review of Biophysics and Biomolecular Structure, 36, 63–77. Feigenson, G. W. 2009. Phase diagrams and lipid domains in multicomponent lipid bilayer mixtures. Biochimica et Biophysica Acta, 1788(1), 47–52. Feigenson, G. W. and Buboltz, J. T. 2001. Ternary phase diagram of dipalmitoyl-PC/dilauroyl-PC/cholesterol: Nanoscopic domain formation driven by cholesterol. Biophysical Journal, 80(6), 2775–2788. Fidorra, M. et al. 2006. Absence of luid-ordered/luid-disordered phase coexistence in ceramide/POPC mixtures containing cholesterol. Biophysical Journal, 90(12), 4437–4451. Filippov, A., Oradd, G., and Lindblom, G. 2003. The effect of cholesterol on the lateral diffusion of phospholipids in oriented bilayers. Biophysical Journal, 84(5), 3079–3086. Frazier, M. L. et  al. 2007. Investigation of domain formation in sphingomyelin/cholesterol/POPC mixtures by luorescence resonance energy transfer and Monte Carlo simulations. Biophysical Journal, 92(7), 2422–2433. Fuller, D. M. and Zhang, W. 2009. Regulation of lymphocyte development and activation by the LAT family of adapter proteins. Immunological Reviews, 232(1), 72–83. Fung, B. K. K. and Stryer, L. 1978. Surface density determination in membranes by luorescence energytransfer. Biochemistry, 17(24), 5241–5248. Goh, S. L., Amazon, J. J., and Feigenson, G. W. 2013. Toward a better raft model: Modulated phases in the fourcomponent bilayer, DSPC/DOPC/POPC/CHOL. Biophysical Journal, 104(4), 853–862. Gullapalli, R. R., Demirel, M. C., and Butler, P. J. 2008. Molecular dynamics simulations of DiI-C18(3) in a DPPC lipid bilayer. Physical Chemistry Chemical Physics, 10(24), 3548–3560. Halling, K. K. et al. 2008. Cholesterol interactions with luid-phase phospholipids: Effect on the lateral organization of the bilayer. Biophysical Journal, 95(8), 3861–3871. Haluska, C. K. et al. 2008. Combining luorescence lifetime and polarization microscopy to discriminate phase separated domains in giant unilamellar vesicles. Biophysical Journal, 95(12), 5737–5747. Hammond, A. T. et al. 2005. Crosslinking a lipid raft component triggers liquid ordered-liquid disordered phase separation in model plasma membranes. Proceedings of the National Academy of Sciences of the United States of America, 102(18), 6320–6325.

Liposome-Based Models for Membrane Rafts

163

Hancock, J. F. 2006. Lipid rafts: Contentious only from simplistic standpoints. Nature Reviews Molecular Cell Biology, 7(6), 456–462. Heberle, F. A., 2011. Forster Resonance Energy Transfer in Lipid Bilayers: Theory and Applications for Examining Phase Behavior. (PhD). Cornell University. Heberle, F. A. and Feigenson, G. W. 2011. Phase separation in lipid membranes. Cold Spring Harbor Perspectives in Biology, 3, a004630. Heberle, F. A. et al. 2012. Model-based approaches for the determination of lipid bilayer structure from smallangle neutron and x-ray scattering data. European Biophysical Journal, 41(10), 875–890. Heberle, F. A. et al. 2013. Bilayer thickness mismatch controls domain size in model membranes. Journal of the American Chemical Society, 135(18), 6853–6859. Heberle, F. A. et al. 2010. Comparison of three ternary lipid bilayer mixtures: FRET and ESR reveal nanodomains. Biophysical Journal, 99(10), 3309–3318. Heerklotz, H. 2002. Triton promotes domain formation in lipid raft mixtures. Biophysical Journal, 83(5), 2693–2701. Holowka, D. et al. 2005. Lipid segregation and IgE receptor signaling: A decade of progress. Biochimica et Biophysica Acta, 1746(3), 252–259. Holtta-Vuori, M. et al. 2008. BODIPY-cholesterol: A new tool to visualize sterol traficking in living cells and organisms. Trafic, 9(11), 1839–1849. Honerkamp-Smith, A. R. et al. 2008. Line tensions, correlation lengths, and critical exponents in lipid membranes near critical points. Biophysical Journal, 95(1), 236–246. Ionova, I. V., Livshits, V. A., and Marsh, D. 2012. Phase diagram of ternary cholesterol/palmitoylsphingomyelin/palmitoyloleoyl-phosphatidylcholine mixtures: Spin-label EPR study of lipid-raft formation. Biophysical Journal, 102(8), 1856–1865. Juhasz, J., Sharom, F. J., and Davis, J. H. 2009. Quantitative characterization of coexisting phases in DOPC/ DPPC/cholesterol mixtures: Comparing confocal luorescence microscopy and deuterium nuclear magnetic resonance. Biochimica et Biophysica Acta, 1788(12), 2541–2552. Katsaras, J. 1998. Adsorbed to a rigid substrate, dimyristoylphosphatidylcholine multibilayers attain full hydration in all mesophases. Biophysical Journal, 75(5), 2157–2162. Kiskowski, M. A. and Kenworthy, A. K. 2007. In silico characterization of resonance energy transfer for diskshaped membrane domains. Biophysical Journal, 92(9), 3040–3051. Klopfenstein, W. E. et  al. 1974. Differential scanning calorimetry on mixtures of lecithin, lysolecithin and cholesterol. Chemistry and Physics of Lipids, 13(3), 215–222. Konyakhina, T. M. et al. 2011. Control of a nanoscopic-to-macroscopic transition: Modulated phases in fourcomponent DSPC/DOPC/POPC/Chol giant unilamellar vesicles. Biophysical Journal, 101(2), L8–10. Konyakhina, T. M. et  al. 2013. Phase diagram of a 4-component lipid mixture: DSPC/DOPC/POPC/chol. Biochimica et Biophysica Acta, 1828(9), 2204–2214. Korlach, J. et al. 1999. Characterization of lipid bilayer phases by confocal microscopy and luorescence correlation spectroscopy. Proceedings of the National Academy of Sciences of the United States of America, 96(15), 8461–8466. Kučerka, N. et al. 2005. Structure of fully hydrated luid phase DMPC and DLPC lipid bilayers using x-ray scattering from oriented multilamellar arrays and from unilamellar vesicles. Biophysical Journal, 88(4), 2626–2637. Kučerka, N., Nieh, M. P., and Katsaras, J. 2011. Fluid phase lipid areas and bilayer thicknesses of commonly used phosphatidylcholines as a function of temperature. Biochimica et Biophysica Acta, 1808(11), 2761–2771. Kuo, C. and Hochstrasser, R. M. 2011. Super-resolution microscopy of lipid bilayer phases. Journal of the American Chemical Society, 133(13), 4664–4667. Kuzmin, P. I. et al. 2005. Line tension and interaction energies of membrane rafts calculated from lipid splay and tilt. Biophysical Journal, 88(2), 1120–1133. Ladbrooke, B. D. and Chapman, D. 1969. Thermal analysis of lipids, proteins and biological membranes. A review and summary of some recent studies. Chemistry and Physics of Lipids, 3(4), 304–356. Leslie, M. 2011. Mysteries of the cell. Do lipid rafts exist? Science, 334(6059), 1046–1047. Lin, W. C. et  al. 2007. Lipid domains in supported lipid bilayer for atomic force microscopy. Methods in Molecular Biology, 400, 503–513. Loura, L. M. et al. 2009. FRET analysis of domain formation and properties in complex membrane systems. Biochimica et Biophysica Acta, 1788(1), 209–224. Loura, L. M., Fedorov, A., and Prieto, M. 2001. Fluid-luid membrane microheterogeneity: A luorescence resonance energy transfer study. Biophysical Journal, 80(2), 776–788.

164

Liposomes, Lipid Bilayers and Model Membranes

Loura, L. M. et al. 2008. Effects of luorescent probe NBD-PC on the structure, dynamics and phase transition of DPPC. A molecular dynamics and differential scanning calorimetry study. Biochimica et Biophysica Acta, 1778(2), 491–501. Loura, L. M. and Ramalho, J. P. 2009. Fluorescent membrane probes’ behavior in lipid bilayers: Insights from molecular dynamics simulations. Biophysical Reviews, 1(3), 141–148. Marsh, D. 2009. Cholesterol-induced luid membrane domains: A compendium of lipid-raft ternary phase diagrams. Biochimica et Biophysica Acta, 1788(10), 2114–2123. Marsh, D. and Smith, I. C. 1973. An interacting spin label study of the luidizing and condensing effects of cholesterol on lecithin bilayers. Biochimica et Biophysica Acta, 298(2), 133–144. Masui, T., Urakami, N., and Imai, M. 2008. Nano-meter-sized domain formation in lipid membranes observed by small angle neutron scattering. European Physical Journal E, 27(4), 379–389. Mateo, C. R., Acuna, A. U., and Brochon, J. C. 1995. Liquid-crystalline phases of cholesterol lipid bilayers as revealed by the luorescence of trans-parinaric acid. Biophysical Journal, 68(3), 978–987. McMullen, T. P. W., Lewis, R. N. A. H., and McElhaney, R. N. 2004. Cholesterol-phospholipid interactions, the liquid-ordered phase and lipid rafts in model and biological membranes. Current Opinion in Colloid and Interface Science, 8(6), 459–468. Mills, T. T. et al. 2009. Effects of cholesterol and unsaturated DOPC lipid on chain packing of saturated gelphase DPPC bilayers. General Physiology and Biophysics, 28(2), 126–139. Mills, T. T. et al. 2008. Liquid-liquid domains in bilayers detected by wide angle x-ray scattering. Biophysical Journal, 95(2), 682–690. Morales-Penningston, N. F. et al. 2010. GUV preparation and imaging: Minimizing artifacts. Biochimica et Biophysica Acta, 1798(7), 1324–1332. Muddana, H. S. et al. 2011. Atomistic simulation of lipid and DiI dynamics in membrane bilayers under tension. Physical Chemistry Chemical Physics, 13(4), 1368–1378. Munro, S. 2003. Lipid rafts: Elusive or illusive? Cell, 115(4), 377–388. Nyholm, T. K. M. et al. 2011. Construction of a DOPC/PSM/cholesterol phase diagram based on the luorescence properties of trans-parinaric acid. Langmuir, 27(13), 8339–8350. Ono, A. 2010. Relationships between plasma membrane microdomains and HIV-1 assembly. Biology of the Cell, 102(6), 335–350. Pabst, G. et al. 2009. Effect of ceramide on nonraft proteins. Journal of Membrane Biology, 231(2–3), 125–132. Pan, J. et al. 2013. Using small-angle neutron scattering to detect nanoscopic lipid domains. Chemistry and Physics of Lipids, 170–171, 19–32. Pan, J. et al. 2012. Molecular structures of luid phase phosphatidylglycerol bilayers as determined by small angle neutron and x-ray scattering. Biochimica et Biophysica Acta, 1818(9), 2135–2148. Pan, J. et al. 2008. Cholesterol perturbs lipid bilayers nonuniversally. Physical Review Letters, 100(19), 198103. Parton, R. G. and Richards, A. A. 2003. Lipid rafts and caveolae as portals for endocytosis: New insights and common mechanisms. Trafic, 4(11), 724–738. Pathak, P. and London, E. 2011. Measurement of lipid nanodomain (raft) formation and size in sphingomyelin/ POPC/cholesterol vesicles shows TX-100 and transmembrane helices increase domain size by coalescing preexisting nanodomains but do not induce domain formation. Biophysical Journal, 101(10), 2417–2425. Pencer, J. et  al. 2006. Method of separated form factors for polydisperse vesicles. Journal of Applied Crystallography, 39, 293–303. Pencer, J. et al. 2005. Detection of submicron-sized raft-like domains in membranes by small-angle neutron scattering. European Physical Journal E, 18(4), 447–458. Petruzielo, R. S. et  al. 2013. Phase behavior and domain size in sphingomyelin-containing lipid bilayers. Biochimica et Biophysica Acta, 1828(4), 1302–1313. Pokorny, A. et al. 2006. Temperature and composition dependence of the interaction of delta-lysin with ternary mixtures of sphingomyelin/cholesterol/POPC. Biophysical Journal, 91(6), 2184–2197. Rajendran, L. and Simons, K. 2005. Lipid rafts and membrane dynamics. Journal of Cell Science, 118(6), 1099–1102. Repakova, J. et al. 2006. Inluence of pyrene-labeling on luid lipid membranes. Journal of Physical Chemistry B, 110(31), 15403–15410. Sachl, R. et al. 2011. Limitations of electronic energy transfer in the determination of lipid nanodomain sizes. Biophysical Journal, 101(11), L60-L62. Sankaram, M. B. and Thompson, T. E. 1990. Interaction of cholesterol with various glycerophospholipids and sphingomyelin. Biochemistry, 29(47), 10670–10675. Scherfeld, D., Kahya, N., and Schwille, P. 2003. Lipid dynamics and domain formation in model membranes composed of ternary mixtures of unsaturated and saturated phosphatidylcholines and cholesterol. Biophysical Journal, 85(6), 3758–3768.

Liposome-Based Models for Membrane Rafts

165

Semrau, S. et al. 2008. Accurate determination of elastic parameters for multicomponent membranes. Physical Review Letters, 100(8), 088101. Shaw, A. S. 2006. Lipid rafts: Now you see them, now you don’t. Nature Immunology, 7(11), 1139–1142. Sheets, E. D., Holowka, D., and Baird, B. 1999. Critical role for cholesterol in Lyn-mediated tyrosine phosphorylation of FcepsilonRI and their association with detergent-resistant membranes. Journal of Cell Biology, 145(4), 877–887. Silvius, J. R. 2003. Fluorescence energy transfer reveals microdomain formation at physiological temperatures in lipid mixtures modeling the outer lealet of the plasma membrane. Biophysical Journal, 85(2), 1034–1045. Simons, K. and Ikonen, E. 1997. Functional rafts in cell membranes. Nature, 3876633, 569–572. Singer, S. J. and Nicolson, G. L. 1972. The luid mosaic model of the structure of cell membranes. Science, 1754023, 720–731. Skaug, M. J., Longo, M. L., and Faller, R. 2011. The impact of Texas red on lipid bilayer properties. Journal of Physical Chemistry B, 115(26), 8500–8505. Smith, A. K. and Freed, J. H. 2009. Determination of tie-line ields for coexisting lipid phases: An ESR study. Journal of Physical Chemistry B, 113(12), 3957–3971. Stel, M. et  al. 2012. Dynamics and size of cross-linking-induced lipid nanodomains in model membranes. Biophysical Journal, 102(9), 2104–2113. Stockl, M. et  al. 2008. Detection of lipid domains in model and cell membranes by luorescence lifetime imaging microscopy of luorescent lipid analogues. Journal of Biological Chemistry, 283(45), 30828–30837. Tayebi, L. et al. 2012. Long-range interlayer alignment of intralayer domains in stacked lipid bilayers. Nature Materials, 11(12), 1074–1080. Towles, K. B. et al. 2007. Effect of membrane microheterogeneity and domain size on luorescence resonance energy transfer. Biophysical Journal, 93(2), 655–667. Towles, K. B. and Dan, N. 2007. Determination of membrane domain size by luorescence resonance energy transfer: Effects of domain polydispersity and packing. Langmuir, 23(9), 4737–4739. Uppamoochikkal, P., Tristram-Nagle, S., and Nagle, J. F. 2010. Orientation of tie-lines in the phase diagram of DOPC/DPPC/cholesterol model biomembranes. Langmuir, 26(22), 17363–17368. Veatch, S. L. 2007. From small luctuations to large-scale phase separation: Lateral organization in model membranes containing cholesterol. Seminars in Cell and Developmental Biology, 18(5), 573–582. Veatch, S. L., Gawrisch, K., and Keller, S. L. 2006. Closed-loop miscibility gap and quantitative tie-lines in ternary membranes containing diphytanoyl PC. Biophysical Journal, 90(12), 4428–4436. Veatch, S. L. and Keller, S. L. 2003. Separation of liquid phases in giant vesicles of ternary mixtures of phospholipids and cholesterol. Biophysical Journal, 85(5), 3074–3083. Veatch, S. L. and Keller, S. L. 2005a. Miscibility phase diagrams of giant vesicles containing sphingomyelin. Physical Review Letters, 94(14), 148101. Veatch, S. L. and Keller, S. L. 2005b. Seeing spots: Complex phase behavior in simple membranes. Biochimica et Biophysica Acta, 1746(3), 172–185. Veatch, S. L. et  al. 2004. Liquid domains in vesicles investigated by NMR and luorescence microscopy. Biophysical Journal, 86(5), 2910–2922. Veatch, S. L. et al. 2007a. Fluorescent probes alter miscibility phase boundaries in ternary vesicles. Journal of Physical Chemistry B, 111(3), 502–504. Veatch, S. L. et al. 2007b. Critical luctuations in domain-forming lipid mixtures. Proceedings of the National Academy of Sciences of the United States of America, 104(45), 17650–17655. Vequi-Suplicy, C. C. et  al. 2010. Vesicles with charged domains. Biochimica et Biophysica Acta, 1798(7), 1338–1347. Walde, P. et al. 2010. Giant vesicles: Preparations and applications. ChemBioChem, 11(7), 848–865. Wolber, P. K. and Hudson, B. S. 1979. An analytic solution to the Forster energy transfer problem in two dimensions. Biophysical Journal, 28(2), 197–210. Wu, S. H. W. and McConnell, H. M. 1975. Phase separations in phospholipid membranes. Biochemistry, 14(4), 847–854. Yuan, J. et al. 2009. Solution synchrotron x-ray diffraction reveals structural details of lipid domains in ternary mixtures. Physical Review E, 79(3 Pt 1), 031924. Zhao, J. et al. 2007a. Phase studies of model biomembranes: Complex behavior of DSPC/DOPC/cholesterol. Biochimica et Biophysica Acta, 1768(11), 2764–2776. Zhao, J. et  al. 2007b. Phase studies of model biomembranes: Macroscopic coexistence of Lalpha + Lbeta, with light-induced coexistence of Lalpha + Lo phases. Biochimica et Biophysica Acta, 1768(11), 2777–2786.

9

Nanoscale Membrane Mimetics for Biophysical Studies of Membrane Proteins Catherine J. Baker, Ilia G. Denisov, and Stephen G. Sligar

CONTENTS 9.1 Introduction: The Need for Nanodiscs ................................................................................. 167 9.2 Development of Nanodiscs ................................................................................................... 168 9.3 Preparing Nanodiscs ............................................................................................................. 169 9.4 Understanding and Characterizing Nanodiscs ..................................................................... 171 9.5 Membrane Proteins in Nanodiscs ......................................................................................... 172 9.6 Analytical Techniques .......................................................................................................... 174 9.7 Conclusion ............................................................................................................................ 175 References ...................................................................................................................................... 175

9.1

INTRODUCTION: THE NEED FOR NANODISCS

A large fraction of biological activity, from transport to signaling, takes place at the interface between cells and their surroundings. In particular, the cell membrane hosts a large number of proteins that are responsible for this physiological activity. These proteins are major drug targets and have been the focus of extensive research over the years. Typically, they have hydrophilic domains that reside in the aqueous environment within or outside a cell, as well as hydrophobic regions within the bilayer that anchor them to the cell membrane but also make them dificult to express and purify. Methods effective for aqueous soluble proteins often result in aggregation and denaturation when applied to those that reside within the membrane. This dificulty has resulted in the development of various membrane mimetic techniques that strive to incorporate membrane proteins into lipid environments to maintain the folding and function of the target protein while still producing a robust and simple system. The nanoscale membrane mimetics (called “Nanodiscs”) discussed in this chapter are effective for such efforts because they are composed of small discs of lipid bilayer (~10 nm in diameter) stabilized and rendered aqueous soluble by an amphipathic protein around the outer edge of the hydrophobic acyl chains (see Figure 9.1 for illustration). Nanodiscs are water soluble, small enough to contain a protein in its monomeric form, can provide a native-like environment for the protein, and are stable and easy to produce. Additionally, as discs, they are accessible from both sides without a closed interior. Nanodiscs have proven to be an excellent membrane mimetic and solution to the dificulties of working with membrane proteins. The development of Nanodiscs, the approach by which they are made, and examples of the types of proteins and techniques that have found success are described in this chapter.

167

168

Liposomes, Lipid Bilayers and Model Membranes

FIGURE 9.1 (See color insert.) An illustration of the Nanodisc, composed of a portion of lipid bilayer whose outer acyl chains are surrounded by two amphipathic membrane scaffold proteins (MSP).

9.2

DEVELOPMENT OF NANODISCS

The concept of Nanodiscs originated with the high-density lipoproteins (HDLs) that are responsible for the transportation of cholesterol through the bloodstream. HDL is primarily made up of phosphatidylcholine lipids, cholesteryl esters, and apolipoprotein A-I (apoA-I), a lexible, amphipathic protein primarily made up of helices that help stabilize the HDL structure. The recently formed and lipidated HDL is devoid of cholesteryl esters and is roughly discoidal in shape. As the cholesteryl esters are added to the HDL core, the particle becomes spherical (Jonas 2000). Recombinant HDL (rHDL) can be formed in vitro through the combination of lipids and apolipoproteins (Jonas 1986). HDL and rHDL have been the focus of extensive research but the heterogeneity of the resulting particles makes them nonideal for membrane mimetic applications in which monodispersity and homogeneity are key. To address these concerns, the membrane scaffold protein MSP1 was engineered based on apoA-1 with 43 N-terminal residues removed (Bayburt et al. 2002). Further systematic N-terminal truncations from the original MSP1 sequence revealed which portions are required for the assembly of monodisperse Nanodiscs with high yield, and an optimized sequence of MSP1D1 was derived and characterized (Denisov et al. 2004). To aid in the puriication and utility of MSP, epitope tags and cleavage sites were incorporated. MSP has been produced with hexa-histidine or FLAG tags as well as Tobacco Etch virus (TEV) or factor X cleavage recognition sites (Bayburt et al. 2002, Denisov et al. 2004, Grinkova et al. 2010a). To further increase the utility of Nanodiscs, longer MSPs were created by the addition of one or more 22-mer amphipathic helices to the protein (Denisov et al. 2004, Grinkova et al. 2010a). Since the encircling MSPs determine the size of the resulting discoidal lipoprotein particle at optimum lipid loading, the longer MSPs assemble to larger Nanodiscs. Larger lipoprotein particles can accommodate larger membrane proteins, oligomers, or complex assemblies for study. The two commonly used types of MSP are MSP1D1 and MSP1E3. MSP1D1 excludes an additional 11 residues from the MSP1 sequence that do not participate in Nanodisc formation and includes a hexa-histidine tag and TEV recognition site. MSP1E3 includes three additional helices to the middle of the sequence, as well as the afinity and cleavage tags (Grinkova et al. 2010a). Different lipids (and mixtures of lipids) can comprise the lipid bilayer within Nanodiscs and, along with the variations of MSP available, make Nanodiscs a lexible system and useful in a wide variety of biochemical systems and analytical applications (Bayburt and Sligar 2010, Borch and Hamann 2009, Nath et al. 2007, Ritchie et al. 2009). The Nanodiscs themselves are created by self-assembly (Bayburt and Sligar 2002); the organization of the components into the discoidal shape minimizes the interaction of hydrophobic domains of the lipids and MSP with the aqueous environment. A bilayer arrangement of lipids is formed so that the hydrophilic head groups face outward and the hydrophobic tails are inside, away from the water.

169

Nanoscale Membrane Mimetics for Biophysical Studies of Membrane Proteins

The boundaries of this bilayer structure are stabilized by the MSP, which wraps around the perimeter of the disc and shields the hydrophobic portions from the solvent. In this way, the Nanodisc is hydrophilic and soluble in water, while also including an environment suitable to house the hydrophobic domains of membrane proteins (Bayburt and Sligar 2010, Denisov et al. 2004, 2005).

9.3

PREPARING NANODISCS

Nanodiscs are formed by self-assembly from a reconstitution mixture, which is composed of lipids, MSP, detergent, and the target membrane to be incorporated (if applicable). A schematic of the preparation procedure of Nanodiscs is illustrated in Figure 9.2. The most important factors for the optimal assembly of membrane proteins into Nanodiscs include the lipid–protein stoichiometry and the careful choice of detergent and temperature. An optimal lipid–protein stoichiometric ratio can be estimated based on geometric considerations, as has been described (Bayburt and Sligar 2010, Denisov et al. 2004, Grinkova et al. 2010a). The average number of lipids per Nanodisc relects the mean surface area per lipid molecule (0.67–0.70 nm2 for 1-palmitoyl-2-oleoyl-sn-glycero-3phosphocholine (POPC) at ambient conditions) and the diameter of Nanodisc, which is in turn determined by the length of the corresponding scaffold proteins. The choice of lipid or a mixture of lipids may be dictated by the speciic goals of experiments (lipid charge, chemical structure, or composition of lipid mixture). When natural membranes are used (Civjan et al. 2003, Duan et al. 2004), the self-assembly mixture can be doped with extra lipid to optimize the lipid–protein stoichiometric ratio during the assembly, which provides for better yield. The best detergents for Nanodisc self-assembly are those with high critical micelle concentrations (CMCs), which can be easily dialyzed or removed from solution by adsorption to biobeads (Bayburt and Sligar 2010). The ability of the detergent to solubilize lipids and target protein in an optically clear micellar solution is important. Cholate is good for lipid solubilization at a molar ratio of 2:1 or higher (and absolute concentration near CMC or higher), but not all membrane proteins are amenable to cholate solubilization. In this case, mixed detergent systems can be used, where cholate is used for initial solubilization of lipids and MSP, and other detergents (alkyl maltoside or glucoside, polyoxyethylene glycols, phosphocholines, CHAPS, etc.) are used for the target protein. Temperature is also an important parameter for the successful incorporation of membrane proteins into Nanodiscs. Multiple experiments have revealed that the best results and highest yield can be achieved by performing the self-assembly process and detergent removal near the temperature of the main phase transition of the corresponding lipid bilayer, which is 270 K for POPC and 297 K for 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC). The formation of homogeneous discs is dependent on the proper ratio of lipids to MSP, and hence, lipid concentration should be measured precisely in the reconstitution mixture. The stock solution of lipid is quantiied by phosphate analysis. The solvent is then dried from the lipids in a test tube using a stream of nitrogen gas while rotating the tube such that the lipids form a thin, even (a)

(b)

(c)

(d)

5.5 nm

10 nm

FIGURE 9.2 (See color insert.) Depiction of the Nanodisc formation process. (a) Detergent-solubilized MSP and lipids, (b) the removal of detergent, (c) the fractionation and puriication through SEC, and (d) a schematic of the fully formed Nanodisc.

170

Liposomes, Lipid Bilayers and Model Membranes

coating around the sides and bottom. The avoidance of a large clump of dried lipid at the bottom of the test tube will aid in the ensuing detergent solubilization process. The test tube is then put under vacuum overnight to remove any remnant solvent. A solution of cholate is added to the dried lipids in a hot water bath, and vortexing and sonication are used to solubilize the lipids so that no visible ilm or particles remain. A cholate-to-lipid ratio of 2:1 should be used to ensure adequate solubilization. To this mixture, MSP in its aqueous buffer is added according to the optimal ratio based on the lipids used and the size of the MSP (see Table 9.1). The mixture is allowed to incubate for at least 15 min at a temperature above the transition temperature for the lipids. If membrane proteins are to be incorporated into the Nanodiscs, they must irst also be detergent solubilized. The optimal detergent and protein concentrations are not generic and must be determined for each protein individually. The solubilized protein is added to the disc reconstitution mixture at low ratio (e.g., 1:20) compared to MSP. Two MSPs are necessary to form each Nanodisc and an excess of discs increases the probability that the membrane proteins will be incorporated monomerically. However, this ratio is dependent on the goals of the study. If oligomerization is desired, the ratio may be varied. At this stage, everything is solubilized, but Nanodiscs have not yet formed; for the self-assembly to occur, the detergents must be removed. This causes the components to rearrange themselves to minimize the interactions of the hydrophobic domains with the aqueous solvent. Dialysis or polystyrene beads, such as biobeads, can be used to remove the detergents from the reconstitution mixture. The beads are added to the mixture and allowed to incubate with shaking for several hours and then separated from the mixture by iltration or centrifugation. The removed beads should be washed with buffer to remove any disc components that may have remained at the bead surface to ensure maximal yield of Nanodisc formation. To separate the Nanodiscs from any excess components that may not have been properly incorporated, and to ensure the homogeneity of the Nanodiscs, the mixture is fractionated using size-exclusion chromatography (SEC). Protein standards with known Stokes’s diameters such as bovine thyroid thyroglobulin (17 nm), horse spleen ferritin (12.2 nm), bovine liver catalase (10.4 nm), and bovine serum albumin (7.1 nm) are used to calibrate the column. Bare Nanodiscs using MSP1D1 should elute at about the same time as catalase (Bayburt et al. 2002, Denisov et al. 2004, Grinkova et al. 2010a). As mentioned earlier, to achieve homogeneous disc populations, optimal lipid–protein ratios are necessary. This ratio is dependent on both the MSP length and the lipid to be incorporated. If insuficient amount of lipid is used, heterogeneous particles are formed and free MSP is present. Alternatively, having excess lipids above the optimal ratio yields a broadened size distribution and aggregates of unincorporated lipids (Bayburt et al. 2002, Denisov et al. 2004). The optimal ratio for selected lipids is presented in Table 9.1 (Bayburt and Sligar 2010); for other lipids, the ratio can be estimated by calculating the number of lipids that would it into the area of the Nanodisc based on the mean surface area per lipid (Denisov et al. 2004, Grinkova et al. 2010a). Analysis by SEC should show one narrow peak with minimal shouldering or additional aggregates (Figure 9.3). TABLE 9.1 Reconstitution Ratios and Temperatures for the Formation of Nanodiscs for Three Commonly Used Lipids DPPC DMPC POPC

Optimal Ratios for MSP1D1:Lipid

Optimal Ratios for MSP1E3:Lipid

Incubation Temperature

90:1 80:1 65:1

170:1 150:1 130:1

37°C 25°C 4°C

Source: Data from Bayburt, T. H. and Sligar, S. G. 2010. FEBS Lett., 584(9), 1721–1727.

171

Nanoscale Membrane Mimetics for Biophysical Studies of Membrane Proteins Stokes’s diameter (nm)

Absorbance at 280 nm

1.0

>20

17 12

10 7

0.8 0.6 0.4 0.2 0.0 0

10

20

30

40

50

Time (min)

FIGURE 9.3 An example of an SEC of Nanodiscs on a calibrated column.

9.4 UNDERSTANDING AND CHARACTERIZING NANODISCS A variety of different techniques have conirmed the basic structure and dimensions of Nanodiscs. Scintillation counting of tritiated lipids determined the average number of lipids per Nanodisc, which remained consistent across an SEC elution proile (Denisov et al. 2004), indicating that the ratio between lipids and MSP is consistent and reproducible at constant lipid and MSP ratios. As mentioned earlier, the size of the Nanodisc is largely dictated by the length of the MSP, and therefore, about the same number of lipids will be present in Nanodiscs made with the same components (Denisov et al. 2004). The diameters of the Nanodiscs were determined by SEC using standard proteins to create a calibration curve. Accordingly, the Nanodiscs formed with MSP1D1 were estimated to have a Stokes diameter of ~9.7 nm (Bayburt et  al. 2002, Denisov et  al. 2004). Initial size estimates were also made by scanning probe microscopy (SPM) performed on Nanodiscs adhered in a single layer to atomically lat mica, which suggested a width of ~10 nm and a height equivalent to the thickness of a single lipid bilayer (5.5 nm) (Bayburt et al. 2002). Small-angle x-ray scattering (SAXS) further conirmed a diameter of ~10 nm, as well as the discoidal shape of the particles (Denisov et al. 2004, Wlodawer et al. 1979). More recent SAXS and neutron scattering data at different lipid:MSP ratios suggest an elongated oval disc (Skar-Gislinge et al. 2010). Further insight into Nanodisc structure has been revealed by molecular dynamics (MD) (Aksimentiev et al. 2008, Kijac et al. 2010, Shih et al. 2005, 2006, 2008). In the irst study (Shih et al. 2005), four different all-atom simulations were run: three simulations that varied the length of the MSP and one in which the alignment of the MSPs relative to each other around the Nanodisc was changed. The experimentally determined number, 160 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) molecules per Nanodisc (Denisov et al. 2004), was used for each of the simulations. A simulation with MSP1D1 where the irst 11 N-terminal residues are removed formed stable discs. The scaffold protein with 22 N-terminal residues removed had worse vertical alignment of the MSPs and indicated that too many residues may have been deleted. In contrast, Nanodiscs formed with the original MSP1 sequence with no truncations were distorted due to an insuficient number of lipids to ill out the belt. This result indicates that not all the residues in the MSP sequence are necessary for the belt stabilization of the Nanodiscs, conirming what had been experimentally determined through sequence deletions (Denisov et al. 2004). It was also determined that it is unlikely that the Nanodiscs form with the MSP gaps aligned because such a simulation yielded a distorted disc (Shih et al. 2005), although the optimal mutual orientation of the belts could not be determined de novo. Early on, there were several different hypotheses regarding the coniguration of the amphipathic protein around the phospholipids; among them are the picket fence (Phillips et al. 1997) and the

172

Liposomes, Lipid Bilayers and Model Membranes

double-belt conigurations (Wlodawer et al. 1979). Results obtained using polarized internal relection infrared spectroscopy indicated that the two proteins encircled the lipids about the perimeter in accordance with the belt coniguration (Koppaka et  al. 1999). This model was conirmed by magic-angle-spinning solid-state nuclear magnetic resonance (NMR) wherein the arrangement of the prolines in the MSP sequence correlates to what would be expected for the double-belt model, but not that of the picket fence (Li et al. 2006). The results of these independent experiments have built conidence in a model of Nanodiscs as circular portions of lipid bilayer with MSP encircling in the double-belt conformation.

9.5

MEMBRANE PROTEINS IN NANODISCS

The functional properties of membrane proteins can be evaluated through their assembly into Nanodiscs. Several eukaryotic cytochromes P450 have been successfully studied using Nanodiscs (Denisov and Sligar 2011). Detailed biophysical and biochemical characterizations of the stable monomeric human cytochrome P450 (CYP) 3A4 in a phospholipid bilayer were possible with its incorporation into Nanodiscs (Baas et  al. 2004, Denisov et  al. 2006, 2007). These experiments revealed the full spin-state conversion upon substrate binding (with bromocriptine and testosterone), fast oxygen binding and autoxidation, and signiicantly higher stability with respect to the formation of the inactive P420 form. The coincorporation of CYP3A4 and nicotinamide adenine dinucleotide phosphate (NADPH)-cytochrome P450 reductase (CPR) in Nanodiscs made possible the global analysis of the steady-state kinetics and resulted in deconvolution of the stepwise binding constants and fractional contributions of binding intermediates to the overall NADPH consumption and product formation (Denisov et al. 2007, 2009, Frank et al. 2011). The successful coincorporation of CYP73A5 and NADPH P450 reductase into Nanodiscs outlined the general approaches to the high-throughput screening of multiple eukaryotic cytochromes P450 expressed in baculosomes (Duan et al. 2004). Later, it was shown that the simple addition of full-length CPR to CYP3A4-Nanodiscs results in facile incorporation of reductase into Nanodiscs and the formation of functionally active complexes (Grinkova et  al. 2010b). Equilibration of the system takes only 5–10 min at 37°C as monitored by the rate of steady-state NADPH consumption. This approach is fast and straightforward and provides an opportunity for the development of quick and eficient protocols for in vitro screening of mammalian cytochromes P450, especially when the availability of puriied P450 isozyme is a limiting factor. The same approach can be easily extended to other membrane-bound enzymes that require redox partners. Several integral membrane proteins such as G protein-coupled receptors (GPCRs), channels, and transporters have been incorporated in Nanodiscs in monomeric form for structural and functional studies (Alvarez et al. 2010, Bayburt et al. 2007, Velez-Ruiz and Sunahara 2011, Whorton et al. 2008). Predominantly, monomeric states of target proteins in Nanodiscs are easily achieved by using a large molar excess of lipid and scaffold protein to facilitate “dilution” of oligomers and favor monomerization (Bayburt and Sligar 2010). This protocol was irst developed using bacteriorhodopsin from Halobacterium salinarum (Bayburt and Sligar 2003) and was later used for β-2 adrenogenic receptor (Leitz et al. 2006), rhodopsin (Bayburt et al. 2007, 2011, Whorton et al. 2008), proteorhodopsin (Ranaghan et  al. 2011), and other receptors (Inagaki et  al. 2012, Kuszak et  al. 2009). These studies provided for the irst time a stable system to directly address the physicochemical and functional properties of monomeric proteins that normally exist in dimers or oligomers in vivo. Deciphering the role of supramolecular organization of these oligomeric receptors is critically important for the mechanistic understanding of cell-signaling processes as well as for the development of new drugs targeting GPCRs (Gurevich et al. 2012). This goal cannot be achieved without deconvolution of the functional properties of monomers and the role of membrane structure and composition in oligomer formation and function. For these studies, Nanodiscs represent an ideal system that is being used currently in many studies of integral membrane proteins (Dalal et al. 2012, Ishmukhametov et al. 2010, Näsvik Öjemyr et al. 2012, Ye et al. 2010, 2012). Complex biological

Nanoscale Membrane Mimetics for Biophysical Studies of Membrane Proteins

173

machines found in membranes can also be assembled into a Nanodisc, for example, light-harvesting complex II (LHCII) has both luorescent and quenched conformations that are sensitive to aggregation that alters the luorescence signal. LHCII trimers were incorporated into larger Nanodiscs (utilizing the longer scaffold protein) and the absorbance and luorescence spectra indicated that the complex was properly folded and functional within these Nanodiscs without the formation of larger aggregates (Pandit et al. 2011). Similarly, chemoreceptors exist as homodimers, but further oligomerization of these dimers occurs. With controlled stoichiometry, Nanodiscs containing different numbers of dimers were formed and their activities were tested (Boldog et al. 2006). The functions of ligand binding, adaptational modiication, and transmembrane signaling could be accomplished by a single dimer of the chemoreceptor; however, the activation of chemotaxis histidine kinase required the higher-order interaction of three dimers. In previous membrane mimetic techniques, such as vesicles, the number of chemoreceptors could not be controlled but through Nanodiscs, the functionality of the different forms could be determined (Boldog et al. 2006, Li and Hazelbauer 2011). Similarly, Nanodiscs were used to characterize the clusters of histidine kinase CheA, coupling protein CheW, and chemoreceptor Tar suggesting that the core unit of the signaling complex consists of 2(Tar2)3:2 CheW:1 CheA2 (Li et al. 2011). Protein translocation across a cell membrane was studied in Nanodiscs and was found to be accomplished through an assembly of three membrane proteins, SecY, SecE, and SecG. This assembly has been dificult to study ex vivo due to the dificulty in controlling oligomeric states in lipid vesicles (Dalal and Duong 2010). Previous studies led to conlicting results regarding the subunits necessary to accomplish the different tasks of the assembly. By utilizing Nanodiscs, various roles of the proteins could be differentiated. Although SecY can bind the preprotein, a dimer of SecY is necessary for transport (Dalal et al. 2012). Similar techniques were used for rhodopsin, another membrane protein for which it had been unclear if a dimer is necessary for the function. Monomeric rhodopsin was incorporated into Nanodiscs and it was found that a dimer is unnecessary for the phosphorylation by rhodopsin kinase (GRK1) (Bayburt et al. 2011, Velez-Ruiz and Sunahara 2011, Whorton et al. 2008). The structure of Nanodiscs also facilitates the study of some membrane proteins due to the Nanodiscs’ lack of an interior and exterior. For this reason, liposomes and detergent micelles had been ineffective for studying the four-helix bundle histidine kinases, adenylate cyclases, methyl accepting proteins, and phosphatases (HAMP) domain switches responsible for transmitting signals from receptors to output domains (Wang et al. 2012). In Nanodiscs, the incorporated proteins are accessible from both sides of the bilayer, which allowed the different domains to be labeled and their accessibility and conformation to be determined through Förster’s resonance energy transfer (FRET) and electron paramagnetic resonance (EPR) spectroscopy that had been precluded by previous membrane methods (Wang et al. 2012). The F0F1-ATP synthase is a large, transmembrane complex that transports protons across the cell membrane in the synthesis of adenosine triphosphate (ATP). Incorporating F0F1-ATP synthase into Nanodiscs has allowed the mechanism of transport to be elucidated for this large molecular motor. The complex could transport a proton across the Nanodisc bilayer and mutations to the proteins helped elucidate the mechanism by which this translocation occurs (Ishmukhametov et al. 2010). Nanodiscs with photosynthetic reaction centers were used in conjunction with carbon nanotubes. The Nanodiscs aligned along the carbon nanotubes and formed a complex that disassembles in the presence of detergents but reassembles back into a functional complex at the removal of the detergent, which mimics the natural self-repair process (Ham et al. 2010). Voltage-dependent anion channel (VDAC) regulates the exchange between cytosol and the mitochondria, particularly of metabolites and ions. The complex includes a β-barrel made up of 19 transmembrane strands. After incorporation into Nanodiscs, solution NMR was carried out to identify the sites of nicotinamide adenine dinucleotide (NADH) binding and other structural features (Yu et al. 2012). The effect of lipid environment on the photophysics of green proteorhodopsin was determined through the variation of Nanodisc lipids (Ranaghan et al. 2011). Similarly, the head groups of surrounding lipids were found to control the blood-clotting cascade (Tavoosi et al. 2011). Thus, a wide

174

Liposomes, Lipid Bilayers and Model Membranes

variety of membrane proteins have been successfully incorporated into Nanodiscs and the resulting studies have demonstrated signiicant insight into their functional behavior.

9.6

ANALYTICAL TECHNIQUES

Nanodiscs have also opened up a number of analytical and biophysical techniques previously unavailable to the study of insoluble proteins. Membrane proteins have been dificult to study through mass spectrometry (MS) because the detergents that aid in the solubilization of the proteins interfere with the resulting spectra. MS is, however, a valuable proteomic technique. While Nanodiscs alleviate the solubility and detergent problems, the conventional sample preparation technique produces matrix-assisted laser desorption/ionization time of light (MALDI-TOF) spectra with prevalent MSP peaks that overwhelm those of the target membrane protein. An ultrathin layer method, however, was found to remove a signiicant portion of the MSP so that a clear signal can be obtained for the protein target (Marty et al. 2012). Through this methodology, MS is available to Nanodisc-encapsulated membrane proteins. The structure of integral membrane proteins is dificult to acquire from crystallography or NMR due to the combination of their size and tendency to aggregate rather than to form crystals. Hydrogen exchange coupled with MS can provide information about the structure of the membrane proteins despite the previously mentioned dificulties. The amide hydrogens in the peptide backbone are able to exchange with the surrounding solvent. A deuterated solvent exchanges deuterium into the protein backbone and this additional mass can be detected by MS. The protein is then digested and the segments with additional mass indicate which residues are in contact with the solvent or are present in an area of the protein structured against solvent contact. While this method gives less detailed information than x-ray crystallography, some preliminary structural information can be determined through hydrogen exchange. Engen et al. have performed this technique on both MSP itself and the membrane protein γ-glutamyl carboxylase (GGCX) within Nanodiscs (Hebling et al. 2010, Morgan et al. 2011). The MSP was conirmed to have a different structure between its free and Nanodisc conformations, and a predicted topology map was created for GGCX according to the structural information determined by hydrogen exchange. Surface studies and imaging of Nanodiscs and their incorporated proteins can be accomplished through atomic force microscopy (AFM), where cations induce Nanodiscs to lie lat, in a single layer, along the mica or silicon oxide surfaces (Bayburt et al. 1998, Carlson et al. 1997). This is key to being able to faithfully image the discs and proteins and get an accurate measure of their heights. In this way, when imaged using AFM, the thickness of the discs above the mica is indicative of a single Nanodisc layer oriented such that the acyl chains of the lipids are perpendicular to the mica surface and the bilayer of the Nanodisc runs parallel (Carlson et al. 1997). NADPH-CPR and cytochrome P450 2B4 (CYP2B4) are examples of proteins that were studied through incorporation in rHDL particles that were laid down on mica and studied with AFM. Peaks above the rHDL surface along with activity measurements indicated the successful insertion of active P450 reductase into the rHDL bilayer (Bayburt et al. 1998) and AFM of CYP2B4 showed similar peaks 3.5 nm above the rHDL surface (Bayburt et al. 2002). CPR contains a trypsin cleavage site between the catalyst domain and the membrane anchor and the rHDL containing CPR protrusions was reimaged after treatment with trypsin. The results showed a removal of the majority of the topology peaks, providing evidence that the P450 reductase was responsible for the height variation (Bayburt et al. 1998). Nanodiscs provide a convenient manner in which the surface analysis of membrane proteins can be achieved and characterized. Nanodiscs are useful for other analytical techniques such as surface plasmon resonance (SPR) because the encircling MSP prevents lateral diffusion and aggregation of the proteins as well as denaturation or “poisoning” of the surface. This is a key difference between Nanodiscs and other membrane capture methods, such as vesicles, in which this diffusion and the resulting aggregation is a concern. Borch et al. used cholera toxin subunit B (CTB) and glycolipid receptor GM1 to test the

Nanoscale Membrane Mimetics for Biophysical Studies of Membrane Proteins

175

effectiveness of SPR on a Biacore system (Borch et al. 2008). They found that chips coated with antibodies, particularly the anti-tetra-His, were effective for immobilizing Nanodiscs tagged with hexa-histidine to the low cells. Nanodiscs containing GM1 were immobilized in the sample low cell and in order to account for nonspeciic binding, empty discs were immobilized in the reference low cell. When the solution containing CTB was lowed over the cells, the response increased in the sample cell, implying that CTB had bound to the GM1. Binding through the tetra-His antibody prevented leakage from the sensor chip, indicating that the Nanodiscs remained immobilized in the cells over the course of the experiment (Borch et al. 2008). SPR, however, is only useful for the binding of large molecules that cause a signiicant change in the bulk refractive index. Localized surface plasmon resonance (LSPR) can be used to detect the binding of small drug molecules to a protein. A common drug target, CYP3A4, was incorporated into Nanodiscs and drug binding was measured through the coupling of the electronic transitions of the target protein with the surface plasmon. This technique was shown to be of use in large-scale screening of potential drugs to P450 proteins (Das et al. 2009, Zhao et al. 2006). Other structural methods, such as solution and solid-state NMR (Glueck et al. 2009, Kijac et al. 2007, Shenkarev et al. 2009, 2010), have been successfully adapted for working with Nanodiscs. The variety of techniques available allows a wide range of information to be gathered about membrane proteins.

9.7 CONCLUSION We have presented just a few examples of the utilization and application of Nanodiscs. Nanodiscs have found wide application in biophysical studies with a large variety of proteins and analytical techniques. Several review articles describing the Nanodisc technology have been published (Bayburt and Sligar 2010, Borch et al. 2009, Denisov et al. 2011, Nath et al. 2007, Ritchie et al. 2009). Nanodiscs represent an eficient tool for resolving a number of biochemical questions; their ease to both create and manipulate, as well as their dependability and reproducibility have allowed and stimulated signiicant progress in the ield of membrane protein biochemistry. When once many of these areas were closed due to their experimental dificulty, the addition of Nanodisc technology to the toolbox of biochemists permits the exploration of previously excluded areas of study.

REFERENCES Aksimentiev, A. et al. 2008. Computer modeling in biotechnology: A partner in development. Meth. Mol. Biol. (Totowa, NJ, USA), 474(nanostructure design), 181–234. Alvarez, F. J. D., Orelle, C., and Davidson, A. L. 2010. Functional reconstitution of an ABC transporter in nanodiscs for use in electron paramagnetic resonance spectroscopy. J. Am. Chem. Soc., 132(28), 9513–9515. Baas, B. J., Denisov, I. G., and Sligar, S. G. 2004. Homotropic cooperativity of monomeric cytochrome P450 3A4 in a nanoscale native bilayer environment. Arch. Biochem. Biophys., 430(2), 218–228. Bayburt, T. H., Carlson, J. W., and Sligar, S. G. 1998. Reconstitution and imaging of a membrane protein in a nanometer-size phospholipid bilayer. J. Struct. Biol., 123(1), 37–44. Bayburt, T. H., Grinkova, Y. V., and Sligar, S. G. 2002. Self-assembly of discoidal phospholipid bilayer nanoparticles with membrane scaffold proteins. Nano Lett., 2(8), 853–856. Bayburt, T. H. and Sligar, S. G. 2002. Single-molecule height measurements on microsomal cytochrome P450 in nanometer-scale phospholipid bilayer disks. Proc. Natl. Acad. Sci. USA, 99(10), 6725–6730. Bayburt, T.H. and Sligar, S. G. 2003. Self-assembly of single integral membrane proteins into soluble nanoscale phospholipid bilayers. Protein Sci., 12(11), 2476–2481. Bayburt, T. H. and Sligar, S. G. 2010. Membrane protein assembly into nanodisks. FEBS Lett., 584(9), 1721–1727. Bayburt, T. H. et al. 2007. Transducin activation by nanoscale lipid bilayers containing one and two rhodopsins. J. Biol. Chem., 282(20), 14875–14881. Bayburt, T. H. et al. 2011. Monomeric rhodopsin is suficient for normal rhodopsin kinase (GRK1) phosphorylation and arrestin-1 binding. J. Biol. Chem., 286(2), 1420–1428.

176

Liposomes, Lipid Bilayers and Model Membranes

Boldog, T. et al. 2006. Nanodiscs separate chemoreceptor oligomeric states and reveal their signaling properties. Proc. Natl. Acad. Sci. USA, 103(31), 11509–11514. Borch, J. and Hamann, T. 2009. The nanodisc: A novel tool for membrane protein studies. Biol. Chem., 390(8), 805–814. Borch, J. et al. 2008. Nanodiscs for immobilization of lipid bilayers and membrane receptors: Kinetic analysis of cholera toxin binding to a glycolipid receptor. Anal. Chem., 80(16), 6245–6252. Carlson, J. W., Jonas, A., and Sligar, S. G. 1997. Imaging and manipulation of high-density lipoproteins. Biophys. J., 73(3), 1184–1189. Civjan, N. R. et al. 2003. Direct solubilization of heterologously expressed membrane proteins by incorporation into nanoscale lipid bilayers. Biotechniques, 35(3), 556–563. Dalal, K. and Duong, F. 2010. Reconstitution of the secy translocon in nanodiscs. Meth. Mol. Biol. (NY, USA), 619(protein secretion), 145–156. Dalal, K. et al. 2012. Two copies of the secy channel and acidic lipids are necessary to activate the seca translocation ATpase. Proc. Natl. Acad. Sci. USA, 109(11), 4104–4109. Das, A. et al. 2009. Screening of type I and II drug binding to human cytochrome P450-3A4 in nanodiscs by localized surface plasmon resonance spectroscopy. Anal. Chem., 81(10), 3754–3759. Denisov, I. G., Frank, D. J., and Sligar, S. G. 2009. Cooperative properties of cytochromes P450. Pharmacol. Ther., 124(2), 151–167. Denisov, I. G. and Sligar, S. G. 2011. Cytochromes P 450 in nanodiscs. Biochim. Biophys. Acta, Proteins Proteomics, 1814(1), 223–229. Denisov, I. G. et al. 2004. Directed self-assembly of monodisperse phospholipid bilayer nanodiscs with controlled size. J. Am. Chem. Soc., 126(11), 3477–3487. Denisov, I. G. et al. 2005. Thermotropic phase transition in soluble nanoscale lipid bilayers. J. Phys. Chem. B, 109(32), 15580–15588. Denisov, I. G. et  al. 2006. The ferrous-dioxygen intermediate in human cytochrome P450 3A4—Substrate dependence of formation and decay kinetics. J. Biol. Chem., 281(33), 23313–23318. Denisov, I. G. et al. 2007a. Cooperativity in cytochrome P450 3A4—Linkages in substrate binding, spin state, uncoupling, and product formation. J. Biol. Chem., 282(10), 7066–7076. Denisov, I. G. et al. 2007b. The one-electron autoxidation of human cytochrome P450 3A4. J. Biol. Chem., 282(37), 26865–26873. Duan, H. et al. 2004. Co-incorporation of heterologously expressed arabidopsis cytochrome P450 and P450 reductase into soluble nanoscale lipid bilayers. Arch. Biochem. Biophys., 424(2), 141–153. Frank, D. J., Denisov, I. G., and Sligar, S. G. 2011. Analysis of heterotropic cooperativity in cytochrome P450 3A4 using alpha-naphtholavone and testosterone. J. Biol. Chem., 286(7), 5540–5545. Glueck, J. M. et al. 2009. Integral membrane proteins in nanodiscs can be studied by solution NMR spectroscopy. J. Am. Chem. Soc., 131(34), 12060–12061. Grinkova, Y. V., Denisov, I. G., and Sligar, S. G. 2010a. Engineering extended membrane scaffold proteins for self-assembly of soluble nanoscale lipid bilayers. Protein Eng. Des. Sel., 23(11), 843–848. Grinkova, Y. V., Denisov, I. G., and Sligar, S. G. 2010b. Functional reconstitution of monomeric CYP3A4 with multiple cytochrome P450 reductase molecules in nanodiscs. Biochem. Biophys. Res. Commun., 398(2), 194–198. Gurevich, E. V. et al. 2012. G protein-coupled receptor kinases: More than just kinases and not only for GPCRs. Pharmacol. Ther., 133(1), 40–69. Ham, M.-H. et  al. 2010. Photoelectrochemical complexes for solar energy conversion that chemically and autonomously regenerate. Nat. Chem., 2(11), 929–936. Hebling, C. M. et al. 2010. Conformational analysis of membrane proteins in phospholipid bilayer nanodiscs by hydrogen exchange mass spectrometry. Anal. Chem., 82(13), 5415–5419. Inagaki, S. et al. 2012. Modulation of the interaction between neurotensin receptor Nts1 and Gq protein by lipid. J. Mol. Biol., 417(1–2), 95–111. Ishmukhametov, R. et  al. 2010. Direct observation of stepped proteolipid ring rotation in E. coli F0F1-ATP synthase. EMBO J., 29(23), 3911–3923. Jonas, A. 1986. Reconstitution of high-density lipoproteins. Meth. Enzymol., 128, 553–582. Jonas, A. 2000. Lecithin cholesterol acyltransferase. Biochim. Biophys. Acta, 1529(1–3), 245–256. Kijac, A. Z. et al. 2007. Magic-angle spinning solid-state NMR spectroscopy of nanodisc-embedded human CYP3A4. Biochemistry, 46(48), 13696–13703. Kijac, A. et al. 2010. Lipid–protein correlations in nanoscale phospholipid bilayers determined by solid-state nuclear magnetic resonance. Biochemistry, 49(43), 9190–9198.

Nanoscale Membrane Mimetics for Biophysical Studies of Membrane Proteins

177

Koppaka, V. et al. 1999. The structure of human lipoprotein A-I. Evidence for the “Belt” model. J. Biol. Chem., 274(21), 14541–14544. Kuszak, A. J. et  al. 2009. Puriication and functional reconstitution of monomeric mu-opioid receptors: Allosteric modulation of agonist binding by Gi2. J. Biol. Chem., 284(39), 26732–26741. Leitz, A. J. et  al. 2006. Functional reconstitution of beta2-adrenergic receptors utilizing self-assembling nanodisc technology. Biotechniques, 40(5), 601,602,604,606,608,610,612. Li, M. and Hazelbauer, G. L. 2011. Core unit of chemotaxis signaling complexes. Proc. Natl. Acad. Sci. USA, 108(23), 9390–9395. Li, M. et al. 2011. Chemotaxis kinase chea is activated by three neighbouring chemoreceptor dimers as effectively as by receptor clusters. Mol. Microbiol., 79(3), 677–685. Li, Y. et al. 2006. Structural analysis of nanoscale self-assembled discoidal lipid bilayers by solid-state NMR spectroscopy. Biophys. J., 91(10), 3819–3828. Marty, M. T., Das, A., and Sligar, S. G. 2012. Ultra-thin layer MALDI mass spectrometry of membrane proteins in nanodiscs. Anal. Bioanal. Chem., 402(2), 721–729. Morgan, C. R. et al. 2011. Conformational transitions in the membrane scaffold protein of phospholipid bilayer nanodiscs. Mol. Cell. Proteomics, 10(9), 1–11. Näsvik Öjemyr, L. et al. 2012. Reconstitution of respiratory oxidases in membrane nanodiscs for investigation of proton-coupled electron transfer. FEBS Lett., 586(5), 640–645. Nath, A., Atkins, W. M., and Sligar, S. G. 2007. Applications of phospholipid bilayer nanodiscs in the study of membranes and membrane proteins. Biochemistry, 46(8), 2059–2069. Pandit, A. et  al. 2011. Assembly of the major light-harvesting complex II in lipid nanodiscs. Biophys. J., 101(10), 2507–2515. Phillips, J. C. et al. 1997. Predicting the structure of apolipoprotein A-I in reconstituted high-density lipoprotein disks. Biophys. J., 73(5), 2337–2346. Ranaghan, M. J. et al. 2011. Green proteorhodopsin reconstituted into nanoscale phospholipid bilayers (nanodiscs) as photoactive monomers. J. Am. Chem. Soc., 133(45), 18318–18327. Ritchie, T. K. et  al. 2009. Reconstitution of membrane proteins in phospholipid bilayer nanodiscs. Meth. Enzymol., 464(liposomes, part F), 211–231. Shenkarev, Z. O. et al. 2009. Lipid–protein nanodiscs: Possible application in high-resolution NMR investigations of membrane proteins and membrane-active peptides. Biochemistry (Moscow), 74(7), 756–765. Shenkarev, Z. O. et al. 2010. Lipid–protein nanodiscs as reference medium in detergent screening for highresolution NMR studies of integral membrane proteins. J. Am. Chem. Soc., 132(16), 5628–5629. Shih, A. Y. et al. 2005. Molecular dynamics simulations of discoidal bilayers assembled from truncated human lipoproteins. Biophys. J., 88(1), 548–556. Shih, A. Y. et al. 2006. Coarse grained protein–lipid model with application to lipoprotein particles. J. Phys. Chem. B, 110(8), 3674–3684. Shih, A. Y. et al. 2008. Molecular modeling of the structural properties and formation of high-density lipoprotein particles. Comput. Modeling Membr. Bilayers, 60, 313–342. Skar-Gislinge, N. et al. 2010. Elliptical structure of phospholipid bilayer nanodiscs encapsulated by scaffold proteins: Casting the roles of the lipids and the protein. J. Am. Chem. Soc., 132(39), 13713–13722. Tavoosi, N. et  al. 2011. Molecular determinants of phospholipid synergy in blood clotting. J. Biol. Chem., 286(26), 23247–23253. Velez-Ruiz, G. A. and Sunahara, R. K. 2011. Reconstitution of G protein-coupled receptors into a model bilayer system: Reconstituted high-density lipoprotein particles. Meth. Mol. Biol., 756, 167–182. Wang, J. et  al. 2012. HAMP signal relay mechanism in a sensory rhodopsin–transducer complex. J. Biol. Chem., 287(25), 25316–25325. Whorton, M. R. et al. 2008. Eficient coupling of transducin to monomeric rhodopsin in a phospholipid bilayer. J. Biol. Chem., 283(7), 4387–4394. Wlodawer, A. et al. 1979. High-density lipoprotein recombinants: Evidence for a bicycle tire micelle structure obtained by neutron scattering and electron microscopy. FEBS Lett., 104(2), 231–235. Ye, F. et al. 2010. Recreation of the terminal events in physiological integrin activation. J. Cell Biol., 188(1), 157–173. Yu, T.-Y. et al. 2012. Solution NMR spectroscopic characterization of human VDAC-2 in detergent micelles and lipid bilayer nanodiscs. Biochim. Biophys. Acta, Biomembranes, 1818(6), 1562–1569. Zhao, J. et  al. 2006. Resonance surface plasmon spectroscopy: Low molecular weight substrate binding to cytochrome P450. J. Am. Chem. Soc., 128, 11004–11005.

Microemulsions

10 Biomimetic Systems for Characterization of Biomembranes and Their Associated Biomolecules Douglas G. Hayes CONTENTS 10.1 Introduction .......................................................................................................................... 179 10.2 Microemulsions: An Overview ............................................................................................. 180 10.3 Reversed Micelles as Biomimetic Systems .......................................................................... 184 10.4 Bicontinuous Microemulsions: A Potentially Valuable Bimimetic System ......................... 189 10.5 Conclusions ........................................................................................................................... 194 References ...................................................................................................................................... 194

10.1 INTRODUCTION Research in the life sciences from a holistic perspective demonstrates biological events are quite complex, and require understanding on the nanoscale level. Such an understanding of the biological systems will allow twenty-irst century scientists and engineers to harness the power enveloped within nature to address many societal problems: the need for sustainable fuels, chemicals and materials, the ability to detect and combat cancer and other diseases which threaten our ever-aging and increasingly urbanized population with new therapeuticals and new ways to deliver them, and to improve the environmental quality of our resources. Many biological events occur in the vicinity of interfaces, such as reactions and transport of metabolites across membranes, signaling between cells, and the unfolding and refolding of proteins. A deeper understanding of molecular events has inspired scientists in recent years to develop new approaches on the nanoscale level to preparing materials, microelectronics, and biomedical devices, to name a few (Karlsson et  al. 2004). Therefore, the study of “biomimetics” is a worthwhile pursuit, where this term is deined as “. . . the attempt to learn from nature; it deals with development of innovations on the basis of investigation of natural, evolutionarily optimized biological structures, functions, processes, and systems” (von Gleich et al. 2010). A similar deinition was provided in a review: “. . . a science which employs the principles of biochemical organization (i.e., the principles of structural organization, functioning, and regulation of biological systems at levels corresponding to biomacromolecules, supramolecular complexes, and subcellular structures) for the construction of artiicial systems with predetermined properties or for conferring desired properties on natural biochemical systems with the help of artiicial elements” (Kurganov and Topchieva 1991). Membrane proteins: their characterization and utilization is another major application of biomimetic systems. Membrane proteins account for approximately one-third of proteins encoded by the human genome, and are targets for 50–60% of therapeutical agents. Moreover, membrane proteins are

179

180

Liposomes, Lipid Bilayers and Model Membranes

involved with many diseases, such as cystic ibrosis, cancer, diabetes, and kidney diseases (Fagerberg et  al. 2010, Klammt et  al. 2012). Membrane proteins play essential roles in biochemistry, such as molecular transport, signaling, biocatalysis, and interaction and fusion between cells. They are dificult to characterize due to their poor solubility in aqueous media and their need for a biomembrane to obtain their structural integrity. Membrane proteins are dificult to crystallize, making x-ray crystallographic determination of their structure unlikely to be successful; moreover, membrane proteins account for only 3% of the three-dimensional structures (many of which are not to atomic resolution, but to intermediate resolution) determined to date, accounting for only 30–40 structures within the Protein Data Bank (Fagerberg et al. 2010, Bill et al. 2011, Warschawski et al. 2011). Therefore, nuclear magnetic resonance (NMR), a viable alternative to x-ray crystallography for structural characterization, is being investigated (reviewed in Warschawski et al. 2011, Klammt et al. 2012). Due to the complexity of biological systems, several different biomimetic systems have been developed to provide simpliied systems to allow for meaningful scientiic discovery on the nanoscale level to occur while manipulating one variable at a time [reviewed in Monnard et al. (2008)]. Surfactant self-assembly systems are an important tool in the biomimetic science toolbox, to simulate biological membranes. As stated in a recent review, “colloidal systems offer the possibility to have a welldeined chemical composition and precisely characterized structure” (Chopineau et  al. 1998). In addition, the self-assembly process may be a key component of Origin of Life (reviewed in Luisi 2000, Walde 2006). The cited reviews suggest many surfactants may have resided during the prebiotic era, formed under prebiotic conditions or brought to earth via meteorites, such as fatty acids, fatty alcohols, and phospholipids. Surfactant self-assembly structures may have led to the formation of cells and organelles. The occurrence of compartmentalization may have played a role in chemical reactions, such as the formation of RNA, due to the partitioning of substrates and products between the bulk phase and the interface (with hydrophobic interactions perhaps playing an important role), and the role of a charged interface in catalyzing chemical reactions. The self-replication and selfrepair of surfactant self-assembly structures is also relevant (Monnard et al. 2008). Surfactant self-assembly structures employed in biomimetic research include bilayer- and surfactant/water-based self-assembly structures (vesicles, liposomes and giant unilamellar vesicles micelles), micelles, liquid crystalline phases [e.g., lamellar (L α), hexagonal (H1), and bicontinuous cubic (QII) (Nicot et al. 1996, Libster et al. 2011, Tenchov and Koynova 2012)], macroemulsions (typically water-in-oil, or “w/o”), and microemulsions (Monnard et al. 2008). Lamellar systems are well-known as potentially useful systems for solubilizing membrane proteins for their subsequent study and characterization (Nicot et al. 1996). Of the surfactant self-assembly structures listed above, microemulsions possess many attributes that render their use as biomimetic systems: isotropic, thermodynamically stable, optically transparent, and high dynamicity (e.g., rapid exchange of contents between aggregates upon collision), to name a few. In addition, microemulsions have been useful as delivery vehicles for drugs, nutraceuticals, and other food and pharmaceutical-related agents (Garti 2003, Spernath and Aserin 2006). The purpose of this chapter is to irst provide a brief overview of microemulsions: their formation, their properties, and their applications. Subsequently, the employment of w/o-microemulsions (reversed micelles) as biomimetic systems will be described. (o/w-Microemulsions are employed as biomimetic systems much less frequently, Krieg and Whitten 1984a,b.) In addition, bicontinuous microemulsions, a potentially valuable biomimetic system that has been underutilized for this role, will be described, with the brief history of this unique system in relation to biomimetics being discussed.

