Transmembrane Dynamics of Lipids (Wiley Series in Protein and Peptide Science)

TRANSMEMBRANE DYNAMICS OF LIPIDS WILEY SERIES ON PROTEIN AND PEPTIDE SCIENCE VLADIMIR N. UVERSKY, Series Editor Metal...

Author: Andreas Herrmann

40 downloads 710 Views 18MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

TRANSMEMBRANE DYNAMICS OF LIPIDS

WILEY SERIES ON PROTEIN AND PEPTIDE SCIENCE VLADIMIR N. UVERSKY, Series Editor Metalloproteomics • Eugene A. Permyakov Instrumental Analysis of Intrinsically Disordered Proteins: Assessing Structure and Conformation • Vladimir Uversky, Sonia Longhi Protein Misfolding Diseases: Current and Emerging Principles and Therapies • Marina Ramirez-Alvarado, Jeffery W. Kelly, Christopher M. Dobson Calcium Binding Proteins • Eugene A. Permyakov and Robert H. Kretsinger Protein Chaperones and Protection from Neurodegenerative Diseases • Stephan Witt Transmembrane Dynamics of Lipids • Philippe Devaux and Andreas Herrmann

INTRODUCTION TO THE WILEY SERIES ON PROTEIN AND PEPTIDE SCIENCE Proteins and peptides are the major functional components of the living cell. They are involved in all aspects of the maintenance of life. Their structural and functional repertoires are endless. They may act alone or in conjunction with other proteins, peptides, nucleic acids, membranes, small molecules, and ions during various stages of life. Dysfunction of proteins and peptides may result in the development of various pathological conditions and diseases. Therefore, the protein/peptide structure–function relationship is a key scientific problem lying at the junction point of modern biochemistry, biophysics, genetics, physiology, molecular and cellular biology, proteomics, and medicine. The Wiley Series on Protein and Peptide Science is designed to supply a complementary perspective from current publications by focusing each volume on a specific protein- or peptide-associated question and endowing it with the broadest possible context and outlook. The volumes in this series should be considered required reading for biochemists, biophysicists, molecular biologists, geneticists, cell biologists, and physiologists, as well as those specialists in drug design and development, proteomics, and molecular medicine with an interest in proteins and peptides. I hope that each reader will find in the volumes within this book series interesting and useful information. First and foremost, I would like to acknowledge the assistance of Anita Lekhwani of John Wiley & Sons, Inc., throughout this project. She has guided me through countless difficulties in the preparation of this book series, and her enthusiasm, input, suggestions, and efforts were indispensable in bringing the Wiley Series on Protein and Peptide Science into existence. I would like to take this opportunity to thank everybody whose contribution in one way or another has helped and supported this project. Finally, special thank you goes to my wife, sons, and mother for their constant support, invaluable assistance, and continuous encouragement. Vladimir N. Uversky September 2008

TRANSMEMBRANE DYNAMICS OF LIPIDS EDITED BY PHILIPPE F. DEVAUX ANDREAS HERRMANN

The Wiley Series in Protein and Peptide Science Series Editor: Vladimir N. Uversky

A JOHN WILEY & SONS, INC. PUBLICATION

The cover shows a transparency used by P.D. in presentations given in the days before PowerPoint was available. The cartoons illustrate the principal of the ascorbate assay to assess the transbilayer motion and distribution of spin-labeled lipids in membranes taking the plasma membrane of red blood cells as an example (see Preface and Chapters 1 and 6). Lower cartoon: Spin-labeled lipids (red) were incorporated into the outer leaflet of the plasma membrane and redistributed between both leaflets. Upper cartoon: Ascorbate was added to the cell suspension reducing spin-labeled lipids selectively on the outer leaflet. By comparing the signal intensity of spin-labeled lipids without and with ascorbate, the transbilayer distribution of labeled lipid analogs can be measured. The transparency has been used many times (see the spread colors on the left). Just by chance, it was rediscovered among an amazing pile of reprints in the office of P.D. Copyright © 2012 by John Wiley & Sons, Inc. All rights reserved Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley. com/go/permissions. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: Transmembrane dynamic lipids / edited by Philippe Devaux, Andreas Herrmann. â•…â•…â•… p. cm. – (The Wiley series in protein and peptide science ; 9) â•… Includes index. â•…â•… ISBN 978-0-470-38845-7 (hardback) â•‡ 1.â•‡ Membrane lipids.â•… 2.â•‡ Cell membranes.â•… 3.â•‡ Membrane transport.â•… I.â•‡ Devaux, Philippe F.â•… II.â•‡ Herrmann, Andreas. â•… QP752.M45T73 2012 â•… 572'.577–dc23 2011021424 Printed in the United States of America eISBN: 9781118120088 ePub: 9781118120101 oISBN: 9781118120118 MOBI: 9781118120095 10â•… 9â•… 8â•… 7â•… 6â•… 5â•… 4â•… 3â•… 2â•… 1

CONTENTS

INTRODUCTION LIST OF CONTRIBUTORS

xiii xxiii

PART Iâ•… ASSESSING TRANSMEMBRANE MOVEMENT AND ASYMMETRY OF LIPIDS

1

â•‡ 1â•… Methods for the Determination of Lipid Transmembrane Distribution and Movement in Biological Membranes

3

Philippe F. Devaux and Andreas Herrmann

1.1 1.2 1.3 1.4 1.5

Introduction,â•‡ 3 Development of Assays for Distribution and Translocation of Lipids across Membranes,â•‡ 4 Overview on Assays for Measuring Distribution and Translocation of Lipids across Cellular Membranes,â•‡ 7 Main Techniques Used to Determine Transbilayer Distribution of Endogenous Lipids in Cell Membranes,â•‡ 9 Main Techniques Used to Determine Transbilayer Distribution of Lipid Analogs in Cell Membranes,â•‡ 12 Abbreviations,â•‡ 21 References,â•‡ 21

â•‡ 2â•… Detection and Measurement of Unlabeled Lipid Transmembrane Movement

25

Iván López-Montero, Marisela Vélez, and Philippe F. Devaux

2.1 2.2

Introduction,â•‡ 25 Measurement of Transmembrane Flip-Flop of Unlabeled Lipids by Shape Change of GUVs,â•‡ 27 v

vi

CONTENTS

2.3 2.4

Measurement of Transmembrane Flip-Flop of Unlabeled Lipids Using AFM,â•‡ 36 Conclusions,â•‡ 41 Acknowledgments,â•‡ 41 Abbreviations,â•‡ 41 References,â•‡ 42

PART IIâ•… LIPID ASYMMETRY IN CELL MEMBRANES

45

â•‡ 3â•… New Insights in Membrane Lipid Asymmetry in Animal and Plant Cells

47

Alain Zachowski

3.1 3.2 3.3

Lipid Asymmetry in Animal Membranes,â•‡ 47 Creating, Maintaining, or Randomizing the Membrane Phospholipid Distribution: Phospholipid Transporters,â•‡ 49 What about Lipid Asymmetry and Translocation in Plant Cell Membranes?,â•‡ 50 Abbreviations,â•‡ 61 References,â•‡ 61

â•‡ 4â•… Sphingolipid Asymmetry and Transmembrane Translocation in Mammalian Cells

65

Gerrit van Meer, Sylvia Neumann, and Per Haberkant

4.1 4.2 4.3 4.4 4.5 4.6

Introduction,â•‡ 65 Sphingosine, Sphingosine-1-Phosphate, and Ceramide,â•‡ 67 Ceramide,â•‡ 68 Glycosphingolipids,â•‡ 68 Sphingomyelin,â•‡ 70 Future Perspectives,â•‡ 71 Abbreviations,â•‡ 71 References,â•‡ 71

â•‡ 5â•… Transbilayer Movement and Distribution of Cholesterol Peter Müller, Anna Pia Plazzo, and Andreas Herrmann

5.1 5.2 5.3 5.4 5.5

Introduction,â•‡ 75 Physicochemical Features of Cholesterol,â•‡ 76 Methods for Measuring Cholesterol Transbilayer Movement and Distribution,â•‡ 77 Transbilayer Movement of Cholesterol in Model Membranes,â•‡ 81 Transbilayer Movement of Cholesterol in Biological Membranes,â•‡ 82

75

vii

CONTENTS

5.6 5.7 5.8 5.9

Transbilayer Distribution of Cholesterol in Lipid and Biological Membranes,â•‡ 82 Cholesterol Flip-Flop: Fast or Slow?,â•‡ 87 Role of Proteins in the Transport of Cholesterol across Membranes,â•‡ 88 Concluding Remarks,â•‡ 90 Acknowledgment,â•‡ 92 Abbreviations,â•‡ 92 References,â•‡ 93

PART IIIâ•… ENERGY-INDEPENDENT PROTEIN-MEDIATED TRANSMEMBRANE MOVEMENT OF LIPIDS

97

â•‡ 6â•… Phospholipid Flip-Flop in Biogenic Membranes

99

Anant K. Menon and Andreas Herrmann

6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10

Introduction,â•‡ 99 Assays for Measuring Transbilayer Distribution of Endogenous Phospholipids,â•‡ 100 Assays for Measuring Transbilayer Distribution and Movement of Phospholipid Analogs,â•‡ 102 Shape Changes of GUVs as a Tool to Measure Flip-Flop,â•‡ 106 Transbilayer Movement of Phospholipids in the ER,â•‡ 108 Transbilayer Movement of Phospholipids in the Bacterial Inner Membrane,â•‡ 110 Mechanism of Rapid Lipid Flip-Flop in Biogenic Membranes,â•‡ 112 Efforts to Identify Phospholipid Flippases,â•‡ 113 Flipping of Isoprenoid-Based Glycolipids,â•‡ 115 Conclusion,â•‡ 115 Abbreviations,â•‡ 116 References,â•‡ 116

â•‡ 7â•… Phospholipid Scramblase: When Phospholipid Asymmetry Goes Away Edouard M. Bevers and Patrick L. Williamson

7.1 7.2 7.3 7.4 7.5 7.6

Introduction,â•‡ 119 Historical Overview,â•‡ 120 Physiological Importance of Lipid Scrambling,â•‡ 122 Characteristics of the Phospholipid Scrambling Process,â•‡ 124 Toward Identification: Proposed Candidate Proteins and Mechanisms,â•‡ 132 Concluding Remarks,â•‡ 139 Abbreviations,â•‡ 139 References,â•‡ 140

119

viii

CONTENTS

PART IVâ•… ENERGY-DEPENDENT LIPID TRANSPORT ACROSS MEMBRANES

147

â•‡ 8â•… Flip or Flop: Mechanism and (Patho)Physiology of P4-ATPase-Catalyzed Lipid Transport

149

Patricia M. Verhulst, Joost C.M. Holthuis, and Thomas G. Pomorski

8.1 8.2 8.3 8.4 8.5 8.6

Introduction,â•‡ 149 P4-ATPases are Prime Candidate Phospholipid Translocases,â•‡ 152 Mechanism of P4-ATPase-Catalyzed Lipid Transport: Role of Accessory Subunits,â•‡ 156 Role of P4-ATPases in Vesicle-Mediated Protein Transport,â•‡ 161 P4-ATPase Dysfunction and Disease,â•‡ 162 Future Challenges,â•‡ 166 Acknowledgments,â•‡ 166 Abbreviations,â•‡ 166 References,â•‡ 167

â•‡ 9â•… Coupling Drs2p to Phospholipid Translocation, Membrane Asymmetry, and Vesicle Budding

171

Xiaoming Zhou, Paramasivam Natarajan, Baby-Periyanayaki Muthusamy, Todd R. Graham, and Ke Liu

9.1 9.2 9.3 9.4 9.5

Introduction,â•‡ 171 P4-ATPases in Budding Yeast,â•‡ 172 Evidence That Drs2p Is a Flippase,â•‡ 175 Drs2p in Protein Transport and Vesicle Budding,â•‡ 183 Concluding Remarks,â•‡ 191 Abbreviations,â•‡ 192 References,â•‡ 193

10â•… Substrate Specificity of the Aminophospholipid Flippase

199

Shelley M. Cook and David L. Daleke

10.1 10.2 10.3 10.4 10.5 10.6 10.7

Introduction,â•‡ 199 Substrate Specificity of the PM Aminophospholipid Flippase,â•‡ 200 Identification and Substrate Specificity of Candidate Aminophospholipid Flippases,â•‡ 205 Is the Lipid Specificity of Candidate Aminophospholipid Flippases Unique?,â•‡ 210 Lipid Specificity of Other PS-Binding Proteins,â•‡ 213 Sequence Elements That Bind to PS,â•‡ 215 Conclusions,â•‡ 216 Acknowledgments,â•‡ 217 Abbreviations,â•‡ 218 References,â•‡ 218

ix

CONTENTS

11â•… The Flippase Delusion?

225

Naomi L. Pollock, Petra H.M. Niesten, and Richard Callaghan

11.1 11.2 11.3 11.4 11.5 11.6

ATP-Binding Cassette (ABC) Transporters and Lipid Flip-Flop,â•‡ 225 ABCA4 and Lipid Translocation: Explaining a Phenotype?,â•‡ 228 MsbA and Lipid Translocation: A Key to Survival,â•‡ 230 Drug and Lipid Movement by ABCB1: Is the Mechanism a Flip-Flop?,â•‡ 237 ABCB4: The Forgotten and Likely Lipid Flippase?,â•‡ 240 Conclusions and Perspectives,â•‡ 244 Abbreviations,â•‡ 244 References,â•‡ 245

PART Vâ•… RELEVANCE OF LIPID TRANSMEMBRANE DISTRIBUTION FOR MEMBRANE PROPERTIES AND PROCESSES

251

12â•… Membrane Lipid Asymmetry and Permeability to Drugs: A Matter of Size

253

Adam Blanchard and Cyril Rauch

12.1 12.2 12.3 12.4 12.5

Introduction,â•‡ 253 The Origin of Lipinski’s Second Rule from the Point of View of the Pharmaceutical Industry,â•‡ 254 Solving Lipinski’s Second Rule,â•‡ 257 Lipinski’s Second Law and Potential Application,â•‡ 264 Conclusion,â•‡ 270 Acknowledgment,â•‡ 272 Abbreviations,â•‡ 272 References,â•‡ 273

13â•… Endocytosis and Lipid Asymmetry Nina Ohlwein, Andreas Herrmann, and Philippe F. Devaux

13.1 13.2 13.3 13.4 13.5 13.6

Introduction,â•‡ 275 Bending a Membrane,â•‡ 276 Shape Changes of GUVs Induced by Lipid Asymmetry,â•‡ 278 How Endocytosis Is Linked to Lipid Asymmetry,â•‡ 280 Role of P4-ATPases in the Formation of Endocytic Invaginations,â•‡ 283 Concluding Remarks,â•‡ 284 Acknowledgments,â•‡ 285 Abbreviations,â•‡ 285 References,â•‡ 285

275

x

CONTENTS

PART VIâ•… APOPTOSIS AND DISEASES: CONSEQUENCES OF DISRUPTION TO LIPID TRANSMEMBRANE ASYMMETRY

289

14â•… Membrane Lipid Asymmetry in Aging and Apoptosis

291

Krishnakumar Balasubramanian and Alan J. Schroit

14.1 14.2 14.3 14.4 14.5 14.6 14.7 14.8 14.9

Introduction,â•‡ 291 Phospholipid Transporters,â•‡ 292 Lipid Asymmetry in Erythrocytes,â•‡ 294 Lipid Asymmetry during Apoptosis,â•‡ 297 Ca2+ Homeostasis during Apoptosis,â•‡ 298 Membrane Phospholipid Asymmetry: Static or Dynamic?,â•‡ 299 Regulation of Lipid Asymmetry during Apoptosis,â•‡ 300 Significance,â•‡ 304 Concluding Remarks,â•‡ 306 Abbreviations,â•‡ 306 References,â•‡ 307

15â•… Phosphatidylserine Exposure in Hemoglobinopathies

315

Frans A. Kuypers and Eric Soupene

15.1 15.2 15.3 15.4 15.5 15.6 15.7 15.8

Introduction,â•‡ 315 RBC Phospholipid Organization,â•‡ 316 The RBC Flippase,â•‡ 319 PS Exposure in RBCs,â•‡ 324 PS Exposure in Hemoglobinopathies,â•‡ 328 Consequences of PS Exposure,â•‡ 329 Phospholipid Transbilayer Movement in Hemoglobinopathies,â•‡ 330 Conclusion,â•‡ 332 Abbreviations,â•‡ 333 References,â•‡ 334

16â•… Scott Syndrome: More Than a Hereditary Defect of Plasma Membrane Remodeling Florence Toti and Jean-Marie Freyssinet

16.1 16.2 16.3 16.4 16.5 16.6

Introduction,â•‡ 341 Scott Syndrome Features and Phenotype,â•‡ 342 Cell Biology of Scott Syndrome,â•‡ 343 Candidate Proteins in the Transmembrane Redistribution of PS,â•‡ 345 The Significance of Membrane Vesiculation and of Derived MPs,â•‡ 346 What Can Be Learned from Scott Syndrome?,â•‡ 347

341

xi

CONTENTS

16.7

Conclusion,â•‡ 348 Abbreviations,â•‡ 349 References,â•‡ 350

17â•… ABCA1, Tangier Disease, and Lipid Flopping

353

Ana Zarubica and Giovanna Chimini

17.1

Historical Notes: Tangier Disease (TD) and ATP-Binding Cassette Transporter 1 (ABCA1),â•‡ 353 17.2 The ABCA1 Gene and the Regulation of Its Expression,â•‡ 354 17.3 The ABCA1 Protein and Its Interactions,â•‡ 356 17.4 ABCA1: Mutations and Clinical Signs,â•‡ 358 17.5 Targeted Inactivation and Overexpression of ABCA1 in Animal Models,â•‡ 360 17.6 Liver and Macrophage ABCA1: Lipid Efflux and HDL Formation,â•‡ 362 17.7 ABCA1 and Membrane Function,â•‡ 363 17.8 ABCA1: Lipid Flop and Lipid Efflux,â•‡ 364 17.9 ABCA1 and the Lipid Microenvironment at the Membrane,â•‡ 366 17.10 Conclusions,â•‡ 368 Acknowledgments,â•‡ 369 Abbreviations,â•‡ 369 References,â•‡ 371 INDEX

379

Note: Color versions of many of the black and white figures in this book can be viewed at ftp://ftp.wiley.com/public/sci_tech_med/transmembrane_dynamics. Please read the figure captions to learn which figures are available on the ftp site.

INTRODUCTION

HISTORICAL PERSPECTIVES: WHO DID WHAT AND WHAT’S NEXT? Ole Mouritsen, in his recent monograph entitled “Lipids—As a Matter of Fat,” summarized with humor the views of many biologists concerning lipids, as follows: “Lipids appear to play a fairly non-specific role, being rather dull and anonymous compared to fashionable stuff like the proteins that catalyze all biochemical reactions and the genes that contain the information needed to produce proteins” [1]. The present book, which is addressed to researchers, teachers, and students in cell biology and in biochemistry, has the goal of convincing all scientists that lipids, on the contrary, have sophisticated behaviors and play multiple important roles in living organisms. It is also addressed to physicists fascinated by the various spontaneous self-organization of lipids in water (lipid polymorphism) to warn them that lipids in biological systems are not always at thermal equilibrium, and that phase separations and lateral or transmembrane domains seen in model systems can differ fundamentally from biological situations. Indeed, molecule segregation in biological systems results often from the work of ATPases, like the flippases, or is the result of a molecule sorting by “protein gates” (see the “fence and picket model” of Kusumi and collaborators [2]). Such mechanisms are difficult to mimic in model systems. In any case, all lipids are not equivalent and their chemical heterogeneity, for example, between the two sides of a biomembrane, is the result of a long selection during evolution, which allows lipids to fulfill different functions, from that of a fluid hydrophobic medium for membrane proteins to that of selective messenger molecules and enzyme cofactors. In the latter case, they xiii

xiv

INTRODUCTION

have to find their partners in a cell, hence to move rapidly in a very anisotropic environment. To many biologists, lipids form the third class of molecules of living organism after proteins and nucleic acids. Yet, lipids were probably not the third in the evolution nor are they third in importance, since a cell and even many viruses cannot exist without a membrane. The fact is that lipids form the building blocks of biological membranes. They determine the boundary of all living organisms as well as the compartmentalization of organelles in eukaryotes. Regarded as passive molecules forming only viscous cement that holds membrane proteins, filtering out hydrophilic molecules, the lipid bilayer is in reality a sophisticated structure capable of a remarkable polymorphism in water. The physical characteristic of a lipid bilayer permits not only protein movement but also membrane deformations and, coupled to the cytoskeleton, provides the cell membrane with mechanical properties. Not the least astonishing is the bilayer’s ability to divide in two compartments during cell division without losing molecules in the plasma due to efficient self-sealing capacities. Nonetheless, there are still mysteries concerning lipids, which are matters of research, speculation, and controversy. (1) Biophysicists have succeeded in making stable membranes (liposomes) with only one type of lipids, in suspension in water, for example, with egg phosphatidylcholine (PC), while biological membranes harbor several hundred different lipids. Why are there so many chemically different lipids coexisting in nature? (2) Why is the lipid composition of various membranes of eukaryotes different and sometimes even the two sides of biological membranes different (asymmetrical)? This requires numerous specific enzymes for the synthesis and ultimately for the shuttling to the right destination of newly formed lipids. Is such a multiplicity necessary for a fine-tuning of membrane-bound enzymes or is the variety of lipids used to give specific messages to specific proteins? Is the detailed chemical structure of lipids without real importance and does it reflect only the precursor molecules available? Not only do eukaryotic membranes have many chemically different lipids if one considers chain length, unsaturation, and polar head group, but also the lipids are not homogenously distributed within the various organelles and even between the different sides of one membrane. This lipid heterogeneity, a “complication of Nature,” was transmitted more than a million years in eukaryotic cells and has survived the filter of evolution, suggesting that the lipid composition and distribution within a cell is neither accidental nor inconsequential for the activity of cells. Although cells tolerate certain variability in lipid composition, many human diseases have been associated with the inability of mutated cells to synthesize specific lipids or to recycle particular lipids from the nutriments or to address specific lipids to their correct destination. Alternatively, the excess of certain lipids such as cholesterol or saturated phospholipid chains can be poisonous. In the late 1960s, V. Luzzati, in a pioneer work carried out in France, showed by X-ray crystallography that lipids extracted from biological membranes form, in water, lamellar phases, giving rise spontaneously to large multilamel-

INTRODUCTION

xv

lar (onion-style) liposomes made of a superposition of bilayers [3, 4]. Physicists characterized the bilayers as liquid crystals that could be in a fluid state or in a more viscous, gel state. In the early 1970s, the concept of lipid bilayer emerged as the basic model of biomembranes and was popularized in the famous model of “fluid mosaic membrane” of S.J. Singer and G.L. Nicolson [5]. Although the concept of “mosaicity” implies the presence of heterogeneous lateral domains, and in spite of the work carried out by several physical chemists such as H. McConnell, it was only in 1997 (almost 30 years after the initial work of Luzzati and McConnell) that the importance of lateral domains began to be popular among membranologists and that biological functions associated with lateral domains (or rafts) were highlighted (see the work of K. Simons and E. Ikonen [6]). Indeed, the two monolayers of biomembranes form distinct lipid domains: M. Bretscher in England demonstrated in the early 1970s the asymmetrical transmembrane distribution of phospholipids in the plasma membrane of human erythrocytes [7]. Bretscher used the chemical labeling of the amino groups of phosphatidylserine (PS) or phosphatidylethanolamine (PE) and showed that aminophospholipids are principally in the membrane inner monolayer, while PC and sphingomyelin (SM) are essentially in the outer monolayer of human red cells. Subsequent investigation in the laboratory of L.L.M. van Deenen in The Netherlands based on phospholipases and sphingomylinases assays [8, 9] confirmed Bretscher’s results and demonstrated that the transmembrane asymmetry of red cells is an ubiquitous property of the plasma membrane of eukaryotes. In model systems, on the other hand, no transmembrane lipid segregation was found to form spontaneously. Sonication allows one to achieve a lipid sorting between inner and outer monolayers in small unilamellar vesicles (SUVs), but the latter structures are not physiological because of their small size compared with that of vesicles produced in vivo (∼20-nm diameter for SUVs vs. ∼200â•¯nm for endocytic vesicles). Thus, lipid sorting observed in biomembranes had to be caused by a process that does not exist in liposomes and is not a mere thermodynamic equilibrium. Initially, the segregation of aminophospholipids was believed to be due to the topology of enzymes responsible for lipid synthesis or to lipid–cytoskeleton interactions (J.A.F. Op den Kamp [10]). However, Bretscher had the remarkable intuition to postulate the existence of specific lipid enzymes that he named “phospholipid flippase,” which would be responsible for the establishment of the asymmetrical lipid organization at the expense of ATP hydrolysis. In practice, it was later found necessary to specify the orientation of the postulated lipid carrier and the requirement or absence of requirement for ATP hydrolysis. This explains why the habit is now to differentiate among flippase, floppase, and scramblase (Fig. I.1). A prerequisite for stable lipid segregation between the two monolayers of a membrane is a priori a slow transmembrane diffusion. In 1971, R.D. Kornberg and H.M. McConnell at Stanford University demonstrated for the first time, with spin-labeled lipids, the very slow transmembrane diffusion of

xvi

INTRODUCTION

extracellular or lumenal

ATP

flippase

cytoplasm

ATP

floppase

scramblase

flip-flop

Figure I.1.â•‡ Definition of the various lipid transporters in eukaryotic cell membranes. Note that the scramblase is calcium dependent and that “flippase” is a term that is used sometimes to designate an enzyme that catalyzes lipid flip-flop in both directions (inward or outward), for example, in the endoplasmic reticulum.

phospholipids in sonicated lipid vesicles, where the “flip-flop” between the two monolayers was found to require several hours at 30°C [11]. It is now admitted that the spontaneous transmembrane diffusion of lipids is very slow in liposomes of any size as well as in biological membranes. A few exceptions to this rule were discovered recently. Cholesterol, ceramide, phospatidic acid, diacylglycerol, and free fatty acids or esters have a rapid spontaneous diffusion (τ1/2 less than 1 minute). The absence of real polar head groups in such lipids probably explains this unusual result. It was only in 1984, that is, more than 10 years after Bretscher’s hypothesis, that the existence of a phospholipid flippase was demonstrated in France by M. Seigneuret and P.F. Devaux in the human erythrocyte membrane using spin-labeled analogs of naturally occurring phospholipids [12] and the year after by D.L. Daleke and W.H. Huestis, who provided confirmation using an elegant technique involving nonlabeled lipids [13], while A. Schroit’s group [14] took advantage of fluorescent analogs to prove the existence of an erythrocyte aminophospholipid transporter. The requirement of hydrolyzable Mg2+ATP was demonstrated as being necessary for the rapid transport of aminophospholipids, and the specificity was carefully investigated; however, no proteins were identified initially. In 1989, an ATP-dependent flippase activity in chromaffin granules from bovine adrenal medulla was reported by the Paris laboratory and attributed to the so-called ATPase II [15]. This was the first report of aminophospholipid translocase activity in the inner membranes of the eukaryotic cell. The transport observed was in fact from the lumen to the cytosol of the granules but was classified as a flippase activity. In 1996, P. Williamson and R.A. Schlegel’s groups in the United States showed that this granule flippase was homolog to a yeast ATPase (called Drs2p), and studied a mutant deprived of Drs2 that was unable to flip aminophospholipids [16]. The phospholipid flippase seemed to be discovered. However, in 1999 and 2003, the groups of T. Graham in the United States [17] and G. van Meer and J. Holthuis in The Netherlands [18] showed that Drs2p is in fact localized in the yeast trans-Golgi and not in the plasma mem-

INTRODUCTION

xvii

brane, and that five homologs of this protein exist: two in the plasma membrane (Dnf1p and Dnf2p), two in the trans-Golgi (Dnf3p and Drs2p), and one (Neo1p) in endosomes or cis-Golgi. Furthermore, these P-type ATPases seem to be associated with other proteins playing the role of chaperones (CDC50p) or are necessary for the proper targeting to their final destination of the newly formed proteins [19]. In 2006, P. Natarajan and T. Graham [20] showed a flippase activity with fluorescent lipids in yeast Golgi membranes, which they could attribute to the Drs2p. Interestingly, the triple knockout of the Drs2p homologs in yeast led to viable cells, but they were deprived of endocytic activity [18]. In conclusion, the various P-type ATPases may have different specificities but may also be partially redundant. Thus, after about 20 years of research in different laboratories throughout the world, it became obvious that the ubiquitous eukaryotic flippase was in reality a combination of several proteins, including four ATPases called P4ATPase, actually forming a family of five proteins in yeast. In humans, it was predicted from genomic investigation that 14 P4-ATPases were members of the family and could be involved in lipid transport. The purification of specific P4-ATPases and of Drs2p from chromaffin granules or after expression in various systems (yeast and insect cells) is in progress. However, so far the purification has not been achieved on a large enough scale to allow unambiguous tests of lipid transport in reconstituted lipoproteins. Other membrane proteins were reported to have an ATP-dependent lipid translocation activity and correspond to the so-called floppases (see Fig. I.1) with an ATP-binding cassette (ABC). Suggested originally by C.F. Higgins and M.M. Gottesman in 1992 [21], the laboratories of G. van Meer and of P. Borst in The Netherlands [22] showed in 1996 that the ABC transporter P-glycoprotein, also called MDR1, which is responsible for multidrug resistance and is a serious obstacle in cancer therapy, was able to transport fluorescent phospholipids from the inner monolayer to the outer monolayer of the plasma membrane of eukaryotic cells. The low specificity of the P-glycoprotein suggested that this protein could be involved in the transport of SM and PC toward the outer monolayer of the plasma membrane, hence play an important role in the transmembrane lipid asymmetry of the eukaryotic plasma membrane. Other members of the ABC protein family seemed to be responsible for the specific outward transport of PC in transfected epithelial cells [22]. An important point is that ABC proteins are also found in prokaryotes and could be implicated in lipid translocation in bacteria [23]. Besides ATP-dependent flippases, which were found essentially in the plasma membrane of eukaryotes, other ATP-independent proteins also called flippases were postulated to be in specific organelle membranes (endoplasmic reticulum) and could explain the rapid flip-flop observed by several groups. Their primary function would be to facilitate the transmembrane diffusion of lipids in the membranes specialized in lipid synthesis. Already in 1985, W.R. Bishop and R.M. Bell [24] suggested the existence of ATP-independent flippase, catalyzing the diffusion of PC in the endoplasmic reticulum. Since

xviii

INTRODUCTION

then, several researchers have attempted to isolate the protein(s) responsible (A. Menon in the United States [25] and A. Herrmann and collaborators in Germany [26]). Other researchers have attempted to prove that any transmembrane protein should accelerate the flip-flop of lipids, making it unnecessary to search for specific proteins (B. de Kruijff and A. Killian and collaborators in The Netherlands [27, 28]). In practice, the identification of the ATPindependent flippase of low lipid specificity seems even more difficult than it is for the ATP-dependent selective flippase, precisely because the test of ATP requirement cannot be used to discover the latter transporter. BIOLOGICAL ADVANTAGES OF LIPID ASYMMETRY The complexity involved in the regulation of lipid topology, requiring ATP hydrolysis, raises the question of the biological function(s) of such an elaborate system. Actually, one might rephrase this question differently: The lipid composition of a biological membrane is always a mixture of many different lipids. The actual justification of this fact is not obvious, since a stable lipid bilayer can be achieved in liposomes with a single phospholipid species. So what is the biological advantage of the synthesis of many different lipids? A reasonable hypothesis would be that lipid asymmetry is used to tag the two sides of a membrane and to optimize their functionality, which is obviously different. Indeed, the cell outer environment differs fundamentally from the cytosol. One of the first indications of the physiological importance of lipid asymmetry came from the observation by A. Schroit and collaborators (United States) who showed in 1983 and 1985 that the presence of a very small percentage of PS (∼1% of the total lipid composition) in the outer monolayer of red cells was used in vivo as a signal of cell aging and led in the blood circulation to the elimination of aged cells by macrophages [29]. These conclusions, which came originally from experiments associated with the introduction of exogenous PS in the outer monolayer of red cells, were confirmed later by the detection of natural PS with fluorescent Annexin V by J.F. Tait and D. Gibson in 1994 [30]. The exposure of PS in the outer monolayer of platelets is also associated with the formation of clots that stop bleeding (R. Zwaal and collaborators [31]). Thus, lipid flip-flop concerns directly at least two important physiological problems: (1) blood coagulation, which is triggered in vivo by the exposure of PS, a cofactor required for the conversion of prothrombin into thrombin, and (2) elimination of aged and/or apoptotic cells by macrophages. The lipid randomization, that is, loss of lipid asymmetry that is used in vivo as a signal for cell elimination, can be triggered artificially by penetration of calcium ions in the cytosol of platelets, erythrocytes, or lymphocytes with calcium ionophores, and results in “lipid scrambling,” that is, lipid randomization between the two leaflets. This phenomenon is associated with a so far unknown protein named

INTRODUCTION

xix

“scramblase.” (Note: Suzuki et al. [40] identified the protein TMEM16F as an essential component for Ca2+-dependent exposure of PS.) A rare but severe disease, called “Scott syndrome,” is characterized by the absence of PS redistribution upon calcium entry and has been investigated by R. Zwaal’s group in The Netherlands [32] and by J.-M. Freyssinet and collaborators [33] in France. Various severe diseases such as cancer and Alzheimer’s disease were also reported to be accompanied by defects in lipid asymmetry [34]. ABCA1, another lipid transporter of the family of ABC-ATPases, was considered to be responsible for Tangier disease, characterized by impaired efflux of cholesterol and phospholipids from peripheral cells onto apolipoproteins such as Apo A-1. Cholesterol accumulation in macrophages and apolipoprotein degradation lead to tissue deposition of cholesterol esters and increase the risk of arteriosclerosis in patients. G. Chimini and collaborators in Marseilles studied this particular defect associated with a lipid transporter [35]. In humans, several mutated ABC proteins reputed to be responsible for lipid transport are believed to cause metabolism disorders such as Stargardt syndrome (a genetic disease of vision), progressive intrahepatic cholestasis, pseudoxanthoma elasticum, adrenoleukodystrophy, or sitosterolemia. In 1999, E. Farge and collaborators, in A. Dautry-Varsat’s laboratory, provided evidence of a biological role played by a lipid transporter during the first step of endocytosis [36]. It was shown that the transport of PS and PE from the outer to the inner monolayer by the ATP-dependent flippase is a stimulation of endocytosis and could be the molecular motor of membrane bending involved in the first step of endocytosis. The explanation proposed was that the excess of lipids in one monolayer triggers membrane invagination, as shown in model systems [37]. The yeast knockout experiments mentioned above [18] confirmed that in the absence of flippase proteins, endocytosis was blocked. There are also reports suggesting that PS is important for fusion; hence, it could be useful in the inner monolayer for exocytosis and not only for the regulation of inner leaflet proteins. There are certainly many other enzymes that require specific lipids at specific positions in a cell. Actually, the difference in head group of the lipids from the two sides of a membrane is not the only difference between inner and outer leaflets lipids. Indeed, there is evidence of difference in unsaturation, which is associated with differences in membrane viscosity, as observed in erythrocytes with spin-labeled lipids [38] and with fluorescent lipids [39]: The inner monolayer is more fluid; the outer is more rigid, hence more resistant. It is very likely that this feature is associated with the activity of proteins. When transmembrane lipid asymmetry was demonstrated in red cells and soon after in the plasma membranes of all animals, it was assumed that this feature was a general property of living organisms. This may be true in animal

xx

INTRODUCTION

and in plant cells, which are both eukaryotes. But the evidence regarding prokaryotes is limited and often concerns rare lipids. PROSPECTS What kind of progress can be expected in a reasonable time? Clearly, the bottleneck for progress in understanding the mechanism of lipid translocation by membrane proteins in eukaryotes has been the difficulty in assigning, isolating, and overexpressing the protein(s) responsible for this process; studying the properties of proteoliposomes; and crystallizing a flippase. Crystallization will be a necessary step for ultimately understanding the mechanism that allows a hydrophobic transmembrane protein to accumulate against a gradient amphiphilic molecule. There are some reports, at low resolution, on the structure of ABC proteins possibly involved in lipid transport. With P4-type ATPases, the data obtained with Ca2+-ATPase can be used as first-order approximation to stimulate the speculations of researchers, but the difference between a lipid and a calcium ion is so large that the detailed analysis of the mechanism is presumptuous. In any case, the determination of the structure and molecular mechanism of a flippase is a challenge for the coming years. It is therefore an objective that cannot be forsaken. Progress in the molecular biology and purification of the P4-type ATPases will lead to this achievement. Other objectives are as follows: 1. isolation of protein(s) responsible for ATP-independent rapid lipid flipflop in the endoplasmic reticulum; 2. isolation of protein(s) responsible for calcium-induced lipid scrambling (scramblase); 3. deeper understanding of all the consequences of lipid asymmetry, including recognition of the diseases caused specifically by a defect (impairment) in flippase activity. ORGANIZATION OF THIS BOOK As shown in the Table of Contents of this book, each chapter concentrates on one particular aspect of lipid asymmetry in biomembranes. However, we are not yet in a situation to give a complete and rational picture. As a consequence, one of the main difficulties in assembling this book was to choose a rational order for the chapters. Although the various chapters are closely linked to each other, there were no compelling reasons to decide which subjects deserved to be first or second. Hence, some repetition is unavoidable and the order of the chapters is rather arbitrary. Nevertheless, we must apologize for this weakness. On the other hand, each chapter can stand alone and does not necessarily require the reading of other chapters.