10.2

MICROEMULSIONS: AN OVERVIEW

The fundamentals of microemulsion systems have been thoroughly reviewed elsewhere (Sjöblom et  al. 1996). In this section, a brief introduction and overview of microemulsions will be given. Microemulsions are isotropic nanometer-sized thermodynamically stable dispersions of oil or water formed via surfactant monolayers. As depicted in Figure 10.1, dispersions consist of nanodroplets of

181

Microemulsions o/wMicroemulsions

w/oMicroemulsions

Bicontinuous Microemulsions w/oMicroemulsions

o/wMicroemulsions

Winsor-I

Winsor-II

Winsor-III

Bicontinuous Microemulsions

Lamellar (Lα) Winsor-IV

FIGURE 10.1 (See color insert.) Winsor-type microemulsion systems.

oil-in-water (o/w-) or water-in-oil (w/o-), or bicontinuous microemulsions (the properties of which will be discussed below). Microemulsions can be monophasic, or exist in two- or three-phase systems according to the nomenclature identiied by Winsor over 60 years ago (Figure 10.1), with interfacial tensions between phases frequently being ultralow, 1 occurs as a 0 is decreased and v increased, resulting in negative curvature and therefore the formation of w/o- microemulsions. Therefore, w/o-microemulsions form when v is increased, for example, by employing surfactants with multiple alkyl chains, branching in their chains, or employment of co-surfactants, which solubilize in the tail region. When 0.5 < P ~ 1, more planar interfaces (zero curvature) occur, leading to the formation of bicontinuous microemulsions or L α phases. In addition to the molecular architecture, the relative HLB of surfactant can be strongly affected by environmental parameters. Temperature is particularly important for alkyl ethoxylates. As temperature is increased, ethylene oxide groups rearrange, leading to the loss of water molecules of hydration, hence to a decrease of hydrophilicity and of a 0, resulting in an increase of P. In contrast, an increase of temperature increases the hydrophilicity of ionic surfactants. Ionic surfactants are more strongly affected by an increase of salinity, becoming more lipophilic due to the Debye shielding of head groups, analogous to an increase of temperature for alkyl ethoxylates. To further understand the physicochemical nature of surfactants, several generalized phase diagrams of a water/oil/surfactant system are depicted in Figure 10.3. Figure 10.3a illustrates a

183

Microemulsions (a)

Surfactant (and cosurfactant)

Liquid crystal (e.g., lamellar: Lα) w/o-Microemulsions (reversed micelles)

Bicontinuous Microemulsions

o/wMicroemulsions

Winsor-I Winsor-II Winsor-III Water

Oil

(b)

Temperature

Winsor-II Winsor-III

Liquid crystal (e.g., lamellar: Lα)

Winsor-I Winsor-IV

(c) Winsor-II Temperature

Lα

W

in

so r

-IV

Lα

0

Winsor-I

Oil/(water + oil)

1

FIGURE 10.3 Generalized depiction of phase diagrams for water/surfactant/oil systems. (a) Ternary phase diagram at a constant temperature, (b) “ish” phase diagram of surfactant versus temperature at a ixed water/ oil ratio (typically 1:1 g/g), (c) oil fraction (surfactant-free basis) versus temperature at a ixed surfactant concentration. (Reprinted from Curr. Sci., 80(8), Paul, B. K. and Moulik, S. P., Uses and applications of microemulsions, 990–1001, Copyright (2001); Adv. Colloid Interface Sci., 65, Sjoblom, J., Lindberg, R., and Friberg, S. E., Microemulsions–phase equilibria characterization, structures, applications and chemical reactions, 125–287, Copyright (1996), with permission from and Current Science Association (Bangalore, India) and Elsevier (Amsterdam, the Netherlands), respectively.)

184

Liposomes, Lipid Bilayers and Model Membranes

water-surfactant-oil ternary phase diagram at a ixed temperature. Actual phase diagrams may differ signiicantly depending on the chemical nature of the components, temperature, and other environmental factors. As indicated in the igure, o/w- and w/o-microemulsions (Winsor-IV systems) form near the water and oil vertices, respectively, in the near-absence of oil and water, respectively. Multiphasic microemulsion systems form at low surfactant concentrations and oil and water mass fractions far away from 1.0. Winsor-I and -II triangular phase regions merge together at a water:oil ratio near 1.0, forming a one-phase bicontinuous microemulsion (Winsor-IV) phase. As the surfactant concentration is increased, multiphasic systems form, as well as liquid crystalline phases (e.g., L α and hexagonal phases), the latter particularly in low-oil media. Figure 10.3b depicts a “ish” phase diagram for an alkyl ethoxylate surfactant-based system, temperature versus surfactant concentration at a ixed oil–water ratio (near 1:1 g/g), so named due to the shape of the phase boundary. Low surfactant concentrations (above a minimal, or “critical” surfactant concentration) yield a Winsor-III phase region, with its phase boundaries resembling the head of a ish. A Winsor-I to Winsor-III to Winsor-II transition occurs in this region as temperature is increased due to the decreased hydrophilicity of the ethoxylate group, hence to increased lipophilicity for the surfactant. The Winsor-III phase region diminishes to a point as the surfactant concentration is increased. The temperature corresponding to this point is referred to as the HLB temperature, THLB. A further increase of surfactant yields either Winsor-IV (particularly bicontinuous microemulsions) or L α phases (the latter near THLB), resembling the tail of a ish. Figure 10.3c depicts the effect of temperature and the oil mass fraction (on a surfactant-free basis) for a ixed surfactant concentration. The main feature of this phase diagram is a Winsor-IV region extending diagonally from low temperature and low oil to higher temperatures and oil fractions. For alkyl ethoxylates, the Winsor-IV phase consists of o/w-microemulsions, bicontinuous microemulsions, and w/o-microemulsions as at low, moderate, and high temperature, respectively (Figure 10.3c). L α phases are more likely to form at either very high water or oil fractions. Of the microemulsion systems depicted in Figure 10.1, w/o-microemulsions (Winsor-IV) have attracted the greatest interest and employment as biomimetic systems, examples of which will be described in the next section.

10.3

REVERSED MICELLES AS BIOMIMETIC SYSTEMS

Reversed micelles, w/o-microemulsions possessing relatively low volume fractions of water, have been employed frequently as biomimetic systems due to several advantageous properties. Reversed micelles are optically transparent, isotropic solutions of typically monodisperse populations of spherical aqueous nanodroplets dispersed in apolar media via surfactant monolayers. Therefore, reversed micelles are useful for spectroscopic investigations. However, ellipsoidal and rod-like shapes can also be formed, suggesting a possible relationship between how the size plus geometry of both reversed micelles and cellular compartments in vivo control the size and geometry of aggregates formed inside of them (Pileni et al. 1997). Reversed micelles form spontaneously from mixing appropriate amounts of water, oil, and surfactant (as described in the previous section), resulting in a thermodynamically stable system, with self-assembly driven entropically. Reversed micellar solutions are dynamic, with events such as droplets colliding, coalescing, exchanging materials, and then separating, occurring on the timescale of microseconds, typically. The radius of the inner aqueous spherical cores of reversed micelles is typically proportional to the water-surfactant mole ratio, allowing the scientist exquisite control of their size (and also their curvature). Several reversed micellar systems have been developed which are temperature-insensitive over a large temperature range, particularly when employing mixed surfactant systems. The ability of reversed micelles to solubilize biomolecules has been under investigation for over 30 years, with particular interest paid to hosting enzymatic reactions, especially those that involve water-insoluble substrates and/or enzymes, for investigating biomolecules in apolar environments, for extractive puriication of proteins from aqueous broths via formation of Winsor-II systems, and for the refolding of

Microemulsions

185

denatured proteins (Nicot and Waks 1996, Melo et al. 2001, Fadnavis and Deshpande 2002, Orlich and Schomacker 2002, Tonova and Lazarova 2008). Reversed micelles have many additional applications (Sjöblom et al. 1996), including tertiary oil recovery, templating agents for nanoparticle synthesis (Eastoe et al. 2006, Pileni 2006), and hosting chemical reactions requiring multiple phases (Holmberg 2007). Reversed micelles frequently serve as biomimetic systems to simulate the environment near biomembranes. Since reversed micelles and proteins are of the same dimensions, ~1–10 nm, conditions of proteins localized near membranes (e.g., in the interstitial aqueous layer residing between bilayers) are readily mimicked. In both reversed micelles and aqueous environments near surfactant and phospholipid head groups, water behaves quite differently compared to bulk-phase water, possessing lower polarity and greater nucleophilicity, due to the disruption of hydrogen bonds between water molecules by the head groups. Under such “crowded” conditions, proteins, amphiphiles, and other membrane-localized molecules undergo competition for water molecules to enable hydration of their charged moieties. The placement of proteins into geometrically restricted environments often increases their stability, including the retention of secondary structural motifs such as α-helices. The differing behavior of water when conined to such crowded conditions may play a key underlying role in the increased stability often reported. For instance, a recent molecular modeling study of polypeptide octa-alanine in reversed micelles suggests the helical folding for the polypeptide occurs when the nanodroplets are small, due to an insuficient population of free water molecules available for fully hydrating the polypeptide (Abel et al. 2006). In addition, reversed micelle-like aggregates may serve as part of the underlying mechanism for translocation across biomembranes. A further biomimetic application is in the area of Origins of Life. Bachman and Luisi proposed the rapid self-assembly of surfactants to prepare reversed micelles serves as an effective biomimetic system to support autopoiesis theory: the self-creation of membranes as a requirement for life to exist in all living organisms (Bachmann et  al. 1991). Expanding outward, the self-assembly of reversed micelles at a macroscopic interface, for example, into rod-like reversed micelles, may serve as a biomimetic system for self-organization of molecules for forming multilayer ilms (Nelson et al. 2005). Several different biomimetic related-applications are listed in Table 10.1. The list is divided into three subgroups: reactions and binding interactions that occur near biomembranes, large molecular weight bioinspired materials, drug delivery-related applications, and membrane protein characterization. Several of the biochemical reactions involve oxidation and photochemical reactions, which often catalyzed by membrane-associated enzymes. Another example involves the binding between DNA and a luorescent probe DAPI (4′,6′-diamidino-2-phenylindole). Through luorescence spectroscopic analysis of DNA–DAPI binding in reversed micelles, changes in the luorescence of DAPI were identiied, suggesting in restricted environments that such environmentally triggered changes in luorescence can be misinterpreted as binding with DNA (Banerjee and Pal 2008). Protein unfolding and refolding in reversed micelles has been investigated by Waks and coworkers. Their results demonstrate that for several proteins (e.g., cytochrome c, lysozyme, and myelin basic protein) their structural integrity is greatest when solubilized in reversed micelles possessing small size, observed through high compressibility and small volume (Valdez et al. 2001). Taking advantage of recent improvements in molecular dynamic simulations to model the structure and dynamics of reversed micelles, Abel et al. used reversed micelles to predict that cytochrome c undergoes unfolding as the water-surfactant ratio, hence the average size of the microemulsions, increased, and that the unfolding is initiated at loops located near the heme group, leading to a partial opening of the heme crevice (Abel et al. 2010). Several polymeric or larger molecular weight materials have been produced in reversed micelles that mimic the synthesis of similar materials in nature (Table 10.1). Proposed as a biomimetic system for the synthesis of lignin, John and coworkers investigated the peroxidase- or laccase-catalyzed polymerization of 4-ethylphenol and other phenol derivatives in reversed micelles (Karayigitoglu et al. 1995). Due to the phenol substrates residing at the interface, the resultant polymerization also occurs at the interface, yielding polyphenolic spherical nanoparticles (Karayigitoglu et al. 1995). Other examples of large molecular weight biomimetic materials synthesized in reversed micelles as

186

Liposomes, Lipid Bilayers and Model Membranes

TABLE 10.1 Employment of Reversed Micelles as Biomimetic Systems: Examples from the Literature Topic

Peroxidase-catalyzed oxidation of olive oil Oxidation reactions of nicotine, indole derivatives, ibuprofen, and lavanone Photochemical reactions involving porphyrin derivatives Oxidation of cholesterols Chlorophyll

Dynamics of a luorescent probe known to bind with genomic DNA

Unfolding/refolding of proteins and polypeptides Proteolysis of myelin basic protein

Biomimetic Application Reactions and Binding Interactions Naturally occurring enzymatic oxidation in olive oil Role of interface in enhancing oxidation Photochemical reactions in compartmentalized systems in vivo Mimetic of in vivo oxidation of cholesterol Chlorophyll aggregates are a biomimetic of the primary donor for the photosystem II reaction center Helping to differentiate between direct interactions between probe and DNA versus environmental factors in the restricted aqueous environment of microemulsions Role of interfaces in protein unfolding/refolding Role of protection from proteolysis for myelin basic protein when solubilized in the interstitial water layer residing between bilayers

References

Papadimitriou et al. (2011) Chaudhary et al. (1998), Chauhan and Sahoo (1999), Chauhan et al. (1996, 1997) Kurreck (1996), Szelinski et al. (1996), Milanesio et al. (2008) Patel and Mishra (2006), Chauhan et al. (1992) Ceglie et al. (1993)

Banerjee and Pal (2008)

Valdez et al. (2001), Nicot and Waks (1996), Abel et al. (2006, 2010) Nicot et al. (1993)

Large Molecular Weight Bioinspired Materials Peroxidase-catalyzed polymerization Role of interfaces to align Karayigitoglu et al. (1995) of phenols using H2O2 as oxidation monolignols in the formation of agent lignin Formation of calcium bilirubinate in Role of amino acids and interfaces in Shen et al. (2005) the presence of amino acids the formation of gall stones Formation of calcium carbonate Formation of abalone shells and Ganguli et al. (2007) (aragonite) pearls typically occurring under hyperbolic conditions under the sea

Spectroscopic characterization of norharmane (an alkaloid) Photodynamic inactivation of model compounds (e.g., 9,10-dimethylanthracene and tryptophan) by synthesized agents Interaction between drug mimics and HSA near interfaces Serotonin–HSA interactions

Drug Delivery-Related Norharmane as an agent to combat phototoxicity or photocarcinogenicity Photodynamic inactivation of microorganisms (e.g., fungi that cause candidiasis) and cancer cells Role of HSA-drug binding in drug delivery and toxicity Physiology and biochemistry of interactions between serotonin and membrane-associated receptor proteins in mental disorders

Mallick et al. (2011)

Mora et al. (2010), Scalise and Durantini (2004), Tempesti et al. (2008) Desfosses et al. (1991), Zhang et al. (2008a,b,c) Sengupta et al. (2004)

187

Microemulsions

TABLE 10.1 (continued) Employment of Reversed Micelles as Biomimetic Systems: Examples from the Literature Topic Bis(maltolato_oxovanadium(iv), BMOV)

Myelin-associated proteins: Folch-Pi proteolipid myelin basic protein Cytochrome P450/electron transfer protein fusion proteins Membrane proteinsa dissolved in reversed micelles, including those using near-critical short-chain alkanes

a

Biomimetic Application Role of interfaces in the control of insulin receptors’ activity by BMOV (diabetes) Membrane Proteins Localization of proteins encountered in the central nervous system System to investigate redox reactions of heme-containing monooxygenases Characterization of membrane proteins via NMR

References Winter et al. (2012)

Merdas et al. (1998), Nicot et al. (1993) Hirakawa et al. (2010)

Gratkowski et al. (2002), Kielec et al. (2010), Peterson et al. (2005), Valentine et al. (2010), Flynn et al. (2007)

Membrane-soluble peptide MS1, lavodoxin, myristoylated HIV-1 matrix protein and recoverin, KCSAΔC35, and gramicidin A.

listed in Table 10.1 include gallstones, mimicked by the biomineralization of calcium bilirubinate (to better understand their formation) (Shen et al. 2005), and abalone shells and pearls, mimicked by the biomineralization of calcium carbonate (Ganguli et al. 2007). Several drug delivery-related examples are also listed in Table 10.1, with the aim to test behavior of therapeuticals in biomimetic environmental conditions and their binding and/or interactions with drug targets. The employment of reversed micelles as vehicles for oral and transdermal delivery of drugs and nutraceuticals is reviewed elsewhere (Garti 2003, Spernath and Aserin 2006). Chen’s group has investigated the binding of several drugs, or drug models (e.g., gallic acid, esculin, and bergenin), to human serum albumin (HSA) in reversed micelles (Zhang et al. 2008a,b,c). HSA, the most abundant protein in blood plasma, serves as a vehicle for the transport of drugs. It is important that drug-HSA binding is not too strong; else, the concentration of free active ingredient in the bloodstream will be too low, and the release of other HSA-bound drugs via displacement is promoted, thereby leading to possible interactions between the two different drugs. In addition to enhancing solubilization of the drugs, employment of reversed micelles represents an additional biomimetic model for drug delivery: the transport of drugs across cell membranes to interact with encapsulated proteins. The results from Chen et al. are useful for distinguishing the nature of the interaction between drug and HSA (electrostatic, hydrophobic, or both), identifying the site of binding on HSA via luorescence spectroscopy and demonstrating that HSA resides near the reversed micellar interface (Zhang et al. 2008a,b,c). An earlier study by Desfosses et al. on the binding of HSA with oxyphenylbutazone, dansylsarcosine, and hemin provided evidence for the hypothesis that the protein undergoes a surface-induced conformational change, with the surface-induced HSA subpopulation possessing a different binding afinity than the noninterfacially bound HSA subpopulation (Desfosses et al. 1991). HSA was also employed as a model receptor for serotonin in reversed micelles, to mimic the in vivo binding of serotonin, an important neurotransmitter, to receptor proteins localized in biomembranes (Sengupta et al. 2004). In contrast to its formation of different rotamers in aqueous media, in reversed micelles serotonin resided in a single conformation, based on luorescence decay studies (Sengupta et al. 2004). Using reversed micelles as a biomimetic system, Mallick et al. found that norharmane, an alkaloid important as a photosensitizing agent employed in the photodynamic therapy treatment of cancer, producing a cytotoxic effect on cells via production of singlet molecular oxygen, is much more eficient in water-restricted environments such as

188

Liposomes, Lipid Bilayers and Model Membranes

reversed micelles (Mallick et al. 2005). Durantini et al. also employed reversed micelles to evaluate porphyrin derivatives as photodynamic therapy agent to combat tumors and fungal infections such as candidiasis (Scalise and Durantini 2004, Tempesti et al. 2008, Mora et al. 2010). As described above, the pursuit of biomimetic media for solubilizing membrane proteins, to enable their characterization and utilize them as biocatalysts, is worthy. Reversed micelles have been used to successfully solubilize and characterize membrane proteins. Waks and coworkers have investigated proteins contained in myelin within reversed micellar solution for many years, including the Folch-Pi proteolipid, a water-insoluble transmembrane protein (actually, a protein– lipid complex) (Nicot and Waks 1996). In one study, the authors found the Folch-Pi proteolipid when incorporated into reversed micelles prepared using C12E4 retained its helical content (observed via circular dichroism), demonstrating reversed micellar encapsulation did not strongly affect the structure of the protein (Merdas et al. 1998). In contrast, the incorporation of the Folch-Pi proteolipid had a profound effect on the structure and behavior of the reversed micelles: greatly increased attractive interactions and change of the ratio of semiaxes for the ellipsoidal-shaped aggregates, with both trend supporting the hypothesis that the proteins “bridge” together two or more reversed micelles (Merdas et  al. 1998). Waks has also investigated myelin basic protein, which, although being water-soluble and not a membrane protein per se, is encountered in the aqueous interstitial space between adjacent bilayers, and readily unfolds when dissolved in aqueous media (Valdez et al. 2001). Solubilization of myelin basic protein in reversed micelles of smallest size led to the restoration of its inherent structure in vivo, presumably due to the similarities of water molecules near the bilayer surface and near surfactant head groups in reversed micelles (Valdez et al. 2001), and helped protect the protein from proteolysis, presumably mimicking the same protection that would occur in the aqueous interstitial space between biomembranes (Nicot et al. 1993). Hirakawa et al. demonstrated a fusion protein containing cytochrome P450 and its two electron transfer proteins, putidaredoxin and putidaredoxin reductase, successfully catalyzed the hydroxylation of d-camphor (Hirakawa et al. 2010). The addition of alcohol dehydrogenase to the reversed micelles allowed for cofactor regeneration during the hydroxylation reaction (Hirakawa et al. 2010). A more recent pathway of utilization for has been NMR characterization of reverse micellar-solubilized membrane proteins, to determine structural information and site-speciic dynamics (e.g., using 1H, 13C, and 15N—homonuclear or heteronuclear multidimensional NMR approaches). The employment of reversed micelles negates the need for selective deuteration of proteins, as required by other NMR techniques (Flynn et al. 2007). A key advantage of reversed micelles compared to aqueous media to achieve this purpose is the increased resolution (e.g., narrower line widths) due to the decrease of viscosity. However, ultralow-viscosity solvents such as near- or supercritical solvents (e.g., short-chain alkanes, Xe, and CO2) may be necessary to obtain short molecular correlation times for large proteins (Kielec et al. 2010, Warschawski et al. 2011), thereby requiring high pressure (100 bar or greater). Surfactants employed for this purpose should be able to form micelles in aqueous media, to form a complex with membrane proteins and detergents contained in the aqueous preparations, as well as being able to form reversed micelles (Kielec et al. 2010). Successful surfactants for this dual purpose include the cationics DTAB and CTAB (Kielec et al. 2010). The concentration of surfactant should be optimized; a high-surfactant concentration is needed to increase the concentration of reversed micelles, hence the protein concentration, but not overly high, which can cause percolation and strong interdroplet attractive forces, leading to increased viscosity, and/or to perturbation of the relaxation properties of the protein (Kielec et al. 2010). The application of reversed micelles for solution-based NMR analysis of nonmembrane-associated proteins has been reviewed (Flynn et al. 2007). As described in the cited review, ubiquitin has been resolved to the Angstrom level using 15N, 13C multidimentional NMR; also, the motional dynamics of polypeptides has been investigated using 15N-NMR relaxation. The cited paper also describes NMR analysis of their unfolding and refolding, particularly for low temperature-induced denaturation. Several examples of NMR analysis of reversed micellar-encapsulated membrane proteins have been reported in the literature (Table 10.1). Wand and coworkers have successfully solubilized

Microemulsions

189

several membrane proteins in reversed micelles to enable NMR analysis: potassium channel KCSA, myristoylated recoverin (involved with the Ca2+ signaling pathway in vision), myristoylated HIV-1 matrix protein, and lavodoxin (Gratkowski et  al. 2002, Peterson et  al. 2005, Kielec et  al. 2010, Valentine et al. 2010). Methods such as 1H-, 15N- HSQC suggest similarity in structure for the listed membrane proteins between reversed micellar and free solution-solubilized environments. Flynn and coworkers have successfully refolded gramicidin A, an ion channel-functional membrane proteins, in AOT-reversed micelles, and through nuclear Overhauser effect spectroscopy (NOESY) NMR technique found the membrane protein forms homodimers (Van Horn et al. 2008). Van Horn and coworkers have demonstrated that reversed micelles are valuable for investigating proteins in a low-water environment that mimics the environment near biomembranes via NMR (Van Horn et al. 2009). To further enhance the fractional occupancy of proteins in microemulsions, the Van Horn group has taken advantage of the precipitation of major proportions of the reversed micellar-encapsulated water by quenching the solutions possessing low ionic strength solution aqueous subphases at subzero temperature (Van Horn et al. 2009). Despite the large body of work involving reversed micelles as biomimetic systems, they possess several key disadvantages. First, it can be dificult to obtain suficiently high protein concentrations in reversed micelles, given limitations of water solubilization and the occurrence of precipitation between proteins and surfactants, particularly for ionic surfactants. Likewise, although the concentration of encapsulated protein can be increased through increasing the concentration of reversed micelles, through adding additional aqueous solution and surfactant, this process will frequently increase the viscosity of reversed micellar due to clustering between reversed micellar droplets, which ultimately leads to phase separation. The increase of viscosity will cause poor signal-to-noise spectroscopic measurements for reversed micellar-encapsulated proteins. The disadvantage is further exacerbated by the strong absorbance of most surfactants in the low-ultraviolet spectral region (Nicot and Waks 1996). Second, it is frequently not clear if protein-containing, or “illed,” reversed micelles differ in properties such as size and geometry compared to “unilled” reversed micelles. Typically, measured properties of reversed micelles (e.g., size and geometry obtained via dynamic light scattering or small-angle scattering, and inherent chemistry of surfactant head groups or micellar-encapsulate water via NMR, FTIR, or other spectroscopic methods) relect the entire population of nanodroplets. It is very challenging to obtain reversed micellar systems where the fraction of illed reversed micelles is greater than 1%; therefore, measured physical properties more strongly relect unilled reversed micelles. Whether differences exist in physical properties between illed and unilled reversed micelles is a controversial issue that is quite challenging to resolve (Sheu et al. 1986). Third, reversed micelle formation can be quite sensitive to salinity and ion type, particularly at relatively high water and protein concentrations, often leading to challenges to form one-phase Winsor-IV systems from fermentation broths or from solutions that contain crudely puriied proteins.

10.4

BICONTINUOUS MICROEMULSIONS: A POTENTIALLY VALUABLE BIMIMETIC SYSTEM

Bicontinuous microemulsions consist of a three-dimensional network of interconnected nanodomains of oil and of water that are separated by semilexible monolayers of surfactant existing at near-zero curvature. Bicontinuous microemulsions are considered to be “. . . molten precursors of the lyotropic liquid crystalline phase that are adjacent to microemulsions in phase diagrams” (Zemb 2009). A recent review describes that several different bicontinuous microemulsion structures can form, such as interconnected spheres, cylinders, or lamellae; however, discrete spheres or aggregates of other geometry are not present (Zemb 2009). The typical size of bicontinuous structures is on the order 10–100 nm (Komura 2007). The existence of bicontinuous structures has been veriied most directly using freezer fracture electron microscopy, showing oil and water nanodomains that are randomly interconnected (Burauer et  al. 2003, Komura 2007). An example is shown in Figure 10.4. In this igure, the oil nanodomains appear as being shadowed due to the nature of the

190

Liposomes, Lipid Bilayers and Model Membranes

e

200 nm

FIGURE 10.4 Freeze fracture electron micrograph (104 × instrument magniication) of bicontinuous microemulsions formed by the water/C12E5/n-octane system using an equal volume fraction of water and oil and a mass fraction of surfactant equal to 0.06, prepared at 32.62°C. (Reprinted from Colloids Surf., A, 228(1–3), Burauer, S. et al., Bicontinuous microemulsions revisited: A new approach to freeze fracture electron microscopy (FFEM), 159–170, Copyright (2003), with permission from Elsevier (Amsterdam, the Netherlands).)

experimental procedure employed. The water and oil subdomains are clearly distinguishable, with their respective volume fractions being similar, and the geometry of the nanodomains being random but interconnected. Small-angle neutron and x-ray scattering (SANS and SAXS, respectively) are also useful, with the itting of the Teubner–Strey model to the I versus Q data providing evidence of bicontinuity (Figure 10.5) (Komura 2007, Freiberger et al. 2007). Another indicator is the measurement of self-diffusion coeficients for dispersed water and oil via Fourier’s transform-pulsed 40,000

No protein

I (Q), 1/cm

30,000

Bacteriorhodopsin, 1 g/L

20,000

10,000

0 0.003

0.004

Protein No protein Bacteriorhodopsin

0.005 0.006 0.007 0.008 Momentum, transfer, Q, 1/A Quasi-Perioidic Repeat Distance, Å 1407 1147

0.009

0.01

Correlation Length, Å 364 335

FIGURE 10.5 (See color insert.) SANS data for the middle, bicontinuous microemulsion, phase of WinsorIII systems formed using D2O/H2O/AOT/CK-2,13/heptane in the presence and absence of bacteriorhodopsin at 25°C. Proportions of solvents: D2O (0.390% NaCl)/H2O (0.533%)/heptane 1:1:2 v/v/v. Concentration of surfactant (AOT/CK-2,13 1:1 w/w): 2.0 wt%. Curves depict form factor-structure factor model it using the Teubner–Strey form factor. Values of the model-derived parameters are given at the bottom of the igure. Inset depicts Winsor-III system formed using cytochrome c.

191

Microemulsions

gradient spin-echo NMR spectroscopy (Burauer et al. 2003). In o/w- or w/o-microemulsions, the diffusion coeficient for the dispersed phase is signiicantly lower than for the same material existing in bulk phase. However, for bicontinuous microemulsions, the diffusion coeficient for water and oil are of a similar value to their respective values as bulk-phase solvents. Bicontinuous microemulsions have been reported for several surfactant (as well as block copolymer-based) systems (Table 10.2). The incorporation of HSA at low water-surfactant ratios (1 wt% was employed, the extraction of BSA appears to be driven by hydrophobic interactions. For example, even though both AOT and BSA are negatively charged at neutral pH, BSA is extracted favorably under these conditions, supported by several studies demonstrating strong attractive interactions between AOT and BSA in reversed micelles (discussed in Gomez del Rio and Hayes 2011). To recover the proteins from the bicontinuous microemulsions, referred to as “back-extraction,” the following procedure was employed. After the bottom, aqueous excess phase was removed, an aqueous stripping solution of high ionic strength (5.0 wt%) and/or of pH 12.0 was added, with both salinity and an increase of pH negating the attractive interactive force, thereby releasing the protein to the bottom, aqueous excess phase. For some microemulsion systems additional surfactant must be added to maintain Winsor-III formation. Employing this approach, both cytochrome c and α-chymotrypsin were back-extracted from the middle phase at >75% by mass, with the speciic activity of recovered α-chymotrypsin being >90% of its original value. Winsor-III extraction is a viable alternative to the extraction of proteins using Winsor-II-based protein extraction. Developed in the 1980s (reviewed in Mazzola et al. 2008, Ono and Goto 2005, Pires et al. 1996, Tonova and Lazarova 2008), proteins are extracted into the top, w/o-microemulsion phase via speciic interactions between surfactant and protein. The latter method has several disadvantages that can be potentially overcome by WinsorIII-based protein extraction. Higher protein concentrations can occur in the microemulsion phase, thereby enhancing the overall eficiency for the process, particularly for extraction of larger and more hydrophobic proteins (e.g., membrane proteins). Also, the lower interfacial curvature can lower the susceptibility to denaturation and enhance the effectiveness of back-extraction. A disadvantage is the need to operate within a narrow temperature window near the HLB temperature, the latter of which may differ slightly from empty microemulsions when protein is present. To verify the occurrence of bicontinuous microemulsions in the middle phase of the Winsor-III systems formed by water (D2O + H2O)/AOT/CK-2,13/n-heptane in the presence and absence of proteins, SANS was employed. The system composition was modiied to allow for a larger volume fraction of middle phase than that depicted in the inset of Figure 10.5, of 20–30%, to enable isolation of the middle phase into the neutron beam. The SANS data are generally it well by the Teubner–Strey form factor-based model (Figure 10.5). One of the parameters obtained from the Teubner–Strey model, the quasiperiodic repeat distance, d, the distance across one oil layer plus one water layer, is on the large end of the spectrum for bicontinuous microemulsions, 1407 Å. The addition of protein, particularly when the initial aqueous phase protein was ≥1 g L −1, led to a signiicant decrease of d and in the correlation length, ε, the latter being proportional to the surface area per volume. This information demonstrates the incorporation of proteins into the bicontinuous microemulsions, perhaps within the surfactant monolayers. An additional, indirect method suggesting incorporation of proteins into the bicontinuous microemulsions is a signiicant decrease of the volume fraction for middle phases that contain proteins. Figure 10.5 depicts SANS data for a middle, bicontinuous microemulsion phase containing bacteriorhodopsin, a water-insoluble membrane protein. The Winsor-III system that formed after mixing the organic solvent solution and an aqueous suspension of bacteriorhodopsin was initially cloudy due to the presence of undissolved protein. However, after a few minutes of gentle agitation, a Winsor-III system with three clear phases formed. As indicated in Figure 10.5, the incorporation of bacteriorhodopsin led to

194

Liposomes, Lipid Bilayers and Model Membranes

a decrease of d (observed by a concurrent increase of the Q position of the peak maximum for the I versus Q data) and a decrease of ε, consistent with the above-mentioned trends. Therefore, this example demonstrates the potential ability to incorporate membrane proteins in bicontinuous microemulsions, in Winsor-III (and perhaps/Winsor-IV) systems, the former of which may serve as a useful medium for hosting biochemical reactions occurring at biomembranes. [Actually, middle phases formed using AOT and CK-2,13 under the conditions listed in Figure 10.5 underwent multiple elastic coherent scattering, due to the intense scattering produced by many bicontinuous microemulsions, leading to broadened I versus Q proiles, hence to a poorer it of the Teubner– Strey equation to the data (Silas and Kaler 2003). As directed by the cited reference, a series of microemulsion samples were prepared using differing ratio of D2O and H2O as polar phase to obtain several different degrees of neutron contrast across the surfactant monolayer. Values of a Teubner–Strey parameter for a given set of microemulsion conditions were plotted versus the scattering length density squared, yielding a nearly linear plot, with the actual value of the parameter determined from extrapolating to a scattering length density squared value of 0. Using this approach, d and ε for the protein-free microemulsion were 1364 Å and 436 Å, respectively (manuscript in preparation).]

10.5

CONCLUSIONS

Microemulsions, particularly reversed micelles, have been useful for the investigation of biomimetic systems that involve biomembranes. Moreover, the localization of biomolecules near the interface, an environment where water molecules behave quite differently than bulk due to the disruption of hydrogen bonding networks between solvent molecules, resembles the microenvironment near phospholipid bilayers. The ability of reversed micelles to form spontaneously, their dynamicity, their optical transparency, and isotropic property render reversed micelles useful for spectroscopic characterization of biomolecules. The self-assembly of surfactants leading to formation of reversed micelles may mimic prebiotic events leading to formation of cell membranes and subcellular compartmentalization. Several different biomimetic applications have been hosted in reversed micelles, such as membrane protein characterization via NMR, photochemical and oxidation reactions, protein refolding, and transport of drugs in blood. Bicontinuous microemulsions have been underutilized as biomimetic media, and have many inherent advantages compared to reversed micelles, such as the potential ability to obtain higher solubility of biomolecules and the possession of planar interfaces that more closely resemble the curvature of bilayers, while maintaining most of the same advantages of reversed micelles listed above.

REFERENCES Abel, S., Waks, M., and Marchi, M. 2010. Molecular dynamics simulations of cytochrome c unfolding in AOT reverse micelles: The irst steps. Eur. Phys. J. E Soft Matter Biol. Phys., 32(4), 399–409. Abel, S. et al. 2006. Structure, stability, and hydration of a polypeptide in AOT reverse micelles. J. Am. Chem. Soc., 128(2), 382–383. Alkhatib, M. H., Hayes, D. G., and Urban, V. S. 2009. Characterization of microemulsion systems formed by a mixed 1,3-dioxolane ethoxylate/octyl glucoside surfactant system. J. Surf. Deterg., 12(3), 277–283. Bachmann, P. A. et al. 1991. Self-replicating micelles: Aqueous micelles and enzymatically driven reactions in reverse micelles. J. Am. Chem. Soc., 113(22), 8204–8209. Backlund, S. and Rantala, M. 1995. Lecithin/propanol-based microemulsions used as media for a cholesterol oxidase-catalyzed reaction. Colloid Polym. Sci., 273(3), 293–297. Banerjee, D. and Pal, S. K. 2008. Excited-state solvation and proton transfer dynamics of DAPI in biomimetics and genomic DNA. J. Phys. Chem. A, 112(32), 7314–7320. Bill, R. M. et al. 2011. Overcoming barriers to membrane protein structure determination. Nat. Biotechnol., 29(4), 335–340. Burauer, S. et al. 2003. Bicontinuous microemulsions revisited: A new approach to freeze fracture electron microscopy (FFEM). Colloids Surf., A, 228(1–3), 159–170.

Microemulsions

195

Cameron, N. R., Flook, K. J., and Wren, S. A. C. 2003. Polymerised bicontinuous microemulsions as stationary phases for capillary electrochromatography. Chromatographia, 57(3/4), 203–206. Ceglie, A. et  al. 1993. NMR study of AOT microemulsion with acetone in the presence of chlorophyll a: Distribution of acetone and role of chlorophyll. Colloids Surf., A, 72, 285–293. Chaudhary, S., Awasthi, A., and Chauhan, S. M. S. 1998. Biomimetic oxidation of nicotine with hydrogen peroxide and 5-ethyllavin mononucleotide perchlorate. Indian J. Chem., Sect. B Org. Chem. Incl. Med. Chem., 37B(3), 294–297. Chauhan, S. M. S., Kalra, B., and Gulati, A. 1997. Biomimetic oxidation of lavanone with hydrogen peroxide catalyzed by 5,10,15,20-tetra-(2′,6′-dichlorophenyl)porphyrinatoiron(III) chloride in CTAB reverse micelles. J. Surf. Sci. Technol., 13(1), 49–57. Chauhan, S. M. and Sahoo, B. B. 1999. Biomimetic oxidation of ibuprofen with hydrogen peroxide catalysed by horseradish peroxidase (HRP) and 5,10,15,20-tetrakis-(2′,6′-dichloro-3′-sulphonatophenyl) porphyrinatoiro n(III) and manganese(III) hydrates in AOT reverse micelles. Bioorg. Med. Chem., 7(11), 2629–2634. Chauhan, S. M. S. et al. 1992. Biomimetic oxidation of cholesterol and related sterols by chemical model for horseradish peroxidase (HRP) in AOT reverse micelles. Indian J. Chem., Sect. B, 31B(12), 837–843. Chauhan, S. M. S. et al. 1996. Biomimetic oxidation of indole-3-acetic acid and related substrates with hydrogen peroxide catalyzed by 5,10,15,20-tetrakis(2′,6′-dichloro-3′-sulfonatophenyl)porphyrinatoiron(III) hydrate in aqueous solution and AOT reverse micelles. J. Mol. Catal. A Chem., 113(1–2), 239–247. Chen, S.-H. and Choi, S. 1998. Mesoscopic scale structures in self-organized surfactant solutions determined by small-angle neutron scattering. Supramol. Sci., 5(3–4), 197–206. Chopineau, J. et al. 1998. Self-evolving microstructured systems upon enzymic catalysis. Biochimie, 80(5–6), 421–435. Chow, P. Y. and Gan, L. M. 2005. Microemulsion polymerizations and reactions. Adv. Polym. Sci., 175(Polymer Particles), 257–298. Desfosses, B. et al. 1991. Ligand binding at membrane mimetic interfaces. Human serum albumin in reverse micelles. Eur. J. Biochem., 199(1), 79–87. Eastoe, J., Hollamby, M. J., and Hudson, L. 2006. Recent advances in nanoparticle synthesis with reversed micelles. Adv. Colloid Interface Sci., 128–130, 5–15. Ellison, C. J. et  al. 2009. Bicontinuous polymeric microemulsions from polydisperse diblock copolymers. J. Phys. Chem. B, 113(12), 3726–3737. Fadnavis, N. W. and Deshpande, A. 2002. Synthetic applications of enzymes entrapped in reverse micelles & organo-gels. Curr. Org. Chem., 6(4), 393–410. Fagerberg, L. et al. 2010. Prediction of the human membrane proteome. Proteomics, 10(6), 1141–1149. Flynn, P. F., Simorellis, A. K., and Van Horn, W. D. 2007. NMR studies of encapsulated macromolecules. Annu. Rep. NMR Spectrosc., 62, 179–219. Frank, C. et  al. 2007. Nonionic surfactants with linear and branched hydrocarbon tails: Compositional analysis, phase behavior, and ilm properties in bicontinuous microemulsions. Langmuir, 23(12), 6526–6535. Freiberger, N. et al. 2007. An attempt to detect bicontinuity from SANS data. J. Colloid Interface Sci., 312(1), 59–67. Ganguli, A. K. et al. 2007. Mimicking the biomineralization of aragonite (calcium carbonate) using reversemicelles under ambient conditions. J. Nanosci. Nanotechnol., 7(6), 1760–1767. Garti, N. 2003. Microemulsions as microreactors for food applications. Curr. Opin. Colloid Interface Sci., 8(2), 197–211. Gomez del Rio, J. A. and Hayes, D. G. 2011. Protein extraction by Winsor-III microemulsion systems. Biotechnol. Prog., 27(4), 1091–1100. Gomez del Rio, J. A., Hayes, D. G., and Urban, V. S. 2010. Partitioning behavior of an acid-cleavable, 1,3-dioxolane alkyl ethoxylate, surfactant in single and binary surfactant mixtures for 2- and 3-phase microemulsion systems according to ethoxylate head group size. J. Colloid Interface Sci., 352, 424–435. Gratkowski, H. et al. 2002. Cooperativity and speciicity of association of a designed transmembrane peptide. Biophys. J., 83(3), 1613–1619. Hathout, R. M. et al. 2010. Microemulsion formulations for the transdermal delivery of testosterone. Eur. J. Pharm. Sci., 40(3), 188–196. Hirakawa, H. et al. 2010. Artiicial self-suficient P450 in reversed micelles. Molecules, 15, 2935–2948. Holmberg, K. 2007. Organic reactions in microemulsions. Eur. J. Org. Chem., (5), 731–742. Holmberg, K., Oh, S. G., and Kizling, J. 1996. Microemulsions as reaction medium for a substitution reaction. Prog. Colloid Polym. Sci., 100(Trends in Colloid and Interface Science X), 281–285.

196

Liposomes, Lipid Bilayers and Model Membranes

Jha, S. K. et al. 2011. Formulation development & characterization of microemulsion drug delivery systems containing antiulcer drug. Int. J. Drug Dev. Res., 3(4), 336–343. Johannsson, R., Almgren, M., and Schomaecker, R. 1993. Diffusion-controlled reactions in bicontinuous microemulsions of AOT and C12E5 (pentaoxyethylene monododecyl ether) studied by triplet-state deactivation. Langmuir, 9(5), 1269–1273. Jones, B. H. and Lodge, T. P. 2009. High-temperature nanoporous ceramic monolith prepared from a polymeric bicontinuous microemulsion template. J. Am. Chem. Soc., 131(5), 1676–1677. Jones, B. H. et al. 2012. Nanoporous poly(3,4-ethylenedioxythiophene) derived from polymeric bicontinuous microemulsion templates. Macromolecules (Washington, DC, U.S.), 45(1), 599–601. Karayigitoglu, C. F. et al. 1995. The morphology of phenolic polymers enzymatically synthesized in surfactant microstructures. Mater. Sci. Eng., C, C2(3), 165–171. Karlsson, M. et al. 2004. Biomimetic nanoscale reactors and networks. Annu. Rev. Phys. Chem, 55, 613–649. Kawano, S. et al. 2009. Propagation of polymer nanosheets from silica opal membrane gaps by thermal polymerization of bicontinuous microemulsions. Chem. Commun. (Cambridge, U.K.), (13), 1688–1690. Kawano, S. et al. 2010. Construction of continuous porous organogels, hydrogels, and bicontinuous organo/hydro hybrid gels from bicontinuous microemulsions. Macromolecules (Washington, DC, U.S.), 43(1), 473–479. Kielec, J. M., Valentine, K. G., and Wand, A. J. 2010. A method for solution NMR structural studies of large integral membrane proteins: Reverse micelle encapsulation. Biochim. Biophys. Acta, Biomembr., 1798(2), 150–160. Klammt, C. et al. 2012. Facile backbone structure determination of human membrane proteins by NMR spectroscopy. Nat. Methods, 9(8), 834–839. Klostermann, M. et  al. 2012. Structure and dynamics of balanced supercritical CO2-microemulsions. Soft Matter, 8(3), 797–807. Kluge, K. et al. 2001. Temperature-insensitive microemulsions formulated from octyl monoglucoside and alcohols: Potential candidates for practical applications. Tenside, Surfactants, Deterg., 38(1), 30–34, 37–40. Komura, S. 2007. Mesoscale structures in microemulsions. J. Phys. Condens. Matter, 19(46), 463101/463101–463101/463130. Krieg, M. and Whitten, D. G. 1984a. Self-sensitized photo-oxidation of protoporphyrin IX and related porphyrins in erythrocyte ghosts and microemulsions: A novel photo-oxidation pathway involving singlet oxygen. J. Photochem., 25(2–4), 235–252. Krieg, M. and Whitten, D. G. 1984b. Self-sensitized photooxidation of protoporphyrin IX and related freebase porphyrins in natural and model membrane systems. Evidence for novel photooxidation pathways involving amino acids. J. Am. Chem. Soc., 106(8), 2477–2479. Kurganov, B. I. and Topchieva, I. N. 1991. Problems of biomimetic chemistry. Biokhimiya (Moscow), 56(11), 1946–1959. Kurreck, H. 1996. EPR studies of photoinduced electron transfer reactions of cyclohexylene and butylene linked porphyrin quinones in reversed micelles and liquid crystals. Mol. Phys. Rep., 13, 37–54. Laupheimer, M. et al. 2011. Bicontinuous microemulsion as reaction medium for ω-transaminase catalysed biotransformations. Tenside, Surfactants, Deterg., 48(1), 28–33. Libster, D., Aserin, A., and Garti, N. 2011. Interactions of biomacromolecules with reverse hexagonal liquid crystals: Drug delivery and crystallization applications. J. Colloid Interface Sci., 356(2), 375–386. Liu, H. et al. 2006. Bicontinuous water-AOT/Tween85-isopropyl myristate microemulsion: A new vehicle for transdermal delivery of cyclosporin A. Drug Dev. Ind. Pharm., 32(5), 549–557. Luisi, P. L. 2000. The relevance of supramolecular chemistry for the origin of life. Adv. Supramol. Chem., 6, 287–307. Lutter, S. et al. 2009. Polymer-modiied bicontinuous microemulsions used as a template for the formation of nanorods. J. Dispersion Sci. Technol., 30(6), 745–752. Mallick, A., Das, P., and Chattopadhyay, N. 2011. Photophysics of norharmane in solution phase: From homogeneous to microheterogeneous environments. J. Photochem. Photobiol., C, 11(2–3), 62–72. Mallick, A. et  al. 2005. Photophysical study of 3-acetyl-4-oxo-6,7-dihydro-12H-indolo[2,3-a]quinolizine in biomimetic reverse micellar nanocavities: A spectroscopic approach. J. Phys. Chem. B, 109(30), 14675–14682. Mazzola, P. G. et al. 2008. Liquid-liquid extraction of biomolecules: An overview and update of the main techniques. J. Chem Tech. Biotechnol., 83(2), 143–157. Melo, E. P., Aires-Barros, M. R., and Cabral, J. M. S. 2001. Reverse micelles and protein biotechnology. Biotechnol. Annu. Rev., 7, 87–129. Merdas, A. et al. 1998. Bridging of nonionic reverse micelles by a myelin transmembrane protein. J. Phys. Chem. B, 102(3), 528–533.

Microemulsions

197

Milanesio, M. E. et  al. 2008. Photophysical characterization and photodynamic activity of metallo 5-(4-(trimethylammonium)phenyl)-10,15,20-tris(2,4,6-trimethoxyphenyl)porphyrin in homogeneous and biomimetic media. Photochem. Photobiol. Sci., 7(8), 963–972. Monduzzi, M., Caboi, F., and Moriconi, C. 1997. On the bicontinuous microstructure induced by a guest protein in a typical AOT microemulsion. Colloids Surf., A, 129,130, 327–338. Monnard, P. A., DeClue, M. S., and Ziock, H. J. 2008. Organic nano-compartments as biomimetic reactors and protocells. Curr. Nanosci., 4(1), 71–87. Moon, D.-S. and Lee, J.-K. 2012. Tunable synthesis of hierarchical mesoporous silica nanoparticles with radial wrinkle structure. Langmuir, 28(33), 12341–12347. Mora, S. J. et al. 2010. The photodynamic activity of a novel porphyrin derivative bearing a luconazole structure in different media and against Candida albicans. Dyes Pigm., 87(3), 234–240. Morkved, T. L. et al. 2001. Static and dynamic scattering from ternary polymer blends: Bicontinuous microemulsions, Lifshitz lines, and amphiphilicity. J. Chem. Phys., 114(16), 7247–7259. Mortensen, K. 2001. PEO-related block copolymer surfactants. Colloids Surf., A, 183–185, 277–292. Naoui, W. et  al. 2011. Microemulsion microstructure inluences the skin delivery of an hydrophilic drug. Pharm. Res., 28(7), 1683–1695. Nelson, M. et  al. 2005. Periodic arrays of interfacial cylindrical reverse micelles. Langmuir, 21(22), 9799–9802. Nicot, C. and Waks, M. 1996. Proteins as invited guests of reverse micelles: Conformational effects, signiicance, applications. Biotechnol. Genet. Eng. Rev., 13, 267–314. Nicot, C. et al. 1993. Limited proteolysis of myelin basic protein in a system mimetic of the myelin interlamellar aqueous space. J. Neurochem., 60(4), 1283–1291. Nicot, C. et al. 1996. Squeezing of oil-swollen surfactant bilayers by a membrane protein. Phys. Rev. Lett., 77(16), 3485–3488. Ono, T. and Goto, M. 2005. Bioseparation through liquid–liquid interfaces. In: Watarai, H., Teramae, N., and Sawada, T. (eds.), Interfacial Nanochemistry: Molecular Science and Engineering at Liquid–Liquid Interfaces, Chapter 14, Springer, Berlin. Orlich, B. and Schomacker, R. 2002. Enzyme catalysis in reverse micelles. Adv. Biochem. Eng./Biotechnol., 75(History and Trends in Bioprocessing and Biotransformation), 185–208. Papadimitriou, V. et al. 2011. Microemulsions based on virgin olive oil: A model biomimetic system for studying native oxidative enzymatic activities. Colloids Surf., A, 382(1–3), 232–237. Patel, S. and Mishra, B. K. 2006. Oxidation of cholesterol by a biomimetic oxidant, cetyltrimethylammonium dichromate. J. Org. Chem., 71(9), 3522–3526. Paul, B. K. and Moulik, S. P. 2001. Uses and applications of microemulsions. Curr. Sci., 80(8), 990–1001. Peterson, R. W., Lefebvre, B. G., and Wand, A. J. 2005. High-resolution NMR studies of encapsulated proteins in liquid ethane. J. Am. Chem. Soc., 127(29), 10176–10177. Pileni, M. P. 2006. Reverse micelles used as templates: A new understanding in nanocrystal growth. J. Exp. Nanosci., 1(1–4), 13–27. Pileni, M. P., Tanori, J., and Filankembo, A. 1997. Biomimetic strategies for the control of size, shape and selforganization of nanoparticles. Colloids Surf., A, 123–124, 561–573. Pires de Campos Araujo, L. M., Thomazine, J. A., and Fonseca Vianna Lopez, R. 2010. Development of microemulsions to topically deliver 5-aminolevulinic acid in photodynamic therapy. Eur. J. Pharm. Biopharm., 75(1), 48–55. Pires, M. J., Aires-Barros, M. R., and Cabral, J. M. S. 1996. Liquid-liquid extraction of proteins with reversed micelles. Biotechnol. Prog., 12(3), 290–301. Reimer, J. et al. 2003. Microstructure of Alkyl glucoside microemulsions: Control of curvature by interfacial composition. Langmuir, 19(26), 10692–10702. Reyes, P. Y. et  al. 2010. Synthesis of silver nanoparticles by precipitation in bicontinuous microemulsions. J. Nanomater., Article 948941 (7pp.). Sathishkumar, M. et al. 2008. Bicontinuous microemulsion as reaction medium for the β-glucosidase-catalyzed synthesis of n-hexyl-β-d-glucopyranoside. Enzyme Microb. Technol., 42(3), 252–258. Sathishkumar, M. et al. 2010. Role of bicontinuous microemulsion in the rapid enzymatic hydrolysis of (R,S)ketoprofen ethyl ester in a micro-reactor. Bioresour. Technol., 101(20), 7834–7840. Scalise, I. and Durantini, E. N. 2004. Photodynamic effect of metallo 5-(4-carboxyphenyl)-10,15,20-tris(4methylphenyl) porphyrins in biomimetic AOT reverse micelles containing urease. J. Photochem. Photobiol., A, 162(1), 105–113. Schubert, K. V. and Strey, R. 1991. Small-angle neutron scattering from microemulsions near the disorder line in water/formamide-octane-CiEj systems. J. Chem. Phys., 95(11), 8532–8545.

198

Liposomes, Lipid Bilayers and Model Membranes

Sengupta, B. et al. 2004. Characterization of serotonin in protein and membrane mimetic environments: A spectroscopic study. Chem. Biodivers, 1(6), 868–877. Shen, Y. et al. 2005. Biomimetic synthesis of calcium bilirubinate in different inverse microemulsions. Synth. React. Inorg., Met.-Org., Nano-Met. Chem., 35(5), 359–364. Sheu, E. et al. 1986. Small-angle neutron scattering studies of protein-reversed micelle complexes. Biotechnol. Prog., 2(4), 175–186. Silas, J. A. and Kaler, E. W. 2003. Effect of multiple scattering on SANS spectra from bicontinuous microemulsions. J. Colloid Interface Sci., 257(2), 291–298. Sjöblom, J., Lindberg, R., and Friberg, S. E. 1996. Microemulsions—phase equilibria characterization, structures, applications and chemical reactions. Adv. Colloid Interface Sci., 65, 125–287. Spernath, A. and Aserin, A. 2006. Microemulsions as carriers for drugs and nutraceuticals. Adv. Colloid Interface Sci., 128–130, 47–64. Stubenrauch, C. et  al. 2008. Gelled polymerizable microemulsions. 2. Microstructure. Langmuir, 24(16), 8473–8482. Szelinski, H. et al. 1996. Porphyrins linked to high acceptor strength cyano quinones as models for the photosynthetic reaction center. Tetrahedron, 52(25), 8497–8516. Tamhane, V. A. et  al. 2012. Enhanced permeation, leaf retention, and plant protease inhibitor activity with bicontinuous microemulsions. J. Colloid Interface Sci., 383(1), 177–183. Tempesti, T. C., Stockert, J. C., and Durantini, E. N. 2008. Photosensitization ability of a water soluble Zinc(II) tetramethyltetrapyridinoporphyrazinium salt in aqueous solution and biomimetic reverse micelles medium. J. Phys. Chem. B, 112(49), 15701–15707. Tenchov, B. and Koynova, R. 2012. Cubic phases in membrane lipids. Eur Biophys J., 41(10), 841–850. Tonova, K. and Lazarova, Z. 2008. Reversed micelle solvents as tools of enzyme puriication and enzymecatalyzed conversion. Biotechnol. Adv., 26(6), 516–532. Valdez, D. et al. 2001. Hydration and protein folding in water and in reverse micelles: Compressibility and volume changes. Biophys. J., 80(6), 2751–2760. Valentine, K. G. et al. 2010. Reverse micelle encapsulation of membrane-anchored proteins for solution NMR studies. Structure (Cambridge, MA, U.S.), 18(1), 9–16. Van Horn, W. D., Ogilvie, M. E., and Flynn, P. F. 2008. Use of reverse micelles in membrane protein structural biology. J. Biomol. NMR, 40(3), 203–211. Van Horn, W. D., Ogilvie, M. E., and Flynn, P. F. 2009. Reverse micelle encapsulation as a model for intracellular crowding. J. Am. Chem. Soc., 131(23), 8030–8039. von Gleich, A. et al. 2010. Potentials and Trends in Biomimetics. Heidelberg, Germany: Springer-Verlag. Walde, P. 2006. Surfactant assemblies and their various possible roles for the Origin(S) of life. Origins Life Evol. Biospheres, 36(2), 109–150. Walsh, D. and Mann, S. 1996. Chemical synthesis of microskeletal calcium phosphate in bicontinuous microemulsions. Chem. Mater., 8(8), 1944–1953. Warschawski, D. E. et al. 2011. Choosing membrane mimetics for NMR structural studies of transmembrane proteins. Biochim. Biophys. Acta, Biomembr., 1808(8), 1957–1974. Wellert, S. et al. 2011. The DFPase from Loligo vulgaris in sugar surfactant-based bicontinuous microemulsions: Structure, dynamics, and enzyme activity. Eur. Biophys. J., 40(6), 761–774. Winter, P. W. et al. 2012. The anti-diabetic bis(maltolato)oxovanadium(iv) decreases lipid order while increasing insulin receptor localization in membrane microdomains. Dalton Trans., 41(21), 6419–6430. Yang, J. H. K., Burban, J. H., and Cussler, E. L. 1995. Copper selective adsorption with a microemulsion-based resin. AIChE J., 41(5), 1165–1170. Yu, Z.-J. and Neuman, R. D. 1995. Reversed micellar solution-to-bicontinuous microemulsion transition in sodium bis(2-ethylhexyl) phosphate/n-heptane/water system. Langmuir, 11(4), 1081–1086. Zemb, T. 2009. Flexibility, persistence length and bicontinuous microstructures in microemulsions. C. R. Chim., 12(1–2), 218–224. Zhang, Y. et  al. 2008a. Studies on the interaction of gallic acid with human serum albumin in membrane mimetic environments. Talanta, 76(2), 246–253. Zhang, Y. et al. 2008b. Characterization of interaction between bergenin and human serum albumin in membrane mimetic environments. J. Fluoresc., 18(3–4), 661–670. Zhang, Y. et al. 2008c. Characterization of interaction between esculin and human serum albumin in membrane mimetic environments. J. Mol. Struct., 889(1–3), 119–128.

11

Locations of Small Biomolecules in Model Membranes Drew Marquardt and Thad A. Harroun

CONTENTS 11.1 Introduction .......................................................................................................................... 199 11.2 Localization of MBM ...........................................................................................................200 11.2.1 X-Ray Scattering....................................................................................................... 201 11.2.2 Neutron Scattering ....................................................................................................202 11.2.3 NMR Spectroscopy .................................................................................................. 203 11.2.4 Fluorescent Spectroscopy .........................................................................................205 11.2.4.1 Fluorescence Resonance Energy Transfer .................................................207 11.2.5 ESR Spectroscopy ....................................................................................................208 11.2.6 Differential Scanning Calorimetry ...........................................................................208 11.3 Location, Location, and Location......................................................................................... 210 11.3.1 Sterols ....................................................................................................................... 210 11.3.2 Vitamins ................................................................................................................... 211 11.3.3 Anesthetics................................................................................................................ 212 11.3.4 Antimicrobials and Other Drugs .............................................................................. 212 11.4 Concluding Remarks ............................................................................................................ 213 References ...................................................................................................................................... 213

11.1 INTRODUCTION Although great progress has been made in our understanding of the cell membrane since the cell wall of plants was irst directly visualized by the earliest seventeenth century microscopes, even our current understanding is still being informed by theories of Overton over a century ago. Searching for small, organic molecules that can be readily taken up into cells osmotically, Overton (1895) noted that “the living protoplasm of all basic organisms (plant cells, protozoa, ciliary, and epithelial cells) is readily permeable for various organic solutes (p. 181).” The permeability of the cell membrane was not determined by the size of the solute, but the membrane permeability coeficient of a solute can be simply correlated to its oil/water partition coeficient. This observation led to the simple representation of the membrane as a homogeneous oil slab, and based on this idea, a simple bulk solubility–diffusion model of membrane permeability was proposed. It is beyond the scope of this chapter to trace the discovery of the molecular structure of membranes; we assume the reader is familiar with the complex heterogeneity of the lipid bilayer. Notably, along each principle direction within the membrane (i.e., perpendicular and lateral to the plane), the density distributions, order parameters, and diffusion rates show characteristic properties that are simply not observed in bulk oil systems. Moreover, lipid bilayers contain highly polar moieties, such as the headgroup and the glycerol/ester regions, which do not have any counterpart in oil solvents. 199

200

Liposomes, Lipid Bilayers and Model Membranes

Nevertheless, the Overton paradigm still governs much of our thinking of the membrane today (Missner et al., 2008), and much more sophisticated models built from the Overton foundation can predict the permeability of pure lipid bilayers to a wide range of solutes (Lomize et al., 2011). The selective permeability of the membrane is obviously important to life’s processes, but without the assistance of membrane proteins, only small, nonpolar molecules are able to diffuse through the bilayer at any appreciable rate. Aside from the process of active transport of ions, nutrients, and other important molecules across the membrane by proteins, many small molecules still have some direct afinity for the lipid phase. Importantly, these interactions may alter the physical properties of the bilayer, which in turn may change the function of the proteins embedded therein. Despite their importance, our knowledge of the interactions between small molecules and lipid bilayers is still rather limited. Perhaps, the most important example of such small molecules is aliphatic alcohols, originally studied by Overton in the nineteenth century (Overton, 1895). Even today, the anesthetic properties of simple ethanol continue to focus its effects on the lipid bilayer. The key to our current understanding is providing direct experimental evidence for either a binding interaction between ethanol and lipid bilayers or for the location of this interaction on, or within, the bilayer (Barry and Gawrisch, 1994; Toppozini et al., 2012). The afinity of a molecule for a lipid phase can be determined via a straightforward protocol of equilibrium dialysis and spectrophotometry (Tipping et al., 1979). However, the method says nothing of the mechanics of binding, or the resulting effects of the bilayer itself. The heterogeneity present inside membranes affects the partitioning and diffusion of a solute as a function of its position inside the membrane, and is extremely dificult to measure experimentally. Experiments on lipid bilayer systems are complicated by the disordered nature of the system. In this chapter, we introduce the reader to several experimental techniques that have been developed to tackle the problem of localizing the partition of small membrane-bound molecules (MBM) within a lipid bilayer. Here, we deine a small biological molecule as an organic molecule that has a molecular mass on the order of a phospholipid or less (≤~800 g/mol). We choose a phospholipid as a “soft” limit because it is the monomer unit of a bilayer. There are a variety of molecules that it our deinition of an MBM and many of these molecules role in vivo is still unknown. For example, vitamin E (tocopherol) is a commercially used, fat soluble, antioxidant, but its true biological function is unknown. Another good example is cholesterol, arguably the most studied MBM in the ield of biochemistry and biophysics. Despite the extensive research on this MBM, its true role is illusive and debated today. In an effort to solve the mystery surrounding these molecules, groups have begun to draw relationships between the function and the MBM’s location along the normal to the membrane plane (Kucˇerka et al., 2009a; Komljenovic et al., 2010; Marquardt et al., 2013).

11.2

LOCALIZATION OF MBM

A wide variety of experimental techniques have been used to locate an MBM within the lipid bilayer. We survey some of the most important techniques here and highlight their strengths and weaknesses. The goal of each technique, however, is to localize the interaction of the MBM in regard to the depth of the membrane; that is, is the interaction primarily through the head groups, or is the MBM solubilized in the hydrocarbon core? The balance of free energy from steric interactions, polar, and nonpolar groups will determine the location of an MBM partitioned into lipids. Understanding this location may explain other effects, such as access to membrane or cytosolic proteins, or changes in the physical state of the bilayer. These techniques can be characterized by the source of the experimental signal. Many techniques require the presence of additional molecules or chemical groups serving as probes and reporters of the local environment (e.g., luorescence and electron spin resonance [ESR]), whereas other techniques measure the structure of the native bilayer directly (e.g., x-ray diffraction and 31P nuclear magnetic resonance [NMR]).

Locations of Small Biomolecules in Model Membranes

201

One can also classify these methods by how the source of the experimental signal is incorporated into the MBM/bilayer system. The signal may come directly from the lipids, and in general, such experiments report only on their current physical state. In this case, the independent variable is often the concentration of MBM, and the measurement follows how the presence of the MBM alters the physical state of the lipids. Alternatively, the probe may be part of the MBM itself. In this case, the dependent variable may be how the signal changes as the MBM partitions at different depths in the bilayer, which can be open for interpretation. Finally, of course, the probe molecule may be a third separate component, and it can serve either of the previous two roles. The signal of the probe molecule is generally used to either report on the physical state of the lipids, as above, or report on the proximity of the MBM in question. Clearly, the ideal experiment is one that mimics the biologically relevant situation most closely, and it is up to the researcher to determine whether the perturbation caused by the addition of a probe to the system is a hindrance, or whether the payoff in further physical insight is important. In some techniques, it is fortuitous that the probe is minimally invasive, such as assumed to be with deuterium substituting for hydrogen. In the examples that follow, we will try to highlight how, or even whether, the perturbations caused by a foreign probe were accounted for in the experiments.

11.2.1

x-ray scatterIng

Bragg’s diffraction of x-rays from multilamellar stacks of lipid bilayers was one of the irst techniques used to determine their molecular structure, building from techniques used to study smectic liquid crystals (Papahadjopoulos and Miller, 1967). In this way, the one-dimensional repeating unit cell is formed from a single bilayer and associated waters. To have a repeating structure to satisfy Bragg’s law, this style of experiment is performed on oriented stacks of hydrated phospholipid bilayers (Tristram-Nagle, 2007). X-ray scattering experiments can be designed to examine the physical state of the lipids, but they can also be designed where the signal comes from a probe. Much work is done on membrane and molecular interactions without the use of a “label” (an electron-rich atom); however, examples exist where guest molecules are halogenated or labeled with other electron-rich atoms (Katsaras et al., 1991; Hristova and White, 1998). In this case, analysis proceeds similar to the case of neutron diffraction, discussed below. The real-space information one gets from an x-ray diffraction is an electron density proile, the spatial distribution electron density.* An EDP of a lipid bilayer (without a high electron density label atom) highlights the electron-rich phosphate groups of a phospholipid, as well as the electron-poor center of the bilayer. From these points of reference, the MBMs can be localized to various regions in the membrane. The canonical resolution of an x-ray diffraction experiment is d/hmax, where d is the unit cell repeat spacing and hmax is the highest-order Bragg peak; thus, a typical experimental resolution is ~7 Å. The assigned uncertainty to the location of the MBM is also on the same order. We will see later that the use of contrast variation in, for example, neutron scattering can improve this standard diffraction resolution even further. Typically, x-ray experiments are conducted without the use of labeled guest molecules to look for changes in the EDP such as perturbations to the membrane and added electron density. From these observations, the location and interaction of the MBM can be deduced. Many MBMs contain oxygen (16 e−) and nitrogen (14 e−) in their chemical structures; the added electron density and the location of these atoms are often detected in the EDP (Barrett et al., 2012; Toppozini et al., 2012). For example, Barrett et al. (2012) found that acetylsalicylic acid (aspirin) preferentially resides in the headgroup region of a 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) bilayer. The in-plane structure (bilayer surface) and the out-of-plane structure (bilayer cross section) can be determined by using highly oriented bilayers on a Si substrate. Barrett et al. (2012) suggested that the acid group of aspirin is located at the headgroups of the phospholipids, while the authors were also able to infer the *

For a detailed procedure for generating EDP of x-ray diffraction data, refer to Franks and Lieb (1979).

202

Liposomes, Lipid Bilayers and Model Membranes

lateral organization of cholesterol, phospholipids, and aspirin. It was noted that there is a particularly high ordering effect on the presence of 50 mol% aspirin.

11.2.2

neutron scatterIng

In 1932, James Chadwick made his Nobel Prize-winning discovery of the neutron (Chadwick, 1932, 1933; Mason et al., 2013). The neutron has no charge thus making its interaction with matter a shortranged interaction, and allows it to penetrate deep into condensed matter, whether it is hard or soft. The interaction of neutrons with matter is either nuclear via the strong force, or magnetic; however, the magnetic dipole properties of neutrons play little role in determining biomembrane structure. Neutron scattering is a method of examining the structure of various materials without the use of bulky probes and provides insights that other methods cannot (Zaccai et al., 1975; Harroun et al., 2006a). Neutron scattering is advantageous to the other structural scattering techniques of x-ray because there is no trend in an atom’s neutron scattering power, providing large scattering contrasts between neighboring atoms on the periodic table, and even isotopes of the same element. The contrast in x-ray scattering increases with electron density and thus atomic number, making the contrast between atoms present in biological materials small, and making hydrogen virtually invisible. The study of membranes and MBMs requires precise conditions, such as temperature, pressure, and hydration, to simulate physiological conditions. Neutrons provide a valuable method for examining these systems with technical obstacles because many materials (i.e., aluminum) are nearly transparent to neutrons, allowing for sample environments to be easily constructed and implemented without the need for special windows (Harroun et al., 2009). The true power of neutron scattering as a probe for the study of the location of MBMs is the large scattering length difference between hydrogen (1H) and deuterium (2H); see Table 11.1. The substitution of deuterium atoms for hydrogen, at selective locations, provides contrast between two different samples (a “labeled” and an “unlabeled”). The difference in neutron scattering length density between the “labeled” and “unlabeled” sample yields the location and distribution of the “label” (Harroun et al., 2006b, 2008; Kucˇerka et al., 2009a, 2010; Komljenovic et al., 2010; Atkinson et al., 2010; Marquardt et al., 2013). The resolution of the experimental data can be increased through the

TABLE 11.1 Atomic Nuclei Present in Biology

Atom Hydrogen Deuterium Carbon Nitrogen Oxygen Fluorine Phosphorous Chlorine

Nucleus H H 12C 14N 16O 19F Pb Cl 1

2

Scattering Length (10−12 cm) −0.374 0.667 0.665 0.930 0.580 0.556 0.513 0.958

Coherent Cross Section (10−24 cm2) 1.76 5.56 5.56 11.1 4.23 4.03 3.31 11.53

Incoherent Cross Sectiona (10−24 cm2) 79.7 2.01 0 0 0 0 0 5.9

Absorbance Cross Section (10−24 cm2) 0.33 0 0 1.88 0 0 0.17 33.6

f X-ray (10−12 cm) 0.28 0.28 1.69 1.97 2.25 2.53 4.22 4.74

Source: With kind permission from Springer Science+Business Media: Neutron Scattering in Biology Techniques and Applications, Neutron scattering for biology, 2006a, pp. 1–18, Harroun, T., G. Wignall, and J. Katsaras. Chapters 1–18, In: J. Fitter, T. Gutberlet, and J. Katsaras (eds.), Heidelberg, Germany. a Values of the absorption cross section (σ ) are a function of wavelength and are given at λ = 1.8 Å. abs b Values are an average over the relative natural abundance of the stable isotopes.

Locations of Small Biomolecules in Model Membranes

203

use of a speciic deuterium labeling, with the end result being the near-atomic accurate determination of the label’s position along the bilayer (Gordeliy and Chernov, 1997). Gordeliy and Chernov theoretically determined that the spatial resolution of the diffraction data can be improved from the canonical resolution of d/hmax down to ~1 Å. Neutron scattering studies of biologically relevant samples have an intrinsic advantage due to the abundance of hydrogen atoms that can be replaced (labeled) by deuterium. For example, the difference in neutron scattering length density (NSLD) proiles from lipid bilayers containing labeled and unlabeled cholesterol has been used to accurately determine the position of the cholesterol’s hydroxyl group (Harroun et al., 2006b). In addition, the external contrast variation is a useful method to determine the phases of scattering form factors, which are a very common “speed bump” in the analysis of crystallographic data.* Harroun et al. (2006b, 2008) unequivocally located cholesterol in the center of di-20:4 PC bilayers, by the use of “head” and “tail” deuterated cholesterol and the employment of neutron scattering. Combined with NMR experiments, the absolute orientation of cholesterol in the polyunsaturated fatty acid (PUFA) bilayer was revealed, that is, if it was coplanar with the bilayer plane or possessed free axial rotation (Brzustowicz et al., 1999, 2002). It was determined that cholesterol undergoes fast axial motion as it lies in the center of a PUFA bilayer (Harroun et al., 2008). This noncanonical behavior of cholesterol was independently observed by Haldar et al., through the observation of membrane dipole potentials (Haldar et al., 2012). Further investigation into cholesterol’s behavior in PUFA bilayer was conducted. Doping highly unsaturated bilayers with saturated phospholipids yielded cholesterol reorienting itself to its generally accepted upright position in the bilayer. With as little as 5% saturated lipid, Kucˇerka et al. (2009a, 2010) demonstrated this cholesterol lip-lop action, which could provide insight into potential cholesterol transport mechanisms and/or cellsignaling mechanisms of cholesterol.

11.2.3

nMr spectroscopy

H NMR is particularly useful in the study of biological systems, speciically MBMs, because it is sensitive to the local orientations of the carbon–deuterium (C–D) bond in a lipid tail. It is this sensitivity that makes it possible for the determination of molecular motion in speciic locations, thus leading to inferences about the molecular structure of the local environment. 2H is not the only NMR-active nuclei useful in the study of biological materials, see Table 11.2. For locating MBM in vesicles, the irst approximation is to differentiate the hydrophilic (polar) and hydrophobic (nonpolar) environments and to quantify the order parameters of the hydrophobic chains (Xu et al., 2002). After addition of the MBM, the order parameters will change, either in a speciic location or more broadly, if the perturbation to the bilayer is more widely spread (Figure 11.1). Likewise, order parameters of a 2H label on the guest MBM are particularly useful, as they can be compared to those of the membrane environment (Saint-Laurent et al., 1998; Vermeer et al., 2007; Leftin and Brown, 2011). Order parameters, (SCD), describe the conformationally averaged angular luctuations of the C–D bond with respect to the bilayer normal (Shaikh et al., 2006). Intuitively, it becomes apparent that the order parameters decrease with depth along the phospholipid, indicating that there is increased molecular motion toward the terminal methyl group of the phospholipid acyl chain compared to the environment near the carbonyl carbons of the glycerol backbone (Falck et al., 2006). Quite often, NMR order parameter proiles are used to corroborate other location-determining techniques such as x-ray or neutron scattering (Marsan et al., 1999; Shaikh et al., 2006). Saturated acyl chains and headgroup deuterated phospholipids are readily available, making the order 2

*

As with x-ray diffraction, refer to Preparation of Oriented, Fully Hydrated Lipid Samples for Structure Determination Using X-Ray Scattering for making reproducible oriented samples (Tristran-Nagle, 2007) and for an in-depth discussion on the analysis of oriented small-angle neutron diffraction, refer to Kucˇ erka et al. (2009b). See Squires’ text, Introduction to the Theory of Thermal Neutron Scattering, for the foundations in neutron scattering (Squires, 1978).

204

Liposomes, Lipid Bilayers and Model Membranes

TABLE 11.2 NMR-Active Nuclei Present in Biology Atom Hydrogen Deuterium Carbon

Nucleus

Spin

1

H

2

H C

1 2 1 1 2 1 1 − 2

13

Nitrogen Nitrogen

14 15

N N

Oxygen

17

O

Fluorine

19

F

Phosphorous

31

P

Chlorine

Cl33

Chlorine

Cl37

−

5 2 1 2 1 2 3 2 3 2

parameter determination at each carbon perpendicular to the plane of the bilayer. However, not all phospholipids are saturated and deuterated unsaturated acyl chains are often expensive, unavailable, or even impossible to synthesize. In the cases of unsaturated lipid bilayers, hydroponically matched fatty acids, which are deuterated, are used to achieve order parameter proiles for comparison (Williams et al., 2012; Marquardt et al., 2013). Other methods of determining the penetration depth of guest molecules involve using the chemical shift and relaxation couplings of NMR-active nuclei present (Boland and Middleton, 2008). DPPC + Chol DPPC

0.25 H H 0.20 |SCD|

OH H

H H

0.15 O 0.10

O

PC O H

0.05

O 16

14

12

10

8 n

6

4

2

0

FIGURE 11.1 (See color insert.) Example of averaged C–D order parameters (SCD) from the SN-1 and SN-2 chains of pure d62-DPPC (squares) and from d62-DPPC with 13 mol% cholesterol. (The data are regenerated from Falck, E. et al. 2006. Biophysical Journal 91 (5), 1787–1799.)

Locations of Small Biomolecules in Model Membranes

205

ET (30)

55.4 MeOH

48.4 Isopropanol

31 Hexane

FIGURE 11.2 (See color insert.) Cartoon representation of a Toc’s location in a phospholipid bilayer along with the resulting polarity of the location and the representative solvent. Polarities along the lipids length were determined by Cohen et al. (2008a,b).

These types of NMR experiments are typically considered as two-dimensional (2D) NMR, which is an umbrella for a family of techniques such as correlation spectroscopy (COSY), exchange spectroscopy (EXSY), and nuclear Overhauser effect spectroscopy (NOESY). A simpler technique for using chemical shift is to use Reichardt’s ET(30) parameters (Reichardt, 1994; Cohen et al., 2008a,b). Reichardt (1994) indexed solvent polarity into units of ET(30) as a way of quantitatively comparing solvents. It turns out that the NMR chemical shift as a function of solvent polarity in ET(30) units trends linearly, allowing for an “NMR-based molecular ruler for determining the depth of intercalates within the lipid bilayer” (Cohen et al., 2008a,b). Figure 11.2 illustrates the ET(30) at various points along a phospholipid and how it compares with common organic solvents. Cholesterol interaction with lipid bilayers has been studied for years with a wide variety of phospholipids order parameter proile (Huster et al., 1998; Trouard et al., 1999; Falck et al., 2006; Vermeer et  al., 2007; Clarke et  al., 2009). Not only the penetration depth of a molecule can be determined by 2H labeling and order parameters, but, often, tilt angles of certain molecules can be achieved. Numerous studies have been conducted where the tilt angle of cholesterol was determined (Davis, 1993; Kessel et al., 2001; Shaikh et al., 2006). It has been found that cholesterol is experimentally very sensitive to its lipid environment. In early 2H NMR work, a profound difference in order parameter was observed for [3α-2H1]cholesterol added to dispersions of PUFA PCs and saturated PC phospholipids (Brzustowicz et al., 1999). It was this NMR work that sparked investigation into cholesterol orientation in highly disordered lipid bilayers.

11.2.4 Fluorescent spectroscopy The most straightforward and standard luorescent technique when studying the location of guest molecules is the parallax method (Kachel et al., 1998; Boldyrev et al., 2007; Kondo et al., 2008;

206

Liposomes, Lipid Bilayers and Model Membranes

Shrivastava et  al., 2009). Briely, the principle of this technique measures relative luorescence quenching eficiencies of a series of quenching probe-labeled phospholipids. These probes can be selectively placed at different locations within the lipid bilayer based on where it is attached to the probe lipid, as shown in Figure 11.3. Analysis by use of the three nitroxide probes in Figure 11.3 allows calculation of the depth of a luorophore at the 12 Å level of resolution (Kachel et al., 1998). Although this technique can give angstrom resolution, the use of the bulky probe may produce less than accurate results; the aforementioned “lip-lop” of cholesterol exempliies the effect pollutants to the native bilayer can have. This style of experiment has been extensively used in the study of vitamin E (Fukuzawa, 2008). Many attempts have been made to determine the location of vitamin E in a bilayer using the parallax method. Typically, n-DOXYL or anthroyloxyl stearic acid would be used exploiting the intrinsic luorescence of vitamin E (Fukuzawa et al., 1992, 1993; Fukuzawa, 2008). The resulting quenching eficiencies suggest that vitamin E resides high in the phospholipid bilayer. In addition, the location of cholesterol can be examined in a similar manner; however, cholesterol has no intrinsic luorescence. Figure 11.4c shows a full cholesterol molecule with the addition of a boron-dipyrromethene (BODIPY) luorophore on the tail. The antithesis to the above parallax method is having the luorophores as the probes attached to the lipid and the guest MBM being the quenching agent. Figure 11.4 illustrates a head-labeled and tail-labeled phospholipid and a labeled cholesterol. When comparing the relative size of the stable radical probes of Figure 11.3 and the probes of Figure 11.4, it is apparent that the stable radicals contribute less-steric bulk, thus maintaining a more unperturbed membrane environment. This experimental method is well suited for the study of radical penetration into the lipid bilayer, which is required for a complete understanding of oxidative stress on a membrane.

(a)

O O O H

O O P O O

N+

N O

O TEMPO (b)

O

O O O H

O

O P O O

N+

O

N O

DOXYL

(c)

O

O O O H

O

N O

O P O O–

N+

O

DOXYL

FIGURE 11.3 Examples of the commonly used and readily available spin probes (a) is the headgroup stable radical TEMPO, (b) and (c) are the DOXYL stable radical; DOXYL can be incorporated at a variety of positions down the hydrophobic chain allowing the probing of the membrane to various depths.

207

Locations of Small Biomolecules in Model Membranes O (a)

O O H

O O P O O– + NH4

O NHS O

O

N DANSYL

O (b)

O–

O O

N+

NH

O

O N

O NBD

H

O P O O

N+

O

N

N+

F B– F N

BODIPY

H

(c)

H H

H

HO

FIGURE 11.4 Examples of the commonly used and readily available luorescent probes. (a) 18:1 DANSYL PE, (b) 14:0–12:0 NBD PC, and (c) is the BODIPY cholesterol.

Studies that attempt to address the obvious critique of the use of bulky probes have discovered orientation properties of the BODIPY luorescent probe, as well as ways to address the problem (Armendariz et al., 2012). A correlation was observed between the probes behavior as a function of surface pressure identifying how placing the probe on the SN-1 chain, between C-9 and C4, is the optimal placement of the probe within the hydrophobic region to maximize its sensitivity (Armendariz et al., 2012). In addition to the orientation, the problem arises of the probes inluencing the lipid organization; for example, driving the formation of rafts—that is, probe-rich domains and probe-poor domains (Shaw et al., 2006). Fluorescent probes are being used in in vivo samples already, speciically in cholesterol traficking in living cells (Wustner et al., 2011). The future may provide instrumentation and new probe technologies that will allow for the equilibrium location of many MBMs to be determined in vivo. 11.2.4.1 Fluorescence Resonance Energy Transfer Fluorescence (or Förster) resonance energy transfer (FRET) studies the energy transfer from one chromophore (donor), in an electronic excited state, to another chromophore (acceptor), through a proximity-dependent resonance process. Through this energy transfer, the acceptor emits a photon at a characteristic wavelength signifying the donor molecule is in the proximity of the acceptor molecule. The energy transfer falls off as r6, where r is the distance between the donor and acceptor (Zheng, 2006). Taraska et al. (2009) state the typical Förster distance, the distance that yields 50% energy transfer, is a proximity of 20–80 Å; however, using a transition metal can better this resolution to 5–30 Å. Stella et al. (2007), with the use of FRET, determined that alamethicin does not translocate from the outer to the inner lealet of the membrane, where FRET acceptors were located. To determine this lack of translocation, the authors created three types of vesicles: inner lealet, outer lealet, and inner and outer lealet labeled, with the FRET acceptor NBD, the donor label was a luorenyl

208

Liposomes, Lipid Bilayers and Model Membranes

located on the alamethicin. Cholesterol’s lip-lop between lealets has also been studied using FRET (Silvius, 2003; Bruckner et al., 2009). In Bruckner et al. (2009), dehydroergosterol (DHE) (a luorescing cholesterol derivative) was added to a lipid bilayer as the FRET donor in this system and a water-soluble FRET acceptor was added to the bulk water. The curvature of the vesicles containing DHE was changed via osmotic pressure and the curvature-induced lip-lop was measured. It was determined that a threshold area per lipid in the outer lealet was required to shift the DHE equilibrium location to the outer lealet preferentially (Bruckner et al., 2009). Owing to the resolution of the technique, it is only useful for nanoscale location determination.

11.2.5

esr spectroscopy

ESR, also known as electron paramagnetic resonance spectroscopy (EPR), studies spin-active materials with unpaired electrons. The technique is very similar to NMR, whereby spin populations are excited by a magnetic ield; however, electron spin states are excited in ESR and atomic nuclei are excited in NMR. The sensitivity of an ESR signal from a radical is 600–1000 times that of a proton NMR, thus requiring signiicantly lower concentrations of probes (Berliner, 2010). ESR spectroscopy is particularly powerful at determining the penetration depth of radicals (unpaired electrons) into the cell membrane (Bodner et  al., 2010). There are various ways to determine the penetration depth of these potentially damaging molecules. Owing to the increasing attention to the antioxidant properties of vitamin E and oxidative stress in general, ESR has been utilized to determine the location of reactive-free radicals and other reactive oxygen species (ROS) in model membranes (Subczynski et  al., 2009; Spasojevic, 2011). On the basis of spin-labeled phospholipid acyl-chain comparison with neutron diffraction, Subczynski et al. (2009) have concluded that the time-averaged position of nitroxide spin labels, similar to those in Figure 11.3 is accurate to within 1.5–3.4 Å. The ESR spectra of 5-NS and 16-NS labeled in DMPC or DMPC-dicetylphosphate (DCP) liposomes were not changed by the addition of ascorbic acid (AsA) in the buffer solution of pH 7.0, indicating that negatively charged AsA could not penetrate into the membrane (Fukuzawa et al., 1993). One use of spin label is the comparison of the spin probes behavior with and without the MBM present. If the presence of the MBM has a signiicant effect on the ESR spectra of the selectively located probe, it can be inferred that the MBM is to some extent present in that location. Another method utilizes the ET (30) that scales similar to the NMR molecular ruler (Reichardt, 1994; Bodner et al., 2010). In this style of experiment, the researcher relies on the micropolarity at different locations within the membrane. Figure 11.2 highlights the primary locations along a phospholipid long axis. Different polar environments affect the hyperine splitting constant (aβ−H) of the stable radical probe; for example, Bodner et al. (2010) report their aβ−H decreases linearly with an increase in solvent polarity (increase in ET (30)).

11.2.6

dIFFerentIal scannIng calorIMetry

Differential scanning calorimetry (DSC) is a common technique proving the association of MBM with lipid bilayers, and even loose inference of the location of the MBM within the bilayer. DSC works under the principle that there will be a power difference between the sample and a reference when the sample undergoes some thermotropic change if the two cells are kept at the same temperature and heated or cooled at the same rate. For lipids, this generally means thermotropic phase transitions such as the chain-melting transition or lamellar to nonlamellar morphological changes. Some of the earliest differential thermal analyses of phospholipids were performed by Chapman and Collin (1965). In this section, much of the literature presented and referenced is quite old; this technique has not changed over the past 50 years. In many eyes, the information collected by this technique is not considered precise; however, we view DSC as an indispensable tool for quick determination of the penetration of an MBM into the bilayer. For a comprehensive review of theoretical principles and applications of DSC, refer to Differential Scanning Calorimetry, 2nd edition (Höhne et al., 2003).

209

Locations of Small Biomolecules in Model Membranes

TABLE 11.3 Qualitative Effects on DSC Thermogram Caused by Different Membrane Locations Interaction 1 2 3

Main Chain Melting Temperature (Tm)

Enthalpy of Main Chain Melting (ΔHm)

Increase Decrease Decrease

Increase or none Decrease None

Source: Developed by Papahadjopoulos, D., M. Moscarello, and E. E. T. Isac. 1975. Biochimica et Biophysica Acta 401, 317–335.

Heat flow

The location determination of MBMs using DSC is at a very low resolution, providing a qualitative view of where a molecule resides in the bilayer. In fact, one can characterize the qualitative locations into three categories (Papahadjopoulos et al., 1975; Maghraby et al., 2005), see Table 11.3. Category 1 is the molecules that interact only with the headgroups of the bilayer lipids and the MBM adsorbs to the surface of the bilayer. Category 2 is the partial penetration of an MBM, which is typically as deep as the glycerol backbone of the bilayer lipids. Finally, category 3 is MBMs that penetrate deep in the hydrophobic core completely disrupting the lipids acyl chains and hindering cooperativity of the phase transition. These categories were initially developed to characterize peptide and protein interactions; however, it is our experience that these categories are useful and accurate for small MBMs as well. Figure 11.5 shows the DSC traces for DMPC and the guest molecule chlorhexidine, a clinically used antimicrobial, which demonstrates a decrease in both Tm and ΔHm. This molecule is a prime example of category 2 MBM. Neutron scattering experiments have unequivocally located chlorhexidine’s hexane chain at the glycerol backbone and with a little biophysical reasoning, it was concluded that the remaining portion of the molecule is located at the headgroup water interface (Komljenovic et al., 2010).

0%

0%

5%

2.5%

10%

5%

15%

14.3%

25 30 35 40 45 50 55 Temperature (°C)

30 35 40 45 50 55 60 Temperature (°C)

FIGURE 11.5 DSC traces of DPPC with varying concentrations of cholesterol (left) and DPPC with increasing concentrations of the broad-spectrum pesticide ethyl-azinphos (right). (The data are reproduced from Videira, R. A., M. C. Antunes-Madeira, and V. M. Madeira. 1999. Chemistry and Physics of Lipids 97 (2), 139–153.)

210

Liposomes, Lipid Bilayers and Model Membranes

TABLE 11.4 Common DSC Appropriate Phospholipids Phospholipid 14:0/14:0 PC 16:0/16:0 PC 18:0/18:0 PC 18:0/18:1 PC 14:0/14:0 PE 16:0/16:0 PE 18:0/18:0 PE 14:0/14:0 PS 16:0/18:1 PS 14:0/14:0 PG 18:0/18:0 PG

Main (°C) 23 43 55 6 50 63 74 35 14 23 55

Other Transitions (°C) 14a 36a

118b 100b

Note: The actual phase transition values vary in literature depending on sample preparation and conditions. The reported values are from Avanti Polar Lipids product data. a Pretransition. b Hexagonal phase transition.

Cholesterol is an example of MBM that does not follow this characterization method. Many other techniques demonstrate that cholesterol resides with its hydroxyl group near the glycerol backbone, with the fused ring system and tail extending deep into the hydrophobic core (Dufourc et al., 1984; Léonard et al., 2001; Kessel et al., 2001; Harroun et al., 2006b, 2008; Kucˇerka et al., 2009a, 2010; Shrivastava et al., 2009; Subczynski et al., 2009). The DSC thermogram of cholesterol in 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) and DMPC yields a decrease in Tm with cholesterol concentration, indicative of interactions at the glycerol backbone. There is also an apparent decrease in ΔHm with the addition of cholesterol that is a slight deviation from the above-mentioned categorization (Singer and Finegold, 1990; Kyrikou et al., 2004; Mavromoustakos, 2007; Demetzos, 2008). In a typical DSC phospholipid sample, the vesicles are left as multilamellar vesicles. The concentration of these samples varies from calorimeter to calorimeter depending on the sensitivity of the instrument. However, hydration plays a critical role in the thermodynamics of the phase transition, thus maintaining enough water per lipid and always running a pure lipid control with the same conditions is the key (Mavromoustakos, 2007). DSC relies on the ability to observe the main-chain melting phase transition. Many phospholipids undergo the main-chain melting transitions at temperatures below 0°C, predominantly unsaturated chain phospholipids, limiting the number of phospholipid systems that can be studied using this technique. Table 11.4 lists such phospholipids appropriate for DSC.

11.3

LOCATION, LOCATION, AND LOCATION

In this inal section, we will highlight just some of the important MBMs that have shaped our understanding of the lipid bilayer.

11.3.1

sterols

Cholesterol (Figure 11.6a) is a modulator of the physical properties of a membrane by increasing ordering. The hypothesized consequences of this ordering are the inhibition or activation of the function of membrane-bound proteins. The canonical location of cholesterol in a cell and model membrane has been determined by many biophysical techniques. All these results agree that

211

Locations of Small Biomolecules in Model Membranes

(a) H H H

H

HO H (b)

HO

CH3

H3 C CH3 H3C

CH3

CH3

O CH3

CH3

Cl NH NH

(c)

NH

NH

NH NH

NH NH

NH

NH

Cl

FIGURE 11.6 Example of three types of MBMs location in model membranes has been studied (Harroun et al., 2006b, 2008; Kucˇerka et al., 2009a, 2010; Komljenovic et al., 2010; Atkinson et al., 2010; Marquardt et al., 2013). Structure (a) is the steroid cholesterol, one of the most studied molecules in biology, structure (b) is the lipid-soluble vitamin, vitamin E (α-tocopherol), and (c) is the clinically used antimicrobial chlorhexidine.

cholesterol sits upright parallel with the lipid hydrocarbon chains (Franks and Lieb, 1979; Dufourc et al., 1984; Léonard et al., 2001; Kessel et al., 2001; Harroun et al., 2006b, 2008; Kucˇerka et al., 2009a, 2010; Shrivastava et al., 2009; Subczynski et al., 2009). This location allows cholesterol to modulate the membrane properties by steric interactions with the lipid hydrocarbon tails. In some cases, one technique is not enough to give the whole picture, especially if results are unintuitive or different from generally accepted results. In these instances, often, methods are used to complement each other. For example, combining the NMR with the neutron scattering experiments, the absolute orientation of cholesterol in the PUFA bilayer was determined. The culmination of these experiments deduced that cholesterol was coplanar with the bilayer plane or possessed free axial rotation (Brzustowicz et al., 1999, 2002). It was determined that cholesterol undergoes fast axial motion as it lies in the center of a PUFA bilayer (Harroun et al., 2008).

11.3.2

vItaMIns

Vitamins come in a variety of forms and solubilities. The most notable fat-soluble vitamins are A, D, E, and K, with vitamin E being a focus in the study of MBMs. Vitamin E (Figure 11.6b) is the only essential vitamin for which it is not known why it is essential and many scientists are working on making a connection between its location and function in the cell membrane and model membranes (Katsaras et al., 1991; Afri et al., 2004; Fukuzawa, 2008; Atkinson et al., 2010; Marquardt et al., 2013). Throughout these studies, many of the aforementioned experimental techniques were used. For example, Afri et al. (2004) utilized NMR and the chemical shift dependence on polarity to determine the depth of vitamin E and several antioxidants. It was determined that vitamin E resides near the water–lipid interface, where he found the ubiquinol deeper in the hydrophobic slab (Cohen et al., 2008a,b). Fukuzawa (2008) has relied heavily on years of luorescent work in the elucidation of vitamin E’s location and role in a membrane. Katsaras et al. (1991), Atkinson et al. (2010), and Marquardt et al. (2013) have determined, by both x-ray and neutron scattering experiments, that

212

Liposomes, Lipid Bilayers and Model Membranes

vitamin E stands “upright” in the bilayer, with the hydroxyl group at or above the glycerol backbone, depending on the lipid species. It is not surprising that these researchers have determined, qualitatively, the same location of the membrane-bound vitamin with the use of very different experimental techniques. The culmination of this work and much biochemistry has led to location–function mechanism for vitamin E. It has been postulated that vitamin E’s depth in a lipid bilayer relates to its antioxidant capability (Marquardt et al., 2013).

11.3.3

anesthetIcs

An anesthetic is a molecule that elicits the physiological response of insensitivity to pain. Some of these molecules include nitrous oxide, halogenated hydrocarbons (chloroform), and alcohols (ethanol). The actual mechanism of anesthetics is still debated. There are two current theories: one theory hypothesizes that the anesthetic interacts directly with the protein, speciically an ion channel (Cantor, 1997). The other theory, which deals with membrane–anesthetic interactions, suggests that the anesthetic interacts with the cell membrane. The interactions with anesthetics change the physical properties of the membrane such as the lateral surface pressure, lipid packing, and change the bilayer thickness (Leonenko et al., 2007). As a result of these bilayer physical changes, the function of integral proteins is greatly affected as well as the cells ability to communicate with other cells (Leonenko et al., 2007). For decades, anesthetics interaction with membranes has been studied (Franks and Lieb, 1979). Many of Franks’ observations with anesthetics were conducted using x-ray and neutron scattering techniques to observe the effects on the membrane. Many of the observations yielded null results, that is, there was no signiicant physical change in the bilayer (Franks and Lieb, 1979; Dickinson et al., 1994). However, other studies supported the membrane–anesthetic interaction hypothesis. A primary example of an anesthetics location and membrane-altering properties is exempliied by Anbazhagan et al. (2010), who added phenylethanol (a local anesthetic) to model membranes. It was observed that phenylethanol increased the membrane luidity and also severely altered the oligomerization of certain membrane-bound proteins. In addition, Toppozini et al. (2012) have observed that ethanol has a substantial interaction with model membranes. Ethanol molecules were observed located in the headgroup region at concentrations as high as 1.6 ethanol molecules per lipid.

11.3.4

antIMIcrobIals and other drugs

Many antimicrobials exist whose primary target is the cell membrane, and whose mechanism is the disruption of lipid membranes by the insertion of these amphiphilic molecules. Many of these antimicrobials include peptides in the form of helices and polymyxins, and surfactants that are commonly quaternary ammonium salts. An example of a noncanonical antimicrobial is chlorhexidine (Figure 11.6c). Investigation by neutron scattering determined the penetration depth and a hypothesized mechanism of interaction with the cell membrane. Neutron scattering and the use of deuterium label revealed that chlorhexidine anchors itself into the bilayer with the hexamethyl chain penetrating past the glycerol backbone (Komljenovic et al., 2010). This interaction mimics that of an axe wedge. An example of a nonantimicrobial drug that interacts with the membrane would be triluoperazine (TFP), an antipsychotic. One method to try to determine the location of various drugs in lipid structures is to differentiate between the hydrophilic region (headgroups/backbones) and the hydrophobic region (hydrocarbon chains); this can be accomplished by quantifying the ordering parameters of the hydrophobic chains (Boland and Middleton, 2008). Boland and Middleton (2008) used 13C NMR to determine the depth of the antipsychotic TFP. The authors located TFP’s CF3substituted ring being situated among the hydrocarbon core, whereas the rest of the right structure extends toward the headgroups of the bilayer. This is a primary example of how pharmacologically and toxicologically active small molecules interact with the cell membrane.

Locations of Small Biomolecules in Model Membranes

11.4

213

CONCLUDING REMARKS

We have discussed a variety of biophysical techniques and demonstrated how these techniques can complement each giving a more complete picture of the interaction of small molecules inside lipid bilayers. With the increased sensitivity of instrumentation comes the ability to study systems of increased complexity and physiological concentrations. For example, luorine NMR may be used in the future to study peptides in vivo at physiological concentrations. Because of its high sensitivity and low background (Buer et al., 2010), it could be used to label and study MBMs in vivo. It is the authors’ opinion that the surface is just scratched in the study of MBMs and makes the connection between the function and location.

REFERENCES Afri, M., B. Ehrenberg, Y. Talmon, J. Schmidt, Y. Cohen, and A. A. Frimer. 2004. Active oxygen chemistry within the liposomal bilayer part III: Locating vitamin E, ubiqinol and ubiquinone and their derivatives in the lipid bilayer. Chemistry and Physics of Lipids 131, 107–121. Anbazhagan, V., C. Munz, L. Tome, and D. Schneider. 2010. Fluidizing the membrane by a local anesthetic: Phenylethanol affects membrane protein oligomerization. Journal of Molecular Biology 404 (5), 773–777. Armendariz, K. P., H. A. Huckabay, P. W. Livanec, and R. C. Dunn. 2012. Single molecule probes of membrane structure: Orientation of BODIPY probes in DPPC as a function of probe structure. Analyst 137 (6), 1402–1408. Atkinson, J., T. Harroun, S. R. Wassall, W. Stillwell, and J. Katsaras. 2010. The location and behavior of a-tocopherol in membranes. Molecular Nutrition and Food Research 54, 641–651. Barrett, M. A., S. Zheng, G. Roshankar, R. J. Alsop, R. K. R. Belanger, C. Huynh, N. Kucˇ erka, and M. C. Rheinstaedter. 2012. Interaction of aspirin (acetylsalicylic acid) with lipid membranes. PLOS ONE 7 (4), e34357. Barry, J. A. and K. Gawrisch. 1994. Direct NMR evidence for ethanol binding to the lipid–water interface of phospholipid bilayers. Biochemistry 33, 8082–8088. Berliner, L. J. 2010. From spin-labeled proteins to in vivo EPR applications. European Biophysics Journal with Biophysics Letters 39 (4, SI), 579–588. Bodner, E., M. Afri, and A. A. Frimer. 2010. Determining radical penetration into membranes using ESR splitting constants. Free Radical Biology and Medicine 49 (3), 427–436. Boland, M. P. and D. A. Middleton. 2008. The dynamics and orientation of a lipophilic drug within model membranes determined by C-13 solid-state NMR. Physical Chemistry Chemical Physics 10 (1), 178–185. Boldyrev, I. A., X. Zhai, M. M. Momsen, H. L. Brockman, R. E. Brown, and J. G. Molotkovsky. 2007. New BODIPY lipid probes for luorescence studies of membranes. Journal of Lipid Research 48 (7), 1518–1532. Bruckner, R., S. Mansy, A. Ricardo, L. Mahadevan, and J. Szostak. 2009. Flip-lop-induced relaxation of bending energy: Implications for membrane remodeling. Biophysical Journal 97 (12), 3113–3122. Brzustowicz, M., W. Stillwell, and S. Wassall. 1999. Molecular organization of cholesterol in polyunsaturated phospholipid membranes: A solid state H-2 NMR investigation. FEBS Letters 451 (2), 197–202. Brzustowicz, M. R., V. Cherezov, M. Zerouga, M. Caffrey, W. Stillwell, and S. R. Wassall. 2002. Controlling membrane cholesterol content. A role for polyunsaturated (docosahexaenoate) phospholipids. Biochemistry 41 (41), 12509–12519. PMID: 12369842. Buer, B. C., J. Chugh, H. M. Al-Hashimi, and E. N. G. Marsh. 2010. Using luorine nuclear magnetic resonance to probe the interaction of membrane-active peptides with the lipid bilayer. Biochemistry 49 (27), 5760–5765. Cantor, R. S. 1997. The lateral pressure proile in membranes: A physical mechanism of general anesthesia. Biochemistry 36 (9), 2339–2344. Chadwick, J. 1932. Possible existence of a neutron. Nature 129 (3252), 312. Chadwick, J. 1933. Bakerian lecture. The neutron. Proceedings of the Royal Society of London, Series A 142 (846), 1–25. Chapman, D. and D. T. Collin. 1965. Differential thermal analysis of phospholipids. Nature 206 (4980), 189–189. Clarke, J. A., J. M. Seddon, and R. V. Law. 2009. Cholesterol containing model membranes studied by multinuclear solid state NMR spectroscopy. Soft Matter 5 (2), 369–378.

214

Liposomes, Lipid Bilayers and Model Membranes

Cohen, Y., M. Afri, and A. A. Frimer. 2008a. NMR-based molecular ruler for determining the depth of intercalants within the lipid bilayer part II: The preparation of a molecular ruler. Chemistry and Physics of Lipids 155, 114–119. Cohen, Y., E. Bodner, M. Richman, M. Afri, and A. A. Frimer. 2008b. NMR-based molecular ruler for determining the depth of intercalants within the lipid bilayer part I. Discovering the guidelines. Chemistry and Physics of Lipids 155 (2), 98–113. Davis, J. H. 1993. The molecular dynamics, orientational order, and thermodynamics phase equilibria of cholesterol/phosphatidylcholine mixtures: 2H nuclear magnetic resonance. In: L. Finegold (ed.), Cholesterol in Membrane Models, Chapter 4, pp. 67–135. Boca Raton, FL: CRC. Demetzos, C. 2008. Differential scanning calorimetry (DSC): A tool to study the thermal behavior of lipid bilayers and liposomal stability. Journal of Liposome Research 18 (3), 159–173. Dickinson, R., N. Franks, and W. Lieb. 1994. Can the stereoselective effects of the anesthetic isolurane be accounted for by lipid solubility? Biophysical Journal 66 (6), 2019–2023. Dufourc, E., E. Parish, S. Chitrakorn, and I. Smith. 1984. Structural and dynamical details of cholesterol lipid interaction as revealed by deuterium NMR. Biochemistry 23 (25), 6062–6071. Falck, E., J. T. Hautala, M. Karttunen, P. K. J. Kinnunen, M. Patra, H. Saaren-Seppala, I. Vattulainen, S. K. Wiedmer, and J. M. Holopainen. 2006. Interaction of fusidic acid with lipid membranes: Implications to the mechanism of antibiotic activity. Biophysical Journal 91 (5), 1787–1799. Franks, N. P. and W. R. Lieb. 1979. The structure of lipid bilayers and the effects of general anaesthetics. Journal of Molecular Biology 133, 469–500. Fukuzawa, K. 2008. Dynamics of lipid peroxidation and autoxidation of alpha-tocopherol. Journal of Nutritional Science and Vitaminology 54, 273–285. Fukuzawa, K., W. Ikebata, A. Shibata, I. Kumadaki, T. Sakanaka, and S. Urano. 1992. Location and dynamics of a-tocopherol in model phospholipid membranes with different charges. Chemistry and Physics of Lipids 63, 69–75. Fukuzawa, K., W. Ikebata, and K. Sohmi. 1993. Location, antioxidant and recycling dynamics of alpha-tocopherol in liposome membranes. Journal of Nutritional Science and Vitaminology 39, 9–22. Gordeliy, V. and N. Chernov. 1997. Accuracy of determination of position and width of molecular groups in biological and lipid membranes via neutron diffraction. Acta Crystallographica Section D—Biological Crystallography 53, 377–384. Haldar, S., R. K. Kanaparthi, A. Samanta, and A. Chattopadhyay. 2012. Differential effect of cholesterol and its biosynthetic precursors on membrane dipole potential. Biophysical Journal 102 (7), 1561–1569. Harroun, T., G. Wignall, and J. Katsaras. 2006a. Neutron scattering for biology. In: J. Fitter, T. Gutberlet, and J. Katsaras (eds.), Neutron Scattering in Biology Techniques and Applications, Chapters 1–18, pp. 1–18. Heidelberg, Germany: Springer. Harroun, T. A., J. Katsaras, and S. R. Wassall. 2006b. Cholesterol hydrozyl group is found to reside in the center of a polyunsaturated lipid membrane. Biochemistry 45, 1227–1233. Harroun, T. A., J. Katsaras, and S. R. Wassall. 2008. Cholesterol is found to reside in the center of a polyunsaturated lipid membrane. Biochemistry 47 (27), 7090–7096. PMID: 18543943. Harroun, T. A., N. Kucˇ erka, M.-P. Nieh, and J. Katsaras. 2009. Neutron and x-ray scattering for biophysics and biotechnology: Examples of self-assembled lipid systems. Soft Matter 5 (14), 2694–2703. Höhne, G., W. F. Hemminger, and H.-J. Flammersheim. 2003. Differential Scanning Calorimetry (2nd ed.). New York: Springer. Hristova, K. and S. White. 1998. Determination of the hydrocarbon core structure of luid dioleoylphosphocholine (DOPC) bilayers by x-ray diffraction using speciic bromination of the double-bonds: Effect of hydration. Biophysical Journal 74 (5), 2419–2433. Huster, D., K. Arnold, and K. Gawrisch. 1998. Inluence of docosahexaenoic acid and cholesterol on lateral lipid organization in phospholipid mixtures. Biochemistry 37 (49), 17299–17308. Kachel, K., E. Asuncion-Punzalan, and E. London. 1998. The location of luorescence probes with charged groups in model membranes. Biochimica et Biophysica Acta—Biomembranes 1374 (1–2), 63–76. Katsaras, J., R. Stinson, J. Davis, and E. Kendall. 1991. Location of two antioxidants in oriented model membranes: Small-angle x-ray diffraction study. Biophysics Biophysical Journal 59, 645–653. Kessel, A., N. Ben-Tal, and S. May. 2001. Interactions of cholesterol with lipid bilayers: The preferred coniguration and luctuations. Biophysical Journal 81 (2), 643–658. Komljenovic, I., D. Marquardt, T. A. Harroun, and E. Sternin. 2010. Location of chlorhexidine in DMPC model membranes: A neutron diffraction study. Chemistry and Physics of Lipids 163, 480–487. Kondo, M., M. Mehiri, and S. L. Regen. 2008. Viewing membrane-bound molecular umbrellas by parallax analyses. Journal of the American Chemical Society 130 (41), 13771–13777.

Locations of Small Biomolecules in Model Membranes

215

Kucˇ erka, N., D. Marquardt, T. A. Harroun, M.-P. Nieh, S. R. Wassall, D. H. de Jong, L. V. Schafer, S. J. Marrink, and J. Katsaras. 2010. Cholesterol in bilayers with PUFA chains: Doping with DMPC or POPC results in sterol reorientation and membrane-domain formation. Biochemistry 49, 7485–7493. Kucˇ erka, N., D. Marquardt, T. A. Harroun, M.-P. Nieh, S. R. Wassall, and J. Katsaras. 2009a. The functional signiicance of lipid diversity: Orientation of cholesterol in bilayers is determined by lipid species. Journal of the American Chemical Society 131, 16358–16359. Kucˇ erka, N., M.-P. Nieh, J. Pencer, J. N. Sachs, and J. Katsaras. 2009b. What determines the thickness of a biological membrane. General Physiology and Biophysics 28 (2), 117–125. Kyrikou, I., I. Daliani, T. Mavromoustakos, H. Maswadeh, C. Demetzos, S. Hatziantoniou, S. Giatrellis, and G. Nounesis. 2004. The modulation of thermal properties of vinblastine by cholesterol in membrane bilayers. Biochimica et Biophysica Acta—Biomembranes 1661 (1), 1–8. Leftin, A. and M. F. Brown. 2011. An NMR database for simulations of membrane dynamics. Biochimica et Biophysica Acta—Biomembranes 1808 (3), 818–839. Léonard, A., C. Escriv, M. Laguerre, E. Pebay-Peyroula, W. Néri, T. Pott, J. Katsaras, and E. J. Dufourc. 2001. Location of cholesterol in DMPC membranes. A comparative study by neutron diffraction and molecular mechanics simulation. Langmuir 17, 2019–2030. Leonenko, Z. V., E. Finot, and D. T. Cramb. 2007. Atomic force microscopy to study interacting forces in phospholipid bilayers containing general anesthetics. In: A. M. Dopico (ed.), Methods in Membrane Lipids, Chapter 40, pp. 601–609. Totowa, New Jersey: Humana Press Inc. Lomize, A. L., I. D. Pogozheva, and H. I. Mosberg. 2011. Anisotropic solvent model of the lipid bilayer. 2. Energetics of insertion of small molecules, peptides, and proteins in membranes. Journal of Chemical Information and Modeling 51, 930–946. Maghraby, G. E., A. Williams, and B. Barry. 2005. Drug interaction and location in liposomes: Correlation with polar surface areas. International Journal of Pharmaceutics 292 (12), 179–185. Marquardt, D., J. A. Williams, N. Kucˇ erka, J. Atkinson, S. R. Wassall, J. Katsaras, and T. A. Harroun. 2013. Tocopherol activity correlates with its location in a membrane: A new perspective on the antioxidant vitamin E. Journal of the American Chemical Society 135, 7523–7533. Marsan, M., I. Muller, C. Ramos, F. Rodriguez, E. Dufourc, J. Czaplicki, and A. Milon. 1999. Cholesterol orientation and dynamics in dimyristoylphosphatidylcholine bilayers: A solid state deuterium NMR analysis. Biophysical Journal 76 (1, Part 1), 351–359. Mason, T. E., T. J. Gawne, S. E. Nagler, M. B. Nestor, and J. M. Carpenter. 2013. The early development of neutron diffraction: Science in the wings of the Manhattan project. Acta Crystallographica Section A 69 (Part 1), 37–44. Mavromoustakos, T. M. 2007. The use of differential scanning calorimetry to study drug–membrane interaction. In: A. M. Dopico (ed.), Methods in Membrane Lipids, Chapter 39, pp. 587–600. Totowa, New Jersey: Humana Press Inc. Missner, A., P. Kügler, Y. N. Antonenkod, and P. Pohl. 2008. Passive transport across bilayer lipid membranes: Overton continues to rule. Proceedings of the National Academy of Sciences of the United States of America 105 (52), E123. Overton, E. 1895. About the osmotic properties of living plant and animal cell. Vierteljahrsschr Naturforsch Ges Zuerich 40, 159–201. Papahadjopoulos, D. and N. Miller. 1967. Phospholipid model membranes. I. Structural characteristics of hydrated liquid crystals. Biochimica et Biophysica Acta (BBA)—Biomembranes 135 (4), 624–638. Papahadjopoulos, D., M. Moscarello, and E. E. T. Isac. 1975. Effects of proteins on thermotropic phase transitions of phospholipid membranes. Biochimica et Biophysica Acta 401, 317–335. Reichardt, C. 1994. Solvatochromic dyes as solvent polarity indicators. Chemical Reviews 94, 2319–2358. Saint-Laurent, A., P. Boudreau, R. Gaudreault, P. Poyet, and M. Auger. 1998. Interaction between lipid bilayers and a new class of antineoplastic agents derived from arylchloroethylurea: A H-2 solid-state NMR study. Biochemistry and Cell Biology—Biochimie et Biologie Cellulaire 76 (2–3), 465–471. Shaikh, S., V. Cherezov, M. Caffrey, S. Soni, D. LoCascio, W. Stillwell, and S. Wassall. 2006. Molecular organization of cholesterol in unsaturated phosphatidylethanolamines: X-ray diffraction and solid state H-2 NMR reveal differences with phosphatidylcholines. Journal of the American Chemical Society 128 (16), 5375–5383. Shaw, J., R. Epand, Z. Li, R. Bittman, and C. Yip. 2006. Correlated luorescence–atomic force microscopy of membrane domains: Structure of luorescence probes determines lipid localization. Biophysical Journal 90 (6), 2170–2178. Shrivastava, S., S. Haldar, G. Gimpl, and A. Chattopadhyay. 2009. Orientation and dynamics of a novel luorescent cholesterol analogue in membranes of varying phase. Journal of Physical Chemistry B 113 (13), 4475–4481.

216

Liposomes, Lipid Bilayers and Model Membranes

Silvius, J. 2003. Fluorescence energy transfer reveals microdomain formation at physiological temperatures in lipid mixtures modeling the outer lealet of the plasma membrane. Biophysical Journal 85 (2), 1034–1045. Singer, M. and L. Finegold. 1990. Cholesterol interacts with all of the lipid in bilayer-membranes—Implications for models. Biophysical Journal 57 (1), 153–156. Spasojevic, I. 2011. Free radicals and antioxidants at a glance using EPR spectroscopy. Critical Reviews in Clinical Laboratory Sciences 48 (3), 114–142. Squires, G. L. 1978. Introduction to the Theory of Thermal Neutron Scattering. Mineola, New York: Courier Dover Publications. Stella, L., M. Burattini, C. Mazzuca, A. Palleschi, M. Venanzi, I. Coin, C. Peggion, C. Toniolo, and B. Pispisa. 2007. Alamethicin interaction with lipid membranes: A spectroscopic study on synthetic analogues. Chemistry and Biodiversity 4 (6), 1299–1312. Subczynski, W. K., J. Widomska, and J. B. Feix. 2009. Physical properties of lipid bilayers from EPR spin labeling and their inluence on chemical reactions in a membrane environment. Free Radical Biology and Medicine 46 (6), 707–718. Taraska, J. W., M. C. Puljung, and W. N. Zagotta. 2009. Short-distance probes for protein backbone structure based on energy transfer between bimane and transition metal ions. Proceedings of the National Academy of Sciences of the United States of America 106 (38), 16227–16232. Tipping, E., B. Ketterer, and L. Christodoulides. 1979. Interactions of small molecules with phospholipid bilayers. Binding to egg phosphatidylcholine of some organic anions. Biochemical Journal 180, 327–337. Toppozini, L., C. L. Armstrong, M. A. Barrett, S. Zheng, L. Luo, H. Nanda, V. G. Saka, and M. C. Rheinstädter. 2012. Partitioning of ethanol into lipid membranes and its effect on luidity and permeability as seen by x-ray and neutron scattering. Soft Matter 8, 11839–11849. Tristram-Nagle, S. A. 2007. Preparation of oriented, fully hydrated lipid samples for structure determination using x-ray scattering. In: A. M. Dopico (ed.), Methods in Membrane Lipids, Volume 400 of Methods in Molecular Biology, Chapter 5, pp. 63–75. New Jersey: Humana Press Inc. Trouard, T., A. Nevzorov, T. Alam, C. Job, J. Zajicek, and M. Brown. 1999. Inluence of cholesterol on dynamics of dimyristoylphosphatidylcholine bilayers as studied by deuterium NMR relaxation. Journal of Chemical Physics 110 (17), 8802–8818. Vermeer, L. S., B. L. de Groot, V. Reat, A. Milon, and J. Czaplicki. 2007. Acyl chain order parameter proiles in phospholipid bilayers: Computation from molecular dynamics simulations and comparison with H-2 NMR experiments. European Biophysics Journal with Biophysics Letters 36 (8), 919–931. 20th Joint Biannual Meeting of the Societe-Francaise-de-Biophysiques/Groupe-d’Etudes-des-InteractionsMolecules-Membranes/Groupe-de-Recherche-en-Ingenierie-des-Proteines, Anglet Biarritz, France, October 14–19, 2006. Videira, R. A., M. C. Antunes-Madeira, and V. M. Madeira. 1999. Biophysical perturbations induced by ethylazinphos in lipid membranes. Chemistry and Physics of Lipids 97 (2), 139–153. Williams, J. A., S. E. Batten, M. Harris, B. D. Rockett, S. R. Shaikh, W. Stillwell, and S. R. Wassall. 2012. Docosahexaenoic and eicosapentaenoic acids segregate differently between raft and nonraft domains. Biophysical Journal 103 (2), 228–237. Wustner, D., L. Solanko, E. Sokol, O. Garvik, Z. Li, R. Bittman, T. Korte, and A. Herrmann. 2011. Quantitative assessment of sterol trafic in living cells by dual labeling with dehydroergosterol and BODIPYcholesterol. Chemistry and Physics of Lipids 164 (3), 221–235. Xu, Y., V. Yushmanov, and P. Tang. 2002. NMR studies of drug interaction with membranes and membraneassociated proteins. Bioscience Reports 22 (2), 175–196. Zaccai, G., J. K. Blasie, and B. P. Schoenborn. 1975. Neutron diffraction studies on the location of water in lecithin bilayer model membranes. Proceedings of the National Academy of Sciences 72 (1), 376–380. Zheng, J. 2006. Spectroscopy-based quantitative luorescence resonance energy transfer analysis. In: J. D. Stockand and M. S. Shapiro (eds.), Ion Channels: Methods and Protocols, Volume 337 of Methods in Molecular Biology, Chapter 6, pp. 65–77. Totowa, NJ: Humana Press Inc.

12

Membrane Medicine Georg Pabst and Karl Lohner

CONTENTS 12.1 Introduction .......................................................................................................................... 217 12.2 Membrane Architecture and Composition ........................................................................... 218 12.3 Physics of Membrane Function............................................................................................. 219 12.3.1 Intramembrane Forces and Pressure Fields .............................................................. 219 12.3.1.1 How Lipids Affect Proteins ....................................................................... 219 12.3.1.2 How Proteins Affect Lipids ....................................................................... 221 12.3.2 Domain Formation.................................................................................................... 223 12.4 Membranes: Hot Spots as Mediator of Diseases and Drug Target .......................................224 12.4.1 Membrane Lipid Metabolism and Diseases .............................................................224 12.4.1.1 Altered Membrane Lipid Metabolism in Cancer .......................................224 12.4.1.2 The Role of Cholesterol and Gangliosides in Neurodegenerative Diseases ..................................................................................................... 226 12.4.1.3 Lipid Oxidation: Impact on Membrane Structure and Function ............... 227 12.4.1.4 Lipolytic Products: The Example of Sphingomyelinase Action ................ 229 12.4.2 Membrane Targeting................................................................................................. 230 12.4.2.1 The Action of Anesthetics: Protein or Lipid Mediated? ............................ 230 12.4.2.2 Membrane Discrimination by Antimicrobial Peptides.............................. 232 12.4.2.3 Anticancer Peptides Targeting Exposed Phosphatidylserine ..................... 234 12.4.2.4 Crossing Membrane Borders: Cell Penetrating Peptides ........................... 234 12.5 Conclusions and Future Directions....................................................................................... 237 References ...................................................................................................................................... 237

12.1 INTRODUCTION The role of membranes in diseases has been discussed since decades (Goldberg and Riordan 1986, Escriba et al. 2008, Ashrafuzzaman and Tuszynski 2013). Disruption of lipid metabolic enzymes and pathways has been shown to be involved in diseases such as cancer, neurodegenerative, and some forms of diabetes and obesity. Considering the multifunctional role of membranes, compartmentalization of cells and organelles, anchoring cell wall components or cytoskeleton, signaling and regulating transport across the lipid membrane, etc., it is conceivable that membrane abnormalities affect its physiological behavior. However, there is no coherent scheme by which diseases caused by changes in cell membranes can be classiied owing to the diversity of perturbation, which can be due to changes occurring in the lipid bilayer, membrane proteins or in the cytoskeleton (PetitZeman 2004, Ashrafuzzaman and Tuszynski 2013). Given the importance of membrane lipid composition to maintain the topology, mobility, or activity of membrane proteins and to ensure normal cell physiology, cells have developed robust mechanisms to maintain membrane lipid homeostasis (Oresic et al. 2008). Hence concepts for medical compounds that interfere with one or several of the above named functions may be developed. In general, direct drug/protein interactions are in the focus of drug development schemes to unravel the structural base of so-called key-lock mechanisms, which is underlined by the fast 217

218

Liposomes, Lipid Bilayers and Model Membranes

growing number of structures (>80.000) deposited in the RCSB Protein Data Bank. However, only a minor fraction of about 6% of these structures relates to membrane proteins. On the other hand, membranes are composed of thousands of different proteins and lipids and it is the collective properties of all these membrane constituents that determine its function. Therefore, another strategy appears to be appropriate, namely to inluence the collective membrane properties instead of considering only isolated molecules. This requires a detailed knowledge of membrane lipid composition and possible modiications. Thus, one major challenge of membrane medicine will be the identiication of small differences between lipids under normal and pathological conditions. Novel analytical approaches especially liquid chromatography and mass spectrometry allow for quantitative estimation of a wide range of lipids, and are essential prerequisites to gain new insights into lipid-associated mechanisms of pathology (Wenk 2010, Subramaniam et al. 2011). Today, two large-scale initiatives, the European Lipidomics Initiative and the American LIPID Metabolites And Pathways Strategy (LIPID MAPS), drive the determination of lipid patterns and metabolic pathways in biological systems, with a focus on lipid associated diseases, such as obesity, cardiovascular disease, hypertension, stroke, or cancer to name but a few. So far, several thousand distinct molecular lipid species have been identiied (Dennis et al. 2010). Another challenge of membrane medicine comes from the fact that drugs should have no side effects, which seems to be dificult to be achieved, when the membrane itself is considered to be a rather unspeciic composite. Within this chapter, we will show that membranes have speciic features, which can be targeted and we will review physical concepts that should be applied when developing membrane-active drugs.

12.2 MEMBRANE ARCHITECTURE AND COMPOSITION Compartmentalization is one of the major functions of biomembranes, which enables segregation of speciic processes and deines cells. Prokaryotes usually have either one cytoplasmic membrane (Gram-positive bacteria), or two, an outer and an inner (cytoplasmic) membrane (Gram-negative bacteria), while eukaryotic organisms have, in addition to the cytoplasmic membrane, membranes that enclose the nucleus and organelles. Although the overall structure of membranes is a highly dynamic, liquid–crystalline phospholipid bilayer (Bloom et  al. 1991), which has the feature of a water/oil/water interface, they fulill different functionalities determined by their lipid/protein composition. The concept of a characteristic lipid composition for a given cell membrane is well accepted, although changes in lipid composition may occur depending on patho(physiological) conditions. Here lies the key to regulate speciic functions and hence to the development of membrane medicine. The main lipids of biomembranes, their topology and the spatial organization of their metabolism have been reviewed recently (van Meer et al. 2008). In brief, glycerophospholipids, sphingolipids, glycolipids, and nonpolar sterols are the main membrane lipid classes of eukaryotes, whereby bacteria contain predominantly glycerophospholipids (Figure 12.1). The cytoplasmic bacterial membrane contains mostly phosphatidylethanolamine (PE), phosphatidylglycerol (PG), and diphosphatidylglyercol (DPG or cardiolipin), while phosphatidylcholine (PC) is the most abundant lipid in eukaryotes (Lohner 2001, van Meer et al. 2008). Gram-positive bacteria rather contain a high amount of PG, while PE represents the major phospholipid class of Gram-negative bacteria. PGs are anionic at neutral pH. PEs, in turn, are zwitterionic and have an inverted cone-like molecular shape and therefore are prone to form nonplanar phases, such as the inverted hexagonal phase, imposing a curvature stress onto the membrane (Seddon 1990, Rappolt et al. 2003). Mammalian plasma membranes were shown to exhibit an asymmetric distribution of membrane lipids, where negatively charged phosphatidylserine (PS) is exclusively found in the inner lealet as well as the major fraction of PE (Zwaal and Schroit 1997). In turn, the outer lealet contains mainly the zwitterionic choline phospholipids, PC, and sphingomyelin (SM). Both lipids roughly adopt a cylindrical molecular shape and therefore are lamellar phase formers. However, these lipids differ in backbone chemistry. In particular, the amino group present in SM allows for tighter packing due to the formation of H-bonds. Moreover, because of preferential interaction of cholesterol with SM,

219

Membrane Medicine

LPS Gram-negative bacteria

Human erythrocyte

E. coli: OM PE/PG/DPG 91/3/6 E. coli: IM PE/PG/DPG 82/6/12

PC/SM/ 31/24 PE/PS/PI/PA 29/13/1/2

Outer membrane (OM)

Peptidoglycan layer Gram-positive bacteria

LTA

S. aureus: PG/DPG, 95/5 Inner (cytoplasmic) membrane (IM) Cholesterol

FIGURE 12.1 Scheme of plasma membranes highlighting characteristics of Gram-negative and -positive bacteria as well as erythrocytes. Phospholipid composition of typical representatives is given. (Adapted from Lohner, K. and Prenner, E.J. 1999. Biochim. Biophys. Acta 1462, 141–156; Lohner, K. 2001. In: Development of Novel Antimicrobial Agents: Emerging Strategies, pp. 149–165 and references therein.) Note that Staphylococcu aureus can modify PG by attaching l-lysine to the headgroup and that in erythrocytes anionic lipids are conined to the inner membrane lealet like the largest fraction of PE, while the predominant amount of PC and SM is located in the outer membrane lealet. Lipopolysaccharides (LPS) are the predominant component of the outer lealet of the outer membrane of Gram-negative bacteria. A peptidoglycan layer is located in the periplasmic space between OM and IM and forms together with lipoteichoic acid (LTA) the cell wall of Gram-positive bacteria, respectively.

cholesterol might also be enriched in the outer lealet of mammalian plasma membrane, although data on this issue are still controversial (van Meer et al. 2008). Interaction of cholesterol with sphingolipids is also important to the formation of functional platforms, termed rafts, as discussed in detail in Section 12.3.2. In summary, this means that the surface of plasma membranes, the irst site of drug interaction, has a very distinct lipid organization differing between cells and organisms in a number of properties such as charge, headgroup chemistry, lipid packing and organization, which can be exploited for the design of speciic compounds as has been demonstrated, for example, for novel antibiotics and anticancer agents (see Sections 12.4.2.2 and 12.4.2.3).

12.3

PHYSICS OF MEMBRANE FUNCTION

Within the central layer membrane lipids and proteins forms a compound with given collective properties determined by the properties of the lipids and proteins and their mutual interactions (Marsh 2008b). Chapter 1 of this book gives detailed insight on the physics of membranes, for recent review, see also Pabst (2013). Here, we give only a brief review of those aspects relevant for the subject of this chapter.

12.3.1 IntraMeMbrane Forces and pressure FIelds 12.3.1.1 How Lipids Affect Proteins Membranes are highly dynamic systems over a broad range of length and time scales, that is, from local bond vibrational modes in molecules to large-scale bending luctuation and molecular

220

Liposomes, Lipid Bilayers and Model Membranes

diffusion modes. So in many senses, lipid bilayers are vibrant systems with myriads of molecular collisions. There is no way of talking about single molecules; it is always the collective behavior of hundreds to thousands of lipid molecules that dictate the overall physical appearance of bilayers. At the same time we emphasize, however, that the collective behavior of membranes relates to the individual properties of the constituent lipids. It is this dual picture one has to bear in mind, especially if one thinks about modifying the collective properties by membrane active compounds, that is, drugs. Because of a variation of lateral intramolecular interactions along the bilayer normal, a lateral pressure ield emerges, known as the lateral pressure proile (Figure 12.2). The relevance of the lateral pressure proile becomes immediately apparent, when considering its coupling to protein function. Imagine, for example a trans-membrane ion channel protein, whose function is to selectively allow for the passive transport of ions through a central pore. Opening of the pore, which may be induced for example by ligands or voltage changes, leads to a lateral expansion of the protein. This lateral expansion needs to take place against the lateral pressure ield. Similar arguments can also be used to calculate the work required for protein insertion. Then the expansion of the bilayer needed to incorporate a protein of a given shape is considered. Hence, one can derive the inluence of collective membrane properties on protein sorting, something that is known to occur along the Golgi apparatus. Analogously, one can also derive the preferential partitioning of proteins in speciic lipid environments, for example, rafts.

(a)

p (bar)

(b)

600 400 200 0 –200 –400 –600 –800 –1000

p z (kBT/nm2)

(c)

p z2 (kBT/nm)

(d)

20 0 –20 –40 40 20 0 –20 –40 –60 –80 –3

–2

–1

0 z (nm)

1

2

3

FIGURE 12.2 Lateral pressure proiles in lipid bilayers. (a) A schematic with repulsive pressures in the headgroup and acyl chain regime (dark arrows) and attractive interactions at the polar/nonpolar interface (bright arrow). (b) The pressure proile in palmitoyl phosphatidylcholine bilayers calculated from MD simulations. (Calculated from Jerabek, H. et al. 2010. J. Am. Chem. Soc. 132, 7990–7997.) (c) and (d) Give the corresponding integrands of the irst and second integral moments.

221

Membrane Medicine r

(a) –

0

p(z) +

+

K0 =

[r]0 [t]0

† +

+ +

+ + +

z

(b) –

p(z) 0 +

+

K=

[r] [t]

+

+ +

+ + +

z Closed channel

Open channel

FIGURE 12.3 Mechanical coupling of bulk membrane properties to the conformational equilibrium of an ion channel. The initial equilibrium K0 is characterized by the opening of the central pore of the channel protein (change from state r to t) requiring a lateral expansion of the protein against the lateral pressure proile p(z) (a). If a compound inserts into the membrane the lateral pressure proile may be changed, leading to a new conformational equilibrium K of the pore protein (b). This will be the most effective, if the compound inserts close to the polar/apolar interface, where p(z) exhibits the largest changes.

As an example, consider a protein in a given lipid environment, whose collective properties have been changed, for example, either by the insertion of a membrane active compound (drug) or by changes in lipid composition, for example, by enzymatic activity. We now ask how these changes in bulk membrane properties effect protein function. For the sake of simplicity, consider again an ion channel, which exhibits over time and over several identical ion channels a distribution of open and closed states (Figure 12.3), described by a thermodynamic equilibrium constant, irst considered by Cantor (Cantor 1997). The distribution of closed to open states changes in a modiied ield of lateral pressures and consequently also the equilibrium constant. 12.3.1.2 How Proteins Affect Lipids Whenever, a protein interacts with a lipid membrane, either by adsorbing or inserting (partially or completely), it interferes with the structural and elastic properties of the lipid bilayer. Additionally electrostatic interactions are of importance. In this section, we focus on elastic interactions only. For example, large proteins adsorbing to the membrane surface may act as scaffolds to introduce global membrane curvature, which is energetically optimized with bilayer elasticity, for review see (Zimmerberg and Kozlov 2006). Insertion of proteins into lipid membranes in general leads to a local deformation of the bilayer. For example, partial insertion of amphiphathic peptides (e.g., antimicrobial peptides, AMPs) causes the formation of a local dimple (Figure 12.4a) (Huang 2000). Integral membrane proteins, or pore-forming peptides, in turn may have a hydrophobic length that does not match the hydrophobic thickness of the lipid bilayer (Figure 12.4b) (Dan et al. 1993). Hydrophobic matching, then may lead to local distortions of the bilayer in the vicinity of the inserted protein or to tilting or conformational changes of the transmembrane helices or even protein aggregation (Dan and Safran 1998, Killian 2003). In fact, the activity of various membrane-bound enzymes and transporters is reported to depend on lipid chain length (see, e.g. Johannsson et  al. 1981a,b,

222

Liposomes, Lipid Bilayers and Model Membranes (b) (a) 0

dB

dB

dP

(c)

FIGURE 12.4 Examples of protein induced changes in bilayer structure. Local deformations may be induced by partially inserted proteins, for example, amphiphathic peptides (a) or integral proteins, whose projected hydrophobic length does not match the bilayer thickness (b). A case where the hydrophobic length of the protein projected onto the bilayer normal (dP) is larger than the bilayer thickness (d B0 ), dP > d B0 . In lipid membranes with high spontaneous curvatures also global curvature changes may be induced (c). Hydrophobic protein surfaces are colored in black.