INTRODUCTION

xxi

ACKNOWLEDGMENTS We would like to express our gratitude to the contributing authors, in particular for being so flexible and sympathetic. Many of the contributing authors belonged to an international research and training network, “Flippases,” which was funded by the European Union from 2005 to 2008. This network provided an important trigger for assembling the book. We thank Professor Sophie Cribier (Paris) and Professor Daniel Picot (Paris) for their help in speeding up the final steps for assembling the book. Finally, we are very much indebted to Anita Lekhwani, Senior Acquisitions Editor at John Wiley and Sons, and to Catherine Odal, Assistant to the Senior Acquisitions Editor, for their efficient collaboration and support as well as for being flexible and handling unforeseen problems pleasantly. November 2010

Philippe F. Devaux Andreas Herrmann

REFERENCES â•‡ 1â•… O. Mouritsen, Lipid—As a Matter of Fat. The Merging Science of Lipidomics, Springer, Berlin, 2005. â•‡ 2â•… C. Nakada, K. Ritchie, Y. Oba, M. Nakamura, Y. Hotta, R. Iino, R. S. Kasai, K. Yamaguchi, T. Fujiwara, A. Kusumi, Nat. Cell Biol. 2003, 5, 626–632. â•‡ 3â•… R. P. Rand, V. Luzzati, Biophys. J. 1968, 8, 125–137. â•‡ 4â•… V. Luzzati, F. Reiss-Husson, E. Rivas, T. Gulik-Krzywicki, Ann. N.Y. Acad. Sci. 1966, 137, 409–413. â•‡ 5â•… S. J. Singer, G. L. Nicolson, Science 1972, 175, 720–731. â•‡ 6â•… K. Simons, E. Ikonen, Nature 1997, 387, 569–572. â•‡ 7â•… M. S. Bretscher, Science 1973, 181, 622–629. â•‡ 8â•… A. J. Verkleij, R. F. A. Zwaal, B. Roelofsen, P. Comfurius, D. Kastelijn, L. L. M. van Deenen, Biochim. Biophys. Acta 1973, 323, 178–193. â•‡ 9â•… R. F. A. Zwaal, B. Roelofsen, P. Comfurius, L. L. M. van Deenen, Biochim. Biophys. Acta 1975, 406, 83–96. 10â•… J. A. F. Op den Kamp, Biochemistry 1979, 48, 47–71. 11â•… R. D. Kornberg, H. M. McConnell, Biochemistry 1971, 10, 1111–1120. 12â•… M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. 13â•… D. L. Daleke, W. H. Huestis, Biochemistry 1985, 24, 5406–5416. 14â•… J. Connor, A. J. Schroit, Biochemistry 1987, 26, 5099–5105. 15â•… A. Zachowski, J. P. Henry, P. F. Devaux, Nature 1989, 340, 75–76. 16â•… X. J. Tang, M. S. Halleck, R. A. Schlegel, P. Williamson, Science 1996, 272, 1495–1497. 17â•… C.-Y. Chen, M. F. Ingram, P. H. Rosal, T. R. Graham, J. Cell Biol. 1999, 147, 1223–1236. 18â•… T. Pomorski, R. Lombardi, H. Riezman, P. F. Devaux, G. van Meer, J. C. Holthuis, Mol. Biol. Cell 2003, 14, 1240–1254.

xxii

INTRODUCTION

19â•… K. Saito, K. Fujimura-Kamada, N. Furuta, U. Kato, M. Umeda, K. Tanaka, Mol. Biol. Cell 2004, 15, 3418–3432. 20â•… P. Natarajan, T. R. Graham, Methods 2006, 39, 163–168. 21â•… C. F. Higgins, M. M. Gottesman, Trends Biochem. Sci. 1992, 17, 18–21. 22â•… A. van Helvoort, A. J. Smith, H. Sprong, I. Fritzsche, A. H. Schinkel, P. Borst, G. van Meer, Cell 1996, 87, 507–517. 23â•… A. Pohl, P. F. Devaux, A. Herrmann, Biochim. Biophys. Acta 2005, 1733, 29–52. 24â•… W. R. Bishop, R. M. Bell, Cell 1985, 42, 51–60. 25â•… A. Menon, W. E. Watkins, III, S. Hrafnsdóttir, Curr. Biol. 2000, 10, 241–252. 26â•… S. Vehring, L. Pakkiri, A. Schroer, N. Alder-Baerens, A. Herrmann, A. K. Menon, T. Pomorski, Eukaryot. Cell 2007, 6, 1625–1634. 27â•… M. A. Kol, A. I. P. M. de Kroon, J. A. Killian, B. de Kruijff, Biochemistry 2004, 43, 2673–2681. 28â•… M. A. Kol, A. I. P. M. de Kroon, D. T. S. Rijkers, J. A. Killian, B. de Kruijff, Biochemistry 2001, 40, 10500–10506. 29â•… A. J. Schroit, J. W. Madsen, Y. Tanaka, J. Biol. Chem. 1985, 260, 5131–5138. 30â•… J. F. Tait, D. Gibson, J. Lab. Clin. Med. 1994, 123, 741–748. 31â•… E. M. Bevers, P. Comfurius, R. F. A. Zwaal, Biochim. Biophys. Acta 1983, 736, 57–66. 32â•… E. M. Bevers, T. Wiedmer, P. Comfurius, S. J. Shattil, H. J. Weiss, R. F. A. Zwaal, P. J. Sims, Blood 1992, 79, 380–388. 33â•… N. Bettache, P. Gaffet, N. Allegre, L. Maurin, F. Toti, J.-M. Freyssinet, A. Bienvenue, Br. J. Haematol. 1998, 101, 50–58. 34â•… A. Castegna, C. M. Lauderback, H. Mohmmad-Abdul, D. A. Butterfield, Brain Res. 2004, 1004, 193–197. 35â•… Y. Hamon, C. Broccardo, O. Chambenoit, M.-F. Luciani, F. Toti, S. Chaslin, J.-M. Freyssinet, P. F. Devaux, J. Neish, D. Marguet, G. Chimini, Nat. Cell Biol. 2000, 2, 399–406. 36â•… E. Farge, D. M. Ojcius, A. Subtil, A. DautryVarsat, Am. J. Physiol. Cell Physiol. 1999, 45, C725–C733. 37â•… E. Farge, P. Devaux, Biophys. J. 1992, 61, 347–357. 38â•… M. Seigneuret, A. Zachowski, A. Herrmann, P. F. Devaux, Biochemistry 1984, 23, 4271–4275. 39â•… G. Morrot, S. Cribier, P. F. Devaux, D. Geldwerth, J. Davoust, J. F. Bureau, P. Fellmann, P. Herve, B. Frilley, Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 6863–6867. 40â•… Suzuki et al., Nature 2010, 468, 834–838.

LIST OF CONTRIBUTORS

Krishnakumar Balasubramanian, University of Pittsburgh, Pittsburgh, PA 15219; Email: [email protected] Edouard M. Bevers, Department of Biochemistry, Cardiovascular Research Institute Maastricht, Maastricht University, The Netherlands; Email: [email protected] Adam Blanchard, School of Veterinary Medicine and Science, University of Nottingham, Sutton Bonington Campus, College Road, Sutton Bonington, Leicestershire LE12 5RD, UK Richard Callaghan, Nuffield Department of Clinical Laboratory Sciences, John Radcliffe Hospital, University of Oxford, Oxford, UK; Email: [email protected] Giovanna Chimini, Centre d’Immunologie de Marseille-Luminy, INSERMCNRS-Université de La Méditerranée, Parc Scientifique de Luminy, 13288, Marseille, France; Email: [email protected] Shelley M. Cook, Department of Biochemistry and Molecular Biology, Medical Sciences, Bloomington, Indiana University School of Medicine, Bloomington, IN 47405 David L. Daleke, Department of Biochemistry and Molecular Biology, Medical Sciences, Bloomington, Indiana University School of Medicine, Bloomington, IN 47405; Email: [email protected] Philippe F. Devaux, Institut de Biologie Physico-Chimique, 13 rue Pierre et Marie Curie 75005 Paris , France; Email: [email protected] xxiii

xxiv

LIST OF CONTRIBUTORS

Jean-Marie Freyssinet, U. 770 INSERM, Hôpital de Bicêtre, Le KremlinBicêtre, France; Faculté de Médecine, Université Paris-Sud, Le KremlinBicêtre, France; and Faculté de Médecine, Institut d’Hématologie & Immunologie, Université de Strasbourg, Strasbourg, France; Email: [email protected] Todd R. Graham, Department of Biological Sciences, Vanderbilt University, Nashville, TN 37235; Email: [email protected] Per Haberkant, EMBL, Heidelberg, Germany Andreas Herrmann, Department of Biology, Humboldt-University Berlin, Invalidenstr. 42, D-10115 Berlin, Germany; Email: andreas.herrmann@ rz.hu-berlin.de Joost C.M. Holthuis, Department of Membrane Enzymology, Bijvoet Center and Institute of Biomembranes, Utrecht University, 3584 CH Utrecht, The Netherlands; Email: [email protected] Frans A. Kuypers, Children’s Hospital Oakland Research Institute, 5700 Martin Luther King Way, Oakland, CA 94609; Email: [email protected] Ke Liu, NIH Chemical Genomics Center, Bethesda, MD 20892 Iván López-Montero, Departamento de Química Física I, Universidad Complutense de Madrid, 28040 Madrid, Spain; Email: [email protected] Anant K. Menon, Department of Biochemistry, Weill Cornell Medical College, New York, NY 10065 Peter Müller, Department of Biology, Humboldt-University Berlin, InvaliÂ� denstr. 42, D-10115 Berlin, Germany; Email: [email protected] Baby-Periyanayaki Muthusamy, Department of Biological Sciences,Vanderbilt University, Nashville, TN 37235 Paramasivam Natarajan, Department of Biological Sciences, Vanderbilt University, Nashville, TN 37235 Sylvia Neumann, Department of Cell Biology, The Scripps Research Institute, La Jolla, CA Petra H.M. Niesten, Nuffield Department of Clinical Laboratory Sciences, John Radcliffe Hospital, University of Oxford, Oxford, UK Nina Ohlwein, Department of Biology, Humboldt University Berlin, Invalidenstr. 42, D-10115 Berlin, Germany; Email: [email protected] Anna Pia Plazzo, Department of Biology, Humboldt-University Berlin, Invalidenstr. 42, D-10115 Berlin, Germany Naomi L. Pollock, Nuffield Department of Clinical Laboratory Sciences, John Radcliffe Hospital, University of Oxford, Oxford, UK

LIST OF CONTRIBUTORS

xxv

Thomas G. Pomorski, Department of Plant Biology and Biotechnology, Faculty of Life Sciences, University of Copenhagen, DK-1871 Frederiksberg C, Denmark Cyril Rauch, School of Veterinary Medicine and Science, University of Nottingham, Sutton Bonington Campus, College Road, Sutton Bonington, Leicestershire LE12 5RD, UK; Email: [email protected] Alan J. Schroit, Department of Pharmacology, The University of Texas Southwestern Medical Center, Dallas, TX 75390; Email: alan.schroit@ utsouthwestern.edu Eric Soupene, Children’s Hospital Oakland Research Institute, 5700 Martin Luther King Way, Oakland, CA 94609 Florence Toti, U. 770 INSERM, Hôpital de Bicêtre, Le Kremlin-Bicêtre, France; Faculté de Médecine, Université Paris-Sud, Le Kremlin-Bicêtre, France; and Faculté de Médecine, Institut d’Hématologie & Immunologie, Université de Strasbourg, Strasbourg, France Gerrit van Meer, Faculty of Science, Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands; Email: [email protected] Marisela Vélez, Instituto de Catálisis y Petroleoquímica, Consejo Superior de Investigaciones Científicas, 28049 Madrid, Spain; and IMDEA Nanociencia, Facultad de Ciencias, Universidad Autonóma de Madrid, 28049 Madrid, Spain; Patricia M. Verhulst, Department of Membrane Enzymology, Bijvoet Center and Institute of Biomembranes, Utrecht University, 3584 CH Utrecht, The Netherlands Patrick L. Williamson, Department of Biology, Amherst College, Amherst, MA; Email: [email protected] Alain Zachowski, Laboratory of “Physiologie Cellulaire et Moléculaire des Plantes,” Université Pierre et Marie Curie—Paris 6 (UR 5) and Centre National de la Recherche Scientifique (EAC 7180); Email: alain.zachowski@ upmc.fr Ana Zarubica, Centre d’Immunologie de Marseille-Luminy, INSERMCNRS-Université de La Méditerranée, Parc Scientifique de Luminy, 13288, Marseille, France Xiaoming Zhou, Department of Biological Sciences, Vanderbilt University, Nashville, TN 37235

PART I ASSESSING TRANSMEMBRANE MOVEMENT AND ASYMMETRY OF LIPIDS

1 METHODS FOR THE DETERMINATION OF LIPID TRANSMEMBRANE DISTRIBUTION AND MOVEMENT IN BIOLOGICAL MEMBRANES Philippe F. Devaux Institut de Biologie Physico-Chimique, Paris, France

Andreas Herrmann Department of Biology, Humboldt-University Berlin, Berlin, Germany

1.1â•… INTRODUCTION To access the transbilayer distribution and movement of lipids and, perhaps in particular, to convince all readers that lipids really are distributed asymmetrically between the two leaflets of eukaryotic cell plasma membranes and to explain how lipid transporters were discovered in biomembranes, it is necessary to give an overview of the main techniques that were and—in many cases—are still used. Our objective here is in fact limited to give solely an overview as well as an indication of the limits of the techniques that have been used during the last 40 years. In this book, it is not possible to detail the various aspects and weakness of each technique. Details can be found in specialized publications, but also in various chapters of this book, to which

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

3

4

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

the readers will be referred. These various techniques, because they have given not always identical but still very consistent results, have allowed some definitive conclusions to be drawn. We will respond to criticisms against the use of lipid analogs carrying a reporter moiety which “could introduce artifacts.” We are well aware, and will outline below, that quantitative but rarely qualitative differences with respect to the behavior of endogenous lipids can be caused by reporter moieties covalently attached to lipids. However, even assays characterizing the transbilayer organization of endogenous lipids can cause modifications of the membrane no longer comparable to the unperturbed, original situation. Nevertheless, looking back over decades of research in this field, analogs in conjunction with assays that may affect membrane properties have provided milestones in understanding the dynamics of transbilayer distribution of lipids. For example, by using spin-labeled phospholipids, two major discoveries in this field were made. First, early in the 1970s, Kornberg and McConnell were able to give for the first time quantitative data on kinetics of passive transbilayer movement (flip-flop) of phospholipids in a bilayer [1]. Second, in 1984, Seigneuret and Devaux discovered by using short-chain spin-labeled lipids that the inward translocation of aminophospholipids in red blood cell membranes is ATP dependent, pointing to an energy-dependent lipid transporter that may also be typical for the plasma membrane of other mammalian or even all eukaryotic cells [2]. Indeed, subsequently similar conclusions of an ATP-dependent inward translocation of aminophospholipids were obtained from studies with fluorescent [3–5] as well as nonlabeled short-chain exogenous [6, 7] or even radioactive long-chain lipids [8]. Hence, being aware that labeling of lipids by reporter moieties affects their properties, lipid analogs provided significant insights into the transbilayer movement and distribution of lipids. Nevertheless, independent complementary methods, in particular those based on endogenous lipids, or at least long-chain lipids with nonperturbing labels (e.g., radioactive), are not only desirable but even mandatory to reach confident conclusions. However, applications of labeled lipid analogs are typically easier to perform as techniques relying on natural lipids. Hence, starting with lipid analogs to address questions on transbilayer lipid organization might efficiently pave the way to apply or even to develop subsequent techniques based on endogenous lipids. 1.2â•… DEVELOPMENT OF ASSAYS FOR DISTRIBUTION AND TRANSLOCATION OF LIPIDS ACROSS MEMBRANES Discoveries of essential aspects of transbilayer lipid organization, in particular of plasma membranes, have been driven by the development of new methods. Likewise, questions that could not be solved with the available repertoire of methods initiated new methods. Early studies on transbilayer organization of lipids addressed whether phospholipids are asymmetrically distributed across

DEVELOPMENT OF ASSAYS AND TRANSLOCATION OF LIPIDS

5

the membrane. Those studies investigated essentially the transbilayer distribution of endogenous lipids in the plasma membrane of red blood cells, for example, by phospholipases or by chemical modifications of lipids (see below). They revealed that aminophospholipids phosphatidylserine (PS) and phosphatidylethanolamine (PE) are preferentially localized on the cytoplasmic leaflet of red blood cells, while phosphatidylcholine (PC) and sphingomyelin (SM) are predominantly on the exoplasmic leaflet (see reviews by Op den Kamp [9, 10]). In particular, PS, contributing to about 10% of the total phospholipid content of the plasma membrane, was almost exclusively shielded from the external leaflet. These studies also implicated that the asymmetric phospholipid distribution in the plasma membrane is typical for mammalian cells. Of course, this unique distribution immediately raised the question about the molecular mechanism not only generating but also preserving the asymmetric distribution, for example, in the case of the red blood cell circulating for about 120 days in human blood vessels. So, the question was, can lipid asymmetry be created by spontaneous segregation of lipids between the two leaflets of a membrane? Lateral segregation of lipids in domains, for example, in cholesterol- and sphingolipid-enriched domains, so-called rafts [11–14], is a thermodynamic phase separation. In contrast, transversal segregation of lipids does not exist spontaneously in a pure lipid bilayer. However, several factors may trigger such segregation. Membrane bending can cause a spontaneous segregation determined by the size of the polar head group of the lipids, which are mixed initially. In fact, a topological asymmetry exists due to membrane curvature. If one mixes lipids with a large head group and lipids with a small head group and prepares small unilamellar vesicles by sonification, a segregation of lipids is generated due to the large membrane curvature; the preference of lipids with a large head group is to occupy the external side where the curvature allows more space (see the review by Op den Kamp [10]). Other factors spontaneously generating lipid asymmetry are feasible in biological membranes with an asymmetrical environment on each side of a membrane.Transmembrane potential generally creates an electric field that polarizes each side of a membrane. Membrane proteins are not structurally symmetrically organized across the membrane. ATPases of the plasma membrane, for example, have charged residues usually on the cytoplasmic leaflet where ATP binds. Other proteins bind on the external surface or on the cytoplasmic monolayer where the cytoskeleton is attached. One could speculate that positively charged residues of those proteins may interact preferentially with lipids carrying head groups of opposite charges such as PS, or negatively charged amino acid residues with PS via the divalent calcium ion, eventually giving rise to an asymmetric transbilayer arrangement of these lipids. This is how the asymmetrical organization of lipids was initially explained with PS being trapped on the cytoplasmic leaflet of the red blood cell plasma membrane by its negatively charged head group interacting with the cytoskeleton proteins [9, 10]. However, this concept left many questions unanswered.

6

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

A new twist in understanding the generation and maintenance of lipid asymmetry in the plasma membrane of red blood cells (and other mammalian cells) was provided by the introduction of spin-labeled and fluorescent lipid analogs. When spin-labeled aminophospholipids were inserted into the exoplasmic leaflet of human red blood cells, surprisingly, they rapidly redistributed to the cytoplasmic leaflet [2, 15]. PS analogs disappeared almost completely from the external leaflet within a few minutes at 37°C, while PE reached a stationary distribution with about 80% of the analog on the inner leaflet within about 40 minutes [2]. In contrast, spin-labeled PC and SM moved only very slowly to the cytoplasmic monolayer and remained essentially on the external side. These studies provided two exciting results: a more technical one and a heuristic one. First, the stationary distribution of spin-labeled phospholipids was very similar to that of endogenous phospholipids already known, demonstrating that lipid analogs could qualitatively mimic their endogenous counterparts. Second, there is an energy-dependent transport of specific phospholipids in the plasma membrane. The directed and fast inward redistribution of aminophospholipids immediately provided an explanation for the generation and maintenance of lipid asymmetry in the plasma membrane of human red blood cells and—as implicated by later studies—typically for mammalian cells. In contrast to the model explaining lipid asymmetry by a specific interaction of lipids with the cytoskeleton, the finding of a directed transport could explain how a cell can rapidly repair or readjust any perturbation of lipid asymmetry caused by, for example, endo- and exocytotic processes. Moreover, as mentioned above, lipid asymmetry has to remain during the lifetime of the cells, which lasts sometimes more than several days. The transverse diffusion of lipids or lipid flip-flop, even slow, should lead finally to an equilibration of the lipid distribution between the two monolayers in the absence of a mechanism of lipid distribution repair, which in practice is carried out by a transporter and a flippase protein, and requires ATP as a source of energy. These and other results [16] obtained with lipid analogs also made clear why assays based on chemical labeling or phospholipase treatment of endogenous lipids in the way they were performed could not recognize the rapid inward motion of aminophospholipids. Comparing the time required to perform the assays with the characteristic time of inward and outward motion of phospholipids, the experimental approaches would not have been able to detect the rapid inward transport of aminophospholipids. The observation of an energy-consuming transport of lipids triggered exciting but very challenging research on the identification and characterization of lipid transporters. Although many important details still have to be unraveled, today, we know that for the rapid inward transport of PS and PE in the plasma membrane of eukaryotic cells, a P-Type ATPase is responsible (see in particular Chapters 8–10). Meanwhile, several members of the ATP-binding cassette (ABC) transporters have been identified also to mediate transport of lipids at the expense of ATP, not only phospholipids but also other lipids, for example,

OVERVIEW ON ASSAYS DISTRIBUTION OF LIPIDS

7

sterols (see Chapters 11 and 17). It is important to mention that there are other functions of lipid transport besides generating/preserving lipid asymmetry. Lipids can be secreted to the extracellular space via transport across the membrane (see Chapters 5, 11, and 17). Another function of lipid transporter could be to generate bending of the membrane by creating a difference of the surface area between the two monolayers of the membrane (see Chapters 2, 12, and 13). Finally, essentially based on the use of spin-labeled and fluorescent analogs, the study of transbilayer motion of lipids showed that a fast, rather unspecific rapid transbilayer scrambling of phospholipids in the plasma membrane of mammalian cells can be triggered by activation of a putative scramblase (see Chapter 7). In biogenic membranes, flip-flop of lipids is typically mediated very fast by membrane proteins (flippases) mediating an energyindependent and nonspecific redistribution of lipids (see Chapter 6). 1.3â•… OVERVIEW ON ASSAYS FOR MEASURING DISTRIBUTION AND TRANSLOCATION OF LIPIDS ACROSS CELLULAR MEMBRANES Several prerequisites have to be matched by an assay to generate credible results. Of course, each assay has its limits, and the choice of an assay always depends on the information in which one is interested. That is, while an assay might be useful to detect the transbilayer distribution of endogenous lipids or lipid analogs in a cellular membrane, it may not be useful to measure the kinetics of transbilayer motion of lipids. Several reports and reviews have considered in detail which criteria have to be fulfilled in order to determine the transbilayer distribution and movement of (phospho)lipids [17–22] (see also Chapters 5 and 6). Here, we will only shortly summarize the criteria: (1) The assay has to recognize the lipid of interest on the surface of the membrane in a quantitative manner. (2) The approach must distinguish between lipid species located on one leaflet and those on the opposite leaflet of the membrane. Hence, the recognizing reagent or enzyme must not have access to both sides of the membrane; that is, it must be impermeable. (3) Exchange/ redistribution of lipids between both monolayers should not occur while the assay is performed. That is, the time required to assay a specific lipid quantitatively in a leaflet must be shorter in comparison to the characteristic time of lipid transbilayer movements. (4) The amount of lipid of interest should not change during the assay, for example, due to delivery of new lipids or removal of (already modified by the assay) lipids via endocytosis and exocytosis, respectively. (5) Treatment of lipids during the assay should not modify membranes leading to a perturbation and, hence, an enhanced transbilayer redistribution of lipids during assay. In Figure 1.1, most relevant assays for assessing transbilayer movement and/ or distribution of lipid analogs as well as of endogenous lipids are summarized schematically. We will now briefly describe the various assays, focusing first on

8

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

(a)

(c)

(b.1)

(d.1) Enzyme

Ext Red t=0

(b.2)

t>0

(d.2) Enzyme III.

1.

Ext 2.

II.

5.

3.

Red

3. 4.

I.

Figure 1.1.â•‡ Assays for the detection of lipid transbilayer distribution (modified from Pohl et al. [81]). (a) Chemical modification assays for endogenous lipids. Endogenous lipids present on the outer plasma membrane leaflet are typically modified on the level of the head group. Reagents frequently used for modification are trinitrobenzene sulfonic acid (TNBS, specific for PE) and fluorescamine. Assays can also be applied to endogenous lipids synthesized in the presence of radioactive precursors in the cell (see b.2). (b) Enzymatic assays for endogenous lipids. (b.1) Phospholipase A2 (enzyme) treatment converts phospholipids in the outer plasma membrane leaflet to lysolipid and fatty acid. Lipid products are then analyzed by chromatography and can be compared with samples untreated with enzyme. An analogous technique is used for SM, employing sphingomyelinase. (b.2) Enzymatic assays have also been applied to assess transbilayer organization of endogenous lipids that are synthesized in the presence of radioactive precursors (marked with asterisks) in the cell (I) and localize to various cellular membranes, that is, also to the cytoplasmic leaflet of the plasma membrane, due to vesicular or monomeric transport (II), and move to the extracellular leaflet of the plasma membrane, for example, due the presence of transporter proteins (III). Upon appearance on the outer membrane surface, lipids are converted by enzyme treatment (enzyme) (see b.1).

TRANSBILAYER DISTRIBUTION OF ENDOGENOUS LIPIDS

9

Figure 1.1.â•‡ (Caption Continued) (c) Antibody-, peptide-, or protein-binding assay for endogenous lipids. Specific antibodies, peptides (e.g., Ro09-198, binding to PE), or proteins (e.g., Annexin V, binding to PS) with a high affinity for a particular lipid head group bind to endogenous lipids present on the outer plasma membrane leaflet. The amount of bound antibody/peptide/protein is quantified. (d) Albumin-extraction, dithionite, and ascorbate assays for fluorescent and spin-labeled lipid analogs. (d.1) At time tâ•¯=â•¯0, the outer (accessible) leaflet of the membrane is labeled with short-chain spinlabeled or fluorescent analogs. Analogs redistribute to the inner leaflet of the membrane leaflet by passive flip-flop or active transport (tâ•¯>â•¯0). To assess transbilayer distribution, analogs are extracted from the outer leaflet (Ext), for example, phospholipids by albumin or cholesterol by methyl-β-cyclodextrin, followed by separation of cells and media. By comparing the extracted amount of analogs with that of analogs remaining in the membrane, the transbilayer distribution can be estimated. If aliquots of the sample are investigated at different time points after labeling, the transbilayer movement of the analog can be assessed. Alternatively to albumin extraction, fluorescence of lipid analogs on the outer leaflet can be quenched using dithionite, or the spin-label signal can be reduced using ascorbate (Red). (d.2) The short-chain lipid analog precursor integrates into the outer membrane leaflet (1), crosses the plasma membrane (e.g., by passive flip-flop) (2), and distributes to different intracellular membranes (e.g., by monomeric transport) (3). Enzymes of the endoplasmic reticulum or Golgi convert part of the lipid analog precursor to the lipid analog of interest (4), which can distribute back to the cytoplasmic leaflet of the plasma membrane, where it becomes available to outward transport by transporter proteins (5). Upon appearance on the outer leaflet, lipid analog is extracted (Ext) or reduced (Red) (see d.1). Color version on the Wiley web site.

endogenous lipids and subsequently on lipid analogs. Note that assays based on exogenous but nonlabeled lipids are introduced in Chapter 2. 1.4â•… MAIN TECHNIQUES USED TO DETERMINE TRANSBILAYER DISTRIBUTION OF ENDOGENOUS LIPIDS IN CELL MEMBRANES 1.4.1â•… Chemical Labeling The first technique used to investigate the localization of phospholipids in the plasma membrane of a eukaryotic cell, the human red cell membrane, was carried out by Mark Bretscher in England in the early 1970s [23]. It was based on the chemical labeling of aminophospholipids by an NH2 reactive reagent (Fig. 1.1a; see also Chapter 6). Typical NH2 reactive reagents are trinitrobenzene sulfonic acid (TNBS) and fluorescamine. To match the condition that reagents do not permeate to the opposite membrane leaflet, those experiments are usually performed at lower temperatures, between 4 and 10°C. Bretscher demonstrated that PS and PE could react only if the cell membrane was made permeable to those essentially nonpermeable reagents. He was the first to claim and to demonstrate the asymmetrical organization of lipids in the human red cell membrane with the aminophospholipids PS and PE located principally

10

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

in the inner leaflet of the plasma membrane, while the choline-containing phospholipids PC and SM were principally in the outer monolayer. The bilayer concept had already been suggested a long time before by Gorter and Grendel in 1925 [24], who came to that conclusion by an evaluation of the area covered by the erythrocyte lipids. However, they could not infer the lipid asymmetry because they assumed that all lipids were identical and corresponded to phosphatidylglycerol (PG). 1.4.2â•… Enzymatic Treatment Almost at the same time when Bretscher published his results with chemical labeling of aminophospholipids, the Dutch group of Laurence Van Deenen in Utrecht [25, 26] developed a completely different technique to study the transmembrane distribution of phospholipids in the plasma membrane of eukaryotes, based on lipid degradation by phospholipases and sphingomyelinases (Fig. 1.1b; see also Chapter 6). For example, phospholipase A2 treatment by addition of the enzyme to cells converts phospholipids in the outer plasma membrane leaflet to lysolipids and fatty acids. This technique seems a priori hazardous since there may be slow destruction of the membrane under investigation; analysis of the lipid composition in the time course of the assay proves that the external leaflet was indeed perturbed. Yet the results found by this assay are similar to those obtained by chemical labeling of lipids [23], namely, a preferential location of aminophospholipids in the inner leaflet of the plasma membrane. A similar distribution was also found afterward in other cells of the blood circulation such as platelets and, in fact, in all eukaryotic cells investigated with similar techniques. Small variations could be reported for different animals such as ruminants, which have essentially no PC (but a larger fraction of SM). More data corresponding to different eukaryotic cells including plant cells can be found in Chapter 3 of this book (see Table 3.1). Another, very specific enzymatic treatment of phospholipid head group is the use of a decarboxylase, which transforms PS into PE [27, 28]. These approaches can also be applied to radioactively labeled phospholipids, allowing a more sensitive quantification of reaction products. It is very reasonable that radioactive labeling does not affect the behavior of lipids with respect to endogenous, nonlabeled lipids. Labeling of membranes with radioactively labeled phospholipids with two long fatty acid chains can be achieved by lipid transfer proteins [29–31]. However, labeling of the membrane by this technique is a rather slow process. While the approach should be useful to study the stationary distribution of lipids across the membrane, it might be difficult to resolve fast transbilayer movement of lipids. An interesting variant is the measurement of cell surface exposure of lipids that have been labeled radioactively intracellularly. Kälin et al. incubated red blood cells with [14C]-labeled fatty acids, which—after uptake—were incorporated into lyso-PC via two enzymatic steps [32] (Fig. 1.1b). Subsequently,

TRANSBILAYER DISTRIBUTION OF ENDOGENOUS LIPIDS

11

exposure of [14C]PC on the cell surface was assessed by hydrolysis via phospholipase A2 (see above). Results were in agreement with a complementary assay based on the exchange of surface-exposed PC (including [14C]PC) on PC from added liposomes by a lipid transfer protein [32]. The enzymatic approach has also been applied to lipids other than phospholipids. The transbilayer distribution of cholesterol was assessed using cholesterol oxidase [33–35] (see Chapter 5). Several enzymes and chemical modifications are available to specifically detect the transbilayer organization of glycosphingolipids (see Reference 20; Chapter 4). 1.4.3â•… Protein-Binding Assay for Endogenous Lipids Peptides or proteins that bind noncovalently to phospholipids by recognizing specific parts of their head groups provide noninvasive approaches to assess exposure of endogenous lipids on the membrane surface without modification of lipids (Fig. 1.1c). In particular, fluorescent variants of these molecules allow detection of binding by fluorescence microscopy and—as very often used—by flow cytometry analysis in a rather easy way. The most prominent example is the specific binding of Annexin V to PS, which was first employed by Thiagarajan and Tait [36]. Using Annexin V, they could demonstrate the exposure of endogenous PS on the cell surface of activated platelets (see also Chapter 7). Since then, it has found numerous applications (see Chapters 7, 15, and 16). For example, Annexin V binding is used for the detection of PS exposure as an early event of cell apoptosis [37, 38] (see Chapter 15). While binding of Annexin V to PS requires the presence of Ca2+, the cell adhesion glycoprotein lactadherin recognizes PS on cell surfaces without this cation [39, 40]. Lactadherin, also known as milk fat globuleepidermal growth factor (EGF) 8 (MFG-E8), is secreted by macrophages for mediating engulfment of PS-expressing apoptotic cells. Since the activity of prothrombinase essential for blood coagulation is dependent on binding to PS on the cell surface of thrombocytes, this has been used to assay exposure of PS on the membranes [41, 42]. However, interpretation of protein binding to membranes must always consider affinity to other lipids, for example, in case of Annexin V to other negatively charged lipids. Cinnamycin Ro09-0198 is a cyclic peptide that has been isolated from Streptoverticillium griseoverticillatum. This peptide specifically recognizes the head group of PE [43] and forms an equimolar complex with PE on biological membranes. To use the peptide as a probe for analyzing the surface exposure of PE, the peptide has been fluorescently labeled, preserving its reactivity and specificity [44]. For example, it has been applied to assess the exposure of PE on the cell surface of yeast cells [45]. However, one has to be aware that this peptide is able to induce transbilayer lipid movement, as has been shown for HeLa cells as well as model membranes [46]. To initiate flip-flop by Ro09-0198, the presence of PE is required.

12

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

1.5â•… MAIN TECHNIQUES USED TO DETERMINE TRANSBILAYER DISTRIBUTION OF LIPID ANALOGS IN CELL MEMBRANES 1.5.1â•… Spin-Labeled and Fluorescent Lipid Analogs 1.5.1.1 Early Studies Using Spin-Labeled Analogs The spin-labeling technique was invented by Harden McConnell, a professor of chemistry in Stanford University, in the early 1970s. (Roger Kornberg, one of his students, went on to win a Nobel Prize in 2006, on a very different subject.) In 1971 Kornberg and McConnell published a paper [1] on the slow spontaneous lipid diffusion from one leaflet to the opposite (flip-flop) in sonicated phospholipids vesicles using a modified PC (dipalmitoyl phosphatidylcholine [DPPC]), which contains in the head group a nitroxide ring (TEMPO) replacing a single CH2. The transbilayer redistribution was measured by selective reduction of spin-labeled lipids on the outer leaflet (see below). The half-time of diffusion measured with this probe was of the order of 6 hours at 30°C [1]. Because there was no value known at the time, this number was very important to obtain. However, one could guess that TEMPO is likely to slow down the transmembrane diffusion of the lipid analog because of the size of the paramagnetic moiety and of the polar character of the probe itself. It was only in 2005 that Liu and Conboy (see below) proved by sum frequency vibrational spectroscopy (SFVS) that the TEMPO–DPPC flip-flop is indeed one order of magnitude slower compared with pure DPPC [47]. The SFVS technique is rather sophisticated but has the advantage of measuring the transmembrane diffusion of nonlabeled lipids (if one admits that deuterated lipids are indeed perfect representatives of natural lipids). The value measured by Liu and Conboy is certainly important to know, but the information obtained by the McConnell laboratory in 1971 (34 years before!) was nevertheless an extremely useful hallmark. 1.5.1.2 Spin-Labeled and Fluorescent Analogs with a Short Fatty Acid Chain The spin-labeled TEMPO–DPPC used by Kornberg and McConnell is not soluble in water because of the long chains and must be added before vesicles are formed, for example, by sonication. To label biological membranes, they have to be fused with those vesicles, or they have to be incorporated with phospholipid transfer proteins. This is a limitation that hampers the use of such probe with natural membranes even with plasma membranes like erythrocyte cell membrane. To overcome this problem, in the early 1980s, fluorescent and spin-labeled lipid analogs, in particular, phospholipid analogs, were developed, which are slightly water soluble due to a short fatty acid chain replacing one of the natural long chains—typically in the sn2 position (Fig. 1.2) [2, 15, 48, 49]. The short chain usually has 5 or 6 or sometimes up to 12 carbon atoms. They actually form in water micelles and monomers, which is an essential property for efficient and rapid labeling of intact/preformed membranes. A mere addition to a suspension or monolayer of cells is sufficient to label the plasma mem-

13

TRANSBILAYER DISTRIBUTION OF LIPID ANALOGS (a)

(b)

H

H

O

P

P O

N

O O

O

O

R-N

NBD

N

N O

(c)

(d)

N

+

N

O O P – O O

+

O O P – O

O

O

O

O

O

O O

O

O

+

–

N -O

O N N O

+

N O N – O

Figure 1.2.â•‡ Structure of short-chain spin-labeled (a,c) and fluorescent (b,d) phospholipid analogs. Principal structure of spin-labeled (a) and fluorescent (b, NBD moiety) analogs. Chemical structures of spin-labeled (c) and NBD-labeled (d) phosphatidylcholine analogs. The label moiety is attached to the sn2 chain. P, phosphate group; H, head group. Color version on the Wiley web site.

brane because the monomers will incorporate spontaneously. This process could be very rapid. For example, spin-labeled analogs insert within a few seconds into the outer layer of the plasma membrane [18, 19]. As a dynamic equilibrium between micelles and monomers exists, insertion of monomers into the membrane leads to a depletion and finally disappearance of micelles and monomers from the medium. Since the amount of analogs used corresponds typically to 1â•›molâ•›% or even less of endogenous lipids, in many cases,

14

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

almost all analogs incorporate into the membrane. However, this has always to be checked carefully. Of course, analogs are also very useful to assess the transbilayer distribution of organelles and reconstituted systems (see, e.g., Chapters 6, 8, and 9). The reporter moiety should be in principle as small as possible to avoid or minimize steric perturbations. Originally, most studies based on fluorescent analogs used lipids with a short chain of six carbon atoms to which terminally the fluorescent group 7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD) is attached (C6-NBD analogs). Later on, C5-BODIPY analogs were introduced, which have several advantages such as the BODIPY moiety is integrated into the lipid structure along the fatty acid chain, they are more apolar, and they have better fluorescence properties [50]. It has to be underlined that not only the fluorescent moieties but also the spin-labeled groups are of polar nature that may perturb the normal hydrophobic environment of a lipid chain. As a consequence, a looping of the fatty acid chain to which the reporter moiety (NBD) is attached (usually the short chain) to the polar interface of the membrane is facilitated [51, 52]. The possibility to attach the probe on the polar head group exists, but it raises a new problem: The analog may not be recognized by a membrane protein that is supposed to interact selectively with a lipid. In particular, a lipid has to be recognized and transported by a lipid transporter (flippase), which is usually very specific for one or a few substrates. So in practice, the reporter moieties, in particular, fluorescent groups that are generally rather big, can cause artifacts. In other words, a test of biological activity has to be done before using those probes to verify that the degree of perturbation, which is introduced by the probe, is acceptable. Alternatively, observations obtained by using lipid analogs should be verified by assessing the behavior of endogenous lipids. Indeed, there are multiple examples proving that, in particular, fluorescent probe can cause serious perturbations. Examples of artifacts due to size and polarity of fluorescent probes are given in the articles by Devaux and collaborators [22, 53]. Another important issue of those analogs is their short-chain fatty acids. Early studies have shown that the fatty acid chain may have a significant influence on lipid translocation. In 1986, Middelkoop and collaborators showed that by exposing red cell membranes to exogenous phospholipases A2, phospholipids with at least one unsaturated chain experience a more rapid flip-flop than saturated lipids [54]. Half-times varied from 26.3 to 2.9 hours for 1,2 dipalmitoyl-PC and 1-palmitoyl-2-linenoyl-PC, respectively. Recently, various cholesterol analogs have been studied showing large variations in the potential to mimic endogenous cholesterol [55] (for structure of spin-labeled and fluorescent sterol/cholesterol analogs, see Chapter 5). 1.5.1.3 Assessing Transbilayer Distribution and Movement of Spin-Labeled and Fluorescent Lipid Analogs In principle, there are two ways to assess the transbilayer distribution and movement of lipid analogs, either by chemical reduction of the label moieties on the outer/accessible leaflet of the membrane

TRANSBILAYER DISTRIBUTION OF LIPID ANALOGS

15

by a nonpermeant agent or by extraction of analogs from the outer leaflet (Fig. 1.1d). In any case, treatment of analogs must be leaflet specific. The transbilayer distribution can be estimated from the amount of reduced/extracted analogs, that is, analogs of the outer/accessible leaflet, and the amount of nonmodified analogs, that is, analogs on the inner/nonaccessible leaflet. However, one has always to control the amount of total analogs in the membrane under investigation during the assay procedure, which in the ideal case should not alter (see below). To assess the kinetics of transbilayer distribution, aliquots of the labeled sample will be measured after different times of labeling. 1.5.1.3.1â•… Reduction Assayâ•… Spin labels and fluorescent NBD analogs are typically reduced by addition of ascorbate [1, 22] and dithionite [56–58], respectively (Fig. 1.1d). Both agents are, in principle, nonpermeable to membranes. Addition of the reducing agent to the suspension medium destroys the reporter moiety and, hence, the signal of labeled probes present on the outer monolayer. The remaining signal comes from probes that have flipped to the inner monolayer, which can be determined by electron paramagnetic resonance (EPR) or fluorescence spectroscopy. This allows one to determine the percentage of probes at time t that were exposed to the outer and inner monolayer. However, in case of biological membranes, and in particular at a higher temperature (37°C), reducing agents may cross the membrane reacting also with analogs on the intracellular leaflet. If so, reduction has to be performed at a low temperature, for example, 4°C. In any case, the nonpermeability of agents has to be controlled carefully (for more details, see Chapter 6). By measuring aliquots of the labeled sample at different time points after labeling the membrane, the transbilayer movement of lipid analogs can be assessed provided that the movement is significantly smaller in comparison to the time required to perform the assay, that is, reduction of analogs on the outer leaflet (for critical discussion, see Chapters 5 and 6). In principle, those assays are also applicable to analogs with two long fatty acid chains as long as the reporter moiety is accessible to reducing agent. This depends essentially on the localization of reporter moiety. If the moiety is deeply buried into the hydrophobic phase, access to them by polar reducing agents such as ascorbate or dithionite is strongly impaired. Hence, reduction may take much longer and may even be incomplete. Reporter moieties attached to short-chain fatty acids are known to loop back to the membrane surface because of its (partial) polar character (see above) becoming easily accessible to reducing agents. 1.5.1.3.2â•… Back-Exchange Assayâ•… Fatty acids as well as short-chain phospholipids including labeled analogs can be easily extracted from the outer/ accessible membrane leaflet by albumin, typically bovine serum albumin (Fig. 1.1d), while phospholipids with two long fatty acid chains, as is typical for endogenous ones, are not removed from the membrane by albumin [16, 48, 59, 60]. Similar as described for the reduction assay, the transbilayer distribution of short-chain phospholipids analogs can be measured at a given time t after