Montecucco et al. 1982, Froud et al. 1986, Perozo et al. 2002), that is, optimum activity is found for a given lipid composition. Thus, there is an elastic feedback of the membrane to protein inclusion that may couple again to protein function. The major contributions to membrane deformation are due to local compression-expansion and splay-distorsion, whereas contributions from Gaussian curvature modes have been estimated to be insigniicant and are neglected consequently (Huang 1986, Nielsen and Andersen 2000). An additional term for incomplete adaption can be also considered (Marsh 2008a). The lateral extension of the adaptation region of the lipid curvature to the protein depends on the extent of hydrophobic mismatch. Of particular interest are, however, the irst-shell lipids (Marsh 2007, 2008a) at the protein/lipid interface often referred to as annular lipids, which act as solvents for the protein inclusions (Lee 2004). The reason for this interest is that lipids beyond the irst shell were reported to not inluence the association energies with the proteins (Powl et al. 2007). Hence, the largest contribution to hydrophobic coupling is thought to originate from irst-shell lipids (Marsh 2008a,b). Annular lipids need to be distinguished from lipids, which bind to speciic sites on the protein (Lee 2005). Hence, annular lipids interact with the protein nonspeciically. They do not bind, but may exchange with outer-shell lipids, although their lateral diffusion is slowed down signiicantly due to grooves on the protein surface such that they move essentially with the protein. This zone of correlated protein/lipid movement extends beyond the irst shell lipids and MD-simulations reported strongly correlated low within a distance of 5–6 nm from the protein surface (Niemela et al. 2010), which appear to contrast experimental indings (Powl et al. 2007). Presently, all these considerations do not include the possibility of alleviating deformation energies originating from hydrophobic mismatch by diffusion of lipids into the lipid/protein interface that by their molecular shape (chain length, spontaneous curvature) it ideally and minimize if not nullify this contribution, a situation that is highly plausible in natural systems. Besides the local deformation interactions of protein inclusions in membranes, there is signiicant evidence that proteins have long-range (global) effects on membrane elasticity and structure. For example, pore formation by alamethicin in lipid membranes leads to a four-fold decrease of the membrane rigidity (Pabst et al. 2007), even if there is no decrease in membrane thickness, that is,

Membrane Medicine

223

bilayer thickness matches the hydrophobic length of alamethicin without deformation (Pan et al. 2009). This indicates that the protein surface and in particular the amino acid side chains may affect the lateral pressure proile signiicantly. In line with this considerations are frequently observed shifts in main transition temperatures of lipid bilayers in the presence of amphiphatic peptides and with a broadening of the transition regime (Lohner and Blondelle 2005), which signiies a decrease of cooperativity. For lipid bilayers with signiicantly stored curvature frustration energy, that is, consisting of lipids with cone-like molecular shapes and consequently high spontaneous curvature, such as, for example, PEs, insertion of peptides may release the curvature frustration energy and lead a global change in membrane curvature (Figure 12.4c). In particular, nonlamellar phases, such as inverted hexagonal or cubic phases, are frequently observed (see, e.g. van der Wel et al. 2000, Hickel et al. 2008), see also Chapter 2. In summary, effects of proteins on membranes depend on protein structure and conformational lexibility and in particular on the distribution of hydrophilic to hydrophobic surfaces relative to a given bilayer structure and elasticity.

12.3.2

doMaIn ForMatIon

Natural membranes contain thousands of different lipid species as outlined in Section 12.2. Hence, within a bilayer these lipids will interact to minimize their free energy that arises from their individual properties (structure, packing and chain order, hydrogen bond forming abilities, entropy, etc.). Neglecting headgroup interactions, it can be stated qualitatively that lipids prone to form gel phases (saturated lipid species) and lipids prone to form luid phases (unsaturated lipid species) will phase separate over a broad range of temperatures and compositions. Cholesterol and its preferred interactions with saturated lipid species in a way enhance this effect such that ternary mixtures of unsaturated lipid, saturated lipid, and cholesterol in many cases exhibit domains with sizes on the order of a few microns (Figure 12.5d) (Marsh 2009), see also chapter eight. Mixtures mimicking the outer lealet of plasma membranes typically contain cholesterol and either PC or SM or both. In the late 1980s, beginning 1990s irst phase diagrams of binary mixtures of PCs with cholesterol were published (Ipsen et al. 1987, Vist and Davis 1990), showing the famous miscibility gap, where a cholesterol-poor, liquid disordered phase, Ld, would coexist with a cholesterol-rich

FIGURE 12.5 (See color insert.) Schematics of different mixing scenarios of lipid A (green squares) and lipid B (blue squares). (a) Ideal (random) mixing, (b) nonrandom mixing (phase luctuations), (c) nanoscopic domains, (d) macroscopic domains, and (e) domain formation via (e.g., electrostatic) interactions with membrane active compounds (red rectangle).

224

Liposomes, Lipid Bilayers and Model Membranes

liquid ordered phase, Lo. It turned out that these phase diagrams, which are well-established in many scientists’ memory, were lawed by the experimental window of the applied techniques. Questioned originally in 2005 (Veatch and Keller 2005), the view has now shifted to a phase diagram with a gradual evolution from Ld to Lo, without the presence of a miscibility gap (Marsh 2010). Another issue of mammalian membranes relates to asymmetry (Section 12.2), which is maintained by an array of lipid translocases (lippases). Cholesterol interacts differently with these membrane lipids, thus domains (rafts) will form in the outer SM-enriched membrane lealet, while lipid mixtures of the inner lealet do not exhibit phase separation on their own. So the question is whether there is a coupling between the outer and the inner lealet of the plasma membrane that allows to transmit signals related to raft formation. Here, luorescence interference contrast microscopy on solid-supported membranes was instrumental to show that in many lipid mixtures—the more complex, the better—the domain structure of the outer membrane lealet was projected exactly onto the inner one (Kiessling et al. 2009). Recently, this was also conirmed in asymmetric lipid vesicles (Chiantia and London 2012). Owing to all the buzz about functional domains in mammalian membranes, domains in bacterial membranes have received far less attention. There is, however, evidence for domain formation of PGs and PEs in few bacterial strains (Epand and Epand 2009c), which have been implicated to be important for certain regulatory functions of the cells (Norris and Fishov 2001, Vanounou et al. 2003, Mileykovskaya and Dowhan 2005, Matsumoto et al. 2006). Indeed, biophysical studies on model membranes composed of PGs and PE have demonstrated nonideal mixing behavior (Garidel and Blume 2000, Lohner et al. 2001, Pozo et al. 2005). Knowledge on the size and stability of these domains has presently by far not reached that of mammalian membrane mimetic systems. Domain formation in bacterial membranes is known to occur also for another reason. If a polyvalent ion, such as a cationic peptide, interacts with anionic membranes, electrostatic forces will cause a lipid de-mixing into peptide-enriched and peptidepoor domains (Figure 12.5e). There is ample evidence for such scenarios from studies on host-defense peptides (Lohner and Blondelle 2005, Lohner et al. 2008). In fact, such de-mixing might be lethal to cell function. Again more research on peptide-induced lipid domains is of need.

12.4 12.4.1

MEMBRANES: HOT SPOTS AS MEDIATOR OF DISEASES AND DRUG TARGET MeMbrane lIpId MetabolIsM and dIseases

As outlined in Section 12.1, there is no coherent scheme to classify membrane diseases but ample evidence that changes in lipid metabolism affecting membrane organization and function is involved in diseases such as cancer, Alzheimer, diabetes, and inlammation. Here, we focus on cancer and neurodegenerative diseases and stress response. 12.4.1.1 Altered Membrane Lipid Metabolism in Cancer Alterations in lipid metabolism play a central role in the pathogenesis of cancer (Hirsch et al. 2010, Cairns et al. 2011, Santos and Schulze 2012, Zhang and Du 2012). Increased lipogenesis, considered as a hallmark of cancers (Kuhajda et al. 1994, Swinnen et al. 2006), as well as lipolytic remodeling of lipid species, respectively, have been reported (Nomura et al. 2011). The expression and activity of a number of enzymes involved in fatty acid synthesis are up-regulated in many types of cancer to meet among other functions the need of membrane biogenesis in the highly proliferating cancer cells (Zhang and Du 2012). Therefore, it is of interest to gain knowledge on the physicochemical properties of the newly generated lipids, as they serve as important building blocks for membranes determining their mechanical properties as well as domain organization, which in turn govern membrane function as described in Section 12.3. Modiications that affect membrane luidity and organization relate particularly to changes in the fatty acid proile of cancer membrane lipids, as well as cholesterol content of cancer plasma

225

Membrane Medicine

membranes. Normally the cholesterol content of cell membranes is tightly regulated and dysregulation of these mechanisms has pathological implications as shown in many malignancies (Cortes et al. 2013). Loss of feedback regulation of cholesterol absorption and synthesis resulted in up-regulation of the cholesterol synthesis pathway and in an increase of LDL receptor expression in prostate and other types of cancer cells (Hentosh et  al. 2001, Chen and Hughes-Fulford 2001). Consequently, accumulation of cholesterol has been reported in various solid tumor tissues, including prostate, breast, and oral cancer as well as hepatoma (Van Blitterswijk et al. 1984, Hager et al. 2006, Li et al. 2006). Increased levels of cholesterol will result in decreased membrane luidity, that is, rigid cancer cell membranes. Although studies which directly link fatty acid and phospholipid synthesis are scarce, recent mass spectrometric phospholipid analysis of tumor tissues indicated that cancer cells rather contain higher amounts of saturated fatty acids than normal cells (Rysman et al. 2010, Hilvo et al. 2011). Moreover, differentiation had a strong impact on the lipid proiles, whereby most aggressive tumors exhibited the highest overall amount of saturated membrane phospholipids making cancer cell membranes again less luid (Hilvo et al. 2011). Decreased membrane luidity was connected with less sensitivity of tumor cells to oxidative stress and chemotherapeutics such as doxorubicin (Rysman et al. 2010) or vinblastine (May et al. 1988). Higher cholesterol levels were also found in multidrug-resistant ovarian cancer cells (Mazzoni and Trave 1993). On the other hand, there are studies that revealed an increase in cancer membrane luidity (lymphomas, lung carcinomas, neural tumors, metastatic cells) as compared with their healthy counterpart (Nakazawa and Iwaizumi 1989, Sherbet 1989, Campanella 1992, Rodrigues et al. 2002, Sok et al. 2002). However, in this context it is important to consider that membrane luidity itself is not uniform all over the membrane (Sherbet and Jackson 1986, Lingwood and Simons 2010) and a more differentiated view will be needed. Thus, a comprehensive understanding of lipid metabolism and its inluence on membrane properties such as the physical state of the membrane (lateral pressure proile, curvature strain, elasticity) are needed for the development of novel types of anticancer drugs. Finally, a characteristic of normal mammalian cell membranes is the coninement of PS to the inner membrane lealet (Sections 12.2 and 12.3.2). In contrast, loss of PS asymmetry has been reported for a number of cancer cells (Riedl et al. 2011b). PS exposure normally signals the start of apoptosis, but can be circumvented by cancer cells by different pathways (Miyashita and Reed 1993, Soengas et al. 2001). We were also able to demonstrate that cancer-speciic PS exposure was not a signature of apoptosis. The amount of PS being transferred to the outer membrane lealet of tumorigenic melanoma cell lines correlated with tumor progression and cells from primary cancer cell cultures already exposed PS (Riedl et al. 2011a) (Table 12.1). Hence, PS exposure is a promising TABLE 12.1 Exposure of Phosphatidylserine on Cancer Cells in Relation to Noncancer Cells Measured by Annexin V Binding Cell Typea Melanocytes pcc, SBcl2 cl WM35 cl WM9 cl, metastasis WM164 cl, metastasis LNCaP cl 769-P cl TE671 cl

Annexin V Binding 1.0 ± 0.4 4.1 ± 0.6 7.0 ± 1.1 7.6 ± 2.1 11.0 ± 3.9 4.2 ± 0.5 6.7 ± 0.3 14.9 ± 4.1

Comment Noncancer, healthy control cells Melanoma Tumor progression with respect to SBcl2 cl ↓ Tumor progression with respect to SBcl2 cl ↓↓ Tumor progression with respect to SBcl2 cl ↓↓↓ Prostate cancer Renal cancer Rhabdomyosarcoma

Source: Data from Riedl, S. et al. 2011a. Biochim. Biophys. Acta 1808, 2638–2645. a pcc, primary cell culture; cl, cell line.

226

Liposomes, Lipid Bilayers and Model Membranes

overall marker for cancer. Additionally, loss of membrane lipid asymmetry affects the net surface charge and lateral pressure proile of cancer cell membranes and can be targeted by speciically designed cationic membrane-active host defense peptides (see Section 12.4.2.3). 12.4.1.2 The Role of Cholesterol and Gangliosides in Neurodegenerative Diseases A common motif of neurodegenerative diseases like Alzheimer’s, Parkinson’s, or Huntington’s disease is the cell surface deposition of extended, β-sheet peptide ibrils, also termed amyloid (Burke et  al. 2013). The initiation of the earliest potential event of peptide aggregation, as well as the mechanism that triggers neuronal dysfunction, is still unclear (Naeem and Fazili 2011). However, it is well established that membrane surfaces can (i) alter the peptide structure, (ii) affect nucleation and rate of aggregation, (iii) exert inluence on aggregation state by promoting speciic polymorphs, and (iv) stabilize potentially toxic aggregates (Williams and Serpell 2011, Burke et al. 2013). Here, we focus exemplary on Alzheimer’s disease. In amyloid plaques of Alzheimer’s disease, usually two proteolytic fragments of the amyloid precursor protein, a receptor-like transmembrane protein, Aβ(1-40) and Aβ(1-42), are observed. Aβ(1-42) generally forms ibrils more rapidly than Aβ(1-40), which was attributed to its two additional hydrophobic amino acid residues at the C-terminus (Kim and Hecht 2005). The aggregation of these peptides is promoted by their amphipathicity, which also enables the peptides to interact with membrane surfaces, that is, to penetrate into membranes and to modulate membrane functions (Williams and Serpell 2011, Burke et al. 2013). Therefore, strong efforts have been devoted to gain knowledge on the interaction of Aβ and lipid membranes to understand their toxicity mechanism. It was supposed that cell selectivity depended strongly on the content of cholesterol and presence of anionic lipid components (Wakabayashi and Matsuzaki 2007, De Felice et al. 2008, Lin et al. 2008). The presence of cholesterol enhanced the ability of Aβ(1-40) to bind and to perturb lipid bilayers (Yip et al. 2001). This experimental observation is supported by recent all-atom molecular dynamics simulation, which showed that upon incorporation of cholesterol in a POPC bilayer the increased hydrocarbon chain order and consequently decreased membrane luidity promotes the adsorption of Aβ(1-42) (Yu and Zheng 2012). Neutron diffraction experiments showed that insertion of oligomeric forms of Aβ(1-42) into cholesterol-containing PC membrane induced a displacement of membrane cholesterol toward the polar surfaces of the bilayer, changing acyl chain order and hence mechanical properties of the bilayer, which can affect a wide range of membrane processes including intercellular signaling (Ashley et al. 2006). Micro-dissection of senile plaques of Alzheimer’s patients revealed an enrichment of cholesterol to concentrations similar to Aβ itself (Panchal et al. 2010) and supports the results derived from membrane mimetic studies, which suggest a role of cholesterol for the pathogenicity of Alzheimer’s disease. Therefore, age-related changes in membrane composition/mechanics may facilitate an increased cellular susceptibility to Aβ and hence may be associated with an increased risk of Alzheimer’s disease, the most common type of senile dementia in aging populations. In addition, lipid composition of neuronal membranes not only triggers binding of Aβ, but also the conversion of the soluble, nontoxic monomeric Aβ to aggregated toxic β-sheet ibrils (Figure 12.6). Of particular interest are gangliosides, glycolipids, which comprise 5–10% of lipids within the outer membrane lealet and regulate various physiological events at the cell surface (Alberts et al. 2007). Raft-like lipid bilayers composed of GM1, sphingomyelin and cholesterol mediated the formation of amyloid ibrils of Aβ, whereby self-aggregation was accompanied by a conformational transition from an α-helix rich to a β-sheet rich form depending on the Aβ/GM1 molar ratio (Ikeda et al. 2011). The amyloid ibrils formed on membranes were composed of antiparallel β-sheets, while those formed in solution contained parallel β-sheets (Okada et al. 2008). Moreover, ibrils formed on membranes exhibited greater surface hydrophobicity (Fukunaga et al. 2012). Fibrils of Aβ(1-42) appeared to be short and laterally co-assembled on the membrane, while ibrils of Aβ(1-40) were long and extruded to the aqueous phase (Ogawa et al. 2011). Although both peptides have similar lipid speciicity and afinity, the toxicity of Aβ(1-42) is about ten times higher than that of Aβ(1-40). The molecular mechanism(s) of amyloid toxicity is still a matter of debate.

227

Membrane Medicine

In solution

Less toxic fibrils Toxic fibrils

Aβ

α-Helix

GM1 cluster

β-sheet aggregate (~15 mer)

Seed

Aβ/GM1 GM1

Cholesterol

Sphingomyelin

FIGURE 12.6 Working model of gangliosode (GM1)-induced formation of amyloid ibrils. Upon binding to GM1 Aβ changes its conformation from α-helix to β-sheet initiating peptide aggregation. For details, see text. (Reprinted with permission from Ikeda, K. et al. 2011. Mechanism of amyloid beta-protein aggregation mediated by GM1 ganglioside clusters. Biochemistry 50, 6433–6440. Copyright 2011, American Chemical Society.)

At a cellular level, there is growing evidence for a correlation between dysregulation of cellular Ca2+ homeostasis and synaptic degeneration. Several mechanisms have been discussed including speciic membrane receptors for Aβ, carpeting of the peptides at the outer membrane lealet, detergent-like action, and amyloid-speciic stable membrane pores/channels (Williams and Serpell 2011). Electrophysiological recordings of planar lipid bilayers, atomic force microscopy, and molecular dynamics simulations demonstrated that the d-isomer of Aβ forms channel-like structures and exhibits an ion conductance behavior in the bilayer, which is indistinguishable from the natural l-isomer (Capone et  al. 2012, Connelly et  al. 2012). These results are strongly in favor of a nonreceptor-mediated mechanism via amyloids, leading to the formation of Ca2+ -leaking, unregulated channels. Another nonreceptor-related mechanism proposed impairment of the activity of the plasma membrane Ca2+ -ATPase, a key regulator of neural Ca2+ homeostasis. In this model amyloids induce changes in lipid raft composition and/or disruption of the lipid raft structure, in which the Ca2+ -ATPase is embedded (Mata et al. 2011). 12.4.1.3 Lipid Oxidation: Impact on Membrane Structure and Function Recent efforts in lipidomics demonstrated an essential role for lipid oxidation in cell behavior and pathology showing that oxidized phospholipids formed as a consequence of oxidative stress are signiicantly increased in apoptosis as well as in inlammation and are involved in a number of pathological conditions like atherosclerosis, cancer, inlammation, type 2 diabetes, and neurological diseases, with the detailed mechanisms remaining to be established (Kinnunen et  al. 2012, Volinsky and Kinnunen 2013). On the other hand, oxidized phospholipids exhibit also protective properties, as reported, for example, endotoxin-induced tissue damage (Kadl et al. 2004, Bochkov 2007). Presently, we are far from understanding these Janus-faced biological activities. The structural motifs and properties of oxidized phospholipids have been described (Subbanagounder et al. 2000, Fruhwirth et al. 2007, Stemmer and Hermetter 2012a). Major stable products of oxidative fragmentation of polyunsaturated phospholipid oxidation are aldehyde- and carboxy-phospholipids. These lipids contain a long hydrophobic and a short polar substituent, respectively, which folds back to the membrane interface or aqueous surface (Podrez

228

Liposomes, Lipid Bilayers and Model Membranes

+N

O O

O H O

O

O

O

O

O H

O

–O

O

–O

O

P

P

O

O

N+

Aqueous phase

PazePC

HO

H

O

O

PoxnoPC

Hydrophobic core

FIGURE 12.7 Conformation and molecular shape of oxidized phospholipids: Carboxyphospholipid (PazePC) adopting an extended conformation with the negatively charged sn-2 carboxylate group stretching into the water phase; aldehydphospholipid (PoxnoPC) with its neutral but polar sn-2 aldehyde group remaining in the membrane interface. Both lipids adopt a conical molecular shape due to the expanded headgroup area being more pronounced for the aldehydphospholipid. (Presentation of molecular dynamics data of oxidized phospholipids in a bilayer is taken from Biophys. J., 96, Khandelia, H. and Mouritsen, O.G., Lipid gymnastics: Evidence of complete acyl chain reversal in oxidized phospholipids from molecular simulations, 2734–2743, Copyright (2009), with permission from Elsevier.)

et al. 2002, Greenberg et al. 2008) as also shown by in silico studies (Wong-Ekkabut et al. 2007, Khandelia and Mouritsen 2009) (Figure 12.7). The different conformations between aldehyde- and carboxy-phospholipids result in a microenvironment being less polar and less hydrated for aldehydophospholipids (Pande et al. 2010). In addition, aldehydophospholipids can form a Schiff base with membrane-associated proteins, peptides, and aminophospholipids (Fruhwirth et al. 2007, Stemmer and Hermetter 2012a). These two oxidized phospholipids are also characterized by a different molecular shape as compared with their parent diacyl phospholipid (Stemmer and Hermetter 2012a). While PC molecules adopt a cylindrical molecular shape with similar cross-sectional areas of headgroup and acyl chains, aldehydo- and in particular carboxy-phospholipids exhibit a conical molecular shape owing to the increased headgroup area by the extended sn-2 chain (Figure 12.7). Such lipids promote the formation of positive membrane curvature strain (Smith et al. 2009, Epand et al. 2009b), which may have a number of structural and functional implications. For example, a strongly enhanced lip-lop rate was reported for luorescent labeled PS in the presence of oxidatively modiied lipids, which depended on degree of oxidation and membrane curvature (Volinsky and Kinnunen 2011). It was suggested that this property may be of signiicance as oxidative stress is associated with processes such as apoptosis and cancer, where PS relocates from the inner to the outer membrane lealet (see Section 12.4.1.2). In ternary lipid mixtures mimicking membrane rafts, lipid oxidation has been found to stabilize phase segregation and can cause formation of large domains (Stottrup et al. 2004, Coban et al. 2007). Gradual exchange of PC with its oxidized carboxylate analog postponed the miscibility transition in a concentration-dependent manner and abolished it above a critical concentration (Volinsky et al. 2012). This observation, supported by molecular dynamics simulation and X-ray diffraction studies (Mason et al. 1997, Wong-Ekkabut et al. 2007, Vernier et al. 2009), was attributed to increased hydrophobic mismatch, which is expected owing to the truncation of the sn-2 chain. The decrease in bilayer thickness of the oxidized lipid-enriched phase will repulse lipids such as cholesterol and sphingomyelin from this phase favoring the formation of large extended lipid domains as observed

Membrane Medicine

229

in model membranes (Jacob and Mason 2005). Such changes in two- and three-dimensional organization of the membrane together with the reactive lipid derivatives suggest interference with lipid–protein interactions, inluencing metabolic pathways (Volinsky et  al. 2013). Similar effects on membrane organization have been reported by phospholipase activity (Holopainen et al. 2000), which will be addressed below (see Section 12.4.1.4). Induction of membrane heterogeneity and positive membrane curvature strain may also lead to preferred partitioning of proteins and aminophospholipids in distinct membrane domains, which may alter the activity of membrane proteins (Stemmer and Hermetter 2012a). Indeed, proteomic analysis in living cells showed that Schiff base formation of polypeptides with a luorescent carboxyphospholipid analog is very selective (Stemmer et al. 2012b). Apart from possible effects of oxidized phospholipids on protein structure, the presence of oxidized phospholipids in the membrane bilayer may facilitate the activity of interfacial active proteins or peptides. For example, studies on the activity of phospholipase A2 revealed that aldehydophospholipids stimulated its activity on DPPC (Uchida 2000, Code et al. 2010). Moreover, incorporation of aldehydophospholipids into mono- and bilayers enhanced the binding of various amphipathic AMPs (Mattila et al. 2008). In addition to altered membrane properties, Schiff’s base formation between the amino groups of the peptides and the lipid aldehyde function seemingly plays an important role, because incorporation of peptides into the membrane was strongly attenuated by methoxyamine and was less eficient in carboxyphospholipid containing model membranes. 12.4.1.4 Lipolytic Products: The Example of Sphingomyelinase Action Sphingolipids and especially SM are important lipid components of the outer plasma membrane lealet of mammalian cells. Interest in sphingolipids has been spurred by two aspects. First, sphingolipids have been shown to be pivotal in signal transduction and regulation of cell function. Cell proliferation, differentiation, cell cycle arrest or apoptotic cell death are some of the well-known examples (Levade and Jaffrezou 1999). Second, sphingolipids are important structuring elements in membrane heterogeneity, that is, together with cholesterol they are considered to form the lipid part of membrane rafts (Pike 2006). It is more than likely that these two aspects are linked to each other. Sphingomyelinase, for example, a speciic phospholipase C that is part of the sphingolipid signal transduction pathway, catalyzes the hydrolysis of membrane-bound sphingomyelin to ceramide and water soluble phosphocholine group (Goni and Alonso 2002). Various cellular processes have been related to SMase activity, among which is, most prominently, apoptosis (programmed cell death) (van Blitterswijk et al. 2003, Hannun and Obeid 2008). Ceramide has physical properties, which are very different from sphingomyelin, that is, it does not form a bilayer and melts at ~90°C (Shah et al. 1995), leading to diverse membrane restructuring effects (Goni and Alonso 2006). Besides, ceramide may also play a role as second messenger, but this is still a matter of debate (van Blitterswijk et al. 2003, Hannun and Obeid 2008). Here, we focus on the membrane structure modulatory role of ceramide and how this may couple to ionchannel function. For example, it has been demonstrated in SM containing lipid-only model systems that the activity of SMase leads to vesicle aggregation (Ruiz-Arguello et al. 1996) and blebbing (Holopainen et al. 2000) due to an asymmetric distribution of ceramide. When distributed symmetrically between both membrane lealets, ceramide leads to phase separation into luid and gel domains (Pabst et al. 2009, Boulgaropoulos et al. 2010) (Figure 12.8). On the molecular level, this can be understood on the basis of hydrogen bonds between the amide groups of SM and ceramide (Boulgaropoulos et al. 2011). This interaction is so strong that it even leads to an exclusion of cholesterol from the SM/ceramide domains (Boulgaropoulos et al. 2012). Consequently, the coexisting luid phase is depleted from SM but enriched in cholesterol upon increasing overall ceramide levels. Recently, we asked how such changes in lateral membrane structure would affect the functioning of an ion-channel. To address this issue we irst determined bending rigidities, Gaussian moduli of curvature and spontaneous curvatures of the luid and the gel domains, both in the absence and presence of cholesterol (Boulgaropoulos et al. 2012). These are integral parameters of the lateral

230

log (I) (arb. units)

I (arb. units)

Liposomes, Lipid Bilayers and Model Membranes

SAXS

50 WAXS 40 30 20 10 1.2

Cer = 15 5 μm

Cer = 0

1.4 1.6 q (Å–1)

Cer = 15

Cer = 0 0.0

0.2

0.4

0.6

0.8

q (Å–1)

FIGURE 12.8 Fluid/gel phase coexistence induced by C16:0 ceramide in equimolar mixtures of palmitoyl oleoylphosphatidylcholine and sphingomyelin as evidenced by X-ray scattering and luorescence microscopy. In the absence of ceramide (Cer = 0), the small-angle X-ray pattern (SAXS) shows a single lamellar and the wide-angle pattern a broad diffuse peak of luid acyl chains. In the presence of 15 mol% ceramide (Cer = 15), a second lamellar phase is observed in the SAXS regime (arrows) and a sharp chain packing peak, characteristic for a gel phase in the WAXS regime. In luorescence microscopy, domains with anisotropic boundaries are observed (insert). (With kind permission from Springer Science+Business Media: J. Membr. Biol., Effect of ceramide on nonraft proteins, 231, 2009, 125–132, Pabst, G. et al.)

pressure proile (Section 12.3.1). Then, we assumed a simple geometric model for an ion channel and calculated the changes in the opening probability in the different lipid environments. We found that the ion channels would be completely inhibited in the gel domains. In turn, if the channel would reside within the luid domains, we calculated a strong increase of open states in the absence of cholesterol, whereas changes were only minute in the presence of cholesterol. This demonstrates an important role of cholesterol in maintaining membrane stability and function.

12.4.2

MeMbrane targetIng

12.4.2.1 The Action of Anesthetics: Protein or Lipid Mediated? Anesthesia is an indispensable tool in modern surgery and has a long and successful history of more than 160 years. However, because of several dificulties, our knowledge on the pertaining molecular mechanisms remains very limited, because it is dificult to deine and measure endpoints of analgesia, hypnosis, amnesia, and paralysis which all contribute to anesthesia. Nevertheless, consensus has been reached that general anesthetics act on the central nervous system and modulate a large number of ion channels of the neurotransmission system (Urban 2002, Mashour et al. 2005, Franks 2006, Hendrickx et al. 2008, Eger et al. 2008, Eckenhoff 2008). It is therefore largely undisputed that these membrane proteins are targets of anesthetics and that their change of activity proile inally causes anesthesia. Recently, the crystalline structure of a pentameric ligand-gated ion channel was reported showing propofol and deslurane bound to speciic sites of the protein (Nury et al. 2011). Alternatively, Heimburg and coworkers have invoked a model based on soliton propagation in nerve membranes that does not involve speciic protein/drug interactions (Heimburg and Jackson 2005, 2007, Villagran-Vargas et al. 2013). Thus, it is still unclear whether anesthetics change protein function by interacting directly with the proteins or indirectly via a membranemediated mechanism.

231

Membrane Medicine

However, protein- or lipid-mediated anesthesia are not necessarily mutually excluding each other. Several studies suggest that primary points of action for anesthetics are located within transmembrane domains of proteins (Trudell and Bertaccini 2002). This means that the anesthetic-binding sites are directly adjacent to the phospholipid bilayer and even if the drug binds to the protein, it is very likely that it reaches the interaction site by diffusion through the lipid bilayer, thus modulating membrane properties. Therefore, a concerted mechanism of membrane modulation and binding to proteins leading inally to protein/protein interactions may be relevant and underlines the complexity of anesthesia. Following this hypothetical low of events, the lipid membrane would be the irst site of interaction with anesthetics. The modulation of membrane properties by anesthetics and its coupling to ion-channel function via lateral pressures was introduced by Cantor (1997) (see also Section 12.3.1). Importantly, the proposed mechanism is not simply correlated to the overall lipophilicity of anesthetics, but rather to the speciic location of the drug within the membrane, and the consequent changes of the lateral pressure proile of the membrane. Hence, drug insertion is proposed to lead to depth-dependent changes in the lateral pressure proile, which then couple mechanically to ion channel activity. Recently, we were able to show by a combination of molecular dynamics simulations and X-ray scattering that ketamine leads to signiicant shifts of lateral pressures to an extent that it affects the conformational equilibrium of simple geometric models for nACh and NMDA receptors (Jerabek et al. 2010). Depending on the used model, the found IC50 values were as low as 2 mol% and it was estimated that this was well within the range of clinical concentrations (Figure 12.9).

Closed (r) ru ϕa

Open (t)

rl ru ϕb

rc

120

Inhibition (%)

100

Tilted Bent

80 60 40 20 0 –20 0.1

1

10

100

Ketamine (mol%)

FIGURE 12.9 Inhibition of geometric ion-channel models (tilted helices, bent helices) in palmitoyl oleoylphosphatidylcholine bilayers as a function of ketamine concentration. Thin lines show the IC50 concentrations. (Reprinted with permission from Jerabek, H. et al. 2010. Membrane-mediated effect on ion channels induced by the anesthetic drug ketamine. J. Am. Chem. Soc. 132, 7990–7997. Copyright 2010, American Chemical Society.)

232

Liposomes, Lipid Bilayers and Model Membranes

12.4.2.2 Membrane Discrimination by Antimicrobial Peptides We face a worldwide re-emergence of infectious diseases and a rapid increase in pathogenic bacteria that are multi-resistant to commercially available antibiotics causing a global health problem with a strong social and economic impact. As a consequence, the WHO ranked antibiotic resistance as a priority disease and encouraged strongly the development of novel antibiotics to counteract the steady decline of approved antibiotics since the early 80s with linezolid and daptomycin being the only new classes introduced (Nordberg et al. 2005). In light of this situation, alternative agents with completely novel mechanisms of action have to be developed. One highly promising strategy is based on AMPs, effector molecules of innate immunity that provide a irst line of defense against a substantial array of pathogenic microorganisms (Hancock 2001, Zasloff 2002) compiled in the Antimicrobial Data Base (Wang et al. 2009). As outlined previously, the concept of a characteristic lipid composition for a given cell membrane is well accepted and hence the different physicochemical properties of the lipids found in biological membranes allow AMPs to discriminate between bacterial and mammalian cell membranes (Latal et al. 1997, Lohner et al. 1997, Lohner 2001). AMPs mostly kill bacteria within minutes by interacting nonspeciically with the target membrane leading to disruption, which makes the occurrence of resistance unlikely. Thereby, the positive charge of the peptide is essential for initial binding to the negatively charged bacterial membrane surface and its hydrophobicity is needed for insertion into and disruption of the membrane. Both parameters determine the window of activity (Findlay et al. 2010). Thus, the amphipathic topology of AMPs, that is, their physicochemical properties rather than a speciic amino acid sequence, is responsible for their biological activities. Numerous studies demonstrated that AMPs interfere with the integrity of bacterial membranes via diverse mechanisms (for reviews see, for example, Bechinger and Lohner 2006, Bechinger 2009, Lohner 2009, Wimley and Hristova 2011). However, the common ground of these mechanisms is a lethal physical interference with the barrier function of bacterial membranes. The most frequently discussed mode of action includes the formation of toroidal pores (Matsuzaki et  al. 1996, Ludtke et  al. 1996) and the coverage of the membrane surface by peptides (carpet model; Shai 2002). In the toroidal pore model peptides, together with the lipid, assemble into a supra-molecular arrangement of high curvature forming a water-illed pore. Although a number of molecules including α-helical and β-sheet-type peptides have been shown to adopt such a pore under the given experimental conditions (Huang 2006), there is little evidence that the majority of these peptides act in vivo via trans-membrane pores (Wimley 2010). In the carpet model peptides accumulate at the membrane being aligned parallel to the bilayer surface. Above a certain threshold concentration the amphipathic peptides insert into the membrane, which destabilizes the integrity of the membrane bilayer resulting in membrane permeabilization and disruption (Shai 2002). At the molecular level different scenarios may apply, which were summarized in a recent review (Wimley and Hristova 2011). Briely, membrane permeabilization has been described in the form of 1. Lipid clustering or phase separation, creating domains with different physicochemical properties between lipid bulk and peptide-enriched domains creating packing defects at their boundary (Latal et  al. 1997, Lohner 2009, Arouri et  al. 2009, Epand and Epand 2009a, Epand et al. 2010, 2011, Zweytick et al. 2011). 2. The free volume model for peptides being aligned parallel to the membrane plane. This creates “voids” in the hydrophobic core of the membrane, which is compensated by increased trans-gauche isomerization of the acyl chains of the lipids or by moving the opposing membrane lealet toward the hydrophobic face of the peptide. This induces a quasi-interdigitated structure in the gel phase and membrane thinning/dimple formation in the luid phase (Huang 2000, Sevcsik et al. 2007, Sevcsik et al. 2008, Lohner 2009). 3. The sinking raft model describes AMP activity in terms of binding, insertion, and perturbation and allows a formal thermodynamic analysis to predict activity (Pokorny and Almeida 2004, Gregory et al. 2008, Almeida and Pokorny 2009).

233

Membrane Medicine

4. The interfacial activity model, deined as the propensity of imperfectly amphipathic peptides to partition into the membrane interface and to disrupt the normally strict segregation of polar and nonpolar groups of the lipids (Wimley 2010, Rathinakumar and Wimley 2010). 5. Detergent-like dissolution of the membrane in particular at high peptide concentration (Hristova et al. 1997, Bechinger and Lohner 2006). 6. Molecular shape models, in which the interaction of AMPs and lipids are depicted with phase diagrams, which besides of electrostatic interactions also account for the molecular shape of components, to describe the propensity of AMPs to permeabilize a membrane by disrupting the lipid packing (Bechinger and Lohner 2006, Bechinger 2009). Models (i)–(v) may be considered as special cases within the complex generic phase diagram, relecting the impact of the nature of AMP and lipid composition (Figure 12.10). Gram-negative bacteria contain high amounts of nonlamellar phase forming lipids such as PE or cardiolipin (Section 12.2). PE itself can adopt an inverted hexagonal (HII) phase (Rappolt et al. 2003), which was diversely affected in the presence of AMPs: promotion of HII (El Jastimi and Laleur 1999, Epand et al. 2003, Willumeit et al. 2005), abolishment of HII (Hickel et al. 2008), or induction of biocontinous cubic phases (Keller et al. 1996, Hickel et al. 2008). Furthermore, we have shown that AMPs induce cubic phases also in lipid extracts of Escherichia coli and Acholeplasma laidlawii (Staudegger et al. 2000, Zweytick et al. 2008, 2011). This is of particular interest because both bacteria regulate precisely their lipid composition in a narrow window close to a lamellar to nonlamellar phase boundary (McElhaney 1992, Rilfors et al. 1993, Morein et al. 1996). These mixtures are composed of cylindrical lipids (PG) and inverted cone-shaped lipids (PE). Consequently, such bilayers are characterized by high membrane curvature strain (see also Section 12.3.1.2), which can be released upon interaction with AMPs, that is, lowering the lamellar to nonlamellar phase boundary and giving rise to the formation of cubic or inverted hexagonal phases Perforated sheets (wormholes)

Peptide concentration

Micelles

Nonlamellar phases Bilayer Pn3m

Bicelles Toroidal pore

Carpet

Clustering

Interfacial activity HII phase/ nonbilayer phases

PC, SM, PG–

Lipid composition (molecular shape)

PE, DPG–

FIGURE 12.10 Scheme of a generic phase diagram of antimicrobial peptide/phospholipid mixtures. Here two lipids with different molecular shapes as indicated in the igure are mixed and their effect on the macroscopic phase in the presence of AMPs is shown. A variety of macroscopic phases as described in the text are obtained as a function of peptide concentration and lipid composition. For a description of peptide activities on mammalian plasma membranes, cholesterol would have to be considered.

234

Liposomes, Lipid Bilayers and Model Membranes

(Lohner and Blondelle 2005). Gross changes in the lateral pressure proile may not only affect membrane structure but also membrane function considering that PE supposedly controls the conformation of membrane proteins (Gruner 1985) (see Section 12.4.2.1). Similarly, exclusion of certain lipids due to their preferential interaction with AMPs may perturb lipid mixing, such as, for example, existing membrane domains/rafts in bacteria (Section 12.3.2), and therefore inluence function of membrane proteins, which may in turn affect the viability of bacteria (Epand and Epand 2009a). The multiple facets of activities of AMPs promise to provide a strong basis for the development of novel antimicrobial compounds. 12.4.2.3 Anticancer Peptides Targeting Exposed Phosphatidylserine Cancer treatment with conventional chemotherapeutics still has severe side effects, mostly because of insuficient selectivity against healthy mammalian cells (Cassidy and Misset 2002, Kalyanaraman et al. 2002). Additionally, conventional chemotherapeutics can induce drug resistance (Perez-Tomas 2006, Gillet and Gottesman 2010). Both factors adversely affect their successful application. Anticancer peptides (ACPs) derived from host defense peptides, which target selectively differences between cancer and noncancer cell membranes, may circumvent these deicits and drawbacks, and hence represent an alternative strategy for the development of novel anticancer agents (Papo and Shai 2005, Hoskin and Ramamoorthy 2008, Schweizer 2009, Riedl et al. 2011b). Like AMPs, these cationic peptides speciically interact with anionic cell membrane components. As discussed in Section 12.4.1.1, negatively charged PS is transferred from the inner to the outer membrane lealet of cancer cells representing the predominant target of ACPs. Selective binding to PS was demonstrated by co-localization of PS and synthetic ACPs on the outside of cancer cells (Papo et al. 2006). Two general mechanisms, killing by necrosis or apoptosis, have been proposed for ACPs as a consequence of their membrane-related mode of action (Papo and Shai 2005, Schweizer 2009). Both necrosis via cell membrane lysis and apoptosis via the mitochondrial lytic effect are dependent on the presence of PS. As shown for AMPs the molecular mechanism(s) of membrane damage mutually depends on the nature of both peptides and membrane lipid composition (Sevcsik et al. 2008) and thus models analogous to those discussed in Section 12.4.2.2 have been proposed for ACPs (Hoskin and Ramamoorthy 2008, Riedl et al. 2011b). It should be mentioned that on the surface of cancer cells in addition to PS, frequently also other levels of negatively charged components, such as sialic acid residues (Kufe 2009, Bafna et al. 2010) and heparan sulfate (Fadnes et al. 2009), are enhanced. This may facilitate accumulation of positively charged ACPs (Risso et al. 1998, Fadnes et al. 2011). Increased cancer cell surface area owing to increased microvilli will further contribute to binding of a larger amount of peptides per cell (Zwaal and Schroit 1997, Papo and Shai 2005). On the other hand, membrane penetration of ACPs will be mainly governed by membrane luidity. As outlined in Section 12.4.1.1, there is strong evidence that cancer cell membranes are more rigid, which can be overcome by an increase in hydrophobicity of the peptide. However, a subtle balance of hydrophobicity to electrostatic interactions is needed in order to act speciically on cancer cell membranes. Therefore, the differences in lipid composition of cell membranes will determine the interaction with ACPs and hence susceptibility of different types of cancer cells against a certain ACP (Leuschner and Hansel 2004, Mader et al. 2005). Based on successful in vitro experiments, a number of peptides have also been tested successfully in vivo against various cancers. Optimizing the use of ACPs as a cancer drug includes the development of strategies for improved serum stability, as well as administration, where linking ACPs to homing domains, which transport them across the cytoplasmic membrane into the cytosolic compartment of cancer cells, is a promising approach (Riedl et al. 2011b) (see Section 12.4.2.4). 12.4.2.4 Crossing Membrane Borders: Cell Penetrating Peptides Membranes of eukaryotic and prokaryotic cells, including all organelles, are permeable to small and/or hydrophobic compounds, while they constitute a serious barrier for hydrophilic molecules

Membrane Medicine

235

like nucleotides, proteins, and peptides. This is one of the most restrictive factors in the development of drugs (Lipinski et al. 2001) and the transport of biologically active compounds into cells to interact with intracellular targets, which would expand their therapeutic space, remains a major challenge for drug development (Milletti 2012, Keller et al. 2013, Ruzza et al. 2013). Several techniques have been developed to overcome the delivery problem (Zorko and Langel 2005, Heitz et al. 2009). A new avenue was opened about 20 years ago, when peptides were discovered that were able to cross cellular membranes and to translocate different cargos into cells (Frankel and Pabo 1988, Green and Loewenstein 1988, Fawell et al. 1994). Since then, more than hundred peptides have been reported to have cell penetrating activity, whereby Tat, penetratin, transportan, and octaarginine (R8) represent extensively studied variants (Jones and Sayers 2012). For a classiication of cell-penetrating peptides (CPPs) based on their physical–chemical properties (cationic, primary or secondary amphipathic, hydrophobic) or based on their origin the reader is referred to a recent review by Milletti (2012). Cell-penetrating peptides share some general properties with AMPs: they are typically composed of 5–30 amino acids and mostly cationic, have in general no sequence homology, exhibit amphipathic character and frequently show structural plasticity. Therefore, it is not surprising, that one may ind cell-penetrating peptides that exhibit antimicrobial activity and vice versa AMPs that translocate through cell membranes (Henriques et  al. 2006, Splith and Neundorf 2011). Furthermore, the molecular mechanism of membrane translocation still is controversial; partly due to different experimental approaches and conditions applied in different laboratories (Heitz et  al. 2009, Raagel et  al. 2010). Nevertheless, CPPs apparently can use two pathways to enter cells: (i) direct energy-independent translocation through the membrane, also termed spontaneous membrane translocation, and (ii) internalization by endocytosis (Heitz et al. 2009, Raagel et al. 2010, Schmidt et al. 2010, Madani et al. 2011, Marks et al. 2011). Although a general mechanism for peptide membrane translocation does not exist, there is consensus that binding of CPPs to the membrane surface, analogous to AMPs and ACPs, is dominated by electrostatic interactions and that the pathway depends on peptide concentration and membrane lipid composition of a particular cell type (Heitz et al. 2009). Below a certain threshold concentration, arginine-rich peptides like Tat, penetratin, and poly-R were shown to enter primarily by endocytosis and above via direct translocation (Duchardt et al. 2007, Tunnemann et al. 2008, Kosuge et al. 2008), whereby the presence of cell surface proteoglycans such as heparan sulfate triggers the binding of CPPs and their accumulation at the membrane surface (Duchardt et al. 2007, Ziegler and Seelig 2008). A number of mechanisms were proposed for nonendocytotic, direct translocation of CPPs through membranes. Some of the suggested models such as formation of inverted micelles (Derossi et al. 1994), or pore formation (Suzuki et al. 2002) resemble features described above for the activity of AMPs (see Section 12.4.2.2). An “adaptive translocation” was proposed (Wender et  al. 2008), which involves the clustering of anionic cell surface components by the cationic transporter molecule facilitating diffusion into the membrane and subsequently into the cytosolic compartment of the cell (Rothbard et al. 2004). The driving force for the translocation was attributed to the membrane potential, which varies between cell types emphasizing the importance of knowing the membrane composition for ine tuning of such a process. The importance of arginine residues, speciically the hydrogen-bond formation of the guanidinium moiety in arginine with phosphate of lipids or sulfates of cellular components, for peptide membrane translocation was early recognized (Rothbard et  al. 2002, 2004, Sakai et  al. 2005, Wender et  al. 2008) and thus novel guanidinium-rich transporters were designed (Rothbard et  al. 2002, 2004, Wender et al. 2008). Primarily based on studies of Tat peptide and synthetic analogs it was shown that arginine-rich peptides induce negative membrane curvature in membrane mimetics, manifested in the generation of the bicontinuous cubic phase Pn3m (Mishra et al. 2008). This type of curvature is characterized by saddle-shaped deformations and is facilitated by the bidentate hydrogen bond formation between the guanidinium group and the phosphates thereby crosslinking multiple

236

Liposomes, Lipid Bilayers and Model Membranes

FIGURE 12.11 (See color insert.) Saddle-shaped membrane curvature (a) promoted by bidentate H-bonding between guanidinium residue of Arg and phosphates of lipids shown in the inset. Such curvature is topologically necessary for the formation of membrane pores (b) and membrane protrusions such as those in macropinocytosis (c). (Adapted and modiied with permission from Schmidt, N.W. et al. 2012. Molecular basis for nanoscopic membrane curvature generation from quantum mechanical models and synthetic transporter sequences. J Am Chem. Soc. 134, 19207–19216. Copyright (2010) Elsevier and Copyright (2012) American Chemical Society.)

lipids (Schmidt et al. 2012) (Figure 12.11). In agreement with quantum mechanical calculations, synthetic guanidine polymers showed that the generation of curvature is very sensitive to the guanidinium group spacing (Mishra et al. 2011). It was suggested that small cargo molecules can be directly transported by generation of negative membrane curvature inducing pore-like structures. In case of larger cargo molecules (larger than a few nm), the Tat peptide interacts on the cytoplasmic side with the cytoskeleton remodeling it to induce cellular uptake (Mishra et al. 2011, Maniti et al. 2012). Using plasma membrane spheres, which represent an elaborated cellular model for plasma membranes, it was shown that peptide-induced deformation in membrane structure indeed resulted in tubular formation resembling the endosomal uptake in the absence of ATP (macropinocytosis). Therefore, the authors termed this energy-independent pathway “physical endocytosis” (Maniti et al. 2012). Direct translocation of cell-penetrating peptides would have the advantage to deliver the cargo right into the cytosolic compartment of the cell, while in case of endocytosis the cargo-loaded peptides accumulate in endosomes, which can lead to degradation and inability to reach the cytosolic target, if an escape route is missing (Milletti 2012). Having this in mind, Wimley and co-workers developed chemical tools and orthogonal high-throughput screening methods for the discovery of peptides, which spontaneously translocate through model membranes (Marks et al. 2011). Only 18 peptides out of the 24,000 library members based on a rationally designed nine-residue combinatorial segment showed the desired properties: aqueous solubility, very high translocation rate, and no membrane leakage. This behavior was conirmed in living CHO cells. The spontaneous translocation through the bilayer membrane was explained by their “interfacial activity model” established previously for AMPs (Rathinakumar and Wimley 2010, Hristova and Wimley 2011, Wimley and Hristova 2011). Again, the interaction of the cationic arginine side chain with the anionic lipid phosphate group plays a special role and it was suggested that the phosphates may act as chaperones for the amino acids enabling a deeper penetration into the membrane and eventually translocation (Hessa et al. 2005, Hristova and Wimley 2011, Schow et al. 2011).

Membrane Medicine

237

12.5 CONCLUSIONS AND FUTURE DIRECTIONS We propose in extension to the concept of membrane-lipid therapy (Escriba 2006, Escriba et al. 2008) two strategies that either interfere with the collective behavior of membranes or interact with speciic membrane lipid components. The concept of membrane–lipid therapy is based on the observation that the activity of membrane proteins can be regulated by lipids, that is, lipid treatments could modulate the activities of proteins and/or associated signaling mechanisms. This approach has been used to develop molecules with low toxicity and high eficacy, to treat cancer, cardiovascular diseases, obesity, etc. (Escriba et al. 2008). Compounds that act on the collective behavior of membranes may represent a more general approach of membrane–lipid therapy as we show through several examples in this chapter. In this respect, membrane-active compounds may exert their activity upon inluencing the lateral pressure proile, which results in a physically different microenvironment for membrane proteins. Furthermore, membrane-active compounds may also interfere with the organization of membrane rafts. Modulating these domains represents another way to control membrane function. Thus, in general, membrane medicine opens new windows for the design of membrane-active drugs for future application in diverse diseases.

REFERENCES Alberts, B., Johnson, A., Lewis, J., Raff, M., Roberts, K., and Walter, P. 2007. Molecular Biology of the Cell (New York: Garland Science, Taylor & Francis Group New York). Almeida, P.F. and Pokorny, A. 2009. Mechanisms of antimicrobial, cytolytic, and cell-penetrating peptides: From kinetics to thermodynamics. Biochemistry 48, 8083–8093. Arouri, A. et al. 2009. Peptide induced demixing in PG/PE lipid mixtures: A mechanism for the speciicity of antimicrobial peptides towards bacterial membranes? Biochim. Biophys Acta 1788, 650–659. Ashley, R.H. et al. 2006. Autoinsertion of soluble oligomers of Alzheimer’s Abeta(1-42) peptide into cholesterol-containing membranes is accompanied by relocation of the sterol towards the bilayer surface. BMC. Struct. Biol. 6, 21. Ashrafuzzaman, M. and Tuszynski, J. 2013. Membrane biophysics. Membrane-related diseases. In: Ashrafuzzaman, M. and Tuszynski, J. (eds), 151–170. Biological and Medical Physics, Biomedical Engineering (Berlin-Heidelberg: Springer). Bafna, S. et al. 2010. Membrane-bound mucins: The mechanistic basis for alterations in the growth and survival of cancer cells. Oncogene 29, 2893–2904. Bechinger, B. 2009. Rationalizing the membrane interactions of cationic amphipathic antimicrobial peptides by their molecular shape. Curr. Opin. Colloid Interf. Sci. 14, 349–355. Bechinger, B. and Lohner, K. 2006. Detergent-like actions of linear amphipathic cationic antimicrobial peptides. Biochim. Biophys. Acta 1758, 1529–1539. Bloom, M. et al. 1991. Physical properties of the luid lipid-bilayer component of cell membranes: A perspective 1. Q. Rev. Biophys. 24, 293–397. Bochkov, V.N. 2007. Inlammatory proile of oxidized phospholipids. Thromb. Haemost. 97, 348–354. Boulgaropoulos, B. et al. 2010. Implication of sphingomyelin/ceramide molar ratio on the biological activity of sphingomyelinase. Biophys. J. 99, 499–506. Boulgaropoulos, B. et al. 2011. Stable and unstable lipid domains in ceramide containing membranes. Biophys. J. 100, 2160–2168. Boulgaropoulos, B. et  al. 2012. Lipid sorting by ceramide and the consequences for membrane proteins. Biophys. J. 102, 2031–2038. Burke, K.A. et al. 2013. Biophysical insights into how surfaces, including lipid membranes, modulate protein aggregation related to neurodegeneration. Front Neurol. 4, 17. Cairns, R.A. et al. 2011. Regulation of cancer cell metabolism. Nat. Rev. Cancer 11, 85–95. Campanella, R. 1992. Membrane lipids modiications in human gliomas of different degree of malignancy. J. Neurosurg. Sci. 36, 11–25. Cantor, R.S. 1997. The lateral pressure proile in membranes: A physical mechanism of general anesthesia. Biochemistry 36, 2339–2344. Capone, R. et al. 2012. All-d-Enantiomer of beta-amyloid peptide forms ion channels in lipid bilayers. J. Chem. Theory. Comput. 8, 1143–1152.

238

Liposomes, Lipid Bilayers and Model Membranes

Cassidy, J. and Misset, J.L. 2002. Oxaliplatin-related side effects: Characteristics and management. Semin. Oncol. 29, 11–20. Chen, Y. and Hughes-Fulford, M. 2001. Human prostate cancer cells lack feedback regulation of low-density lipoprotein receptor and its regulator, SREBP2. Int. J. Cancer 91, 41–45. Chiantia, S. and London, E. 2012. Acyl chain length and saturation modulate interlealet coupling in asymmetric bilayers: Effects on dynamics and structural order. Biophys. J. 103, 2311–2319. Coban, O. et al. 2007. Transition from nanodomains to microdomains induced by exposure of lipid monolayers to air. Biophys. J. 92, 2842–2853. Code, C. et  al. 2010. Activation of phospholipase A2 by 1-palmitoyl-2-(9′-oxo-nonanoyl)-sn-glycero-3phosphocholine in vitro. Biochim. Biophys. Acta 1798, 1593–1600. Connelly, L. et al. 2012. Atomic force microscopy and MD simulations reveal pore-like structures of all-denantiomer of Alzheimer’s beta-amyloid peptide: Relevance to the ion channel mechanism of AD pathology. J Phys. Chem. B 116, 1728–1735. Cortes, V.A. et al. 2013. Advances in the physiological and pathological implications of cholesterol. Biol. Rev. Camb. Philos. Soc. 88, 825–843. Dan, N. et al. 1993. Membrane-induced interactions between inclusions. Langmuir 9, 2768–2771. Dan, N. and Safran, S.A. 1998. Effect of lipid characteristics on the structure of transmembrane proteins. Biophys. J. 75, 1410–1414. De Felice, F.G. et al. 2008. Alzheimer’s disease-type neuronal tau hyperphosphorylation induced by A beta oligomers. Neurobiol. Aging 29, 1334–1347. Dennis, E.A. et al. 2010. A mouse macrophage lipidome. J. Biol. Chem. 285, 39976–39985. Derossi, D. et  al. 1994. The third helix of the Antennapedia homeodomain translocates through biological membranes. J Biol. Chem. 269, 10444–10450. Duchardt, F. et al. 2007. A comprehensive model for the cellular uptake of cationic cell-penetrating peptides. Trafic. 8, 848–866. Eckenhoff, R.G. 2008. Why can all of biology be anesthetized? Anesth. Analg. 107, 859–861. Eger, E.I. et  al. 2008. Is a new paradigm needed to explain how inhaled anesthetics produce immobility? Anesth. Analg. 107, 832–848. El Jastimi, R. and Laleur, M. 1999. Nisin promotes the formation of non-lamellar inverted phases in unsaturated phosphatidylethanolamines. Biochim. Biophys. Acta 1418, 97–105. Epand, R.F. and Epand, R.M. 2009a. Lipid domains in bacterial membranes and the action of antimicrobial agents. Biochim. Biophys. Acta 1788, 289–294. Epand, R.M. and Epand, R.F. 2009c. Domains in bacterial membranes and the action of antimicrobial agents. Mol. Biosyst. 5, 580–587. Epand, R.M. and Epand, R.F. 2011. Bacterial membrane lipids in the action of antimicrobial agents. J Pept. Sci. 17, 298–305. Epand, R.F. et al. 2003. Interactions of the antimicrobial beta-peptide beta-17 with phospholipid vesicles differ from membrane interactions of magainins. Eur. J. Biochem. 270, 1240–1248. Epand, R.F. et al. 2009b. Anti-inlammatory peptides grab on to the whiskers of atherogenic oxidized lipids. Biochim. Biophys. Acta 1788, 1967–1975. Epand, R.F. et al. 2010. Probing the “charge cluster mechanism” in amphipathic helical cationic antimicrobial peptides. Biochemistry 49, 4076–4084. Escriba, P.V. 2006. Membrane-lipid therapy: A new approach in molecular medicine. Trends Mol. Med. 12, 34–43. Escriba, P.V. et al. 2008. Membranes: A meeting point for lipids, proteins and therapies. J. Cell Mol. Med. 12, 829–875. Fadnes, B. et al. 2009. The anticancer activity of lytic peptides is inhibited by heparan sulfate on the surface of the tumor cells. BMC. Cancer 9, 183. Fadnes, B. et al. 2011. Small lytic peptides escape the inhibitory effect of heparan sulfate on the surface of cancer cells. BMC Cancer 11, 116. Fawell, S. et al. 1994. Tat-mediated delivery of heterologous proteins into cells. Proc. Natl. Acad. Sci. U. S. A 91, 664–668. Findlay, B. et al. 2010. Cationic amphiphiles, a new generation of antimicrobials inspired by the natural antimicrobial peptide scaffold. Antimicrob. Agents Chemother. 54, 4049–4058. Frankel, A.D. and Pabo, C.O. 1988. Cellular uptake of the tat protein from human immunodeiciency virus. Cell 55, 1189–1193. Franks, N.P. 2006. Molecular targets underlying general anaesthesia. Br. J. Pharmacol. 147 Suppl 1, S72–S81.

Membrane Medicine

239

Froud, R.J. et al. 1986. Effects of lipids and long-chain alkyl derivatives on the activity of (Ca2+-Mg2+)-ATPase. Biochemistry 25, 7544–7552. Fruhwirth, G.O. et al. 2007. Oxidized phospholipids: From molecular properties to disease. Biochim. Biophys. Acta 1772, 718–736. Fukunaga, S. et  al. 2012. GM1 cluster mediates formation of toxic Abeta ibrils by providing hydrophobic environments. Biochemistry 51, 8125–8131. Garidel, P. and Blume, A. 2000. Miscibility of phosphatidylethanolamine-phosphatidylglycerol mixtures as a function of pH and acyl chain length. Eur. Biophys. J. 28, 629–638. Gillet, J.P. and Gottesman, M.M. 2010. Mechanisms of multidrug resistance in cancer. Methods Mol. Biol. 596, 47–76. Goldberg, D.M. and Riordan, J.R. 1986. Role of membranes in disease. Clin. Physiol. Biochem. 4, 305–336. Goni, F.M. and Alonso, A. 2002. Sphingomyelinases: Enzymology and membrane activity. FEBS Lett. 531, 38–46. Goni, F.M. and Alonso, A. 2006. Biophysics of sphingolipids I. Membrane properties of sphingosine, ceramides and other simple sphingolipids. Biochim. Biophys. Acta 1758, 1902–1921. Green, M. and Loewenstein, P.M. 1988. Autonomous functional domains of chemically synthesized human immunodeiciency virus tat trans-activator protein. Cell 55, 1179–1188. Greenberg, M.E. et al. 2008. The lipid whisker model of the structure of oxidized cell membranes. J. Biol. Chem. 283, 2385–2396. Gregory, S.M. et al. 2008. A quantitative model for the all-or-none permeabilization of phospholipid vesicles by the antimicrobial peptide cecropin A. Biophys. J. 94, 1667–1680. Gruner, S.M. 1985. Intrinsic curvature hypothesis for biomembrane lipid composition: A role for nonbilayer lipids. Proc. Natl. Acad. Sci. U. S. A 82, 3665–3669. Hager, M.H. et  al. 2006. The role of cholesterol in prostate cancer. Curr. Opin. Clin. Nutr. Metab Care 9, 379–385. Hancock, R.E. 2001. Cationic peptides: Effectors in innate immunity and novel antimicrobials. Lancet Infect. Dis. 1, 156–164. Hannun, Y.A. and Obeid, L.M. 2008. Principles of bioactive lipid signalling: Lessons from sphingolipids. Nat. Rev. Mol. Cell Biol. 9, 139–150. Heimburg, T. and Jackson, A.D. 2005. On soliton propagation in biomembranes and nerves. Proc. Natl. Acad. Sci. U.S.A. 102, 9790–9795. Heimburg, T. and Jackson, A.D. 2007. The thermodynamics of general anesthesia. Biophys. J. 92, 3159–3165. Heitz, F. et al. 2009. Twenty years of cell-penetrating peptides: From molecular mechanisms to therapeutics. Br. J. Pharmacol. 157, 195–206. Hendrickx, J.F. et al. 2008. Is synergy the rule? A review of anesthetic interactions producing hypnosis and immobility. Anesth. Analg. 107, 494–506. Henriques, S.T. et  al. 2006. Cell-penetrating peptides and antimicrobial peptides: How different are they? Biochem. J. 399, 1–7. Hentosh, P. et al. 2001. Sterol-independent regulation of 3-hydroxy-3-methylglutaryl coenzyme A reductase in tumor cells. Mol. Carcinog. 32, 154–166. Hessa, T., White, S.H., and von Heijne, G. 2005. Membrane insertion of a potassium-channel voltage sensor. Science 307, 1427. Hickel, A. et al. 2008. Inluence of antimicrobial peptides on the formation of nonlamellar lipid mesophases. Biochimica et Biophysica Acta—Biomembranes 1778, 2325–2333. Hilvo, M. et al. 2011. Novel theranostic opportunities offered by characterization of altered membrane lipid metabolism in breast cancer progression. Cancer Res. 71, 3236–3245. Hirsch, H.A. et al. 2010. A transcriptional signature and common gene networks link cancer with lipid metabolism and diverse human diseases. Cancer Cell 17, 348–361. Holopainen, J.M. et al. 2000. Vectorial budding of vesicles by asymmetrical enzymatic formation of ceramide in giant liposomes. Biophys. J. 78, 830–838. Hoskin, D.W. and Ramamoorthy, A. 2008. Studies on anticancer activities of antimicrobial peptides. Biochim. Biophys. Acta 1778, 357–375. Hristova, K. et al. 1997. Critical role of lipid composition in membrane permeabilization by rabbit neutrophil defensins. J. Biol. Chem. 272, 24224–24233. Huang, H.W. 1986. Deformation free energy of bilayer membrane and its effect on gramicidin channel lifetime. Biophys J 50, 1061–1070. Huang, H.W. 2000. Action of antimicrobial peptides: Two-state model. Biochemistry 39, 8347–8352.

240

Liposomes, Lipid Bilayers and Model Membranes

Huang, H.W. 2006. Molecular mechanism of antimicrobial peptides: The origin of cooperativity. Biochim. Biophys. Acta 1758, 1292–1302. Ikeda, K. et al. 2011. Mechanism of amyloid beta-protein aggregation mediated by GM1 ganglioside clusters. Biochemistry 50, 6433–6440. Ipsen, J.H. et al. 1987. Phase equilibria in the phosphatidylcholine-cholesterol system. Biochim. Biophys. Acta 905, 162–172. Jacob, R.F. and Mason, R.P. 2005. Lipid peroxidation induces cholesterol domain formation in model membranes. J. Biol. Chem. 280, 39380–39387. Jerabek, H. et al. 2010. Membrane-mediated effect on ion channels induced by the anesthetic drug ketamine. J. Am. Chem. Soc. 132, 7990–7997. Johannsson, A. et al. 1981a. The effect of bilayer thickness and n-alkanes on the activity of the (Ca2+ + Mg2+)dependent ATPase of sarcoplasmic reticulum. J. Biol. Chem. 256, 1643–1650. Johannsson, A. et al. 1981b. The effect of bilayer thickness on the activity of (Na+ + K+)-ATPase. Biochim. Biophys. Acta 641, 416–421. Jones, A.T. and Sayers, E.J. 2012. Cell entry of cell penetrating peptides: Tales of tails wagging dogs. J. Control Release 161, 582–591. Kadl, A. et al. 2004. Apoptotic cells as sources for biologically active oxidized phospholipids. Antioxid. Redox. Signal. 6, 311–320. Kalyanaraman, B. et  al. 2002. Doxorubicin-induced apoptosis: Implications in cardiotoxicity. Mol. Cell Biochem. 234–235, 119–124. Keller, A.A. et al. 2013. Relationship between cargo, cell penetrating peptides and cell type for uptake of noncovalent complexes into live cells. Pharmaceuticals 6, 184–203. Keller, S.L. et al. 1996. Small concentrations of alamethicin induce a cubic phase in bulk phosphatidylethanolamine mixtures. Biochim. Biophys. Acta 1278, 241–246. Khandelia, H. and Mouritsen, O.G. 2009. Lipid gymnastics: Evidence of complete acyl chain reversal in oxidized phospholipids from molecular simulations. Biophys. J. 96, 2734–2743. Kiessling, V. et al. 2009. Domain coupling in asymmetric lipid bilayers. Biochim. Biophys. Acta 1788, 64–71. Killian, J.A. 2003. Synthetic peptides as models for intrinsic membrane proteins. FEBS Lett. 555, 134–138. Kim, W. and Hecht, M.H. 2005. Sequence determinants of enhanced amyloidogenicity of Alzheimer A{beta}42 peptide relative to A{beta}40. J. Biol. Chem. 280, 35069–35076. Kinnunen, P.K. et al. 2012. Protein-oxidized phospholipid interactions in cellular signaling for cell death: From biophysics to clinical correlations. Biochim. Biophys. Acta 1818, 2446–2455. Kosuge, M. et  al. 2008. Cellular internalization and distribution of arginine-rich peptides as a function of extracellular peptide concentration, serum, and plasma membrane associated proteoglycans. Bioconjug. Chem. 19, 656–664. Kufe, D.W. 2009. Mucins in cancer: Function, prognosis and therapy. Nat. Rev. Cancer 9, 874–885. Kuhajda, F.P. et al. 1994. Fatty acid synthesis: A potential selective target for antineoplastic therapy. Proc. Natl. Acad. Sci. U. S. A. 91, 6379–6383. Latal, A. et  al. 1997. Structural aspects of the interaction of peptidyl-glycylleucine-carboxyamide, a highly potent antimicrobial peptide from frog skin, with lipids. Eur. J. Biochem. 248, 938–946. Lee, A.G. 2004. How lipids affect the activities of integral membrane proteins. Biochim. Biophys. Acta 1666, 62–87. Lee, A.G. 2005. How lipids and proteins interact in a membrane: A molecular approach. Mol. Biosyst. 1, 203–212. Leuschner, C. and Hansel, W. 2004. Membrane disrupting lytic peptides for cancer treatments. Curr. Pharm. Des. 10, 2299–2310. Levade, T. and Jaffrezou, J.P. 1999. Signalling sphingomyelinases: Which, where, how and why? Biochim. Biophys. Acta 1438, 1–17. Li, Y.C. et al. 2006. Elevated levels of cholesterol-rich lipid rafts in cancer cells are correlated with apoptosis sensitivity induced by cholesterol-depleting agents. Am. J. Pathol. 168, 1107–1118. Lin, M.S. et al. 2008. Examining the levels of ganglioside and cholesterol in cell membrane on attenuation the cytotoxicity of beta-amyloid peptide. Colloids Surf. B Biointerf. 65, 172–177. Lingwood, D. and Simons, K. 2010. Lipid rafts as a membrane-organizing principle. Science 327, 46–50. Lipinski, C.A. et al. 2001. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv. Drug Deliv. Rev. 46, 3–26. Lohner, K. 2001. The role of membrane lipid composition in cell targeting and their mechanism of action. In: Lohner, K. (ed), Development of Novel Antimicrobial Agents: Emerging Strategies, (Wymondham: Horizon Scientiic Press), pp. 149–165. Lohner, K. 2009. New strategies for novel antibiotics: Peptides targeting bacterial cell membranes. General Physio. Biophys. 28, 105–116.

Membrane Medicine

241

Lohner, K. and Blondelle, S.E. 2005. Molecular mechanisms of membrane perturbation by antimicrobial peptides and the use of biophysical studies in the design of novel peptide antibiotics. Comb. Chem. High Throughput. Screen. 8, 241–256. Lohner, K. and Prenner, E.J. 1999. Differential scanning calorimetry and X-ray diffraction studies of the speciicity of the interaction of antimicrobial peptides with membrane-mimetic systems. Biochim. Biophys. Acta 1462, 141–156. Lohner, K. et al. 1997. Differential scanning microcalorimetry indicates that human defensin, HNP-2, interacts speciically with biomembrane mimetic systems. Biochemistry 36, 1525–1531. Lohner, K. et al. 2001. Packing characteristics of a model system mimicking cytoplasmic bacterial membranes. Chem. Phys. Lipids 111, 177–192. Lohner, K., Sevcsik, E., and Pabst, G. 2008. Liposome-based biomembrane mimetic systems: Implications for lipid-peptide interactions. In: Leitmannova-Liu, A. (ed), Advances in Planar Lipid Bilayers and Liposomes (Vol. 6) (Amsterdam: Elsevier), pp. 103–137. Ludtke, S.J. et al. 1996. Membrane pores induced by magainin. Biochemistry 35, 13723–13728. Madani, F. et al. 2011. Mechanisms of cellular uptake of cell-penetrating peptides. J. Biophys. 2011, 414729. Mader, J.S. et al. 2005. Bovine lactoferricin selectively induces apoptosis in human leukemia and carcinoma cell lines. Mol. Cancer Ther. 4, 612–624. Maniti, O. et  al. 2012. Metabolic energy-independent mechanism of internalization for the cell penetrating peptide penetratin. Int. J. Biochem. Cell Biol. 44, 869–875. Marks, J.R. et al. 2011. Spontaneous membrane-translocating peptides by orthogonal high-throughput screening. J. Am. Chem. Soc. 133, 8995–9004. Marsh, D. 2007. Lateral pressure proile, spontaneous curvature frustration, and the incorporation and conformation of proteins in membranes. Biophys. J. 93, 3884–3899. Marsh, D. 2008a. Energetics of hydrophobic matching in lipid-protein interactions. Biophys. J. 94, 3996–4013. Marsh, D. 2008b. Protein modulation of lipids, and vice-versa, in membranes. Biochim. Biophys. Acta 1778, 1545–1575. Marsh, D. 2009. Cholesterol-induced luid membrane domains: A compendium of lipid-raft ternary phase diagrams. Biochim. Biophys. Acta 1788, 2114–2123. Marsh, D. 2010. Liquid-ordered phases induced by cholesterol: A compendium of binary phase diagrams. Biochim. Biophys. Acta 1798, 688–699. Mashour, G.A. et al. 2005. Mechanisms of general anesthesia: From molecules to mind. Best. Pract. Res. Clin. Anaesthesiol. 19, 349–364. Mason, R.P. et al. 1997. Effect of oxidative stress on membrane structure: Small-angle X-ray diffraction analysis. Free Radic. Biol. Med. 23, 419–425. Mata, A.M. et al. 2011. Impairment of the activity of the plasma membrane Ca(2)(+)-ATPase in Alzheimer’s disease. Biochem. Soc. Trans. 39, 819–822. Matsumoto, K. et al. 2006. Lipid domains in bacterial membranes. Mol. Microbiol. 61, 1110–1117. Matsuzaki, K. et al. 1996. An antimicrobal peptide, magainin 2, induced rapid lip-lop of phospholipids coupled with pore formation and peptide translocation. Biochemistry 35, 11361–11368. Mattila, J.P. et  al. 2008. Oxidized phospholipids as potential molecular targets for antimicrobial peptides. Biochim. Biophys. Acta 1778, 2041–2050. May, G.L. et al. 1988. Plasma membrane lipid composition of vinblastine sensitive and resistant human leukaemic lymphoblasts. Int. J. Cancer 42, 728–733. Mazzoni, A. and Trave, F. 1993. Cytoplasmic membrane cholesterol and doxorubicin cytotoxicity in drugsensitive and multidrug-resistant human ovarian cancer cells. Oncol. Res. 5, 75–82. McElhaney, R.N. 1992. Membrane structure. In: Maniloff, J., McElhaney, R.N., Finch, L.R., and Baseman, J.B. (eds), Mycoplasma: Molecular Biology and Pathogenesis (Washington DC: American Society for Microbiology), pp. 113–155. Mileykovskaya, E. and Dowhan, W. 2005. Role of membrane lipids in bacterial division-site selection. Curr. Opin. Microbiol. 8, 135–142. Milletti, F. 2012. Cell-penetrating peptides: Classes, origin, and current landscape. Drug Discov. Today 17, 850–860. Mishra, A. et al. 2008. HIV TAT forms pores in membranes by inducing saddle-splay curvature: Potential role of bidentate hydrogen bonding. Angew. Chem. Int. Ed Engl. 47, 2986–2989. Mishra, A. et al. 2011. Translocation of HIV TAT peptide and analogues induced by multiplexed membrane and cytoskeletal interactions. Proc. Natl. Acad. Sci. U. S. A. 108, 16883–16888. Miyashita, T. and Reed, J.C. 1993. Bcl-2 oncoprotein blocks chemotherapy-induced apoptosis in a human leukemia cell line. Blood 81, 151–157.

242

Liposomes, Lipid Bilayers and Model Membranes

Montecucco, C. et al. 1982. Bilayer thickness and enzymatic activity in the mitochondrial cytochrome c oxidase and ATPase complex. FEBS Lett. 144, 145–148. Morein, S. et al. 1996. Wild-type Escherichia coli cells regulate the membrane lipid composition in a “window” between gel and non-lamellar structures. J. Biol. Chem. 271, 6801–6809. Naeem, A. and Fazili, N.A. 2011. Defective protein folding and aggregation as the basis of neurodegenerative diseases: The darker aspect of proteins. Cell Biochem. Biophys. 61, 237–250. Nakazawa, I. and Iwaizumi, M. 1989. A role of the cancer cell membrane luidity in the cancer metastases: An ESR study. Tohoku J. Exp. Med. 157, 193–198. Nielsen, C. and Andersen, O.S. 2000. Inclusion-induced bilayer deformations: Effects of monolayer equilibrium curvature. Biophys. J. 79, 2583–2604. Niemela, P.S. et al. 2010. Membrane proteins diffuse as dynamic complexes with lipids. J. Am. Chem. Soc. 132, 7574–7575. Nomura, D.K. et al. 2011. Monoacylglycerol lipase exerts dual control over endocannabinoid and fatty acid pathways to support prostate cancer. Chem. Biol. 18, 846–856. Nordberg, P., Monnet, D. L., and Cars, O. 2005. Antibacterial drug resistance: Options for concerned action. WHO, Department of Medicines Policy and Standards. http://archives.who.int/prioritymeds/report/ background/antibacterial.doc. Norris, V. and Fishov, I. 2001. Hypothesis: Membrane domains and hyperstructures control bacterial division. Biochimie 83, 91–97. Nury, H. et al. 2011. X-ray structures of general anaesthetics bound to a pentameric ligand-gated ion channel. Nature 469, 428–431. Ogawa, M. et  al. 2011. Ganglioside-mediated aggregation of amyloid beta-proteins (Abeta): Comparison between Abeta-(1-42) and Abeta-(1-40). J. Neurochem. 116, 851–857. Okada, T. et al. 2008. Formation of toxic Abeta(1-40) ibrils on GM1 ganglioside-containing membranes mimicking lipid rafts: Polymorphisms in Abeta(1-40) ibrils. J. Mol. Biol. 382, 1066–1074. Oresic, M. et al. 2008. Lipidomics: A new window to biomedical frontiers. Trends Biotechnol. 26, 647–652. Pabst, G. 2013. Coupling of membrane elasticity and structure to protein function. In: Iglic, A. and Kulkarni, C.V. (eds), Advances in Planar Lipid Bilayers and Liposomes (Vol. 18),(Academic Press: Burlington), pp. 81–109. Pabst, G. et  al. 2007. Entropy-driven softening of luid lipid bilayers by alamethicin. Langmuir 23, 11705–11711. Pabst, G. et al. 2009. Effect of ceramide on nonraft proteins. J. Membr. Biol. 231, 125–132. Pan, J. et  al. 2009. Alamethicin in lipid bilayers: Combined use of X-ray scattering and MD simulations. Biochim. Biophys. Acta 1788, 1387–1397. Panchal, M. et  al. 2010. Enrichment of cholesterol in microdissected Alzheimer’s disease senile plaques as assessed by mass spectrometry. J. Lipid Res. 51, 598–605. Pande, A.H. et al. 2010. Oxidatively modiied fatty acyl chain determines physicochemical properties of aggregates of oxidized phospholipids. Biochim. Biophys. Acta 1798, 442–452. Papo, N. and Shai, Y. 2005. Host defense peptides as new weapons in cancer treatment. Cell Mol. Life Sci. 62, 784–790. Papo, N. et al. 2006. Inhibition of tumor growth and elimination of multiple metastases in human prostate and breast xenografts by systemic inoculation of a host defense-like lytic peptide 1618. Cancer Res. 66, 5371–5378. Perez-Tomas, R. 2006. Multidrug resistance: Retrospect and prospects in anti-cancer drug treatment. Curr. Med. Chem. 13, 1859–1876. Perozo, E. et al. 2002. Physical principles underlying the transduction of bilayer deformation forces during mechanosensitive channel gating. Nat. Struct. Biol. 9, 696–703. Petit-Zeman, S. 2004. Membranes and the Diseases within. http://www.nature.com/horizon/livingfrontier/ background/diseases.html. Pike, L.J. 2006. Rafts deined: A report on the keystone symposium on lipid rafts and cell function. J. Lipid Res. 47, 1597–1598. Podrez, E.A. et al. 2002. Identiication of a novel family of oxidized phospholipids that serve as ligands for the macrophage scavenger receptor CD36. J. Biol. Chem. 277, 38503–38516. Pokorny, A. and Almeida, P.F. 2004. Kinetics of dye eflux and lipid lip-lop induced by delta-lysin in phosphatidylcholine vesicles and the mechanism of graded release by amphipathic, alpha-helical peptides. Biochemistry 43, 8846–8857. Powl, A.M. et al. 2007. Different effects of lipid chain length on the two sides of a membrane and the lipid annulus of MscL. Biophys. J. 93, 113–122. Pozo, N.B. et al. 2005. Composition dependence of vesicle morphology and mixing properties in a bacterial model membrane system. Biochim. Biophys. Acta 1716, 40–48.

Membrane Medicine

243

Raagel, H. et al. 2010. Peptide-mediated protein delivery-which pathways are penetrable? Biochim. Biophys. Acta 1798, 2240–2248. Rappolt, M. et al. 2003. Mechanism of the lamellar/inverse hexagonal phase transition examined by high resolution x-ray diffraction. Biophys. J. 84, 3111–3122. Rathinakumar, R. and Wimley, W.C. 2010. High-throughput discovery of broad-spectrum peptide antibiotics. FASEB J. 24, 3232–3238. Riedl, S. et al. 2011a. In search of a novel target—Phosphatidylserine exposed by non-apoptotic tumor cells and metastases of malignancies with poor treatment eficacy. Biochim. Biophys. Acta 1808, 2638–2645. Riedl, S. et al. 2011b. Membrane active host defense peptides challenges and perspectives for the development of novel anticancer drugs. Chem. Phys. Lipids 164, 766–781. Rilfors, L. et al. 1993. Regulation and physicochemical properties of the polar lipids in Acholeplasma laidlawii. Subcell. Biochem. 20, 109–166. Risso, A. et al. 1998. Cytotoxicity and apoptosis mediated by two peptides of innate immunity. Cell Immunol. 189, 107–115. Rodrigues, C.M. et al. 2002. Perturbation of membrane dynamics in nerve cells as an early event during bilirubin-induced apoptosis. J. Lipid Res. 43, 885–894. Rothbard, J.B. et al. 2002. Arginine-rich molecular transporters for drug delivery: Role of backbone spacing in cellular uptake. J. Med. Chem. 45, 3612–3618. Rothbard, J.B. et al. 2004. Role of membrane potential and hydrogen bonding in the mechanism of translocation of guanidinium-rich peptides into cells. J. Am. Chem. Soc. 126, 9506–9507. Ruiz-Arguello, M.B. et al. 1996. Different effects of enzyme-generated ceramides and diacylglycerols in phospholipid membrane fusion and leakage. J. Biol. Chem. 271, 26616–26621. Ruzza, A. et al. 2013. Cell-penetrating peptides: A comparative study on lipid afinity and cargo delivery properties. Pharmaceuticals 3, 1045–1062. Rysman, E. et al. 2010. De novo lipogenesis protects cancer cells from free radicals and chemotherapeutics by promoting membrane lipid saturation. Cancer Res. 70, 8117–8126. Sakai, N. et al. 2005. Direct observation of anion-mediated translocation of luorescent oligoarginine carriers into and across bulk liquid and anionic bilayer membranes. Chembiochem. 6, 114–122. Santos, C.R. and Schulze, A. 2012. Lipid metabolism in cancer. FEBS J. 279, 2610–2623. Schmidt, N. et al. 2010. Arginine-rich cell-penetrating peptides. FEBS Lett. 584, 1806–1813. Schmidt, N.W. et  al. 2012. Molecular basis for nanoscopic membrane curvature generation from quantum mechanical models and synthetic transporter sequences. J. Am. Chem. Soc. 134, 19207–19216. Schow, E.V., Freites, J.A., Cheng, P., Bernsel, A., von Heijne, G., White, S.H., and Tobias, D.J. 2011. Arginine in membranes: The connection between molecular dynamics simulations and translocon-mediated insertion experiments. J. Membr. Biol. 239, 35–48. Schweizer, F. 2009. Cationic amphiphilic peptides with cancer-selective toxicity. Eur. J. Pharmacol. 625, 190–194. Seddon, J.M. 1990. Structure of the inverted hexagonal (HII) phase, and non-lamellar phase transitions in lipids. Biochim. Biophys. Acta 1031, 1–69. Sevcsik, E. et  al. 2007. How lipids inluence the mode of action of membrane-active peptides. Biochim. Biophys. Acta 1768, 2568–2595. Sevcsik, E. et al. 2008. Interaction of LL-37 with model membrane systems of different complexity: Inluence of the lipid matrix. Biophys. J. 94, 4688–4699. Shah, J. et al. 1995. Structural and thermotropic properties of synthetic C16:0 (palmitoyl) ceramide: Effect of hydration. J. Lipid Res. 36, 1936–1944. Shai, Y. 2002. Mode of action of membrane active antimicrobial peptides. Biopolymers 66, 236–248. Sherbet, G.V. 1989. Membrane luidity and cancer metastasis. Exp. Cell Biol. 57, 198–205. Sherbet, G.V. and Jackson, S. 1986. Temperature-dependent surface charge modulation, membrane luidity, and metastatic ability of L51178Y lymphoma. Anticancer Res. 6, 129–134. Smith, H.L. et al. 2009. Early stages of oxidative stress-induced membrane permeabilization: A neutron relectometry study. J. Am. Chem. Soc. 131, 3631–3638. Soengas, M.S. et al. 2001. Inactivation of the apoptosis effector Apaf-1 in malignant melanoma. Nature 409, 207–211. Sok, M. et al. 2002. Cell membrane luidity and prognosis of lung cancer. Ann. Thorac. Surg. 73, 1567–1571. Splith, K. and Neundorf, I. 2011. Antimicrobial peptides with cell-penetrating peptide properties and vice versa. Eur. Biophys. J. 40, 387–397. Staudegger, E. et  al. 2000. X-ray studies on the interaction of the antimicrobial peptide gramicidin S with microbial lipid extracts: Evidence for cubic phase formation. Biochim. Biophys. Acta 1468, 213–230.

244

Liposomes, Lipid Bilayers and Model Membranes

Stemmer, U. and Hermetter, A. 2012a. Protein modiication by aldehydophospholipids and its functional consequences. Biochim. Biophys. Acta 1818, 2436–2445. Stemmer, U. et al. 2012b. Uptake and protein targeting of luorescent oxidized phospholipids in cultured RAW 264.7 macrophages. Biochim. Biophys. Acta 1821, 706–718. Stottrup, B.L. et  al. 2004. Nonequilibrium behavior in supported lipid membranes containing cholesterol. Biophys. J. 86, 2942–2950. Subbanagounder, G. et al. 2000. Bioactive products of phospholipid oxidation: Isolation, identiication, measurement and activities. Free Radic. Biol. Med. 28, 1751–1761. Subramaniam, S. et al. 2011. Bioinformatics and systems biology of the lipidome. Chem. Rev. 111, 6452–6490. Suzuki, T. et al. 2002. Possible existence of common internalization mechanisms among arginine-rich peptides. J. Biol. Chem. 277, 2437–2443. Swinnen, J.V. et al. 2006. Increased lipogenesis in cancer cells: New players, novel targets. Curr. Opin. Clin. Nutr. Metab. Care 9, 358–365. Trudell, J.R. and Bertaccini, E. 2002. Molecular modelling of speciic and non-speciic anaesthetic interactions. Br. J. Anaesth. 89, 32–40. Tunnemann, G. et al. 2008. Live-cell analysis of cell penetration ability and toxicity of oligo-arginines. J. Pept. Sci. 14, 469–476. Uchida, K. 2000. Role of reactive aldehyde in cardiovascular diseases. Free Radic. Biol. Med. 28, 1685–1696. Urban, B.W. 2002. Current assessment of targets and theories of anaesthesia. Br. J. Anaesth. 89, 167–183. Van Blitterswijk, W.J. et  al. 1984. Differences in membrane lipid composition and luidity of transplanted GRSL lymphoma cells, depending on their site of growth in the mouse. Biochim. Biophys. Acta 778, 521–529. van Blitterswijk, W.J. et  al. 2003. Ceramide: Second messenger or modulator of membrane structure and dynamics? Biochem. J. 369, 199–211. van der Wel, P.C. et al. 2000. Tryptophan-anchored transmembrane peptides promote formation of nonlamellar phases in phosphatidylethanolamine model membranes in a mismatch-dependent manner. Biochemistry 39, 3124–3133. van Meer, G. et al. 2008. Membrane lipids: Where they are and how they behave. Nat. Rev. Mol. Cell Biol. 9, 112–124. Vanounou, S. et  al. 2003. Phosphatidylethanolamine and phosphatidylglycerol are segregated into different domains in bacterial membrane. A study with pyrene-labelled phospholipids. Mol. Microbiol. 49, 1067–1079. Veatch, S.L. and Keller, S.L. 2005. Seeing spots: Complex phase behavior in simple membranes. Biochim. Biophys. Acta 1746, 172–185. Vernier, P.T. et al. 2009. Electroporating ields target oxidatively damaged areas in the cell membrane. PLoS. One. 4, e7966. Villagran-Vargas, E. et al. 2013. Variations in interpulse interval of double action potentials during propagation in single neurons. Synapse 67, 68–78. Vist, M.R. and Davis, J.H. 1990. Phase equilibria of cholesterol/dipalmitoylphosphatidylcholine mixtures: 2H nuclear magnetic resonance and differential scanning calorimetry. Biochemistry 29, 451–464. Volinsky, R. et al. 2011. Oxidized phosphatidylcholines facilitate phospholipid lip-lop in liposomes. Biophys. J. 101, 1376–1384. Volinsky, R. and Kinnunen, P.K. 2013. Oxidized phosphatidylcholines in membrane level cellular signaling from biophysics to physiology and molecular pathology. FEBS J. 280, 2806–2816. Volinsky, R. et al. 2012. Oxidized phosphatidylcholines promote phase separation of cholesterol-sphingomyelin domains. Biophys. J. 103, 247–254. Wakabayashi, M. and Matsuzaki, K. 2007. Formation of amyloids by Abeta-(1-42) on NGF-differentiated PC12 cells: Roles of gangliosides and cholesterol. J. Mol. Biol. 371, 924–933. Wang, G. et al. 2009. APD2: The updated antimicrobial peptide database and its application in peptide design. Nucleic Acids Res. 37, D933–D937. Wender, P.A. et al. 2008. The design of guanidinium-rich transporters and their internalization mechanisms. Adv. Drug Deliv. Rev. 60, 452–472. Wenk, M.R. 2010. Lipidomics: New tools and applications. Cell 143, 888–895. Williams, T.L. and Serpell, L.C. 2011. Membrane and surface interactions of Alzheimer’s Abeta peptide— Insights into the mechanism of cytotoxicity. FEBS J. 278, 3905–3917. Willumeit, R. et  al. 2005. Structural rearrangement of model membranes by the peptide antibiotic NK-2. Biochim. Biophys. Acta 1669, 125–134.

Membrane Medicine

245

Wimley, W.C. 2010. Describing the mechanism of antimicrobial peptide action with the interfacial activity model. ACS Chem. Biol. 5, 905–917. Wimley, W.C. and Hristova, K. 2011. Antimicrobial peptides: Successes, challenges and unanswered questions. J. Membr. Biol. 239, 27–34. Wong-Ekkabut, J. et  al. 2007. Effect of lipid peroxidation on the properties of lipid bilayers: A molecular dynamics study. Biophys. J. 93, 4225–4236. Yip, C.M. et al. 2001. Cholesterol, a modulator of membrane-associated Abeta-ibrillogenesis and neurotoxicity. J. Mol. Biol. 311, 723–734. Yu, X. and Zheng, J. 2012. Cholesterol promotes the interaction of Alzheimer beta-amyloid monomer with lipid bilayer. J. Mol. Biol. 421, 561–571. Zasloff, M. 2002. Antimicrobial peptides of multicellular organisms. Nature 415, 389–395. Zhang, F. and Du, G. 2012. Dysregulated lipid metabolism in cancer. World J. Biol. Chem. 3, 167–174. Ziegler, A. and Seelig, J. 2008. Binding and clustering of glycosaminoglycans: A common property of monoand multivalent cell-penetrating compounds. Biophys. J. 94, 2142–2149. Zimmerberg, J. and Kozlov, M.M. 2006. How proteins produce cellular membrane curvature. Nat. Rev. Mol. Cell Biol. 7, 9–19. Zorko, M. and Langel, U. 2005. Cell-penetrating peptides: Mechanism and kinetics of cargo delivery. Adv. Drug Deliv. Rev. 57, 529–545. Zwaal, R.F. and Schroit, A.J. 1997. Pathophysiologic implications of membrane phospholipid asymmetry in blood cells. Blood 89, 1121–1132. Zweytick, D. et  al. 2008. Membrane curvature stress and antibacterial activity of lactoferricin derivatives. Biochem. Biophys. Res. Commun. 369, 395–400. Zweytick, D. et al. 2011. Studies on lactoferricin-derived Escherichia coli membrane-active peptides reveal differences in the mechanism of N-acylated versus nonacylated peptides. J. Biol. Chem. 286, 21266–21276.

13

Structural Diversity of DNA– Phospholipid Aggregates Daniela Uhríková and Petra Pullmannová

CONTENTS 13.1 Introduction .......................................................................................................................... 247 13.2 Polymorphic Behavior of DNA–GS–Phospholipid Complexes ........................................... 250 13.2.1 GSs in Gene Delivery ............................................................................................... 250 13.2.2 Helper Lipids ............................................................................................................ 251 13.2.3 Lamellar-to-Hexagonal Phase Transitions in DNA–DOPE–CnGSm ...................... 253 13.2.4 Condensed Lamellar Phase in DNA–DOPE–CnGSm ............................................. 254 13.2.4.1 Effect of Surface Charge Density .............................................................. 254 13.2.4.2 Effect of the CnGSm Spacer on DNA Packing in LCα ................................ 255 13.2.5 Structural Variety in DNA–PC–CnGSm.................................................................. 257 13.3 Aggregates of DNA–Neutral Phospholipid–Metal Cation Systems ..................................... 258 13.3.1 Ca2+ in DNA–DPPC Interactions .............................................................................260 13.3.2 DNA–DPPC–Transition Metal Cations .................................................................... 262 13.4 Concluding Remarks ............................................................................................................ 263 Acknowledgment ...........................................................................................................................264 References ......................................................................................................................................264

13.1

INTRODUCTION

The main objective of gene therapy is a successful in vivo transfer of genetic materials to the targeted tissues. Virus-based gene carriers involve potential risks of immunogenicity, and chromosomal insertion of viral genome. In contrast, nonviral vectors such as lipoplexes are preferable for their many advantages, including lower toxicity, nonimmunogenicity, and nononcogenicity, greater nucleic acid packaging capacity, and easier and cheaper preparation. However, from the therapeutic point of view, a level of transfection suffers low eficiency when lipoplexes are used as gene delivery vectors. The establishment of a correlation between transfection eficiency and various characteristics of the lipoplexes requires thus in-depth research (Madhusudhana Rao and Gopal 2006, Li and Huang 2006, Zuhorn et al. 2007, Donkuru et al. 2010). In a brief description of cationic lipid-mediated gene delivery, DNA (deoxyribonucleic acid) appears as a linear polyanion at a physiologically relevant pH. Its interactions with cationic liposomes result spontaneously in DNA–cationic liposome aggregates called lipoplexes. Negatively charged phosphate fragments of DNA create the binding sites for formation of aggregates due to electrostatic interactions. Neutralization of more than 90% of the DNA’s negative charge by cationic liposomes results in a condensation of DNA, allowing thus a close approach of the DNA strands. In this packing, DNA is protected against degradation in plasma (Rolland 1998). The further requirement for an effective transfection vector is the neutral or positive surface charge of the aggregate to avoid its electrostatic repulsion by negatively charged cell membrane (Felgner et al. 1987). The aggregates of DNA–cationic liposomes have sizes in the range of hundreds of nanometers, and are 247

248

Liposomes, Lipid Bilayers and Model Membranes

taken up by the cells largely by endocytosis (Zabner et al. 1995). A structural design of the carrier is thus obviously a crucial step in ensuring the success of gene delivery. The following carriers of DNA can be self-assembled: 1. DNA–cationic liposomes (prepared by mixing neutral phospholipids and cationic amphiphiles). Felgner et al. (1987) have found an ability of cationic liposomes to transfer DNA into the cell with successful expression. These cationic liposomes were prepared from a hydrated mixture of N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium chloride (DOTMA) and dioleoylphosphatidylethanolamine (DOPE) (DOTMA:DOPE = 1:1 wt.) using ultrasound for homogenization. Amphiphilic molecules with a DOTMA-like structure are also called cationic lipids or cytofectins. With the aim to ind an “ideal” delivery vector for DNA, new cytofectins were synthesized too (Miller 1998, 2003, Madhusudhana Rao and Gopal, 2006, Donkuru et al. 2010, Shirazi et al. 2011), some of which are commercially available (see, e.g., Miller 2003). Recent molecular design research leads to the development of transfection molecules that are biodegradable, less toxic (Shirazi et al. 2011), and possess pH sensitivity (Byk et al. 2000, Wetzer et al. 2001, Chen et al. 2012); also many pharmaceutically relevant molecules (peptides, antibiotics, etc.) were tested as delivery vectors (Pachuk et al. 2000, Duzgunes et al. 2003, Bombelli et al. 2009, Rosada et al. 2012). 2. DNA–cationic surfactants. DNA interacts with micelles or vesicles of cationic surfactants while forming aggregates that show some characteristic symmetries in their packing (Ghirlando et al. 1992, Miguel et al. 2003). These supramolecular assemblies can serve as DNA delivery vectors. Gemini surfactants (GS) are the leader in this ield of transfection experiments. ENGEMS (European Network of Gemini Surfactants, 1997–2001) organization has tested 250 GS for transfection activity (in vitro) while most of them revealed high transfection eficiency (~90%) and low toxicity (Camilleri et  al. 2000, Bell et  al. 2003, Kirby et al. 2003). 3. DNA–neutral/anionic phospholipid liposomes–metal cations (Mg2+, Ca2+, etc.). Finally, divalent metal cations can serve as a mediator in the binding of DNA and liposomes prepared either from neutral or anionic phospholipids. The capability of these aggregates for DNA delivery has been demonstrated in several studies (Kovalenko et al. 1996, Zhdanov et al. 1997, Sato et al. 2005). The lipoplexes composed of nontoxic anionic lipids, plasmid DNA, and divalent Ca2+ bridges have shown high transfection eficiency comparable to those of cationic lipoplexes, but with signiicantly lower toxicity (Patil et al. 2004, Srinivasan and Burgess 2009). The aforementioned cationic vesicles interact with DNA polyanion electrostatically, forming aggregates. The interaction is endothermic and rapid (Barreleiro et al. 2000), and takes place during the aggregation process, while both phospholipid and DNA are undergoing a complete topological transformation into compact quasispherical particles with roughly 0.2 mm diameter. They can easily form string-like colloids with an ordered internal structure (Gershon et al. 1993, Koltover et al. 1999). For example, the following microstructures have been identiied: (i) the condensed lamellar phase LC (sandwich structure) with DNA strands ordered regularly between lipid bilayers (Figure 13.1a) (Lasic et al. 1997, Radler et al. 1997); (ii) the condensed-inverted hexagonal phase H IIC (honeycomb structure) with DNA molecules inserted in tubules formed by inverted lipid monolayers and arranged on a hexagonal lattice (Figure 13.1b) (Koltover et al. 1998); (iii) hexagonal phase HI formed by cylindrical micelles with DNA strands intercalated between hexagonally packed cylinders (Figure 13.1c) (Ewert et al. 2006). It should be mentioned that also (iv) cubic phases have been reported and tested as gene delivery vectors (Koynova and Tenchov 2009, Leal et al. 2010). Originally, Felgner (Felgner et al. 1987, Felgner and Rhodes 1991) proposed a “bead-on-string” structure of the cationic liposomes–DNA aggregates and pictured the DNA strand decorated with distinctly attached liposomes. Techniques of electron microscopy have identiied a variety

Structural Diversity of DNA–Phospholipid Aggregates (a)

249

d

dDNA

(b)

a

(c)

a

FIGURE 13.1 (See color insert.) Structural model of condensed lamellar phase LCα (a), condensed inverted hexagonal phase H IIC (b), and columnar hexagonal phase HI (c) with lattice parameters depicted in blue color.

of structures including string-like structures, and indicated also the fusion of liposomes (Gershon et al. 1993), oligolamellar structures (Gustafsson et al. 1995), and tube-like images possibly depicting lipid bilayer-covered DNA (the so-called spaghetti-like structure) (Sternberg et al. 1994). Radler et al. (1997) conducted optical microscopy and x-ray diffraction studies of lipoplexes formed due to DNA interactions with cationic liposomes, and have conirmed highly ordered multilamellar

250

Liposomes, Lipid Bilayers and Model Membranes

structures with DNA sandwiched between cationic bilayers. Theoretical models testing the stability of lipoplexes (Dan 1996, 1998, May and Ben Shaul 1997) have also supported the picture of the aforementioned structures possessing long-range order that is not only intermediated or metastable during the aggregation process. The packing and bending constraints of long DNA molecules, charge, and mass ratio of DNA-to-lipid may play roles there. Most importantly, aggregates showing some degree of such symmetry (in one-, two-, or three-dimensions) in their packing (the so-called long-range order) have been reported as effective in DNA transfection. In this chapter, we review the structural diversity of lipoplexes detected by small-angle x-ray diffraction (SAXD). We pay a special attention to two groups of gene delivery vectors: (a) lipoplexes formed with GS and (b) lipoplexes formed by DNA interactions with neutral phospholipids in the presence of divalent cations. While the attraction of the latter group is in the nontoxicity of such carriers, the appealing speciicity of GS in the former group is due to the spacer that modulates physicochemical properties of molecules. We illustrate the role of this spacer in the aggregation process of DNA–cationic vesicles by studying cationic vesicles prepared as mixtures of helper lipids and GS.

13.2 POLYMORPHIC BEHAVIOR OF DNA–GS–PHOSPHOLIPID COMPLEXES 13.2.1

gss In gene delIvery

The GS structure contains three parts—cationic headgroups, spacer, and tail groups—allowing a great lexibility in modulating the structure according to the speciic requirement. In addition to GS, also the so-called “gemini lipids” (GL) are recognized in the ield of gene delivery. Their importance is borne by the similarities with cardiolipin, as GL comprises of four or more hydrophobic chains and multiple headgroups (for review see, e.g., Bombelli et al. 2009, Kumar et al. 2010). In general, GS show greatly enhanced surfactant properties relative to the corresponding monomeric surfactants. Results of differential scanning calorimetry (DSC), polarizing optical microscopy, x-ray diffraction, and neutron scattering experiments have conirmed their ability to form lyotropic mesophases with long-range organization at higher concentrations (Alami et al. 1993b). GS-DNA complexes as promising transfectors were introduced by the group of Kirby (Camilleri et al. 2000, Kirby et al. 2003). In this short review, we focus on a simple GS molecule of alkane-α, ω-diyl-bis(alkyldimethylammonium bromide) (CnGSm) with two cationic headgroups, and two hydrophobic chains (each having m carbons) connected by a polymethylene spacer (having n carbons) as shown in Figure 13.2. The CnGSm proved to have a powerful bactericidal activity (Imam et al. 1983, Devínsky et al. 1985) with the correlation between its structure, activity, and critical micellar concentration studied in further detail (Devínsky et al. 1987, Hirata et al. 1995). The CnGSm have been shown to be powerful plasmid curing agents (Belicová et al. 1995), and were used to increase the eficiency of DNA transfer into bacterial cells (Horniak et al. 1990). Nevertheless, DNA–CnGSm complexes on their own were found ineficient for in vitro transfection (Fisicaro et al. 2005, Foldvari et al. 2006). In the reported transfection experiments, the presence of helper lipid DOPE has been found crucial to raise the transfection eficiency (Foldvari et al. 2006). The highest transfection activity in vitro was then detected at CnGSm compounds with the spacer of three carbons (Foldvari et al. 2006, Wettig

CH3

CH3

CH3

N+ (CH2)n

N+

(CH2)m–1

(CH2)m–1

CH3

CH3

CH3

2 Br–

FIGURE 13.2 The structure of GS alkane-α, ω-diyl-bis(alkyldimethylammonium bromide) (CnGSm).

Structural Diversity of DNA–Phospholipid Aggregates

251

et al. 2007). Generally, GS with short spacers were reported for their enhanced transfection activity. For example, C2GS14/cholesterol/DOPE = 2:1:1 mol/mol/mol (Cardoso et  al. 2011), C4GS16/ DMPC = 1:1 mol/mol (Bombelli et al. 2005b) or derivatives of N,N-bisdimethyl-1,2-ethanediamine with n = 2 and m = 12 in a mixture with DOPE (1:2 mol/mol) have shown signiicantly higher transfection activity in comparison to standard commercial transfection reagents (Fisicaro et al. 2005). It should be stressed that the works reported so far have studied the use of simplest GS molecules in transfection experiments, and as such represent only a minor fraction of a great structural variety of GS molecules used in gene delivery. One can ind a preview of the large variety of GS structures applied in this ield in reviews published recently (Bombelli et al. 2009, Kumar et al. 2010). An example worth mentioning in this regard is a recent work of Donkuru et al. (2012), who have found a signiicant enhancement of the transfection eficiency in keratinocytes when incorporating a pHactive amine group within the GS spacer.

13.2.2

helper lIpIds

In many (but not all) cases, the inclusion of “zwitterionic helper” lipids in the cationic liposomes improved the eficiency of lipoplexes (Hui et al. 1996, Hirsch-Lerner et al. 2005, Koynova et al. 2005, Cardoso et al. 2011). Typically, volume fraction of the helper lipid included in the lipoplexes is signiicant, and it affects their electrostatics and self-assembling properties; that is, structure of the lipoplex, its level of hydration, and the secondary and tertiary structure of DNA (Zuidam and Barenholz 1997, 1998, Hirsch-Lerner and Barenholz 1998). Some of x-ray structural studies have shown the rate of DNA release from lipoplexes as well as transfection activity correlating well with nonlamellar phase progressions observed in cationic mixtures with membrane lipids (Hafez et al. 2001, Koynova et al. 2006, Koynova and Tenchov 2009). In particular, two groups of zwitterionic phospholipids—phosphatidylcholines (PCs) and phosphatidylethanolamines (PEs)—are often employed as helper lipids. Let us introduce briely their self-assembling properties. PEs have a smaller, less hydrated headgroup than PCs, and readily form intermolecular hydrogen bonds. This reduces the area per headgroup resulting in a cone-shaped molecule, especially when the molecule’s hydrophobic part consists of unsaturated acyl chains. The shape of the molecule then determines its critical packing parameter v/A0 lc to be greater than 1 (where v is the chain volume per molecule, A0 is the equilibrium area/molecule at the hydrocarbon–water interface, and lc is the critical chain length, which is smaller than the length of the fully extended all-trans chain (Israelachvili 1992)). The tendency for spontaneous curvature of such molecules at the water–lipid interface is to bend toward the water region, resulting in a negative curvature of membrane surface. Indeed, PEs in excess water are known to adopt columnar-inverted hexagonal phase (HII) above some temperature, which may be very high for saturated species (Seddon 1990). The most frequently used helper lipid DOPE forms HII phase (see Figure 13.1b) in excess water at room temperature, with the typical SAXD diffractogram shown in Figure 13.3a. Collected scattering data then provide complex information about its structure. The lattice parameter a = 7.58 nm was extracted from the positions of peak maxima according to the relation shk =2(h2 + hk + k2)1/2/√3a, where h and k are Miller indices. The radius of water cylinders inside the tubules was determined by Tate and Gruner (1989) to be 2.16 nm at 20°C, the thickness of the lipid layer 1.62–2.20 nm, while the area per DOPE molecule equaled to 0.474 nm2. This structure of the assembly is known not to be modiied by temperature increase, however, it yields a decrease in unit cell size. On the other hand, one can observe a transition to a lamellar phase while cooling the sample below 20°C, although it is dificult to predict the temperature of the HII–L α transition (TLH). Literature values of temperature range from −4°C to 16°C (Koynova and Caffrey 1994) and TLH have been found to depend on the rate of temperature change along the heating/cooling processes (Toombes et al. 2002). Finally, the presence of additives can shift TLH to higher values as well. Among phosphatidylcholines, dioleoylphosphatidylcholine (DOPC) is frequently utilized as a helper lipid, while successful transfection in lipoplexes with CnGSm has been reported also for

252

Liposomes, Lipid Bilayers and Model Membranes 52 48 TLH (°C)

LC(1) H(1, 0)

44 40 36 0

LC(2) H(1, 1)

LC(1)

12

n

H(2, 0) H(1, 0)

8

4

LC(3)

LC(2)

(e)

(d)

DNA

(c) H(1, 0) H(1, 1) (b)

H(2, 0)

H(2, 1)

H(3, 0) (a)

0.1

0.2

0.3

0.4

s (nm–1)

FIGURE 13.3 SAXD diffractograms of DOPE (a), and DNA-DOPE-CnGS12 complexes at molar ratio CnGS12:DOPE = 0.15 and with n = 2 (b), n = 4 (c), n = 10 (d), and n = 12 (e) while T = 20°C. Intensities are plotted in logarithmic scale as a function of s = (2/ λ ) sin(θ), where λ is the wavelength and 2θ is the scattering angle. Inset: HII–L α transition temperature (TLH) vs. the number of carbons (n) in spacer.

saturated dimyristoylphosphatidylcholine (Bombelli et al. 2005a, 2005b). The methylation of the phospholipid terminal ammonium group in PCs has a strong effect on the molecular packing. The bulky phosphorylcholine headgroup requires an area very similar to that occupied by the chains, and forms almost cylindrical shape of the molecule. The critical packing parameter v/A0 lc ~ 1 then predicts assembling of such molecules into bilayers (Israelachvili 1992). PCs in excess water indeed form vesicles of multilamellar bilayers. SAXD patterns usually show 2–3 equidistant diffraction peaks characteristic to a lamellar phase. Depending on the PC and temperature used, the repeat distance d takes up the values between 5.7 and 6.5 nm.

Structural Diversity of DNA–Phospholipid Aggregates

13.2.3

253

laMellar-to-hexagonal phase transItIons In dna–dope–cngsm

In this section, we demonstrate “the readiness” of DOPE to adopt a lamellar phase in the presence of additives, namely, GS. Structural changes are monitored through SAXD patterns shown in Figure 13.3. The results illustrate a modulation of TLH in DNA–CnGS12–DOPE (n = 2–12) lipoplexes prepared at the ixed ratio of cationic quaternary ammonium group to DOPE membrane surface charge density σm (R4N+/DOPE = 0.3). Neither GS with very short spacer (n = 2) nor those with long spacer (n = 10–12) had stabilized lamellar structure completely, as part of the mixture still adopted the H IIC phase (Figure 13.3b,d,e). On the other hand, we have detected a condensed lamellar phase LCα at 20°C for intermediate spacer lengths of n = 3–8 (Figure 13.3c). In fact, the TLH behaved nonlinearly as a function of CnGSm spacer length, showing a maximum at n ~ 6 (Figure 13.3, inset), while it increased almost linearly with the molar fraction of CnGS12 in DNA–DOPE– CnGS12 lipoplexes (Pullmannová 2011). We also note for comparison that monoalkylammonium bromide (C12TMABr) at the same ratio of cationic groups (i.e., C12TMA+:DOPE = 0.3) shifted the TLH considerably up to about 54°C in the case of DNA–DOPE–C12TMA+ complexes. Our model system then reveals how minor changes in the amphiphile design can tune polymorphic behavior of lipoplexes. At the molecular level, the intercalation of CnGS12 between DOPE molecules induces changes in two regions: electrostatic interactions between cationic ammonium groups and P−N+ dipoles of DOPE result in a lateral expansion of the interface, while a mismatch between the length of GS (m = 12) alkyl substituents and the length of hydrocarbon acyl chains of DOPE (18 carbons) creates defects in the hydrophobic membrane core. The latter defects are eliminated by chains bending or through their trans-gauche isomerization, leading to a decrease in the thickness of membrane hydrophobic region (Balgavý and Devínsky 1996, Dubnicˇková et al. 2004). The addition of amphiphiles to the lipid mixture affects also the membrane bending rigidity (Sainya et al. 1989), while the increase of the surface charge inhibits the formation of inverted nonlamellar phases (Lewis and McElhaney 2000). Indeed, the hydration of DOPE–CnGS12 mixtures in our model system with 150 mM NaCl solution has resulted in a homogenous dispersion of uni- or oligolamellar cationic vesicles. These complexes were formed due to the interactions of DNA with vesicles, while the isoelectric composition of the mixtures was kept constant (DNA:CnGS12 = 1 mol bp/mol). It is worth reminding that in addition to the elastic properties of DOPE–CnGS12 membrane, electrostatic interactions between the cationic membrane and DNA also play an important role in the aggregation process. As a result, the preferred complexation geometry is generally dictated by a nontrivial interplay between electrostatic and elastic contributions to the complex formation free energy as shown previously (May et al. 2000). Let us inspect structural properties of CnGSm to understand the observed polymorphism. The area per C2GS12 in LCα phase is 0.697 nm2 (Alami et al. 1993a), and is smaller than the area occupied by two C12TMA+ molecules which is equal to 2 × 0.480 = 0.960 nm2 (Lu et al. 1995), while possessing the same nominal σm. The area per CnGS12 molecule increases progressively with the length of spacer up to n = 12 (Alami et al. 1993a, Espert et al. 1998, Pisárcˇik et al. 2005). At the same time, the distance between cationic moieties of short spacers (n = 2, 3) is ixed, whereas the longer spacers (n = 10–12) are lexible and hydrophobic enough to fold into the membrane’s hydrophobic region (Alami et al. 1993a, Pisárčik et al. 2005). Thus both the suficiently large area per molecule and the rigidity of spacers with intermediate lengths (n = 3–8) stabilize lamellar phase in DNA– DOPE–CnGSm lipolexes even at low σm. In other words, lamellar phase is formed when the shape of the molecule reaches the critical packing parameter of about 1. Too long spacers (n = 10–12), on the other hand, folding into the membrane’s hydrophobic region increase its elasticity and also the tendency to form nonlamellar phases. In terms of the packing parameter, the average volume per chain in hydrophobic region v increases, while lc decreases, shifting thus the value of the expression above 1 (i.e., v/A0 lc > 1).

254

Liposomes, Lipid Bilayers and Model Membranes

Most importantly, and in spite of low surface charge density, DOPE–CnGS12 cationic liposomes condense DNA extremely well in all the various structures mentioned before. Diffractograms of lipoplexes with CnGS12 of intermediate spacers (n = 3–8) that form LCα phase have revealed a diffraction peak related to the regularly packed DNA strands with lateral periodicity of about 5.5 nm (Figure 13.3c, marked by arrow). The repeat distance of LCα decreases slightly with increasing length of spacer (n = 2–12) in the range 7.10–6.85 nm. The lattice parameter of condensed hexagonal phase H IIC formed in the case of n = 2 and 12 also decreases with the length of spacer (a = 7.15 and 6.85 nm, respectively), while the values are smaller compared to HII phase formed by DOPE itself (a = 7.56 nm). Additionally, reported radius of the water-illed DOPE cylinders (RW ~ 2.16 nm (Tate and Gruner 1989)) is reduced in the H IIC phase to 1.3–1.4 nm (Koltover et al. 1998, May et al. 2000). Each DNA is then surrounded by a highly (negatively) curved lipid monolayer. This cylindrically concentric geometry provides eficient neutralization of the DNA charges by the cationic surface charges—especially at the isoelectric point (May and Ben Shaul 1997)—and explains the condensation eficiency of this system.

13.2.4

condensed laMellar phase In dna–dope–CnGSm

13.2.4.1 Effect of Surface Charge Density The highest transfection activity in vitro in DNA–DOPE–CnGSm complexes has been detected for CnGSm compounds with a spacer of three carbons (Foldvari et al. 2006). Polymethylene spacer with three carbons is rigid, and causes thus a progressive stabilization of LCα phase, as shown in Figure 13.4. At the isoelectric composition of DNA–cationic liposome complexes in LCα phase, the DNA interhelical spacing dDNA is inversely related to the lipid membrane surface charge density σm (dDNA = e− /(l0σm), where l0 is the average distance per anionic charge along the DNA backbone (Koltover et al. 1999)), and its values have been found in the range 2.45–5.71 nm (Radler et al. 1997). It is worth noting that the most compressed spacing of 2.45 nm approaches the short-range repulsive hard-core interactions of the DNA rods (Podgornik et al. 1989). In our model system, the DNA spacing was decreasing almost linearly with C3GS12:DOPE molar ratio increasing up to 0.35 (i.e., increasing σm), where it reached a minimum of 3.7 nm (Figure 13.6, full symbols). The next increase in σm did not bring DNA strands closer, and we have observed “a plateau” in dDNA most likely resulting from interplay between steric repulsive and hydration forces and DNA propensity for the effective charge screening. A further increase of cationic amphiphiles in the membrane inally resulted in species demixing, which induced a lateral phase separation as demonstrated in Figure 13.4 at C3GS12:DOPE = 1 mol/mol. In addition to the LCα phase, there appeared an LαB phase enriched in C3GS12 molecules and with considerably smaller periodicity (dB ~ 5.07 nm). High σm and the lateral phase demixing lead also to tighter DNA packing yielding dDNA = 2.93 nm. Interestingly, the species demixing was successfully avoided when employing GS with longer alkyl chain (i.e., m = 14). The diffractogram of this complex prepared at lower molar ratio C3GS14:DOPE = 0.8 (Figure 13.4) has suggested almost the same DNA packing, showing dDNA = 2.90 nm. Finally, elongation of the alkyl substituent of CnGSm is known to affect also the elasticity of the membrane hydrophobic core, causing the cationic bilayers becoming more rigid (May et al. 2000) and condensing the DNA strands more eficiently. Structural parameters relect these changes particularly at higher σm. For example, at a molar ratio C3GS14:DOPE = 0.5, we have detected only a small increase in the bilayer repeat distance (Δd = 0.24 nm), while a signiicant reduction in the DNA spacing (Δd DNA = 0.60 nm) is observed when comparing complexes with C3GS12 and C3GS14. This experiment then illustrates the modulation of structural parameters (d, d DNA) through the changes in membrane bending rigidity that in turn is tuned by the composition of lipid–surfactant mixture.

255

Structural Diversity of DNA–Phospholipid Aggregates LC(1)

LC(2)

DNA

C3GS14/DOPE (mol/mol) 0.80

LC(1)

0.30 LB(1) C3GS12/DOPE (mol/mol)

LC(2) LB(2)

1.00

0.30

H(1, 0) 0.15 H(1, 1) 0.10 0.1

0.2

0.3

0.4

s (nm–1)

FIGURE 13.4 SAXD diffractograms of DNA–DOPE–C3GSm (m = 12 and 14) complexes at selected molar ratios and T = 20°C. Intensities are plotted in logarithmic scale.

13.2.4.2 Effect of the CnGSm Spacer on DNA Packing in LCα The DNA–DNA distance in LCα phase results from a natural tendency of complexes to minimize free energy due to matching between the charges of differently charged species. This then also allows a release of (previously conined) counterions into aqueous solution (May and Ben Shaul 2004), which is accompanied by an entropy gain (Wagner et al. 2000). A theoretical model of LCα phase involves two contributions: the electrostatic (charging) free energy and the (in-plane) lipid-mixing entropy, assuming the absence of curvature or lipid thickness modulations. DNA strands are assumed to be “stiff rods” with a uniform distribution of negative charge. This is justiied by the fact that the persistence length of double-stranded DNA (~50 nm) is much larger than its diameter (~2 nm), and also much larger when compared to membrane dimensions (both the lipid and water layer thicknesses). The

256

Liposomes, Lipid Bilayers and Model Membranes

cationic membrane in liquid-crystalline state represents a dynamic system allowing lateral diffusion of all membrane components, and the one-dimensional DNA lattice, due to the effective screening of its negative charges, can induce a spatial modulation (or “polarization”) of the cationic charges minimizing thus the electrostatic energy of the system (Harries et al. 1998, May and Ben Shaul 2004). In complexes with GS, it is the length of spacer and its lexibility modulating the freedom of R4N+ group arrangements in the vicinity of DNA. We have studied the role of the spacer’s length (n = 2–12) in DNA–DNA packing in DNA–CnGS12–DOPE lipoplexes at different σm. Figure 13.5 adapted from Pullmannová et al. (2012b) summarizes the obtained results. The plot of DNA spacing (dDNA) as a function of the number of carbons (n) in the spacer of CnGS12 relects the ability of each GS to adopt a conformation leading to effective DNA charge neutralization at the given nominal σm. At low σm (R4N+:DOPE = 0.5 mol/mol), the DNA strands in complexes without restrictions imposed by spacers (i.e., C12TMA+ complexes) favored the tighter DNA packing compared to complexes with a spacer (i.e., CnGS12 complexes). Obviously, cationic moieties must be displaced over relatively large distances to effectively screen the negative DNA charges. On the other hand, their “freedom” is modulated by the spacer’s length and its lexibility. It is worth noting that systems with short spacer (n = 3) and those with long ones (n = 10, 12) brought DNA strands closer (dDNA ~ 4.4 nm) when compared to systems with intermediate spacer lengths. High surface charge density systems (R4N+:DOPE = 0.7–0.8 mol/mol) resulted in closer DNA packing in all complexes when compared to the previous case of low charge density. We have observed CnGS12 with short and rigid spacer being more eficient in the reduction of the DNA– DNA distance in comparison to those possessing longer spacers. Obviously, the long and rather rigid spacers (n = 10, 12) create a steric hindrance in the strands approach (dDNA ~ 4–4.1 nm). dDNA was increasing linearly with n and correlated well with the area per CnGS12 molecule (see Figure 13.5). This correlation observed at a high surface charge density has demonstrated nicely the dominance of electrostatic interactions in the building architecture of complexes. The theoretical model predicts charge segregation in the vicinity of DNA being rather weak at suficiently high σm. Similarly, at energetically optimal composition (ideal charge matching), a model of local charge

dDNA (nm)

4.8 4.4 4.0

ACnGS12 (nm2)

3.6 3 2 1 0

2 4 6 8 10 12 Number of spacer carbons n

FIGURE 13.5 The top panel shows DNA interhelical distance dDNA in DNA–DOPE–CnGS12 complexes at 20°C as a function of number of spacer carbons n at the molar ratios CnGS12/DOPE = 0.4 and C12TMA/ DOPE = 0.8 (◊), CnGS12/DOPE = 0.35 and C12TMA/DOPE = 0.7 (○), CnGS12/DOPE = 0.25 and C12TMA/ DOPE = 0.5 (◼). The bottom panel shows headgroup area vs. number of spacer carbons n (●). (Extracted from literature Alami, E. et al. 1993a. Langmuir 9:1465–1469; Espert, A. et al. 1998. Langmuir 14:4251–4260; Lu, J. R. et al. 1995. J. Phys. Chem. 99(12):4113–4123.) Note that points at n = 0 were obtained using the DNA–DOPE– C12TMA complexes. (Adapted from Pullmannová, P. et al. 2012b. Biochim. Biophys. Acta 1818:2725–2731.)

257

Structural Diversity of DNA–Phospholipid Aggregates

density between two neighboring DNA strands does not show any dramatic differences, if the lipidmixing entropy contribution to free energy is ignored, and the constant electrical potential of the bilayer surfaces is assumed (Harries et al. 1998). It must be stressed that the earlier observed linear dependence of dDNA on n is conditioned not only by the high surface charge density but also by the used phospholipid and DNA. Analogous experiments performed with egg yolk phosphatidylcholine (EYPC) at CnGS:EYPC = 0.5 mol/mol revealed quasiparabolic dependence of dDNA(n) with a maximum at n ~ 6–8 (Uhríková et al. 2012b). The structural difference in phospholipid headgroups apparently modiies both the surface roughness and the charge balance of cationic membranes. We will discuss the implication of PC lipids in gene delivery and DNA complexes in the following section.

13.2.5

structural varIety In dna–pc–cngsm

There are several methods for the preparation of cationic vesicles from PC lipids. In the simplest approach, a solution of cationic surfactant is added to the dispersion of neutral vesicles, where molecules of surfactant intercalate between zwitterionic PC molecules, and the vesicles become positively charged (Haydon and Myers 1973). Generally, large unilamellar vesicles are formed if σm exceeds 1–2 µC/cm2 (Hauser 1993). Cationic vesicles prepared using either saturated or unsaturated PCs (e.g., DLPC, DPPC, DOPC) and CnGSm (where, m = 12–16) condense DNA forming a condensed lamellar phase LC. The lipids in these complexes are in the gel or liquid-crystalline state, depending on the temperature (Uhríková et al. 2002, 2004a). Proiting from our model system, it is interesting to compare structural parameters of lipoplexes prepared with the same experimental protocol (in 150 mM NaCl) while using phospholipids with two different headgroups (i.e., DOPE and DOPC). We observed small differences in the DNA–DNA packing (dDNA), together with the bilayer repeat distance in DOPE system being ~0.3 nm smaller in comparison to lipoplexes with DOPC (Figure 13.6). However, these variations could be accounted for by the differences in the size and hydration of the two lipid headgroups. We conclude therefore that structural parameters of LCα phase formed by the DNA interaction with C3GS12–phospholipid vesicles prepared with either DOPE or DOPC are, for the most part, equivalent.

8

d, dDNA (nm)

7

d

6 5 dDNA 4 3 0.1

0.2

0.3

0.4

0.5

C3GS12/phospholipid (mol/mol)

FIGURE 13.6 Structural parameters of DNA–phospholipid–C3GS12 complexes at 20°C. The dependence of the repeat distance d, and DNA–DNA spacing dDNA on the molar ratios of C3GS12 to DOPE (full symbols) and DOPC (empty symbols). (Extracted from Pullmannová, P. et al. 2012a. Biophys. Chem. 160(1):35–45; Pullmannová, P. et al. 2012b. Biochim. Biophys. Acta 1818:2725–2731.)

258

Liposomes, Lipid Bilayers and Model Membranes

Similar to the PE systems discussed previously, a key parameter for the transfection eficiency of LCα phase-forming PC complexes is the membrane surface charge density σm. Lin et al. (2003) have deined the optimal membrane surface charge density σm* ≈ 1.04 e − /nm 2 for complexes to be stable yet allowing successful escape of DNA through their fusion with endosomal membranes. The area per molecule of CnGS12, which depends on the spacer length, ranges between 0.7 and 2.3 nm 2 (Alami et  al. 1993a). As such, the C2GS12 molecule with the smallest area carries 2.9 e − /nm 2 while the C12GS12 molecule with the long spacer carries only 0.9 e − /nm 2. Further, the need for a helper lipid in transfection experiments (Fisicaro et al. 2005, Bombelli et al. 2005b, Foldvari et al. 2006) causes “a dilution” of R4N+ cationic moieties at the membrane surface. The advantage of CnGSm with short spacers to get optimal σ m* at their lower volume fraction in the membrane is thus obvious. On the other hand, the volume fractions of CnGSm in phospholipid membranes in various transfection experiments were rather high, with different authors reporting molar ratios of GS:phospholipid from 0.5:1 up to 1:1 (Fisicaro et al. 2005, Bombelli et al. 2005a). Let us then inspect shortly the structural variety in DNA–PC–CnGSm lipoplexes at high CnGSm:PC molar ratios and at isoelectric composition. We detected a condensed lamellar phase in DNA–CnGS12–DPPC, while utilizing short spacers (n = 2, 3), up to a molar ratio of CnGS12:DPPC = 0.5 (Figure 13.7b). Interestingly, the lattice parameters (d = 6.34 nm, dDNA = 3.40 nm) were close to those observed in lipoplexes with DOPE. However, further increase in the surfactant’s portion in the membrane as well as the spacer’s elongation induced a demixing of species yielding the coexistence of two lamellar phases, as illustrated by SAXD diffractograms in Figure 13.7a and d. We note the progression of the effect of species demixing in our system with the phospholipid’s acyl chains shortening and their saturation. For example, in the case of DNA–C4GS12–DLPC system, we detected the coexistence of two phases already for molar ratios starting at 0.35 (e.g., Figure 13.7a) (Uhríková et al. 2004a), while the use of same surfactant and unsaturated EYPC showed better miscibility with the phase coexistence identiied at 1:1 molar ratio (Figure 13.7d) (Uhríková et al. 2012b). It should be mentioned that the elongation of the spacer to n = 6–12 changed the morphology completely. For both the helper lipids, either saturated DPPC or unsaturated EYPC, hexagonal symmetry was found to be dominant in the lipoplex structure (Figure 13.7c and e). In regards to the spacer length (n), the lattice parameter of DNA–EYPC–CnGS12 (CnGS12:EYPC = 1:1 mol/ mol) decreased nonlinearly from a = 6.31 nm at n = 6 down to a = 5.86 nm at n = 12. For comparison, Figure 13.7f shows the diffractogram of C4GS14–DNA complexes where the HI phase has been formed due to DNA interactions with C4GS14 micelles. The lattice parameter of this phase increased in the range 4.80–5.27 nm with increasing C4GS14/DNA molar ratio (Uhríková et  al. 2005b). Unfortunately, the hexagonal symmetries of arrangements in normal HI and inverted HII phases provide the same SAXD patterns of relections, which does not allow for the two structures to be distinguished. To arbitrate between intercalate H IIC phases, the discussion should turn back to the elasticity of the membrane’s hydrophobic region and surface-bending rigidity.

13.3 AGGREGATES OF DNA–NEUTRAL PHOSPHOLIPID–METAL CATION SYSTEMS More than three decades ago Budker et al. (1978) reported the ability of divalent metal cations to mediate the interactions between DNA and PC vesicles. Different experimental methods used to study physicochemical properties of DNA–PC–ion2+ aggregates include turbidimetry, microcalorimetry, ESR, IR, and so on (Budker et al. 1980, Vojcˇíková and Balgavý 1988, Vojcˇíková et al. 1989, Bruni et  al. 1997, 2001, Khusainova et  al. 1999, Balgavý et  al., 2002). Initially, electron freeze fracture micrographs of mixtures with Ca2+ suggested structures with long-range organization (Tarahovsky et al. 1996, Khusainova et al. 1999). SAXD revealed later the presence of a condensed lamellar phase (LC) with DNA strands regularly packed between lipid bilayers in

259

Structural Diversity of DNA–Phospholipid Aggregates (c)

H(1, 0)

(f)

H(1, 0)

H(1, 1) H(2, 0) LC(1)

H(1, 1)

H(1, 0)

(b)

H(2, 0)

(e)

DNA

LC(2)

LC(3)

LC(1)

LC(1) LB(1)

H(1, 1)

H(2, 0)

(a) (d)

LB(1)

LC(2) LC(2)

LB(2)

0.1

0.2

0.3 s (nm–1)

0.4

LB(2)

0.5

0.1

0.2

0.3 s (nm–1)

0.4

0.5

FIGURE 13.7 SAXD diffractograms of DNA–phosphatidylcholine–CnGS12 complexes for selected PCs, temperatures, and molar ratios. C4GS12:DLPC = 0.35 at 20°C (a), C3GS12:DPPC = 0.5 at 60°C (b), C6GS12:DPPC = 0.6 at 60°C (c), C4GS12:EYPC = 1 at 20°C (d), C6GS12:EYPC = 1 at 20°C (e), and C4GS14/ DNA = 2.4 mol/base at 20°C (f). Intensities are plotted in logarithmic scale.

DNA–dipalmitoylphosphatidylcholine–Mg2+ and Ca2+ systems (Uhríková et al. 2001, McManus et  al. 2003). Finally, in addition to the biologically relevant divalent cations of Ca2+ and Mg2+, also Fe2+, Co2+, and Mn2+ have shown the ability to mediate DNA–neutral phospholipid binding (Francescangeli et al. 2003). Generally, in the system consisting of DNA, neutral phospholipid, and divalent cations, in addition to a formation of triple complex (DNA–phospholipid–ion2+), one must consider also the interactions of DNA–ion2+ and phospholipid–ion2+. Alkaline earth metal cations (Ca2+, Mg2+) preferentially interact with the phosphate groups of DNA, thereby reducing the charge repulsion between opposing strands of the double helix (Luck and Zimmer 1972) and stabilizing the polynucleotide molecule. Transition metal cations on the other hand interact more extensively with DNA

260

Liposomes, Lipid Bilayers and Model Membranes

bases causing disruption of base pairing and destabilization of DNA molecule (Eichhorn and Shin 1968, Luck and Zimmer 1972). According to Duguid et al. (1993), the afinity of divalent cations to DNA bases decreases in the order Hg2+ > Cu2+ > Pb2+ > Cd2+ > Zn2+ > Mn2+ > Ni2+, Co2+ > Fe2+ > Ca2+ > Mg2+, Ba2+. Fluorescence and UV–VIS experiments have documented the ability of Ca2+, Mg2+, Zn2+, Co2+, and Ni2+ to condense DNA in the presence of lipid bilayer, and to protect it against thermal denaturation in the order Mg2+ ~ Ca2+ ~ Zn2+ > Ni2+ > Co2+ (Lengyel et  al. 2011). The binding site for cations is near the negative phosphate group of the P– –N+ dipole in the phospholipid headgroup (Shepherd and Buldt 1978, Izumitani 1994). By neutralizing the negative charge in the phosphate group, the lipid bilayer becomes positively charged, and starts to form aggregates due to electrostatic interactions with the negatively charged phosphate groups of DNA. Even though this binding scenario looks simple, binding equilibria between individual components are complex. Divalent metal cations bind rather weakly to zwitterionic lipids such as PC and PE (Altenbach and Seelig 1984), the preference for cation binding weakens with the degree of hydrocarbon chain unsaturation, and it depends on the thermodynamic phase of the phospholipid. As a result, one can ind binding constants for calcium to range from 1 to 400 M−1 depending on the lipid and experimental method utilized. The presence of anionic polyelectrolyte between lipid bilayers increases the binding constant signiicantly (Huster and Arnold 1998). The aggregates pack “softly,” and their resistance against the dissolution is lower when compared to those prepared with cationic surfactants or cationic lipids (Uhríková et al., 2004b). Whereas the isoelectric point of DNA–cationic amphiphile lipoplexes can be determined by simple calculations, the evaluation of the isoelectric point of DNA–neutral phospholipid–divalent metal cation aggregates is not a trivial task. Because of the high mobility of metal cations, DNA binding to the lipid bilayers is a complicated balance of metal cation bridging and charge screening. The mechanism of DNA–phospholipid–metal cation interactions and binding stoichiometry are still under discussion (Kharakoz et al. 1999, McManus et al. 2003, Gromelski and Brezesinski 2004, Mengistu et al. 2009, Bruni et al. 2011). The fraction of DNA incorporated in aggregates with DOPC reaches 40–45% of the total DNA volume in the solution, and the binding capacity as a function of the Ca2+ concentration increases up to 15 mM of CaCl2 (Rajnohová et al. 2010, Lengyel et al. 2011). For comparison, the average DNA condensation eficiency with CnGSm–DOPE is about 98% (Pullmannová et al. 2012b). Thus, the ensemble of effects including low afinity of zwitterionic phospholipids to cations, their high mobility, and rather weak interactions between all components results in a large structural diversity of lipoplexes, as discussed below.

13.3.1

ca2+ In dna–dppc InteractIons

The most studied and documented system among DNA–PC–Ca2+ complexes is that with dipalmitoylphosphatidylcholine (DPPC). DPPC is a synthetic zwitterionic phospholipid with a well-known phase diagram. It forms the tilted gel phase (L β′) at temperatures below 35°C, rippled gel phase (Pβ) below 42°C, and liquid-crystalline phase (L α) above 42°C (Albon and Sturtevant 1978, Stumpel et al. 1983). Figure 13.8a shows SAXD diffractogram resulting from DPPC lamellar phase with a periodicity of 6.33 nm as obtained at 20°C and in excess water condition (Uhríková et al. 2007). Representative diffractograms of the structures observed in DNA–DPPC–ion2+ aggregates are also shown in Figure 13.8. Let us follow structural changes resulting from the increasing concentration of Ca2+ ions. The diffractogram of the DPPC:DNA aggregates prepared at 3:1 mol/base ratio and in the presence of 1 mM Ca2+ (Figure 13.8b) represents a superposition of two lamellar phases. Note, we have not observed any peak relecting DNA–DNA lateral packing, and had to conclude this indirectly as follows. The irst phase (LDPPC) is characterized by a repeat distance of 6.49 nm, which is consistent with pure DPPC phase at 20°C. Subtraction of lipid bilayer thickness of 5.38 nm (Uhríková et al.