16

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

labeling the membrane by incubation of the labeled sample, for example, of cells or organelles, with albumin for a short time (on the order of 1–2 minutes or even less [see Reference 17]). During this incubation, all analogs from the outer, albumin-accessible leaflet are extracted. Upon subsequent rapid centrifugation of the samples, the amount of analogs in the supernatant corresponding to albumin-extracted analogs and in the sediment containing the sample with the analogs on the inner, albumin nonaccessible leaflet can be measured again via EPR or fluorescence spectroscopy. If the amount of labeled lipid analogs initially incorporated into the outer leaflet (=â•¯100% of analogs) remained constant during the whole procedure, it is sufficient to measure at different time points only one fraction, typically that in the supernatant. However, as often observed for biological samples, this is not the case. Apart from the fact that the analogs can be metabolically converted or degraded into other (lipid) molecules (see below), the reporter moiety could be destroyed as well. In particular, spin labels are rapidly reduced by intracellular redox systems as gluthathione [15, 61]. As analogs modified on the inner leaflet can also redistribute back to the outer leaflet, these lipids with a destroyed reporter moiety may occur also on the outer leaflet and will not be detected by the spectroscopical measurement. However, as shown for spin labels, those signals can be recovered by addition of appropriate reagents [16, 60]. The back-exchange assay can also be applied to assess the kinetics of the transbilayer movement. As in any case, the time resolution depends on the time to carry out the assay procedures. Usually, incubation of a sample with albumin and subsequent centrifugation with standard laboratory equipment takes at least 2 minutes; that is, any transbilayer movement on the order of 2 minutes or even faster cannot be resolved adequately. Figure 1.3 shows the transbilayer movement of spin-labeled phospholipids in red blood cells [62]. Those measurements have been performed with standard laboratory equipment providing a time resolution even sufficient to resolve the rapid inward movement of the spin-labeled PS at 37°C with a half-time of about 5 minutes. However, the inward redistribution of spin-labeled PS is much faster in the plasma membrane of sperm cells or osteoblasts with a half-time of ≤2 minutes [63, 64] (for reviews, see References 21 and 65–69). Here, the assay based on standard equipment is on its limit of time resolution. In 1986, Tilley et al. measured the transbilayer redistribution of long-chain radiolabeled phospholipids inserted in the outer membrane leaflet of intact human erythrocytes with a nonspecific lipid transfer protein [8]. The transbilayer mobility and equilibrium distribution of the radiolabeled phospholipids were assessed by treatment of the cells with phospholipase A2. These experiments confirmed the selective ATP-dependent transport of aminophospholipids toward the inner membrane leaflet. Because probe insertion with a phospholipid exchange protein required at least 30 minutes incubation, and because cells and phospholipases also had to be incubated, no real kinetics could be drawn by this method. Nevertheless, partial kinetics data obtained with long-chain phospholipids using this technique [8] were consistent

17

TRANSBILAYER DISTRIBUTION OF LIPID ANALOGS 100

SL-PS SL-PE

SL analogs inside (%)

80

60

40 SL-PC 20 SL-SM 0 0

3

6

9

12

15

18

Time (hours)

Figure 1.3.â•‡ Kinetics of the redistribution of spin-labeled phospholipid analogs in human red cells at 37°C. Analogs were incorporated in the plasma membrane outer monolayer of those cells at tâ•¯=â•¯0 and their redistribution was followed by the backexchange assay. The final transbilayer distributions derived from the plateaus of the curves are in fact identical to the equilibrium distribution of endogenous phospholipids in those cells (see Chapters 3 and 10). Modified from Reference [62]. Color version on the Wiley web site.

with the more detailed results obtained with spin-labeled and fluorescent phospholipids. Buton et al. succeeded in optimizing the above-described albumin backexchange procedure to be performed in about 30 seconds [17]. Thus, they could resolve much better the fast flip-flop of spin-labeled analogs in organelles such as the endoplasmic reticulum or the Golgi. A much better time resolution of the back-exchange assay can be obtained by employing the stopped-flow technique [18, 19] (see Chapter 6). At this point, it might be interesting to compare spin-labeled and fluorescent analogs with respect to transbilayer distribution and movement. Again, although red blood cells are “simple cells,” experiments on them nicely illustrate the differences between analogs. The short-chain fluorescent C6-NBD-PS analog redistributes much slower from the exoplasmic to the cytoplasmic leaflet in comparison with the short-chain spin-labeled PS analog [53, 62]. Furthermore, the asymmetric distribution of C6-NBD-PS is less pronounced than that observed for spin-labeled PS (see above) and endogenous PS (see Chapter 3). Hence, although still transported via the aminophospholipid translocase, the NBD moiety affects the recognition and/or transport of the analog. Interestingly, while spin-labeled PE is efficiently transported to the cytoplasmic leaflet, C6-NBD-PE is almost not transported, indicating that the fluorescent analog is a very poor substrate for the aminophospholipid translocase

18

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

[53]. Very likely, the difference between both types of analogs can be explained by the NBD group being more bulky in comparison with the spin-label moiety. Similar differences between spin-labeled and fluorescent PS analogs were observed for the plasma membrane of other mammalian cells, for example, for fibroblasts [49, 70], sperm cells [63, 71], and hepatocytes [72, 73]. Back-exchange assay has also been performed by using phospholipid vesicles as a donor of phospholipid analogs instead of albumin [74–77]. Usually, studies have been done using fluorescent analogs. Exchange of the fluorescent analog under study between acceptor and donor membranes can be followed by Förster resonance energy transfer to a stably anchored fluorescent lipid in the donor vesicles (see also Chapter 6). While albumin is able to extract short-chain phospholipids and fatty acids, it does not extract cholesterol and sterol (analogs). However, methyl-βcyclodextrin (MβCD) can efficiently remove cholesterol from membranes. Indeed, extraction by MβCD has been used to assess the transbilayer dynamics of cholesterol/sterol analogs. As outlined in Chapter 5, several (additional) limitations of MβCD-mediated removal of cholesterol analogs have to be taken into account. For example, those analogs in many cases barely reflect the behavior of endogenous cholesterol. A major issue is that MβCD removes not only cholesterol analogs but also endogenous cholesterol. Hence, in this case, the composition of the membrane is continuously altered. For more detailed information, see Chapter 5. 1.5.1.3.3â•… Consequences of Intracellular Trafficking of Analogs for Assaysâ•… Initially, phospholipid analogs have been applied successfully to characterize transbilayer distribution and movement of lipids in the plasma membrane of human red blood cells. Although one has to take into account hydrolysis of analogs during experiment, the absence of any endo- and exocytotic activity and of intracellular membranes in red blood cells was very favorable for applied assays. Indeed, a drawback in assessing transbilayer distribution of spin-labeled or fluorescence analogs in the plasma membrane of eukaryotic cells is the removal of analogs by endocytosis and by intracellular redistribution of analogs to organelle membranes [49, 64, 70, 78]. Hence, the amount of analogs in the membrane of interest may not be constant during the assay procedure, violating a criterion given above. Figure 1.4 illustrates these disadvantages. In Figure 1.4a, the endocytic uptake of fluorescent C6-NBD-PC into osteoblasts at 37°C is shown [64]. As known for PC, movement from the exoto the cytoplasmic leaflet is slow. However, at the chosen temperature, endocytosis is a significant process, causing the disappearance of a significant amount if not most of the analog from the plasma membrane as visualized by the fluorescent intracellular spots corresponding to endosomal structures. In contrast, the fluorescent PS analog, which is transported rapidly from the exoto the cytoplasmic leaflet by the aminophospholipid translocase activity, disappears from the cytoplasmic leaflet and, thereby, from the plasma membrane to the cytoplasm due to the partial water solubility of analogs. Finally, it inserts

TRANSBILAYER DISTRIBUTION OF LIPID ANALOGS

19

(a)

(b)

Figure 1.4.â•‡ Uptake of fluorescent lipid analogs into osteoblasts. (a) Intracellular uptake of C6-NBD-PC occurs essentially via endocytosis. The yellow dots correspond to vesicles endocytosed; (b) C6-NBD-PS is transported by the aminophospholipid translocase to the ctyoplasmic leaflet of the plasma membrane. Once the probe is exposed to the cytosol, it can redistribute to other intracellular membranes because of its comparatively large solubility in water. See also Reference 64. Labeling was performed by incubation of analogs with cells. Upon incubation, noninserted analogs as well as analogs on the outer leaflet were removed by washing with albumin (see text).

into intracellular membranes, which leads to a bright intracellular staining (Fig. 1.4b). Due to the very rapid uptake of the PS analog by inward transport, the amount of PS taken up by endocytosis is low. In addition, biochemical modifications—apart from hydrolysis—can reduce the amount of lipid analogs in the membrane during the assay procedure (see Section 1.5.2).

20

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

1.5.2â•… Biosynthetic Labeling Nevertheless, the disadvantage of uptake and intracellular redistribution of (short-chain) analogs when assessing transbilayer organization of lipid analogs in the plasma membrane could be advantageous for biosynthetic labeling of lipids. The idea is that a labeled lipid analog, typically a fluorescent one, is taken up by the cell and, by metabolic processes, converted into another lipid analog (Fig. 1.1d). The trafficking of the latter, in particular, its exposure on the cell surface, is then followed, for example, by back-exchange. Thus, to examine outward transport of short-chain lipid analogs, for example, C6-NBDPC, by the multidrug resistance ABC transporter MDR1 Pgp expressed in the plasma membrane, cells were incubated with C6-NBD-PA. This lipid analog is partially converted into C6-NBD-diacylglycerol at the plasma membrane, which rapidly crosses the plasma membrane and becomes available for intracellular synthesis of not only C6-NBD-PC but also C6-NBD-PE [79–81]. The time-dependent appearance of the fluorescent phospholipids in the extracellular membrane surface was followed by incubating the cells with albumin and measuring the fluorescence of the supernatant after centrifugation. A similar approach has been applied to measure sphingolipid translocation. Short-chain analogs of ceramide, for example, C6-NBD-ceramide, were incorporated into the extracellular leaflet of the plasma membrane. These analogs flipped rapidly across the plasma membrane and—upon intracellular redistribution from the cytoplasmic leaflet—were converted to fluorescent SM and glucosylceramide (GlcCer) by enzymes of the Golgi apparatus. Exposure of these intracellularly synthesized fluorescent lipids on the cell surface was measured by the backexchange assay [20, 79, 82]. Those studies have shown that, for example, MDR1 Pgp is capable of transporting various short-chain GlcCer analogs to the extracellular leaflet of the plasma membrane, while the multidrug resistance transporter MRP1 was rather selectively transporting only C6-NBD-GlcCer [79, 82]. For a detailed description of transbilayer asymmetry and dynamics of glycolipids, see Chapter 4. 1.5.3â•… SFVS To circumvent labeling of lipids with bulky moieties, Liu and Conboy [83] introduced SFVS for measuring transbilayer lipid movement. This coherent nonlinear optical vibrational technique, which has been described in detail [84], takes advantage of the selectivity of infrared (IR) and Raman spectroscopy. Excitation induces dipole oscillations of molecules, for example, of C–H bonds. As molecules behave like an aharmonic oscillator, overtone oscillations are excited, including the second harmonic, which is used here. Since a secondorder nonlinear optical process is forbidden in media of inversion symmetry but not on surfaces with broken symmetry, this technique is surface specific in nature, which—under certain conditions (see below)—makes it also suitable for assessing transbilayer distribution and motion of lipids. Experimentally, the

References

21

excitation is read out by overlapping spatially and temporally a visible and a tunable IR laser on a surface (e.g., membrane surface). As a result, a signal (photon) is generated at the sum of their frequencies, which increases on resonance with a vibrational transition. Conboy and coworkers generated asymmetric lipid bilayers: One of the two monolayers contained phospholipids with a deuterated terminal CH3 group. Hence, the transition dipole moments of the C–H vibrational modes that are oriented antiparallel along the C–C bond are different between the two leaflets and do not cancel out, which can be monitored by SFVS. The difference between the transition dipole and therefore the signal becomes reduced and, finally, disappears upon transbilayer redistribution of phospholipids. Using this approach, the flip-flop of various phospholipids in bilayers of synthetic phospholipids has been measured [47, 83, 85, 86]. For example, a decrease of the fatty acid chain length of PC resulting in a significant increase in the flip-flop rate has been shown. Moreover, it was found that spin labeling can substantially affect the transbilayer dynamics of phospholipids (see above). Although the approach requires deuterated lipids, this modification is small in comparison with tagging of lipids by spin or fluorescent labels, and the behavior of natural lipids should be preserved. However, since the technique requires the assembly of asymmetric bilayers, it is restricted to model membranes and not applicable to native, biological membranes. Nevertheless, peptides (transmembrane domains) [87, 88] and membrane proteins can be reconstituted in asymmetric model membranes to study their influence on lipid transbilayer movements.

ABBREVIATIONS MβCD NBD PC PE PG PS SM

methyl-β-cyclodextrin 7-nitrobenz-2-oxa-1,3-diazol-4-yl phosphatidylcholine phosphatidylethanolamine phosphatidylglycerol phosphatidylserine sphingomyelin

REFERENCES â•‡ 1â•… R. D. Kornberg, H. M. McConnell, Biochemistry 1971, 10, 1111–1120. â•‡ 2â•… M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. â•‡ 3â•… J. Connor, A. J. Schroit, Biochemistry 1987, 26, 5099–5105.

22

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

â•‡ 4â•… â•‡ 5â•… â•‡ 6â•… â•‡ 7â•… â•‡ 8â•…

J. Connor, A. J. Schroit, Biochemistry 1988, 27, 848–851. J. Connor, K. Gillum, A. J. Schroit, Biochim. Biophys. Acta 1990, 1025, 82–86. D. L. Daleke, W. H. Huestis, Biochemistry 1985, 24, 5406–5416. D. L. Daleke, W. H. Huestis, Cell Biol. 1989, 108, 1375–1385. L. Tilley, S. Cribier, B. Roelofsen, J. A. F. Op den Kamp, L. L. M. van Deenen, FEBS Lett. 1986, 194, 21–27. J. A. F. Op den Kamp, New Comp. Biochem. 1981, 1, 84–125. J. A. F. Op den Kamp, Annu. Rev. Biochem. 1979, 48, 47–71. K. Simons, E. Ikonen, Nature 1997, 387, 569–572. E. Ikonen, K. Simons, Semin. Cell Dev. Biol. 1998, 9, 503–509. K. Simons, E. Ikonen, Science 2000, 290, 1721–1726. D. Lingwood, K. Simons, Science 2010, 327, 46–50. M. Seigneuret, A. Zachowski, A. Herrmann, P. F. Devaux, Biochemistry 1984, 23, 4271–4275. M. Bitbol, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 6783–6787. X. Buton, G. Morrot, P. Fellmann, M. Seigneuret, J. Biol. Chem. 1996, 271, 6651–6657. U. Marx, G. Lassmann, K. Wimalasena, P. Müller, A. Herrmann, Biophys. J. 1997, 73, 1645–1654. U. Marx, G. Lassmann, H.-G. Holzhütter, D. Wüstner, P. Müller, A. Hohlig, J. Kubelt, A. Herrmann, Biophys. J. 2000, 78, 2628–2640. D. J. Sillence, R. J. Raggers, G. Van Meer, Assays for transmembrane movement of sphingolipids. In Methods in Enzymology: Part B: Sphingolipid Metabolism and Cell Signaling, Y. A. Hannun, Jr. and A. H. Merrill, eds. Academic Press, San Diego, 2000, 562–579. A. Zachowski, Biochem. J. 1993, 294, 1–14. P. F. Devaux, P. Fellmann, P. Herve, Chem. Phys. Lipids 2002, 116, 115–134. M. S. Bretscher, Science 1973, 181, 622–629. E. Gorter, F. Grendel, J. Exp. Med. 1925, 41, 439–443. A. J. Verkleij, R. F. A. Zwaal, B. Roelofsen, P. Comfurius, D. Kastelijn, L. L. M. van Deenen, Biochim. Biophys. Acta 1973, 323, 178–193. R. F. A. Zwaal, B. Roelofsen, P. Comfurius, L. L. M. van Deenen, Biochim. Biophys. Acta 1975, 406, 83–96. Q. X. Li, W. Dowhan, J. Biol. Chem. 1988, 263, 11516–11522. A. Herrmann, M. J. Clague, A. Puri, S. J. Morris, R. Blumenthal, S. Grimaldi, Biochemistry 1990, 29, 4054–4058. D. B. Zilversmit, Methods Enzymol. 1983, 98, 565–573. R. L. Jackson, J. Westerman, K. W. A. Wirtz, FEBS Lett. 1978, 94, 38–42. J. M. Graham, J. A. Higgins, J. A. Higgins, Methods Mol. Biol. 1994, 27, 125–130. N. Kälin, J. Fernandes, S. Hrafnsdottir, G. van Meer, J. Biol. Chem. 2004, 279, 33228–33236. J. Backer, E. A. Davidowicz, J. Biol. Chem. 1981, 256, 13272–13277. D. L. Brasaemle, A. D. Robertson, A. D. Attie, J. Lipid Res. 1988, 29, 481–489.

â•‡ 9â•… 10â•… 11â•… 12â•… 13â•… 14â•… 15â•… 16â•… 17â•… 18â•… 19â•… 20â•…

21â•… 22â•… 23â•… 24â•… 25â•… 26â•… 27â•… 28â•… 29â•… 30â•… 31â•… 32â•… 33â•… 34â•…

References

23

35â•… Y. Lange, J. Dolde, T. L. Steck, J. Biol. Chem. 1981, 256, 5321–5323. 36â•… P. Thiagarajan, J. F. Tait, J. Biol. Chem. 1990, 265, 17420–17423. 37â•… S. J. Martin, C. P. M. Reutelingsberger, G. A. M. Kuijten, R. M. J. Keehnen, S. T. Pals, M. H. J. van Oers, J. Exp. Med. 1995, 182, 1545–1556. 38â•… F. A. Kuypers, R. A. Lewis, M. Hua, M. A. Schott, D. Discher, J. D. Ernst, B. H. Lubin, Blood 1996, 87, 1179–1187. 39â•… S. K. Dasgupta, P. Guchhait, P. Thiagarajan, J. Lab. Clin. Med. 2006, 148, 19–25. 40â•… J. Shi, Y. Shi, L. N. Waehrens, J. T. Rasmussen, C. W. Heegaard, G. E. Gilbert, Cytometry A 2006, 69, 1193–1201. 41â•… R. F. A. Zwaal, P. Comfurius, E. M. Bevers, Biochim. Biophys. Acta 1998, 1376, 433–453. 42â•… J. M. Graham, J. A. Higgins, P. Comfurius, E. Bevers, R. F. A. Zwaal, Methods Mol. Biol. 1994, 27, 131–142. 43â•… S.-Y. Choung, T. Kobayashi, K. Takemoto, H. Ishitsuka, K. Inoue, Biochim. Biophys. Acta 1988, 940, 180–187. 44â•… Y. Aoki, T. Uenaka, J. Aoki, M. Umeda, K. Inoue, J. Biochem. 1994, 116, 291–297. 45â•… U. Kato, K. Emoto, C. Fredriksson, H. Nakamura, A. Ohta, T. Kobayashi, K. Murakami-Murofushi, T. Kobayashi, M. Umeda, J. Biol. Chem. 2002, 277, 37855–37862. 46â•… A. Makino, T. Baba, K. Fujimoto, K. Iwamoto, Y. Yano, N. Terada, S. Ohno, S. B. Sato, A. Ohta, M. Umeda, K. Matsuzaki, T. Kobayashi, J. Biol. Chem. 2003, 278, 3204–3209. 47â•… J. Liu, J. C. Conboy, Biophys. J. 2005, 89, 2522–2532. 48â•… R. G. Sleight, R. E. Pagano, J. Cell Biol. 1984, 99, 742–751. 49â•… O. C. Martin, R. E. Pagano, J. Biol. Chem. 1987, 262, 5890–5898. 50â•… R. E. Pagano, O. C. Martin, H. C. Kang, R. P. Haugland, J. Cell Biol. 1991, 113, 1267–1279. 51â•… A. Chattopadhyay, E. London, Biochemistry 1987, 26, 39–45. 52â•… D. Huster, P. Müller, K. Arnold, A. Herrmann, Biophys. J. 2001, 80, 822–831. 53â•… M. Colleau, P. Herve, P. Fellmann, P. F. Devaux, Chem. Phys. Lipids 1991, 57, 29–37. 54â•… E. Middelkoop, B. H. Lubin, J. A. F. Op den Kamp, B. Roelofsen, Biochim. Biophys. Acta 1986, 855, 421–424. 55â•… H. A. Scheidt, P. Müller, A. Herrmann, D. Huster, J. Biol. Chem. 2003, 278, 45563–45569. 56â•… J. C. McIntyre, R. G. Sleight, Biochemistry 1991, 30, 11819–11827. 57â•… T. Pomorski, A. Herrmann, A. Zachowski, P. F. Devaux, P. Müller, Mol. Membr. Biol. 1994, 11, 39–44. 58â•… T. Pomorski, A. Herrmann, B. Zimmermann, A. Zachowski, P. Müller, Chem. Phys. Lipids 1995, 77, 139–146. 59â•… W. L. Bergmann, V. Dressler, C. W. M. Haest, B. Deuticke, Biochim. Biophys. Acta 1984, 769, 390–398. 60â•… J.-Y. Calvez, A. Zachowski, A. Herrmann, G. Morrot, P. F. Devaux, Biochemistry 1988, 27, 5666–5670.

24

ASSESSING TRANSBILAYER LIPID DISTRIBUTION AND MOVEMENT

61â•… A. Rousselet, A. Colbeau, P. M. Vignais, P. F. Devaux, Biochim. Biophys. Acta 1976, 426, 372–384. 62â•… G. Morrot, P. Herve, A. Zachowski, P. Fellmann, P. F. Devaux, Biochemistry 1989, 28, 3456–3462. 63â•… K. Müller, T. Pomorski, P. Müller, A. Zachowski, A. Herrmann, Biochemistry 1994, 33, 9968–9974. 64â•… J. Libera, T. Pomorski, O. JosimovicAlasevic, K. G. Fritsch, A. Herrmann, J. Bone Miner. Res. 1999, 14, 690–699. 65â•… P. F. Devaux, Biochemistry 1991, 30, 1164–1173. 66â•… A. J. Schroit, R. F. A. Zwaal, Biochim. Biophys. Acta 1991, 1071, 313–329. 67â•… P. F. Devaux, A. Zachowski, Chem. Phys. Lipids 1994, 73, 107–120. 68â•… A. Menon, Trends Cell Biol. 1995, 5, 355–360. 69â•… T. Pomorski, J. C. M. Holthuis, A. Herrmann, G. van Meer, J. Cell Sci. 2004, 117, 805–813. 70â•… T. Pomorski, P. Müller, B. Zimmermann, K. Burger, P. F. Devaux, A. Herrmann, J. Cell Sci. 1996, 109, 687–698. 71â•… A. Kurz, D. Viertel, A. Herrmann, K. Müller, Reproduction 2005, 130, 615–626. 72â•… P. Müller, T. Pomorski, S. Porwoli, R. Tauber, A. Herrmann, Hepatology 1996, 24, 1497–1503. 73â•… A. Tannert, D. Wüstner, J. Bechstein, P. Müller, P. F. Devaux, A. Herrmann, J. Biol. Chem. 2003, 278, 40631–40639. 74â•… R. E. Pagano, O. C. Martin, A. J. Schroit, D. K. Struck, Biochemistry 1981, 20, 4920–4927. 75â•… J. W. Nichols, R. E. Pagano, Biochemistry. 1981, 20, 2783–2789. 76â•… Z. M. Zhang, J. W. Nichols, Am. J. Physiol. 1994, 267, G80–G86. 77â•… S. Hrafnsdottir, J. W. Nichols, A. K. Menon, Biochemistry 1997, 36, 4969–4978. 78â•… P. Fellmann, P. Herve, T. Pomorski, P. Müller, D. Geldwerth, A. Herrmann, P. F. Devaux, Biochemistry 2000, 39, 4994–5003. 79â•… A. van Helvoort, A. J. Smith, H. Sprong, I. Fritzsche, A. H. Schinkel, P. Borst, G. van Meer, Cell 1996, 87, 507–517. 80â•… A. Pohl, H. Lage, P. Müller, T. Pomorski, A. Herrmann, Biochem. J. 2002, 365, 259–268. 81â•… A. Pohl, P. F. Devaux, A. Herrmann, Biochim. Biophys. Acta 2005, 1733, 29–52. 82â•… R. J. Raggers, A. van Helvoort, R. Evers, G. van Meer, J. Cell Sci. 1999, 112, 415–422. 83â•… J. Liu, J. C. Conboy, J. Am. Chem. Soc. 2004, 126, 8376–8377. 84â•… Y. R. Shen, Nature 1989, 337, 519–525. 85â•… T. C. Anglin, J. C. Conboy, Biochemistry 2009, 48, 10220–10234. 86â•… T. C. Anglin, M. P. Cooper, H. Li, K. Chandler, J. C. Conboy, J. Phys. Chem. B 2010, 114, 1903–1914. 87â•… T. C. Anglin, J. Liu, J. C. Conboy, Biophys. J. 2007, 92, L01–L03. 88â•… T. C. Anglin, K. L. Brown, J. C. Conboy, J. Struct. Biol. 2009, 168, 37–52.

2 DETECTION AND MEASUREMENT OF UNLABELED LIPID TRANSMEMBRANE MOVEMENT Iván López-Montero Departamento de Química Física I, Universidad Complutense de Madrid, Madrid, Spain

Marisela Vélez Consejo Superior de Investigaciones Científicas, Instituto de Catálisis y Petroleoquímica, Madrid, Spain IMDEA Nanociencia, Facultad de Ciencias, Universidad Autonóma de Madrid, Madrid, Spain

Philippe F. Devaux Institut de Biologie Physico-Chimique, Paris, France

2.1â•… INTRODUCTION Since the pioneering work of Kornberg and McConnell in 1971 [1], the measurement of transverse diffusion of lipids is associated with the use of lipid analogs. A large variety of labeled lipids has been developed mainly for electron spin resonance and fluorescence spectroscopy (see Chapters 1 and 6). These techniques require the use of lipid probes, which bear a nitroxide or a fluorescent group generally in one of their acyl chains. For the study of the spontaneous flip-flop on model lipid bilayers (as large unilamellar vesicles [LUVs]), the (spin or fluorescent)-labeled lipids are easily incorporated from

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

25

26

FLIP-FLOP OF UNLABELED LIPIDS

the organic solvent during sample preparation. However, the study of transmembrane diffusion (spontaneous or protein mediated) in cell membranes requires the use of water-soluble probes in order to facilitate their incorporation into the biomembrane from solution. The way to do this is to attach the probe to a modified lipid having a short fatty acid chain. This molecule, which is partly water soluble, spontaneously introduces itself into the membrane carrying the probe along. In addition, the short fatty acid chain enables the selective extraction of the probe from the external leaflet with the backexchange method (see Chapters 1 and 6). Once the probe is extracted, the inward transport of labeled lipid is then determined spectroscopically [2]. Likewise, the outward transmembrane lipid asymmetry can also be easily assessed in situ by chemical reduction of the probes exposed on the outer leaflet. The classical techniques of investigation with labeled lipids [1, 2] have proven to give at least reliable information on comparative values of the flipflop half-times (τ1/2’s) of different phospholipids [3]. However, the main objection to this experimental approximation comes from the possible divergences in the values of τ1/2, which may differ slightly from those corresponding to the endogenous lipid. The presence of the label involves a steric hindrance within the bilayer and could have a strong effect on the lipid–protein interactions involved in active transport carried by flippases. Methods suitable for the quantification of the transmembrane distribution of unlabeled lipids were originally based on exogenous phospholipase attack of the cell surface [4]. The precision of this technique is limited, and in situ lipid degradation can induce artifacts. Another approach consists of using radioactively labeled lipids. However, since these measurements of transmembrane motion are restricted to short-chain lipids and demand long and tedious experiments, studies using these molecular probes are relatively scarce [5]. Very recently, efforts have been made to develop new methodological approaches able to detect and to measure the transbilayer movement of unlabeled and long-chain lipids. First, the transbilayer diffusion of unlabeled molecules has been followed from the shape changes of giant unilamellar vesicles (GUVs) associated with the lipid reorganization [6, 7] and quantified considering the elastic and mechanical properties of lipid bilayers. A second methodology used to measure lipid flip-flop without requiring lipid probes is atomic force microscopy (AFM) [8, 9]. The ability of this technique to image with nanometer resolution enables the monitoring of the time evolution of lipid domains on supported bilayers caused by a lipid exchange between monolayers. Third, sum frequency vibrational spectroscopy has been applied to measure the translocation of unlabeled lipids on lipid bilayers [10, 11]. And very recently, time-resolved, small-angle neutron scattering technique has been applied to vesicle systems to determine the interparticle transfer and the flip-flop of unlabeled phospholipids [12]. In this chapter, we review the potential and the most relevant results obtained with the shape change approach on GUVs (see Section 2.2) and with AFM of supported lipid bilayers (see Section 2.3).

Measurement of Transmembrane Flip-Flop of Unlabeled Lipids

27

2.2â•… MEASUREMENT OF TRANSMEMBRANE FLIP-FLOP OF UNLABELED LIPIDS BY SHAPE CHANGE OF GUVs The shape of vesicles is very sensitive to the relative area between the two monolayers. This is the main principle on which the shape change method performed on GUVs to detect and measure lipid flip-flop is based. We start this section with a brief exposition of the bending theory of vesicle shapes. Some examples of shape changes caused by an asymmetric lipid distribution reported in the literature are then presented. A more detailed description of the shape change approach to measure unlabeled lipid flip-flop follows, and a few experimental results are then described. We conclude with a few remarks on the perspectives of new experiments related to energy-dependent flippases. 2.2.1â•… Bending Theory of Vesicle Shapes Lipid bilayers are elastic objects able to support mechanical deformations as stretching, compression, and bending. The mechanical response of the bilayer against lateral stretching or bending is governed by the material parameters, the stretching modulus ε, and the bending modulus κb, respectively. Stretching a lipid bilayer is energetically more expensive (εâ•¯∼â•¯200â•›mN/m) than bending it, which can be achieved by thermal energy (κbâ•¯∼â•¯10â•›kbT). Thus, in the frame of a morphological theory of lipid vesicles, surface area can be considered as fixed. A symmetrical bilayer would adopt a flat conformation in the absence of edge energy effects. Conversely, asymmetrical bilayers (with a different chemical composition or lipid density between monolayers) can adopt more curved shapes characterized by a spontaneous curvature C0, the preferred curvature of the lipid bilayer at equilibrium. Deviations from the preferred curvature have an energetic cost, expressed by the Hookean expression:

1 2 Eb = κ b dA (C − C0 ) . 2

∫

(2.1)

Equation 2.1 was first introduced by Helfrich [13]. A large variety of vesicular shapes were predicted by minimizing Equation 2.1 under different conditions of osmolarity and surface/volume ratio. However, the early work of Helfrich missed the bilayer nature of membranes and considered the vesicles just as two-dimensional elastic sheets. The second model, derived from Helfrich’s ideas, was implemented some years later [14] with the bilayer architecture of membranes. Indeed, when a closed bilayer is bent, an additional contribution to the pure bending mode (the bending of two monolayers as a fixed area) must be taken into account, that is, the relative stretching and compression of monolayers. This leads to enable changes in the differential area between monolayers given by ΔAâ•¯=â•¯Aoutâ•¯−â•¯Ain, where A represents the surface area of the outer (out) and inner (in) monolayers, respectively. Mathematically, the total energy reads,

28

FLIP-FLOP OF UNLABELED LIPIDS

1 1 ακπ 2 Eb = κ b dA (C − C0 ) + (∆A − ∆A0 )2, 2 2 Ah 2

∫

(2.2)

where ∆A0 = A0out − A0in is the difference of the unstressed monolayer areas and κ is the nonlocal bending modulus. h is the bilayer thickness and α accounts for the relative importance of the two contributions to the total bending energy. α has been estimated to be of the order of one for all phospholipids. Equation 2.2 represents the area difference elasticity (ADE) model [14]. Again, the minimization of Equation 2.2, at given parameters of volume-toarea ratio and preferred differential area, leads to the same diversity of vesicular shapes, but new pear, budded, starfish vesicles or vesicles with holes are now also predicted. Vesicle shape regions of minimal energy can be represented into a diagram, which contains different vesicle classes. An ADE shape diagram is reproduced in Figure 2.1 from Reference 15. Shape changes of vesicles triggered by an asymmetric distribution of lipids constitute an experimental evidence of the bilayer structure included on the ADE model (see next section). This is equivalent to maintain constant the vesicle volume and to modify the preferred area difference along the vertical axis of the shape diagram. Lipid flip-flop represents a very simple mechanism to vary the number difference in the ADE contribution (Eq. 2.2). 2.2.2â•… Asymmetry and Shape Changes of Membranes: Examples Following the shape diagram shown in Figure 2.1, it is evident that one could cross the different shape regions by modifying the spontaneous curvature and the preferred differential area of vesicles. Any asymmetry across the membrane will lead to a change in the effective spontaneous curvature. A large variety of mechanisms for monolayer surface area asymmetry generation can be found in the literature [16–18]. Among all of them, an asymmetric distribution of lipids between leaflets has been extensively explored in cells as well as in model vesicular systems. The influence of the incorporation and redistribution of lipids on erythrocyte morphology was first studied by Daleke and Huestis [19]. Red blood cells show a biconcave (discocyte) shape at physiological conditions of pH and osmolarity. Their shapes are strongly modified when exogenous lipid molecules are added to the plasma membrane. A discocyte-to-echinocyte shape change is induced when phosphatidylserine (PS) and phosphatidylcholine (PC) are incorporated to the external monolayer. In addition, PS-induced echinocytes revert to discocytes and then become invaginate stomatocytes. This second stage is produced by PS translocation into the inner monolayer by the aminophospholipid translocase [20] (Fig. 2.2). An asymmetric amount of lipids between both monolayers can also be easily promoted on pure lipid unilamellar vesicles. More curved shapes (budded vesicles) can be obtained from a prolate giant vesicle by the addition of exogenous lyso-PC on the outer monolayer. Conversely, selective depletion of

Measurement of Transmembrane Flip-Flop of Unlabeled Lipids

29

2.5

2.0 Dpro/pear Lpear pears Cpro/pear 1.5

starfish prolates

∆a0

Cnas/obi

Dpro/nas Aobi 1.0

Csto/obi Dpro/obi Asto

oblates Dsto/obi

0.5 stomatocytes Lsto

0.0 0.0

0.2

0.4

0.6

0.8

1.0

v

Figure 2.1.â•‡ Phase diagram of the ADE model [15]. The horizontal axis is the volumeto-area ratio expressed by the reduced volume v (scaled to that of a sphere). The vertical axis shows the effective differential area expressed by Δa0 (scaled to that of a sphere). Along the vertical axis, vesicles can change their shape by modifying the surface area difference between monolayers. Reprinted from Current Opinion in Colloid & Interface Science, 5, Hans-Günther Döbereiner, Properties of giant vesicles, 256–263, Copyright (2000), with permission from Elsevier.

30

FLIP-FLOP OF UNLABELED LIPIDS

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

Figure 2.2.â•‡ Scanning electron micrographs of erythrocytes. Discocytes (f) change their shape to echinocytes (a–e) when lipids are accumulated in the external monolayer. When PS molecules are incorporated to the membrane, they are finally transported to the inner leaflet by the action of a flippase, producing a shape change to stomatocyte configurations (g–j). Reprinted from Biochemistry, 24, D.L. Daleke and W.H. Huestis, Incorporation and translocation of aminophospholipids in human erythrocytes, 54065416, Copyright © 1985, with permission from American Chemical Society.

symmetrically incorporated lyso-PC molecules, from the outer monolayer of vesicles by incubation with bovine serum albumin (BSA), a fatty acid transporter, makes vesicles to adopt invaginated shapes [21]. Finally, the redistribution of endogenous lipids constitutes another example of morphological changes caused by an accumulation of some negatively charged lipids (as phosphatidylglycerol [PG]) into one monolayer in response to transmembrane pH gradients. Again, the outward transbilayer transport of PG in LUVs made of a mixture of PC and PG [22] transforms initially invaginated LUVs to long narrow tubular structures, or spherical structures with one or more budded vesicles. On the other hand, when PG inward transport is induced, inverting the pH gradient, noninvaginated LUVs reverse to invaginated vesicular structures. All the experiments described above inspired the technique of shape changes performed on giant vesicles to quantify the transbilayer diffusion of unlabeled lipids presented in the next section. 2.2.3â•… Flip-Flop Detection by Shape Changes in Giant Vesicles Shape of GUVs can be easily modified by varying experimentally the parameters that describe their shape: the preferred area difference between monolayers (asymmetry) and the vesicular surface area/volume ratio (see Fig. 2.1). GUVs are very sensitive to the relative area between monolayers and the minimum asymmetry needed to induce a shape change is of the order of 0.1% of the total area of the vesicle [23, 24]. GUVs are consequently, in comparison with other membrane models, excellent objects to detect lipid asymmetry. They are visible under optical microscopy, which facilitates their observation and manipulation, and their internal volume can be modified. This is important

Measurement of Transmembrane Flip-Flop of Unlabeled Lipids

31

since the reduced volume must be lower than 1 (which corresponds to a sphere) in order to have a vesicular surface excess for allowing shape changes. The most extended method to fabricate giant vesicles is electroformation [25]. Giant vesicles obtained by this technique are unilamellar in a very high percentage but have spherical shape. Vesicles with lower reduced volumes can be obtained by allowing the external solution to evaporate for some minutes. Due to an osmotic pressure difference between inside and outside, spherical vesicles deflate and adopt a prolate shape. The “shape change” approach to infer the flip-flop rates (τ1/2) of nonlabeled lipids is based on the incorporation of a few molecules to the external membrane leaflet that increase its area and trigger a shape change from a prolate to a budded shape vesicle. If a transbilayer movement of lipids occurs, the prolate shape is recovered due to the lipid redistribution into both lipid monolayers. Monitoring the times it takes for the shape changes to occur allows one to calculate the τ1/2’s according to the theoretical model described in the next section. 2.2.4â•… Theoretical Model For quantitatively estimating the translocation rate of unlabeled molecules in shape changes experiments performed on GUVs, a very simple kinetic model can be used [6]. This model is based on the time evolution of the asymmetry function c, which is defined as the difference of the unlabeled lipid concentration between both monolayers:

c (t ) = cout (t ) − cin (t ) ,

(2.3)

where cout(t) and cin(t) are the unlabeled lipid concentrations (in percentage of total lipids) in the outer and the inner monolayer, respectively. Obviously, both the outer and inner concentrations must take into account simultaneously the superposition of the kinetics of exogenous nonlabeled lipid insertion in the outer monolayer and their diffusion toward the inner monolayer. As said before, the threshold for shape changes in GUVs is given by a lipid asymmetry of the order of 0.1% of the total area of the vesicle. This means that budding transition and shape recovery would occur if câ•¯=â•¯cthrsâ•¯=â•¯0.1%. In a first approximation, the time evolution of the two different monolayers must explicitly introduce the two rate constants for insertion and flip-flop and can be written as a differential equation system:

dcout = f (t ) − Kout cout + Kin cin , dt

(2.4)

dcin = Kout cout − Kin cin , dt

(2.5)

where f(t) accounts for an empirical function of lipid incorporation to the external monolayer. Kout and Kin are the translocation rates from the outer

32

FLIP-FLOP OF UNLABELED LIPIDS

to the inner monolayer and from the inner to the outer monolayer, respectively. The expression for f(t) is independently measured [6, 7] and depends on two main parameters: a characteristic time of incorporation and the final concentration of amphiphilic molecules incorporated in the outer monolayer, which is expressed in percentage of total lipids in the GUV. Both parameters depend on the nature of the injected molecules. Several approximations have been assumed in order to simplify Equations 2.4 and 2.5: • The threshold of asymmetry is rigorously defined by ∆S / S = 0.1 ⇒ cthrs = 0.1(avl ) /(aincl ) where avl is the area occupied by a vesicular lipid molecule and aincl is that of the incorporated molecule. Both molecular areas have been assumed to be the same, and therefore, cthrsâ•¯=â•¯0.1. Actually, the theoretical threshold for shape change established on 0.1% is a larger source of error. • The outer aqueous volume is very large compared with the volume enclosed by GUVs. Therefore, only a small fraction of the total added lipid will be incorporated into the vesicle, and the outer aqueous concentration of exogenous lipids can be considered as a constant. • The fraction of exogenous molecules in the inner monolayer, which spontaneously partitions into the internal volume of vesicle, is negligible because the internal volume of the vesicles is very small in comparison to the external volume. Likewise, the water–membrane exchange of the incorporated molecules to the outer monolayer has been neglected. For a GUV, both leaflets are equivalent (in terms of curvature and tension) and therefore, Koutâ•¯∼â•¯Kinâ•¯=â•¯K. By taking into account the above considerations, Equations 2.4 and 2.5 read

dc(t ) = f (t ) − 2 Kc(t ). dt

(2.6)

An analytical solution for c(t) can be obtained once f(t) is experimentally measured. c(t) results as a function that depends on two main kinetic parameters: the incorporation rate of exogenous molecules to the outer monolayer and the flip-flop rate K. The time behavior of the monolayer area asymmetry (Fig. 2.3) has a nonmonotonic character and starts from zero (there is no asymmetry at tâ•¯=â•¯0). The initial increase of c(t) corresponds to the incorporation of exogenous molecules to the external monolayer, whereas the decrease of c(t) with time results from re-equilibration of nonlabeled molecules between the external and the inner monolayer. The curve intercepts the threshold of shape change câ•¯=â•¯cthrsâ•¯=â•¯0.1% at t1 and t2, at the budding transition and the shape recovery, respectively. The experimental values of t1 and t2 allow one to fit the value of K and hence τ1/2â•¯=â•¯ln2 / K.