261

Structural Diversity of DNA–Phospholipid Aggregates

(i)

LC(1)

(e) LC(1)

DNA

LC(2) LV(2) DNA LV(3)

LC(2)

(d) (h)

LC(1) (c)

LV(2) LV(3) LC(2)

LDPPC(1)

(g) LDPPC(2) (b)

LDPPC(1) LX(1) (f)

LX(2) LDPPC(2) (a)

0.1

0.2

0.3 s (nm–1)

0.4

0.1

0.2

0.3

0.4

s (nm–1)

FIGURE 13.8 SAXD diffractograms of DPPC at 20°C (a), and DNA–DPPC–cation2+ aggregates in the presence of: 1 mM CaCl2 at 20°C (b), 2 mM CaCl2 at 20°C (c), 2 mM CaCl2 at 37°C (d), 50 mM CaCl2 at 20°C (e), 50 mM CaCl2 at 60°C (f), 20 mM ZnCl2 with the solution’s total ionic strength of 65 mM at 20°C (g), 40 mM ZnCl2 at 20°C (h), and 20 mM ZnCl2 with the solution’s total ionic strength of 122 mM at 20°C (i). Intensities in (c) and (d) panels are plotted in logarithmic scale.

2009b) from the repeat distance provides an interlamellar thickness of 1.2 nm. This is clearly smaller than what would be needed for hydrated DNA (diameter of 2.5 nm), and thus it is indicative of no inclusion of DNA or Ca2+ ions in this phase. The other phase (L X), on the other hand, has the larger repeat distance of 8.04 nm. Applying then the same procedure as before, one inds the interlamellar thickness of 2.66 nm, which offers enough space to accommodate DNA strands intercalated in water layers between the DPPC bilayers. Our previous work (Uhríková et al. 2005a) proved such a structure also in DOPC–DNA aggregates in a wide range of cation concentrations (0–76.5 mM of Ca2+ or Mg2+), and similar structures were observed and discussed in the literature (McManus et al. 2003, Francescangeli et al. 2003). The coexistence of two phases can be explained by a lateral segregation

262

Liposomes, Lipid Bilayers and Model Membranes

of DNA strands and metal cations, similar to DNA–cationic amphiphiles systems suggested in theoretical models (Harries et al. 1998) and conirmed experimentally (Macdonald et al. 2000). Figure 13.8c and d show diffractograms of the same mixtures as earlier but prepared this time in 2 mM Ca2+ solution at 20°C and 37°C, respectively. Smooth full line corresponds to a curve resulting from a superposition of 3 (or 5) Lorentzians atop a linear background obtained by a deconvolution of diffraction maxima. Clearly, a small increase in Ca2+ concentration improved the organization of DNA strands inside the L x phase, although the DNA peak is broad (half-width at the half-maximum of 0.06 nm−1) indicating a rather poor organization of the DNA lattice (dDNA ~ 5 nm). Interestingly, further heating of aggregates caused an intensity decrease in both relections of LDPPC phase, which “dissolved” completely at 37°C (Figure 13.8d). More importantly, however, the increase of Ca2+ concentration to 50 mM showed a condensed lamellar phase ( LCβ ) at 20°C with a well-deined DNA peak (Figure 13.8e). When heating these aggregates to higher temperatures, we observed an increase of the DNA peak width and a decrease of its intensity, gradually merging with the background (see also Uhríková et al. 2007). In the liquid-crystalline state of DNA–DPPC–Ca2+ aggregates at temperatures above 44°C (Khusainova et al. 1999), LCβ appears to convert to the L x phase with DNA strands packed irregularly (Figure 13.8f). We have observed more or less similar temperature behavior when Mg2+, Co2+, Ni2+, or Zn2+ mediate DNA–DPPC binding. In all of these cases, LCβ was created with regular packing of DNA strands between DPPC bilayers while in the temperature range corresponding to the gel state of DPPC. However, the DNA regular packing was disrupted when the lipids transitioned into a liquidcrystalline state, as shown previously (Uhríková et al. 2007, Lengyel 2010). The diffraction peak related to regularly packed DNA strands between DPPC bilayers in the gel state of the aggregate was also reported by McManus et al. (2003), and Pisani et al. (2006), although these authors observed coexistence of two phases at different cation concentrations (Ca2+, Mg2+, Mn2+) and temperatures. As documented earlier, structural variety of DNA–neutral PC–cation2+ aggregates does not depend only on the composition of samples. Due to complex binding equilibria, the way of mixing individual components during the aggregate preparation may play a role as well. McManus et al. (2004) have documented an organization of DNA strands in DNA–DPPC–Ca2+ aggregates in rectangular columnar phase with 2D lattice constants a = 3.53 nm and b = 2d = 15.66 nm, where d is the periodicity of lipid bilayer stacking. However, as these authors have reported, aggregates prepared with other compositions and calcium concentrations did not show such high levels of DNA organization, even a long time after preparation.

13.3.2

dna–dppc–transItIon Metal catIons

When transition metal cations (Co2+, Ni2+) as well as Zn2+ mediate DNA–DPPC binding, structural differences occur. This can be illustrated using zinc. Zinc plays a fundamental role in several critical cellular functions such as protein metabolism, gene expression, structural and functional integrity of biomembranes, and metabolic processes (Christianson 1991). By comparison with other micronutrients, zinc is present in biological systems at high concentrations, particularly in biomembranes. In animal systems, zinc ranges from 10−3 M in some membrane vesicles (Williams 1988). At low concentrations of zinc (cZnCl2 ≤ 20 mM), DNA–DPPC–Zn2+ aggregates showed microstructures similar to those of divalent alkaline earth metals (Figure 13.8g). However, the diffraction pattern of DNA–DPPC dispersed in 40 mM ZnCl2 solution showed two phases (Figure 13.8h): LCβ with lattice parameters d = 8.54 nm and dDNA = 6.14 nm, and a lamellar phase LV with periodicity d = 13.5 nm. The latter periodicity (with water gap dW ~ 8 nm) is too big when compared to the structural parameters of LC phase, either when it is accommodating DNA strands or when formed by neutral lipids themselves. This has indicated a destruction of the longrange lamellar structure by its swelling into excess water, where the periodicity is dictated by the concentration of ions in the solution (Uhríková et al. 2009a).

Structural Diversity of DNA–Phospholipid Aggregates

263

Figure 13.8i displays the diffractogram of DNA–DPPC in 20 mM ZnCl2 solution where NaCl was used to modulate its ionic strength (total ionic strength Is = 122 mM). The observed structural parameters (d = 8.18 nm, dDNA = 5.70 nm, and dV = 13.12 nm) are close to the results of the previously discussed system (i.e., DNA–DPPC in 40 mM ZnCl2). It has been shown that neither the repeat distance nor the bilayer thickness of neutral phospholipids changed with the solution ionic strength ranging between 1 and 500 mM NaCl (Pabst et al. 2007). Moreover, SAXD and SANS experiments on this system have shown only a fraction of lipid being bound by DNA and Zn2+ forming LC phase. The rest of DPPC forms the LV phase, with its periodicity decreasing as the concentration of ions in solution increases, and the phase is macroscopically separated from the LC phase (Uhríková et al. 2009a, 2012a). To understand the observed structural changes, one must consider the afinity of both lipid and DNA to zinc. Zinc cation possesses a high afinity to electronegative groups (Gresh and Šponer 1999), such as ester oxygens and/or carbonyl groups of the lipid headgroup. Accordingly, Zn2+ bridges neighboring zwitterionic lipids while forming a lipid:Zn2+ complexes at the ratios from 2:1 up to 1:1 at saturation (Binder et al. 2001). In addition, the sedimentation of DNA has been proven at millimolar concentrations of zinc (Kejnovsky and Kypr 1998). Thus, at higher zinc concentrations, binding sites of both DNA and DPPC are saturated, and zinc does not mediate the binding any more. The electrostatic screening of Zn2+ charge due to ion accumulation and formation of a diffuse double layer at the membrane surface then leads to a macroscopic phase separation. Finally, a similar behavior has been observed for DNA–DPPC system in the presence of Ni2+ and Co2+ transition metal cations (Lengyel 2010).

13.4 CONCLUDING REMARKS It is well accepted that progress in transfection eficiency requires a full understanding of the roles played by both the physicochemical properties of gene delivery carriers, and cellular processes. The latter represents many complex steps starting with the association of the lipoplex by a cell, and ending with successful DNA transcription in the nucleus. These pathways are not fully understood. Although diverse types of cationic lipid carriers have been designed and tested for gene delivery, there is still a need for systematic research and targeted improvements in the design of carriers. An important knowledge for the targeted designing of lipoplexes concerns the relation between the stereochemistry of used amphiphiles and the resulting morphology of the carrier. In this chapter, we have examined two kinds of lipoplexes, frequently reported because of their lower toxicity. We have highlighted dependencies of structural parameters and resulting morphologies on the increasing volume fraction (concentration) and/or on the small differences in structures of some promising amphiphilic molecules. The lexibility in the design of GS in particular offers a powerful tool to modulate the structure in order to get the required characteristics of an ideal gene delivery carrier. We have demonstrated that by modulating the length of spacer and the amphiphile/ lipid ratio, one can achieve plausible structures such as condensed lamellar phase LCα , condensed inverted hexagonal phase H IIC , or normal hexagonal phase HI. Our model system, on the other hand, has also revealed a limitation in miscibility of GS/lipid systems indicating a high probability of species demixing at concentrations commonly used in transfection experiments. The variety of structures formed due to DNA interactions with phosphatidylcholines in the presence of divalent cations is also very large. We have identiied structures such as condensed lamellar phase LC, lamellar phase L X with irregularly packed DNA strands intercalated between the adjacent lipid bilayers, as well as a coexistence of pure lipid lamellar phase LPC with L X or LC phases. High concentrations of transition metal ions (Co2+, Ni2+) and Zn2+ induced a macroscopic phase separation. In addition to the condensed lamellar phase, a partially ordered lamellar phase LV has been identiied and linked to the electrostatic screening of phospholipid positive surface charge at a high concentration of solution ions. In conclusion, this chapter has illustrated a variety of lyotropic liquid-crystalline mesophases possibly ensuing from the targeted designing of delivery vectors, and that are potentially applicable

264

Liposomes, Lipid Bilayers and Model Membranes

in the ield of gene therapy. The inherent advantages of these vectors over their viral counterparts make the reviewed research imperative to the continued advances in this ield.

ACKNOWLEDGMENT This work is dedicated to Professor F. Devínsky with thanks for many years of fruitful collaboration and for kindly providing the GS chemicals. The authors thank S.S. Funari for his assistance at SAXD experiments and N. Kucˇerka, J. Teixeira, and P. Balgavý for their comments and help in the manuscript preparation. Financial support provided by the European Community’s Seventh Framework Program (FP7/2007–2013) under grant agreement no. 226716 (HASYLAB project II-20100372 EC), by the JINR project 04-4-1069-2009/2014, and by MŠ SR grant VEGA 1/1224/12 is gratefully acknowledged.

REFERENCES Alami, E., Beinert, G., Marie, P., and Zana, R. 1993a. Alkanediyl-alpha, omega-bis(dimethylalkylammoniumbromide) surfactants 3. Behavior at the air-water interface. Langmuir 9:1465–1469. Alami, E., Levy, H., and Skoulios, A. 1993b. Alkanediyl-alpha,omega-bis(dimethylalkylammonium bromide) surfactants. 2. Structure of the lyotropic mesopahses in the presence of water. Langmuir 9(4):940–944. Albon, N. and Sturtevant, J. M. 1978. Nature of the gel to liquid crystal transition of synthetic phosphatidylcholines. Proc. Natl. Acad. Sci. USA 75(5):2258–2260. Altenbach, Ch. and Seelig, J. 1984. Ca2+ binding to phosphatidylcholine bilayers as studied by deuterium magnetic resonance. Evidence for the formation of a Ca2+ complex with two phospholipid molecules. Biochemistry 23:3913–3920. Balgavý, P. and Devínsky, F. 1996. Cut-off effects in biological activities of surfactants. Adv. Colloid. Interface Sci. 66:23–63. Balgavý, P., Vojcˇíková, L., and Švajdlenka, E. 2002. Microcalorimetric study the DNA melting in the presence of phosphatidylcholine liposomes and magnesium ions. Acta Facultatis Pharm. Univ. Comenianae 49:17–25. Barreleiro, P. C. A., Olofsson, G., and Alexandridis, P. 2000. Interaction of DNA with cationic vesicles: A calorimetric study. J. Phys. Chem. B 104(32):7795–7802. Belicová, A., Ebringer, L., and Devínsky, F. 1995. Plasmid curing capacity of bisquaternary ammonium-salts— Comparison of 4 sets of N,N′-bis(alkydimethyl)-3-X-1,5-pentanediammonium dibromides. Biologisches Zentralblatt 114(3):294–298. Bell, P. C., Bergsma, M., Dolbnya, I. P. et al. 2003. Transfection mediated by gemini surfactants: Engineered escape from the endosomal compartment. J. Am. Chem. Soc. 125(6):1551–1558. Binder, H., Arnold, K., Ulrich, A. S., and Zschornig, O. 2001. Interaction of Zn2+ with phospholipid membranes. Biophys. Chem. 90:57–74. Bombelli, C., Caracciolo, G., Di Proio, P. et al. 2005a. Inclusion of a photosensitizer in liposomes formed by DMPC/Gemini surfactant: Correlation between physicochemical and biological features of the complexes. J. Med. Chem. 48(15):4882–4891. Bombelli, C., Faggioli, F., Luciani, P., Mancini, G., and Sacco, M. G. 2005b. Eficient transfection of DNA by liposomes formulated with cationic gemini amphiphiles. J. Med. Chem. 48(16):5378–5382. Bombelli, C., Giansanti, L., Luciani, P., and Mancini, G. 2009. Gemini surfactant based carriers in gene and drug delivery. Curr. Med. Chem. 16(2):171–183. Bruni, P., Cingolani, F., Iacussi, M., Pierfederici, F., and Tosi, G. 2001. The effect of bivalent metal ions on complexes DNA-liposome: A FT-IR study. J. Mol. Struct. 565–566:237–245. Bruni, P., Francescangeli, O., Marini, M., Mobbili, G., Pisani, M., and Smorlesi, A. 2011. Can neutral liposomes be considered as genetic material carriers for human gene therapy? Mini-Rev. Org. Chem. 8:38–48. Bruni, P., Gobbi, G., Morganti, G., Iacussi, M., Maurelli, E., and Tosi, G. 1997. Use and activity of metals in biological systems. I. The interaction of bivalent metal cations with double-stranded polynucleotides and phospholipids. Gazz. Chim. Ital. 127(9):513–517. Budker,V. G., Godovikov, A. A., Naumova, L. P., and Slepneva, I. A. 1980. Interaction of polynucleotides with natural and model membranes. Nucleic Acid Res. 8:2499–2515. Budker, V. G., Kazatchkov, Y. A., and Naumova, L. P. 1978. Polynucleotides adsorb on mitochondrial and model lipid membranes in the presence of bivalent cations. FEBS Lett. 95:143–146.

Structural Diversity of DNA–Phospholipid Aggregates

265

Byk, G., Wetzer, B., Frederic, M. et al. 2000. Reduction-sensitive lipopolyamines as a novel nonviral gene delivery system for modulated release of DNA with improved transgene expression. J. Med. Chem. 43:4377–4387. Camilleri, P., Kremer, A., Edwards, A. J. et al. 2000. A novel class of cationic gemini surfactants showing eficient in vitro gene transfection properties. Chem. Comm. (14):1253–1254. Cardoso, A. M. S., Faneca, H., Almeida, J. A. S. et  al. 2011. Gemini surfactant dimethylene-1,2bis(tetradecyldimethylammonium bromide)-based gene vectors: A biophysical approach to transfection eficiency. Biochim. Biophys. Acta 1808:341–351. Chen, H., Zhang, H., Thor, D., Rahinian, R., and Guo, X. 2012. Novel pH-sensitive cationic lipids with linear ortho ester linkers for gene delivery. Eur. J. Med. Chem. 52:159–172. Christianson, D. W. 1991. Sructural biology of zinc. Adv. Protein. Chem. 42:281–355. Dan, N. 1996. Formation of ordered domains in membrane-bound DNA. Biophys. J. 71:1267–1272. Dan, N. 1998. The structure of DNA complexes with cationic liposomes—cylindrical or lat bilayers? Biochim. Biophys. Acta 1369:34–38. Devínsky, F., Lacko, I., Bittererová, F., and Mlynarcˇík, D. 1987. Quaternary ammonium-salts.18. Preparation and relationship between structure, IR spectral characteristics, and antimicrobial activity of some new bis-quaternary isosters of 1,5-pentanediammonium dibromides. Chem. Papers-Chemicke Zvesti 41(6):803–814. Devínsky, F., Lacko, I., Mlynarcˇík, D., Racˇanský, V., and Krasnec, L. 1985. Relationship between critical micelle concentrations and minimum inhibitory concentrations for some non-aromatic quaternary ammonium salts and amine oxides. Tenside Deterg. 22:10–15. Donkuru, M., Wettig, S. D., Verrall, R. E., Badea, I., and Foldvari, M. 2010. Advancing nonviral gene delivery: Lipid- and surfactant-based nanoparticle design strategies. Nanomedicine 5(7):1103–1127. Donkuru, M., Wettig, S. D., Verrall, R. E., Badea, I., and Foldvari, M. 2012. Designing pH-sensitive gemini nanoparticles for non-viral gene delivery into keratinocytes. J. Mater. Chem. 22(13):6232–6244. Dubnicˇková, M., Yaradaikin, S., Lacko, I., Devínsky, F., Gordeliy, V., and Balgavý, P. 2004. Effects of gemini surfactants on model phosphatidylcholine bilayers in the luid lamellar phase. Colloids Surf. B Biointerfaces 34:161–164. Duguid, J., Bloomield, V. A., Benevides, J., and Thomas, G. J., Jr. 1993. Raman spectroscopy of DNA-metal complexes. I. Interactions and conformational effects of the divalent cations: Mg, Ca, Sr, Ba, Mn, Co, Ni, Cu, Pd, and Cd. Biophys. J. 65(5):1916–1928. Duzgunes, N., de Ilarduya, C. T., Simoes, S., Zhdanov, R. I., Konopka, K., and de Lima, M. C. P. 2003. Cationic liposomes for gene delivery: Novel cationic lipids and enhancement by proteins and peptides. Curr. Med. Chem. 10:1213–1220. Eichhorn, G. L. and Shin, Y. A. 1968. Interaction of metal ions with polynucleotides and related compounds. XII. The relative effect of various metal ions on DNA helicity. J. Am. Chem. Soc. 90:7323–7328. Espert, A., Klitzing, R. V., Poulin, P., Colin, A., and Zana, R. 1998. Behavior of soap ilms stabilized by a cationic dimeric surfactant. Langmuir 14:4251–4260. Ewert, K. K., Evans, H. M., Zidovska, A., Bouxsein, N. F., Ahmad, A., and Sainya, C. R. 2006. A columnar phase of dendritic lipid-based cationic liposome-DNA complexes for gene delivery: Hexagonally ordered cylindrical micelles embedded in a DNA honeycomb lattice. J. Am. Chem. Soc. 128:3998–4006. Felgner, P. L., Gadek, T. R., Holm, M. et  al. 1987. Lipofection: A highly eficient, lipid-mediated DNAtransfection procedure. Proc. Natl. Acad. Sci. USA 84(21):7413–7417. Felgner, P. L. and Rhodes, G. 1991. Gene therapeutics. Nature 349:351. Fisicaro, E., Compari, C., Duce, E., Donofrio, G., Rozycka-Roszak, B., and Woznyak, E. 2005. Biologically active bisquaternary ammonium chlorides: Physico-chemical properties of long chain amphiphiles and their evaluation as non-viral vectors for gene delivery. Biochim. Biophys. Acta 1722:224–233. Foldvari, M., Badea, I., Wettig, S., Verrall, R., and Bagonluri, M. 2006. Structural characterization of novel gemini non-viral DNA delivery systems for cutaneous gene therapy. J. Exp. Nanosci. 1(2):165–176. Francescangeli, O., Stanic, V., Gobbi, L. et al. 2003. Structure of self-assembled liposome–DNA–metal complexes. Phys. Rev. E 67(1):art-011904. Gershon, H., Ghirlando, R., Guttman, S. B., and Minsky, A. 1993. Mode of formation and structural features of DNA cationic liposome complexes used for transfection. Biochemistry 32:7143–7151. Ghirlando, R., Wachtel, E. J., Arad, T., and Minsky, A. 1992. DNA Packaging induced by micellar aggregates a novel in vitro DNA condensation system. Biochemistry 31:7110–7119. Gresh, N. and Šponer, J. 1999. Complexes of pentahydrated Zn2+ with guanine, adenine, and the guanine–cytosine and adenine–thymine base pairs. Structures and energies characterized by polarizable molecular mechanics and ab initio calculations. J. Phys. Chem. B 103:11415–11427. Gromelski, S. and Brezesinski, G. 2004. Adsorption of DNA to zwitterionic DMPE monolayers mediated by magnesium ions. Phys. Chem. Chem. Phys. 6:5551–5556.

266

Liposomes, Lipid Bilayers and Model Membranes

Gustafsson, J., Arvidson, G., Karlsson, G., and Almgren, M. 1995. Complexes between cationic liposomes and DNA visualized by cryo TEM. Biochim. Biophys. Acta 1235:305–312. Hafez, I. M., Maurer, N., and Cullis, P. R. 2001. On the mechanism whereby cationic lipids promote intracellular delivery of polynucleic acids. Gene Therapy 8:1188–1196. Harries, D., May, S., Gelbart, W. M., and Ben Shaul, A. 1998. Structure, stability, and thermodynamics of lamellar DNA-lipid complexes. Biophys. J. 75(1):159–173. Hauser, H. 1993. Phospholipid vesicles. In: G. Cevc (Ed.), Phospholipid Handbook. Marcel Dekker, Inc., New York, pp. 603–637. Haydon, D. A. and Myers, V. B. 1973. Surface charge, surface dipoles and membrane conductance. Biochim. Biophys. Acta 307(3):429–443. Hirata, H., Hattori, N., Ishida, M., Okabayashi, H., Frusaka, M., and Zana, R.1995. Small-angle neutron scattering study of bis(quaternary ammonium bromide) surfactant micelles in water. Effect of spacer chain length on micellar structure. J. Phys. Chem. 99:17778–17784. Hirsch-Lerner, D. and Barenholz, Y. 1998. Probing DNA-cationic lipid interactions with the luorophore trimethylammonium diphenyl-hexatriene (TMADPH). Biochim. Biophs. Acta 1370(1):17–30. Hirsch-Lerner, D., Zhang, M., Eliyahu, H., Ferrari, M. E., Wheeler, C. J., and Barenholz, Y. 2005. Effect of “helper lipid” on lipoplex electrostatics. Biochim. Biophys. Acta 1714:71–84. Horniak, L., Devínsky, F., Balgavý, P., Lacko, I., and Ebringer, L. Quaternary ammonium halides for increased eficiency of bacterial transformation. [Patent 88/3,560 (CS 269,549)]. 1990. Czechoslovakia. Hui, S. W., Langner, M., Zhao, Y. L., Ross, P., Hurley, E., and Chan, K. 1996. The role of helper lipids in cationic liposome-mediated gene transfer. Biophys. J. 71(2):590–599. Huster, D. and Arnold, K. 1998. Ca2+-mediated interaction between dextran sulfate and dimyristoyl-sn-glycero3-phosphocholine surfaces studied by H-2 nuclear magnetic resonance. Biophys. J. 75(2):909–916. Imam, T., Devínsky, F., Lacko, I., Mlynarcˇík, D., and Krasnec, L. 1983. Preparation and antimicrobial activity of some new bisquaternary ammonium salts. Pharmazie 38:308–310. Israelachvili, J. N., 1992. Intermolecular and Surface Forces. Academic, San Diego. Izumitani, Y. 1994. Cation dipole interaction in the lamellar structure of DPPC bilayers. J. Colloid Interface Sci. 166(1):143–159. Kejnovsky, E. and Kypr, J. 1998. Milimolar concentrations of zinc and other metal cations cause sedimentation of DNA. Nucleic Acids Res. 26:5295–5299. Kharakoz, D. P., Khusainova, R. S., Gorelov, A. V., and Dawson, K. A. 1999. Stoichiometry of dipalmitoylphosphatidylcholine-DNA interaction in the presence of Ca2+: A temperature-scanning ultrasonic study. FEBS Lett. 446(1):27–29. Khusainova, R. S., Dawson, K. A., Rochev, I., Gorelov, A. V., and Ivanitskii, G. R. 1999. Structural changes in DNA-Ca2+-dipalmitoylphosphatidylcholine complexes during changes in the molar ratio of nucleotide/ lipid. Microcalorimetric study. Dokl. Akad. Nauk 367(4):553–556. Kirby, A. J., Camilleri, P., Engberts, J. B. F. N. et al. 2003. Gemini surfactants: New synthetic vectors for gene transfection. Angew. Chem.- Int. Ed. 42(13):1448–1457. Koltover, I., Salditt, T., Radler, J. O., and Sainya, C. R. 1998. An inverted hexagonal phase of cationic liposome–DNA complexes related to DNA release and delivery. Science 281:78–81. Koltover, I., Salditt, T., and Sainya, C. R. 1999. Phase diagram, stability, and overcharging of lamellar cationic lipid–DNA self-assembled complexes. Biophys. J. 77:915–924. Kovalenko, D. V., Shafei, R. A., Zelenina, I. A., Semenova, M. L., Samuilova, O. V., and Zhdanov, R. I. 1996. Metallonucleoliposome complexes as a vehicle for gene delivery to mouse skeletal muscles in vivo. Genetika 32:1299–1301. Koynova, R. and Caffrey, M. 1994. Phases and phase transitions of the hydrated phosphatidylethanolamines. Chem. Phys. Lipids 69:1–34. Koynova, R. and Tenchov, B. 2009. Cationic phospholipids: Structure–transfection activity relationships. Soft Matter 5:3187–3200. Koynova, R., Wang, L., and MacDonald, R. C. 2006. An intracellular lamellar-nonlamellar phase transition rationalizes the superior performance of some cationic lipid transfection agents. Proc. Natl. Acad. Sci. USA 103:14373–14378. Koynova, R., Wang, L., Tarahovsky, Y., and MacDonald, R. C. 2005. Lipid phase control of DNA delivery. Bioconjugate Chem. 16:1335–1339. Kumar, M., Jinturkar, K., Yadav, M., and Misra, A. 2010. Gemini amphiphiles: A novel class of nonviral gene delivery vectors. Crit. Rev. Ther. Drug Carrier Syst. 27(3):237–278. Lasic, D. D., Strey, H., Stuart, M. C. A., Podgornik, R., and Frederik, P. M. 1997. The structure of DNA– liposome complexes. J. Am. Chem. Soc. 119:832–833.

Structural Diversity of DNA–Phospholipid Aggregates

267

Leal, C., Bouxsein, N. F., Ewert, K. K., and Sainya, C. R. 2010. Highly eficient gene silencing activity of siRNA embedded in a nanostructured gyroid cubic lipid matrix. J. Am. Chem. Soc. 132:16841–16847. Lengyel, A., 2010. Interaction of DNA with phospholipid liposomes in presence of divalent cations. PhD thesis Faculty of Pharmacy, Comenius University, Bratislava, Slovakia. Lengyel, A., Uhríková, D., Klacsová, M., and Balgavý, P. 2011. DNA condensation and its thermal stability inluenced by phospholipid bilayer and divalent cations. Colloids Surf. B 86:212–217. Lewis, R. N. A. H. and McElhaney, R. N. 2000. Surface charge markedly attenuates the nonlamellar phaseforming propensities of lipid bilayer membranes: Calorimetric and 31P-Nuclear Magnetic Resonance studies of mixtures of cationic, anionic, and zwitterionic lipids. Biophys. J. 79:1455–1464. Li, S. D. and Huang, L. 2006. Gene therapy progress and prospects: Non-viral gene therapy by systemic delivery. Gene Therapy 13:1313–1319. Lin, A. J., Slack, N. L., Ahmad, A., George, C. X., Samuel, Ch. E., and Sainya, C. R. 2003. Three-dimensional imaging of lipid gene-carriers: Membrane charge density controls universal transfection behavior in lamellar cationic liposome–DNA complexes. Biophys. J. 84:3307–3316. Lu, J. R., Thomas, R. K., Binks, B. P., Fletcher, P. D. I., and Penfold, J. 1995. Structure and composition of dodecane layers spread on aqueous-solutions of dodecyl-and hexadecyltrimethylammonium bromides studied by neutron relection. J. Phys. Chem. 99(12):4113–4123. Luck, G. and Zimmer, C. 1972. Conformational aspects and reactivity of DNA. Effects of manganese and magnesium ions on interaction with DNA. Eur. J. Biochem. 29:528–536. Macdonald, P. M., Crowell, K. J., Franzin, C. M., Mitrakos, P., and Semchyschyn, D. 2000. H-2 NMR and polyelectrolyte-induced domains in lipid bilayers. Solid State Nucl. Magn. Reson. 16(1–2):21–36. Madhusudhana Rao, N. and Gopal, V. 2006. Cationic lipids for gene delivery in vitro and in vivo. Expert Opin. Ther. Patents 16:825–844. May, S. and Ben Shaul, A. 1997. DNA-lipid complexes: Stability of honeycomb-like and spaghetti-like structures. Biophys. J. 73(5):2427–2440. May, S. and Ben Shaul, A. 2004. Modeling of cationic lipid–DNA complexes. Curr. Med. Chem. 11(2):151–167. May, S., Harries, D., and Ben Shaul, A. 2000. The phase behavior of cationic lipid–DNA complexes. Biophys. J. 78(4):1681–1697. McManus, J., Radler, J. O., and Dawson, K. A. 2003. Phase behaviour of DPPC in a DNA–calcium–zwitterionic lipid complex studied by small angle x-ray scattering. Langmuir 19:9630–9637. McManus, J. J., Radler, J. O., and Dawson, K. A. 2004. Observation of a rectangular columnar phase in a DNA–calcium–zwitterionic lipid complex. J. Am. Chem. Soc. 126:15966–15967. Mengistu, D. H., Bohinc, K., and May, S. 2009. Binding of DNA to zwitterionic lipid layers mediated by divalent cations. J. Phys. Chem. B 113:12277–12282. Miguel, M. G., Pais, A. A. C. C., Dias, R., Leal, C., Rosa, M., and Lindman, B. 2003. DNA-cationic amphiphile interactions. Colloids Surf. A. Physicochem. Eng. Asp. 228:43–55. Miller, A. D. 1998. Cationic liposomes for gene therapy. Angew. Chem.-Int. Ed. 37(13–14):1769–1785. Miller, A. D. 2003. The problem with cationic liposome/micelle-based non-viral vector systems for gene therapy. Curr. Med. Chem. 10:1195–1211. Pabst, G., Hodzic, A., Štrancar, J., Danner, S., Rappolt, M., and Laggner, P. 2007. Rigidiication of neutral lipid bilayers in the presence of salts. Biophys. J. 93:2688–2696. Pachuk, C. J., Ciccarelli, R. B., Samuel, M. et al. 2000. Characterization of a new class of DNA delivery complexes formed by the local anesthetic bupivacaine. Biochim. Biophys. Acta 1468(1–2):20–30. Patil, S. D., Rhodes, D. G., and Burgess, D. J. 2004. Anionic liposomal delivery system for DNA transfection. The AAPS J. 6(4):e29. Pisani, M., Bruni, P., Caracciolo, G., Caminiti, R., and Francescangeli, O. 2006. Structure and phase behavior of self-assembled DPPC-DNA-metal cation complexes. J. Phys. Chem. B 110(26):13203–13211. Pisárcˇik, M., Rosen, M. J., Polakovicˇová, M., Devínsky, F., and Lacko, I. 2005. Area per surfactant molecule values of gemini surfactants at the liquid-hydrophobic solid interface. J. Colloid Interface Sci. 289:560–565. Podgornik, R., Rau, D. C., and Parsegian,V. A. 1989. The action of interhelical forces on the organization of DNA double helices: Fluctuation-enhanced decay of electrostatic double-layer and hydration forces. Macromolecules 22(4):1780–1786. Pullmannová, P. 2011. DNA interaction with cationic liposomes. PhD dissertation, Faculty of Chemical and Food Technology, Slovak University of Technology in Bratislava, Slovakia. Pullmannová, P., Bastos, M., Bai, G. et  al. 2012a. The ionic strength effect on the DNA complexation by DOPC-gemini surfactants liposomes. Biophys. Chem. 160(1):35–45. Pullmannová, P., Funari, S. S., Devínsky, F., and Uhríková, D. 2012b. The DNA–DNA spacing in gemini surfactants–DOPE–DNA complexes. Biochim. Biophys. Acta 1818:2725–2731.

268

Liposomes, Lipid Bilayers and Model Membranes

Radler, J. O., Koltover, I., Salditt, T., and Sainya, C. R. 1997. Structure of DNA-cationic liposome complexes: DNA intercalation in multilamellar membranes in distinct interhelical packing regimes. Science 275:810–814. Rajnohová, Z., Lengyel, A., Funari S. S., Uhríková, D. 2010. The structure and binding capacity of lipoplexes. Acta Facultatis Pharm. Univ. Comenianae 57:77–89. Rolland, A. P. 1998. From genes to gene medicines: Recent advances in nonviral gene delivery. Crit. Rev. Ther. Drug Carrier Syst. 15:143–198. Rosada, R. S., Silva, C. L., Andrade Santana, M. H., Nakaie, C. R., and de la Torre, L. G. 2012. Effectiveness, against tuberculosis, of pseudo-ternary complexes: Peptide-DNA-cationic liposome. J. Colloid Interface Sci. 373:102–109. Sainya, C. R., Sirota, E. B., Roux, D., and Smith, G. S. 1989. Universality in interacting membranes: The effect of cosurfactants on the interfacial rigidity. Phys. Rev. Lett. 62:1134–1137. Sato, Y., Kumazawa, N., Yoshikawa, K., and Kurusu, Y. 2005. Transformation of Escherichia coli mediated by neutral phospholipids. Biosci. Biotechnol. Biochem. 69:235–237. Seddon, J. M. 1990. Structure of the inverted hexagonal (HII) phase, and non-lamellar phase transitions of lipids. Biochim. Biophys. Acta 1031:1–69. Shepherd, J. C. and Buldt, G. 1978. Zwitterionic dipoles as a dielectric probe for investigating head group mobility in phospholipid membranes. Biochim. Biophys. Acta 514:83–94. Shirazi, R. S., Ewert, K. K., Leal, C., Majzoub, R. N., Bouxsein, N. F., and Sainya, C. R. 2011. Synthesis and characterization of degradable multivalent cationic lipids with disulide-bond spacers for gene delivery. Biochim. Biophys. Acta 1808:2156–2166. Srinivasan, Ch. and Burgess, D. J. 2009. Optimization and characterization of anionic lipoplexes for gene delivery. J. Control. Release 136(1):62–70. Sternberg, B., Sorgi, F. L., and Huang, L. 1994. New structures in complex formation between DNA and cationic liposomes visualized by freeze-fracture electron microscopy. FEBS Lett. 356:361–366. Stumpel, J., Eibl, H. J., and Nicksch, A. 1983. X-ray analysis and calorimetry on phosphatidylcholine model membranes. Biochim. Biophys. Acta 727:246–254. Tarahovsky, Y. S., Khusainova, R. S., Gorelov, A. V., Dawson, K. A., and Ivanitskij, G. R. 1996. Effect of DNA on the ultrastructural organization of lecithin liposomes in the presence of calcium cations. Dokl. Akad. Nauk 350(3):411–413. Tate, M. W. and Gruner, S. M. 1989. Temperature dependence of the structural dimensions of the inverted hexagonal phase of PE-containing membranes. Biochemistry 28:4245–4253. Toombes, G. E. S., Finnefrock, A. C., Tate, M. W., and Gruner, S. M. 2002. Determination of L–HII phase transition temperature for 1,2-dioleoyl-sn-glycero-3-phosphatidylethanolamine. Biophys. J. 82:2504–2510. Uhríková, D., Hanulová, M., Funari, S. S., Khusainova, R. S., Šeršeň, F., and Balgavý, P. 2005a. The structure of DNA–DOPC aggregates formed in presence of calcium and magnesium ions: A small angle synchrotron X-ray diffraction study. Biochim. Biophys. Acta 1713:15–28. Uhríková, D., Hanulová, M., Funari, S. S., Lacko, I., Devínsky, F., and Balgavý, P. 2004a. The structure of DNA-DLPC-cationic gemini surfactant aggregates: A small angle synchrotron X-ray diffraction study. Biophys. Chem. 111(3):197–204. Uhríková, D., Kovářová, M., and Balgavý, P. 2004b. Interaction of DNA with egg yolk phosphatidylcholine liposomes in the presence of Mg2+ ions: A luorescence study. Acta Facultatis Pharm. Univ. Comenianae 51:217–226. Uhríková, D., Kucˇerka, N., Lengyel, A. et  al. 2012a. Lipid bilayer—DNA interaction mediated by divalent metal cations: SANS and SAXD study. J. Phys. Conf. Ser. 351:012011–9. Uhríková, D., Lengyel, A., Hanulová, M., Funari, S. S., and Balgavý, P. 2007. The structural diversity of DNAneutral phospholipids-divalent metal cations aggregates: A small-angle synchrotron X-ray diffraction study. Eur. Biophys. J. 36(4–5):363–375. Uhríková, D., Pullmannová, P., Bastos, M., Funari, S. S., and Teixeira, J. 2009a. Interaction of short-fragmented DNA with dipalmitoylphosphatidylcholine bilayers in presence of zinc. Gen. Physiol. Biophys. 28(2):146–159. Uhríková, D., Pullmannová, P., Kucˇerka, N., Funari, S. S., Teixeira, J., and Balgavý, P. 2009b. The structural variety of DNA-DPPC-divalent metal cation aggregates: SAXD and SANS study. Eur. Phys. J. Special Top. 167:191–197. Uhríková, D., Pullmannová, P., Šabíková, A., Devínsky, F., and Funari, S. S. 2012b. The microstructure of DNA-egg yolk phosphatidylcholine gemini surfactants complexes: Effect of the spacer length. Drug Metab. Drug Interact. 27:47–54.

Structural Diversity of DNA–Phospholipid Aggregates

269

Uhríková, D., Rapp, G., and Balgavý, P. 2001. Condensation of DNA and phosphatidylcholine bilayers induced by Mg(II) ions—A synchrotron X-ray diffraction study. In: M. Melník and A. Sirota (Eds), Challenges for Coordination Chemistry in the New Century. Slovak Technical University Press, Bratislava, pp. 219–224. Uhríková, D., Rapp, G., and Balgavý, P. 2002. Condensed lamellar phase in ternary DNA–DLPC–cationic gemini surfactant system: A small-angle synchrotron X-ray diffraction study. Bioelectrochemistry 58:87–95. Uhríková, D., Zajac, I., Dubnicˇková, M. et  al. 2005b. Interaction of gemini surfactants butane-1,4-diyl-bis (alkyldimethylammonium bromide) with DNA. Coll. Surf. B 42(1):59–68. Vojcˇíková, L. and Balgavý, P. 1988. Interaction of DNA with dipalmitoylphosphatidylcholine model membranes: A microcalorimetric study. Stud. Biophys. 125:5–10. Vojcˇíková, L., Švajdlenka, E., and Balgavý, P. 1989. Spin label and microcalorimetric studies of the interaction of DNA with unilamellar phosphatidylcholine liposomes. Gen. Physiol. Biophys. 8(4):399–406. Wagner, K., Harries, D., May, S., Kahl, V., Radler, J. O., and Ben Shaul, A. 2000. Direct evidence for counterion release upon cationic lipid–DNA condensation. Langmuir 16(2):303–306. Wettig, S. D., Badea, I., Donkuru, M., Verrall, R. E., and Foldvari, M. 2007. Structural and transfection properties of amine-substituted gemini surfactant-based nanoparticles. J. Gene Med. 9(8):649–658. Wetzer, B., Byk, G., Frederic, M. et  al. 2001. Reducible cationic lipids for gene transfer. Biochem. J. 356:747–756. Williams, R. J. P. 1988. An introduction in biochemistry of zinc. In: Zinc in Human Biology. Springer-Verlag, London, UK, pp. 15–31. Zabner, J., Fasbender, A. J., Moninger, T., Poellinger, K. A., and Welsh, M. J. 1995. Cellular and molecular barriers to gene transfer by a cationic lipid. J. Biol. Chem. 270:18997–19007. Zhdanov, R. I., Podobed, O. V., Buneeva, O. A., Kutsenco, N. G., Tsvetkova, T. A., and Lavrenova, T. P. 1997. Gene transfer into eukaryotic cells using non-cationic (neutral) liposomes. Vop. Meditsinskoj Khimii 43:212–216. Zuhorn, I. S., Engberts, J. B. F. N., and Hoekstra, D. 2007. Gene delivery by cationic lipid vectors: Overcoming cellular barriers. Eur. Biophys. J. 36:349–362. Zuidam, N. J. and Barenholz, Y. 1997. Electrostatic parameters of cationic liposomes commonly used for gene delivery as determined by 4-heptadecyl-7-hydroxycoumarin. Biochim. Biophys. Acta 1329(2):211–222. Zuidam, N. J. and Barenholz, Y. 1998. Electrostatic and structural properties of complexes involving plasmid DNA and cationic lipids commonly used for gene delivery. Biochim. Biophys. Acta 1368(1):115–128.

14

An Update on Active Membranes David Lacoste and Patricia Bassereau

CONTENTS 14.1 Introduction .......................................................................................................................... 271 14.2 Experiments on Active Membranes...................................................................................... 272 14.2.1 Reconstituted Systems Containing Activable Ion Pumps ........................................ 272 14.2.2 Red Blood Cells ........................................................................................................ 274 14.2.3 Nonequilibrium Membranes in the Presence of Lipid Fluxes .................................. 277 14.3 Theoretical Models ............................................................................................................... 277 14.3.1 Hydrodynamic Models ............................................................................................. 278 14.3.2 Electrokinetic Models ............................................................................................... 281 14.4 Conclusion ............................................................................................................................284 Acknowledgments.......................................................................................................................... 285 References ...................................................................................................................................... 285

14.1 INTRODUCTION Bilayer membranes formed from phospholipid molecules are an essential component of cellular membranes. While the properties of equilibrium membranes are well understood using the Helfrich Hamiltonian constructed from the two elastic moduli, the surface tension and the curvature modulus (for a review, see, for instance, Seifert, 1997), the nonequilibrium behavior of real biological membranes of living cells has not reached an equal level of understanding. In this chapter, we are interested in active membranes, which share one of the important characteristics of membranes of living cells: the sustained uptake and dissipation of energy. Active membranes can be thought of as composite systems, which contain some active elements in addition to a membrane. These active elements exert either directly or in an effective way, nonconservative forces on the membrane surface. These forces are generalized in the sense that they may not be necessarily mechanical in origin: they can be produced by a chemical reaction such as actin polymerization, mechanical stresses, or by electric ields for instance. A common feature in active membranes is that some kind of energy input has to be brought from a source that is exterior to the membrane itself. This energy input (brought by adenosine triphosphate (ATP) hydrolysis, or light for instance) is precisely what drives the membrane out of equilibrium. As a result, the luctuations of an active membrane are different from the thermal luctuations of passive, equilibrium membranes. Classical models of active membranes have been reviewed in Ramaswamy and Rao (2001), so that, in this chapter, we can focus on more recent developments in the ield. Over the years, active membranes have served as a test ground for theoretical models of more complex active systems, such as active luids or active gels (Kruse et al., 2005) and there has been a strong feedback between experiments and theoretical models. These days, there are many branches of research on active luids, which focus, for instance, on the collective motions of active particles, or on speciic rheological properties. As a proof of this tendency, we would like to recommend three recent

271

272

Liposomes, Lipid Bilayers and Model Membranes

reviews on active matter systems, written by researchers who originally made key contributions in the ield of active membranes (Ramaswamy, 2010; Menon, 2010; Marchetti et al., 2013). Nerve cells are a particularly interesting example of active membranes because the chemical energy fed into the system is used to transmit information, in the form of electrical signals (action potentials) (Hille, 2001). According to the standard Hodgkin–Huxley model, this propagation is achieved by the voltage-dependent opening and closing of ion channels, but the speciicity of the membrane physical properties does not play any role. In view of this, the Hodgkin–Huxley model has been challenged in particular by the group of T. Heimburg, who suggested that some lipid phase transition occurs in the membrane in the course of the action potential. This phase transition makes the membrane more conductive to ions and is thus relevant even in the absence of ion channels in the membrane (Heimburg, 2010). At this time of writing, these ideas are not completely accepted by the community of biologists and electrophysiologists, and the role of the membrane composition in the action potential is still not established. In this chapter, we focus mainly on active membranes, deined as membranes that contain active inclusions in the form of ion channels or pumps. These elements are called active because they are able to transport ions from one side of the membrane to the other in a selective way when a source of energy (light, ATP, etc.) is provided. The ion transport itself is called active because it requires an energy source, which can be the hydrolysis of ATP or light for instance. This energy source is required to induce protein conformational changes and consequently ion transfer, in contrast to passive transport, which does not require energy input. The main purpose of this chapter is to review a number of recent experimental and theoretical reports in the ield of active membranes. The chapter is organized as follows: in Section 14.2, we present some experiments on active membranes, while in Section 14.3, we review a number of different theoretical works on active membranes.

14.2 14.2.1

EXPERIMENTS ON ACTIVE MEMBRANES reconstItuted systeMs contaInIng actIvable Ion puMps

Biomimetic active membranes have been experimentally prepared and studied mainly at the Curie Institute (Paris), in the form of giant unilamellar vesicles (GUV) containing ion pumps such as bacteriorhodopsin (BR) (Figure 14.1), or calcium ATPase (Manneville et  al., 1999, 2001; Girard et al., 2005; El Alaoui Faris et al., 2009). BR is a light-activable proton pump, which can switch from a passive to active state when receiving light of the appropriate wavelength. It can then transfer protons through a lipid bilayer. This system is particularly interesting as it allows to compare the active and passive behaviors of the same system. Calcium ATPase is a calcium pump that uses the energy of ATP hydrolysis for performing a conformational change and allows for instance to build up Ca2+ gradients across the membrane of muscle cells. In this case, experiments in the presence or in the absence of ATP are compared. Not many other in vitro active membrane experiments have been performed so far, probably because it requires to reconstitute ion channels or ion pumps in Actinic light: active BR H+ Red light: passive

FIGURE 14.1 (See color insert.) Example of active membranes. GUV contains light-activable proton pumps. (BR, bacteriorhodopsin).

273

An Update on Active Membranes

model membrane systems and to keep them functional. Different methods have been established for reconstituting transmembrane proteins in GUVs (Kahya et al., 2001; Girard et al., 2004b; Doeven et al., 2005; Aimon et al., 2011) that preserve protein activity, but sample preparation still remains a milestone of such experiments. They are usually based on the reconstitution of the proteins in small liposomes as a preliminary step, followed by gentle electroformation of the partially dried ilm formed by the proteoliposomes. Alternative methods may have to be found so that the puriication and the reconstitution steps could be avoided in the future. Indeed, GUVs have now been obtained directly from puriied native membranes (Méléard et al., 2009) or from detachment of blebs from plasma membranes of cells (plasma membrane spheres [PMS] or giant plasma membrane vesicle [GPM]) (Sezgin et al., 2012). Vesicles containing active pumps have been studied using the microaspiration technique (Manneville et al., 1999, 2001; Girard et al., 2005) developed by E. Evans (Evans and Rawicz, 1990) (Figure 14.2a). In thermal equilibrium, the following relation exists between the excess area (measured from the length of the aspirated tongue in the pipette) Δα and the applied tension (deduced from the aspiration pressure) ∆α =

A0 − Ap k T σ  = B ln   , 8πκ  σ 0  A0

(14.1)

(a) 10 μm

(b) 5

Dp

Dv

In (σ)

ΔP

4 Passive Active

3 2 1

Lp

Te – T T

(c)

2

3

4

5

α (%) 2.0 1.5 1.0 0.5 0

0

0.1 0.2 0.3 – Normalized protein density φ

FIGURE 14.2 Micropipette aspiration experiments on active GUVs. (a) Differential interference contrast (DIC) image of a GUV of diameter D v aspirated in a micropipette of diameter D p. The membrane tension can be deduced from the difference of pressure ΔP, D v and, D p, the corresponding excess area from the length of the tongue in the micropipette Lp. (b) When the BR is activated by actinic light (empty symbols), the variation of the logarithm of the membrane tension with excess area remains linear as compared to the same GUV in a passive state illuminated with red light. However, the slope is lowered by a factor on the order of 2. This has been interpreted as an effective temperature larger than the thermal temperature. (From Manneville, J.-B. et al. 1999. Activity of transmembrane proteins induces magniication of shape luctuations of lipid membranes. Phys. Rev. Lett. 82 (21), 4356–4359. Copyright 1999 by The American Physical Society.) (c) Relative evolution of the effective temperature T e of an ATP-activated GUV containing Ca 2+ – ATPase as a function of the normalized protein density in the membrane. (Reprinted with permission from Girard, P., J. Prost, and P. Bassereau. 2005. Passive or active luctuations in membranes containing proteins. Phys. Rev. Lett. 94, 088102. Copyright 2005 by The American Physical Society.)

274

Liposomes, Lipid Bilayers and Model Membranes

where A0 is the optically measured area at the minimal tension σ0 allowing to grab the vesicle, and Ap is the projected area for a tension σ. For passive GUVs, this relation—for tensions typically below 10−5 N/m—allows to measure the bending rigidity of the membrane κ. Experiments have been performed on the same vesicle in the passive and active state for BR, and on a statistically relevant number of GUVs in the absence and in the presence of ATP for the Ca–ATPase. As a result, a relative excess area/stress relation, formally identical to Equation 14.1, has been found in the low-tension regime, but with a renormalized prefactor. The effect of protein activity has been described by assigning an effective temperature Te to the active membrane, but it could also have been interpreted as an effective bending rigidity of the membrane. In both cases, an ampliication of the GUV luctuations has been observed when the proteins are activated, corresponding to a reduction of the slope of the log (tension) versus excess area plots. The effect of protein activity is rather strong and an effective temperature of twice the room temperature has been found for BR (Figure 14.2b) (Manneville et al., 1999, 2001), and is even much stronger, of the order of 1000 K, for Ca– ATPase (Girard et al., 2005). The nonequilibrium nature of the effect was further demonstrated by showing that this effective temperature depends on the viscosity of the bulk medium (it decreases as the viscosity of the bulk medium increases), an effect that does not occur in membranes at thermodynamical equilibrium. The ampliication of luctuations is directly related to the density of active proteins in the membrane (Figure 14.2c) (Girard et al., 2005). Videomicroscopy has been used more recently (El Alaoui Faris et al., 2009) as a different technique to analyze the luctuation spectrum of BR membranes without applying external tension to the vesicle. With this technique, the contour of freely loating GUVs was detected using phase contrast microscopy with a resolution Fd3m (above 1.7) > H2 (~1.7) > V2 (values

Drug Formulations Based on Self-Assembled Liquid Crystalline Nanostructures

345

range from 1.27 to 1.31) > L α (~1) (Larsson, 1989, Hyde et al. 1997). Thus, the phase behavior of monoglyceride-based systems under the inluence of a hydrophobic additive and/or changing temperature or pressure can be attributed to a corresponding change in the CPP values. This explains for instance the tendency upon heating or the addition of a hydrophobic agent to form monolayers with negative spontaneous curvature (H0 < 0). Especially the discontinuous phase (H2, Fd3m, and L2) regions can be signiicantly extended in the investigated phase diagrams (de Campo et al. 2004, Yaghmur et al. 2005, 2006a, 2010a, 2012a) as a consequence of solubilizing hydrophobic guest molecules due to the signiicant reduction in water uptake in these nonplanar phases (expelling water) (de Campo et al. 2004, Yaghmur et al. 2005, 2012a). It is worth noting that increasing temperature reduces the a 0 value due to the headgroup’s dehydration and simultaneously enhances the vs value (de Campo et  al. 2004, Yaghmur et  al. 2005). In other words, increasing temperature induces a monotonously decrease in the monolayer thickness and the water core radii. However, the CPP concept alone is not fully explaining the tendency to form the discontinuous H2, cubic Fd3m, and L2 phases in the presence of hydrophobic additive. Various studies suggest the need to overcome the main energy barrier for the formation of these phases by solubilizing a hydrophobic additive (oil, hydrophobic drug, or surfactant-like peptide) (Engström and Engström 1992, Shearman et al. 2006, Yaghmur et al. 2007, 2010a, Mares et al. 2008, Yaghmur et al. 2011a, 2012a). These guest materials are of assistance to ill out the interstitial regions in the hydrophobic matrices of the self-assembled systems and to release the packing frustrations. It is also possible by changing the lipid composition to have a counter effect to that of solubilizing oil (Yaghmur et al. 2006a). This can be achieved for instance by mixing a bilayer-forming surfactant such as diglycerol monooleate (DGMO) or dioleoyl-phosphatidylglycerol (DOPG) with a lipid favoring the formation of nonlamellar phases (such as MO) (Yaghmur et  al. 2006a). The partial replacement of MO by either DGMO or DOPG makes the mean curvature of the ilm layer less negative and, therefore, it induces at high content of DOPG or DGMO nonlamellar-to-lamellar phase transitions (Yaghmur et al. 2006a, 2008b, 2011b, Awad et al. 2005). These tunable and nanostructured self-assembled media are potentially safe with high interfacial areas and, therefore, have the potential to be used as reservoirs for the solubilization of hydrophobic, hydrophilic, and amphiphilic drugs (Engström et al. 1999, Shah et al. 2001, Shah and Paradkar 2007, Fong et al. 2009, Costa-Balogh et al. 2010, Phan 2011, Yaghmur et al. 2012a). Figure 17.3 illustrates

FIGURE 17.3 An example on the possible localization of a solubilized drug in an inverted-type micellar solution. (a) Hydrophilic drugs localize in the hydrophilic core of the inverted micelles. (b) Hydrophobic drugs are preferentially localized in the hydrophobic continuous domains. (c) Surface-active (amphiphilic) drugs incorporate into the water–lipid interfacial area.

346

Liposomes, Lipid Bilayers and Model Membranes

the potential localization of different drug candidates in inverted-type (L2) micellar solution. The three compartments for the accommodation of solubilized drug in the L2 phase are as following: (a) the internal core (or the water pool), (b) the external (hydrophobic continuous) phase, and (c) the interfacial ilm.

17.3 DYNAMICS OF LIPIDIC STRUCTURAL TRANSITIONS Energetic theory models for the self-assembly of surfactant-like lipids into inverted-type LLC phases or micellar solutions work on the basis that each lipid molecule tends to adopt its inherent shape, that is, the demand of each monolayer to bend according to its spontaneous curvature with a simultaneous need to pack in this monolayer at a constant density (Kozlov et al. 1994, Templer, 1998). This means that the driving force for the formation of nonlamellar phases is given by the bending energy and the energy barrier by possible packing frustrations (Sadoc and Charvolin 1986, Kozlov et  al. 1994). In the ideal stress-free planar bilayer, the lateral repulsive pressures of the hydrophobic chain and the headgroup are perfectly counterbalanced by the interfacial pressure. In this case, the monolayer has a zero spontaneous curvature. However, the lateral monolayer forces generally are not well balanced and hence the monolayer tends to curve. In principle, starting from the L α phase, two different transition pathways leading to the formation of nonlamellar phases are documented in the literature (Figure 17.4) (Rappolt 2006, Shearman et al. 2006, Yaghmur et al. 2008b). The irst pathway describes the classical fusion steps of two opposed membranes and bases on the formation of point defects (Hui 1981): after adhesion of two vesicles (A), a stalk intermediate forms (B), which then converts into an extended area of intermembrane contact (C) from which a pore evolves (D). Finally, the creation of numerous pores is assumed to condense into the bicontinuous primitive cubic (CP, Im3m) phase (Siegel and Banschbach 1990). Often this cubic phase further transforms under increasing temperature into the diamond cubic (CD, Pn3m) phase (Yaghmur et al. 2008a). Ideally, the averaged Gaussian curvatures of the involved cubic phases should remain the same (at least under near equilibrium conditions) and hence the unit-cell parameter should scale exactly with the Bonnet ratio aIm3m /aPn3m of 1.279 (Hyde, 1996). A reasonable phase transition mechanism has been irst proposed by Sadoc and Charvolin (1989), in which two four-fold nodes of the CD phase are obtained by deforming (stretching) one six-fold node of the CP -phase into a connection rod (see Figure 17.4, bottom). Finally, the transition from the CD -phase to the H2 phase takes place with enhancing further the negative spontaneous curvature of the monolayer. This transformation is not well understood at the molecular level, but has been observed in various lipid/water systems (Seddon, 1990, Shearman et al. 2006). In contrast, the second principle pathway bases on the formation of line-defects. The different schemes concerning the rod formation between two opposed bilayers have been deduced from experimental data (Rappolt et al. 2008): starting from two bilayers in close contact (E), a line defect spontaneously forms between them releasing the inherent spontaneous monolayer curvature (F) (Figure 17.4). Here, the lipid molecules are allowed to adopt their intrinsic shape, that is, to splay their chains and shorten accordingly. Next, the deiciency of water content in this line defect is adjusted (G). Last, a irst inverted lipid tube pinches off (H) inducing six nearest-neighbor line defects (*), and thus, further locations for a cooperative rod formation are given. So far, we have outlined the classical lamellar to nonlamellar phase transitions, but disregarding one important point, that is, the nonequilibrium phenomena. The transition pathways in LLC systems far from equilibrium can be investigated for instance with temperature- and pressure-jump techniques (Laggner et  al. 2005, Winter and Jeworrek 2009) using high brilliance synchrotron light sources (for recent reviews, see Angelova et  al. 2012, Yaghmur and Rappolt 2012). This kind of investigations is of outmost importance for life science studies and pharmaceutical applications, since biological reactions and also drug administration processes are usually taking place under nonequilibrium conditions. Phase heterogeneity (Rappolt et al. 2000) or the formation of intermediate phases (Laggner and Kriechbaum 1991, Squires et al. 2005) are often obscured near equilibrium, but become accessible under extreme temperature or pressure rates. Different reaction pathways under nonequilibrium conditions are

Drug Formulations Based on Self-Assembled Liquid Crystalline Nanostructures Point defect route

347

Line defect route

(a)

(e)

(b)

(f)

(c)

(g)

(d)

(h)

*

*

*

* *

*

H2-phase

Im3m→Pn3m (V2-phases)

FIGURE 17.4 (See color insert.) Two schematics of the proposed pathways from the bilayer to the inverted monolayer tube transition. On the left hand side, the classical vesicle fusion route is depicted. The formation of pores is widely believed to be the prerequisite for the formation of bicontinuous cubic (V2) nanostructures (Shearman et al. 2006), which upon further curvature frustration may transform into self-assembled monolayer tubes (H2 phase). On the right hand side, the direct formation of an inverse lipid nanotube between two opposed bilayers is illustrated. For better understanding of the structural conversions, the headgroups of opposed monolayers are shown in light blue whereas the rest are depicted in blue.

caused by the different time scales of the involved processes, for example, for water diffusion, ion binding, chain melting, and so forth. For instance, heating slowly through the main transition of a binary lipid/water system, the water diffusion and the lipid chain melting processes are practically taking place at the same time. However, the slower water diffusion gets decoupled from the chain melting event, when applying rapid-temperature perturbations. In Figure 17.5a, the optothermally induced structural changes in multi-lamellar vesicles of N-methylated dioleoylphosphatidylethanolamine (DOPE-Me) loaded with hydrophilic gold nanoparticles having a size of 4 nm are shown. The structural conversion from vesicles with luid lamellar (L α) membranes to H2 phase through an intermediate state of uncorrelated luid bilayers during in situ UV activation was observed. It should be noted that such a disordered lamellar phase did not occur under near equilibrium conditions. A detailed description of the molecular rearrangements during the transition from the well-ordered multi-lamellar vesicles in the luid L α phase to the H2 phase via the formation of the disordered intermediate phase was put forward in our recent study (Yaghmur et al. 2010b). In Figure 17.5b, an example for a pressure-induced phase transition is

348

Liposomes, Lipid Bilayers and Model Membranes

(a)

(b)

Scattering angle

Intensity (a.u.)

Low High

e-density (a.u.)

100 Å

*

1

Time

(c)

1200 bar, H2 1100 bar, H2 1000 bar, H2 900 bar, Fd3m plus H2 800 bar, Fd3m 700 bar, Fd3m 600 bar, Fd3m 500 bar, Fd3m 400 bar, Fd3m 300 bar, Fd3m 200 bar, Fd3m 103 bar, Fd3m 1 bar, Fd3m

2 q (nm–1)

3

(d)

Ca2+

Capillary X-ray beam

+H2O

Sponge phase

FIGURE 17.5 (See color insert.) (a) Optothermically induced structural changes in multi-lamellar vesicles (MLVs) loaded with hydrophilic gold nanoparticles (NPs). Synchrotron time-resolved SAXS experiments combined with a UV light source irradiation demonstrated that the structure pathway from the luid lamellar (L α) phase to an inverted hexagonal (H2) phase passes through an intermediate state of uncorrelated membranes. The electron density proile of the L α phase is shown at the far left, and the electron density map of the H2 phase is shown on the far right. For the H2 phase, the electron density values are colour coded. (The igure was adapted from Yaghmur, A. et al. 2010b. J. Phys. Chem. Lett. 1(6), 962–966.) (b) SAXS patterns of the fully hydrated tetradecane-loaded nondispersed MO system at different pressures. (Data taken with permission from Yaghmur, A. et al. 2010a. Langmuir 26(2), 1177–1185.) At 20°C in a pressure range from 1 to 1200 bar, the SAXS patterns reveal the structural transitions from the Fd3m via a biphasic system of Fd3m coexisting with H2 to the H2 phase. (c) Scheme of a stop-low mixing of set-up. The sponge phase to cubic Pn3m phase transition was triggered with the rapid addition of calcium ions. (Adapted from Yaghmur, A., Sartori, B., and Rappolt, M. 2011b. Phys. Chem. Chem. Phys. 13(8), 3115–3125.) (d) Electron density reconstruction of cubic gyroid and cubic diamond surfaces (left side) and their corresponding two interwoven water networks (right side) are presented. The transition was induced by the controlled addition of water. The maximum electron density maps are displayed in viewing direction perpendicular to the (220)-planes (G-phase) and perpendicular to the (111)-planes (D-phase), respectively. (Adapted from Rappolt, M. et al. 2006. Europhys. Lett. 75(2), 267–273.)

given. Increasing pressure actually decreases the membrane curvature. It should be noted that the discontinuous cubic Fd3m phase of the investigated fully hydrated sample consisting of the binary MO/tetradecane mixture at a weight ratio of 70/30 transforms at 20°C into the H2 phase (Yaghmur et al. 2010a), which exhibits a lower mean membrane curvature as compared with the micellar cubic phase. Thus, pressure causes the opposite effect of temperature, that is augmenting pressure decreases the chain volume as well as it increases the molecular area per lipid at the polar/apolar interface.

Drug Formulations Based on Self-Assembled Liquid Crystalline Nanostructures

349

While dynamic temperature and pressure SAXS experiments help to characterize the biophysical aspects of model membrane systems, various set-ups designed for studying chemical reactions and material transfer in both dispersed and nondispersed LLC systems are especially important for a real-time monitoring of the structural changes during drug-delivery processes. Among these SAXS set-ups, the stop-low experiments performed by coupling SAXS to a stop-low apparatus or a microluidic device allow (i) investigating the effect of mixing of dispersions with solutions on the self-assembled nanostructures, (ii) probing the effect of environmental changes, and (iii) monitoring the inluence of chemical reactions on the stability of the investigated self-assembled systems (Yaghmur and Rappolt 2012). In the given example in Figure 17.5c, a transition from a sponge phase (approximately described as a swollen and disordered bicontinuous phase) to a well-ordered cubic Pn3m phase was triggered by rapid mixing with a low concentration of calcium ions. It is important to remark that the fast binding of Ca2+ ions to the negatively charged DOPG/MO membrane induced rapid dehydration and concomitant condensation of the phospholipid bilayer membrane due to the screened electrostatic repulsive forces (Yaghmur et al. 2008b, 2011b). Fast injection/immersion SAXS experiments enable monitoring the evolution of the selfassembled nanostructures in situ during the hydration processes in viscous liquid crystalline systems (Yaghmur et al. 2011a, Yaghmur and Rappolt 2012). In Figure 17.5d, the gyroid cubic (CG, Ia3d symmetry) to the diamond (CD, Pn3m symmetry) phase transition in the fully hydrated MO system under nonequilibrium conditions is illustrated. This particular cubic-cubic transformation was induced by fast hydration of the initial cubic CG -phase (Rappolt et al. 2006). As also found in rapid pressure-jump experiments on this system (Squires et al. 2005), an intermediate, likely an H2 phase, was observed. On the left-hand side of Figure 17.5d, electron density maps displaying minimal surfaces of the CG and CD phases are shown, while on the right-hand side the corresponding water channel networks are depicted (maximum density). The maps are displayed in viewing direction perpendicular to the (220) planes (CG phase) and perpendicular to the (111) planes (CD phase), respectively, which are believed to form an epitaxial relationship (Rappolt et al. 2006).