Measurement of Transmembrane Flip-Flop of Unlabeled Lipids

33

0.12 t1

t2

c (% of total lipid)

0.10

0.08

0.06

0.04

0.02

0.00

0

2

4

6

8

time (minutes)

Figure 2.3.â•‡ Experimental time evolution of the asymmetry function c for C6-ceramide (see Section 2.5) incorporated to a prolate shape egg-PC GUV. At times before t1, the asymmetry is lower than cthrs (solid horizontal line) and the vesicle shape remains as a prolate. When câ•¯>â•¯cthrs, the created asymmetry is high enough to induce the budding transition, which is reversed at t2 when ceramides are redistributed into both monolayers by spontaneous flip-flop. From t1 and t2 (dashed vertical lines), flip-flop half-times can be deduced.

2.2.5â•… Examples A good example for illustrating the use of the “shape change” methodology is the measure of the τ1/2 of unlabeled ceramides [6]. In those experiments, ceramides with different chain lengths (C6-, C10-, and C16-ceramides) were added externally with a micropipette to GUVs made of egg-PC. Figure 2.4 illustrates the shape change pathways followed by giant vesicles when ceramides are injected. Unlabeled ceramides trigger first a budding transition at t1. After a short period of time (t2), a sudden reopening process takes place (Fig. 2.4f), and the initial shape is recovered. Time t1, of the order of 1 minute, varies from one ceramide to another and is longer for longer acyl chains. Likewise, the reopening of the neck was observed 9–15 minutes after t1. From t1 and t2, the τ1/2 for ceramides was calculated to be of the order of 1 minute at 20°C (1.2, 1.6, and 2.4 minutes for C6-, C10-, and C16-ceramides, respectively). Shape change experiments can also be performed at different temperatures ranging from the melting temperature (Tm) of lipids, which conform the giant vesicle (in case Tm was lower than the melting temperature of water, 0°C would obviously be the lower limit temperature) to the highest temperature at which

34

FLIP-FLOP OF UNLABELED LIPIDS

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

(k)

(l)

Figure 2.4.â•‡ Shape changes induced by C6-ceramide externally added to an egg-PC giant vesicle. A budding transition is first formed when molecules are accumulated into the outer monolayer (a–e). The prolate shape is recovered when ceramides are redistributed into both monolayers by spontaneous flip-flop (f–l). Scale bar is 10â•›µm. From Reference 6.

shape changes are observable (in case that Kâ•¯>>â•¯incorporation rate, no shape change will be observed on vesicles). For ceramides, the spontaneous transbilayer half-times shifted to values smaller than 1 minute at 37°C (18 and 57 seconds for C6- and C16-ceramides, respectively). It is worth noting that the incorporation of lipids having a very slow transbilayer movement, as lyso-PC [7] or C6-PC, produces a stable budding transition (vesiculation) in GUVs made of PC, as already shown in the previous experiments commented in Section 2.2.2. In a second example [7] illustrating the potentiality of the shape change approximation to detect flip-flop of unlabeled lipids, the prolate shape produced by those lipids was reversed on GUVs prepared with membrane protein extracts from the endoplasmic reticulum (ER) of yeast (see Chapter 6). In eukaryotes, the cytoplasmic leaflet of the ER membrane is the major site of phospholipid biosynthesis. To ensure stable membrane growth, energyindependent flippases mediate rapid, bidirectional, and rather unspecific phospholipid flip-flop with half-times of minutes or less. When egg-lyso-PC or C6-PC was externally added to the reconstituted vesicles, a budding transition

Measurement of Transmembrane Flip-Flop of Unlabeled Lipids

35

occurred within the first 5 minutes, and subsequently, the vesicle regained the prolate shape within 5–10 minutes. Therefore, the reconstituted ER flippase activity triggers the rapid flip-flop that diminishes the surface area difference between the two leaflets (see Chapter 6). 2.2.6â•… Perspectives In comparison with other methods presented in this book, the main advantages of the “shape change” approach are the measurement of flip-flop of unlabeled molecules and long double acyl chain lipids. However, the low solubility in water of the latter molecules arises as the main shortcoming of the method. First, organic solvents are required to accelerate slow diffusion in water of long-chain lipids and moreover to facilitate their incorporation into vesicles. One could argue about possible artifacts associated to the use of organic solvents, but control experiments indicate that the same amount of organic solvent in the absence of the nonlabeled molecule does not trigger any budding transition. Second, a slow rate of incorporation relatively to flipflop disables the triggering of the budding transition. If all incorporated molecules are rapidly redistributed into both monolayers, the accumulation in the outer leaflet needed to induce changes is not reached. Finally, this method is limited to fluid giant vesicles. Indeed, giant vesicles in a liquid-ordered phase can be made of sphingomyelin (SM)/cholesterol (Chol) mixtures. The addition of molecules into the external leaflet triggers the budding transition, but fission of the budded vesicle is produced. This phenomenon prevents the shape recovery and therefore the lipid flip-flop quantification. However, the present approach should also allow investigation of energydependent flippases by reconstituting them into giant vesicles [26]. Although flippases are very likely reconstituted to a similar extent in both opposing directions, unidirectional lipid transport can be ensured by allowing ATP to access only one membrane leaflet. Characterizing the activity of energydependent flippases would not require the addition of lipids. Lipid species to be transported can already be incorporated during GUV formation. Because of the unidirectional fashion of oriented lipid transport, area differences would be created leading to a shape change. This is different from the visualization of energy-independent flippase activity in GUVs where shape changes have to be triggered first, for example, by supplementing the external leaflet with additional lipids. This type of experiment, which requires neither labeled lipids nor short-chain lipids, is illustrated in Figure 2.5 [27]. In this experiment, GUVs were fabricated from erythrocyte membrane extracts lacking the cytoskeleton proteins (spectrin and actin). The addition of magnesium-adenosine triphosphate (Mg-ATP) in the vicinity of a giant prolate vesicle containing solely the endogenous lipids of a red cell membrane triggered the budding transition by the active outward transport of endogenous PS and phosphatidylethanolamine (PE) by the aminophospholipid translocase [20]. This kind of experiments will be important for the identification of the lipid specificity recognition by lipid flippases [28].

36

FLIP-FLOP OF UNLABELED LIPIDS

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

Figure 2.5.â•‡ Giant vesicle made from erythrocyte membrane extract. The activation of aminophospholipid translocase by external addition of Mg-ATP (1â•›mM) induces a budding transition from a prolate vesicle. PS and PE are transported outward and accumulated into the outer monolayer. Time after Mg-ATP injection: (a) tâ•¯=â•¯0 seconds, (b) tâ•¯=â•¯45 seconds, (c) tâ•¯=â•¯46 seconds, (d) tâ•¯=â•¯47 seconds, (e) tâ•¯=â•¯48 seconds, (f) tâ•¯=â•¯52 seconds, (g) tâ•¯=â•¯1 minute, (h) tâ•¯=â•¯1 minute and 1 second, and (i) tâ•¯=â•¯1 minute and 2 seconds. Scale bar represents 10â•›µm. Temperature: 20°C. From Reference 27.

2.3â•… MEASUREMENT OF TRANSMEMBRANE FLIP-FLOP OF UNLABELED LIPIDS USING AFM AFM is a surface characterization technique that provides topographic information of surfaces [29, 30]. Developed in the mid-1980s within the physics community, the technique has spread among the biologist/biophysicist given its unique capacity to characterize surfaces with nanoscopic resolution while

Transmembrane Flip-Flop of Unlabeled Lipids Using AFM

37

8 7 6

Z (Å)

5 4 3 2 1 0 0

100 200 300 400 500 600 700 X (nm)

Figure 2.6.â•‡ Regions with different lipid-packing density in a supported lipid bilayer made of an Escherichia coli lipid extract (left panel). The height profile (indicated with a solid line on the AFM image) shows the height difference (0.5â•›nm) between both regions.

maintaining the surface immersed in solution. Supported biological membranes have been widely characterized structurally using this technique [31, 32]. The height resolution detects lipid domains that differ in as little as 0.5â•›nm, allowing then to unambiguously identify regions with different lipid aggregation state or packing density (Fig. 2.6). This, combined with its lateral resolution, serves to identify nanometer-range lipid-ordered domains [33]. This size range fills the gap between the micrometer scale accessible through optical microscopy and the high resolution obtained by electron microscopy. The direct topographic information it provides cannot be accessed easily through other techniques available. Since the image contrast is provided by height differences in the membranes, the AFM has the additional advantage that it does not require the use of external labels to provide domain information. Another quality of the AFM microscope that makes it convenient for characterizing biological samples is its ability to image surfaces immersed in a liquid environment. It allows time-resolved monitoring of a surface undergoing modifications. The surface remodeling local effect of lipid-degrading enzyme activities [9, 34, 35] has been detected within the minute time resolution accessible with the AFM. The existing limitations to imaging at higher speeds are due to the mechanical stability of the microscope and not to limitations in the basic operating principles. Great progress has been made recently in developing fast scanning probe equipment [36] that promise an exciting future for this technique. The sample preparation requirement of supporting the membrane onto a solid surface does however impose certain limitations to the type of questions

38

FLIP-FLOP OF UNLABELED LIPIDS

that can be addressed with the AFM in the study of membranes. Anchorage to a solid support requires that the lipid heads of the bottom monolayer interact directly with the solid surface. This interaction however does not completely immobilize the lipids. It is well documented that, although restricted, the lipids are still free to diffuse laterally on supported lipid membranes on different substrates [37]. The water monolayer trapped between the polar head groups and the solid surface is enough to allow them to diffuse. Although it is not obvious if this membrane configuration required by the AFM supports lipid flip-flop of individual lipids from the upper layer to the lower layer of vice versa, there have been a few cases in which phospholipids exchange between the monolayers has been reported [8]. There are basically two circumstances in which this lipid reorganization is clearly observed: (1) when asymmetric supported bilayers are formed initially [8] and (2) when the lipid asymmetry is induced by an enzyme activity affecting only the monolayer exposed to the solution [9]. The nanometric structural information of unlabeled membranes provided by AFM complements the dynamic and spectroscopic information provided by the use of fluorescent techniques, making their combined use of particular interest for understanding at the molecular level the effect of lipid composition on structure and rearrangement of membranes. A good illustration of the contribution of AFM to the study of flip-flop is provided by the study of Lin et al. [8]. Asymmetric supported membranes of two lipids with different transition temperatures, 1,2-dilauroyl-sn-glycero3-phosphocholine (DLPC) and 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) lipids, were formed. The membranes were prepared by varying the thermal history of the vesicles or the substrate–vesicle suspension temperature differential during deposition. Depending on the detailed temperature history of the sample preparation, three types of membranes differing in the domain height distribution and in stability were observed. Whereas two of the samples were very stable, one of the configurations was observed to undergo domain height redistribution within a few hours. A combination of fluorescent techniques, fluorescence quenching and fluorescence recovery after photobleaching (FRAP), was able to associate the domains to an asymmetric lipid composition in the two membrane leaflets. Symmetric, symmetric/asymmetric, and asymmetric domains of DLPC/DSPC were identified (Fig. 2.7). The reorganization of the domains was interpreted as being due to lipid reorganization induced by lipid flip-flop between the two membranes. The possibility provided by AFM to image with nanometer resolution the time evolution of the domains allowed addressing the question of whether this lipid exchange occurred within the domains or if it occurred preferentially on their perimeter. The evolution of the domain perimeter was quantified, and the values obtained were compatible with the estimates obtained by assuming that the flip-flop occurs at the perimeter and not with the estimates considering exchange along the whole domain area. A rate constant of 76â•¯±â•¯17 per hour was estimated.

Transmembrane Flip-Flop of Unlabeled Lipids Using AFM

(a)

(b)

39

(c)

Figure 2.7.â•‡ Supported lipid bilayers made of DLPC and DSPC and prepared with different methods, which lead to three different configurations: symmetric (a), symmetric/ asymmetric (b), and asymmetric domains (c). The height profiles are denoted with a black dotted line on the AFM image. Reprinted from Biophysical Journal, 90, WanChen Lin, Craig D. Blanchette, Timothy V. Ratto, and Marjorie L. Longo, Lipid asymmetry in DLPC/DSPC-supported lipid bilayers: A combined AFM and fluorescence microscopy study, 228–237, Copyright (2006), with permission from Elsevier.

Another example of the use of AFM in combination with fluorescence techniques to observe the molecular details of membrane modifications induced by lipid flip-flop is given by the work by Lopez-Montero et al. [9]. Bilayers formed by egg-PCâ•›:â•›egg-PEâ•›:â•›SMâ•›:â•›Chol (1:1:1:1â•›mol) present liquidordered domains rich in SM. Exposure of the external monolayer to sphingomyelinase activity induces an in situ asymmetrical SM conversion to ceramide. The smaller area occupied by the ceramide head compared with the SM reduces the area of the external leaflet relative to the inner one, and this mismatch generates a tension. In GUVs, this tension is enough to collapse the vesicles even at low percentage of SM (≈â•¯5%â•›mol). When supported bilayers of the same composition and under the same treatment were analyzed with AFM, nanometer-sized holes and the appearance of higher domains were observed (Fig. 2.8). The results were interpreted as indicating that the pores formed due the lipid area mismatch allowed the redistribution of ceramide across the two leaflets with the subsequent formation of ceramide-rich domains slightly protruding from the surface.

40

FLIP-FLOP OF UNLABELED LIPIDS

(a)

(b)

(c)

(d) 1.4 1.2

Z (Å)

1 0.8 0.6 0.4 0.2 0 0

0.2

0.4 0.6 X (mm)

0.8

1

Figure 2.8.â•‡ Direct observation of nanoscale defects produced by SM to ceramide conversion. Time after injection of sphingomyelinase: (a) 0 minute, (b) 60 minutes, and (c) 65 minutes. The scanned region in (a) and (b) is the same as that delimited by a dashed square in (c). Dimensions of the dashed square are 5â•¯×â•¯5â•›µm. The nanometer-sized holes that appear upon enzymatic treatment are pointed out by arrowheads (b). The ceramide-rich domains indicated by white arrows are also observables in (c). The profile from the black bar on AFM images indicating the relative heights of different kinds of domains is shown in (d). From Reference 9.

These two examples illustrate that the time-resolved topographic information provided by AFM is an invaluable tool to complement the micrometer spectroscopic and dynamic information provided by fluorescence. Observing lipid rearrangements at the nanoscale occurring in the minute timescale contributes to understand the molecular mechanism underlying lipid scrambling and its implications for membrane structure.

Abbreviations

41

2.4â•… CONCLUSIONS Spontaneous lipid flip-flop on biological membranes is a slow process that can take from hours to days. The translocations most relevant to signal transduction and other important cell metabolic functions take place however much faster, more within the minute timescales (Chol, fatty acids, ceramides, or PG). The two ways of detecting unlabeled membrane asymmetry referred to in this chapter, shape changes in GUVs and lipid rearrangement detected by AFM, can provide information within this faster scale that can prove useful to understand the mechanism and regulation of unlabeled lipid translocation in artificial membranes. However, the above-mentioned lipids represent a little percentage of the large variety of lipids present in cells. Therefore, both methodological approaches shown here would be extended to the study of lipid transbilayer movement mediated by energy-dependent flippases. The experiments carried out on GUVs shown in Section 2.2 constitute stimulating evidence that questions referred to unlabeled lipid specificity recognition by proteins for instance, can be approached with both methodologies. The development of such techniques (among others that do not require marked lipids) can give a useful information, which can be, in many cases, complementary to that obtained for labeled lipids during the last four decades.

ACKNOWLEDGMENTS I.L.-M. thanks CAM and MICINN for financial support from NANOBIO-M and from Juan de la Cierva program 2007, respectively. This work was supported by MEC under grants BFU2005-0487-C02-01 and C02-02, S-0505/MAT-0283, and BIO200804478-C03-02.

ABBREVIATIONS ADE AFM BSA C10- C16- C6- Chol DLPC DSPC ER FRAP

area difference elasticity atomic force microscopy bovine serum albumin C10-acylC16-acylC6-acylcholesterol 1,2-dilauroyl-sn-glycero-3-phosphocholine 1,2-distearoyl-sn-glycero-3-phosphocholine endoplasmic reticulum fluorescence recovery after photobleaching

42

GUV kb LUV lyso-PC Mg-ATP PC PE PG PS SM T τ1/2

FLIP-FLOP OF UNLABELED LIPIDS

giant unilamellar vesicle Boltzmann constant large unilamellar vesicle lyso-phosphatidylcholine magnesium-adenosine triphosphate phosphatidylcholine phosphatidylethanolamine phosphatidylglycerol phosphatidylserine sphingomyelin temperature flip-flop half-time

REFERENCES â•‡ 1â•…R. D. Kornberg, H. M. McConnell, Biochemistry 1971, 10, 1111–1120. â•‡ 2â•… P. Fellmann, A. Zachowski, P. F. Devaux, Methods Mol. Biol. 1994, 27, 161–175. â•‡ 3â•… A. Pohl, I. López-Montero, F. Rouviere, F. Giusti, P. F. Devaux, Mol. Membr. Biol. 2009, 26, 194–204. â•‡ 4â•… A. J. Verkleij, R. F. Zwaal, B. Roelofsen, P. Comfurius, D. Kastelijn, L. L. van Deenen, Biochim. Biophys. Acta 1973, 323, 178–193. â•‡ 5â•… C. J. Kirby, G. Colin, Biochem. J. 1977, 168, 575–577. â•‡ 6â•… I. López-Montero, N. Rodriguez, S. Cribier, A. Pohl, M. Vélez, P. F. Devaux, J. Biol. Chem. 2005, 280, 25811–25819. â•‡ 7â•… A. Papadopulos, S. Vehring, I. López-Montero, L. Kutschenko, M. Stöckl, P. F. Devaux, M. Kozlov, T. Pomorski, A. Herrmann, J. Biol. Chem. 2007, 282, 15559–15568. â•‡ 8â•… W. C. Lin, C. D. Blanchette, T. V. Ratto, M. L. Longo, Biophys. J. 2006, 90, 228–237. â•‡ 9â•… I. López-Montero, M. Vélez, P. F. Devaux, Biochim. Biophys. Acta 2007, 1768, 553–561. 10â•… T. C. Anglin, J. Liu, J. C. Conboy, Biophys. J. 2007, 92, L01–L03. 11â•… T. C. Anglin, J. C. Conboy, Biophys. J. 2008, 95, 186–193. 12â•… M. Nakano, M. Fukuda, T. Kudo, N. Matsuzaki, T. Azuma, K. Sekine, H. Endo, T. Handa, J. Phys. Chem. B 2009, 113, 6745–6748. 13â•… W. Helfrich, Z. Naturforsch. 1973, 28c, 693–703. 14â•… L. Miao, U. Seifert, M. Wortis, H. G. Döbereiner, Phys. Rev. E 1994, 49, 5389–5407. 15â•… H. G. Dobereiner, Curr. Opin. Colloid Interface Sci. 2000, 5, 256–263. 16â•…R. Lipowsky, Europhys. Lett. 1995, 30, 197–202. 17â•… J. O. Rädler, I. Koltover, T. Salditt, C. R. Safinya, Science 1997, 275, 810–814.

References

43

P. G. Petrov, J. Lee, H. G. Döbereiner, Europhys. Lett. 1999, 48, 435. D. L. Daleke, W. H. Huestis, Biochemistry 1985, 24, 5406–5416. M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. E. Farge, P. F. Devaux, Biophys. J. 1992, 61, 347–357. B. L. Mui, H. G. Döbereiner, T. D. Madden, P. R. Cullis, Biophys. J. 1995, 69, 930–941. 23â•… K. Berndl, J. Käs, R. Lipowsky, E. Sackmann, U. Seifert, Europhys. Lett. 1990, 13, 659–664. 24â•… E. Farge, P. F. Devaux, J. Phys. Chem. 1993, 97, 2958–2961. 25â•… M. L. Angelova, D. S. Dimitrov, Faraday Discuss. Chem. Soc. 1986, 81, 303–311. 26â•… P. F. Devaux, Cell. Mol. Biol. Lett. 2002, 7, 227–229. 27â•… P. F. Devaux, I. López-Montero, S. Bryde, Chem. Phys. Lipids 2006, 141, 119–132. 28â•… T. Pomorski, A. K. Menon, Cell. Mol. Life Sci. 2006, 63, 2908–2921. 29â•…R. J. Colton, D. R. Baselt, Y. F. Dufrêne, J. B. Green, G. U. Lee, Curr. Opin. Chem. Biol. 1997, 1, 370–377. 30â•… D. Muller, Biochemistry 2008, 47, 7986–7998. 31â•… P. L. Frederix, P. B. Brosshart, A. Engel, Biophys. J. 2009, 96, 329–338. 32â•… E. I. Goksu, J. M. Vanegas, C. D. Blanchette, W. C. Lin, M. L. Longo, Biochim. Biophys. Acta 2009, 1788, 254–266. 33â•… W. C. Lin, T. V. Ratto, M. L. Longo, Methods Mol. Biol. 2007, 400, 503–513. 34â•… M. C. Giocondi, B. Seantier, P. Dosset, P. E. Milhiet, C. Le Grimellec, Pflugers Arch. 2008, 456, 179–188. 35â•… L. Ira, J. Johnston, Biochim. Biophys. Acta 2008, 1778, 185–197. 36â•… T. Ando, T. Uchihashi, N. Kodera, D. Yamamoto, A. Miyagi, M. Taniguchi, H. Yamashita, Pflugers Arch. 2008, 456, 211–225. 37â•… Y. H. Chan, S. G. Boxer, Curr. Opin. Chem. Biol. 2007, 11, 581–587. 18â•… 19â•… 20â•… 21â•… 22â•…

PART II LIPID ASYMMETRY IN CELL MEMBRANES

3 NEW INSIGHTS IN MEMBRANE LIPID ASYMMETRY IN ANIMAL AND PLANT CELLS Alain Zachowski Laboratory of “Physiologie Cellulaire et Moléculaire des Plantes,” Université Pierre et Marie Curie—Paris 6 (UR 5) and Centre National de la Recherche Scientifique (EAC 7180), Paris, France

3.1â•… LIPID ASYMMETRY IN ANIMAL MEMBRANES In the early 1970s, the fact that phospholipids could not be randomly distributed across a biological membrane was established. The human erythrocyte membrane served as a model of plasma membrane, and the methods used were either chemical modification or enzymatic hydrolysis of phospholipids exposed at the cell surface [1–3]. The asymmetrical distribution was striking: ca. 70% of the phosphatidylcholine and more than 85% of the sphingomyelin were localized in the outer membrane leaflet; conversely, more than 80% of the phosphatidylethanolamine and almost all of the phosphatidylserine were found in the cytoplasmic leaflet. This distribution holds for almost all normal erythrocytes, except for ruminant ones that are deficient in phosphatidylcholine and very rich in sphingomyelin, and where phosphatidylethanolamine is almost exclusively found in the inner leaflet [4]. This distribution can be perturbed in pathological (sickle cells, thalassemic cells) or senescent red blood cells, as described in other chapters. When the same methods were used with

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

47

48

ARE PLANT MEMBRANE LIPIDS ASYMMETRICALLY DISTRIBUTED?

other types of cells, the interpretation of the results was more complex: if erythrocyte membrane is rather impermeant to the chemicals used and resistant to phospholipase-induced lysis and to perturbations induced by phospholipase products, this is not always the case for plasma membranes from nucleated cells. In addition, plasma membrane has to be isolated in order to have access to its total lipid composition. Cell lysis, which is thus the primary step in the isolation protocol, can induce a partial redistribution of the lipids within the bilayer [5]. Finally, if any transmembrane movement of lipids is faster than the reaction time required for labeling or hydrolysis, the description of asymmetry can be erroneous (see Chapters 1 and 6). However, all the data can be interpreted in the same way, namely an enrichment of the outer leaflet in choline-containing phospholipids and an accumulation of aminophospholipids in the inner leaflet. A crucial question was to know whether phosphatidylserine is totally excluded from the outer cell membrane leaflet, or whether it can be present in very small amounts on the cell surface. In other words, were the methods used and the detection of the modified lipid sensitive enough to reveal the presence of a small fraction of a low abundant phospholipid? This was not the case with chemical labeling of the amino group or formation of lysoderivatives by phospholipases A2. A more sensitive method is the prothrombinase assay that can reveal the presence of few percents of phosphatidylserine in a bilayer [6]. However, all these methods work on a cell population and cannot distinguish variations at the single cell level. A very interesting new technical approach to the question of phosphatidylserine asymmetry has been based on the use of fluorescent Annexin V. This protein specifically binds to the lipid in the presence of calcium [7], and if some phosphatidylserine is present on the plasma membrane outer leaflet, living cells become fluorescent (for reviews, see References 8 and 9). Thus, one has access to the situation existing at a single cell level. This noninvasive labeling is now a well-recognized assay and largely confirms that, in most of the “normal, resting” cells, phosphatidylserine is absent from the plasma membrane external layer. On the contrary, it has been shown that apoptotic cells are labeled, indicating that some phosphatidylserine has migrated into the plasma membrane outer layer. Even if some peptides, such as duramycin, can bind phosphatidylethanolamine [10], they have been seldom utilized to label cell surface and to check for the presence of this phospholipid. Determination of lipid distribution in organelle membranes has been much less studied than in plasma membranes. A reason is certainly that one has to isolate these membranes as closed, impermeable vesicles before assaying lipid distribution by either chemical labeling or enzymatic assays (see Chapter 6). On the other hand, the presence of some flippases in given organelles might indicate whether or not a phospholipid asymmetry exists in the membrane. For instance, the presence of an aminophospholipid translocase in the transGolgi network [11] might indicate that phosphatidylserine is asymmetrically distributed in this membrane. Similarly, the presence of a scramblase (see

Phospholipid Transporters

49

Chapter 7) would explain a rather symmetrical distribution of phospholipids in organelles such as exosomes [12]. 3.2â•… CREATING, MAINTAINING, OR RANDOMIZING THE MEMBRANE PHOSPHOLIPID DISTRIBUTION: PHOSPHOLIPID TRANSPORTERS Over the past two decades, numbers of reviews have dealt with the various proteins responsible for lipid translocation within the bilayer, and some of these “flippases” are the subject of other chapters. For some recent reviews, see References 13–17. So, I will just make a brief summary of what has been described in animal or yeast cells. The situation in plant cells will be exposed later in this review. 3.2.1â•… Aminophospholipid Translocase In the 1980s, an ATP-dependent, inward translocation of aminophospholipids in the plasma membrane of erythrocytes and some nucleated cells has been largely documented [1, 3, 18]. Because of the phospholipid specificity, this transporter was named “aminophospholipid translocase.” The transport requires Mg2+ besides ATP, is inhibited by vanadate and calcium, and is sensitive to protein-modifying reagents. The identification of the active peptide, suspected to be identical to the bovine chromaffin granule ATPase II [19], has been established by Tang et al. [20], who showed that this ATPase is able to complement a yeast mutant deficient in the aminophospholipid transport. This was the first member of a large P-type ATPase family, called P4-ATPases, that is present in animals, plants, or yeasts [21] (see Chapters 8 and 9). The P4ATPases maintain aminophospholipid asymmetry and are also involved in membrane trafficking in the endocytic and secretory pathways [22, 23]. A recent question is whether the ATPases can support by themselves the lipid transport or they need a β-subunit to fulfill this function [24]. Peptide homologs to the protein Cdc50p found in yeast [25] would be excellent candidates. 3.2.2â•… ATP-Binding Cassette (ABC) Transporters These transporters would function in an opposite direction to the aminophospholipid translocase. The lipids would move toward the outer membrane leaflet. Few examples of transport of endogenous, long-chain phospholipids have been reported. For instance, MDR3 (ABCB4) is a specific transporter of phosphatidylcholine [26] in bile canalicular membranes [27]. As to MDR1 (ABCB1), it has been reported to catalyze the outward transport of shortchain and 7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD)-labeled phospholipids and sphingolipids [26]. A question is whether the lipids are recognized per se, or

50

ARE PLANT MEMBRANE LIPIDS ASYMMETRICALLY DISTRIBUTED?

whether they are just xenobiotic “drugs” that are expelled from the cell for detoxification (see Chapters 11 and 17). 3.2.3â•… Scramblase In the presence of calcium, such a protein reorganizes membrane phospholipids and partially or totally quenches the asymmetry. Two putative scramblases have been proposed [28, 29], but that they really work as scramblases is still discussed [30] (see Chapter 7). 3.2.4â•… Flippases of Biogenic Membranes In the endoplasmic reticulum membrane, a bidirectional movement of phosphatidylcholine [31] and other phospholipids [32] takes place (see Chapter 6). It is protein dependent [33] and, until now, has never been described in plasma membranes. The role of this transporter would be to equilibrate newly synthesized phospholipids, which are released on one membrane face, between the two membrane leaflets.

3.3â•… WHAT ABOUT LIPID ASYMMETRY AND TRANSLOCATION IN PLANT CELL MEMBRANES? Looking at lipid asymmetry is a little bit more challenging with plant cells than with animal cells. In fact, plant cells are surrounded by a cell wall. It is composed of polysaccharides (e.g., cellulose), some phenolic compounds (e.g., lignin), and proteins. As plant cells are living in a hypotonic environment, they are turgescent because of water influx. The cell wall prevents the cells from indefinitely swelling, until bursting. It is also an exoskeleton that gives a plant its shape. Due to its selective permeability for relatively low-molecular-weight compounds, the determination of the plasma membrane asymmetry based on phospholipase attack cannot be performed. On the other hand, the cell wall is not a totally hydrophilic structure, so fluorescent or spin-labeled lipid probes can neither be used according to the well-established protocols. Anyway, it is possible to get rid of the cell wall by enzymatic digestion and obtain so-called protoplasts, which resembles animal cells in accessibility from the outer medium. However, protoplasts are very fragile and stressed structures that have to be manipulated very cautiously. 3.3.1â•… Lipid Asymmetry in the Plasma Membrane (Plasmalemma) Until now, no extensive determination of phospholipid distribution in the plasma membrane of protoplasts has been carried out, to the exception of one study on mung bean (Vigna radiata) protoplasts [34]. The protoplasts were not

WHAT ABOUT LIPID ASYMMETRY AND TRANSLOCATION?

51

labeled by fluorescent Annexin V, showing that all the phosphatidylserine was sequestered in the inner membrane leaflet. This is thus in accordance with the general trend found in animal cell plasma membranes. However, it has to be noted that phosphatidylserine is much less abundant in plant cells than in animal cells [35, 36] and that its asymmetric distribution represents a weak potential perturbation in the mass equilibrium that must exist between the two monolayers. Using phospholipase A2 and right-side-out plasma membrane vesicles isolated from the same protoplasts, it has been found that phosphatidylcholine, phosphatidylethanolamine, and phosphatidic acid were almost symmetrically distributed across the bilayer. That this represents the genuine distribution in intact membranes or a partially perturbed state due to membrane rupture that partially randomizes phospholipids [5] is still an open question. Finally, no sphingomyelin asymmetrical distribution can be described, as this lipid is absent from plants. The asymmetric distribution of phosphatidylserine has also been described in plasma membrane from other plants (tobacco, apple), always on the basis of the quasi-absence of Annexin V binding [37, 38]. Interestingly, this asymmetry is severely perturbed during progression of apoptosis, and phosphatidylserine becomes exposed on the outer leaflet of the plasma membrane [38–40]. 3.3.2â•… Lipid Asymmetry in Intracellular Membranes Organization of lipids in organelle membranes has been studied by enzymatic assays, using various phospholipases, and by chemical labeling of the exposed groups. However, interpretation of data in terms of asymmetric distribution may be difficult due to the inaccessibility of the different lipid species to the enzymes or the chemical. A good example is what was found in mitochondria from castor bean (Ricinus communis) endosperm [41]. Approximately, 45% and 60% of the phosphatidylcholine was inaccessible to phospholipase A2 in outer or inner mitochondrial membranes, respectively. These numbers dropped to 25% and 12% for phosphatidylethanolamine and 25% and zero for phosphatidylinositol (and eventually phosphatidylserine, if present). As location of these subpopulations within the bilayer is unknown, the true distribution of each lipid within the bilayer cannot be rigorously established. Anyway, on the basis of the accessible fraction of each phospholipid class, some asymmetry appeared. For instance, all the phosphatidylinositol of the inner membrane is localized in the inner leaflet, while it is twice more abundant in the outer leaflet than in the inner leaflet of the outer membrane. Phosphatidylethanolamine is always preferentially located in the outer leaflet of the outer membrane (44 vs. 32) or of the inner one (70 vs. 18). As to phosphatidylcholine, it is almost symmetrically distributed. The situation is simpler in glyoxysomal membrane, as the inaccessible fraction is very low [41]. The outer leaflet is the preferential location of phosphatidylethanolamine (67 vs. 16 in the inner leaflet), while phosphatidylcholine

52

ARE PLANT MEMBRANE LIPIDS ASYMMETRICALLY DISTRIBUTED?

and phosphatidylinositolâ•¯+â•¯phosphatidylserine are mainly or almost exclusively, respectively, found in the inner leaflet. Two other organelles, specific of plant cells, have also been studied for the lipid distribution in their membranes. The first one is the vacuole. The vacuole is an organelle whose volume represents approximately 90% of the cell volume. Its main role is to maintain a well-balanced water equilibrium in the cell (it acts as a water reservoir) and to keep the turgescent state. The vacuole is surrounded by a membrane (also known as tonoplast). In sycamore (Acer pseudoplatanus) cells [42], this membrane is made of phospholipids (45%), glycolipids (40%) including sterylglucosides, and neutral lipids (15%). Among the phospholipids, phosphatidylcholine accounts for 30%, phosphatidylethanolamine for 47%, and phosphatidylinositol for 15%. When the membrane was treated by chemical reagents (trinitrobenzene sulfonic acid [TNBS] or fluorescamine) or by phospholipases (either A2, C, or D), the relatively large proportion of inaccessible molecules made it difficult to interpret data in terms of phospholipid asymmetry [43]. The inaccessible molecules represented 10– 20% of phosphatidylcholine and 16–34% of phosphatidylethanolamine. Sequential degradation by phospholipase A2 and phospholipase C reduced these numbers to 10%, or less, allowing one to draw a better scheme of monolayer composition. Phosphatidylcholine is equally distributed across the bilayer, while phosphatidylethanolamine is slightly more present (by 20%) in the outer leaflet. Another membrane lipid, namely phosphatidylinositol, was a poor substrate of phospholipase A2, even in the presence of detergent, most probably because palmitic acid (16:0) represented more than three-quarters of fatty acids. The other organelle is the chloroplast. It is the intracellular compartment where photosynthesis takes place. The photosynthetic complexes are located in the thylakoid membrane, which is arranged in a complicated system of membranous cisternae and lamellae. The thylakoid membrane can account for up to 90% of the membranes found in a green leaf. The chloroplast is surrounded by two bilayers, the envelopes, which isolate the chloroplast interior from the cytoplasm. The outer envelope is freely permeable to small molecules, while the inner envelope behaves as a permeability barrier. Plastid membranes are characterized by the presence, in large amounts, of polar lipids with a galactose-containing head group, named galactolipids (Fig. 3.1). Besides, some phospholipids are present, but in low quantities [44]. MonogalactosylÂ� diacylglycerol (MGDG) and digalactosyldiacylglycerol (DGDG), in a ratio close to 2:1, represent more than 80% of the thylakoid membrane lipids. Phosphatidylglycerol is the major phospholipid of this membrane. The composition of the inner envelope membrane resembles that of the thylakoid one. The outer envelope membrane is rich in DGDG and phosphatidylcholine, MGDG being less present than in the other two membranes. It has to be noted that phosphatidylethanolamine is completely absent from these bilayers. There is a marked asymmetry of phospholipids in the outer envelope: phosphatidylcholine is almost completely (>80%) located in the outer (cytoplasmic) leaflet,

53

WHAT ABOUT LIPID ASYMMETRY AND TRANSLOCATION? OH

OH OH

OH O H2C

CH

O

O

R1

R2

OH OH

O

O

CH2OH

CH2

H2C

MGDG

CH

O

O

R1

R2

O

CH2

CH2SO3–

SQDG

OH OH

OH O H2C O R1

CH

CH2

OH

O CH2

O

O

OH CH2OH

O R2

OH

DGDG

Figure 3.1.â•‡ Structure of the major glycolipids found in plastidial membranes. GalactoÂ� lipids contain one or two galactose molecules in sn3 position of a glycerol backbone. The monogalactosyldiacylglycerol (MGDG) is a 1,2-diacyl-3-O-(β-D-galactopyranosyl)sn-glycerol. The digalactosyldiacylglycerol (DGDG) is a 1,2-diacyl-3-O-(α-Dgalactopyranosyl-(1→6)-O-β-D-galactopyranosyl)-sn-glycerol. Galactolipids contain a high amount of polyunsaturated fatty acids. Very often, linolenic acid (18:3) is present at both the sn1 and sn2 positions of the glycerol backbone. In some plants, a palmitolenic acid (16:3) is present at sn2 position of MGDG (and not at sn1 position). The proportions of C18/C18 and C18/C16 MGDG vary among plants [71]. The much abundant sulfolipid is the sulfoquinovosyldiacylglycerol (SQDG), which is a 1′,2′-diacyl-3′O-(6-deoxy-6-sulfo-α-D-glucopyranosyl)-sn-glycerol. Some of it is dipalmitoylated (16:0/16:0), but the major part contains both 16:0 and 18:3 fatty acids. Interestingly, the 16:0 fatty acid can be linked to sn1 or sn2 position of the glycerol backbone.

where phosphatidylglycerol is totally absent [45]. The distribution of the galactolipids in the envelopes was never extensively determined. As to thylakoids, there is a clear asymmetry of all the lipids. That theses structures were isolated from spinach (Spinacia oleracea) or from oat (Avena sativa), phosphatidylglycerol was found more present on the outer leaflet (60–70%) on the basis of susceptibility to exogenous phospholipase A2 [46, 47]. The transmembrane distribution of galactolipids was also determined by lipase attack of thylakoids purified from leaves of lettuce, pea, barley, oat, spinach, or black nightshade [48, 49]. MGDG was always found enriched in the outer leaflet (53–65%) while

54

ARE PLANT MEMBRANE LIPIDS ASYMMETRICALLY DISTRIBUTED?

DGDG resides predominantly in the inner leaflet (78–90%). A reliable distribution of sulfoquinovosyldiacylglycerol could not be reached, as a significant part (30%) was inaccessible to the acyl hydrolase used for the assay. However, as only 10% of this lipid can be hydrolyzed in the outer leaflet, and even if the nonaccessible fraction is also located in this leaflet, it will be more present in the inner leaflet. The transmembrane galactolipid asymmetry already exists in prothylakoids [49], that is, before incorporation of the chlorophyll–protein complexes. So, the asymmetry has to be established very soon in membrane life. In fact, it appears that it is the lipid synthesis in the inner envelope and the lipid transfer into the thylakoid membrane that is at the origin of the asymmetry [50, 51]. Once the asymmetry is created, it is easy to conceive that galactolipids keep their distribution, as their transverse diffusion rate is certainly very low, due to the bulky, fully hydrated head group. 3.3.3â•… Aminophospholipid Translocase(s) in Plant Cells The phosphatidylserine asymmetry described in plant cell plasma membranes (see above) makes it probable that an aminophospholipid translocase is active in this bilayer. Indeed, such an activity was found in Arabidopsis thaliana, and called aminophospholipid ATPase (ALA)1 [52]. No direct measure of lipid translocation was made on intact cells or protoplasts. But two assays showed that ALA1 was capable to transport phosphatidylserine within a membrane. The first one was to compare NBD-labeled phospholipid internalization in yeasts, either wild-type, or drs2, or ALA1-expressing drs2. Results showed that phosphatidylserine translocation was reestablished in the latter strain, similar to what was described with bovine chromaffin granule ATPase II [20]. The second one was to measure ATP-dependent translocation of NBDphospholipids in vesicles made from microsomal yeast membranes and soybean lipids. Wild-type reconstituted vesicles showed an ATP-dependent inward movement of NBD-phosphatidylserine. This movement was much reduced in vesicles prepared from drs2 yeasts, and restored in vesicles from ALA1-expressing drs2 yeasts. The same type of results was obtained for NBDphosphatidylethanolamine, although lower in amplitude. As to movement of NBD-phosphatidylcholine, it was identical in the three preparations and insensitive to ATP, as expected. Thus, the ALA1 protein possesses all the characteristics of an aminophospholipid translocase present in plant cells. Another function of flippases would be to create an asymmetry in the lipid content of each monolayer by selectively transporting a given phospholipid from one monolayer to the other one. This excess in population can generate a local membrane curvature that, if it is amplified, will lead to vesicle formation and budding [53, 54] (see Chapter 2). Such a role has been experimentally confirmed in yeasts that lack plasma membrane aminophospholipid translocases (here, Dnf1p and Dnf2p) and also present a defect in endocytosis [23]. In addition, aminophospholipid translocase is not limited to plasma membrane, but can also be found in Golgi membranes [55, 56]. Here again, lack of

WHAT ABOUT LIPID ASYMMETRY AND TRANSLOCATION?

55

activity (Drs2p defective mutant cells) blocks the formation of a subclass of post-Golgi secretory vesicles [57], a phenomenon that can be explained by the importance of lipid transport for vesicle budding. A very similar case has been described in plants. Another isoform (ALA3) of P4-ATPase found in A.â•›thaliana is required for secretory vesicle formation [24] and is localized to the Golgi membrane. To complement yeast mutants defective in Drs2p, Dnf1p, and Dnf2p, expression of ALA3 alone is not sufficient. The presence of a protein from the Cdc50p family (here called ALIS1) is also required. Under these conditions, internalization of NBD-labeled phospholipids is reestablished, especially for NBD-phosphatidylethanolamine, rather than for NBDphosphatidylserine. Surprisingly, the co-expressing yeasts are also capable to internalize NBD-phosphatidylcholine at a higher rate than the defective mutant ones. This is not the case when Drs2p is expressed, showing some differences between the apparent orthologs in plants and yeasts. More recently, the requirement of a second subunit associated to ALA peptide to ensure the correct transport function has also been demonstrated for the ALA2 isoform [58]. In Arabidopsis, there are 12 putative aminophospholipid translocases, ALA1 to ALA12 (Table 3.1). They are all around 130â•›kDa in mass (1107–1247 amino acids), and it is predicted that they contain 10 transmembrane helices. Differences appear in the calculated isoelectric point of these isoforms. They can be distributed in three groups: proteins with a pI below 6 (ALA1, 2, 9, 10, and 12), those with a pI comprised between 6 and 7 (ALA5, 8, and 11), and those with a pI above 7 (ALA3, 4, 6, and 7). Comparing the sequences, it appears that there is at least 30% of identity between two isoforms, showing fairly high sequence conservation within the family (Table 3.2). It has to be pointed out that the identity can reach values as high as ca. 90% (ALA4/ ALA5, ALA6/ALA7, ALA10/ALA11). This is more clearly illustrated by the phylogenic tree (Fig. 3.2) showing that eight isoforms are distributed as four pairs, and that ALA1 and ALA2 are the more distant peptides. Putative aminophospholipid translocases have also been found in other plants whose entire genome is known. Twelve flippases could exist in rice (Oryza sativa), five in grape (Vitis vinifera), seven in a moss (Physcomitrella patens), and two in an algae (Chlamydomonas reinhardtii). It is striking that a high percentage of identity (>25%) exists among these different sequences. In fact, this observation does not concern solely the plants; it holds through all reigns. Comparing a random isoform from various animals, plants, and yeast shows that identity is very often comprised between 30% and 40% (Table 3.3). One can wonder why this enzyme is so conserved, at least in sequence. A reason for maintaining phospholipid asymmetry in animal cell plasma membrane can be found by considering what would happen in its absence (see Chapters 7 and 14–16). The presence of phosphatidylserine on the outer membrane leaflet makes the cell surface procoagulant, which can be deleterious for health. Moreover, the presence of external phosphatidylserine is a signal for cell clearance by monocytes/macrophages. In plants, these responses do

56

ARE PLANT MEMBRANE LIPIDS ASYMMETRICALLY DISTRIBUTED?

TABLE 3.1.â•‡ Properties of the 12 Putative Aminophospholipid Translocase Found in Arabidopsis thaliana Genome Ordered locus name Swiss-Prot accession Number of amino acids Mass (Da) pI

ALA1

ALA2

ALA3

ALA4

ALA5

At5g04930 P98204 1158 130,329 5.92

At5g44240 P98205 1107 124,836 5.97

At1g59820 Q9XIE6 1213 137,753 7.84

At1g17500 Q9LNQ4 1216 138,189 7.53

At1g72700 Q9SGG3 1228 139,342 6.89

The mass and pI are calculated from the amino acid sequence of each peptide (without any potential posttranslational modifications).

TABLE 3.2.â•‡ Percentages of Identity between the 12 Isoforms of Putative Aminophospholipid Translocase Found in Arabidopsis thaliana Genome ALA1

ALA2

ALA3

ALA4

ALA5

ALA6

ALA7

31.3

34.2 31.8

35.0 31.4 44.0

35.6 31.1 43.3 87.7

34.6 30.9 42.2 76.6 75.9

35.7 30.7 42.6 77.3 75.8 88.2

0.993

1 0.998 0.99

sp_P98204_ALA1 sp_Q9XIE6_ALA3 0.992 sp_Q9SLK6_ALA6 sp_Q9LVK9_ALA7 0.985 sp_Q9LNQ4_ALA4 sp_Q9SGG3_ALA5 sp_Q9LK90_ALA8 sp_Q9LI83_ALA10 1 sp_Q9SAF5_ALA11 1 sp_P57792_ALA12 0.997 sp_Q9SX33_ALA9 sp_P98205_ALA2

0.5

Figure 3.2.â•‡ Phylogeny of the 12 potential aminophospholipid translocases from Arabidopsis thaliana. Tree was obtained at http://www.phylogeny.fr [72].

57

WHAT ABOUT LIPID ASYMMETRY AND TRANSLOCATION?

ALA6

ALA7

ALA8

ALA9

ALA10

ALA11

ALA12

At1g54280 Q9SLK6 1244 140,682 7.10

At3g13900 Q9LVK9 1247 140,928 7.15

At3g27870 Q9LK90 1189 135,309 6.21

At1g68710 Q9SX33 1200 136,045 5.92

At3g25610 Q9LI83 1202 136,279 5.69

At1g13210 Q9SAF5 1203 136,584 6.54

At1g26130 P57792 1184 133,793 5.91

ALA8

ALA9

ALA10

ALA11

ALA12

36.0 31.9 43.4 55.0 54.9 53.6 53.0

35.2 31.2 44.1 55.8 55.3 54.4 55.7 63.3

34.6 31.0 43.6 56.4 55.3 53.7 54.9 62.8 74.1

33.4 31.5 44.3 55.9 55.3 53.5 54.7 61.2 73.4 87.0

35.7 31.0 42.7 53.6 53.9 53.2 52.8 63.1 80.9 71.6 70.6

ALA1 ALA2 ALA3 ALA4 ALA5 ALA6 ALA7 ALA8 ALA9 ALA10 ALA11 ALA12

TABLE 3.3.â•‡ Percentages of Identity between Some (Putative) Aminophospholipid Translocase from Various Eukaryotic Organisms: Arabidopsis thaliana (ALA1, Swiss-Prot Number P98204), Man (ATP8B1, Swiss-Prot Number O43520), Ox (Chromaffin Granule ATPase II, ATP8A1, Swiss-Prot Number Q29449), Saccharomyces cerevisiae (DRS2, Swiss-Prot Number P39524), Chlamydomonas reinhardtii (ALA2, Swiss-Prot Number A8IVJ6), Rice (Oryza sativa, Swiss-Prot Number Q10LU3), and Grape (Vitis vinifera, Swiss-Prot Number A5BQL2) Arabidopsis

Man

Ox

Yeast

Chlamydomonas

Rice

Grape

35.1

35.4 39.8

32.5 35.3 39.7

31.9 35.4 36.1 33.2

52.0 32.2 33.1 32.6 29.1

35.4 37.0 37.8 34.0 41.1 34.2

Arabidopsis Man Ox Yeast Chlamydomonas Rice Grape

58

ARE PLANT MEMBRANE LIPIDS ASYMMETRICALLY DISTRIBUTED?

not exist, and eventually apoptotic cells have to be kept (to limit pathogen infection, for instance [59]). One of the roles of lipid translocation through the bilayer is to allow vesicle formation from organelles [24], a role that seems common to yeast, animal cells, and plant cells. In accordance with this role, a mutation in ALA3 gene of A.â•›thaliana provokes severe alterations in the polar growth of the cells, leading to reduced primary root growth, longer root hair, and aberrant trichome expansion [60]. Surprisingly, reduced expression of ALA1 in A.â•›thaliana affects the behavior of the plant exposed to low temperature: modified, antisense plants are smaller than wild-type plants when exposed at 8–12°C. No differences in phenotype are seen at normal temperature. It has to be recalled that the drs2 yeast mutant, defective in a protein of the same family, was also temperature sensitive [61]. How could a defect in aminophospholipid transport affect plant growth at low temperature? During adaptation to cold, plants (and in a broader sense, poikilotherms) modify lipid composition of their membranes, both for lipid class and for fatty acid compositions [62]. As fatty acid desaturases and (major part of) lipid synthases are located in the endoplasmic reticulum, lipid trafficking toward the other organelle membranes has to take place in order to make all the adaptive adjustments. So, one can conceive that quenching an aminophospholipid translocase involved in vesicle formation could hamper the correct response to the new conditions. Another possibility relies on the absence of asymmetry of phosphatidylserine (but it represents only few percents of the membrane lipids) and phosphatidylethanolamine in such a mutant. As the fatty acid composition varies with the phospholipid class, the distribution of these fatty acids in each monolayer will be different in the presence or the absence of flippase. It could be possible that favorable lipidic environment of some proteins is not matched at low temperature in the mutants. It is interesting to check whether levels of transcripts of the various ALAs are sensitive to environmental temperature. Mining Genevestigator databases (http://www.genevestigator.com), one can see that cold treatment is not followed by major changes in the transcript levels of the 11 ALAs tested (Fig. 3.3). In fact, the low level of responsiveness can be found in a great variety of situations: abiotic stresses (drought, osmotic stress, oxidative stress, wounding, light quality and intensity) and presence of chemicals (including nutrients) or hormones. The same is true with biotic stresses (attack by pathogens, for instance). Nonetheless, if one gene has to be considered as the most affected, it would be ALA10 whose expression is sometimes increased by a 2–2.5 factor. When the expression is checked against plant organs, again there are only few cases where there is a significant change. Pollen appears to be an organ where transcripts are abundant, at least for some of the ALAs. When genes are clustered as a function of their expression in the different plant organs (Fig. 3.4), it is clear that the tree obtained is different from the phylogenetic tree, indicating that two very close isoforms (e.g., ALA9/Q9SX33 and ALA12/P57792) are not subjected to the same regulations.

59

WHAT ABOUT LIPID ASYMMETRY AND TRANSLOCATION? –2.0

–1.5

–1.0

–0.5 0.0

1.0

1.5

2.0

2.5

# Arrays (exp/ctrl) 3/3 8/4 6/2 6/6 6/6 6/6 6/6

ALA12 ALA7 ALA3 ALA5 ALA1 ALA4 ALA6 ALA2 ALA11 ALA9 ALA8 ALA10

Arabidopsis thaliana (treatment) Stress: cold_2 Stress: cold_3 Stress: cold_4 Stress: cold_green_early Stress: cold_green_late Stress: cold_roots_early Stress: cold_roots_late

0.5

At1g26130 At3g13900 At1g59820 At1g72700 At5g04930 At1g17500 At1g54280 At1g44240 At5g13210 At1g68710 At3g27870 At3g25610

–2.5

Figure 3.3.â•‡ Microarray analysis of ALA expression in Arabidopsis thaliana enduring a cold stress, as obtained from Genevestigator [73]. The bar at the top is giving the color code for the ratio of induction/repression of gene transcription. Color version on the Wiley web site.

Thus, in A.â•›thaliana, the genes coding for the various aminophospholipid translocase behave (almost) as housekeeping genes, a strong indication that this activity could be essential for cell life. Further studies with plants mutated in each of these genes should help in understanding the role of asymmetry. 3.3.4â•… Other Flippases in Plant Cells: ABC Transporters ABC proteins are extremely abundant in plants [63, 64]. For instance, more than 100 members of this family are presumed to exist in A.â•›thaliana. No in vitro study of lipid translocation by one of these proteins has been performed, so one can just hypothesize that there will certainly be some flippases in the family. Examples are the peroxysomal PAX1/CTS/PED3 protein that could transport fatty acids or acyl-CoAs [65–67] and the plastidial TGD1, 2, 3 multicomponent transporter proposed to mediate the transmembrane movement of polar lipids [68, 69]. 3.3.5â•… Other Flippases in Plant Cells: Endoplasmic Reticulum Flippase Very recently, strong experimental data were presented in favor of the existence of a flippase in endoplasmic reticulum membrane [70]. In proteoliposomes reconstituted from solubilized endoplasmic reticulum from spinach leaves, there is an ATP-independent translocation of phosphatidylcholine and phosphatidylethanolamine. This translocation is sensitive to trypsine and chemical reagents. Based on differential inhibition of translocation of each phospholipid, it is highly probable that more than one flippase is present in

ARE PLANT MEMBRANE LIPIDS ASYMMETRICALLY DISTRIBUTED? Q9SGG3 Q9LNQ4 P98205 Q9XIE6 P98204 Q9SX33 Q9LK90 Q9LI83 Q9SAF5 P57792 Q9LVK9 Q9SLK6

60

callus cell suspension seedling cotyledons hypocotyl radicle imbibed seed inflorescence flower carpel ovary stigma petal sepal stamen pollen pedicel silique seed embryo endosperm stem node shoot apex cauline leaf rosette juvenile leaf adult leaf petiole senescent leaf hypocotyl xylem cork roots lateral root root hair zone root tip elongation zone endodermis endodermis+cortex epid. atrichoblasts lateral root cap stele

References

61

Figure 3.4.â•‡ Clustering of the 12 genes coding for aminophospholipid translocase in Arabidopsis thaliana as a function of their expression in the different organs. The darker the color, the higher is the expression level. This graph was obtained from Genevestigator. Color version on the Wiley web site.

the original membrane. As in other organisms, no molecular identification of the corresponding peptides has been done. In conclusion, plants could represent a good model to shed more light on phospholipid asymmetry and the proteins responsible for it. A genetic (mutant) approach is rather easy with a model such as Arabidopsis, as many knockout and knockdown mutants are readily available, and as stable transformation of the plant is well documented. However, the multiplicity of the isoforms, and a possible redundancy, will make the physiological studies quite tedious. ABBREVIATIONS ABC ALA DGDG MGDG NBD TNBS

ATP-binding cassette aminophospholipid ATPase digalactosyldiacylglycerol monogalactosyldiacylglycerol 7-nitrobenz-2-oxa-1,3-diazol-4-yl trinitrobenzene sulfonic acid

REFERENCES â•‡ 1â•… â•‡ 2â•… â•‡ 3â•… â•‡ 4â•…

â•‡ 5â•… â•‡ 6â•… â•‡ 7â•… â•‡ 8â•… â•‡ 9â•…

10â•…

P. F. Devaux, Annu. Rev. Biophys. Biomol. Struct. 1992, 21, 417–439. J. A. F. Op den Kamp, Annu. Rev. Biochem. 1979, 48, 47–71. A. Zachowski, Biochem. J. 1993, 294, 1–14. J. Florin-Christensen, C. E. Suarez, M. Florin-Christensen, M. Wainszelbaum, W. C. Brown, T. F. McElwain, G. H. Palmer, Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 7736–7741. S. L. Schrier, A. Zachowski, P. Hervé, J. C. Kader, P. F. Devaux, Biochim. Biophys. Acta 1992, 1105, 170–176. P. Comfurius, E. M. Bevers, R. F. A. Zwaal, Methods Mol. Biol. 1994, 27, 131–142. J. F. Tait, D. Gibson, K. Fujikawa, J. Biol. Chem. 1989, 264, 7944–7949. M. van Engeland, L. J. W. Nieland, F. C. S. Ramaekers, B. Schutte, C. P. M. Reutelingsperger, Cytometry 1998, 31, 1–9. C. P. M. Reutelingsperger, E. Dumont, P. W. Thimister, H. van Genderen, H. Kenis, S. van de Eijnde, G. Heidendal, L. Hofstra, J. Immunol. Methods 2002, 265, 123–132. A. Marconescu, P. E. Thorpe, Biochim. Biophys. Acta 2008, 1778, 2217–2224.

62

ARE PLANT MEMBRANE LIPIDS ASYMMETRICALLY DISTRIBUTED?

11â•… P. Natarajan, J. Wang, Z. Hua, T. R. Graham, Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 10614–10619. 12â•… C. Subra, K. Laulagnier, B. Perret, M. Record, Biochimie 2007, 89, 205–212. 13â•… P. F. Devaux, I. Lopez-Montero, S. Bryde, Chem. Phys. Lipids 2006, 141, 119–132. 14â•… G. Lenoir, P. Williamson, J. C. M. Holthuis, Curr. Opin. Chem. Biol. 2007, 11, 654–661. 15â•… T. Pomorski, A. K. Menon, Cell. Mol. Life Sci. 2006, 63, 2908–2921. 16â•… T. Pomorski, J. C. M. Holthuis, A. Herrmann, G. van Meer, J. Cell Sci. 2004, 117, 805–813. 17â•… A. Tannert, A. Pohl, T. Pomorski, A. Herrmann, Int. J. Antimicrob. Agents 2003, 22, 177–187. 18â•… D. L. Daleke, J. V. Lyles, Biochim. Biophys. Acta 2000, 1486, 108–127. 19â•… A. Zachowski, J. P. Henry, P. F. Devaux, Nature 1989, 340, 75–76. 20â•… X. Tang, M. S. Halleck, R. A. Schlegel, P. Williamson, Science 1996, 272, 1495–1497. 21â•… L. R. Poulsen, R. L. Lopez-Marquès, M. G. Palmgren, Cell. Mol. Life Sci. 2008, 65, 3119–3125. 22â•… Z. Hua, P. Fatheddin, T. R. Graham, Mol. Biol. Cell 2002, 13, 3162–3177. 23â•… T. Pomorski, R. Lombardi, H. Riezman, P. F. Devaux, G. van Meer, J. C. M. Holthuis, Mol. Biol. Cell 2003, 14, 1240–1254. 24â•… L. R. Poulsen, R. L. Lopez-Marquès, S. C. McDowell, J. Okkeri, D. Licht, A. Schulz, T. Pomorski, J. F. Harper, M. G. Palmgren, Plant Cell 2008, 20, 658–676. 25â•… K. Saito, K. Fujimura-Kamada, N. Furuta, U. Kato, M. Umeda, K. Tanaka, Mol. Biol. Cell 2004, 15, 3418–3432. 26â•… A. van Helvoort, A. J. Smith, H. Sprong, I. Fritzsche, A. H. Schinkel, P. Borst, G. van Meer, Cell 1996, 87, 507–517. 27â•… A. T. Nies, Z. Gatmaitan, I. M. Arias, J. Lipid Res. 1996, 37, 1125–1136. 28â•… F. Bassé, J. G. Stout, P. J. Sims, T. Wiedmer, J. Biol. Chem. 1996, 271, 17205–17210. 29â•… P. Comfurius, P. Williamson, E. F. Smeets, R. A. Schlegel, E. M. Bevers, R. F. A. Zwaal, Biochemistry 1996, 35, 7631–7634. 30â•… S. K. Sahu, S. N. Gummadi, N. Manoj, G. K. Aradhyam, Arch. Biochem. Biophys. 2007, 462, 103–114. 31â•… W. R. Bishop, R. M. Bell, Cell 1985, 42, 51–60. 32â•… X. Buton, G. Morrot, P. Fellmann, M. Seigneuret, J. Biol. Chem. 1996, 271, 6651–6657. 33â•… S. N. Gummadi, A. K. Menon, J. Biol. Chem. 2002, 277, 25337–25343. 34â•… Y. Takeda, K. Kasamo, Biochim. Biophys. Acta 2001, 1513, 38–48. 35â•… P. Moreau, J. J. Bessoule, S. Mongrand, E. Testet, P. Vincent, C. Cassagne, Prog. Lipid Res. 1998, 37, 371–391. 36â•… E. Delhaize, D. M. Hebb, K. D. Richards, J. M. Lin, P. R. Ryan, R. C. Gardner, J. Biol. Chem. 1999, 274, 7082–7088. 37â•… I. E. W. O’Brien, B. C. Baguley, B. G. Murray, B. A. M. Morris, I. B. Ferguson, Plant J. 1998, 13, 803–814. 38â•… C. J. Xu, K. S. Chen, I. B. Ferguson, J. Zhejiang Univ. Sci. 2004, 5, 137–143.

References

63

39â•… I. E. W. O’Brien, C. P. M. Reutelingsperger, K. M. Holdaway, Cytometry 1997, 29, 28–33. 40â•… S. B. Ning, Y. C. Song, P. van Damme, Electrophoresis 2002, 23, 2096–2102. 41â•… T. M. Cheesbrough, T. S. Moore, Jr., Plant Physiol. 1980, 65, 1076–1080. 42â•… E. Tavernier, D. Lê Quôc, K. Lê Quôc, Biochim. Biophys. Acta 1993, 167, 242–247. 43â•… E. Tavernier, A. Pugin, Biochimie 1995, 77, 174–181. 44â•… M. A. Block, R. Douce, J. Joyard, N. Rolland, Photosyn. Res. 2007, 92, 225–244. 45â•… A. J. Dorne, J. Joyard, M. A. Block, R. Douce, J. Cell Biol. 1985, 100, 1690–1697. 46â•… S. Duchêne, J. Smutny, P. A. Siegenthaler, Biochim. Biophys. Acta 2000, 1463, 115–120. 47â•… P. A. Siegenthaler, C. Giroud, FEBS Lett. 1986, 201, 215–220. 48â•… A. Rawyler, M. D. Unitt, C. Giroud, H. Davies, J. P. Mayor, J. L. Harwood, P. A. Siegenthaler, Photosyn. Res. 1987, 11, 3–13. 49â•… C. Giroud, P. A. Siegenthaler, Plant Physiol. 1988, 88, 412–417. 50â•… A. Rawyler, M. Meylan, P. A. Siegenthaler, Biochim. Biophys. Acta 1992, 1104, 331–341. 51â•… A. Rawyler, M. Meylan-Bettex, P. A. Siegenthaler, Biochim. Biophys. Acta 1995, 1233, 123–133. 52â•… E. Gomès, M. K. Jakobsen, K. B. Axelsen, M. Geisler, M. G. Palmgren, Plant Cell 2000, 12, 2441–2453. 53â•… P. F. Devaux, Biochimie 2000, 82, 497–509. 54â•… P. F. Devaux, A. Herrmann, N. Ohlwein, M. M. Kozlov, Biochim. Biophys. Acta 2008, 1778, 1591–1600. 55â•… C. Y. Chen, M. F. Ingram, P. H. Rosal, T. R. Graham, J. Cell Biol. 1999, 147, 1223–1236. 56â•… N. Alder-Baerens, Q. Lisman, L. Luong, T. Pomorski, J. C. M. Holthuis, Mol. Biol. Cell. 2006, 17, 1632–1642. 57â•… W. E. Gall, N. C. Geething, Z. Hua, M. F. Ingram, K. Liu, S. I. Chen, T. R. Graham, Curr. Biol. 2002, 12, 1623–1627. 58â•… R. L. Lopez-Marqués, L. R. Poulsen, S. Hanish, K. Meffert, M. J. Buch-Pedersen, M. K. Jacobsen, T. G. Pomorski, M. G. Palmgren, Mol. Cell. Biol. 2010, 21, 791–801. 59â•… L. A. Mur, P. Kenton, A. J. Lloyd, H. Ougham, E. Prats, J. Exp. Bot. 2008, 59, 501–520. 60â•… X. Zhang, D. G. Oppenheimer, Plant J. 2009, 60, 195–206. 61â•… T. Ripmaster, G. Vaughn, J. L. Woolford, Mol. Cell. Biol. 1993, 13, 7901–7912. 62â•… E. Ruelland, M. N. Vaultier, A. Zachowski, V. Hurry, Adv. Bot. Res. 2009, 49, 35–150. 63â•… P. A. Rea, Annu. Rev. Plant Biol. 2007, 58, 347–375. 64â•… P. J. Verrier, D. Bird, B. Burla, E. Dassa, C. Forestier, M. Geisler, M. Klein, Ü. Kolukisaoglu, Y. Lee, E. Martinoia, A. Murphy, P. A. Rea, L. Samuels, B. Schulz, E. J. Spalding, K. Yazaki, F. L. Theodoulou, Trends Plant Sci. 2008, 13, 151–159. 65â•… M. Hayashi, K. Nito, R. Takei-Hoshi, M. Yagi, M. Kondo, A. Suenaga, T. Yamaya, M. Nishimura, Plant Cell Physiol. 2002, 43, 1–11.

64

ARE PLANT MEMBRANE LIPIDS ASYMMETRICALLY DISTRIBUTED?

66â•… F. L. Theodoulou, K. Job, S. P. Slocombe, S. Footitt, M. Holdsworth, A. Baker, T. R. Larson, I. A. Graham, Plant Physiol. 2005, 137, 835–840. 67â•… B. K. Zolman, I. D. Silva, B. Bartel, Plant Physiol. 2001, 127, 1266–1278. 68â•… B. Lu, C. Xu, K. Awai, A. D. Jones, C. Benning, J. Biol. Chem. 2007, 282, 35945–35953. 69â•… C. Xu, J. Fan, J. Froehlich, K. Awai, C. Benning, Plant Cell 2005, 17, 3094–3110. 70â•… S. K. Sahu, S. N. Gummadi, Biochemistry 2008, 47, 10481–10490. 71â•… E. Heinz, In Lipids and Lipid Polymers in Higher Plants, M. Tevini and H. K. Lichtenthaler, eds. Springer-Verlag, Berlin, 1977, 102–120. 72â•… A. Dereeper, V. Guignon, G. Blanc, S. Buffet, F. Chevenet, J. F. Dufayard, S. Guindon, V. Lefort, M. Lescot, J. M. Claverie, O. Gascuel, Nucleic Acid Res. 2008, 36, 465–469. 73â•… T. Hruz, O. Laule, G. Szabo, F. Wessendorp, S. Bleuler, L. Oertle, P. Widmayer, W. Gruissem, P. Zimmermann, Adv. Bioinformatics 2008, 2008, 420747.

4 SPHINGOLIPID ASYMMETRY AND TRANSMEMBRANE TRANSLOCATION IN MAMMALIAN CELLS Gerrit van Meer Faculty of Science, Utrecht University, Utrecht, The Netherlands

Sylvia Neumann Department of Cell Biology, Scripps Research Institute, La Jolla, CA

Per Haberkant EMBL, Heidelberg, Germany

4.1 INTRODUCTION Sphingolipids are a typical feature of eukaryotic cells, and indeed, they have been found to fulfill a number of intra- and intercellular functions that are specific for eukaryotes. Membrane sphingolipids are organized in specialized membrane domains [1] that are involved in the sorting of membrane proteins and lipids along the cellular vesicular transport pathways. In addition, the domains have been invoked in various types of signaling events, like the formation of the T-cell receptor complex [2] and the formation of cell–cell signaling domains [3]. On the other hand, individual sphingolipids act as lipid second messengers, the clearest examples being sphingosine-1-phosphate [4] and ceramide [5]. Sphingolipids act at discrete locations, and they are synthesized and degraded at defined locations. These are not always on the same side of the membrane, which necessitates transmembrane transport. The sites of transmembrane translocation, the molecular mechanism, and its possible regulation are the topic of the present chapter (Fig. 4.1). Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

65

66

Cer

Cer

?

SM

SM

DHSph1P

DHSph

DHSph

Lysosome

DHCer

DHCer Cer

Cer

Sph

SphH+

Sph

Sph

Chol

Cer

SM

Cer

Sph

Sph

Chol

NPC1

GlcCer

Endosome

Lysosome

GlcCer

GlcCer GlcCer GlcCer

FAPP2

GlcCer

GlcCer

GM3

GlcCer

GlcCer

GlcCer

GlcCer

SM

GlcCer SM GM3

ER

Golgi

Cer

PM

Cer

Cer

Cer

Cer

CERT

Figure 4.1.â•‡ Transmembrane translocation of sphingolipids. For a detailed explanation of the various pathways, see the text. SM, sphingomyelin; PM, plasma membrane. Color version on the Wiley web site.

DHSph

3-keto DHSph

palmitoylCoA + Ser

DHSph

Sphingosine, Sphingosine-1-Phosphate, and Ceramide

67

4.2 SPHINGOSINE, SPHINGOSINE-1-PHOSPHATE, AND CERAMIDE The first committed step in sphingolipid synthesis is the condensation of serine and palmitoyl-CoA to 3-ketosphinganine. Its reduction yields sphinganine, which serves as a substrate for various (dihydro)ceramide synthases [6]. Most of the dihydroceramide is subsequently desaturated to ceramide by dihydroceramide desaturase [7]. Actually, most ceramide may be synthesized in the endoplasmic reticulum (ER) directly from sphingosine that is derived from sphingolipid breakdown in the lysosomes, via the salvage pathway [8]. Apart from monomeric transport from the lysosomes to the ER, which is not a problem for the relatively hydrophilic molecule, sphingosine that is generated by ceramidase in the lysosomal lumen must translocate across the lysosomal membrane. It has been reported that sphingosine, which has a pKa of 6.7–7.7 in TX micelles, actually has a pKa of 9.1 in phosphatidylserine-containing membranes [9]. From this, it has been argued that the positively charged sphingosine cannot spontaneously flip out of the lysosome but needs a translocator. Because sphingosine was the first lipid to accumulate in lysosomes upon the induction of a Niemann-Pick type C (NPC) disease phenotype, it has been suggested that NPC1 may be the relevant sphingosine transporter [10]. NPC1 belongs to the heavy metal resistance/nodulation/cell division (RND) permeases, a family of bacterial transporters that are driven by the proton motive force [11]. Instead of transporting a proton (or besides transporting a proton), NPC1 may transport protonated sphingosine toward the cytosolic side. This leaves the interesting possibility that NPC1 countertransports a different substrate in the opposite direction toward the lumen or lumenal leaflet of the lysosome. The fact that defects in NPC2, a soluble cholesterol transporter in the lysosomal lumen, result in the same phenotype as defects in NPC1, which has a cholesterol-binding domain next to a sterol-sensing domain [12], strongly suggests that cholesterol transport out of the lysosome is at least one other function controlled by NPC1 [13]. When added from the outside of the cell, sphingosine must cross the plasma membrane and be transported to the ER to be used in cellular synthesis. Unexpectedly, in yeast, the utilization of exogenous dihydrosphingosine was found to involve phosphorylation and dephosphorylation of the sphingoid base on the lumenal side of the ER [14, 15]. It is unclear whether this has anything to do with translocation of the sphingoid base to the lumenal side of the ER membrane, where the active center of the ceramide synthase is thought to be located [16]: Since all lipids tested have been found to be translocated quickly across the ER membrane [17] by a so far unknown translocator [18, 19] (see Chapter 6), it might be expected that also sphingoid bases freely equilibrate between the two leaflets of the ER membrane. Sphingosine-1-phosphate is not only a metabolic intermediate that can be dephosphorylated by a number of phosphatases, or the final step in sphingolipid degradation by the sphingosine-1-phosphate lyase in the ER membrane

68

TRANSMEMBRANE TRANSLOCATION OF SPHINGOLIPIDS

[20]; it is also an important lipid messenger with targets in the cytosol (unidentified) and targets on the outside of cells: the sphingosine-1-phosphate receptors. Because sphingosine-1-phosphate is generated by the cytosolic sphingosine kinases SphK1 and SphK2, it must translocate across the plasma membrane to reach the receptor. It has been reported that this translocation is mediated by the multidrug resistance-related protein MRP1, that is the ATP-binding cassette (ABC) transporter ABCC1 [21], and ABCG2 [22]. Their activity may be regulated by estradiol.

4.3 CERAMIDE After the synthesis of dihydroceramide in the ER, it is desaturated to ceramide or, depending on the cell type, converted to a smaller or larger extent to 4-hydroxysphinganine-containing ceramide (phytoceramide) by C-4hydroxylase, a side activity of a delta 4-desaturase [7, 23, 24]. Dihydroceramide may be produced on the lumenal side of the membrane [16], which also requires that acyl-CoA is translocated. In contrast, the desaturase acts on the cytosolic surface [25]. Translocation of ceramide across the ER is probably fast, because general lipid translocation across this membrane is fast and ceramide has a much smaller polar head group than most other lipids tested. In contrast to regular membrane phospholipids, ceramide spontaneously translocates across model membranes with half-times of less than a minute for natural ceramide [26] and 20 minutes for a fluorescent analog [27]. In contrast to sphingosine, ceramide that is produced in the lysosome is unable to leave the lysosome and mix with the biosynthetic pool of ceramide in the ER: mutations that inactivate the acid ceramidase result in the sphingolipid storage disorder Farber’s disease [28]. In parallel to the observation that mutations in the lumenal cholesterol transport protein NPC2 prevent export of cholesterol from the lysosome, it is most likely the inability of intact ceramide to be transported from the intralumenal vesicles to the limiting membrane of the lysosome that blocks its exit.

4.4 GLYCOSPHINGOLIPIDS In a number of cell types, ceramide is partially converted to galactosylceramide (GalCer) by the ceramide galactosyltransferase on the lumenal aspect of the ER [29], whereas in all cells, ceramide reaches the cytosolic surface of the Golgi where glucosylceramide (GlcCer) is formed. Spin-labeled GlcCer rapidly crossed isolated Golgi membranes in one study [30], and also a GlcCer analog with a short fatty acyl chain, C6-NBD-GlcCer, or with two shortened fatty tails, C8C8-GlcCer, did move across the Golgi membrane [31–33]. This translocation was found to be sensitive to inhibitors of the multidrug trans-

Glycosphingolipids

69

porter MDR1, the ABC transporter ABCB1 [33], which is in line with earlier studies showing that ABCB1 and ABCC1 can translocate these lipid analogs across the plasma membrane [34, 35]. Indirect studies also suggest that natural GlcCer can be translocated across Golgi membranes [36–38], and it has been suggested that the cytosolic glycolipid-binding protein FAPP2, which binds to phosphatidylinositol-4-phosphate on the Golgi, plays a role in this translocation. However, such a translocation of natural GlcCer was not found in a direct experiment. Instead, it was proposed that GlcCer is transported in a retrograde fashion to the ER where it will then translocate spontaneously like all other lipids [33]. In this study, we also found evidence that the pools of natural GlcCer mixed with those of GalCer, and that these lipids can be translocated across a post-Golgi membrane. The latter was based on the observation that newly synthesized GlcCer and GalCer reach the outside of the plasma membrane in the presence of brefeldin A, a drug that inhibited vesicular transport as defined by the complex glycolipid sialyllactosylceramide, GM3 [33]. A special case of GlcCer transport is found in keratinocytes, where GlcCer is stored with other lipids in lamellar granules destined for secretion and extracellular hydrolysis to form the intercellular ceramide lamellae of the skin. This process was found to be interrupted in ichthyosis patients due to mutations in the ABC transporter ABCA12 [39–41], which has therefore been proposed to be a GlcCer translocator or “extruder” [42]. Unfortunately, only a few assays are available to study the transbilayer organization of GlcCer [33, 43], and they have not yet been applied to the keratinocyte system. Interestingly, ceramide regulates ABCA12 expression in the skin via peroxisome proliferatoractivated receptor delta [44]. On the lumenal surface of the Golgi, GlcCer is converted to complex glycosphingolipids. By a sorting event that most likely involves their lateral concentration in microdomains [1], the glycosphingolipids are transported on the inside of carrier vesicles to the plasma membrane where they are exposed on the surface. From there, endocytosis cycles them through the lumen of the endosomes, whereby during each cycle, some glycosphingolipids end up in the lysosomes where they are degraded. A number of studies have reported the presence of complex glycosphingolipids on structures like the cytoskeleton where they might be exposed to the cytosol [45–47]. Unfortunately, these localization studies often involved the use of detergents, which may have induced displacement from their original site of residence [48]. Apart from that, the methodology to study the sidedness of (sphingo)lipids is generally not straightforward and has many potential pitfalls [49, 50]. Whereas under normal conditions complex glycosphingolipids are not found in membranes that are not connected to the exo- and endocytotic pathways like the mitochondria [51, 52], it has been reported that specific glycosphingolipids, for example, GD3, do reach the mitochondria [53–56]. For this, the glycolipid must have translocated across a membrane of the exo/endocytotic system. Where and how this occurs is presently unclear.

70

TRANSMEMBRANE TRANSLOCATION OF SPHINGOLIPIDS

4.5 SPHINGOMYELIN Ceramide destined for the synthesis of sphingomyelin reaches the Golgi not via vesicular transport but via the ceramide transfer protein CERT [57], which, like the glycolipid-binding FAPP2, binds to phosphatidylinositol-4-phosphate on the trans-Golgi membrane. There, ceramide has to cross the Golgi membrane to reach the sphingomyelin synthase SMS1, which is situated with its active center on the lumenal side [58, 59] of the trans-Golgi [33]. It is not fully clear to what extent sphingomyelin is limited to the noncytosolic surface of the Golgi, plasma membrane, and endosomes. Early work indicated that some 10%–20% of the sphingomyelin was situated on the cytosolic surface of the plasma membrane [60, 61], and of chromaffin granules [62]. Unfortunately, essentially only a single method has been applied so far and that is sphingomyelinase C, which is prone to artifacts [63]. Contradictory results have been reported on the effect of brefeldin A on the transport of newly synthesized sphingomyelin to the cell surface, the outside of the plasma membrane. In CHO cells, brefeldin A essentially blocked sphingomyelin transport [64], which implies that (1) sphingomyelin could not translocate across the membrane of the mixed ER–Golgi compartment induced by brefeldin A [65], which is unlikely in view of the fact that GalCer under those conditions does have access to the cytosolic surface and does make it to the cell surface [33]; (2) sphingomyelin cannot cross the aqueous space between ER–Golgi and plasma membrane; and/or (3) sphingomyelin cannot flip across the plasma membrane. Others [66] reported that brefeldin A did not interfere with transport of newly synthesized sphingomyelin to the surface of hepatocytes and claimed that there was a brefeldininsensitive vesicular pathway from the ER–Golgi to the plasma membrane. Although it is possible that sphingomyelin traveled across the aqueous cytosol via a transfer protein (e.g., PITPß [67] ) and was translocated outward across the plasma membrane, it is more likely that in these cells, the sphingomyelin synthase did not redistribute to the ER but remained in a trans-Golgi/ endosome compartment that maintained vesicular traffic to the plasma membrane [68]. During the stress-induced activation of the neutral sphingomyelinase, a large fraction of the plasma membrane sphingomyelin can be degraded. Because the neutral sphingomyelinase has its active center on the cytosolic side [69, 70], either sphingomyelin must be translocated to the cytosolic leaflet of the membrane [71], or sphingomyelin was already present in the cytosolic leaflet and the enzyme itself was activated. Whether this implies the activation of a nonspecific scrambling mechanism [72] or a sphingomyelin-specific process is unclear. Along similar lines, the finding that sphingomyelin can be degraded by a sphingomyelinase that was targeted to the mitochondria has been interpreted to indicate that sphingomyelin reaches this organelle [73, 74], which must involve a translocation step toward the cytosol.