17.4 DISPERSED LIPIDIC PARTICLES ENVELOPING NANOSTRUCTURES An important milestone on dispersing viscous lyotropic nonlamellar phases was made more than 20 years ago by Larsson (1989, 2009). For the formulation of stable submicron-sized particles, enveloping well-ordered internal nanostructures, surfactant-like lipids with a tendency to form nonlamellar systems were dispersed in excess buffer by applying high-energy input in the presence of an eficient hydrophilic stabilizer. Among these nanostructured aqueous dispersions, cubosomes (aqueous dispersions of a V2 phase) and hexosomes (aqueous dispersions of an H2 phase) have attracted much attention as drug and functional food nanocarriers (Larsson, 2009, Yaghmur and Glatter 2009, Yaghmur and Rappolt 2010). In the literature, most investigations reported on the use of the amphiphilic triblock copolymer Pluronic F127 as an eficient hydrophilic stabilizer for the formation of cubosomes and hexosomes (Larsson 2000, Yaghmur and Glatter 2009). It was demonstrated that the steric stabilization of these colloidal particles was realized by the adsorption of the copolymer’s hydrophobic moieties into the outer surface of the dispersed particles shielding the inverted-type self-assembled lipid nanostructure from the surrounding aqueous medium, whereas the hydrophilic moieties were in a direct contact with water (the continuous phase) (Larsson 2000, Barauskas et al. 2005, Yaghmur and Glatter 2009). Later on, the formation and the characterization of micellar cubosomes (aqueous dispersions of the discontinuous cubic Fd3m phase) as well as emulsified microemulsions (EMEs, aqueous dispersions of the inverted-type microemulsion system) were put into practice (Yaghmur et al. 2005, 2006b), and such set of possible dispersions from inverse mesophases was completed: (i) cubosomes (internal cubic Im3m or Pn3m phase), (ii) hexosomes (internal H2 phase), (iii) micellar cubosomes (internal cubic Fd3m phase), and (iv) EMEs (internal waterin-oil (W/O) microemulsion phase) (Larsson, 2000, Yaghmur et  al. 2005, 2006b, Yaghmur and Rappolt 2011) (see also Figure 17.2). This unique family of dispersed particles enveloping internal

350

Liposomes, Lipid Bilayers and Model Membranes

a

Emulsified microemulsions

b

Micellar cubosomes

100 nm

c

Hexosomes

d

Cubosomes

300 nm

200 nm

Temperature and/or solubilized oil content decrease

self-assembled nanostructures is also known as ISASOMES (internally self-assembled particles or “somes”) (Yaghmur et al. 2005). In Figure 17.6, different cryo-transmission electron microscopy (cryo-TEM) images show prominent examples of these oil-free and oil-loaded sub-micron-sized colloidal particles. Further technical details on the formulation procedures and stabilization can be found in the recent review of Yaghmur and Glatter (2009). As for the corresponding nondispersed bulk samples, it was also demonstrated that the aqueous dispersions of the binary monolinolein (MLO)/water system (de Campo et al. 2004) undergo reversible transitions from cubosomes (an internal cubic Pn3m phase) via hexosomes (an internal H2 phase) to an emulsified L2 phase (an internal inverse micellar solution). It was reported that the reversible structural transitions of the dispersed and the nondispersed samples were practically identical, which in turn meant that the polymeric stabilizer F127 had no signiicant impact on the

100 nm

FIGURE 17.6 Cryo-TEM observations taken for four different types of oil-free and oil-loaded monoglyceridebased aqueous dispersions enveloping self-assembled nanostructures: (a) EMEs (Reprinted with permission from Yaghmur, A. et al. 2005. Emulsiied microemulsions and oil-containing liquid crystalline phases. Langmuir 21(2), 569–577. Copyright 2005, American Chemical Society.), (b) micellar cubosomes (Reprinted with permission from Yaghmur, A. et al. 2006b. Oil-loaded monolinolein-based particles with conined inverse discontinuous cubic structure (Fd3m). Langmuir 22(2), 517–521. Copyright 2006b, American Chemical Society.), (c) hexosomes (Reprinted with permission from Yaghmur, A. et al. 2005. Emulsiied microemulsions and oil-containing liquid crystalline phases. Langmuir 21(2), 569–577. Copyright 2005, American Chemical Society.), and (d) cubosomes. (Reprinted with permission from de Campo, L. et al. 2004. Reversible phase transitions in emulsiied nanostructured lipid systems. Langmuir 20(13), 5254–5261. Copyright 2004, American Chemical Society.)

Drug Formulations Based on Self-Assembled Liquid Crystalline Nanostructures

351

conined nanostructures (de Campo et al. 2004). However, the transition temperature especially for the formation of the inverse micellar phase is quite high (>90°C) (de Campo et al. 2004). Therefore, the main intention subsequently focused on lowering the phase transition temperatures in order to form EMEs already at room temperature. As discussed above, the addition of oil can reduce the packing frustration in liquid crystalline phases quite drastically (Yaghmur et al. 2005, 2010a), and this is particularly evident when inducing the formation of the discontinuous H2, cubic Fd3m and L2 phases. The formation of cubosomes, hexosomes, micellar cubosomes, and EMEs was achieved at room temperature by augmenting the solubilized tetradecane content (Figure 17.6). Briely, solubilizing different amounts of the investigated oil at room temperature in both the dispersed and the corresponding nondispersed phases or increasing the temperature induce the same sequence of phase transition (Yaghmur et al. 2005, 2006a,b). The direct formation of cubosomes from liposomes is an attractive alternative production method. In a recent study, it was demonstrated that heating the aqueous dispersion of monoelaidin (ME, a trans monounsaturated monoglyceride), which is the counterpart of MO (a cis monounsaturated monoglyceride), induces a direct liposomes–cubosomes transition (Yaghmur et al. 2008a). It is clear also as discussed above that the addition of calcium ions to negatively charged membranes such as the vesicles based on the binary lipid mixture consisting of DOPG and MO can be applied for the formation of cubosomes and hexosomes (Yaghmur et al. 2008b, 2011b). In a recent interesting report, Muir et al. (2012) found that cubosomes with low polydispersity can be produced by the addition of salt to cationic liposomes. This emulsiication method ensures the formation of cubosomes without the application of high-energy emulsiication input.

17.5 BICONTINUOUS CUBIC AND HEXAGONAL LIQUID CRYSTALLINE PHASES AS SUSTAINED DRUG RELEASE VEHICLES Potential drug carrier properties of V2 and H2 phases have been investigated following subcutaneous (Fong et al. 2009, Ahmed et al. 2010, Rosenbaum et al. 2010, Yaghmur et al. 2012a), transdermal (Fitzpatrick and Corish 2005, Peng et al. 2010), oral (Shah and Paradkar 2005, Boyd et al. 2007), dental (Norling et  al. 1992), buccal (Nielsen et  al. 1998), ophthalmic (Lindell et  al. 1998), and vaginal (Geraghty et  al. 1996) administrations. Among the different LLC phases, drug delivery systems based on MO have attracted most attention as reviewed by Shah et al. (2001). An example of a marketed product based on an MO formulation is Elyzol® dental gel for treatment of periodontal disease. In the application of LLC phases as matrices in drug delivery systems, the bioadhesive properties might be utilized, for example, for improved buccal delivery (Nielsen et al. 1998). Also after oral administration, the time period for the mucoadhesive formulation to reside in the gastrointestinal tract might be extended as observed after oral administration of cinnarizine dissolved in the slowly digestible oleyl glycerate in rats (Boyd et al. 2007). Although the hydration-induced selfassembling property constitutes the basic principle for the formation of these systems, the amount of water taken up by different lipids formulations might vary (Yaghmur et al. 2005, 2012a). It is therefore possible to prepare drug formulations having different levels of hydration, for example, using fully hydrated drug-loaded systems as well as water-free or low-water containing drug preformulations. High viscosity and stiffness of the V2 and H2 systems may be prohibitive for their use as drug delivery systems due to dificulties in administration, for example, for injections; conversely, low-viscous drug formulations that upon contact with an aqueous medium, for example, by subcutaneous injection, transform into the corresponding LLC phases have been investigated intensively (Engström et al. 1992, Chang and Bodmeier 1998, Shah and Paradkar 2005, Boyd et al. 2006b, Fong et al. 2009, Ahmed et al. 2010, Phelps et al. 2011, Yaghmur et al. 2012a). The precursors to these in situ forming drug delivery systems comprise various lipids in the absence of water (Yaghmur et al. 2012a), containing small amounts of water (Fong et al. 2009, Yaghmur et al. 2011a), or mixtures of water and organic solvents (Chang and Bodmeier 1998, Ahmed et al. 2010, Phelps et al. 2011). In addition, the drug itself may facilitate the formation of a preformulation with low viscosity (Chang

352

Liposomes, Lipid Bilayers and Model Membranes

and Bodmeier 1998). Apparently, the interest in using these LLC phases as drug delivery systems is based on their ability to incorporate drugs with variable sizes and physicochemical properties together with the possibility of accomplishing sustained drug release. To this end, incorporation of peptides and proteins in LLC phases and their corresponding aqueous dispersions might protect them from chemical and physical inactivation (Shah et al. 2001). Design of eficient sustained drug delivery systems using LLC phases depends on a detailed understanding of mechanistically factors related to the drug release process. The in vitro release of various compounds from V2 and/or H2 phases has been investigated in several studies. In general, drug release was found to obey Higuchi’s square root of time release kinetics when studying drug release from the LLC phases having a constant contact surface area (Shah and Paradkar 2005, Fong et al. 2009, Rosenbaum et al. 2010, Negrini and Mezzenga 2011). Although drug release from the V2 and H2 formulations most often is described to be diffusion controlled, the release mechanisms are not fully understood; in particular, the inluence of the physicochemical properties of the solubilized drug and the lipid composition on the release mechanism call for further investigations. For hydrophilic compounds, the release from H2 phases was found to be slower than that from V2 phases (Fong et al. 2009, Rizwan et al. 2009, Phelps et al. 2011, Negrini and Mezzenga 2011), which indicates that the diffusion through the matrix is dependent on the diameter and tortuosity of the hydrophilic channels embedded in the nanostructure. To this end, drug release from LLCs has been induced by switching the self-assembled nanostructure (e.g., from V2 to H2 phase transition) through stimuli-responsive factors as have been demonstrated for hydrophilic model substances by varying the temperature (Fong et al. 2009) or the pH of the aqueous medium (Negrini and Mezzenga 2011). For lipophilic compounds, the release mechanism seems to be more complex, for example, the partitioning of the solubilized substance into the lipid part of the nanostructure (the hydrophobic domain) has been shown to affect the release rate. In a recent study, the release of the lipophilic drug bupivacaine (weak base with pKa of 8.1) from in situ formed LLC phases was investigated. The slower release rate observed upon addition of a water-free precursor (MO-based preformulation) to aqueous buffer at pH 7.4 compared with that at pH 6.0 (Figure 17.7) was in accordance with the higher observed lipid partition coeficient at pH 7.4 (Yaghmur et al. 2012a). Furthermore, in this study it was revealed that bupivacaine release from the H2 phase formed using a precursor based on a binary mixture consisting of MO and medium chain triglycerides was faster as compared with that from the V2 phase formed by using the water-free MO-based precursor (Figure 17.7). Sustained release can also be improved by enhancing the solubilized drug afinity to the lipidic matrix of the LLC phases. This can be achieved by chemically modifying the drug substance as shown by alkylation of the model compound tryptophan (Clogston et al. 2005). The drug partitioning into the lipid bilayers of cubic phases has been studied in detail by Engström et al. (1999). As mentioned previously, solubilization of drugs in the LLCs phases may alter the nanostructure as observed for ibuprofen, tetracaine, propanolol, lidocaine, and bupivacaine which all induced a V2-to-H2 phase transition in MO-based systems (Engström and Engström 1992, Chang and Bodmeier 1997, Yaghmur et al. 2011a). Since these structural changes may lead to altered release proiles, the inluence of the nanostructural characteristics of the LLC systems on the release pattern needs to be investigated case by case. Additional important parameters potentially affecting the release behavior include the initial water content in the formulation as well as the water swelling rate. Contradictory results can be found in the literature (Burrows et al. 1994, Lara et al. 2005, Rizwan et al. 2009). In some studies, increase in drug release rate with increasing initial water content of the MO-based system was observed (Chang and Bodmeier 1997, Lara et  al. 2005), whereas in other investigations no correlation between the two parameters was seen (Burrows et al. 1994, Rizwan et al. 2009). It has been suggested that the effect of variation of the precursor MO/water weight ratio on the release rate depends on the afinity of the drug substance to the hydrophobic domains of the LLC phases (Carr et al. 1997). In the assessment of in vitro release results focusing on the inluence of the selfassembled nanostructure-related as well as the drug-related factors on the release rate, it is crucial to employ a feasible in vitro release model capable of mimicking the environment for in vivo drug

Drug Formulations Based on Self-Assembled Liquid Crystalline Nanostructures

353

(a)

BUP release (%)

100

50

0

0

20

40

60

Time (h) (b)

After buffer hydration

Pn3m

H2

L2

Increase in the drug release rate

FIGURE 17.7 (See color insert.) (a) Bupivacaine (BUP) release proiles obtained upon addition of a MO-based precursor to phosphate buffer at pH 6 (triangles) and at pH 7.4 (diamonds) in the rotating dialysis cell model. The release patterns were compared to an aqueous buffer solution of BUP (pH 6.0) (circles). The full lines were obtained by itting the data to irst-order kinetics. Bars represent standard deviations for experiments in triplicate. (b) Schematic illustration of the effect of lipid composition on the release rates of different bupivacaine-loaded lipid formulations based on the binary MO/ medium chain triglycerides precursors. (Reprinted with permission from Yaghmur, A. et al. 2012a. Characterization of bupivacaine-loaded formulations based on liquid crystalline phases and microemulsions: The effect of lipid composition. Langmuir, 28(5), 2881–2889. Copyright 2012, American Chemical Society.)

release. As mentioned, many release studies were performed by investigating drug release from a constant surface area of the formulation; however, the employment of different stirring conditions (and thus hydrodynamics) and differences in the degree of sink condition make comparison of the obtained release data dificult. To better understand the in vivo drug release mechanism, additional factors need to be taken into account, for example, maintenance or formation of the nanostructure in biological luids, volume-to-surface area upon administration of the formulation, and the diffusionconvention conditions in vivo. Thus, further studies in this research area are needed.

17.6

NANOSTRUCTURED COLLOIDAL LIPIDIC PARTICLES AS DRUG NANOCARRIERS

In recent years, there has been an increasing interest in exploring nanoparticulate drug delivery systems for optimizing drug performance. As compared with liposomes, the research area on the pharmaceutical uses of the nonlamellar nanostructured lipid dispersions is still in its infancy. The majority of the new drug candidates are poorly water-soluble small molecules and therefore present a challenge to formulate. The unique structural characteristics of colloidal lipidic particles

354

Liposomes, Lipid Bilayers and Model Membranes

enveloping well-deined nanostructures potentially allow the design of effective formulations with desired properties (Malmsten 2006, Yaghmur and Glatter 2009). Nanostructured liquid crystalline particles not only result in eficient loading of poorly water-soluble drugs, but may also result in formulating nanocarriers with a sustained release property (Lai et al. 2009, Nguyen et al. 2011). PHYT-based cubosomes provided increased bioavailability and sustained plasma concentrations of cinnarizine upon oral administration to rat as compared with conventional delivery systems (suspensions and oleic acid emulsions) (Nguyen et  al. 2011). The in vivo results were somewhat surprising based on the expected burst release character of the cubosomal formulation. It was proposed that the stomach of the rat constituted a nonsink release environment due to the small volume of the gastric liquid present and, thus, partitioning effects slowed the absorption, in addition to the prolonged retention of the cubosomes in the rat stomach. It has also been reported that cubosomes based on MO provide an increase in the oral availability of simvastatin (a poorly water-soluble drug) upon administration to dog, the relative availability was 241% as compared to a simvastatin crystal suspension (Lai et al. 2009). An area of particular interest is formulating anti-cancer drugs administered in chemotherapy. Promising results were recently reported on the use of a liquid crystal nanoparticle formulation of docetaxel for potential prostate cancer treatment (Cervin et al. 2010). Intravenous injection of different formulations of paclitaxel has demonstrated higher plasma levels of paclitaxel loaded in liquid crystalline nanoparticles compared with the commercial formulation Taxol (Zeng et al. 2012). Amphiphilic molecules may also be formulated using cubosomes as demonstrated for chemotherapeutic ruthenium complexes (Mangiapia et al. 2011). Dexamethasone (a drug compound having a steroidal structure) has been formulated for ophthalmic delivery (Gan et al. 2010). In this case, the drug vehicle was composed of cubosomes and was shown to be superior to other formulations principles. The oromucosal delivery route of progesterone was explored using hexosomes (Swarnakar et al. 2007). It was found that the use of hexosomes led to signiicantly increased drug release. A cubosomal dispersion was shown to be well suited as a delivery system for the percutaneous administration of the nonsteroidal anti-inlammatory drug compound indomethacin (Esposito et al. 2005). A prolonged anti-inlammatory effect was observed suggesting that the cubosomal formulation may have a depot effect. In ophtalmic delivery of lurbiprofen (Han et al. 2010), a cubosomal formulation was shown to combine low irritancy and high bioavailability. This study emphasizes the large un-explored potential for cubosomes as nanostructured drug vehicles with limited toxic side effects. Drummond and co-workers (Sagnella et al. 2011) described the use of lipid nanoparticles obtained from the emulsiication of self-assembled nanostructures based on 5-luorouracil lipid prodrugs in the presence of the polymeric stabilizer F127 as means for obtaining a drug delivery system with sustained release properties and target-selective activation upon oral administration in a mouse tumour model. A 3-step enzymatic reconversion sequence was suggested to provide the regeneration of the parent 5-luorouracil in a target selective manner upon absorption of the nanoparticles to the blood stream. There is also an increasing interest in loading peptides and proteins to these nanostructured aqueous dispersions. In this context, preservation of the physical and chemical stability is a main challenge in the delivery of protein-based drugs. To this end, Chung et al. (2002) prepared the socalled nanocubicles at room temperature thus avoiding the emulsiication at high temperatures that would compromise insulin. These nanocarriers were found to lower the blood glucose concentration in diabetic rats upon oral administration. Whether the effect of the colloidal nanocarrier was due to an increased penetration of the intestinal barrier or due to a protection of insulin from proteolytic enzymes was not resolved. In the assessment of lipidic nanocarriers as vaccine delivery systems, Boyd and co-workers (Rizwan et al. 2011) prepared PHYT-based cubosomes loading ovalbumin using the solvent precursor dilution approach developed by Spicer et al. (2001). This was done in order to avoid exposing the potentially labile protein to methods of preparation requiring highenergy input. A more sustained release of FITC-ovalbumin was observed for the cubosomes prepared by the solvent precursor dilution method as compared with those prepared by fragmentation

Drug Formulations Based on Self-Assembled Liquid Crystalline Nanostructures

355

using an Ultra-Turrax. Liquid crystalline nano-carriers were useful for prolonging the plasma halflife of the peptide hormone somatostatin upon intravenous injection into rats (Cervin et al. 2009). The mechanism behind the extended residence time was not elucidated in full but the formulation was suggested to protect somatostatin against peptidase activity. The study also assessed the pharmacokinetic behavior observed for dispersions where somatostatin was encapsulated within or simply adsorbed onto the nanoparticles; the former provided higher AUC and longer a half-life. It would be interesting to have more information on the fate of the nanoparticles in order to extract additional mechanistic information. Bentley and co-workers (Lopes et al. 2006) documented the usefulness of hexosomes for improving the topical delivery of cyclosporine A by following in vitro and in vivo experiments. Importantly the vehicle did not cause skin irritation. Similar to other nanoparticulate delivery systems, thorough characterization of cubosomes and hexosomes is challenging, though highly needed in order to ensure that the lipidic nanocarriers encompass the required and expected properties for their speciic application. This relates both to formulation development and quality control. In addition to the structural characterization of these aqueous dispersions by SAXS and cryo-TEM discussed above, other techniques can be used to obtain further important information on the size of the dispersed cubosomal and hexosomal droplets, the drug solubilization capacity, and the effect of aging on the stability of the emulsiied phases. Measurements of size and polydispersity index are commonly determined using dynamic light scattering (de Campo et al. 2004, Cervin et al. 2010), laser diffraction (Johnsson et al. 2006), photo correlation spectroscopy (Rizwan et al. 2011), and sedimentation ield low fractionation (Esposito et al. 2012). These methods are together with zeta-potential measurements widely used to obtain information on the effect of aging on the stability and the aggregation behavior of various dispersed particles (e.g., Boyd et  al. 2006a, Cervin et  al. 2009) as is common practice for liposomes. The solubilization capacity of poorly water-soluble drug substances has been assessed by ultrailtration (Rosenblatt et al. 2007) as well as gel permeation chromatography (Lai et al. 2009) and sedimentation ield low fractionation followed by HPLC analysis (Esposito et al. 2012). For studying the in vitro drug release properties of the nanostructured aqueous dispersions, the applied in vitro method should take into account the mechanism of drug release from the delivery system as well as the conditions relevant for the intended route of administration. In this context, the selection of the appropriate in vitro release method is central in the predication of the in vivo performance (Larsen et al. 2009). A number of release methods have been applied for the examination of the in vitro release behavior including dynamic dialysis (Lai et al. 2009, Han et al. 2010, Zeng et al. 2012), pressure ultrailtration method (Boyd, 2003, Rosenblatt et al. 2007), differential pulse polarography (Rosenblatt et al. 2007), and sample and separate methods (Rizwan et al. 2011). Dynamic dialysis methods should be used with caution (Boyd, 2003, Rosenblatt et al. 2007) since the membrane limits the diffusion and the release from the dialysis chamber/cell relects drug partitioning between the aqueous phase and the colloidal carrier in the donor phase rather than release from the formulation. This has been pointed out for these lipidic nanostructural formulations by Boyd (2003) and Rosenblatt et al. (2007) comparing data obtained by pressure ultrailtration and differential pulse polarography to those taken from equilibrium dialysis. Based on the drug diffusivities within the LCC phases, the surface area, and the volume (Boyd, 2003), it was estimated that the release from cubosomes can be expected to be fast, that is, have burst release character when sink conditions are present. In order to obtain time-resolved information on the extent of drug release, methods allowing a fast separation of the drug from these colloidal drug nanocarriers are warranted. However, as shown above-sustained release properties have been associated with nanostructured liquid crystalline lipidic particles under particular conditions (Nguyen et al. 2011).

17.7

SUMMARY

Recent progress in the formation, the characterization, and the potential pharmaceutical applications of lipidic lyotropic nonlamellar liquid crystalline phases and their corresponding nanostructured

356

Liposomes, Lipid Bilayers and Model Membranes

aqueous dispersions is summarized. Mainly, the described V2 and H2 phases in both the nondispersed and the dispersed states represent a unique class of nanomaterials that holds a promise in meeting the needs for safe and biocompatible nanocarriers that can be designed to control drug release and even to target speciic tissues after administration. The number of studies on their use in biomedical applications is still limited. Therefore, fundamental research to be carried out for a detailed thorough understanding of the physicochemical properties, and the interactions with the biological milieu is essential for the in vivo potential of these self-assembled systems. This complex task requires a multidisciplinary approach that includes a focus on the structural characterization under nonequilibrium conditions upon the exposure of these self-assembled nanocarriers to the biological environment, and the combination of relevant in vitro and in vivo investigations. The surface modiication of the cubosomal and the hexosomal particles could also be an attractive strategy in the design of nanocarriers for imaging and targeted drug applications.

REFERENCES Ahmed, A. R., Dashevsky, A., and Bodmeier, R. 2010. Drug release from and sterilization of in situ cubic phase forming monoglyceride drug delivery systems. Eur. J. Pharm. Biopharm. 75(3), 375–380. Almsherqi, Z. A., Kohlwein, S. D., and Deng, Y. 2006. Cubic membranes: A legend beyond the Flatland of cell membrane organization. J. Cell Biol. 173(6), 839–844. Almsherqi, Z. A. et al. 2009. Cubic membranes: The missing dimension of cell membrane organization. Int. Rev. Cell Mol. Biol. 274, 275–342. Angelova, A. et al. 2011. Self-assembled multicompartment liquid crystalline lipid carriers for protein, peptide, and nucleic acid drug delivery. Accounts Chem. Res. 44(2), 147–156. Angelova, A. et al. 2012. Small-angle X-ray scattering investigations of biomolecular coninement, loading, and release from liquid-crystalline nanochannel assemblies. J. Phys. Chem. Lett. 3, 445–457. Awad, T. S. et al. 2005. Formation of cubic phases from large unilamellar vesicles of dioleoylphosphatidylglycerol/monoolein membranes induced by low concentrations of Ca2+. Langmuir 21(25), 11556–11561. Barauskas, J. and Landh, T. 2003. Phase behavior of the phytantriol/water system. Langmuir 19(23), 9562–9565. Barauskas, J., Johnsson, M., and Tiberg, F. 2005. Self-assembled lipid superstructures: Beyond vesicles and liposomes. Nano Lett. 5(8), 1615–1619. Boyd, B. J. 2003. Characterisation of drug release from cubosomes using the pressure ultrailtration method. Int. J. Pharm. 260(2), 239–247. Boyd, B. J. et al. 2006a. Hexosomes formed from glycerate surfactants—Formulation as a colloidal carrier for irinotecan. Int. J. Pharm. 318(1–2), 154–162. Boyd, B. J. et al. 2006b. Lyotropic liquid crystalline phases formed from glycerate surfactants as sustained release drug delivery systems. Int. J. Pharm. 309(1–2), 218–226. Boyd, B. J. et al. 2007. A lipid-based liquid crystalline matrix that provides sustained release and enhanced oral bioavailability for a model poorly water soluble drug in rats. Int. J. Pharm. 340(1–2), 52–60. Briggs, J. and Caffrey, M. 1994. The temperature-composition phase diagram of monomyristolein in water: Equilibrium and metastability aspects. Biophys. J. 66(3), 573–587. Burrows, R., Collett, J. H., and Attwood, D. 1994. The release of drugs from monoglyceride-water liquidcrystalline phases. Int. J. Pharm. 111(3), 283–293. Carr, M. G., Corish, J., and Corrigan, O. I. 1997. Drug delivery from a liquid crystalline base across Visking and human stratum corneum. Int. J. Pharm. 157(1), 35–42. Cervin, C. et al. 2009. A combined in vitro and in vivo study on the interactions between somatostatin and lipidbased liquid crystalline drug carriers and bilayers. Eur. J. Pharm. Sci. 36(4–5), 377–385. Cervin, C. et al. 2010. Properties and effects of a novel liquid crystal nanoparticle formulation of docetaxel in a prostate cancer mouse model. Eur. J. Pharm. Sci. 41(2), 369–375. Chang, C. M. and Bodmeier, R. 1997. Binding of drugs to monoglyceride-based drug delivery systems. Int. J. Pharm. 147(2), 135–142. Chang, C. M. and Bodmeier, R. 1998. Low viscosity monoglyceride-based drug delivery systems transforming into a highly viscous cubic phase. Int. J. Pharm. 173(1–2), 51–60. Cherezov, V. et al. 2003. The kinetics of non-lamellar phase formation in DOPE-Me: Relevance to biomembrane fusion. J. Membrane Biol. 195(3), 165–182. Cherezov, V. et al. 2006. Room to move: Crystallizing membrane proteins in swollen lipidic mesophases. J. Mol. Biol. 357(5), 1605–1618.

Drug Formulations Based on Self-Assembled Liquid Crystalline Nanostructures

357

Cherezov, V. and Caffrey, M. 2007. Membrane protein crystallization in lipidic mesophases. A mechanism study using X-ray microdiffraction. Faraday Discuss. 136, 195–212. Chung, H. et  al. 2002. Self-assembled “nanocubicle” as a carrier for peroral insulin delivery. Diabetologia 45(3), 448–451. Clogston, J. et al. 2000. Phase behavior of a monoacylglycerol: (myverol 18–99 K)/water system. Chem. Phys. Lipids 107(2), 191–220. Clogston, J. et al. 2005. Controlling release from the lipidic cubic phase by selective alkylation. J. Control. Release 102(2), 441–461. Conn, C. E. et al. 2006. Dynamics of structural transformations between lamellar and inverse bicontinuous cubic lyotropic phases. Phys. Rev. Lett. 96(10), 108102. Conn, C. E. et al. 2008. A pressure-jump time-resolved X-ray diffraction study of cubic-cubic transition kinetics in monoolein. Langmuir 24(6), 2331–2340. Costa-Balogh, F. O. et al. 2010. Drug release from lipid liquid crystalline phases: Relation with phase behavior. Drug Dev. Ind. Pharm. 36(4), 470–481. Czeslik, C. et al. 1995. Temperature- and pressure-dependent phase behavior of monoacylglycerides monoolein and monoelaidin. Biophys. J. 68(4), 1423–1429. de Campo, L. et al. 2004. Reversible phase transitions in emulsiied nanostructured lipid systems. Langmuir 20(13), 5254–5261. Dong, Y. D. et al. 2010. Nonequilibrium effects in self-assembled mesophase materials: Unexpected supercooling effects for cubosomes and hexosomes. Langmuir 26(11), 9000–9010. Drummond, C. J. and Fong, C. 1999. Surfactant self-assembly objects as novel drug delivery vehicles. Curr. Opin. Colloid Interf. Sci. 4(6), 449–456. Engström, S. and Engström, L. 1992. Phase-behavior of the lidocaine-monoolein-water system. Int. J. Pharm. 79(2–3), 113–122. Engström, S. et al. 1992. A Study of polar lipid drug carrier systems undergoing a thermoreversible lamellarto-cubic phase-transition. Int. J. Pharm. 86(2–3), 137–145. Engström, S., Norden, T. P., and Nyquist, H. 1999. Cubic phases for studies of drug partition into lipid bilayers. Eur. J. Pharm. Sci. 8(4), 243–254. Esposito, E. et al. 2005. Cubosome dispersions as delivery systems for percutaneous administration of indomethacin. Pharm. Res. 22(12), 2163–2173. Esposito, E. et al. 2012. Nanoparticulate lipid dispersions for bromocriptine delivery: Characterization and in vivo study. Eur. J. Pharm. Biopharm. 80(2), 306–314. Faivre, V. and Rosilio, V. 2010. Interest of glycolipids in drug delivery: From physicochemical properties to drug targeting. Expert Opin. Drug Del. 7(9), 1031–1048. Fitzpatrick, D. and Corish, J. 2005. Release characteristics of anionic drug compounds from liquid crystalline gels—I: Passive release across non-rate-limiting membranes. Int. J. Pharm. 301(1–2), 226–236. Fong, W. K., Hanley, T., and Boyd, B. J. 2009. Stimuli responsive liquid crystals provide ‘on-demand’ drug delivery in vitro and in vivo. J. Control Release 135(3), 218–226. Gan, L. et al. 2010. Self-assembled liquid crystalline nanoparticles as a novel ophthalmic delivery system for dexamethasone: Improving preocular retention and ocular bioavailability. Int. J. Pharm. 396(1–2), 179–187. Geraghty, P. B. et al. 1996. The in vitro release of some antimuscarinic drugs from monoolein/water lyotropic liquid crystalline gels. Pharm. Res. 13(8), 1265–1271. Han, S. et al. 2010. Novel vehicle based on cubosomes for ophthalmic delivery of lurbiprofen with low irritancy and high bioavailability. Acta Pharmacol. Sin. 31(8), 990–998. Huang, Z., Seddon, J. M., and Templer, R. H. 1996. An inverse micellar Fd3m cubic phase formed by hydrated phosphatidylcholine/fatty alcohol mixtures. Chem. Phys. Lipids 82(1), 53–61. Hui, S. W. et al. 1981. Membrane fusion through point defects in bilayers. Science 212(4497), 921–923. Hyde, S. T. 1996. Bicontinuous structure in lyotropic liquid crystals and crystalline hyperbolic surfaces. Curr. Opin. Solid State Mater. Sci. 1, 653–662. Hyde, S. et al. 1997. The Language of Shape: The Role of Curvature in Condensed Matter: Physics, Chemistry, Biology. Amsterdam: Elsevier. Israelachvili, J. N. 1991. Intermolecular and Surface Forces. London: Academic Press. Johnsson, M. et al. 2006. Physicochemical and drug delivery aspects of lipid-based liquid crystalline nanoparticles: A case study of intravenously administered propofol. J. Nanosci. Nanotechnol. 6(9–10), 3017–3024. Kaasgaard, T. and Drummond, C. J. 2006. Ordered 2-D and 3-D nanostructured amphiphile self-assembly materials stable in excess solvent. Phys. Chem. Chem. Phys. 8(43), 4957–4975. Kitamoto, D. et al. 2009. Self-assembling properties of glycolipid biosurfactants and their potential applications. Curr. Opin. Colloid Interface Sci. 14(5), 315–328.

358

Liposomes, Lipid Bilayers and Model Membranes

Kozlov, M. M., Leikin, S., and Rand, R. P. 1994. Bending, hydration and interstitial energies quantitatively account for the hexagonal-lamellar-hexagonal reentrant phase-transition in dioleoylphosphatidylethanolamine. Biophys. J. 67(4), 1603–1611. Kulkarni, C. V. 2011. Nanostructural studies on monoelaidin-water systems at low temperatures. Langmuir 27(19), 11790–11800. Kulkarni, C. V. et al. 2011. Monoolein: A magic lipid? Phys. Chem. Chem. Phys. 13(8), 3004–3021. Laggner, P. and Kriechbaum, M. 1991. Phospholipid phase transitions: Kinetics and structural mechanisms. Chem. Phys. Lipids 57(2–3), 121–145. Laggner, P. et  al. 2005. Pharmaceutical solid-state characterization by small-and wide-angle X-ray scattering. In: A. and M. Zakrzewski, eds. Solid State Characterization of Pharmaceuticals, Danbury: Assa International, 407–448. Lai, J. et al. 2009. Glyceryl monooleate/poloxamer 407 cubic nanoparticles as oral drug delivery systems: I. in vitro evaluation and enhanced oral bioavailability of the poorly water-soluble drug simvastatin. Aaps Pharmscitech 10(3), 960–966. Lara, M. G., Bentley, M. V. L. B., and Collett, J. H. 2005. in vitro drug release mechanism and drug loading studies of cubic phase gels. Int. J. Pharm. 293(1–2), 241–250. Larsen, C. et al. 2009. Role of in vitro release models in formulation development and quality control of parenteral depots. Expert Opin. Drug Del. 6(12), 1283–1295. Larsson, K. 1989. Cubic lipid-water phases: Structures and biomembrane aspects. J. Phys. Chem. 93(21), 7304–7314. Larsson, K. 2000. Aqueous dispersions of cubic lipid-water phases. Curr. Opin. Colloid Interface Sci. 5(1–2), 64–69. Larsson, K. 2009. Lyotropic liquid crystals and their dispersions relevant in foods. Curr. Opin. Colloid Interface Sci. 14(1), 16–20. Lindell, K. et al. 1998. Inluence of a charged phospholipid on the release pattern of timolol maleate from cubic liquid crystalline phases. Progr. Colloid Polym. Sci. 108, 111–118. Lopes, L. B. et al. 2006. Reverse hexagonal phase nanodispersion of monoolein and oleic acid for topical delivery of peptides: In vitro and in vivo skin penetration of cyclosporin A. Pharm. Res. 23(6), 1332–1342. Luzzati, V. et al. 1992. Lipid polymorphism: A correction. The structure of the cubic phase of extinction symbol Fd—consists of two types of disjointed reverse micelles embedded in a three-dimensional hydrocarbon matrix. Biochemistry 31(1), 279–285. Luzzati, V. 1997. Biological signiicance of lipid polymorphism: The cubic phases. Curr. Opin. Struct. Biol. 7(5), 661–668. Malmsten, M. 2006. Soft drug delivery systems. Soft Matter 2(9), 760–769. Mangiapia, G. et al. 2011. Cubosomes for ruthenium complex delivery: Formulation and characterization. Soft Matter 7(22), 10577–10580. Mares, T. et al. 2008. Role of phospholipid asymmetry in the stability of inverted hexagonal mesoscopic phases. J. Phys. Chem. B 112(51), 16575–16584. Masum, S. M. et al. 2005. Effect of positively charged short peptides on stability of cubic phases of monoolein/ dioleoylphosphatidic acid mixtures. Langmuir 21(12), 5290–5297. Misquitta, Y. and Caffrey, M. 2001. Rational design of lipid molecular structure: A case study involving the C19:1c10 monoacylglycerol. Biophys. J. 81(2), 1047–1058. Muir, B. W. et al. 2012. Salt induced lamellar to bicontinuous cubic phase transitions in cationic nanoparticles. J. Phys. Chem. B 116(11), 3551–3556. Negrini, R. and Mezzenga, R. 2011. pH-Responsive lyotropic liquid crystals for controlled drug delivery. Langmuir 27(9), 5296–5303. Nguyen, T. H. et al. 2011. Nanostructured liquid crystalline particles provide long duration sustained-release effect for a poorly water soluble drug after oral administration. J. Control. Release 153(2), 180–186. Nielsen, L. S., Schubert, L., and Hansen, J. 1998. Bioadhesive drug delivery systems—I. Characterisation of mucoadhesive properties of systems based on glyceryl mono-oleate and glyceryl monolinoleate. Eur. J. Pharm. Sci. 6(3), 231–239. Norling, T. et al. 1992. Formulation of a drug delivery system based on a mixture of monoglycerides and triglycerides for use in the treatment of periodontal-disease. J. Clin. Periodontol. 19(9), 687–692. Paasonen, L. et al. 2010. Gold-embedded photosensitive liposomes for drug delivery: Triggering mechanism and intracellular release. J. Control. Rel. 147(1), 136–143. Patton, J. S. and Carey, M. C. 1979. Watching fat digestion. Science 204(4389), 145–148. Peng, X. S. et al. 2010. Design and in vitro evaluation of capsaicin transdermal controlled release cubic phase gels. Aaps Pharmscitech 11(3), 1405–1410.

Drug Formulations Based on Self-Assembled Liquid Crystalline Nanostructures

359

Phan, S. et al. 2011. Evaluating the link between self-assembled mesophase structure and drug release. Int. J. Pharm. 421(1), 176–182. Phelps, J., Bentley, M. V., and Lopes, L. B. 2011. In situ gelling hexagonal phases for sustained release of an anti-addiction drug. Colloids Surface B 87(2), 391–398. Qiu, H. and Caffrey, M. 1999. Phase behavior of the monoerucin/water system. Chem. Phys. Lipids 100(1–2), 55–79. Qiu, H. and Caffrey, M. 2000. The phase diagram of the monoolein/water system: Metastability and equilibrium aspects. Biomaterials 21(3), 223–234. Rappolt, M. et al. 2000. New evidence for gel-liquid crystalline phase coexistence in the ripple phase of phosphatidylcholines. Eur. Biophys. J. Biophy. 29(2), 125–133. Rappolt, M. et al. 2003. Mechanism of the lamellar/inverse hexagonal phase transition examined by high resolution X-ray diffraction. Biophys. J. 84(5), 3111–3122. Rappolt, M. 2006. The biologically relevant lipid mesophases as “seen” by X-rays. In: A. Leitmannova, ed. Planar Lipid Bilayers and Liposomes. Amsterdam: Elsevier, 253–283. Rappolt, M. et al. 2006. Non-equilibrium formation of the cubic Pn3m phase in a monoolein/water system. Europhys. Lett. 75(2), 267–273. Rappolt, M. et  al. 2008. Conformational and hydrational properties during the L(beta)- to L(alpha)- and L(alpha)- to H(II)-phase transition in phosphatidylethanolamine. Chem. Phys. Lipids 154(1), 46–55. Rizwan, S. B. et  al. 2009. Liquid crystalline systems of phytantriol and glyceryl monooleate containing a hydrophilic protein: Characterisation, swelling and release kinetics. J. Pharm. Sci. 98(11), 4191–4204. Rizwan, S. B. et al. 2010. Bicontinuous cubic liquid crystals as sustained delivery systems for peptides and proteins. Expert Opin. Drug Del. 7(10), 1133–1144. Rizwan, S. B. et al. 2011. Preparation of phytantriol cubosomes by solvent precursor dilution for the delivery of protein vaccines. Eur. J. Pharm. Biopharm. 79(1), 15–22. Rosa, M. et al. 2006. Spontaneous formation of vesicles and dispersed cubic and hexagonal particles in amino acid-based catanionic surfactant systems. Langmuir 22(13), 5588–5596. Rosenbaum, E., Tavelin, S., and Johansson, L. B. A. 2010. A characterisation study on the application of inverted lyotropic phases for subcutaneous drug release. Int. J. Pharm. 388(1–2), 52–57. Rosenblatt, K. M., Douroumis, D., and Bunjes, H. 2007. Drug release from differently structured monoolein/ poloxamer nanodispersions studied with differential pulse polarography and ultrailtration at low pressure. J. Pharm. Sci. 96(6), 1564–1575. Sadoc, J. F. and Charvolin, J. 1986. Frustration in bilayers and topologies of liquid-crystals of amphiphilic molecules. J. Phys.-Paris 47(4), 683–691. Sadoc, J. F. and Charvolin, J. 1989. Ininite periodic minimal-surfaces and their crystallography in the hyperbolic plane. Acta Crystallogr. A 45, 10–20. Sagnella, S. S. et  al. 2011. Nanostructured nanoparticles of self-assembled lipid pro-drugs as a route to improved chemotherapeutic agents. Nanoscale 3(3), 919–924. Salentinig, S. et al. 2011. Transitions in the internal structure of lipid droplets during fat digestion. Soft Matter 7(2), 650–661. Seddon, J. M. 1990. Structure of the inverted hexagonal (Hii) phase, and non-lamellar phase-transitions of lipids. Biochim. Biophys. Acta 1031(1), 1–69. Seddon, J. M. et  al. 1996. An Fd3m lyotropic cubic phase in a binary glycolipid/water system. Langmuir 12(22), 5250–5253. Seddon, J. M. et al. 2000. Inverse micellar phases of phospholipids and glycolipids. Phys. Chem. Chem. Phys. 2(20), 4485–4493. Shah, J. C., Sadhale, Y., and Chilukuri, D. M. 2001. Cubic phase gels as drug delivery systems. Adv. Drug Del. Rev. 47(2–3), 229–250. Shah, M. H. and Paradkar, A. 2005. Cubic liquid crystalline glyceryl monooleate matrices for oral delivery of enzyme. Int. J. Pharm. 294(1–2), 161–171. Shah, M. H. and Paradkar, A. 2007. Effect of HLB of additives on the properties and drug release from the glyceryl monooleate matrices. Eur. J. Pharm. Biopharm. 67(1), 166–174. Shearman, G. C. et al. 2006. Inverse lyotropic phases of lipids and membrane curvature. J. Phys.-Condens. Mat. 18(28), S1105–S1124. Shearman, G. C. et al. 2009. A 3-D hexagonal inverse micellar lyotropic phase. J. Am. Chem. Soc. 131(5), 1678–1679. Siegel, D. P. and Banschbach, J. L. 1990. Lamellar inverted cubic (L-alpha Qii) phase-transition in N-methylated dioleoylphosphatidylethanolamine. Biochemistry 29(25), 5975–5981. Spicer, P. T. et  al. 2001. Novel process for producing cubic liquid crystalline nanoparticles (cubosomes). Langmuir 17(19), 5748–5756.

360

Liposomes, Lipid Bilayers and Model Membranes

Squires, A. M. et al. 2005. Kinetics and mechanism of the interconversion of inverse bicontinuous cubic mesophases. Phys. Rev. E 72(1), 011502. Swarnakar, N. K. et  al. 2007. Enhanced oromucosal delivery of progesterone via hexosomes. Pharm. Res. 24(12), 2223–2230. Templer, R. H. 1998. Thermodynamic and theoretical aspects of cubic mesophases in nature and biological amphiphiles. Curr. Opin. Colloid Interface Sci. 3, 255–263. Teres, S. et al. 2008. Oleic acid content is responsible for the reduction in blood pressure induced by olive oil. P. Natl. Acad. Sci. USA 105(37), 13811–13816. Tilley, A. et al. 2011. Transfer of lipid and phase reorganisation in self-assembled liquid crystal nanostructured particles based on phytantriol. Phys. Chem. Chem. Phys. 13(8), 3026–3032. Warren, D. B. et al. 2011. Real time evolution of liquid crystalline nanostructure during the digestion of formulation lipids using synchrotron small-angle X-ray scattering. Langmuir 27(15), 9528–9534. Winter, R. et  al. 1999. Inverse bicontinuous cubic phases in fatty acid/phosphatidylcholine mixtures: The effects of pressure and lipid composition. Phys. Chem. Chem. Phys. 1(5), 887–893. Winter, R. and Jeworrek, C. 2009. Effect of pressure on membranes. Soft Matter 5(17), 3157–3173. Yaghmur, A. et al. 2005. Emulsiied microemulsions and oil-containing liquid crystalline phases. Langmuir 21(2), 569–577. Yaghmur, A. et al. 2006a. Control of the internal structure of MLO-based isasomes by the addition of diglycerol monooleate and soybean phosphatidylcholine. Langmuir 22(24), 9919–9927. Yaghmur, A. et al. 2006b. Oil-loaded monolinolein-based particles with conined inverse discontinuous cubic structure (Fd3m). Langmuir 22(2), 517–521. Yaghmur, A. et al. 2007. Tuning curvature and stability of monoolein bilayers by designer lipid-like peptide surfactants. PloS One 2(5), e479. Yaghmur, A. et al. 2008a. Self-assembly in monoelaidin aqueous dispersions: Direct vesicles to cubosomes transition. PloS One 3(11), e3747. Yaghmur, A. et al. 2008b. Calcium triggered L alpha-H2 phase transition monitored by combined rapid mixing and time-resolved synchrotron SAXS. PloS One 3(4), e2072. Yaghmur, A. and Glatter, O. 2009. Characterization and potential applications of nanostructured aqueous dispersions. Adv. Colloid Interface Sci. 147–148, 333–342. Yaghmur, A. and Rappolt, M. 2010. Liquid crystalline nanoparticles as drug nanocarriers. In: M. Fanun, ed. Colloids in Drug Delivery. Karnataka: Taylor & Francis Group Press, 339–355. Yaghmur, A. et al. 2010a. Effects of pressure and temperature on the self-assembled fully hydrated nanostructures of monoolein-oil systems. Langmuir 26(2), 1177–1185. Yaghmur, A. et al. 2010b. Structural elucidation of light activated vesicles. J. Phys. Chem. Lett. 1(6), 962–966. Yaghmur, A. and Rappolt, M. 2011. Recent advances in the characterization of lipid-based nanocarriers. In: N. Garti and I. Amar-Yuli, eds. Nanotechnologies for Solubilization and Delivery in Foods, Cosmetics and Pharmaceuticals, Lancaster: DEStech Publications, 187–208. Yaghmur, A. et al. 2011a. In situ characterization of lipidic bupivacaine-loaded formulations. Soft Matter 7(18), 8291–8295. Yaghmur, A., Sartori, B., and Rappolt, M. 2011b. The role of calcium in membrane condensation and spontaneous curvature variations in model lipidic systems. Phys. Chem. Chem. Phys. 13(8), 3115–3125. Yaghmur, A. and Rappolt, M. 2012. Structural characterization of lipidic dispersions under non-equilibrium conditions. Eur. Biophys. J. Biophy. 41(10), 831–840. Yaghmur, A. et  al. 2012a. Characterization of bupivacaine-loaded formulations based on liquid crystalline phases and microemulsions: The effect of lipid composition. Langmuir 28(5), 2881–2889. Yaghmur, A., Sartori, B., and Rappolt, M. 2012b. Self-assembled nanostructures of fully hydrated monoelaidin-elaidic acid and monoelaidin-oleic acid systems. Langmuir. 28(26), 10105–10119. Zeng, N. et  al. 2012. Preparation and characterization of paclitaxel-loaded DSPE-PEG-liquid crystalline nanoparticles (LCNPs) for improved bioavailability. Int. J. Pharm. 424(1–2), 58–66.

18

Tethered Lipid Membranes Wolfgang Knoll, Renate L. C. Naumann, and Christoph Nowak

CONTENTS 18.1 Introduction .......................................................................................................................... 361 18.2 Peptide-Tethered Bilayer Lipid Membrane ........................................................................... 362 18.3 Oligo-Oxy-Ethylene (OEO)-Tethered Bilayer Lipid Membrane........................................... 362 18.4 Protein-Tethered Bilayer Lipid Membrane ...........................................................................364 18.5 Electronic Wiring of CcO Embedded in the ptBLM ........................................................... 367 18.6 FTIR Spectroscopy of Membrane Proteins .......................................................................... 371 References ...................................................................................................................................... 378

18.1 INTRODUCTION Membrane proteins constitute roughly one-third of all gene products (Tatulian 2003) and play a key role in cell adhesion, recognition, motility, energy production, transport of nutrients, and cholesterol. Nevertheless, the knowledge of the structure–function relationship for membrane proteins is still very limited, and lags behind that of soluble proteins (Popot and Engelman 2000, White and Wimley 1999, White et al. 2001). The handling of membrane proteins requires a lipid environment that closely matches the conditions in the living cell. Many questions pertaining to membrane processes or conformational changes of a membrane-based protein, due to an external perturbation, can be addressed by solid-supported or tethered lipid bilayers (Naumann et  al. 2003a, Schiller et al. 2003). These are novel model membrane platforms that allow for a simultaneous characterization of the structural and the functional aspects of membrane processes and the evaluation of the correlation between both (Knoll et al. 2004). Planar lipid bilayers on solid supports appear as promising alternative to liposomes, which are not amiable for the application of surface analytical techniques to investigate proteins in a functionally active form. However, developments were originally directed more toward understanding lipid bilayers rather than incorporating proteins (Schonherr et al. 2004). The obvious drawback of lipid bilayers, directly coating planar surfaces, is that membrane proteins often have domains that extend into the aqueous phase and consequently require to be displaced from the solid surface by a certain distance in order to avoid destructive interactions which may occur upon contact with the solid phase. Moreover, active proteins are often involved in the transport of ions, water, and small molecules. Hence, an aqueous submembrane reservoir is also necessary to accommodate these entities (Tanaka and Sackmann 2005). In order to solve these problems, linker or tether molecules were developed, which are attached on the one side to the proximal lealet of the lipid bilayer and on the other side provide functional groups for attaching a linker molecule to the surface. On gold surfaces this can be, for example, a sulfur functionality group. These so-called thiolipids self-assemble on the solid surface into a monolayer, which upon fusion with liposomes form a tethered lipid bilayer (tBLM) (Terrettaz et al. 2003). A tBLM schematic is shown in Figure 18.1.

361

362

Liposomes, Lipid Bilayers and Model Membranes

FIGURE 18.1 Schematics of a tBLM with the proximal and the distal lealet of the lipid bilayer composed of a self-assembled monolayer of a thiolipid and a conventional phospholipid, respectively.

18.2

PEPTIDE-TETHERED BILAYER LIPID MEMBRANE

In a irst attempt to construct a functional tBLM, we used thiopeptides as tether molecules to form peptide-tethered bilayer lipid membranes. The advantage of these systems is the rigidity of the spacer molecules, which were shown in some cases to be arranged in an α-helical conformation. This resulted in well-deined layer thicknesses of the submembrane and hence also a distinct membrane phase, as shown by surface plasmon resonance (SPR) spectroscopy. Large membrane proteins, such as the cytochrome c oxidase (CcO), isolated from bovine heart were readily inserted into the preformed bilayer by dilution of the solubilized enzyme below the critical micelle concentration (CMC) (Naumann et  al. 1999). Alternatively, F0F1 ATPase from chloroplasts as well the acetylcholine receptor (AChR) were incorporated by reconstituting the proteins irst in liposomes and then fusing the proteoliposomes with the thiolipid monolayer (Naumann et al. 1995, 1997, Schmidt et al. 1998). However, electrical impedance spectroscopy (EIS) showed insuficient sealing properties of the peptide-tethered BLM due to lateral packing incompatibilities of lipids with the attached peptide moieties. Nevertheless, CcO and F0F1 ATPase were both shown to be catalytically active and to transport protons across the membrane using square wave voltammetry, chronoamperometry, and EIS (Naumann 1997, 1999, 2002). Further, binding of a primary antibody and the speciic inhibitor α-bungarotoxin, respectively, to the AChR were demonstrated by surface-plasmon-enhanced luorescence and SPR spectroscopy. However, due to the poor electrical properties of these systems, any attempts to arrive at a quantitative understanding of transport phenomena were bound to fail.

18.3

OLIGO-OXY-ETHYLENE (OEO)-TETHERED BILAYER LIPID MEMBRANE

Regarding electric behavior of tBLM OEO, linkers have been shown to be much more appropriate as compared to tBLMs based on thiopeptides (Raguse et al. 1998). Therefore, we designed the archaea analog thiolipid 2,3-di-O-phytanyl-sn-glycerol-1-tetraethylene glycol-d,l-α-lipoic acid ester (DPTL) with a terminal sulfur functionality (Figure 18.2). Using DPTL, we obtained well-deined hydrophobic monolayers on ultrasmooth gold surfaces, which, upon fusion with liposomes, formed tBLMs with electric properties that closely matched the ones of freely suspended bilayer lipid membranes (BLMs) (Schiller et al. 2003). The formation of the tBLMs was

363

Tethered Lipid Membranes O O

O

O

O

O

O O

S S

FIGURE 18.2 Structure of the archaea analog thiolipid 2,3-di-O-phytanyl-sn-glycerol-1-tetraethylene glycol-d,l-α-lipoic acid ester (DPTL). (Reprinted with permission from Naumann, R. et al. 2003a. Tethered lipid bilayers on ultralat gold surfaces. Langmuir, 19(13), 5435–5443. Copyright 2003, American Chemical Society.)

followed by EIS (Figure 18.3), SPR, and QCM (quartz crystal microbalance) (Naumann et al. 2003a). Valinomycin reconstituted into such tBMLs showed a drop in resistance by four orders of magnitude upon increasing the concentration of K+ ions (Figure 18.4). The kinetics of K+ transport were modeled by electrical network simulations in terms of a kinetic scheme, originally developed for potassium transport by an ion carrier through BLMs (Naumann et  al. 2003b). Kinetic constants obtained from these early studies as well as electrical data of the tBLM measured before the addition of valinomycin were applied in this simulation. We found that the response of the pure tBLMs to 108

100

107

80 60

105 104

40

103

20

102 101

10–3 10–2 10–1 100 101 102 103 Frequency (Hz)

θ/deg

Z (Ω)

106

0 104

105

106

FIGURE 18.3 Electrical impedance spectra of the DPTL monolayer before (open circles) and after (open triangles) vesicle fusion. The equivalent circuit used for itting the data, Rex is the resistance of the bathing electrolyte solution, Rm and Cm are resistance and capacitance of the lipid membrane, Cdl is the capacitance of the diffuse double layer within the submembrane space. (Reprinted from Journal of Electroanalytical Chemistry, 550–551, Naumann, R. et al. Kinetics of valinomycin-mediated K-ion transport through tethered bilayer membranes, 241–252, Copyright (2003) with permission from Elsevier.)

364

Liposomes, Lipid Bilayers and Model Membranes

107

Rmem (Ωcm–2)

106

Membrane resistance as a function of K+ concentration

105

104

103 –0.5

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 49 50 51 cK+ (mM)

FIGURE 18.4 Membrane resistance Rm from electrical impedance spectra of a tBLM doped with valinomycin as a function of potassium ion concentration added to a sodium chloride solution (100 mM).

electric ields is very similar to the one reported previously for BLMs. Electrical impedance spectra of the tBLMs were measured as a function of externally applied potentials. Resistance changes were modeled by electrical network simulations in terms of permeability changes due to passive ion transport across the membrane. The dependence of the permeability on the potential had to be taken into account in order to it the experimental results (Robertson et al. 2008). This strongly suggests that a carrier model of passive ion transport is also valid in BLMs. A further beneit of these network simulations is that not only the lipid bilayer is taken into account, but the entire tBLM architecture, including the submembrane space given by the linker molecule. This gave us the opportunity to estimate, for example, the evolution of the drop of the electric potential across the different layers as a function of different parameters, such as time, presettings of the potential, or concentration of ions in the aqueous phase. Hence, a tBLM was constructed, whose properties closely matched the ones of freely suspended BLM. This conclusion was supported by the inding that similar OEO-based tBLMs could be realized on Hg (Becucci et al. 2005) as well as silicon semiconductor surfaces (Atanasov et al. 2005). These results seemed quite promising. However, our expectations that the OEO spacer would mimick a water-illed submembrane layer were not fulilled. A thorough investigation of DPTL by polarization modulation-infrared relection–adsorption spectroscopy (PM-IRRAS) revealed that the OEO spacer is a densely packed hydrophobic moiety, in which very few water molecules can be accommodated (Leitch et al. 2009). This explained why we were not able to incorporate proteins more complex than the carrier valinomycin into the DPTL-based tBLMs. Even transport of K+ ions by valinomycin seems to be an unlikely event, unless they are able to partition into the hydrophobic OEO structure. Thus, even small channel proteins such as gramicidin and melittin were not successfully reconstituted in the hitherto described tBLMs in a functionally active form. Incorporation of these channels was accomplished, only when monolayers of thiolipids were mixed with an OEO-terminated dilution molecule (Baumgart et al. 2003, He et al. 2005). However, electrical properties of these mixed layers were too poor to justify any further quantitative treatment of transport processes by simulations.

18.4

PROTEIN-TETHERED BILAYER LIPID MEMBRANE

To overcome the problems described in the previous section, we developed a completely different strategy, using the membrane protein of interest as the essential building block. This idea resulted in the concept of the protein-tethered bilayer lipid membrane (ptBLM), which was irst developed and

Tethered Lipid Membranes

365

veriied by Giess et al. (2004). It is based on the immobilization of a large membrane protein such as cytochrome c oxidase (CcO) from R. sphaeroides provided with His-tags attached to subunit (SU) I. The necessary modiication of the substrate surface is based on the well-established concept of metal ion afinity chromatography. Ion-chelating nitrilotriacetic acid groups bearing a terminal amino group (ANTA) are coupled in situ to an N-hydroxy succinimide (NHS) ester functionalized surface. The resulting nitrilotriacetate (NTA) surface is activated by complexation of Cu2+ or Ni2+ ions rendering it capable of reversibly immobilizing His-tagged proteins. The His-tagged membrane protein is irst attached to the surface in its detergent solubilized form. In the second step, the detergent molecules are substituted by lipid molecules by in situ dialysis, thus forming a lipid bilayer that is tethered to the support by the protein itself (Figure 18.5). The coupling by His-tags provides suficient intramolecular lexibility to the reconstituted protein to become active. Above all, it renders the proposed method universally applicable to all Histagged membrane proteins. Further beneits are the strict control of the orientation as well as the packing density of the protein by choosing the proper surface concentration of the chelate. Binding of the protein as well as insertion of the lipids was measured by SPR, QCM, and EIS. The thicknesses of the dithiobis(succinimidyl propionate) self-assembled monolayer (DTSP-SAM), and the thickness changes due to the coupling of ANTA, respectively, are both too small to be accurately determined by SPR. However, the formation of these layers can be clearly detected by surface-enhanced infrared absorption spectroscopy (SEIRAS) (Ataka et al. 2004). The electrical properties of the ptBLMs measured by impedance spectroscopy (Figure 18.6) showed resistances well in the MΩ cm2 range known from BLMs and tBLMs. However, the capacitances were one order of magnitude higher. This is understood in terms of the large portion of proteins in these layers. The surface concentration of CcO was, for example, determined to be more than 90% in independent measurements (see further in the following text). Hence only a few lipid molecules are inserted in between the proteins as shown by the decrease of the capacitance after in situ dialysis. Hence, the capacitance is dominated by the dielectric constant of the protein (ε = 10) rather than the lipid (ε = 2). To test whether or not the reconstituted CcO is functionally active under these conditions, reduced cytochrome c was added to the oxygenated solution. The impedance spectrum showed a drastic decrease in resistance, indicating that the protein actively transports protons during the

FIGURE 18.5 (See color insert.) Formation of the ptBLM: Solubilized CcO (detergent molecules are marked in blue), His-tagged to subunit I are immobilized on an amino-nitrilotriacetic acid (ANTA)functionalized gold surface, micelles of solubilized phospholipid are added and dialysis is performed in situ by adding biobeads (lipid molecules are indicated by red polar heads and yellow lipid tails). As a consequence of detergent removal, phospholipids assemble around the protein to form a lipid bilayer. (With kind permission from Springer Science + Business Media: Soft Matter, Conformational transitions and molecular hysteresis of cytochrome c oxidase: Varying the redox state by electronic wiring, 6(21), 2010, 5523–5532, Nowak, C. et al.)

366

Liposomes, Lipid Bilayers and Model Membranes (a) 7

2

5 4 3

Reflectivity (a.u.)

Thickness (nm)

6

1

2 1

0.8 0.6 0.4 0.2 0.0 46 48 50 52 54 56 58 60 62 θ (degrees)

0

50 100 150 200

600 1200 Time (min)

1800

(b) 0.0

5.0 × 10–6 Y'' (ω)

3 2

1.0 × 10–5

1 1.5 × 10–5 0.0

5.0 × 10–6 1.0 × 10–5 Y' (ω)

(c) 108

–100 b

1

–80

IZI

106 –60

105 104 103

3

2

–40 –20

Phase angle (°)

107

1.5 × 10–5

102

0 101 10–4 10–3 10–2 10–1 100 101 102 103 104 105 106 Frequency (Hz)

FIGURE 18.6 (a) Kinetic trace of an SPR spectrum recorded during CcO binding (1) and reconstitution (2) at a constant angle of incidence (55°) and transferred into an optical thickness by using the Fresnel equation. Impedance spectra, frequency normalized admittance (b), and Bode’s plots (c), of the Ni-ANTA modiied surface before (1) after binding of CcO before (2) and after reconstitution of the protein (3). Dotted lines represent experimental data, solid lines show the curves itted to an equivalent circuit. Parameter values of CcO before and after reconstitution: Cm = 15.1 ± 1.9 and 7.3 ± 0.5 µF cm−2, Rm = 0.3 ± 0.08 and 12 ± 7 MΩ cm2, respectively. Rm and Cm are the resistance and capacitance of the protein-membrane layer. (Reprinted from Journal of Electroanalytical Chemistry, 649, Schach, D. et al. Modeling direct electron transfer to a multi-redox center protein: Cytochromecoxidase, 268–276, Copyright (2010) with permission from Elsevier.)

367

Tethered Lipid Membranes

catalytic redox cycle (Giess et al. 2004). This was corroborated by cyclic voltammetry (CV), showing catalytic currents in the presence of reduced and oxidized cytochrome c within the oxygenated solution (Friedrich et al. 2008). Simulations of the voltammograms employing DigiSim, a commercially available program, indicated relatively high turnover rates.

18.5

ELECTRONIC WIRING OF CcO EMBEDDED IN THE ptBLM

The concept of the ptBLM was irst applied by Friedrich and coworkers to the CcO from R. sphaeroides with the His-tag attached to SU II (Friedrich et al. 2008). In this case, the enzyme is immobilized with the cytochrome c binding side directed toward the electrode. Electrons can be expected to be transferred to CuA and from there to the remaining redox centers and inally to oxygen (Figure 18.7). Moreover, under aerobic conditions protons would be pumped in the direction opposite to electron transfer such that they could be reduced at the electrode to form hydrogen. Electrochemically controlled reduction of the heme centers has irst been indicated by surfaceenhanced resonance Raman spectroscopy (SERRS) of CcO incorporated in a ptBLM under strictly anaerobic conditions (Figure 18.8). SERR spectra measured on the surface were consistent with the ones measured in solution regarding sensitivity and resolution, before and after electrochemical and chemical reduction, respectively (Friedrich et al. 2004). Different redox states of CcO were characterized particularly by the ν4 modes at 1358 and 1370 cm−1 originating from both hemes in the reduced (−350 mV) and oxidized states (−150 mV), respectively (Friedrich et al. 2008, Heibel et al. 1993). Other characteristic bands are the stretching modes between 1610 and 1680 cm−1, indicative of conjugated vinyl and formyl substituents of type-a hemes, the bands at 1663 and 1671 cm−1 due to the formyl substituent of heme a3, and the characteristic marker band for the reduced heme a at 1518 cm−1 (ν11) not present in the oxidized state. We concluded that the ptBLM offers direct access to the electron transfer (ET) pathway of the protein. SERR spectroscopy, however, indicates redox states of the hemes rather than the CuA and CuB centers. In order to ind out whether or not all four redox centers are reduced and oxidized by electrochemical excitation, the electrochemical behavior of CcO under conditions of direct ET was investigated in more detail. CVs were taken with CcO from R. sphaeroides with the His-tag attached to SU II under strictly anaerobic conditions and before the enzyme had any contact with oxygen (Figure 18.9a). Measurements showed peaks in the range of 200–600 mV, corresponding to midpoint potentials of redox centers known from independent measurements, whereas CcO immobilized in the reverse orientation with the His-tag attached to subunit I exhibited no peak under

Heme a3

CuB O2 Heme a CuA e– H+ Au/Ag

FIGURE 18.7 Pathway of electrons and protons within CcO from R. sphaeroides embedded into a ptBLM when CcO is immobilized via the His-tag attached to SU II.

368

Liposomes, Lipid Bilayers and Model Membranes 1370

A 1585 1570

1647 1635

1671

I (a.u.)

B

C

1358

1611 1584 1621 1567

D

1518

1663

1473

1350

1400

1450

1500 Δv

1550

1600

1650

(cm–1)

FIGURE 18.8 SERR spectra of CcO tethered to a Ni-NTA functionalized Ag electrode and reconstituted with DPGPC, compared with the RR spectra of CcO in solution. Panel a: RR, fully oxidized; panel b: SERR, open circuit; panel C: SERR, −0.65 V; panel D: RR, fully reduced CcO by addition of dithionite. SERR and RR spectra were obtained with 413 nm excitation using a LabRam spectrograph (HR800; Jobin-Yvon). The laser beam (2.8–5.0 mW) was focused with a 20× objective onto a lat disk-shaped stationary Ag electrode in a home-built electrochemical cell. The electrode was in contact with a buffered solution containing 100 mM KCl and 50 mM potassium phosphate (pH 8). The same buffer solution but containing 25 µM CcO and 0.1% DDM was used for the RR measurements that were carried out with the set-up described previously (Naumann et al., 2003). (Friedrich et al. 2004. Active site structure and redox processes of cytochrome c oxidase immobilised in a novel biomimetic lipid membrane on an electrode. Chemical Communications, 21, 2376–2377. Reproduced by permission of The Royal Society of Chemistry.)

anaerobic conditions. Changing to an air-saturated solution, the enzyme started to work under turnover conditions. Two peaks appeared in the negative potential range, slightly shifting in the positive direction with successive scans (Figure 18.9b). Finally, peak potentials under steady state conditions of 242 and 530 mV were reached. These peaks were attributed to repeated electron and proton transfer, characterized by the ampliied current density (Friedrich et al. 2008, Schach et al. 2010). Upon returning to anaerobic conditions, the electron transfer peak persisted, whereas the proton transfer peak disappeared as expected. In accordance with previous indings, the electron transfer peak

369

Tethered Lipid Membranes (a)

90

j/μA cm–2

60 30 0 –30 –60 –0.6 (b)

–0.4

–0.2

0.0 0.2 E/V vs. SHE

0.4

0.6

0.8

50 0

j/μA cm–2

–50 –100 –150 –200 –250 –300 –1.0 –0.8 –0.6 –0.4 –0.2 0.0 E/V vs. SHE (c)

0.2 0.4

0.6

0.4

0.6

60 40

j/μA cm–2

20 0 –20 –40 –60 –80 –0.6

–0.4

–0.2

0.0 0.2 E/V vs. SHE

FIGURE 18.9 Cyclic voltammograms of CcO immobilized via a His-tag on subunit II, that is, with CuA directed toward the electrode. Panel a shows data under anaerobic conditions before activation (scan rate/V s−1 0.05, 0.1, 02, 0.4, 0.8, 1.6), panel b upon evolution of the protein catalytic activity under aerobic conditions, that is, activation (1st, 5th, 10th, and 20th scan with scan rate 0.05 V s−1), and panel c under anaerobic conditions after activation (scan rate/V s−1 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, and 1 V). Peak heights in Figure 18.9a and c increase linearly with scan rate in accordance with ET of the adsorbed species. (Reprinted from Biophysical Journal, 94, Friedrich, M.G. et al., Electronic wiring of a multi-redox site membrane protein in a biomimetic surface architecture, 3698–3705, Copyright (2008) with permission from Elsevier.)