References

71

4.6 FUTURE PERSPECTIVES It is clear that we still have a lack of understanding of a number of aspects of the organization and trafficking of sphingolipids. Notably, the transmembrane distribution and translocation of the basic sphingolipids sphingosine, sphingomyelin, and the monohexosylceramides are enigmatic. This is mostly due to a lack of appropriate technology. It will be a challenge to unravel these issues and to identify the translocators involved. From the fact that these lipids are essential for mammalian survival, it can be predicted that such findings will have their impact on the identification and cure of lipid-linked disorders as has already been shown for the ichthyoses and lipid storage diseases mentioned in the chapter. ABBREVIATIONS C6-NBD- N-6-NBD-aminohexanoylGalCer galactosylceramide GlcCer glucosylceramide REFERENCES â•‡ 1â•…G. van Meer, D. R. Voelker, G. W. Feigenson, Nat. Rev. Mol. Cell Biol. 2008, 9, 112–124. â•‡ 2â•… T. Harder, C. Rentero, T. Zech, K. Gaus, Curr. Opin. Immunol. 2007, 19, 470–475. â•‡ 3â•… A. Regina Todeschini, S. I. Hakomori, Biochim. Biophys. Acta 2008, 1780, 421–433. â•‡ 4â•… S. Pyne, S. C. Lee, J. Long, N. J. Pyne, Cell. Signal. 2009, 21, 14–21. â•‡ 5â•…Y. A. Hannun, L. M. Obeid, Nat. Rev. Mol. Cell Biol. 2008, 9, 139–150. â•‡ 6â•…Y. Pewzner-Jung, S. Ben-Dor, A. H. Futerman, J. Biol. Chem. 2006, 281, 25001–25005. â•‡ 7â•… P. Ternes, S. Franke, U. Zahringer, P. Sperling, E. Heinz, J. Biol. Chem. 2002, 277, 25512–25518. â•‡ 8â•… K. Kitatani, J. Idkowiak-Baldys, Y. A. Hannun, Cell. Signal. 2008, 20, 1010–1018. â•‡ 9â•… F. Lopez-Garcia, V. Micol, J. Villalain, J. C. Gomez-Fernandez, Biochim. Biophys. Acta 1993, 1153, 1–8. 10â•… E. Lloyd-Evans, A. J. Morgan, X. He, D. A. Smith, E. Elliot-Smith, D. J. Sillence, G. C. Churchill, E. H. Schuchman, A. Galione, F. M. Platt, Nat. Med. 2008, 14, 1247–1255. 11â•… T. T. Tseng, K. S. Gratwick, J. Kollman, D. Park, D. H. Nies, A. Goffeau, M. H. Saier, Jr., J. Mol. Microbiol. Biotechnol. 1999, 1, 107–125. 12â•…R. E. Infante, M. L. Wang, A. Radhakrishnan, H. J. Kwon, M. S. Brown, J. L. Goldstein, Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 15287–15292.

72

TRANSMEMBRANE TRANSLOCATION OF SPHINGOLIPIDS

13â•… D. E. Sleat, J. A. Wiseman, M. El-Banna, S. M. Price, L. Verot, M. M. Shen, G. S. Tint, M. T. Vanier, S. U. Walkley, P. Lobel, Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5886–5891. 14â•… K. Funato, R. Lombardi, B. Vallee, H. Riezman, J. Biol. Chem. 2003, 278, 7325–7334. 15â•… A. Kihara, T. Sano, S. Iwaki, Y. Igarashi, Genes Cells 2003, 8, 525–535. 16â•… B. Vallée, H. Riezman, EMBO J. 2005, 24, 730–741. 17â•… A. Herrmann, A. Zachowski, P. F. Devaux, Biochemistry 1990, 29, 2023–2027. 18â•… A. Papadopulos, S. Vehring, I. Lopez-Montero, L. Kutschenko, M. Stockl, P. F. Devaux, M. Kozlov, T. Pomorski, A. Herrmann, J. Biol. Chem. 2007, 282, 15559–15568. 19â•… S. Sanyal, C. G. Frank, A. K. Menon, Biochemistry 2008, 47, 7937–7946. 20â•… H. Fyrst, J. D. Saba, Biochim. Biophys. Acta 2008, 1781, 448–458. 21â•… P. Mitra, C. A. Oskeritzian, S. G. Payne, M. A. Beaven, S. Milstien, S. Spiegel, Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 16394–16399. 22â•… K. Takabe, R. H. Kim, J. C. Allegood, P. Mitra, S. Ramachandran, M. Nagahashi, K. B. Harikumar, N. C. Hait, S. Milstien, S. Spiegel, J. Biol. Chem. 2010, 285, 10477–10486. 23â•… F. Omae, M. Miyazaki, A. Enomoto, A. Suzuki, FEBS Lett. 2004, 576, 63–67. 24â•… F. Omae, M. Miyazaki, A. Enomoto, M. Suzuki, Y. Suzuki, A. Suzuki, Biochem. J. 2004, 379, 687–695. 25â•… C. Michel, G. van Echten-Deckert, FEBS Lett. 1997, 416, 153–155. 26â•… I. López-Montero, N. Rodriguez, S. Cribier, A. Pohl, M. Velez, P. F. Devaux, J. Biol. Chem. 2005, 280, 25811–25819. 27â•… J. Bai, R. E. Pagano, Biochemistry 1997, 36, 8840–8848. 28â•… J. H. Park, E. H. Schuchman, Biochim. Biophys. Acta 2006, 1758, 2133–2138. 29â•… H. Sprong, B. Kruithof, R. Leijendekker, J. W. Slot, G. van Meer, P. van der Sluijs, J. Biol. Chem. 1998, 273, 25880–25888. 30â•… X. Buton, P. Herve, J. Kubelt, A. Tannert, K. N. Burger, P. Fellmann, P. Muller, A. Herrmann, M. Seigneuret, P. F. Devaux, Biochemistry 2002, 41, 13106–13115. 31â•… K. N. Burger, P. van der Bijl, G. van Meer, J. Cell Biol. 1996, 133, 15–28. 32â•… H. Lannert, K. Gorgas, I. Meißner, F. T. Wieland, D. Jeckel, J. Biol. Chem. 1998, 273, 2939–2946. 33â•… D. Halter, S. Neumann, S. M. van Dijk, J. Wolthoorn, A. M. de Maziere, O. V. Vieira, P. Mattjus, J. Klumperman, G. van Meer, H. Sprong, J. Cell Biol. 2007, 179, 101–115. 34â•… A. van Helvoort, A. J. Smith, H. Sprong, I. Fritzsche, A. H. Schinkel, P. Borst, G. van Meer, Cell 1996, 87, 507–517. 35â•…R. J. Raggers, A. van Helvoort, R. Evers, G. van Meer, J. Cell Sci. 1999, 112, 415–422. 36â•… P. Lala, S. Ito, C. A. Lingwood, J. Biol. Chem. 2000, 275, 6246–6251. 37â•… M. F. De Rosa, D. Sillence, C. Ackerley, C. Lingwood, J. Biol. Chem. 2004, 279, 7867–7876.

References

73

38â•…G. D’Angelo, E. Polishchuk, G. Di Tullio, M. Santoro, A. Di Campli, A. Godi, G. West, J. Bielawski, C. C. Chuang, A. C. van der Spoel, F. M. Platt, Y. A. Hannun, R. Polishchuk, P. Mattjus, M. A. De Matteis, Nature 2007, 449, 62–67. 39â•… C. Lefevre, S. Audebert, F. Jobard, B. Bouadjar, H. Lakhdar, O. BoughdeneStambouli, C. Blanchet-Bardon, R. Heilig, M. Foglio, J. Weissenbach, M. Lathrop, J. F. Prud’homme, J. Fischer, Hum. Mol. Genet. 2003, 12, 2369–2378. 40â•… M. Akiyama, Y. Sugiyama-Nakagiri, K. Sakai, J. R. McMillan, M. Goto, K. Arita, Y. Tsuji-Abe, N. Tabata, K. Matsuoka, R. Sasaki, D. Sawamura, H. Shimizu, J. Clin. Invest. 2005, 115, 1777–1784. 41â•… D. P. Kelsell, E. E. Norgett, H. Unsworth, M. T. Teh, T. Cullup, C. A. Mein, P. J. Dopping-Hepenstal, B. A. Dale, G. Tadini, P. Fleckman, K. G. Stephens, V. P. Sybert, S. B. Mallory, B. V. North, D. R. Witt, E. Sprecher, A. E. Taylor, A. Ilchyshyn, C. T. Kennedy, H. Goodyear, C. Moss, D. Paige, J. I. Harper, B. D. Young, I. M. Leigh, R. A. Eady, E. A. O’Toole, Am. J. Hum. Genet. 2005, 76, 794–803. 42â•…G. van Meer, D. Halter, H. Sprong, P. Somerharju, M. R. Egmond, FEBS Lett. 2006, 580, 1171–1177. 43â•… D. J. Sillence, R. J. Raggers, D. C. Neville, D. J. Harvey, G. van Meer, J. Lipid Res. 2000, 41, 1252–1260. 44â•…Y. J. Jiang, Y. Uchida, B. Lu, P. Kim, C. Mao, M. Akiyama, P. M. Elias, W. M. Holleran, C. Grunfeld, K. R. Feingold, J. Biol. Chem. 2009, 284, 18942–18952. 45â•… K. Sakakibara, T. Momoi, T. Uchida, Y. Nagai, Nature 1981, 293, 76–79. 46â•… V. Chigorno, M. Valsecchi, D. Acquotti, S. Sonnino, G. Tettamanti, FEBS Lett. 1990, 263, 329–331. 47â•… B. K. Gillard, L. T. Thurmon, D. M. Marcus, Glycobiology 1993, 3, 57–67. 48â•… S. Hoetzl, H. Sprong, G. van Meer, J. Neurochem. 2007, 103(Suppl. 1),3–13. 49â•… J. A. F. Op den Kamp, Ann. Rev. Biochem. 1979, 48, 47–71. 50â•… D. J. Sillence, R. J. Raggers, G. van Meer, Methods Enzymol. 2000, 312, 562–579. 51â•… I. L. van Genderen, G. van Meer, J. W. Slot, H. J. Geuze, W. F. Voorhout, J. Cell Biol. 1991, 115, 1009–1019. 52â•… M. A. Kiebish, X. Han, H. Cheng, A. Lunceford, C. F. Clarke, H. Moon, J. H. Chuang, T. N. Seyfried, J. Neurochem. 2008, 106, 299–312. 53â•… M. R. Rippo, F. Malisan, L. Ravagnan, B. Tomassini, I. Condo, P. Costantini, S. A. Susin, A. Rufini, M. Todaro, G. Kroemer, R. Testi, FASEB J. 2000, 14, 2047–2054. 54â•… F. Malisan, R. Testi, Biochim. Biophys. Acta 2002, 1585, 179–187. 55â•… C. Garcia-Ruiz, A. Colell, A. Morales, M. Calvo, C. Enrich, J. C. Fernandez-Checa, J. Biol. Chem. 2002, 277, 36443–36448. 56â•… T. Hasegawa, N. Sugeno, A. Takeda, M. Matsuzaki-Kobayashi, A. Kikuchi, K. Furukawa, T. Miyagi, Y. Itoyama, FEBS Lett. 2007, 581, 406–412. 57â•… K. Hanada, K. Kumagai, S. Yasuda, Y. Miura, M. Kawano, M. Fukasawa, M. Nishijima, Nature 2003, 426, 803–809. 58â•… K. Huitema, J. van den Dikkenberg, J. F. Brouwers, J. C. Holthuis, EMBO J. 2004, 23, 33–44. 59â•… S. Yamaoka, M. Miyaji, T. Kitano, H. Umehara, T. Okazaki, J. Biol. Chem. 2004, 279, 18688–18693.

74

TRANSMEMBRANE TRANSLOCATION OF SPHINGOLIPIDS

60â•… A. J. Verkleij, R. F. A. Zwaal, B. Roelofsen, P. Comfurius, D. Kastelijn, L. L. M. van Deenen, Biochim. Biophys. Acta 1973, 323, 178–193. 61â•… J. A. Post, G. A. Langer, J. A. Op den Kamp, A. J. Verkleij, Biochim. Biophys. Acta 1988, 943, 256–266. 62â•…R. M. Buckland, G. K. Radda, C. D. Shennan, Biochim. Biophys. Acta 1978, 513, 321–337. 63â•… F. X. Contreras, A. V. Villar, A. Alonso, R. N. Kolesnick, F. M. Goñi, J. Biol. Chem. 2003, 278, 37169–37174. 64â•… A. van Helvoort, M. L. Giudici, M. Thielemans, G. van Meer, J. Cell Sci. 1997, 110, 75–83. 65â•…R. D. Klausner, J. G. Donaldson, J. Lippincott-Schwartz, J. Cell Biol. 1992, 116, 1071–1080. 66â•…Y.-J. Shiao, J. E. Vance, J. Biol. Chem. 1993, 268, 26085–26092. 67â•… K. J. de Vries, A. A. J. Heinrichs, E. Cunningham, F. Brunink, J. Westerman, P. J. Somerharju, S. Cockcroft, K. W. A. Wirtz, G. T. Snoek, Biochem. J. 1995, 310, 643–649. 68â•…G. van Meer, W. van’t Hof, J. Cell Sci. 1993, 104, 833–842. 69â•… M. Tani, Y. A. Hannun, FEBS Lett. 2007, 581, 1323–1328. 70â•… M. Tani, Y. A. Hannun, J. Biol. Chem. 2007, 282, 10047–10056. 71â•…N. Andrieu, R. Salvayre, T. Levade, Eur. J. Biochem. 1996, 236, 738–745. 72â•… E. M. Bevers, P. L. Williamson, FEBS Lett. 2010, 584, 2724–2730. 73â•… H. Birbes, S. E. Bawab, Y. A. Hannun, L. M. Obeid, FASEB J. 2001, 15, 2669–2679. 74â•… H. Birbes, C. Luberto, Y. T. Hsu, S. El Bawab, Y. A. Hannun, L. M. Obeid, Biochem. J. 2005, 386, 445–451.

5 TRANSBILAYER MOVEMENT AND DISTRIBUTION OF CHOLESTEROL Peter Müller, Anna Pia Plazzo, and Andreas Herrmann Department of Biology, Humboldt-University Berlin, Berlin, Germany

5.1â•… INTRODUCTION Cholesterol is an important biomolecule of mammalian cells and fulfills various biological functions, for example, membrane biogenesis, steroid hormone and bile salt biosynthesis, and embryonic development. The organism derives cholesterol from dietary sources as well as from de novo synthesis. Synthesis is realized from acetyl coenzyme A in a series of enzymatic steps that are mainly confined to the endoplasmic reticulum [1–3]. Because of the extreme relevance of cholesterol for the organism, its level is highly regulated, and any deviation from its physiological concentration may cause pathological situations. The homeostasis of cholesterol underlies complex processes of trafficking, which include desorption and incorporation from/into membranes, intra- and extracellular transport, and transbilayer movement (see Reference 4). Since cholesterol is poorly soluble in water [5], its transfer across aqueous (intra- and extracellular) compartments is accomplished by transfer proteins and vesicles. Cholesterol constitutes a major component of mammalian plasma membranes. Cell membranes are characterized by a gradient of cholesterol among the different organelles, in which its relative fraction on total lipids increases along the secretory pathway from endoplasmic reticulum to Golgi and to the

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

75

76

TRANSBILAYER MOVEMENT OF CHOLESTEROL

plasma membrane, resulting in a specific enrichment in the latter. The cholesterol content of a membrane is often given in molar relation to the phospholipid concentration and amounts to about 0.8 in mammalian plasma membranes. Apart from its role in stabilizing the membrane bilayer structure, cholesterol has specific and unique physiological functions associated with membrane organization and dynamics. Examples are its role in forming membrane lipid domains [6, 7], modifying the structure and function of membrane proteins [8, 9], and facilitating the interaction of peripheric proteins with the plasma membrane [10]. 5.2â•… PHYSICOCHEMICAL FEATURES OF CHOLESTEROL Similarly to phospholipids, cholesterol has an amphiphilic structure that supports the formation of a lipid bilayer in aqueous solutions. It inserts into a membrane with its hydroxyl group oriented toward the membrane surface and an orientation of the long axis parallel to the normal of the bilayer [11]. However, in comparison to phospholipids, the proportion between the hydrophilic and hydrophobic part is different, as cholesterol contains a small hydrophilic head group represented by the OH moiety and a large hydrophobic structure consisting of a rigid ring system and a short-branched hydrocarbon chain. Due to these unique features, cholesterol has specific impacts on the structure and dynamics of membranes, as it increases the molecular order of the acyl chains of phospholipids enhancing the lipid packing in the membrane [12, 13]. This condensing effect, which has been shown to be much stronger for saturated phospholipids than for unsaturated lipids [14, 15], modulates many membrane properties and functions such as membrane thickness, fluidity, permeability to polar molecules, and lipid phase transition [16–18]. According to the “umbrella model,” it was suggested that in a bilayer, cholesterol relies on the coverage provided by the larger head groups of phospholipid neighbors [19]. Because of its large hydrophobic portion and rather small polar head group, cholesterol can penetrate deeply into the hydrophobic interior of the membrane. This causes its highly dynamic motion parallel to the membrane normal, whereby the molecule may also protrude into the opposite monolayer [20]. The interaction of the ring system of cholesterol with saturated fatty acyl chains mediated by van der Waals interactions and the formation of hydrogen bonds between the hydroxyl groups of cholesterol and the amide group of sphingolipids and ceramides are supposed to cause, in a mixture of lipids, a preferential association of cholesterol with these phospholipid species. This specific lipid–lipid interaction can trigger a lateral separation of lipids into liquid-ordered and liquid-disordered domains of fluid bilayers, which have been extensively characterized in model membranes [21–27]. A segregation of lipids is assumed to be the molecular basis for the formation of lateral domains

Methods for Measuring Cholesterol Transbilayer Movement

77

in biological membranes, in particular of cholesterol and saturated lipidenriched domains called rafts [6]. Those domains have been implicated to play an important physiological role in protein sorting, lipid trafficking, protein– ligand interaction, and signal transduction across membranes [7, 28–32]. For understanding the complex organization and dynamics of cholesterol in membranes, that is, its (1) distribution among membranes, (2) relative amount within a membrane, (3) lateral organization, and (4) its mobility parallel to the membrane normal in a monolayer, the investigation of the transbilayer distribution and translocation (flip-flop) of cholesterol are of particular relevance. The first studies dealing with the latter topics came out in the mid1970s, and meanwhile, a consistent amount of data achieved by different approaches has been published. 5.3â•… METHODS FOR MEASURING CHOLESTEROL TRANSBILAYER MOVEMENT AND DISTRIBUTION The transbilayer movement of cholesterol across the membrane is one important step of its complex trafficking within the organism. As for phospholipids, the basic mechanisms describing how a cholesterol molecule may transverse the membrane are (1) (passive) diffusion across the lipid bilayer, (2) unspecific diffusion facilitated by the presence of membrane proteins or at the boundary of membrane domains, and (3) active protein-mediated transport. The methods that have been developed to measure transbilayer mobility and distribution of cholesterol typically employ chemically nonmodified cholesterol (endogenous, radioactive) or modified cholesterol/steroid molecules, that is, fluorescent or spin-labeled analogs (Fig. 5.1). The rationale of assays characterizing cholesterol movement is to manipulate the molecule of choice specifically in one membrane leaflet (in general, the extracellular or extravesicular), either by its extraction onto acceptor molecules or by its chemical modification, and to measure the kinetics of redistribution of unaffected molecules from the other leaflet (Fig. 5.2). Thus, experimental kinetics consists optimally of two components, one corresponding to the extraction/modification process and the other to the transbilayer movement. However, for obtaining reliable quantitative data of cholesterol flip-flop, the rate constant of the process of modification kmod or extraction kex has to be at least one order of magnitude larger than that of transbilayer movement ki and ko (see Fig. 5.2); otherwise, one can only obtain estimates of the minimum rate constant of cholesterol flip-flop. This relates to the problem that in case the transbilayer movement of cholesterol is very rapid (in the timescale of seconds or faster), assays providing the appropriate time resolution are still lacking (see below). The modification of cholesterol/steroid molecules in one membrane leaflet and the quantification of cholesterol flip-flop rate is achieved by (1) treatment with cholesterol oxidase (modifying the hydroxyl group) and following the

78

TRANSBILAYER MOVEMENT OF CHOLESTEROL

CH3

CH3

CH3

CH3

CH3

OH

CH3

CH3

.ON O

HO

3-doxyl-17b-hydroxy-5a-cholestane (SL-androstane)

cholesterol CH3

CH3

CH3

CH3

CH3

CH3

CH3

CH3

O

CH3

.O

N

.O

N O

HO

25-doxyl-cholesterol

3b--doxyl-5a-cholestane (SL-cholestane)

CH3

+ N

H

CH3

O

N

CH3

N

O

CH3

N

HO

O

N CH3

CH3

O

+ N

CH3

CH3

N

O

-

N O

HO

22-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-23,24bisnor-5-cholen-3b-ol (22-NBD-cholesterol)

25-(N-[(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-methyl]amino)-27norcholesterol (25-NBD-cholesterol)

CH3 CH3

CH3

CH3

CH3

CH3

CH3

CH3

CH3

CH3

CH3

HO

HO

ergosta-5,7,9(11),22-tetraen-3b-ol (DHE) CH3 CH3

CH3

cholesta-5,7,9,(11)-trien-3b-ol (CTL) CH3

CH3 CH3

HO

1-methyl-10-norcholesta-1,3,5(10)-trien-3-ol (sterophenol)

CH3

CH3

CH3 CH3

HS

5-cholestene-3b-thiol (thiocholesterol)

Figure 5.1.â•‡ Chemical structure of cholesterol analogs used for measuring transbilayer movement and distribution of cholesterol.

increase in concentration of the reaction product; (2) reduction of spin-labeled or fluorescent analogs measuring the decline of intact label by electron spin resonance and fluorescence spectroscopy, respectively; and (3) quenching of fluorescent analogs by Förster resonance energy transfer (FRET) or quenchers, measuring the decrease in fluorescence intensity. For extraction of cholesterol or its analogs from membranes, a variety of different acceptors have been employed such as liposomes, red blood cell (RBC) membranes, lipoproteins,

79

kmod

Cholmod

Y

ko

Choli

Cholo

ki

A

kex

0.0

0.2

0.4

0.6

0.8

1.0

0

20

40 time

60

80

100

Choli

Cholo

Figure 5.2.â•‡ Principles of assaying transbilayer movement of cholesterol. Cholesterol molecules or analogs in the inner (Choli) and in the outer (Cholo) leaflet move to the opposite leaflet with time constants ko and ki, respectively. Molecules on the outer leaflet are specifically manipulated by chemical modification into Cholmod with a time constant kmod or by extraction onto an extracellular acceptor (A) with a time constant kex. X and Y are the cholesterol-modifying reactant before and after reaction with cholesterol, respectively. On the right, a typical time course-dependent decrease of signal intensity is shown consisting of a fast component (representing the modification/extraction of molecules localized on the outer membrane leaflet) and a slow component (representing the modification/extraction of molecules originally localized on the inner membrane leaflet).

inside

outside

Cholo

X

signal intensity

80

TRANSBILAYER MOVEMENT OF CHOLESTEROL

bile acids, and cyclodextrins. A combination of steroid modification and extraction has also been employed [33]. Notably, the release (extraction) of cholesterol onto extracellular lipoproteins is an important physiological mechanism by which cells export cholesterol (see below). Transbilayer distribution of cholesterol can be measured provided the assays applied fulfill certain conditions. The transbilayer distribution of cholesterol can be deduced from the translocation kinetics. If the rate constants of modification/extraction are very different from those of translocation process, that is, kmod (or kex)â•¯>>â•¯ki and kmod (or kex)â•¯>>â•¯ko (see Fig. 5.2), the distribution can be estimated from the first plateau-like intensity of the respective signal. This plateau is reached after the initial rapid manipulation of all cholesterol (analogs) in the accessible leaflet. Subsequently, the signal intensity decreases only slowly since the redistribution of cholesterol (analogs) from the nonaccessible to the accessible leaflet is slow. In case kmod (or kex)â•¯>â•¯ki and >ko, values for transbilayer distribution can be obtained by fitting the experimental data to an underlying mathematical model [34]. However, if kmod (or kex)â•¯∼â•¯ki and ∼ko or even kmod (or kex)â•¯<â•¯ki and
Transbilayer Movement of Cholesterol in Model Membranes

81

lesterol, acceptors may extract also phospholipids, additionally influencing membrane composition, as shown for methyl-β-cyclodextrin. 3. As outlined above, a large rate constant of the process of cholesterol modification or extraction compared with that of transbilayer movement prevents the accurate estimation of quantitative data and allows only specifying minimum rate constants of cholesterol flip-flop. 4. The use of cholesterol analogs raises the question of how reliable those analogs and their behavior reflect that of endogenous cholesterol [35]. This concerns both the influence of the label moiety and deviations from the structure of endogenous cholesterol on the intrinsic properties of the molecule as well as on its molecular environment. Comparing the impact of different cholesterol analogs on membrane properties, it has recently been shown that the fluorescent dehydroergosterol (DHE) and cholestatrienol, as well as a specific spin-labeled cholesterol analog, mimic essential properties of endogenous cholesterol in membranes, whereas fluorescent molecules bearing the 7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD) moiety are unfaithful analogs of cholesterol [35]. Using the methods described, cholesterol translocation and distribution has been investigated in lipid membranes serving as model systems for biological membranes as well as in plasma membranes and organelle membranes of cells.

5.4â•… TRANSBILAYER MOVEMENT OF CHOLESTEROL IN MODEL MEMBRANES For getting unbiased information on passive flip-flop of cholesterol in membranes, unilamellar liposomes are very useful. These lipid vesicles have several advantages, as (1) they contain only one bilayer, and they allow (2) the controlled investigation of the influence of lipid composition and (3) of various physicochemical parameters. Moreover, data on passive flip-flop of cholesterol in pure lipid vesicles may provide an estimate for that movement in biological membranes. This may help to distinguish between various mechanisms of transbilayer movement and, in particular, to determine the influence of membrane proteins. However, the use of liposomes is also linked to limitations and disadvantages, which are (1) compared with biological membranes a limited heterogeneity with regard to lipid and protein composition and presence of lateral domains, (2) difficulties in handling due to their small size, and, (3) in case of small unilamellar vesicles (SUVs), a large surface curvature. Some assays, in particular those based on extraction, may require separation of vesicles by centrifugation, which might be time-consuming in case of liposomes. The rather high curvature of SUV membranes may give rise to local defects in the bilayer accelerating passive flip-flop.

82

TRANSBILAYER MOVEMENT OF CHOLESTEROL

Studies performed with phospholipid vesicles using different assays gave contradictory estimates for cholesterol flip-flop, ranging in the scale from seconds [33, 36–40] to hours [41–46] (see Table 5.1). The reasons for these large discrepancies are discussed below. Considering the transbilayer movement of cholesterol in complex biological membranes, measurements on phospholipid vesicles have shown that the lipid composition with regard to the head group and/or fatty acids of phospholipids as well as to the presence of endogenous cholesterol has an influence on the flip-flop kinetics [33, 39, 40, 45, 47]. 5.5â•… TRANSBILAYER MOVEMENT OF CHOLESTEROL IN BIOLOGICAL MEMBRANES Transbilayer movement of cholesterol in biological membranes was mainly investigated on RBC, but data have also been presented for other systems. Notably, in biological membranes, large discrepancies of transbilayer rate constants were observed as for model membranes (see Table 5.2). Some researchers found a very slow transbilayer movement with half-times of flip-flop at 37°C in the range of 50–130 minutes for RBC membranes [48–50], of 4–6 hours to 13 days for virus membranes [51, 52], of 13.6 hours to 18 days for mycoplasma membranes [53, 54], and of 11.5 minutes for axonal membranes [55]. In contrast, other studies revealed a rapid movement with half-times of flipflop for RBC membranes in the range of seconds or even faster [40, 56, 57]. By measuring the cholesterol efflux from different mammalian cell lines onto cyclodextrin, Rothblat and coworkers found two different kinetic compartments of release, which they suggested to arise from cholesterol localized in the plasma membrane [58]. Notably, they discussed the slow pool (having a half-time of about 20 minutes at 37°C) to be caused either by cholesterol transbilayer movement or by an exchange between lateral domains and concluded that a better definition of the physical state of cholesterol in the plasma membrane is required to identify the kinetic pools of cholesterol (see below). 5.6â•… TRANSBILAYER DISTRIBUTION OF CHOLESTEROL IN LIPID AND BIOLOGICAL MEMBRANES Strongly related to its transbilayer movement is the question of the cholesterol distribution between the two leaflets of a bilayer. A rapid flip-flop may indicate a rather symmetric distribution unless specific interactions with other molecules would trap cholesterol in either leaflet. An asymmetric distribution of one lipid could also be a consequence for the asymmetric distribution of other lipid species in the bilayer. To preserve bilayer stability, the surface area of both monolayers has to match each other; otherwise, the membrane would bend and, if the surface area imbalance increases, even fragment (vesiculation)

83

Thiocholesterol

Sterophenol

Radioactive cholesterol Radioactive cholesterol Radioactive cholesterol Radioactive cholesterol Radioactive cholesterol Radioactive cholesterol Thiocholesterol

eggPC-SUV

eggPC-SUV

DPPC/chol (1:0.9)-SUV

eggPC-SUV

SUV of different compositions PC/chol-SUV

PC/chol (1:0.8)-SUV different PC species PC/chol (1:0.9)-SUV different PC species eggPC/chol (1:1)-SUV

Steroid

Membrane

Reaction with sulfhydryl reagent DTNB trapped inside SUV

Transfer between vesicles Cholesterol oxidase

Transfer between SUV

Transfer onto RBC

Transfer onto RBC

Reaction with sulfhydryl reagent DTNB Fluorescence quenching Transfer onto RBC

Assay

[44]

[45] [37]

n.d. n.d. n.d. n.d. n.d.

<1.5 hours (37°C) <6 hours (37°C) for eggPC ≤1 minute (37°C) ≤1 minute (20°C)

(Continued)

[38]

[36]

[47]

[43]

[42]

[41]

Reference

n.d.

Symmetric distribution n.d.

n.d.

Transbilayer Distribution

About 6 days (37°C) 3.1 hours for eggPC Very slow (37°C)a

72 minutes (30°C)

42 minutes (22°C)

t1/2 of Flip-Flop (Temperature)

TABLE 5.1.â•‡ Summary of Data Published on Cholesterol Translocation and Distribution in Lipid Membranes

84 Cyclodextrin-mediated transfer onto LUV Extraction by cyclodextrin in conjunction with FRET Reduction of spin label

Radioactive cholesterol

DHE

Spin-labeled androstane

Transfer onto LUV

Assay

Radioactive cholesterol

Steroid

n.d.

40 seconds (10°C) for POPC-LUV

n.d.

n.d.

<1–2 minutes (37°C)

0.6 minute (4°C)

n.d.

Transbilayer Distribution

About 10 hours (37°C)

t1/2 of Flip-Flop (Temperature)

[40]

[33]

[39]

[46]

Reference

Measurable flip-flop under nonequilibrium conditions. PC, phosphatidylcholine; DPPC, dipalmitoyl phosphatidylcholine; chol, cholesterol; LUV, large unilamellar vesicle; SOPC, 1-stearoyl-2-oleoyl phosphatidylcholine; SOPG, 1-stearoyl-2-oleoyl phosphatidylglycerol; POPC, 1-palmitoyl-2-oleoyl phosphatidylcholine; DMPC, dimyristoyl phosphatidylcholine; DTNB, 5,5′-dithiobis(2-nitrobenzoic acid); n.d., not determined.

aâ•‡

eggPC-SUV

eggPC/chol/cerebroside (0.75:0.1:0.15)-SUV and LUV SOPC/SOPG (0.85:0.15)-LUV SOPC/SOPG/chol (0.85:0.15:0.5)-LUV eggPC, POPC, DMPCSUV, and LUV

Membrane

TABLE 5.1.â•‡ (Continued )

85

RBC

RBC

RBC

RBC, fibroblasts

RBC

Spin-labeled androstane and cholestane

Radioactive cholesterol DHE, NBDcholesterol Fluorescence quenching, photobleaching Reduction of spin label

Cholesterol oxidase

Fluorescence quenching

Cholesterol oxidase

Transfer to plasma

Transfer to bile salts

Radioactive cholesterol Radioactive cholesterol Endogenous cholesterol DHE

RBC

Transfer to lipid vesicles Transfer to SUV

Radioactive cholesterol Radioactive cholesterol

Influenza virus Vesicular stomatitis virus RBC

Assay

Steroid

Membrane

n.d. n.d.

<1 hour (37°C) <50 minutes (37°C) 3 seconds (37°C) n.d.

SL-A: 4 minutes (4°C) SL-C: <0.5 minutes (4°C)

130 minutes (37°C) n.d.

Inside concentrated

4–6 hours (37°C)

Symmetric distribution

Inside concentrated

Fibroblasts: inside concentrated; RBC: outside concentrated Inside concentrated

n.d.

Symmetric distribution

Transbilayer Distribution

13 days (37°C)

t1/2 of Flip-Flop (Temperature)

TABLE 5.2.â•‡ Summary of Data Published on Cholesterol Translocation and Distribution in Biological Membranes

(Continued)

[40]

[69]

[50]

[64]

[56]

[48]

[49]

[52]

[51]

Reference

86 Fluorescence quenching

DHE

DHE cholestatrienol

Synaptic plasma membranes CHO cells n.d.

n.d.

18 days (37°C) 13.6 hours (37°C)

11.5 minutes (21°C) n.d.

Inside concentrated

Inside concentrated

[63]

Mycoplasma gallisepticum: symmetric distribution; Mycoplasma capricolum: outside concentrated Symmetric distribution

[71]

[68, 70]

[53, 54]

[55]

[67]

n.d.

Inside concentrated

[65, 66]

Outside concentrated

n.d.

[57]

n.d.

<1 second (37°C) 25 minutes (37°C)

Reference

Transbilayer Distribution

t1/2 of Flip-Flop (Temperature)

NBD, 7-nitrobenz-2-oxa-1,3-diazol-4-yl; CTL, cholestatrienol; PC, phosphatidylcholine; SL_A, spin-labeled androstane; SL-C, spin-labeled cholestane; n.d., not determined.

Transfer to highdensity lipoproteins or eggPC/ cholesterol SUV Fluorescence quenching

Extraction by cyclodextrin Reduction of NBD fluorescence, fluorescence quenching of CTL Fluorescence quenching Exchange with eggPC/ cholesterol SUV Measurement of filipin binding

Assay

Radioactive cholesterol

Radioactive cholesterol Cholesterol

DHE

Radioactive cholesterol NBD-cholesterol cholestatrienol

Steroid

Mycoplasma membranes

Tumor cell lines Squid axonal membrane Mycoplasma membranes

Platelets

RBC

Membrane

TABLE 5.2.â•‡ (Continued )

Cholesterol Flip-Flop: Fast or Slow?

87

or rupture. Hence, the asymmetric distribution of a lipid could be a consequence to keep the balance of surface areas between both monolayers. An asymmetric distribution of cholesterol could be caused both by leafletspecific interactions with lipids and/or proteins or by differences in the time constants of flip and flop [59]. Evidence, obtained in particular from model systems, suggests that phospholipids localized in the exoplasmic leaflet of the plasma membrane of mammalian cells interact more strongly with cholesterol than lipids in the cytoplasmic leaflet. Very likely, this is related to the head groups (interaction with sphingomyelin) and to the more saturated fatty acids of phospholipids enriched in the outer leaflet [39, 60–62]. Whereas the interaction of cholesterol with phospholipid species has been well characterized, its interaction with proteins and the consequences for its transbilayer distribution deserve more attention in future studies. Experimental results on transbilayer distribution of cholesterol are as discrepant as those for kinetics (see Tables 5.1 and 5.2). Whereas some studies observed a symmetric distribution of cholesterol in RBC, virus, and mycoplasma membranes [51, 54, 63], others revealed an enrichment in the outer leaflet of RBC, platelets, and mycoplasma membranes [64–66] or in the inner leaflet of RBC, viruses, various cell lines, and synaptic membranes [50, 52, 64, 67–71]. Notably, one study found for two different cell species (RBC, fibroblasts) a dissimilar distribution of cholesterol by using the same assay [64]. 5.7â•… CHOLESTEROL FLIP-FLOP: FAST OR SLOW? The physicochemical features of cholesterol, that is, its high insolubility in water and its very small head group, strongly argue for its rapid transbilayer movement. Considering the experimental studies of cholesterol transbilayer movement and distribution obtained on model and biological membranes, the question what are the reasons for the large discrepancies between the data arises. Several suggestions have been made, which are related to the methodological problems already emphasized above. A comparison of the data shows that those studies reporting a very slow flip-flop used assays based on transfer of cholesterol from the membrane of interest to acceptors (RBC membranes, vesicle membranes, lipoproteins) in the absence of any substance (carrier) accelerating the transfer. Since the release of cholesterol from a membrane seems to be a very rare, energetically unfavored process, we surmise that those assays did not provide the time resolution necessary for measuring a rapid cholesterol flip-flop. In contrast, studies that utilized cholesterol transfer onto acceptors such as methyl-β-cyclodextrin (which accelerates the extraction of cholesterol from membranes) reported a very rapid flip-flop of cholesterol in lipid vesicles as well as in RBC (in the timescale of seconds or even faster) [33, 39, 57]. Other studies indicating a rapid cholesterol movement in liposomes and RBC membranes in the order of seconds employed the cholesterol oxidase assay [37, 56] with the exception of Brasaemle et al. [50] who found

88

TRANSBILAYER MOVEMENT OF CHOLESTEROL

with this assay a rather slow translocation in RBC (half-time of 130 minutes at 37°C). In conclusion, recent studies point to a rather fast flip-flop of cholesterol. In particular, results on liposomes are in favor for a very rapid transbilayer motion. However, a systematic and comprehensive investigation on how cholesterol content affects cholesterol flip-flop is awaiting. Likewise, a more careful consideration of cholesterol flip-flop in lipid domains differing in lipid packing is required (see below). It is not known whether cholesterol undergoes similar flip-flop in liquid-disordered and liquid-ordered domains. 5.8â•… ROLE OF PROTEINS IN THE TRANSPORT OF CHOLESTEROL ACROSS MEMBRANES Phospholipids are transported across the plasma membrane by transporters specific for lipid species and direction of transport. For example, aminophospholipids are transported from the exoplasmic to the cytoplasmic leaflet in an adenosine triphosphate (ATP)-dependent manner by an aminophospholipid translocase belonging very likely to the class of P4-ATPases (see Chapters 8 and 10). While passive flip-flop of phospholipids is typically a very slow process being in the order of hours or even days, these transporters enable a rapid translocation from one leaflet to the other. For example, the characteristic time of inward transport of phosphatidylserine (PS) in the RBC membrane is about 5 minutes at 37°C [72]. In other mammalian cells such as sperm cells, it may be even faster [73, 74]. This aminophospholipid translocase activity is essential for maintaining the well-known asymmetric distribution of PS, which is almost entirely confined to the cytoplasmic leaflet. A rapid PS flip-flop in the order of the aminophospholipid translocase activity or even faster would interfere with an asymmetric transbilayer distribution of PS. Hence, a rapid passive flipflop of cholesterol provokes the question whether proteins, and, more specifically, energy-dependent transporters are involved in transbilayer movement (and distribution) of cholesterol at all. Membrane proteins could accelerate the passive diffusion of cholesterol either unspecifically at the interface between the protein and the lipid phase or specifically by facilitating the flipflop of the molecule via a floppase function, as known from phospholipids in biogenic membranes (see Chapter 6). However, not much data are available about this pathway of cholesterol translocation. By comparing results of RBC and lipid vesicles prepared from these cells, it was argued that proteins have no/minor influence on cholesterol flip-flop [40]. In fact, in case passive diffusion of cholesterol is very rapid, a significant influence of proteins, in particular transporters, on cholesterol flip-flop—and transbilayer distribution—might be questionable. Nevertheless, since the release of cholesterol from membranes to acceptors requires distinct membrane proteins, the function of those proteins could be to transport cholesterol across the bilayer to the acceptor-

Role of Proteins in the Transport of Cholesterol

89

accessible leaflet and thereby loading the acceptor with cholesterol (see below). A number of membrane proteins have been implicated in the transport of cholesterol across the plasma membrane. These primarily belong to the ATPbinding cassette (ABC) protein family, whose members also play a role in the transport of phospholipids (see Chapter 11). Mutations in the gene encoding the ABCA1 transporter are the cause of Tangier disease, in which efflux of cholesterol and phospholipids onto the extracellular protein apolipoprotein A-I (Apo A-I) is impaired [75–78]. This is the first and rate-limiting step in the formation of high-density lipoprotein (HDL) particles and in the process of reverse cholesterol transport. Based on animal models and on cell transfection studies, it has been suggested that ABCA1 is directly involved in the transbilayer movement and/or efflux of cholesterol (see Reference 79). However, this hypothesis has been debated by other studies, which rather propose that ABCA1 promotes cholesterol efflux by an indirect mechanism, that is, by mediating the translocation of PS to the outer leaflet of the plasma membrane and hence creating a preferred binding site for Apo A-I [80–83]. Whereas ABCA1 is responsible for the transfer of cholesterol to lipid-free or lipid-poor Apo A-I, ABCG1 and ABCG4 mediate transport of cholesterol from cells to phospholipid-containing acceptors such as HDL particles [84], whereby they could act in concert with ABCA1 for the generation of cholesterol-rich HDL [85]. The half-transporters ABCG5 and ABCG8 are involved in steroid secretion [86–88]. Defects in their genes are the cause of sitosterolemia, a disorder characterized by increased uptake of steroids (among them the plant steroid beta-sitosterol) in the intestine and decreased biliary excretion. Recently, ABCG5/G8 was purified from mouse liver and sterol transfer was reconstituted in membrane vesicles [89]. However, it remains to demonstrate whether and to which extent the transporters facilitate the movement of cholesterol from the inner to the outer leaflet to an acceptor. Notably, it was recently shown that ABCG5/G8-mediated efflux of cholesterol requires bile salts as acceptor [90, 91]. Other proteins for which a role in cholesterol transport/efflux has been controversially discussed are ABCB1 (multidrug resistance protein 1 or P-glycoprotein) (see Reference 92), ABCA7, and proteins of the ABCG family (see Reference 93). Summarizing, in regard to the role of ABC proteins in cholesterol transbilayer movement, studies are primarily based on experimental data describing the protein expression or ATPase activity in dependence on cholesterol content, or measuring the export of cholesterol from cells. However, those data do not allow drawing a detailed mechanism of how ABC proteins are involved in cholesterol transport, that is, in the transbilayer movement and/or release of steroids from the plasma membrane. To assess directly a function of ABC proteins as steroid (cholesterol) transporters, measurements on reconstituted

90

TRANSBILAYER MOVEMENT OF CHOLESTEROL A

ATP

A

ADP + P

A

ATP ADP + P

ATP

ADP + P

Figure 5.3.â•‡ Possible mechanisms for the involvement of ABC proteins in the translocation of cholesterol across the membrane. Proteins may mediate (1) the exposure of cholesterol to an extracellular acceptor (A) subsequent to its rapid passive diffusion across the lipid bilayer (left), (2) a flop of cholesterol from the inner to the outer leaflet and subsequent exposure to an acceptor (mid), or (3) a direct movement of cholesterol from the inner leaflet onto an extracellular acceptor (right). All mechanisms presume the consumption of ATP.

model systems constitute the most attractive approach. First, progress in reconstitution of ABCG5/G8 has been recently reported [89]. Based on the fact that a rapid cholesterol transbilayer movement may not require an energy-dependent protein to support its translocation, a hypothesis was proposed considering the physiological situation in which cholesterol is released from plasma membranes onto extracellular acceptors. According to this, the function of (ABC) proteins is to push upward cholesterol molecules out of the membrane surface, presenting them to extracellular acceptors and thereby enabling the release of cholesterol from the membrane [94]. As cholesterol has a high motion parallel to the membrane normal [20], those proteins may stabilize a more outward-exposed orientation. If this is the case, experiments investigating transbilayer movement of cholesterol in the absence of acceptors would give misleading results. As outlined above, studies using methyl-β-cyclodextrin as cholesterol acceptor revealed very short half-times for cholesterol flip-flop. Therefore, the role of cholesterol transport proteins might be an “exponase” function. Either cholesterol already localized in the acceptor-accessible leaflet could be exposed by those proteins, or exposure and release may be directly linked to the transport of cholesterol from the nonaccessible to the acceptor-accessible leaflet (Fig. 5.3). Anyway, the function of (ABC) proteins as cholesterol exponases awaits experimental proof. 5.9â•… CONCLUDING REMARKS Whereas previous studies concluded a slow flip-flop of cholesterol, recent data indicate that this process is very rapid even in pure phospholipid bilayers. As

Concluding Remarks

91

to whether the energy-independent transbilayer movement of cholesterol in biological membranes is facilitated unspecifically by transmembrane domains of membrane proteins and/or by the presence of flippases or floppases, which specifically accelerate flip-flop, remains to be shown. However, the presence of a rapid energy-independent transbilayer movement implicates only a minor role of energy-dependent transport processes in the transbilayer dynamics and distribution of cholesterol if at all. Nevertheless, evidence has been presented showing that energy-dependent membrane proteins such as ABC proteins are involved in the release of cholesterol from membranes to acceptors. This can be accomplished in that those proteins mediate an exposure of cholesterol out of the membrane surface, thereby facilitating its release from the membrane onto acceptor molecules. Likewise, the transbilayer distribution of cholesterol between both leaflets in a biological membrane, whether symmetric or asymmetric, requires still unequivocal confirmation. Although a rapid flip-flop and the absence of leafletspecific interactions of cholesterol with other lipids or proteins may imply a rather symmetric distribution, even under those conditions an asymmetric distribution is conceivable. A difference of surface area between both leaflets, for example, caused by an asymmetric distribution of distinct lipid species, can be counterbalanced by an asymmetric distribution of other lipid species undergoing freely flip-flop that would otherwise distribute symmetrically. In principle, cholesterol may be considered as such a freely moving lipid species. However, experimental results on shape changes of RBC may not be in line with this hypothesis. For example, the incorporation of lysophosphatidylcholine (which has very slow flip-flop) into the outer membrane leaflet of RBC causes a shape change from discocytes to echinocytes. According to the bilayer couple model [95], surface area changes in one membrane leaflet, for example, by specific incorporation of lipids, lead to tensions between both monolayers and consequential shape changes, as observed for RBC [72, 96–98]. In case cholesterol had a rapid flip-flop and could freely distribute between both leaflets, the area difference induced by lysophosphatidylcholine between both leaflets would be immediately counterbalanced by a redistribution of cholesterol preventing or reversing a shape change [99]. However, the echinocytes formed upon addition of lysophosphatidylcholine are rather stable up to several hours [100], much longer than the characteristic time of cholesterol flip-flop. This observation may indicate that the transbilayer dynamics of cholesterol, which is supposed to be rapid in RBC, is more complex and may be restricted by leaflet-specific lipid and/or protein interactions. Cholesterol plays an important role in the formation of lateral lipid membrane domains (see above). Notably, previous studies found an influence of lipid composition on cholesterol translocation [39, 40]. This also raises the question how cholesterol content and the lateral organization of the membrane affect cholesterol translocation. Cholesterol molecules confined to lipid domains, for example, to liquid-ordered domains, may not easily flip-flop between both leaflets [101].

92

TRANSBILAYER MOVEMENT OF CHOLESTEROL

A

lo

kex

A kexc

ld

kex

lo

kff

ld

kff

kexc liquid-ordered domain (lo)

liquid-disordered domain (ld)

Figure 5.4.â•‡ Scheme of lateral and transbilayer cholesterol dynamics in a membrane having lateral liquid-ordered (lo) and liquid-disordered (ld) domains: kfflo and kffld—rate constant for flip-flop in the lo and in the ld domains, respectively; kexc—rate constant lo ld of lateral exchange of cholesterol between the domains; kex and kex —rate constant of cholesterol release from the lo and from the ld domains, respectively, onto extracellular acceptors.

This would mean that studies reporting on fast flip-flop of cholesterol in RBC and plasma membranes of other cells refer only to cholesterol pools undergoing a fast transbilayer movement. Hence, the molecular mechanisms become more complex in that transbilayer movement as well as release onto extracellular acceptors can occur via different domains and depend on the properties of those domains (Fig. 5.4). ACKNOWLEDGMENT Own work of the authors was generously supported by grants from the Deutsche Forschungsgemeinschaft (MU 1017/5).

ABBREVIATIONS ABC ATP DHE FRET HDL LUV

ATP-binding cassette adenosine triphosphate dehydroergosterol Förster resonance energy transfer high-density lipoprotein large unilamellar vesicle

References

NBD RBC SUV

93

7-nitrobenz-2-oxa-1,3-diazol-4-yl red blood cell small unilamellar vesicle

REFERENCES 1â•… C. Lutton, Biochimie 1991, 73, 1327–1334. 2â•… S. K. Krisans, Ann. N.Y. Acad. Sci. 1996, 804, 8142–8164. 3â•… W. Nickel, B. Brügger, F. T. Wieland, Semin. Cell Dev. Biol. 1998, 9, 493–501. 4â•… B. Mesmin, F. R. Maxfield, Biochim. Biophys. Acta 2009, 1791, 636–645. 5â•… P. F. Renshaw, A. S. Janoff, K. W. Miller, J. Lipid Res. 1983, 24, 47–51. 6â•…K. Simons, E. Ikonen, Nature 1997, 387, 569–572. 7â•… R. G. W. Anderson, K. Jacobson, Science 2002, 296, 1821–1825. 8â•… D. C. Mitchell, M. Straume, J. L. Miller, B. J. Litman, Biochemistry 1990, 29, 9143–9149. â•… 9â•… A. D. Albert, J. E. Young, P. L. Yeagle, Biochim. Biophys. Acta 1996, 1285, 47–55. â•‡ 10â•… J. Ayala-Sanmartin, Biochem. Biophys. Res. Commun. 2001, 283, 72–79. â•‡ 11â•… M. P. Marsan, I. Muller, C. Ramos, F. Rodriguez, E. J. Dufourc, J. Czaplicki, A. Milon, Biophys. J. 1999, 76, 351–359. â•‡ 12â•… G. W. Stockton, I. C. Smith, Chem. Phys. Lipids 1976, 17, 251–263. â•‡ 13â•… E. Oldfield, M. Meadows, D. Rice, R. Jacobs, Biochemistry 1978, 17, 2727–2740. â•‡ 14â•… D. Huster, K. Arnold, K. Gawrisch, Biochemistry 1998, 37, 17299–17308. â•‡ 15â•… S. R. Shaikh, V. Cherezov, M. Caffrey, W. Stillwell, S. R. Wassall, Biochemistry 2003, 42, 12028–12037. â•‡ 16â•… R. A. Demel, B. De Kruyff, Biochim. Biophys. Acta 1976, 457, 109–132. â•‡ 17â•… P. L. Yeagle, Biochim. Biophys. Acta 1985, 822, 267–287. â•‡ 18â•… J. H. Ipsen, O. G. Mouritsen, M. Bloom, Biophys. J. 1990, 57, 405–412. â•‡ 19â•… J. Huang, G. W. Feigenson, Biophys. J. 1999, 76, 2142–2157. â•‡ 20â•… C. Gliss, O. Randel, H. Casalta, E. Sackmann, R. Zorn, T. Bayerl, Biophys. J. 1999, 77, 331–340. â•‡ 21â•… D. Scherfeld, N. Kahya, P. Schwille, Biophys. J. 2003, 85, 3758–3768. â•‡ 22â•… B. Y. van Duyl, D. Ganchev, V. Chupin, B. de Kruijff, J. A. Killian, FEBS Lett. 2003, 547, 101–106. â•‡ 23â•… S. L. Veatch, I. V. Polozov, K. Gawrisch, S. L. Keller, Biophys. J. 2004, 86, 2910–2922. â•‡ 24â•… S. R. Wassall, M. R. Brzustowicz, S. R. Shaikh, V. Cherezov, M. Caffrey, W. Stillwell, Chem. Phys. Lipids 2004, 132, 79–88. â•‡ 25â•… B. L. Stottrup, D. S. Stevens, S. L. Keller, Biophys. J. 2005, 88, 269–276. â•‡ 26â•… B. Ramstedt, J. P. Slotte, Biochim. Biophys. Acta 2006, 1758, 1945–1956. â•‡ 27â•… A. Bunge, P. Müller, M. Stöckl, A. Herrmann, D. Huster, Biophys. J. 2008, 94, 2680–2690. â•… â•… â•… â•… â•… â•… â•… â•…

94

TRANSBILAYER MOVEMENT OF CHOLESTEROL

â•‡ 28â•… E. J. Smart, G. A. Graf, M. A. McNiven, W. C. Sessa, J. A. Engelman, P. E. Scherer, T. Okamoto, M. P. Lisanti, Mol. Cell. Biol. 1999, 19, 7289–7304. â•‡ 29â•…K. Simons, D. Toomre, Nat. Rev. Mol. Cell Biol. 2000, 1, 31–39. â•‡ 30â•… S. Munro, Cell 2003, 115, 377–388. â•‡ 31â•… W. H. Binder, V. Barragan, F. M. Menger, Angew. Chem. Int. Ed. Engl. 2003, 42, 5802–5827. â•‡ 32â•… P. F. F. Almeida, A. Pokorny, A. Hinderliter, Biochim. Biophys. Acta 2005, 1720, 1–13. â•‡ 33â•…K. John, J. Kubelt, P. Müller, D. Wüstner, A. Herrmann, Biophys. J. 2002, 83, 1525–1534. â•‡ 34â•…U. Marx, G. Lassmann, H.-G. Holzhütter, D. Wüstner, P. Müller, A. Höhlig, J. Kubelt, A. Herrmann, Biophys. J. 2000, 78, 2628–2640. â•‡ 35â•… H. A. Scheidt, P. Müller, A. Herrmann, D. Huster, J. Biol. Chem. 2003, 278, 45563–45569. â•‡ 36â•… J. M. Backer, E. A. Dawidowicz, Biochim. Biophys. Acta 1979, 551, 260–270. â•‡ 37â•… J. M. Backer, E. A. Dawidowicz, J. Biol. Chem. 1981, 256, 586–588. â•‡ 38â•… E. A. Dawidowicz, J. M. Backer, Biochim Biophys Acta 1981, 644, 273–275. â•‡ 39â•… R. Leventis, J. R. Silvius, Biophys. J. 2001, 81, 2257–2267. â•‡ 40â•… P. Müller, A. Herrmann, Biophys. J. 2002, 82, 1418–1428. â•‡ 41â•… C.-H. Huang, J. P. Charlton, C. I. Shyr, T. E. Thompson, Biochemistry 1970, 9, 3422–3426. â•‡ 42â•… R. J. Smith, C. Green, FEBS Lett. 1974, 42, 108–111. â•‡ 43â•… M. Poznansky, Y. Lange, Nature 1976, 259, 420–421. â•‡ 44â•… M. J. Poznansky, Y. Lange, Biochim. Biophys. Acta 1978, 506, 256–264. â•‡ 45â•… Y. Nakagawa, K. Inoue, S. Nojima, Biochim. Biophys. Acta 1979, 553, 307–319. â•‡ 46â•… W. V. Rodrigueza, J. J. Wheeler, S. K. Klimuk, C. N. Kitson, M. J. Hope, Biochemistry 1995, 34, 6208–6217. â•‡ 47â•… B. Bloj, D. B. Zilversmit, Biochemistry 1977, 16, 3943–3948. â•‡ 48â•… Y. Lange, C. M. Cohen, M. J. Poznansky, Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 1538–1542. â•‡ 49â•… C. Kirby, C. Green, Biochem. J. 1977, 68, 575–577. â•‡ 50â•… D. L. Brasaemle, A. D. Robertson, A. D. Attie, J. Lipid Res. 1988, 29, 481–489. â•‡ 51â•… J. Lenard, J. E. Rothman, Proc. Natl. Acad. Sci. U.S.A. 1976, 73, 391–395. â•‡ 52â•… E. J. Patzer, J. M. Shaw, N. F. Moore, T. E. Thompson, R. R. Wagner, Biochemistry 1978, 17, 4192–4200. â•‡ 53â•… S. Rottem, G. M. Slutzky, R. Bittman, Biochemistry 1978, 17, 2723–2726. â•‡ 54â•… S. Rottem, D. Shinar, R. Bittman, Biochim. Biophys. Acta 1981, 649, 572–580. â•‡ 55â•… J. A. Steele, M. J. Poznansky, D. C. Eaton, M. S. Brodwick, J. Membr. Biol. 1981, 63, 191–198. â•‡ 56â•… Y. Lange, J. Dolde, T. L. Steck, J. Biol. Chem. 1981, 256, 5321–5323. â•‡ 57â•… T. L. Steck, J. Ye, Y. Lange, Biophys. J. 2002, 83, 2118–2125. â•‡ 58â•… M. P. Haynes, M. C. Phillips, G. H. Rothblat, Biochemistry 2000, 39, 4508–4517. â•‡ 59â•… A. Herrmann, P. Müller, Biosci. Rep. 1986, 6, 185–194.

References

95

â•‡ 60â•… T. Y. Wang, J. R. Silvius, Biophys. J. 2001, 81, 2762–2773. â•‡ 61â•… H. Ohvo-Rekila, B. Ramstedt, P. Leppimaki, J. P. Slotte, Prog. Lipid Res. 2002, 41, 66–97. â•‡ 62â•… J. R. Silvius, Biochim. Biophys. Acta 2003, 1610, 174–183. â•‡ 63â•… R. Bittman, S. Rottem, Biochem. Biophys. Res. Commun. 1976, 71, 318–324. â•‡ 64â•… J. E. Hale, F. Schroeder, Eur. J. Biochem. 1982, 122, 649–661. â•‡ 65â•…K. Boesze-Battaglia, S. T. Clayton, R. J. Schimmel, Biochemistry 1996, 35, 6664–6673. â•‡ 66â•…K. Boesze-Battaglia, R. J. Schimmel, J. Exp. Biol. 1999, 202, 453–460. â•‡ 67â•… A. B. Kier, W. D. Sweet, M. S. Cowlen, F. Schroeder, Biochim. Biophys. Acta 1986, 861, 289–301. â•‡ 68â•… W. G. Wood, F. Schroeder, L. Hogy, A. M. Rao, G. Nemecz, Biochim. Biophys. Acta 1990, 1025, 243–246. â•‡ 69â•… F. Schroeder, G. Nemecz, W. Gibson, C. Joiner, G. Morrot, M. Ayraut-Jarrier, P. F. Devaux, Biochim. Biophys. Acta 1991, 1066, 183–192. â•‡ 70â•…U. Igbavboa, N. A. Avdulov, S. V. Chochina, W. G. Wood, J. Neurochem. 1997, 69, 1661–1667. â•‡ 71â•… M. Mondal, B. Mesmin, S. Mukherjee, F. R. Maxfield, Mol. Biol. Cell 2009, 20, 581–588. â•‡ 72â•… M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. â•‡ 73â•… A. Zachowski, Biochem. J. 1993, 294, 1–14. â•‡ 74â•…K. Müller, T. Pomorski, P. Müller, A. Zachowski, A. Herrmann, Biochemistry 1994, 33, 9968–9974. â•‡ 75â•… M. Walter, U. Gerdes, U. Seedorf, G. Assmann, Biochem. Biophys. Res. Commun. 1994, 205, 850–856. â•‡ 76â•… G. A. Francis, R. H. Knopp, J. F. Oram, J. Clin. Invest. 1995, 96, 78–87. â•‡ 77â•… G. Rogler, B. Trumbach, B. Klima, K. J. Lackner, G. Schmitz, Arterioscler. Thromb. Vasc. Biol. 1995, 15, 683–690. â•‡ 78â•… A. T. Remaley, U. K. Schumacher, J. A. Stonik, B. D. Farsi, H. Nazih, H. B. Brewer, Jr., Arterioscler. Thromb. Vasc. Biol. 1997, 17, 1813–1821. â•‡ 79â•… C. Cavelier, I. Lorenzi, L. Rohrer, A. von Eckardstein, Biochim. Biophys. Acta 2006, 1761, 655–666. â•‡ 80â•… Y. Hamon, C. Broccardo, O. Chambenoit, M. F. Luciani, F. Toti, S. Chaslin, J. M. Freyssinet, P. F. Devaux, J. McNeish, D. Marguet, G. Chimini, Nat. Cell Biol. 2000, 2, 399–406. â•‡ 81â•… N. Wang, D. L. Silver, C. Thiele, A. R. Tall, J. Biol. Chem. 2001, 276, 23742–23747. â•‡ 82â•… N. Alder-Baerens, P. Müller, A. Pohl, T. Korte, Y. Hamon, G. Chimini, T. Pomorski, A. Herrmann, J. Biol. Chem. 2005, 280, 26321–26329. â•‡ 83â•… A. Zarubica, A. Pia Plazzo, M. Stöckl, T. Trombik, Y. Hamon, P. Müller, T. Pomorski, A. Herrmann, G. Chimini, FASEB J. 2009, 23, 1775–1785. â•‡ 84â•… N. Wang, D. Lan, W. Chen, F. Matsuura, A. R. Tall, Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 9774–9779. â•‡ 85â•… A. M. Vaughan, J. F. Oram, J. Lipid Res. 2006, 47, 2433–2443.

96

TRANSBILAYER MOVEMENT OF CHOLESTEROL

â•‡ 86â•…K. E. Berge, H. Tian, G. A. Graf, L. Yu, N. V. Grishin, J. Schultz, P. Kwiterovich, B. Shan, R. Barnes, H. H. Hobbs, Science 2000, 290, 1771–1775. â•‡ 87â•… M. H. Lee, K. Lu, S. Hazard, H. Yu, S. Shulenin, H. Hidaka, H. Kojima, R. Allikmets, N. Sakuma, R. Pegoraro, A. K. Srivastava, G. Salen, M. Dean, S. B. Patel, Nat. Genet. 2001, 27, 79–83. â•‡ 88â•… S. Shulenin, L. M. Schriml, A. T. Remaley, S. Fojo, B. Brewer, R. Allikmets, M. Dean, Cytogenet. Cell Genet. 2001, 92, 204–208. â•‡ 89â•… J. Wang, D. Zhang, Y. Lei, F. Xu, J. C. Cohen, H. H. Hobbs, X.-S. Xie, Biochemistry 2008, 47, 5194–5204. â•‡ 90â•… C. Vrins, E. Vink, K. E. Vandenberghe, R. Frijters, J. Seppen, A. K. Groen, FEBS Lett. 2007, 581, 4616–4620. â•‡ 91â•… B. J. Johnson, J. Y. Lee, A. Pickert, I. L. Urbatsch, Biochemistry 2010, 49, 3403–3411. â•‡ 92â•… A. Pohl, P. F. Devaux, A. Herrmann, Biochim. Biophys. Acta 2005, 1733, 29–52. â•‡ 93â•… H. Kusuhara, Y. Sugiyama, Pflugers Arch. 2007, 453, 735–744. â•‡ 94â•… D. M. Small, Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 4–6. â•‡ 95â•… M. P. Sheetz, S. J. Singer, Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 4457–4461. â•‡ 96â•… B. Deuticke, Biochim. Biophys. Acta 1968, 163, 494–500. â•‡ 97â•… J. E. Ferrell, K. J. Lee, W. H. Huestis, Biochemistry 1985, 24, 2849–2857. â•‡ 98â•… B. Isomaa, H. Hägerstrand, G. Paatero, Biochim. Biophys. Acta 1987, 899, 93–103. â•‡ 99â•… R. Heinrich, M. Brumen, A. Jäger, P. Müller, A. Herrmann, J. Theor. Biol. 1997, 185, 295–312. 100â•… A. Tamura, T. Sato, T. Fujii, Cell Biochem. Funct. 1987, 5, 167–173. 101â•… H. J. Risselada, S. J. Marrink, Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 17367–17372.

PART III ENERGY-INDEPENDENT PROTEINMEDIATED TRANSMEMBRANE MOVEMENT OF LIPIDS

6 PHOSPHOLIPID FLIP-FLOP IN BIOGENIC MEMBRANES Anant K. Menon Department of Biochemistry, Weill Cornell Medical College, New York, NY

Andreas Herrmann Department of Biology, Humboldt-University Berlin, Berlin, Germany

6.1â•… INTRODUCTION Phospholipids are synthesized on the cytoplasmic face of biogenic membranes. In the endoplasmic reticulum (ER), phospholipid biosynthetic enzymes, such as choline phosphotransferase, are membrane proteins that have their active site oriented toward the cytoplasm [1–4]. A similar situation is found for the inner bacterial membrane [5, 6]. Thus, newly synthesized phospholipids are located in the cytoplasmic leaflet of biogenic membranes. In the absence of a compensatory mechanism, this would result in asymmetric lipid enrichment. The resulting expansion of the cytoplasmic leaflet over the luminal or exoplasmic leaflet would interfere with the stability of the membrane. Strong bending and, perhaps, vesiculation would be a consequence, consistent with the bilayer couple hypothesis [7]. Indeed, as has been shown for giant unilamellar vesicles (GUVs) that have dimensions on the order of that of the ER, an asymmetric area expansion of the membrane as low as 0.1% drives bending and, eventually, formation of buds [8, 9] (see Chapter 2). To preserve the stability of biogenic membranes, phospholipid excess on the biosynthetic leaflet has to be relieved by an efficient and rapid mechanism of transbilayer movement and Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

99

100

BIOGENIC MEMBRANES

redistribution. Since transbilayer movement or flip-flop of phospholipids is typically very slow, biogenic membranes must be endowed with a phospholipid transport mechanism. Over the last 30 years, many efforts have been undertaken to elucidate the kinetics and molecular basis of phospholipid flip-flop in biogenic membranes. While initially the focus was on the transbilayer movement of glycerophospholipids, later studies expanded to include flipping of precursors of glycosphingolipids and glycophospholipids involved in protein N-glycosylation and glycosylphophatidylinositol (GPI) anchoring in the ER. In this chapter, we focus mainly on the transbilayer movement of glycerophospholipids across the ER. We describe assays for assessing transbilayer movement and distribution of phospholipids, emphasizing the reliability and specific limitations of the assays. We also summarize recent progress toward the biochemical identification of the transport machinery. Readers may wish to consult a recent review for additional material on this topic [10]. 6.2â•… ASSAYS FOR MEASURING TRANSBILAYER DISTRIBUTION OF ENDOGENOUS PHOSPHOLIPIDS Early studies on transbilayer organization of phospholipids were directed toward characterizing transbilayer distribution rather than transbilayer movement. In particular, it was of interest to know whether phospholipids in biogenic membranes adopted a pronounced asymmetric transbilayer distribution as known for the plasma membranes of red blood cells and other mammalian cells. Hence, assays were applied, which were supposed to measure the transbilayer distribution of phospholipids. Essentially, two assays were used, based on the chemical modifications of endogenous phospholipids: treatment of membrane phospholipids with (1) phospholipases or (2) membraneimpermeant reactive substances. As will be outlined below, these studies were undertaken with the problematic assumption that the transbilayer movement of phospholipids across the ER is rather slow, similar to that of phospholipid flip-flop in liposomes. 6.2.1â•… Phospholipase Assay Treatment of intact membranes with a phospholipase leads to irreversible chemical modifications of phospholipids. Phospholipases A1 or A2 cleave the ester bond of the fatty acid chains in the sn1 and sn2 position, respectively, whereas phospholipases C and D modify the head group. While type C releases the head group including the phosphate group, type D leaves the phosphate group linked to the glycerol backbone of phospholipids. To assess transbilayer distribution, phospholipids are extracted after phospholipase treatment and analyzed, for example, by thin-layer chromatography. For a given phospholipid species, the modified fraction is supposed to originate from the outer leaflet being exposed to the enzyme, while the nonmodified fraction is assigned to

Assays for Measuring Transbilayer Distribution

101

the nonexposed, inner leaflet. The basic assumptions underlying the assay are that the phospholipase affects all phospholipids in the outer leaflet, and that during enzyme treatment transbilayer phospholipid distribution is stable; that is, the rate constant of flip-flop is negligible with respect to that of enzyme reaction and the duration of membrane treatment. These assumptions are questionable. First, as established by subsequent studies (see below), phospholipid flip-flop in biogenic membranes is very fast with a half-time on the order of a minute or maximum of a few minutes. Usually, phospholipase treatment is done at 30–37°C for about 10–30 minutes; that is, during enzyme treatment, a significant amount of phospholipids will redistribute across the membranes. Because of this, phospholipase assays are also not useful as a tool to measure transbilayer movement unless the rate constant of the readout of the assay— phospholipase-mediated modification of lipids—would be large in comparison to that of flip-flop. Second, action of phospholipases could lead to products that modify the properties and shape of lipids and, hence, the organization of the structure and stability of the outer, exposed leaflet. Perturbation of that monolayer could by itself accelerate flip-flop in membranes, allowing enzymatic access to lipids redistributing from the nonaccessible to the accessible leaflet. We note that in addition to phospholipases, other phospholipid-modifying enzymes could be used to probe transbilayer distribution and movement. For example, phosphatidylserine (PS) decarboxylase can be used to modify specifically PS on the accessible monolayer converting it into phosphatidylethanolamine (PE). 6.2.2â•… Chemical Labeling Phospholipids, with a reactive moiety in the head group such as PE, offer the application of assays based on chemical labeling. Substances such as trinitrobenzene sulfonic acid (TNBS) or fluorescamine react with the amino group of PE; the resulting modified PE can be easily detected and separated from nonreacted PE by thin-layer chromatography. To assess unambiguously the transbilayer distribution, the reaction should be quantitative and fast compared with transbilayer movement of assayed lipid species, and labeling should not interfere with membrane stability and organization. Furthermore, an important aspect is that reactive substances must not permeate the membrane having access to the opposite monolayer of the membrane. To match this condition, usually incubation with reactive substances has to be performed at low temperature ensuring a very slow permeation. Of course, assays based on chemical labeling are limited to phospholipids with reactive head groups and cannot be applied to other important lipids of biogenic membranes such as phosphatidylcholine (PC). The TNBS assay, for example, has been applied also to assess flip-flop of PE in Bacillus megaterium [5] and in the ER of yeast cells [11]. Huijbregts et al. [12] measured the flip-flop of phospholipids in the inner membrane of Escherichia coli on the basis of biosynthetically 14C-labeled PE.

102

BIOGENIC MEMBRANES

Cells were briefly incubated with fluorescamine to label only 14C-PE on the external leaflet but not on the inner leaflet (see below). 6.3â•… ASSAYS FOR MEASURING TRANSBILAYER DISTRIBUTION AND MOVEMENT OF PHOSPHOLIPID ANALOGS The assays described above probe the transbilayer location of endogenous phospholipids by modifying them. Since the modified lipids resemble a nonnegligible fraction of total lipids, they may affect membrane stability and organization including transbilayer distribution and movement. Therefore, phospholipid analogs with a short chain in the sn2 position bearing either a paramagnetic nitroxide moiety or a fluorescent group such as 7-nitrobenz-2oxa-1,3-diazol-4-yl (NBD) have been developed and used. Due to the short fatty acyl chain, such analogs are partially soluble in water forming in aqueous solutions an equilibrium between monomers and micelles. Both the presence of monomers and the rather short lifetime of micelles allow a rapid and quantitative incorporation of analogs into model as well as natural membranes [13]. Usually, they incorporate quantitatively into the accessible leaflet of bilayer membranes; that is, initially, they are asymmetrically confined to the outer leaflet. A particular advantage is that the analogs are used in low amounts such that their membrane concentration is in the order of 1â•¯mol % of endogenous lipids or less. At such low concentrations, membrane integrity is preserved. Although spin-labeled and fluorescent lipid analogs have provided much insight into protein-mediated transbilayer dynamics of phospholipids, the bulky reporter moieties may affect the absolute values of the rate of transbilayer lipid movement (for a critical discussion, see Reference 14). Several approaches to measure transbilayer dynamics have been applied using phospholipid analogs. 6.3.1â•… Measuring Transbilayer Dynamics of Lipid Analogs by Extraction to Bovine Serum Albumin (BSA) Redistribution of phospholipid analogs across membranes is monitored by selective extraction of the analogs from the outer leaflet by defatted BSA, the so-called back-exchange assay [15–19]. To distinguish between analogs accessible to BSA and those residing on the inner leaflet, usually a centrifugation step is performed to separate membranes from the suspension medium containing BSA/analog complexes. A precondition of the assay is that BSA extracts only those analogs that are located in the outer membrane leaflet at the time of BSA addition. This assumption holds as long as no appreciable transmembrane movement of analogs occurs during the extraction of analogs with BSA. If, as for example in the case of rat liver microsomes, significant transmembrane movement of analogs occurs during extraction of analogs with BSA and subsequent centrifugation, this conventional back-exchange assay

Movement of Phospholipid Analogs

103

may yield inaccurate results. Despite considerable improvement of the assay to a time resolution of 30 seconds [20], the centrifugation step limits the resolution for detecting kinetics of rapid transbilayer lipid redistribution. Indeed, Buton et al. [20] already pointed out that, due to their comparatively low time resolution, available assays, for example, the back-exchange assay, may not be able to accurately characterize the transbilayer dynamics of phospholipid analogs in microsomal membranes (see “Introduction” and below) and may only allow definition of an upper limit for the half-time of phospholipid transbilayer movement in rat liver microsomal membranes. 6.3.2â•… Stopped-Flow Assays To characterize rapid lipid flip-flop and its underlying mechanism, assays are required that provide a sufficient time resolution for measurement of transbilayer dynamics of phospholipids. Recently, a stopped-flow assay that provides a sufficiently high resolution for the characterization of rapid lipid transbilayer movement in ER and other membranes has been developed [13, 21, 22]. The transbilayer distribution and movement of NBD- and spin-labeled short-chain analogs in rat liver microsomal membranes have been measured by this assay, which is also based on selective extraction of membrane-incorporated analogs “back” to the acceptor BSA. However, the major improvement was to make use of the fact that the signal of both spin-labeled and fluorescent analogs bound to BSA is different from that of the analogs incorporated into membranes. For example, as shown in Figure 6.1a, the electron paramagnetic resonance (EPR) signal of membrane-bound short-chain spin-labeled phospholipid analogs is different from that of the lipids when bound to BSA. Extraction of analogs from membranes to BSA is associated with line shape and, hence, signal intensity changes at a given magnetic field as indicated by arrows in Figure 6.1a. Those intensity changes can be measured and translated quantitatively into extraction kinetics. A similar approach can be applied for fluorescent short-chain lipid analogs by taking advantage of the observation that the quantum yield of analogs is different between inserted into membranes and bound to BSA. Thus, extraction can be followed quantitatively even in the presence of BSA. By that, the formerly obligatory centrifugation step for separation is no longer required. The rapid mixing of labeled membranes with a buffered BSA solution by stopped flow (Fig. 6.1a) prevents a delay too long for detection of rapid transbilayer movement. The delay time due to mixing of the stopped-flow equipment used here is ∼10â•¯ms. This is certainly sufficient to follow transbilayer movement of phospholipid analogs in membranes. Of course, resolving lipid transbilayer movement is limited by the rate constant of analog exchange to BSA; that is, the applicability of this approach depends on the quantitative relation between the rate constants for extraction of analogs by BSA (Fig. 6.1b) and the rate constants for transbilayer movement of analogs (Fig. 6.1c). Only if the extraction step is significantly faster in comparison to the transbilayer movement of analogs, the latter can be characterized by this

104

BIOGENIC MEMBRANES (a) Spin-labeled lipid Reservoirs

BSA

Mixing chamber and measurement of EPR signal

BSA +

Stopped-flow device EPR signal of spin-labeled lipid

Nonextractible SL-PC (%)

(b)

Outside 100

t=0

80

Lumenal

60 40 20

BSA

+

0 0

10 20 30 40 50 60 Time (seconds)

(c) Nonextractible SL-PC (%)

+ BSA

Asymmetrically labeled liposomes Outside

+ BSA

100

t=0

80 60 40 20

Lumenal

Phase I

kex

Phase II

+

BSA

k0

0 0

20 40 60 80 Time (seconds)

100

+ ER from rat liver

BSA

Movement of Phospholipid Analogs

105

Figure 6.1.â•‡ Rapid flip-flop of spin-labeled phospholipids assessed by stopped flow. (a,b) Rapid extraction of lipid analogs (red) from the monolayer accessible by BSA. The EPR spectrum is different between spin-labeled analogs bound to membranes and bound to BSA (a, left). Spin-labeled membranes and BSA solution are kept in two different reservoirs. After rapid mixing by stopped flow at tâ•¯=â•¯0, extraction of analogs to BSA can be followed quantitatively by measuring the intensity at the indicated spectrum position (arrows). In (b), a typical kinetics of extraction is shown. To demonstrate solely the kinetics of extraction, liposomes were labeled on the outer leaflet. Since flip-flop of spin-labeled phospholipid analogs in pure lipid membranes is very slow, all analogs were accessible to and, hence, extractable by BSA. (c) After labeling of ER membranes, analogs were allowed to redistribute across the membrane, before performing the stopped-flow experiment. At time tâ•¯=â•¯0, membranes are rapidly mixed with the BSA solution and extraction kinetics are followed. The initial rapid decline of signal intensity corresponds to extraction of analogs exposed to the BSA-accessible leaflet (Phase I, compare b). The subsequent slower decrease of signal intensity (Phase II) corresponds to analogs, which flipped from the BSA-non-accessible to the accessible leaflet becoming extracted to BSA. Finally, all analogs are extracted to BSA. For details, see Marx et al. [13, 21]. SL-PC, spin-labeled PC. Color version on the Wiley web site.

approach. Both the rate constant (or half-time) of transbilayer movement and the transmembrane distribution of analogs can be deduced from fitting of the respective kinetics of analog extraction to a mathematical model. This technique revealed that previous studies indeed underestimated the transbilayer dynamics of phospholipid analogs across ER membranes (see below). Thus, using fluorescence or EPR spectroscopy (the latter is shown in Fig. 6.1) in conjunction with stopped-flow mixing, allows monitoring of transbilayer movement of analogs in real time, that is, on-line, with a high time resolution. 6.3.3â•… Measuring Transbilayer Dynamics of Lipid Analogs by Selective Chemical Reduction Other methods to trace the transmembrane migration of analogs are based on reduction of exterior-facing spin- or NBD-labeled analogs to diamagnetic hydroxylamine or to nonfluorescent derivatives through the action of ascorbate [23] and dithionite [24], respectively. The EPR or fluorescence signal remaining after reduction is due to analogs that are oriented to the inner membrane leaflet. As shown previously, those assays can be optimized to assess analogs on the outer leaflet within a few seconds [13, 25] enabling also the following of rapid transbilayer movement even of long-chain analogs. In principle, a stopped flow can also be used for this type of assaying lipid analogs. However, a prerequisite is that membranes are impermeable to reducing agents in the time course of the assay. Since the assay does not rely on extractable short-chain lipid analogs, it can also be applied to long-chain spin-labeled or fluorescent analogs, provided the reporter moiety is readily accessible to the chemical reagent.