370

Liposomes, Lipid Bilayers and Model Membranes

was shifted by more than 400 mV toward the negative direction (Figure 18.9c). The cyclic voltammogram returned to its original form (Figure 18.9a), if the protein was incubated under anaerobic conditions for at least 60 min. We deduced from these results that the enzyme is catalytically active and that it takes up four electrons under aerobic and anaerobic conditions, respectively. Moreover, it undergoes a gradual transition from a nonactivated (Figure 18.9a) to an activated conformational state (Figure 18.9c) when it passes through a number of redox cycles during turnover (Figure 18.9b). This is equivalent to the transition from the resting to the pulsed state when CcO, reconstituted in liposomes, is subjected to oxygen pulses (Antonini et al. 1985, Brunori et al. 1985, Jancura et al. 2006). Therefore, we decided to analyze the ET for a multiredox site protein using modeling based on a rigorous electrochemical theory (Schach et al. 2010). The theory, however, does not consider protonations as well as catalytic turnover. As such the theory applies to the electrochemically controlled reduction/ oxidation under strictly anaerobic conditions, where proton pumping is inhibited. The general concept is fully described in Schach et al. (2010). Briely, we discriminate the transfer of electrons from the electrode to a center (denoted as “uptake”) and by exchange of electrons between two centers (denoted as “exchange”). The various redox states of the enzyme are considered as different conformational states, and the k th conformation has the probability pk, hence

∑p

k

= 1,

k

(18.1)

where the sum includes all conformational states. The transition between two states k and l is described by the low Jk,l = kk,l pk – kl,k pl.

(18.2)

The rate coeficients kk,l and kl,k depend on the type of electron transfer and the centers involved. In the case of electron uptake, they also depend on the applied potential E according to kk,l (E) = ke,i exp[(Eo,i – E)/(2φn)] and kl,k (E) = ke,i/exp[(Eo,i – E)/(2φn)], (uptake),

(18.3)

φn = RT/F.

(18.4)

where

In Equation 18.3, Eo,i denotes the standard potential of the ith center (which is reduced in this transfer), and ke,i is the rate constant of the electrochemical reaction. A symmetrical energy barrier is assumed in Equation 18.3. Electron exchange between centers i and j corresponds to a chemical reaction. Hence, the rate coeficients are independent of E. The electron exchange is described by a forward rate constant ki,j and a backward rate constant which follows from detailed balancing (Hill 1977) kk,l = ki,j and kl,k = kj,i exp[(Eo,i – Eo,j)/φn] (exchange).

(18.5)

The low of electrons associated with the ith center, expressed as current density ji, becomes ji = −Γ F [ΣJk,l|→i − ΣJk,l|i→],

(18.6)

where Jk,l|→i and Jk,l|i→ denote the lows into the ith center and out of it, respectively, and Γ the surface coverage. The current density jel lowing through the electrode is then given by

371

Tethered Lipid Membranes

jel = Σi ji = −ne FΣ Jk,l|up,

(18.7)

where Jk,l|up denotes lows pertaining to electron uptake. On the basis of these general principles, two limiting cases were considered. i. Sequential electron transfer (ECCC mechanism). Center 1 can take up electrons from the electrode. The other centers exchange electrons with their neighbors. The kinetic scheme of this electron transfer is shown in Schach et al. (2010). If applied to CcO the centers 1, 2, 3, and 4 represent CuA, heme a, heme a3, and CuB, respectively. ii. Independent electron transfer (EEEE mechanism). All centers can take up electrons directly from the electrode but do not exchange electrons with their neighbors. Since a redox interaction is excluded, this model can be represented by four electron uptakes (cf. Equation 18.4), as shown by the kinetic scheme in Figure 18.2b of Schach et al. (2010). In this case, no unambiguous assignment of centers 1–4 to the redox centers in CcO is possible. Both cathodic and anodic branches of baseline-corrected CVs, taken over a broad range of scan rates between 0.05 and 8 V s−1, were analyzed simultaneously. The parameter values are reported in Schach et al. (2010), from which we mention the electrochemical rate constants to CuA for the reductive and oxidative branches, ke,1 = 368 and 370 s−1, respectively, as well as the surface concentration of CcO molecules of Γ ~ 6 pmol cm−2. The analysis provides strong evidence that direct ET to CcO in the activated state follows the sequential model (ECCC mechanism). Thus, electronic wiring can be considered as equivalent to ET from the genuine electron donor of CcO, cytochrome c. Independent ET to each center separately (EEEE mechanism) can be excluded, since the analysis failed when both branches of the CVs were considered simultaneously. Moreover, the values for the electrochemical rate constants, obtained when analyzing the two branches separately, are signiicantly different, thus violating basic physicochemical principles. In the case of nonactivated CcO, discrimination between the two mechanisms is not possible because of the interference of other processes. However, an E-range restricted analysis showed that the CVs are still compatible with sequential ET. The electrochemical rate constant for ET to CuA is well in the range found for other proteins (Armstrong 2002a,b, Jeuken et al. 2002). The discrepancy in an earlier report (Friedrich et al. 2008) is most certainly due to differences in preparations of CcO, which was obtained from different sources. Intraprotein electron exchange is known to be much faster than ET to CuA, hence the pertinent rate constants could not be itted. ET between these centers is then always close to equilibrium, which can be simulated by assigning large and constant values to the rate constants. The found average value for the surface coverage Γ (≈ 6 pmol cm−2) is in excellent agreement with 6 pmol cm−2 estimated for a densely packed monolayer of CcO. For this estimate, we assumed an ellipsoidal disk of 4.5 × 7.0 nm for the in-plane dimension of CcO, which can be deduced from crystal structure data of R. sphaeroides (Svensson-Ek et al. 2002). Therefore, we conclude that the ptBLM consists of a densely packed monolayer of CcO interspersed with a small number of lipid molecules, as also indicated by EIS and SPR measurements. Information on the exact pathway of electrons in CcO embedded in a ptBLM is important for several aspects. For example, spectro-electrochemical measurements using SEIRAS have revealed conformational changes that crucially depend on the pathway of the electrons through the enzyme (Nowak et al. 2010). Additionally, the analysis of time-resolved spectro-electrochemical measurements performed in our laboratory requires a well-deined model of ET.

18.6 FTIR SPECTROSCOPY OF MEMBRANE PROTEINS The use of FTIR spectroscopy of studying phospholipids or polypeptide architectures, as well as proteins, is well documented (Braiman and Rothschild 1988, Muller et  al. 1996, Susi 1972).

372

Liposomes, Lipid Bilayers and Model Membranes

The label-free beneits of infrared spectroscopic methods enable the observation of biomolecules with minimal perturbation to the membrane assembly (Schultz and Levin 2011). Both lipids and proteins possess various chemical functional groups that allow for the monitoring of component interactions associated with such membrane systems described before. Useful vibrational modes relect, in general, chemical functional group stretching and bending motions (Schultz and Levin 2011). Additionally, hydrogen bonding and other interactions among lipids, proteins, and the membrane–water interface can be observed as broadening effects, peak frequency shifts, and splitting of spectral features (Barth 2000, Nowak et al. 2011a, Schultz and Levin 2011). In addition to chemical interactions, vibrational spectra are sensitive to membrane characteristics such as the lipid phases of the membrane bilayer and the orientation and conformation of the membrane proteins within the lipid bilayer system (Schultz and Levin 2011, Tatulian 2003). In addition to the power and potential of vibrational spectroscopy, direct observation, at times aided by isotopic enrichment, of the vibrational features of speciic moieties within a membrane system does not require perturbing labels that are utilized, for example, in spectroscopic techniques such as luorescence spectroscopy (Schultz and Levin 2011). In the ield of infrared spectroscopy of proteins or biomimetic membrane systems, attenuated total relection IR spectroscopy (ATR-IR) has become the most frequently used technique to characterize such systems (Ataka and Heberle 2004, Baurecht et al. 2002, Giess et al. 2004, Nowak, C. et al. 2009a, 2011a). In ATR-IR experiments, an internal relection plate made from silicon (Si), germanium (Ge), or zinc-selenide (ZnSe), to name the most commonly used materials, is covered with the sample of interest, like the protein–lipid system, and the infrared beam is focused into the plate. The light travels inside the plate by means of a series of internal relections from one surface of the plate to the other, creating an exponentially decaying evanescent radiation ield outside the plate (Tatulian 2003). ATR-IR spectra on supported membranes with reconstituted proteins contain a wealth of information about the structure of the system (Barth 2000). Useful features and theoretical aspects about the method of ATR-FTIR spectroscopy have been reported in previous reviews (Fringeli 1992, Goormaghtigh et al. 1999, Tamm and Tatulian 1997). Because the decay length of the evanescent wave (~0.2–0.6 µm) extends far beyond the dimensions of the largest proteins, there are no concerns regarding the size of the protein (Tatulian 2003). The decay length of the evanescent wave is constant within the measuring volume for the spectra. Therefore, the ATR setup is very useful for measuring samples under aqueous conditions, since the contribution from water, which as a strong absorber in the infrared, can be subtracted. Additionally, because of large wavelengths of infrared light, light scattering problems are no issue as compared to circular dichroism or luorescence experiments in the UV region (Tatulian 2003). One of the most powerful methods for the identiication of small amounts of adsorbed molecules and their molecular structure is SEIRAS (Huo et al. 2005, Osawa 2001, Osawa et al. 1993). Like its counterpart, surface-enhanced Raman scattering (SERS), (Nowak et al. 2009a, Otto et al. 1992), SEIRAS relies on the morphology of the metal surface (Osawa 2001). Thin metallic ilms consisting of metal islands (Pucci 2005), as well as monolayers of nanoparticles (NPs), are good candidates (Nishikawa et al. 1993, Osawa 2001). The surface enhancement is considered in terms of local enhancements of the electronic ield induced by local structures (Nowak et al. 2009a). Local ield enhancement preferably occurs in the narrow gap between single islands or NPs, smaller in size than the wavelength of light, which are kept at distances wide enough not to touch each other (Nowak et al. 2009a). SEIRAS has been applied on colloidal gold ilms assembled on attenuated total relection crystals for in situ monitoring of the adsorption and chemical reaction of monomolecular layers (Ataka et al. 2004, Nishikawa et al. 1993, Pucci 2005). Combining electrochemistry with IR-spectroscopy, we introduced a two-layer gold surface (Nowak et  al. 2009a). The preparation of these substrates begins with a 25-nm-thin gold layer by thermal evaporation on the Si ATR crystal. This thickness provides the limiting condition to achieve an electrical conducting layer of the evaporated gold (Nowak et  al. 2009a). Protruding

373

Tethered Lipid Membranes

structures on the 25-nm-thick gold surface were subsequently used as seed crystals. AuNPs will grow predominantly on these sites rather than illing gaps in the loose underlayer (Nowak et al. 2009a). The size of these NPs and, at the same time, the enhancement factor can be controlled by systematically varying the growth conditions as shown in previous studies (Enders et  al. 2006, Nowak et al. 2009a). The effect of islands percolating to larger structures can be observed by AFM revealing growth times of more than 10 min, shown in Figure 18.10. The optimal enhancement effect was achieved after 10 min of growth time. The enhancement factor with a value of 127.8 was 5.8 times higher than the value reported by others (Ataka and Heberle 2006, Osawa 2001). The preparation of the newly developed two-layer surface was greatly facilitated, while the thickness could be controlled very precisely (Nowak et al. 2009a). Cytochrom c (cc), a commercially available redox protein from the respiratory chain, was used as a benchmark system to test the spectroelectrochemical abilities of the developed two-layer gold surface (Nowak et al. 2009b). After immobilization of cc to the mercaptoethanol-modiied two-layer

t = 2 min

6 4 2 0 –2 –4

t = 6 min

6 4 2 0 –2 –4

t = 10 min

6 4 2 0 –2 –4

Height (nm)

t = 0 min

6 4 2 0 –2 –4

t = 12 min

6 4 2 0 –2 –4 –6 0.0

0.4

0.8 1.2 1.6 Latitude (μm)

2.0

FIGURE 18.10 (See color insert.) AFM images of Au ilms obtained with different growth times (0, 2, 6, 10, and 12 min as indicated on the line scans). (Reproduced from Nowak, C. et al. 2009a. A two-layer gold surface with improved surface enhancement for spectro-electrochemistry using surface-enhanced infrared absorption spectroscopy. Applied Spectroscopy, 63(9), 1068–1074 with permission from the Society for Applied Spectroscopy.)

374

Liposomes, Lipid Bilayers and Model Membranes

gold surface, spectral changes of the redox protein were recorded as a function of the applied potential. Difference spectra were taken at different potentials, using the reduced state (–103 mV vs. SHE (SHE = Standard Hydrogen Electrode)) as reference (Figure 18.11). The potential was varied in steps of 50 mV between +497 and −103 mV vs. SHE. Changes in the spectra were observed at wavenumbers in the range of 1500–1800 cm−1. The strong band at 1692 cm−1 was assigned to the amide I band of the β-turn type III of amino acids 14–19 (including His-18) and 67–70 of reduced cc (Ataka and Heberle 2004, Nowak et al. 2009b). His-18 is the axial ligand of the porphyrine ring of the heme structure inside the cc. Hence, the decrease of this band was considered to directly relect the transition of cc into the reduced state. It was correlated with an increase of the band at 1672 cm−1 which had been attributed to the same type of amide I band of oxidized cc (for the band assignments of cc, see Ataka and Heberle 2004). The strong positive band at 1658 cm−1 and the negative band at 1667 cm−1 were assigned to an amide I band of the β-turn type II of amino acids 32–35 and 35–38 of oxidized and reduced cc, respectively (with slight shifts of the frequencies compared with (Ataka and Heberle 2004)). The positive band at 1552 cm−1 indicated an amide II band of the β-turn type III type of oxidized cc, which is more remote from the heme structure. The small negative band at 1624 cm−1 indicated the extended β-strand of cc in the reduced form. These changes were very similar to the spectra observed on the classical SEIRA surface, at least as far as the band positions are concerned. The relative intensities of the bands measured on the two-layer gold surface, however, are more similar to the ones on mercaptopropionic acid than the ME-modiied classical SEIRA surface reported in Nowak et al. (2009b). This might be due to different morphologies of the two gold layers. The bands are more or less correlated with the redox transition as can be deduced from a plot of the amplitudes vs. potential. For example, the band at 1692 cm−1 yielded a sigmoidic curve, which can be itted to the Nernst equation with a slope of 36 mV vs. absorbance and an inlection point at 217 mV, in reasonable agreement with the theoretical slope of 25 mV and the standard redox potential of cc, E 0 = 230 mV (Haas et  al. 2001). The band at 1552 cm−1 yielded also a sigmoidic function with an inlection point at 220 mV, but a shallow slope of 130 mV vs. absorbance. This is in agreement with a band originating from the peptide group which is located at a greater distance 3.0 × 10–3

1672

1552

1658

Absorbance (a.u.)

2.0 × 10–3

1.0 × 10–3

0.0

–1.0 × 10–3

1667

–2.0 × 10–3

1624

1692 amide | β-turn type III –3.0 × 10–3 1900

1850

1800

1750

1700

1650

1600

1550

1500

Wavenumber (cm–1)

FIGURE 18.11 Potential dependent spectra of cc by ATR-IR-spectroscopy. Potential differences applied across the immobilized protein were varied stepwise from the fully reduced to the fully oxidized state. Difference spectra are calculated vs. the fully reduced state at –103 mV vs. SHE. (Reprinted with permission from Nowak, C. et al. 2009b. Electron transfer kinetics of cytochrome C in the submillisecond time regime using time-resolved surface-enhanced infrared absorption spectroscopy. Journal of Physical Chemistry C, 113(6), 2256–2262. Copyright 2009, American Chemical Society.)

Tethered Lipid Membranes

375

from the redox center and is therefore less correlated with the redox transition. An entirely unspeciic orientational or other change of a peptide in an electric ield would be indicated by a straight line, as was observed previously for molecules, which do not undergo any faradaic process (Noda et al. 1999, Nowak et al. 2009a). Finally, potential-controlled time-resolved (tr)-SEIRA measurements were conducted in the rapid-scan and the step-scan mode with different square-wave excitation frequencies (Nowak et al. 2009a). Regarding these excitations, one should bear in mind that changes (conformational or orientational) of different peptide groups are indicated by tr-SEIRAS at different exciting frequencies. For example, the band at 1692 cm−1 had the highest amplitude at an exciting frequency of 500 Hz (Nowak et  al. 2009a). As discussed earlier, this band directly relects the redox transition of cc, since it indicates a conformational or orientational change of the amino acid directly attached to the central iron atom of the porphyrin ring (Nowak et al. 2009a). Therefore, the frequency of the tr-SEIRA spectrum with the highest amplitude was considered to match the frequency of the redox transition. Changes of other peptide groups were excited at much lower frequencies. For example, the band at 1552 cm−1 was excited at 1.85 Hz. This band is related to a remote peptide group (discussed earlier). It was more prominent at even lower frequencies, for example, during a rapid-scan measurement taken at 0.7 Hz. Changes of peptide groups represented by these bands were regarded as not directly correlated with the redox process. Further, time-dependent changes of the bands not only at 1692 cm−1 but also at 1663, 1644, and 1625 cm−1 were observed in tr-SEIRA spectra obtained at 500 Hz. Their amplitudes clearly exceed the noise level of adjacent frequencies. The entire spectroelectrochemical measurements on cc are described in detail in Nowak et al. (2009a). The same potential-dependent IR measurements were performed on cytochrome c oxidase immobilized in the earlier described biomimetic membrane system (Nowak et al. 2010). However, these IR spectra were more complex than those obtained from cc. Due to strong overlaps of bands within the amide I region, a deconvolution method had to be found. Besides the commonly used methods for deconvoluting those spectra-like Fourier self-deconvolution (Wi et al. 1998) or calculating the second derivative of a given IR spectrum, different experimental techniques can be used to reduce the number of overlapping bands (Muller et al. 1996, Noda 1990). Another possibility is the application of modulation spectroscopy (Muller et al. 1996). In such a modulation experiment, an external parameter like pressure, temperature, concentration or, like in our case, the applied electrical potential is varied. There are two different methods to analyze modulation spectra. One of these, the 2D correlation spectroscopy, was introduced by Noda (1990). The basic concept of 2D IR is somewhat analogous to 2D correlation techniques used in NMR (Noda 1990). However, the speciic experimental procedure developed for 2D IR is substantially different from that used in 2D NMR (Muller et al. 1996, Noda 1990). A detailed description about 2D IR can be found in these references (Czarnecki et al. 1998, Ganim et al. 2008, Jung et al. 2000, Noda 1990, Noda and Ozaki 2004, Ozaki et al. 2003). 2D IR offers increased structural resolution by spreading the spectra over a second frequency dimension, revealing two-dimensional line shapes and cross-peaks (Czarnecki et al. 1998, Ganim et al. 2008, Jung et al. 2000, Noda 1990, Noda and Ozaki 2004, Ozaki et al. 2003). 2D IR spectroscopy is now commonly used especially for time-resolved IR spectroscopy up to the picosecond time regime, making it an excellent choice for understanding protein dynamics (Ganim et al. 2008, Nowak et al. 2011a). Further it is also appropriate to disentangle IR bands, which strongly overlap with each other (Noda 1990, Nowak et al. 2011a). From previous electrochemical results (see section “Electronic wiring of CcO embedded in the ptBLM” within this chapter), we deduced that the enzyme undergoes a gradual transition from a nonactivated to an activated conformational state when the enzyme, under aerobic condition, passes through a number of redox cycles (Ganim et al. 2008, Nowak et al. 2011a). This is equivalent to the transition from the resting to the pulsed state, when the CcO reconstituted in liposomes is subjected to oxygen pulses. This was explained in terms of a conformational transition of CcO, consistent with a change in the environment of the heme and Cu centers, in the course during which we are

376

Liposomes, Lipid Bilayers and Model Membranes

able to capture the “fast” or “pulsed” state of the enzyme (Ganim et al. 2008, Nowak et al. 2011a). In order to characterize all redox transitions, a broad potential range was covered in the potentiometric titrations. In both cases, potential-dependent spectra were dominated by broad positive bands in the amide I region. Smaller bands were found in the ingerprint region. Tentative band assignments were performed according to these works (Gorbikova et al. 2006, Hellwig et al. 1998, 1999, Iwaki et al. 2006) on the basis of potentiometric titrations of the CcO in solution, which had been performed independently by several groups in the presence of mediators using simultaneous UV–VIS and FTIR spectroscopy. Speciic changes in the UV–VIS spectra of the chromophores as a function of the redox state were used to assign vibrational modes in the IR to particular redox centers, including the amide I region. Other applied criteria were the sigmoid potential dependence of the absorbance of these bands, which are characteristic of a redox transition. Unspeciic changes, merely due to the electric ield, do not yield sigmoid dependencies. An attempt was made to separate such unspeciic contributions from speciic entities (Nowak et al. 2010). Synchronous and asynchronous 2D-IR correlation maps were constructed from SEIRA spectra of the CcO in the activated (Figure  18.12) and nonactivated states (Figure 18.13), respectively, accounting separately for the amide I 1550–1750 cm−1 (Figures 18.12a and b, 18.13a and b) and the ingerprint region 1300–1550 cm−1 (Figures 18.12c and d, 18.13c and d). The 2D data were used to deconvolute the 1D SEIRA spectra into single bands. This was performed mostly on the basis of potentiometric (a) 7.0 × 10–3 6.0 × 10–3 1682

5.0 × 10–3 4.0 × 10–3

1700

3.0 × 10–3 2.0 × 10–3 1742

1600 1682

5.0 × 10–3

1623 1586 1565

4.0 × 10–3

1560

1590

1590 Wavenumber (cm–1)

1560

1680

1565

1620 1650 1680

1710

1710

1740

1740 1740 1710 1680 1650 1620 1590 1560 Wavenumber (cm–1)

1586

10–3 0.0

1650

1623

2.0 × 10–3 1742

0.0

1620

1700

3.0 × 10–3

1.0 ×

1.0 × 10–3

Wavenumber (cm–1)

1654

6.0 × 10–3

1600 Absorbance

Absorbance

(b) 7.0 × 10–3

1654

1740 1710 1680 1650 1620 1590 1560 Wavenumber (cm–1)

FIGURE 18.12 Asynchronous (left) and asynchronous (right) 2D correlation maps of activated CcO in the amide I region. Potential applied was 900 mV vs. the SHE for the fully oxidized state and varied from 500 to −700 mV in 100 mV steps for reduced states. No mediators were added. The upper panel shows one of the 1D-SEIRA difference spectra (taken at −700 mV) as an example for the deconvolution. (Nowak, C. et  al. 2011b. 2D-SEIRA spectroscopy to highlight conformational changes of the cytochrome c oxidase induced by direct electron transfer. Metallomics, 3, 619–627. Reproduced with permission of The Royal Society of Chemistry.)

377

Tethered Lipid Membranes (a) 7.0 × 10–3

(b) 7.0 × 10–3

1656

4.0 × 10–3 3.0 × 10–3

1681

1639 1620

1574 1558

Absorbance

5.0 × 10–3

5.0 × 10–3 4.0 × 10–3 3.0 × 10–3 2.0 × 10–3

10–3

1.0 × 10–3

0.0

0.0

1560

1560

1590

1590

Wavenumber (cm–1)

1620

1620

1650

Wavenumber

2.0 × 10–3 1.0 ×

1656

6.0 × 10–3

1607

(cm–1)

Absorbance

6.0 × 10–3

1650

1680 1710

1607 1681

1639 1620

1574 1558

1680 1710 1740

1740 1740 1710 1680 1650 1620 1590 1560 Wavenumber (cm–1)

1740 1710 1680 1650 1620 1590 1560 Wavenumber (cm–1)

FIGURE 18.13 Asynchronous (left) and asynchronous (right) 2D correlation maps of nonactivated CcO. Potential applied was 900 mV vs. the SHE for the fully oxidized state and varied from 500 to −700 mV in 100 mV steps for reduced states. No mediators were added. The upper panel shows one of the 1D-SEIRA difference spectra (taken at −700 mV) as an example for the deconvolution. (Nowak, C. et al. 2011b. 2D-SEIRA spectroscopy to highlight conformational changes of the cytochrome c oxidase induced by direct electron transfer. Metallomics, 3, 619–627. Reproduced with permission of The Royal Society of Chemistry.)

titrations of the CcO followed simultaneously by UV/VIS and FTIR spectroscopy (Hellwig et al. 1996, 1998, 1999, Iwaki et al. 2002, 2006) analogously to Nowak et al. (2010). However, due to the better resolution, a higher number of bands could be resolved. The amide I region shows the characteristic band pattern of secondary structures, which can also be found in proteins without any redox function, particularly using 2D IR (Zheng and Gunner 2009). For example, the band at 1654–1656 cm−1 indicates the α-helical structure whereas the assembly of bands at 1618–1623, 1638–1641, and 1681–1684 cm−1 is a characteristic feature of increased β-sheets. The same bands were associated before (Hellwig et al. 1996, 1998, 1999, Iwaki et al. 2006) with redox transitions of heme a3 and CuA of CcO. Sigmoid functions were constructed from 1D spectra for the bands at 1623 and 1655 cm−1, respectively (Nowak et al. 2010). 2D spectra should not be used for the same purpose, because quantitative analysis should not be performed on deconvoluted spectra (Smith et al. 2001). The band at 1600–1607 cm−1 has usually not been associated with the β-sheet structure. However, on the basis of potentiometric titrations of CcO, this band was attributed to the redox transition of CuA, located in the middle of the β-sheet structure, within the aqueous domain of the protein. Moreover, sigmoidal functions were constructed from this band under conditions of direct and mediated ET (Nowak et al. 2010). Importantly, the activated state is characterized by a higher number of strongly correlated conformational transitions. This suggests that the enzyme under turnover conditions undergoes a more global conformational transition than hitherto expected (Nowak et  al. 2011a). This is consistent with the considerable conformational

378

Liposomes, Lipid Bilayers and Model Membranes

change in response to reduction shown by x-ray crystallography, which was proposed to be of functional signiicance (Nowak et  al. 2011, Qin et  al. 2009). In this context, it should be mentioned that the activated state (AS) of CcO is also generated under reducing conditions. The result of these conformational transitions seems to be a more condensed form of CcO, thus optimizing tunnel pathways for electrons and protons. The assumption of a more condensed form of the AS is consistent with several observations in 2D SEIRAS, such as the narrower bandwidth and higher amplitude of bands characteristic for α-helices vs. those of β-sheets. But above all, the global conformational change is consistent with the shift in the midpoint potentials of the CcO in the AS revealed by both cyclic voltammetry and potentiometric titration (Nowak et al. 2011a). Midpoint potentials were shown to depend strongly on the environment of the heme structure in the case of P450, which is characterized by an enzyme cycle very similar to that of CcO (Udit and Gray 2005, Udit et al. 2005). The large number of correlations in 2D IR spectra also suggests a high degree of cooperativity between single transitions, particularly in the AS. Correlations were found in the case of H bridges, hydrophobic interactions, dipole interactions, which are well known to exhibit cooperative behavior. Nevertheless, it should also be highlighted that further measurements on the complex CcO–ptBLM system are needed to corroborate the indings and interpretations on CcO made by IR- and Raman’s measurements. However, this example shows clearly the applicability of biomimetic membrane systems such as the ptBLM system to investigate complex problems by surface-analytical tools such as IR-spectroscopy or other techniques such as impedance, SPR, or Raman’s spectroscopy to name some of the available tools.

REFERENCES Antonini, G. et al. 1985. Pulsed cytochrome-C oxidase. Journal of Inorganic Biochemistry, 23(3–4), 289–293. Armstrong, F.A. 2002a. Insights from protein ilm voltammetry into mechanisms of complex biological electron-transfer reactions. Journal of the Chemical Society-Dalton Transactions, (5), 661–671. Armstrong, F.A. 2002b. Protein ilm voltammetry: Revealing the mechanisms of biological oxidation and reduction. Russian Journal of Electrochemistry, 38(1), 49–62. Ataka, K. et al. 2004. Oriented attachment and membrane reconstitution of his-tagged cytochrome c oxidase to a gold electrode: In situ monitoring by surface-enhanced infrared absorption spectroscopy. Journal of the American Chemical Society, 126(49), 16199–16206. Ataka, K. and J. Heberle. 2004. Functional vibrational spectroscopy of a cytochrome c monolayer: SEIDAS probes the interaction with different surface-modiied electrodes. Journal of the American Chemical Society, 126(30), 9445–9457. Ataka, K. and J. Heberle. 2006. Use of surface enhanced infrared absorption spectroscopy (SEIRA) to probe the functionality of a protein monolayer. Biopolymers, 82(4), 415–419. Atanasov, V. et al. 2005. Membrane on a chip: A functional tethered lipid bilayer membrane on silicon oxide surfaces. Biophysical Journal, 89(3), 1780–1788. Barth, A. 2000. The infrared absorption of amino acids side chains. Progress in Biophysics and Molecular Biology, 74, 141–173. Baumgart, T. et al. 2003. Fusion of small unilamellar vesicles onto laterally mixed self-assembled monolayers of thiolipopeptides. Journal of Colloid and Interface Science, 258(2), 298–309. Baurecht, D. et al. 2002. A new method of phase sensitive detection in modulation spectroscopy applied to temperature induced folding and unfolding of RNase A. Vibrational Spectroscopy, 30(1), 85–92. Becucci, L. et al. 2005. Potassium ion transport by valinomycin across a Hg-supported lipid bilayer. Journal of the American Chemical Society, 127(38), 13316–13323. Braiman, M.S. and K.J. Rothschild. 1988. Fourier transform infrared techniques for probing membrane protein structure. Annual Reviews of Biophysics and Biophysical Chemistry, 17, 541–570. Brunori, M. et  al. 1985. Stopped-low studies of cytochrome-oxidase reconstituted into liposomes—Proton pumping and control of activity. Journal of Inorganic Biochemistry, 23(3–4), 373–379. Czarnecki, M.A. et al. 1998. 2D FT-NIR and FT-IR correlation analysis of temperature-induced changes of nylon 12. Chemical Physics Letters, 283(5–6), 326–332. Enders, D. et al. 2006. Reversible adsorption of Au nanoparticles on SiO2/Si: An in situ ATR-IR study. Surface Science, 600(6), L71–L75.

Tethered Lipid Membranes

379

Friedrich, M.G. et al. 2004. Active site structure and redox processes of cytochrome c oxidase immobilised in a novel biomimetic lipid membrane on an electrode. Chemical Communications, 21, 2376–2377. Friedrich, M.G. et al. 2008. In situ monitoring of the catalytic activity of cytochrome c oxidase in a biomimetic architecture. Biophysical Journal, 95(3), 1500–1510. Fringeli, U.P. 1992. In: Internal Relection Spectroscopy: Theory and Applications. Mirabella, F.M. Ed., Marcel Dekker Inc., New York, pp. 255–324. Ganim, Z. et  al. 2008. Amide I two-dimensional infrared spectroscopy of proteins. Accounts of Chemical Research, 41(3), 432–441. Giess, F. et al. 2004. The protein-tethered lipid bilayer: A novel mimic of the biological membrane. Biophysical Journal, 87(5), 3213–3220. Goormaghtigh, E. et al. 1999. Attenuated total relection infrared spectroscopy of proteins and lipids in biological membranes. Biochimica Et Biophysica Acta—Reviews on Biomembranes, 1422(2), 105–185. Gorbikova, E.A. et al. 2006. Redox titration of all electron carriers of cytochrome c oxidase by Fourier transform infrared spectroscopy. Biochemistry, 45(17), 5641–5649. Haas, A.S. et al. 2001. Cytochrome c and cytochrome c oxidase: Monolayer assemblies and catalysis. Journal of Physical Chemistry B, 105(45), 11351–11362. He, L.H. et al. 2005. Tethered bilayer lipid membranes based on monolayers of thiolipids mixed with a complementary dilution molecule. 1. Incorporation of channel peptides. Langmuir, 21(25), 11666–11672. Heibel, G.E. et al. 1993. Comparative resonance Raman-study of cytochrome-c-oxidase from beef-heart and Paracoccus denitriicans. Biochemistry, 32(40), 10866–10877. Hellwig, P. et al. 1996. Carboxyl group protonation upon reduction of the Paracoccus denitriicans cytochrome c oxidase: Direct evidence by FTIR spectroscopy. Febs Letters, 385(1–2), 53–57. Hellwig, P. et al. 1998. Electrochemically induced FT-IR difference spectra of the two- and four-subunit cytochrome c oxidase from P-denitriicans reveal identical conformational changes upon redox transitions. Biochimica Et Biophysica Acta—Bioenergetics, 1409(2), 107–112. Hellwig, P. et al. 1999. Electrochemical and ultraviolet/visible/infrared spectroscopic analysis of heme a and a(3) redox reactions in the cytochrome c oxidase from Paracoccus denitriicans: Separation of heme a and a(3) contributions and assignment of vibrational modes. Biochemistry, 38(6), 1685–1694. Hill, T.L., 1977. Free Energy Transduction in Biology. Academic Press, New York. Huo, S.J. et al. 2005. Tunable surface-enhanced infrared absorption on Au nanoilms on Si fabricated by selfassembly and growth of colloidal particles. Journal of Physical Chemistry B, 109(33), 15985–15991. Iwaki, M. et  al. 2002. ATR-FTIR difference spectroscopy of the P-M intermediate of bovine cytochrome c oxidase. Biochimica Et Biophysica Acta—Bioenergetics, 1555(1–3), 116–121. Iwaki, M. et al. 2006. ATR-FTIR characterisation of the P-M intermediate of Paracoccus denitriicans cytochrome c oxidase. BBA—Bioenergetics, 180–181. Jancura, D. et  al. 2006. Spectral and kinetic equivalence of oxidized cytochrome c oxidase as isolated and “activated” by reoxidation. Journal of Biological Chemistry, 281(41), 30319–30325. Jeuken, L.J.C. et  al. 2002. Electron-transfer mechanisms through biological redox chains in multicenter enzymes. Journal of the American Chemical Society, 124(20), 5702–5713. Jung, Y.M. et al. 2000. Two-dimensional infrared, two-dimensional Raman, and two-dimensional infrared and Raman heterospectral correlation studies of secondary structure of beta-lactoglobulin in buffer solutions. Journal of Physical Chemistry B, 104(32), 7812–7817. Knoll, W. et al., 2004. Functional tethered bilayer lipid membranes. In: Ultrathin Electrochemical Chemo- and Biosensors, Wolfbeis, O. S., Mirsky, V. M. Eds., Springer, Berlin, pp. 239–253. Leitch, J. et al. 2009. In situ PM-IRRAS studies of an archaea analogue thiolipid assembled on a Au(111) electrode surface. Langmuir, 25(17), 10354–10363. Muller, M. et al. 1996. 2D-FTIR ATR spectroscopy of thermo-induced periodic secondary structural changes of poly-(L)-lysine: A cross-correlation analysis of phase-resolved temperature modulation spectra. Journal of Physical Chemistry, 100(25), 10810–10825. Naumann, R. et  al. 1995. Incorporation of membrane-proteins in solid-supported lipid layers. Angewandte Chemie—International Edition, 34(18), 2056–2058. Naumann, R. et al. 1997. Coupling of proton translocation through ATPase incorporated into supported lipid bilayers to an electrochemical process. Bioelectrochemistry and Bioenergetics, 42(2), 241–247. Naumann, R. et al. 1999. The peptide-tethered lipid membrane as a biomimetic system to incorporate cytochrome c oxidase in a functionally active form. Biosensors & Bioelectronics, 14(7), 651–662. Naumann, R. et al. 2002. Proton transport through a peptide-tethered bilayer lipid membrane by the H+-ATP synthase from chloroplasts measured by impedance spectroscopy. Biosensors & Bioelectronics, 17(1–2), 25–34.

380

Liposomes, Lipid Bilayers and Model Membranes

Naumann, R. et al. 2003a. Tethered lipid bilayers on ultralat gold surfaces. Langmuir, 19(13), 5435–5443. Naumann, R. et al. 2003b. Kinetics of valinomycin-mediated K+ ion transport through tethered bilayer lipid membranes. Journal of Electroanalytical Chemistry, 550, 241–252. Nishikawa, Y. et al. 1993. Silver island ilms for surface-enhanced infrared-absorption spectroscopy—Effect of island morphology on the absorption enhancement. Vibrational Spectroscopy, 6(1), 43–53. Noda, I. 1990. 2-Dimensional infrared (2d IR) spectroscopy—Theory and applications. Applied Spectroscopy, 44(4), 550–561. Noda, I. and Y. Ozaki. 2004. Two-Dimensional Correlation Spectroscopy, John Wiley & Sons Ltd., Chichester, 2004. Noda, H. et al. 1999. Time-resolved surface-enhanced infra-red study of molecular adsorption at the electrochemical interface. Surface Science, 427–28, 190–194. Nowak, C. et al. 2009a. A two-layer gold surface with improved surface enhancement for spectro-electrochemistry using surface-enhanced infrared absorption spectroscopy. Applied Spectroscopy, 63(9), 1068–74. Nowak, C. et al. 2009b. Electron transfer kinetics of cytochrome C in the submillisecond time regime using time-resolved surface-enhanced infrared absorption spectroscopy. Journal of Physical Chemistry C, 113(6), 2256–2262. Nowak, C. et al. 2010. Conformational transitions and molecular hysteresis of cytochrome c oxidase: Varying the redox state by electronic wiring. Soft Matter, 6(21), 5523–5532. Nowak, C. et al. 2011a. Oriented immobilization and electron transfer to the cytochrome c oxidase. Journal of Solid State Electrochemistry, 15,105–114. Nowak, C. et  al. 2011b. 2D-SEIRA spectroscopy to highlight conformational changes of the cytochrome c oxidase induced by direct electron transfer. Metallomics, 3, 619–627. Osawa, M., 2001. Surface-enhanced infrared absorption spectroscopy. In: Near Field Optics and Surface Plasmon Polaritons, Kawata, S., Ed., Springer, Berlin, pp. 163–184. Osawa, M. et al. 1993. Surface-enhanced infrared-spectroscopy—The origin of the absorption enhancement and band selection rule in the infrared-spectra of molecules adsorbed on ine metal particles. Applied Spectroscopy, 47(9), 1497–1502. Otto, A. et  al. 1992. Surface-enhanced raman-scattering. Journal of Physics-Condensed Matter, 4(5), 1143–1212. Ozaki, Y. et al. 2003. Two-dimensional infrared correlation spectroscopy studies on secondary structures and hydrogen bondings of side chains of proteins. Spectroscopy—an International Journal, 17(2–3), 79–100. Popot, J.L. and D.M. Engelman. 2000. Helical membrane protein folding, stability, and evolution. Annual Review of Biochemistry, 69, 881–922. Pucci, A. 2005. IR spectroscopy of adsorbates on ultrathin metal ilms. Physica Status Solidi B-Basic Solid State Physics, 242(13), 2704–2713. Qin, L. et al. 2009. Redox-dependent conformational changes in cytochrome c oxidase suggest a gating mechanism for proton uptake. Biochemistry, 48(23), 5121–5130. Raguse, B. et  al. 1998. Tethered lipid bilayer membranes: Formation and ionic reservoir characterization. Langmuir, 14(3), 648–659. Robertson, J.W.F. et al. 2008. Modeling ion transport in tethered bilayer lipid membranes. 1. Passive ion permeation. Journal of Physical Chemistry B, 112(34), 10475–10482. Schach, D. et al. 2010. Modeling direct electron transfer to a multi-redox center protein: Cytochrome c oxidase. Journal of Electroanalytical Chemistry, 649, 268–276. Schiller, S.M. et al. 2003. Archaea analogue thiolipids for tethered bilayer lipid membranes on ultrasmooth gold surfaces. Angewandte Chemie-International Edition, 42(2), 208–211. Schmidt, E.K. et  al. 1998. Incorporation of the acetylcholine receptor dimer from Torpedo californica in a peptide supported lipid membrane investigated by surface plasmon and luorescence spectroscopy. Biosensors & Bioelectronics, 13(6), 585–591. Schonherr, H. et  al. 2004. Vesicle adsorption and lipid bilayer formation on glass studied by atomic force microscopy. Langmuir, 20(26), 11600–11606. Schultz, Z.D. and I.W. Levin. 2011. Vibrational spectroscopy of biomembranes. Annual Review of Analytical Chemistry, 4, 343–366. Smith, W.L. et  al. 2001. Membrane protein structural biology minireview series. Journal of Biological Chemistry, 276(35), 32393–32394. Susi, H. 1972. Infrared spectroscopy-conformation. In: Methods in Enzymology, Hirs, C.H.W. and Timasheff, S.N. Eds., Academic Press, New York, vol. 26, pp. 455–472. Svensson-Ek, M. et al. 2002. The X-ray crystal structures of wild-type and EQ(I-286) mutant cytochrome c oxidases from Rhodobacter sphaeroides. Journal of Molecular Biology, 321(2), 329–339.

Tethered Lipid Membranes

381

Tamm, L.K. and S.A. Tatulian. 1997. Infrared spectroscopy of proteins and peptides in lipid bilayers. Quarterly Reviews of Biophysics, 30(4), 365–429. Tanaka, M. and E. Sackmann. 2005. Polymer-supported membranes as models of the cell surface. Nature, 437(7059), 656–663. Tatulian, S.A. 2003. Attenuated total relection Fourier transform infrared spectroscopy: A method of choice for studying membrane proteins and lipids. Biochemistry, 42(41), 11898–11907. Terrettaz, S. et al. 2003. Highly electrically insulating tethered lipid bilayers for probing the function of ion channel proteins. Langmuir, 19(14), 5567–5569. Udit, A.K. and H.B. Gray. 2005. Electrochemistry of heme-thiolate proteins. Biochemical and Biophysical Research Communications, 338(1), 470–476. Udit, A.K. et al. 2005. Protein-surfactant ilm voltammetry of wild-type and mutant cytochrome P450BM3. Inorganic Chemistry, 44(12), 4109–4111. White, S.H. et al. 2001. How membranes shape protein structure. Journal of Biological Chemistry, 276(35), 32395–32398. White, S.H. and W.C. Wimley. 1999. Membrane protein folding and stability: Physical principles. Annual Review of Biophysics and Biomolecular Structure, 28, 319–365. Wi, S. et al. 1998. Predictions of protein secondary structures using factor analysis on Fourier transform infrared spectra: Effect of Fourier self-deconvolution of the amide I and amide II bands. Biospectroscopy, 4, 93–106. Zheng, Z. and A.R. Gunner. 2009. Analysis of the electrochemistry of hemes with E(m)s spanning 800 mV. Proteins-Structure Function and Bioinformatics, 75(3), 719–734.

19

Ion-Transporting Supported and Tethered Lipid Bilayers That Incorporate Biological Membrane Transport Proteins Donald K. Martin, Bruce A. Cornell, Lavinia Liguori, Jean-Luc Lenormand, Jean-Pierre Alcaraz, Gwenaël Scolan, and Philippe Cinquin

CONTENTS 19.1 Introduction .......................................................................................................................... 383 19.2 Lipid Membranes for Protecting Implanted Devices ........................................................... 384 19.3 Structural Elements for Molecularly Engineered Biomimetic Devices ............................... 384 19.3.1 Tethered and Supported Lipid Bilayer Membranes .................................................. 385 19.3.2 Molecules That Provide Transport Capability.......................................................... 387 19.4 Tethered Lipid Bilayer Systems That Incorporate Biological Membrane Transport Proteins ................................................................................................................................. 387 19.5 Supported Lipid Bilayer Systems That Incorporate Biological Transport Proteins for Creating Energy .............................................................................................................. 394 References ...................................................................................................................................... 398

19.1

INTRODUCTION

The fabrication of model lipid membranes possessing some degree of predictability was irst reported in the pioneering work of K.R. Blodgett and I. Langmuir who formed thin monomolecular ilms supported on solid or liquid surfaces (Blodgett 1935, 1939, Blodgett and Langmuir 1937). Prior to that pioneering work, the thickness of suspended bimolecular liquid ilms of soap in air had been investigated and measured (Johonnott 1899, 1906). The extension of that work to produce bimolecular liquid ilms in air led to an understanding of the formation of lipid ilms in water (Folch and Lees 1951). On the basis of these early reports, a major milestone in suspended lipid bilayer research was achieved by the reconstitution of an excitable cell membrane structure in a saline solution, which was in effect the irst report of the production of the black lipid membrane (Mueller et al. 1962a,b). These early reports paved the way for the irst report of liposomes and functional studies of the ionic transport across swollen phospholipids at the Babraham Laboratories in Cambridge (Bangham et al. 1965). These major achievements by pioneering scientists, including E.S. Johonnott, J. Folch, P. Mueller, A.D. Bangham, K.R. Blodgett, and I. Langmuir, created a platform for the development of the understanding of transport processes in lipid bilayer membranes (BLMs). Indeed, over the past two decades, there has been an explosion in the use of model membrane systems as a tool to understand biological systems better. There are numerous reviews of such model membrane systems (see, e.g., Ti Tien and Diana 1968, Ti Tien 1988, Ottova-Leitmannova and Ti Tien 1992, Ottova et al. 1997, 383

384

Liposomes, Lipid Bilayers and Model Membranes

Feller 2000, Castellana and Cremer 2006, Martin 2007, Koeper 2007, Mouritsen 2011). Recent developments, however, have seen the use of model biomimetic membranes as more than an aid to understand biological function better and have now advanced to become a core engineered element in a new generation of biomimetic molecular devices. The challenge for any technology based on synthetic lipid membranes is device stability. Although challenging to engineer, good stability is usually achieved using lipid bilayers that are tethered to or supported by a substrate that may be nonporous (e.g., gold, silicon) or porous (e.g., polymer, gel), rather than using suspended lipid bilayers. In this chapter, we focus on devices incorporating membranes whose function stems from biological transport proteins fabricated into lipid bilayers that are either tethered to a gold substrate or supported on a hydrogel. We describe the technical operation of representative devices and provide examples of typical target applications. The irst reported example of such an ion channel biosensor was a platform diagnostic sensing technology based on a tethered lipid bilayer membrane (tLBM) (Cornell et al. 1997). Many attributes such as the sensor speciicity and sensitivity in that device are drawn from biomimicry and will be discussed in the following sections. Synthetic biomimetic membranes have found many additional applications including drug (e.g., liposomes, nanosomes, vesosomes, virosomes, polymerosomes, and capsosomes) and gene delivery systems, and gene (e.g., lipoplexes and polylipoplexes) delivery systems. A complete description of each of these is beyond the scope of this chapter, and the reader is directed to many excellent reports in the abundant literature that describe these applications (see, e.g., Mangipudi et al. 2009, Shi et al. 2011). However, we start by describing a further important area of application, which is using lipid membranes to create a biocompatible, cell-like surface for implantable devices.

19.2

LIPID MEMBRANES FOR PROTECTING IMPLANTED DEVICES

Lipid membranes can be used for the protection of implants against rejection by the host. When a foreign body is detected, an immune response occurs with inlammation and possible allergic reaction. The use of lipid membranes to coat implants to improve their biocompatibility has been well described across many applications from protecting arterial implants from coagulation to heart valves and cardiac assist devices to dental prosthesis (see, e.g., Martin 2007, Zhan et  al. 2010, Jagoda et al. 2011, Liu et al. 2011). Implanted microelectrode arrays allow for the real-time monitoring of neural activity. The use of a lipid membrane coating has shown encouraging results for the prevention of the formation of a scar tissue around the implant that would limit its functionality (Hayward and Chapman 1984). The properties of stents can be beneicially modulated by a coating of polymer possessing lipid membrane-like phosphoryl choline groups. A lipid membrane coating also improves the hemocompatibility of stents and reduces the thrombogenicity (Gotman 1997, Ishihara and Takai 2009). It has been reported that the lipid-based coating of a metallic implant is an eficient strategy to reduce the risk of bacterial growth and thus the risk of infection (He and Bellamkonda 2005). Lipid membrane coatings slow the rate of corrosion of metallic implants and can even limit the formation of ibrous tissue around implants (Bollo 2007). Lipid membranes have been used to encapsulate Islets of Langerhans resulting in the implantation of a replacement pancreas with no need of an immunosuppressive therapy (Hayward and Chapman 1984) and minimal or no side effects.

19.3 STRUCTURAL ELEMENTS FOR MOLECULARLY ENGINEERED BIOMIMETIC DEVICES The principal building blocks for molecularly engineered biomimetic devices are the lipids and protein molecules that facilitate the transport of ions across the membrane in biology. Both natural protein extracts and synthetic recombinant proteins are widely employed to make devices based on tethered or supported membranes. Electrical measurements of these devices draw heavily on classical patch-clamp electrophysiology that remains the “gold standard” method to monitor ion channel

Ion-Transporting Supported and Tethered Lipid Bilayers

385

proteins, but it requires a substantial investment in terms of time (training) and money (equipment). Considerable room exists for alternative techniques that are more convenient such as artiicial lipid bilayer platforms that allow an impedimetric measurement monitoring of the transport functions of the proteins, for example, in pharmaceutical drug development and in the engineering of ion channel-based biosensors. It is timely to consider the engineering of robust systems that include transporting lipid BLMs with very long lifetimes. This is the biotechnological advancement over reconstituting ion channels in simple free-standing black lipid membranes, which has previously been a convenient method to study their activity and to envision protein–ligand screening, drug discovery, and biosensor design (Ariga et al. 2006).

19.3.1

tethered and supported lIpId bIlayer MeMbranes

In a general manner, a supported lipid bilayer membrane (sLBM) can be deined as a single and continuous phospholipid bilayer on a solid or polymeric substrate. The substrate includes many types of materials, including silica, glass, mica, polymer, or gold (Groves and Boxer 2002, Richter et al. 2006). Gold in particular has been shown as a good support for the formation of well-oriented self-assembly layers. Since the substrate will usually constitute the working electrode in an sLBMbased biosensor, the use of gold has a distinct advantage for electrochemical measurements of conductance of the sLBM. The biomimetic membrane can be linked to its support by (i) direct deposit (the lipid bilayer is adsorbed directly to the support, which could potentially inactivate a membrane protein incorporated in the membrane through contact with the solid support), (ii) cushion support (a polymer forms a cushion between lipid bilayer and solid support, thereby reducing the risk of inactivating incorporated proteins by interaction with the solid support), or (iii) tethering the membranes to the solid support (provides a reservoir between the lipid membrane and the solid support). The tLBM method comprises covalently tethering the lipid membrane to a gold electrode through the covalent attachment of hydrophobic spacer molecules with a hydrophilic linker. The short hydrophilic spacers, which usually contain sulfur groups, are reactive with gold. The same molecule also contains a sequence of polar ethylene glycol groups that provide an aqueous layer at the support surface and a hydrocarbon group that behaves as an ionic barrier and mimics the property of the lipid bilayer. Indeed, as the molecules bind, they pack together and progressively form a two-dimensional lipid bilayer ilm. This bilayer is stabler than an unsupported black lipid bilayer. The reservoir space between the membrane and support is increased in the tLBM due to the presence of the spacers (Cornell et al. 1997). As a result, membrane proteins are more easily integrated and, importantly, retain functionality because the risk of undesired interaction with the support is substantially reduced. However, the reservoir on the inner side of the membrane is still limited in comparison with porous supports, which can restrict the monitoring of ion transport (Reimhult and Kumar 2007). Although gold substrates are the most widely used for the design of tLBM, silicon substrates can also be used (Zagnoni et al. 2007). However, membranes assembled on gold supports have been shown to bring key advantages such as ease of fabrication, reproducibility, availability of materials, chemical stability, and lexibility (Sinner and Knoll 2001, Pohorille and Deamer 2002). In most cases, both the tethering surface and the return electrode are pure gold, which, at the applied measurement voltages of 20–50 mV, has no electrochemistry, and provides a capacitive coupling between the electrons in gold and the ions in solution (Elender et al. 1996). In these coupling capacitors, one plate is formed by the gold charged with electrons and the other plate is formed by ions in solution crowding at the gold surface. This approach overcomes many problems associated with Ag/AgCl electrodes. In particular, proteins can be damaged by silver ions and it is a big advantage when using capacitive coupling since there are no metal ions in solution. Also, the approach avoids signal drift due to variations in electrochemical potential that can depend on both temperature and trace amounts of ionic impurities.

386

Liposomes, Lipid Bilayers and Model Membranes

The tLBM is prepared in two stages. The irst stage is to coat the gold electrode with tethered lipid and tethered spacer molecules (Figure 19.1). The tethered lipid molecules that penetrate the bilayer to form the inner layer lealet of the tLBM are phytanyl bis-tetra-ethyleneglycol, which are chemically bound to the gold substrate using benzyl disulide to provide greater stability against oxidation than would be provided by a simple thiol. Benzyl disulide also provides optimal spacing of the hydroxyterminated-bis-tetra-ethyleneglycol to eliminate conduction limitation of ionic species within the reservoir between the tLBM and the gold substrate. The hydroxyterminated-bistetra-ethyleneglycol, which does not penetrate the bilayer to form any of the lealets, is terminated in an OH group at the surface of the inner lealet. These respectively anchor the lipid bilayer to the electrode surface and provide space to permit the mobile lipid to adjust to the requirements of the bilayer and to allow conformational relaxation of any incorporated protein (Cornell et al. 2001). Spacer groups act as a cushion maintaining the bilayer off the gold surface, which compensates for the surface roughness effects. They also constitute an ion reservoir on the internal side of the lipid membrane. The ratio of the tethering fraction to the mobile fraction may be adjusted to permit the incorporation of proteins of varying size. Usually, the larger the molecular weight of the protein, the lower the fraction of tethers that are used to allow space within the membrane to accommodate large proteins. Apart from the density or fraction of tethering, the chemical composition and length of the spacer molecules are also modular. The right balance has to be determined for each protein. The tLBM then forms an electrically insulating platform allowing the individual monitoring of a protein’s function and activity (Frederix et al. 2009). The membranes thus synthesized have the ability to resist factors that would disrupt liposomes or hydrolyze the acyl chain attachments in another context (Moradi-Monfared et al. 2012). Only a fraction of the membrane lipids is tethered that stabilizes the whole structure and provides it a resistance to the insertion of random material. This increased stability is another property speciic to tLBM. It results in a longer lifetime of several months against several hours for the classic free-spanning black lipid membranes. Tethered biomembranes provide a very unique environment in respect of the minimum requirements for the preservation of membrane proteins in an adequate orientation. Functionalized tLBM can be used for the design of intelligent molecular devices. They can be applied to almost any sensing applications for blood typing and detection of bacteria, viruses, and other biomarkers (e.g., hormones, enzymes, or antibodies), and detection (Sleytr et al. 2007). Membrane proteins embedded in the tLBM can be envisioned as actual sensing units (Cornell et al. 1999, Yin et al. 2003).

+

++

+

Mobile

Tethered lipid 4 nm Ionic insulator

Tethered spacers

4 nm

Ionic conductor +

+

FIGURE 19.1 (See color insert.) Schematic diagram of a tLBM. A 4-nm-thick insulating layer is tethered to the gold surface via a 4-nm-thick polar, “reservoir layer” between the membrane and the gold surface. A fraction of the tethered lipid spans half the 4-nm thickness of the bilayer and acts as anchors. A component (tethered spacer) is tethered but does not penetrate the bilayer and acts as a spacer for the tethers and a cushion on which the bilayer is supported.

Ion-Transporting Supported and Tethered Lipid Bilayers

387

19.3.2 Molecules that provIde transport capabIlIty Membrane proteins represent approximately 30% of the total proteins from an organism and are involved in essential biochemical and physiological processes. Some examples of such proteins include all the transmembrane receptors, integrins, membrane transporters, and ion channel proteins. Limitations in the production of folded membrane proteins represent the major bottleneck for functional and structural studies of this huge category of macromolecules because classical protein overexpression systems based on living cells lead to membrane-protein precipitation, cell toxicity, and low production yield. The development of cell-free protein synthesis systems provides a method of performing biological production without cellular proliferation. They are based on gene expression in the presence of cell extracts, mostly from cells containing highly active transcription/ translation machinery (Liguori et  al. 2007). The optimization of a bacterial cell-free expression system allows to produce folded and functional membrane proteins either embedded in liposomes (lipid vesicles) or under soluble form (in the presence of detergents). This new biotechnology represents the ideal platform for an extensive investigation on membrane proteins and the possibility to use them to approach challenging research topics such as structural studies, engineered therapeutic proteoliposomes and design of biomimetic devices (Liguori et al. 2008, Liguori and Lenormand 2009, Deniaud et al. 2010). However, for the purpose of this chapter, we will introduce ion channel proteins, which we utilize in the devices described later in this chapter. Ion channels are integral parts of many physiological processes and are necessary components of lipid BLMs to provide ion transport capabilities in biological cells. Their puriication and subsequent incorporation within proteoliposomes* that are readily available for incorporation into sLBM or tLBM provides a way to study them to understand their physiological functions and their therapeutic potential better. Since ion channels comprise the main targets for drugs, it is crucial to fully understand their mechanisms and functions to improve the use of medications in the body. Moreover, the high afinity of these channels for particular speciic ligands makes them very attractive for the manufacture of ion channel-based biosensing systems (Lundquist et al. 2010).

19.4 TETHERED LIPID BILAYER SYSTEMS THAT INCORPORATE BIOLOGICAL MEMBRANE TRANSPORT PROTEINS Engineering a biomimetic device requires a lipid bilayer that is robust and stable. Although suspended lipid BLMs have been extensively used as a system to study incorporated membrane proteins, the BLM is unstable and degrades over a time course that can be from minutes to hours. A stabler device is obtained by supporting the lipid bilayer on a substrate, or indeed by tethering the lipid bilayer to the substrate using speciic molecules. For example, a supported lipid bilayer was utilized in an acoustic wave biosensor to provide a direct immunosensing capability (Gizeli et al. 1997). That device was based on an acoustic waveguide geometry that supported a Love wave, with the biorecognition surface formed on a gold layer and consisting of a biotinylated-supported lipid layer that speciically bound streptavidin and, subsequently, biotinylated goat IgG. The modiied surface was used as a model immunosensor and successfully detected rabbit anti-goat IgG in the concentration range 10−8–10−6 M. The irst practical biosensor based on such principles of a tethered lipid bilayer was developed by Cornell et al. (1997) and is now being commercialized by SDX Tethered Membranes Pty Ltd.† In that original tethered membrane biosensor of Cornell et al. (1997), the conductance of a population of gramicidin ion channels incorporated into the tethered lipid bilayer is switched by the binding of a molecule recognized by a receptor attached to the gramicidin ion channels in the outer lealet of the tethered bilayer. The approach mimics biological sensory functions and can be used with most * †

www.synthelis.fr. www.sdxtetheredmembranes.com.

388

Liposomes, Lipid Bilayers and Model Membranes

types of receptor, including antibodies and nucleotides. The technique is very lexible and even in its simplest form, it is sensitive to picomolar concentrations of proteins. The sensing electronic system relies on the change in membrane conductance measured by impedance spectroscopy. These uses can include cell typing, the detection of large proteins, viruses, antibodies, deoxyribonucleic acid (DNA), electrolytes, drugs, pesticides, and other low-molecular-weight compounds. That sensor has undergone continuing development (Woodhouse et al. 1999), with either gramicidin or alamethicin channels incorporated into the tethered BLMs (Yin et al. 2003). The tethered lipid bilayer is also useful as a platform to investigate the properties of membrane proteins that have a transport function. The advantage of the tethered lipid bilayer is that it is stable, it is suitable for incorporating a range of membrane transport proteins, including ion channels, and the conductance of the incorporated membrane transport protein is easily recorded using impedance spectroscopy. The formation of a tLBM occurs in four steps as shown in Figure 19.2. The inner lealet of the lipid bilayer membrane and the tethering molecules are bound to the gold substrate. The tethered lipid molecules that penetrate the bilayer to form the inner layer lealet of the tLBM are phytanyl bis-tetra-ethyleneglycol, which are chemically bound to the gold substrate using benzyl disulide to provide greater stability against oxidation than would be provided by a simple thiol. Benzyl disulide also provides optimal spacing of the hydroxyterminated-bis-tetra-ethyleneglycol to eliminate conduction limitation of ionic species within the reservoir between the tLBM and the gold substrate. The hydroxyterminated-bis-tetra-ethyleneglycol, which does not penetrate the bilayer to form any of the lealets, is terminated in an OH group at the surface of the inner lealet. The mobile lipid species are usually diphytanyletherphosphatidylcholine (DPEPC) and glycerodiphytanylether (GDPE) in the mole ratio 70:30, respectively. That mole ratio achieves an average area per molecule at the hydrocarbon–aqueous interface that is commensurate with the average area per molecule at the center of the bilayer. That ratio can be adjusted for speciic demands for particular incorporated proteins within the membrane, and is best determined empirically for any particular incorporated protein. The mobile species can also be any mixture of lipids such as 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and 1,2-dilinoleoyl-sn-glycero-3-phosphoethanolamine (DOPE), such as provided by the unfolding of proteolipsomes used to incorporate membrane proteins in the tLBM. Changing the mole ratio of the penetrating tethered lipid species (phytanyl bis-tetra-ethyleneglycol) compared to the nonpenetrating tethered lipid species (hydroxyterminated-bis-tetra-ethyleneglycol) provides a means to optimize the incorporation of proteins into the mobile outer lealet, and thence the complete tLBM. This mole ratio can be expressed as the percentage fraction x = penetrating/ nonpenetrating. The basis of this optimization procedure for protein incorporation is a simple volume calculation based on the molecular weight of the membrane protein to be incorporated and an assumed thickness of 4 nm for the tLBM. Nevertheless, all mole ratios, x, ranging from 0 < x < 100 will form bilayers but with a trade-off between the stability and ionic leakage of the tLBM and the volume of Add mobile lipids in ethanol

Step 1

Step 2

Step 3

Rinse in PBS to form t-BLM

Step 4

FIGURE 19.2 Formation of a tLBM. Step 1: Gold is immersed in ethanol–lipid solution. Step 2: The lipids react with the gold and then tether to the gold surface. Step 3: Addition of nontethered lipids dissolved in ethanol. Step 4: Rinse with phosphate buffered saline (PBS) and the tLBM self-assembles.

Ion-Transporting Supported and Tethered Lipid Bilayers

389

luid lipid bilayer available to incorporate the ion channel protein. When designing an experiment, it is preferable to use the highest tether density (i.e., largest x) possible for a particular molecular weight protein being incorporated into the tLBM. The molecular weights are calculated according to this simple geometric model. Of course, this molecular weight may only be a fraction of the molecular weight of a whole ion channel or other membrane-associated compound being studied with the tLBM. The conductance of the tLBM formed in that way can be measured using impedance spectroscopy, such as with the TethaPod supplied by SDX Tethered Membranes.* With this conductance measure from the TethaPod, the properties of the membrane itself or the change in conductance in response to the insertion of ion channels can be studied without a requirement for an in-depth understanding of impedance spectroscopy. The limited quantity of ions available in the reservoir between the tLBM and gold substrate means that the conductance needs to be recorded using alternating current (AC) (e.g., impedance spectroscopy) rather than using direct current (DC) (e.g., patch-clamp electrophysiology). The tLBM has an area that is an order of magnitude greater than the lipid bilayer normally suspended on a patch-clamp micropipette, and therefore, the overall sealed impedance of the tLBM is on the order of a megaohm rather than the gigaohm as for the patch-clamp micropipette. Ions in the saline solution for the tLBM create two capacitors, with one due to the insulating barrier of the membrane (~0.5 µF/cm2), and the second due to ions crowding at the gold surface (~3.5 µF/cm2). An AC excitation, chosen to be between 10 and 50 mV p–p amplitude, is swept in frequency from 0.1 Hz to 1 kHz and applied between the gold electrode and a large area gold return electrode in contact with the ionic solution. Current passes through these two capacitors and the voltage across a reference resistor due to the current low is ampliied and supplied via a universal serial bus (USB) port to a computer for processing and analysis. Although single-channel measurements are not available, the tethered membrane principle provides robust measurements of membrane conductance with sensitivity and lexibility of technique. The impedance spectroscopy provides a measurement of both phase and impedance at a range of frequencies for a range typically between 0.1 Hz and 1 kHz. These recordings of phase and impedance allow calculation of the membrane conductance attributable to the incorporated ion channel proteins (i.e., from the admittance at the frequency for minimum phase) and the thickness of the tLBM (i.e., from the admittance at 1 kHz). In our laboratories, we have assessed the stability of the tethered lipid bilayer by challenging the bilayer with detergents. For example, we have utilized the nonionic dialyzable detergent octyl β-d-glucopyranoside (OG). OG is also the detergent used for the micellization of bacteriorhodopsin during its puriication (Gordeliy et al. 2003). The critical micelle concentration (CMC) of OG is 23 mM (Morandat and El Kirat 2007). In these experiments, the mobile outer lealet of the tLBM was formed with DPEPC and GDPE in the mole ratio 70:30, respectively, and we used the solutions as the following: Buffer 3 TRO buffer

20 mM Na2HPO4/KH2PO4, pH 6.9 150 mM NaCl, 5 mM HEPES, pH 7.3

However, as shown in Figure 19.3, OG at the CMC reduces the impedance of the tethered lipid bilayer to that of the gold substrate electrode, due to micellization of the outer lealet of the bilayer. However, by reducing the concentration of OG to