106

BIOGENIC MEMBRANES

6.3.4â•… Exchange of Lipid Analogs between Vesicles Another assay for continuous measurement of rapid transbilayer movement, originally applied to short-chain NBD-labeled analogs, is based on (1) the rapid exchange of those analogs between donor and acceptor vesicles and on (2) resonance energy transfer to a long-chain nonexchangeable fluorescent phospholipid analog such as N-Rh-PE. Previous studies have demonstrated that N-Rh-PE molecules do not transfer between membrane vesicles, whereas short-chain NBD-phospholipid analog molecules transfer rapidly [26, 27]. Provided this transfer is much faster in comparison to transbilayer movement of short-chain NBD-lipids analogs, the latter can be determined by the assay. The assay has already been successfully applied to brush-border membranes [28] and bacterial membranes [29]. 6.3.5â•… Water-Soluble Analogs There are a few examples where flippase activity has been measured with analogs of such low hydrophobicity that they may be considered to be water soluble. Dibutyroyl phosphatidylcholine (diC4PC), with a water solubility of close to 80â•¯mM, was used to assay phospholipid transport across ER vesicles [30], while mannose–phosphate–citronellol (Man-P-Cit) and glucose– phosphate–citronellol (Glc-P-Cit) were used to assay transport of the corresponding dolichol-based compounds across the ER [31, 32]. While the use of phospholipid analogs with high water solubility has attracted some criticism, they have nevertheless proven to be invaluable. Di(N-acetylglucosaminyl) pyrophosphorylnerol (GlcNAc-PP-nerol), a water-soluble analog of GlcNAcPP-undecaprenol, was used successfully to demonstrate a role for the WzxE protein in glycolipid transport [33]. In the case of Man-P-Cit and Glc-P-Cit, the analogs are recognized as sugar donors by glycosyltransferases involved in assembling the lipid precursor of protein N-glycosylation (see below), suggesting that they would also be recognized by the corresponding glycolipid flippases. In all cases, the analog is presumed to reside principally in the aqueous phase: On interaction with the membrane, it encounters the flippase and is transported to the opposite side where it partitions once again into the aqueous phase. Thus, incubation of water-soluble analogs with ER vesicles results in the transport and equilibration of the analog between the extravesicular and intravesicular spaces.

6.4â•… SHAPE CHANGES OF GUVs AS A TOOL TO MEASURE FLIP-FLOP GUVs are rather easy to prepare. Most often the electroswelling method according to Angelova et al. [34] is used. The diversity of shapes of GUVs has been described in detail [8, 35]. In particular, the shape of GUVs is sensitive

107

SHAPE CHANGES OF GUVs AS A TOOL TO MEASURE FLIP-FLOP

to modifications of their membrane leaflets (see Chapter 2). Even a small difference of surface area between monolayers is accompanied by the formation of a single bud [8, 9, 35, 36]. This bud is stable as long as the surface area difference persists. The very high sensitivity of GUV shape to a small excess of lipids in one monolayer provides an interesting way to study flip-flop of unlabeled lipids. Indeed, when a small amount of lipids, for example, lyso-PC, is inserted into the external leaflet of prolate GUVs (generated from spherical GUVs by increasing the osmolarity of the surrounding medium), a surface area difference is created between the two leaflets, which in turn results in the formation of a bud-like structure (Fig. 6.2a) [8, 9, 35, 36]. When added lipids and/or lipids from GUVs are allowed to redistribute between the two leaflets, (a)

(b)

flippase

lipids lipids

t=0

t = t1

t1 < t < t2 t = t2

t > t2

t1

t2

Figure 6.2.â•‡ Flip-flop of phospholipids assessed by shape changes of GUVs. (a) Insertion of exogenously added phospholipids (e.g., lyso-PC) at tâ•¯=â•¯0 into the outer leaflet of eggPC-GUVs causes budding. In the absence of lipid flip-flop, the bud remains essentially stable as shown by the image series. Both lipids, eggPC and lyso-PC, do not undergo significant flip-flop. GUVs were visualized by differential interference contrast at room temperature. (b) In the presence of a rapid spontaneous or flippase-mediated lipid flip-flop, the bud retracts, and the GUV returned to its original shape. A time series of an eggPC-GUV prepared from proteoliposomes containing yeast microsomal membrane proteins is shown. At time tâ•¯=â•¯0, lyso-LPC was added, causing the formation of a bud. The budding transition was observed at t1. t2 indicates the start of the recovery of the prolate shape. Based on these two time points, a half-time of flip-flop can be calculated (for details, see Reference 36). Color version on the Wiley web site.

108

BIOGENIC MEMBRANES

for example, by a reconstituted energy-independent flippase (see below), the bud is only transiently formed, and the original shape is recovered (Fig. 6.2b) [8, 35]. This can be explained by a relaxation of the monolayer area difference due to flip-flop. Otherwise, in the absence of flipping lipids, the bud remains. The time dependence of shape changes can be used for derivation of the flipflop rate constant [36]. This approach should also allow investigation of energydependent lipid transporters.Although transporters are very likely reconstituted to a similar extent in both opposing directions, unidirectional lipid transport can be ensured by allowing ATP to access only one membrane leaflet. Notably, lipid species to be transported can already be incorporated during GUV assembly. This is different from the visualization of ATP-independent flippase activity in GUVs where shape changes have to be triggered first, for example, by supplementing the external leaflet with additional lipids. 6.5â•… TRANSBILAYER MOVEMENT OF PHOSPHOLIPIDS IN THE ER As already pointed out, the active sites of the enzymes of phospholipid biosynthesis in the ER are located on the cytoplasmic face of the membrane, and consequently, phospholipid biosynthesis is an asymmetric process. Apparently, this observation seemed to be consistent with an asymmetric transbilayer distribution of phospholipids in the ER indicated by earlier studies [37, 38]. These studies assessed transbilayer distribution by hydrolyzing phospholipids of intact ER membranes by externally added phospholipases assuming that only lipids on the outer leaflet are attacked by the enzyme. Results of Higgins and Dawson [38] using phospholipase C suggested an enrichment of PC on the cytoplasmic side while PE was preferentially located on the luminal leaflet. In contrast, Nilsson and Dallner [37] hydrolyzing lipids by phospholipase A2 found a reversed asymmetric distribution of phospholipids, while others using also phospholipase A2 found that PE was equally distributed [39]. Whatever the reason for the discrepancy, studies on transbilayer distribution require that the transbilayer distribution of lipids is not perturbed by the treatment and that the transbilayer motion is slow during assessment (see above). Indeed, subsequently, it was shown that newly synthesized lipids in the ER are transferred rapidly across the membrane to the luminal leaflet [3, 40]. Measurements of transbilayer movement of PE and its methylated derivatives in ER vesicles revealed a characteristic time of a few minutes [3, 40]. Similar values were obtained for PC [30, 41, 42] and lyso-PC [43]. Zilversmit and Hughes reported half-times below 45 minutes for PC, PE, PS, and phosphatidylinositol (PI). In contrast, flip-flop of sphingomyelin (SM) was very slow. Even after 22 hours of measurement, half-times could not be determined [44]. Rapid transbilayer movement of phospholipids was not observed in multilamellar liposomes made from ER lipids [41], indicating a role of proteins in fast flip-flop across the ER.

Transbilayer Movement of Phospholipids in the ER

109

In 1985, Bishop and Bell used a water-soluble PC, diC4PC, and showed that the rapid passage through the ER membrane is sensitive to protein modifications either by proteases or by chemical reagents. This led them to postulate the existence of a PC transporter in the ER. Their hypothesis was sustained by a later report of Backer and Davidowicz [45], who succeeded in reconstituting this “flippase” activity into liposomes, from solubilized ER membranes. However, apart from a comparison between PC and its lyso derivative [43], there are few data concerning the phospholipid specificity of this transporter from ER. Since all the common phospholipids are flipped across the ER membrane (see above), it is widely assumed that the flippase itself is nonspecific; these results could also be explained by proposing a number of distinct flippases, each specific for a particular phospholipid. In support of the proposal that the ER has a single nonspecific phospholipid flippase activity are data showing that non-natural stereoisomers of PI are flipped [22]. By contrast, the aminophospholipid translocase from red cells has a high specificity for aminophospholipids (PS and PE) and does not efficiently transport its lyso derivatives [16] (see Chapters 8 and 10). Another difference is the requirement for cytosolic ATP in the case of the aminophospholipid translocase, while the activity of the microsomal transporter does not seem to require an energy source. Spin-labeled phospholipid analogs to assess flip-flop in the ER of rat liver cells by the conventional BSA back-exchange assay were applied for the first time in 1990 [17]. This study clearly demonstrated that the transbilayer movement is unspecific and energy independent. However, the time resolution of the assay used was too low: The characteristic half-time of flip-flop measured in these studies (∼20 minutes) was too high as became evident from subsequent studies by Buton et al. [20] and Marx et al. [21] using also spin-labeled phospholipid analogs. The latter studies applying assays of higher time resolution observed characteristic times of flip-flop less than 1 minute. While Buton et al. [20] found a half-time of about 30 seconds using a modified backexchange assay by centrifugation, Marx et al. [21] assessing transbilayer movement of spin-labeled phospholipids analogs by back-exchange via stopped flow concluded a half-time of flip-flop in the order of 16 seconds. To get an independent verification for data obtained by the stopped-flow approach, Marx et al. [21] applied the assay based on exchange of short-chain fluorescent lipid analogs between vesicles. For this purpose, freshly prepared rat liver microsomes were preincubated with either C6-NBD-PE or C6-NBDPC and, subsequently, N-Rh-PE-containing acceptor liposomes were added. Addition of the acceptor liposomes to C6-NBD-lipid analog-preincubated microsomes resulted in depletion of C6-NBD-lipid analogs from the microsomal membrane by spontaneous transfer to the acceptor liposomes. The time course of analog transfer from the donor microsomes to the acceptor liposomes could be detected through the decay in NBD fluorescence emission being caused by fluorescence resonance energy transfer from C6-NBD analogs to N-Rh-PE in the acceptor vesicles. The kinetics indicates a biphasic decay qualitatively similar to that shown for spin-labeled analogs in Figure 6.1c. The

110

BIOGENIC MEMBRANES

calculated rate constants for transbilayer movement of NBD-labeled analogs were about six- to eightfold lower than those for transbilayer movement of spin-labeled analogs. Very likely, the reporter moiety has a significant influence on the kinetics of redistribution of short-chain phospholipid analogs in membranes [21]. Recently, the energy-independent flippase activity from yeast ER has been reconstituted into GUVs, and it has been shown that the time course of GUV shape changes can also serve as an approach for quantitative characterization of the lipid transport activity of a flippase [36]. Insertion of lipids into the external leaflet created an area difference between the two leaflets that caused the formation of a bud-like structure as already described above. Upon reconstitution of the energy-independent flippase activity of the yeast ER into GUVs, the initial bud formation was reversible, and the shapes were recovered (Fig. 6.2b). This can be ascribed to a rapid flip-flop, leading to relaxation of the monolayer area difference. Theoretical analysis of kinetics of shape changes provides self-consistent determination of the flip-flop rate and further kinetic parameters. Based on that analysis, the half-time of phospholipid flip-flop in the presence of ER proteins was found to be on the order of about 2 minutes, similar to those that have been determined for flip-flop of fluorescently labeled phospholipid analogs in the ER. In contrast, GUVs reconstituted with influenza virus protein serving for a negative control formed stable buds. The results argue for the presence of specific membrane proteins mediating rapid flip-flop. 6.6â•… TRANSBILAYER MOVEMENT OF PHOSPHOLIPIDS IN THE BACTERIAL INNER MEMBRANE In gram-negative bacteria, phospholipids have to move from the site of their synthesis, the cytoplasmic leaflet of the inner membrane, to the periplasmic leaflet. In addition, phospholipids have to be transported to the inner leaflet of the outer membrane. Both processes have been studied in in vivo and in vitro systems (for a recent review, see Reference 6). The movement of phospholipids between the inner and the outer membrane was first shown by Osborn and Munson [46]. They demonstrated with pulse-labeling studies that PE is synthesized in the inner membrane of Salmonella typhimurium and finally transported to the outer membrane. Moreover, radiolabeled PS, introduced in the outer membrane, was shown to be rapidly transported to the inner membrane, where it becomes accessible to the enzyme PS decarboxylase and transformed to PE within 5 minutes [47, 48]. The resulting PE was transported back to the outer membrane. In the same study, it was shown that the transport of phospholipids in S. typhimurium was head-group independent, since not only the major phospholipids (PE, PG, and cardiolipin [CL]) but also the naturally nonoccurring phospholipid PC was transported to the inner membrane of S. typhimurium. In gram-negative bacteria, membrane contact

Transbilayer Movement of Phospholipids

111

sides between inner and outer membrane, so-called Bayer’s bridges [49], have been suggested to mediate intermembrane lipid transport. Additionally, in E. coli, a rapid bidirectional transport of phospholipids between the inner and the outer membrane was observed on whole cells [50, 51]. The first investigations on transbilayer movement of phospholipids in bacteria were carried out on gram-positive bacteria by Rothman and Kennedy [5, 52]. Similar to the results obtained with ER microsomes, they observed that the translocation of newly synthesized PE from the inner to the outer leaflet in Bacilli occurred with a half-time of 1.5–3 minutes at 37°C. A more recent study showed that short-chain, fluorescently labeled phospholipid analogs translocated rapidly across the B. megaterium membrane with a half-time of ∼30 seconds at 37°C [29]. This transport was demonstrated to be protease sensitive but not head-group dependent. Pulse-labeling studies on separated inner and outer membrane fractions from E. coli demonstrated that newly synthesized PE reached the outer membrane within 2.8 minutes. The transport of anionic phospholipids had a half-time of less than 30 seconds [50, 51]. Investigations on inverted inner membrane vesicles (IIMVs) from E. coli also demonstrated a rapid transbilayer movement of phospholipids across the vesicle membrane [53]. Utilizing short-chain fluorescent analogs of phospholipids, Huijbregts et al. [53] showed that exogenously added analogs rapidly flip across the inner membrane of E. coli with a half-time about 7 minutes at 37°C. This transport was temperature dependent, bidirectional, and not influenced by treatment with sulfhydryl reagents or proteinase K, nor by the presence of ATP or a pH gradient across the membrane of IIMV [53]. Huijbregts and coworkers also studied transmembrane movement of endogenously synthesized phospholipids across the inner membrane of E. coli [12]. Radioactively labeled PE was biosynthetically introduced into IIMV from PE-deficient E. coli strain AD93 by reconstitution with the enzyme phosphatidylserine synthetase (pss) and the addition of wild-type lysate, metabolic substrates and [14C]serine. Another approach utilized right-side-out vesicles, in which the active site of pss is oriented to the lumen of the vesicles. Under these circumstances, the PS conversion took place in the lumen of the vesicles, and the appearance of PE on the outer leaflet was measured. Both approaches demonstrated that the redistribution of newly synthesized radiolabeled PE occurred with a half-time of less than 1 minute. However, these earlier studies did not demonstrate a strong requirement for proteins in the translocation process or lacked the time resolution to measure an accurate translocation rate. Only little is known about the transverse distribution of phospholipids in bacterial membranes. In the plasma membrane of the gram-positive bacterium Micrococcus luteus, the distribution of PG and CL was studied using photoreactive lipid analogs [54]. A slight asymmetric distribution of PG with about 60% of the PG in the outer leaflet was found. CL was equally distributed between the two leaflets. However, in a later report, it was suggested that this distribution strongly depends on cell growth and division [55]. Huijbregts and colleagues, when investigating the transbilayer distribution of phospholipids

112

BIOGENIC MEMBRANES

in IIMV and right-side-out vesicles, detected an asymmetric transbilayer distribution of radiolabeled, newly synthesized PE in the inner membrane with an enrichment (65%) in the periplasmic leaflet [12]. 6.7â•… MECHANISM OF RAPID LIPID FLIP-FLOP IN BIOGENIC MEMBRANES Transbilayer movement in biogenic membranes such as eukaryotic ER or prokaryotic plasma membrane is fast, with flip-flop half-times on the order of minutes or less. Since flipping is extremely slow in liposomes, this suggests that biogenic membranes has a special phospholipid transport capability. Kol et al. [56–58] found that certain single α-helical membrane-spanning peptides promote transbilayer movement of some phospholipid classes in synthetic membranes, suggesting that lipid flip-flop proceeds in the absence of a dedicated flippase. However, data from biochemical reconstitution studies (described below) argue against this possibility. For example, studies on GUVs (see above) clearly demonstrated that the presence of membrane proteins in GUVs per se is not sufficient to facilitate flip-flop of phospholipids [36]. As shown for reconstituted influenza virus hemagglutinin, which exists as a trimer representing three transmembrane domains, the bud formation in GUVs by externally inserted lyso-PC was irreversible, indicating the absence of a rapid flip-flop. However, rapid disappearance of buds was observed when membrane proteins of yeast ER were reconstituted. Thus, these results argue for the presence of specific proteins (flippases) facilitating a fast transbilayer movement of lipids in the ER. This conclusion is supported by recent studies on the ER from rat liver cells and yeast cells suggesting the existence of dedicated flippases [59, 60]. In an attempt to identify the flippase(s) of microsomal membranes, Menon et al. [59] reconstituted protein fractions separated either on a glycerol gradient or by anion-exchange chromatography into proteoliposomes. This approach yielded protein pools of enriched flip-flop activity, suggesting that specific proteins are responsible for the fast lipid movement in the ER. However, very likely, flippase proteins constitute a minor fraction of ER proteins [36]. Thus, it is clear that specific proteins are required. But how do these work? In one model, flippases promote the rearrangement of the lipid bilayer in their proximity into nonbilayer structures that allow transverse movement of lipids. A variation of this idea, the so-called slip-pop model [56], suggests that the flippase functions by increasing the probability of a phospholipid to be found with its long axis lying along the midplane of the membrane; from this position, it could pop into either leaflet. Alternatively, one could consider a pore model in which the phospholipid head group is translocated through a hydrophilic, possibly water-filled region in the core of the flippase protein, while its associated acyl chains travel through hydrophobic regions of the protein or through the bilayer. The protein translocon has the requisite geometry for this flippase model, but depletion studies in yeast and bacteria show

Efforts to Identify Phospholipid Flippases

113

that it is not involved in phospholipid flipping [60, 61]. Many membrane-active peptides, such as magainin, are predicted to form toroidal pores in the membrane that effectively unite the two leaflets of the bilayer; such pores would provide a clear path for the transverse diffusion of lipids [62, 63]. These general models are consistent with the proposal that phospholipid flipping is nonspecific. In the case of isoprenoid-based lipids (see below) where flipping appears to be rather specific, different models must be proposed. Here, it is likely that the flippase has a binding site for the isoprenoid lipid and that a conformational change allows the lipid to be exchanged with a symmetrically placed binding site located on the opposite side of the protein, at the opposite membrane–water interface. 6.8â•… EFFORTS TO IDENTIFY PHOSPHOLIPID FLIPPASES The phospholipid flippase activity has eluded molecular identification thus far, but recent progress with biochemical approaches raises hope that this activity will soon be assigned to one (or more) protein(s). The first description of the biochemical reconstitution of flippase activity from detergent-solubilized rat liver ER proteins appeared in 1987 [45]. This was followed more than a decade later by a series of publications describing the reconstitution of phospholipid flip-flop from detergent-solubilized membrane proteins of rat liver and yeast ER [11, 59, 60], Bacillus subtilis [64], and E. coli inner membrane [65]. Membrane proteins from ER or bacterial inner membrane were solubilized, typically in Triton X-100, and reconstituted into PC vesicles via detergent removal with SM2 Bio-Beads. Flippase activity of the proteoliposomes was assayed in a variety of ways using (1) NBD-phospholipids (Fig. 6.3a) >[22, 60, 66], (2) diC4PC [59], (3) phospholipases [67], and (4) chemical modification [11]. Since none of these assays provided adequate time resolution (the rate at which lipid orientation is probed was typically faster than that of lipid flipping), results were recorded as endpoint measurements (Fig. 6.3b). By varying the amount of protein used for the reconstitution, it was possible to obtain a population of reconstituted vesicles in which none, some, or all possessed a flippase. The amount of protein required to reach the point where all vesicles had a flippase was used to estimate the physiological abundance of the transporter (Fig. 6.3b): This estimate was ∼2% and ∼0.5% by weight of Triton X-100-solubilized yeast and rat liver ER membrane proteins, respectively, roughly 10-fold less abundant than the protein translocon or oligosaccharyltransferase complexes [59, 60]. Velocity sedimentation analysis of Triton X-100-solubilized ER or bacterial membrane proteins indicated that flippase activity could be reconstituted from those fractions that sedimented at ∼4S [59, 60, 64], but that more rapidly sedimenting proteins were inactive. This experiment, as well as chromatographic studies showing that flippase activity can be enriched, supports the proposal that specific proteins are responsible for flippase activity and that not all proteins can provide this function (see above).

114

BIOGENIC MEMBRANES (a)

(b)

Liposome

Proteoliposome

80 % NBD-PC reduced

Out

In S2O42–

70

60

Out 50 0

In

50

75

Protein/phospholipid ratio (mg/mmol)

Membrane NBD

25

Lipid probe

Reduced NBD

Flippase

Figure 6.3.â•‡ Assay of phospholipid flip-flop in reconstituted vesicles. (a) Triton X-100solubilized ER membrane proteins are combined with eggPC and ∼0.3â•›molâ•›% of NBD-labeled phospholipid. The detergent-solubilized mixture is treated with detergentadsorbing SM2 Bio-Beads to yield unilamellar vesicles of diameter 150–175â•›nm. Protein-free vesicles are prepared alongside. On adding dithionite (S2O2− 4 ), a membraneimpermeant dianion, NBD-labeled phospholipids in the outer leaflet become nonfluorescent because dithionite reduces the NBD fluorophore to a nonfluorescent amine. For protein-free liposomes and proteoliposomes that lack the flippase, this results in 50% loss in fluorescence since NBD-phospholipid molecules in the inner leaflet are protected. For flippase-containing vesicles, all phospholipid analogs are reduced since those in the inner leaflet are translocated to the outer leaflet. Since the phospholipid flippase is nonspecific, it is likely that reduced NBD-phospholipids will also be substrates; hence, in flippase-containing vesicles, NBD-phospholipids reduced on the outer leaflet will be able to repopulate the inner leaflet as shown. In a mixed population where some vesicles have a flippase while others do not, the percent of lipid reporters detected will vary from 50% to 100% in proportion to the fraction of flippase-active vesicles. This provides a specific activity measure (the percent of vesicles that are rendered flippase active per milligram of the protein mixture being reconstituted) and also a mean to estimate flippase abundance. (b) Dose–response plot (adapted from Reference 71) indicating the percentage of NBD-PC (C6 or C12) that is reduced in vesicles reconstituted with different amounts of ER membrane proteins (different protein/phospholipid ratios). When no protein is reconstituted, 50% of NBD-PC is reduced corresponding to the outer leaflet population. As more protein is reconstituted, the probability that a particular vesicle has a flippase increases—in flippase-active vesicles, NBD-PC from the inner leaflet will be flipped out and reduced. The point where all vesicles have a flippase (∼25â•¯mg protein/mmol phospholipid) can be used to deduce flippase abundance (see Reference 59). Color version on the Wiley web site.

Conclusion

115

6.9â•… FLIPPING OF ISOPRENOID-BASED GLYCOLIPIDS Although this chapter focuses on membrane phospholipids, it is appropriate to summarize briefly the flip-flop of glycolipids in biogenic membranes. ATPindependent transbilayer transport of dolichol or undecaprenol-based glycolipids is important in the assembly of glycoconjugates in eukaryotes and prokaryotes. For example, the oligosaccharide donor for protein N-glycosylation in the ER lumen, Glc3Man9GlcNAc2-PP-dolichol, is synthesized by stepwise addition of components to dolichyl phosphate: The first seven steps occur on the cytoplasmic face of the ER, while the remaining seven occur in the ER lumen [68]. This necessitates flipping of Man5GlcNAc2-PP-dolichol across the ER membrane. The sugar donors required to convert Man5GlcNAc2-PPdolichol to Glc3Man9GlcNAc2-PP-dolichol are also isoprenoid based: Mannose is donated by mannose-phosphate dolichol and glucose is donated by glucosephosphate dolichol. These lipids are synthesized on the cytoplasmic face of the ER and must be flipped to the ER lumen to supply the glycosyltransferases that elaborate Man5GlcNAc2-PP-dolichol. Flipping of mannose-phosphate dolichol and glucose-phosphate dolichol was assayed using water-soluble analogs as described above [31, 69] and, more recently, by an oxidationbased assay with radiolabeled mannose-phosphate dolichol [70]; flipping of Man5GlcNAc2-PP-dolichol was recently studied in a reconstituted system and, unlike the situation with phospholipid flipping discussed above, shown to be quite specific [10, 71]. Based on genetic evidence, the membrane protein Rft1 was proposed to be the Man5GlcNAc2-PP-dolichol flippase [72]; however, biochemical reconstitution studies did not support this possibility [71, 73, 74]. Isoprenoid-based glycolipids are also involved in the assembly of bacterial O-antigens and peptidoglycan. Biochemical tests using the water-soluble analog GlcNAc-PP-nerol showed that ATP-independent transport of this compound during biosynthesis of the enterobacterial common antigen (a type of O-antigen) required the polytopic membrane protein WzxE [33]. Biosynthesis of bacterial peptidoglycan requires transport of the peptidoglycolipid, lipid II, from the cytoplasmic face of the inner membrane to the periplasmic face. Recent bioinformatics studies identified a candidate flippase for this process, named MurJ/MviN [75], but biochemical tests remain to be performed before definitive assignment can be made. 6.10â•… CONCLUSION We have summarized the approaches that have been used to measure the transbilayer distribution and movement of phospholipids in biogenic membranes. Several of these approaches provide the time resolution to follow rapid phospholipid flip-flop in biogenic membranes. However, although progress has been considerable, the protein(s) required for rapid flip-flop remain to be identified at the molecular level. Biochemical reconstitution approaches

116

BIOGENIC MEMBRANES

hold promise for the identification of these enigmatic proteins by conventional purification. It will be possible to begin understanding their mode of action once the proteins are identified. ABBREVIATIONS BSA GUV NBD PC PE PG PS

bovine serum albumin giant unilamellar vesicle 7-nitrobenz-2-oxa-1,3-diazol-4-yl phosphatidylcholine phosphatidylethanolamine phophatidylglycerol phosphatidylserine

REFERENCES â•‡ 1â•… D. E. Vance, P. C. Choy, S. B. Farren, P. H. Lim, W. J. Schneider, Nature 1977, 270, 268–269. â•‡ 2â•… R. Coleman, R. Bell, J. Cell Biol. 1978, 76, 245–253. â•‡ 3â•… J. L. Hutson, J. A. Higgins, Biochim. Biophys. Acta 1982, 687, 247–256. â•‡ 4â•… R. Bell, L. Ballas, R. Coleman, J. Lipid Res. 1981, 22, 391–403. â•‡ 5â•… J. E. Rothman, E. P. Kennedy, Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 1821–1825. â•‡ 6â•… R. P. H. Huijbregts, A. I. P. M. de Kroon, B. de Kruijff, Biochim. Biophys. Acta 2000, 1469, 43–61. â•‡ 7â•… M. P. Sheetz, S. J. Singer, Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 4457–4461. â•‡ 8â•… K. Berndl, J. Käs, R. Lipowsky, E. Sackmann, U. Seifert, Europhys. Lett. 1990, 13, 659–664. â•‡ 9â•… I. Lopez-Montero, M. Velez, P. F. Devaux, Biochim. Biophys. Acta 2007, 1768, 553–561. 10â•… S. Sanyal, A. K. Menon, Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 767–772. 11â•… T. Nicolson, P. Mayinger, FEBS Lett. 2000, 476, 277–281. 12â•… R. P. H. Huijbregts, A. I. P. M. de Kroon, B. de Kruijff, J. Biol. Chem. 1998, 273, 18936–18942. 13â•… U. Marx, G. Lassmann, K. Wimalasena, P. Müller, A. Herrmann, Biophys. J. 1997, 73, 1645–1654. 14â•… P. F. Devaux, P. Fellmann, P. Herve, Chem. Phys. Lipids 2002, 116, 115–134. 15â•… C. W. M. Haest, G. Plasa, B. Deuticke, Biochim. Biophys. Acta 1981, 649, 701–708. 16â•… G. Morrot, P. Herve, A. Zachowski, P. Fellmann, P. F. Devaux, Biochemistry 1989, 28, 3456–3462. 17â•… A. Herrmann, A. Zachowski, P. F. Devaux, Biochemistry 1990, 29, 2023–2027. 18â•… J. Connor, A. J. Schroit, Biochemistry 1990, 29, 37–43. 19â•… P. Fellmann, A. Zachowski, P. F. Devaux, Synthesis and use of spin-labeled lipids for studies of the transmembrane movement of phospholipids. In Biomembrane

References

117

Protocols II, J. M. Graham and J. A. Higgins, eds. Humana Press, Totowa, NJ, 1994, 161–175. 20â•… X. Buton, G. Morrot, P. Fellmann, M. Seigneuret, J. Biol. Chem. 1996, 271, 6651–6657. 21â•… U. Marx, G. Lassmann, H.-G. Holzhutter, D. Wustner, P. Müller, A. Höhlig, J. Kubelt, A. Herrmann, Biophys. J. 2000, 78, 2628–2640. 22â•… R. A. Vishwakarma, S. Vehring, A. Mehta, A. Sinha, T. Pomorski, A. Herrmann, A. K. Menon, Org. Biomol. Chem. 2005, 3, 1275–1283. 23â•… H. M. McConnell, R. D. Kornberg, Biochemistry 1971, 10, 1111–1120. 24â•… J. C. McIntyre, R. G. Sleight, Biochemistry 1991, 30, 11819–11827. 25â•… T. Pomorski, A. Herrmann, B. Zimmermann, A. Zachowski, P. Müller, Chem. Phys. Lipids 1995, 77, 139–146. 26â•… R. E. Pagano, O. C. Martin, A. J. Schroit, D. K. Struck, Biochemistry 1981, 20, 4920–4927. 27â•… 28â•… 29â•… 30â•… 31â•… 32â•…

J. W. Nichols, R. E. Pagano, Biochemistry 1981, 20, 2783–2789. Z. M. Zhang, J. W. Nichols, Am. J. Physiol. 1994, 267, G80–G86. S. Hrafnsdottir, J. W. Nichols, A. K. Menon, Biochemistry 1997, 36, 4969–4978. W. R. Bishop, R. M. Bell, Cell 1985, 42, 51–60. J. S. Rush, C. J. Waechter, J. Cell Biol. 1995, 130, 529–536. J. S. Rush, K. van Leyen, O. Ouerfelli, B. Wolucka, C. J. Waechter, Glycobiology 1998, 8, 1195–1205.

33â•… P. D. Rick, K. Barr, K. Sankaran, J. Kajimura, J. S. Rush, C. J. Waechter, J. Biol. Chem. 2003, 278, 16534–16542. 34â•… M. Angelova, S. Soleau, P. H. Meleard, F. Faucon, P. Bothorel, Prog. Colloid Polym. Sci. 1992, 89, 127–131. 35â•… J. Käs, E. Sackmann, Biophys. J. 1991, 60, 825–844. 36â•… A. Papadopulos, S. Vehring, I. Lopez-Montero, L. Kutschenko, M. Stöckl, P. F. Devaux, M. Kozlov, T. Pomorski, A. Herrmann, J. Biol. Chem. 2007, 282, 15559–15568. 37â•… O. Nilsson, G. Dallner, FEBS Lett. 1975, 58, 190–193. 38â•… J. A. Higgins, R. M. C. Dawson, Biochim. Biophys. Acta 1977, 470, 342–356. 39â•… R. Sundler, S. L. Sarcione, A. W. Alberts, P. R. Vagelos, Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 3350–3354. 40â•… J. L. Hutson, J. A. Higgins, Biochim. Biophys. Acta 1985, 835, 236–243. 41â•… A. M. H. P. Van Den Besselaar, B. De Kruijff, H. Van Den Bosch, L. L. M. Van Deenen, Biochim. Biophys. Acta 1978, 510, 242–255. 42â•… G. van Duijn, J. Luiken, A. J. Verkleij, B. de Kruijff, Biochim. Biophys. Acta 1986, 863, 193–204. 43â•…Y. Kawashima, R. Bell, J. Biol. Chem. 1987, 262, 16495–16502. 44â•… D. B. Zilversmit, M. E. Hughes, Biochim. Biophys. Acta 1977, 469, 99–110. 45â•… J. Backer, E. A. Davidowicz, J. Biol. Chem. 1981, 256, 13272–13277. 46â•… M. J. Osborn, R. Munson, Separation of the inner (cytoplasmic) and outer membranes of gram-negative bacteria. In Methods in Enzymology: Biomembranes Part A, S. Fleischer and L. Packer, eds. Academic Press, New York, 1974, 642–653.

118

BIOGENIC MEMBRANES

47â•… N. C. Jones, M. J. Osborn, J. Biol. Chem. 1977, 252, 7405–7412. 48â•… N. C. Jones, M. J. Osborn, J. Biol. Chem. 1977, 252, 7398–7404. 49â•… M. E. Bayer, J. Struct. Biol. 1991, 107, 268–280. 50â•… A. M. Donohue-Rolfe, M. Schaechter, Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 1867–1871. 51â•… K. E. Langley, E. Hawrot, E. P. Kennedy, J. Bacteriol. 1982, 152, 1033–1041. 52â•… J. E. Rothman, E. P. Kennedy, J. Mol. Biol. 1977, 110, 603–618. 53â•… R. P. H. Huijbregts, A. I. P. M. de Kroon, B. de Kruijff, Biochim. Biophys. Acta 1996, 1280, 41–50. 54â•… J. De Bony, A. Lopez, M. Gilleron, M. Welby, G. Laneelle, B. Rousseau, J. P. Beaucourt, J. F. Tocanne, Biochemistry 1989, 28, 3728–3737. 55â•… M. Welby, Y. Poquet, J. F. Tocanne, FEBS Lett. 1996, 384, 107–111. 56â•… M. A. Kol, A. I. P. M. de Kroon, J. A. Killian, B. de Kruijff, Biochemistry 2004, 43, 2673–2681. 57â•… M. A. Kol, A. I. P. M. de Kroon, D. T. S. Rijkers, J. A. Killian, B. de Kruijff, Biochemistry 2001, 40, 10500–10506. 58â•… M. A. Kol, A. N. C. van Laak, D. T. S. Rijkers, J. A. Killian, A. I. P. M. de Kroon, B. de Kruijff, Biochemistry 2003, 42, 231–237. 59â•… A. Menon, W. E. Watkins, III, S. Hrafnsdóttir, Curr. Biol. 2000, 10, 241–252. 60â•… S. Vehring, L. Pakkiri, A. Schroer, N. Alder-Baerens, A. Herrmann, A. K. Menon, T. Pomorski, Eukaryot. Cell 2007, 6, 1625–1634. 61â•… W. E. I. Watkins, A. K. Menon, Biol. Chem. 2002, 383, 1435–1440. 62â•… K. Matsuzaki, O. Murase, H. Tokuda, S. Funakoshi, N. Fujii, K. Miyajima, Biochemistry 1994, 33, 3342–3349. 63â•… K. Matsuzaki, Biochim. Biophys. Acta 1998, 1376, 391–400. 64â•… S. Hrafnsdóttir, A. K. Menon, J. Bacteriol. 2000, 182, 4198–4206. 65â•… J. Kubelt, A. K. Menon, P. Müller, A. Herrmann, Biochemistry 2002, 41, 5605–5612. 66â•… Q. Chang, S. N. Gummadi, A. K. Menon, Biochemistry 2004, 43, 10710–10718. 67â•… S. N. Gummadi, A. K. Menon, J. Biol. Chem. 2002, 277, 25337–25343. 68â•… C. B. Hirschberg, M. D. Snider, Annu. Rev. Biochem. 1987, 56, 63–87. 69â•… J. S. Rush, C. J. Waechter, Glycobiology 1998, 8, 1207–1213. 70â•… S. Sanyal, A. K. Menon, Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 11289–11294. 71â•… S. Sanyal, C. G. Frank, A. K. Menon, Biochemistry 2008, 47, 7937–7946. 72â•… J. Helenius, D. T. W. Ng, C. L. Marolda, P. Walter, M. A. Valvano, M. Aebi, Nature 2002, 415, 447–450. 73â•… C. G. Frank, S. Sanyal, J. S. Rush, C. J. Waechter, A. K. Menon, Nature 2008, 454, E3–E4. 74â•… J. S. Rush, N. Gao, M. A. Lehrman, S. Matveev, C. J. Waechter, J. Biol. Chem. 2009, 284, 19835–19842. 75â•… N. Ruiz, Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 15553–15557.

7 PHOSPHOLIPID SCRAMBLASE: WHEN PHOSPHOLIPID ASYMMETRY GOES AWAY Edouard M. Bevers Department of Biochemistry, Cardiovascular Research Institute Maastricht, Maastricht University, Maastricht, The Netherlands

Patrick L. Williamson Department of Biology, Amherst College, Amherst, MA

7.1â•… INTRODUCTION The amphipathic properties of phospholipids generate the membrane bilayers that border and compartmentalize living cells. In the endoplasmic reticulum (ER), the site of membrane biogenesis in eukaryotes (see Chapter 6), most lipid components are randomly distributed between the two leaflets of the membrane. As membranes travel downstream from the ER, the distribution of lipids between the two leaflets of membrane becomes nonrandom, or asymmetric. Multiple factors can contribute to generating this asymmetric distribution of phospholipids including ongoing asymmetric synthesis and modification of membrane lipids, directional transport of particular lipids, and differential biophysical interactions with other membrane components, including both proteins and other lipids. In the plasma membrane of animal cells, the final result of this process is a cytoplasmic leaflet dominated by the aminophospholipids phosphatidylserine (PS) and phosphatidylethanolamine (PE), and an

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

119

120

LOSS OF LIPID ASYMMETRY

exofacial leaflet mainly comprised of the choline-containing phospholipids phosphatidylcholine (PC) and sphingomyelin (SM). Generation and maintenance of this asymmetric lipid distribution is mainly under the control of a specific lipid translocator, the aminophospholipid translocase [1] (see Chapters 8–10); this enzyme catalyzes ATP-dependent transport of PS and PE from the outer to the inner leaflet of the plasma membrane. In addition, there is evidence for ATP-dependent movement of phospholipids from the inner to the outer leaflet of the red cell membrane, catalyzed by ABCC1 (previously MRP1) [2, 3]. Although slow (transport occurs over the course of hours), this movement tends to favor fluorescent phospholipid analogs, but there is evidence that the activity may also affect the distribution of endogenous lipids, since prolonged treatment of red cells with inhibitors of the enzyme results slightly lower levels of phospholipase-accessible PC and SM in the red cell outer leaflet. Together, these results show that the final distribution of phospholipids in a membrane must take account of movements in both directions. Although there is some variation depending on the size, charge, and polarity of the polar head group, the rate of uncatalyzed transbilayer movement of phospholipids is generally slow, with estimated half times of hours to even days [4–7]. As a result, lipid asymmetry is a relatively stable steady state, and lipid translocation is only required to establish asymmetry or to restore it after disturbance, for example, by membrane fusion processes during endo- and exocytosis. However, rapid loss of lipid asymmetry is a physiologically important process that occurs much more quickly than expected from simple inactivation of active lipid transport. Over the last decennia, evidence has accumulated for the existence of one or more proteins that catalyze transbilayer movement of phospholipids at the plasma membrane. Since this movement is bidirectional and nonselective, it causes randomization or scrambling of the phospholipids over both membrane leaflets, and the protein responsible for it was called the phospholipid “scramblase.” Despite considerable efforts, the identity of this protein is still unknown. Unfortunately, a search of protein databases returns at least four “phospholipid scramblases” (PLSCRs) from different species, but as will be outlined below, it is increasingly appreciated that these protein(s) are not scramblases at all. In this chapter, we will review the current knowledge on the properties and most important functions of this elusive protein. The review will be restricted to proteins or mechanisms that enable bidirectional lipid transport in plasma membranes; it is not known whether they are related to (equally elusive) proteins facilitating transbilayer movement in biogenic membranes. These activities are reviewed elsewhere [8]; see also Chapter 6. 7.2â•… HISTORICAL OVERVIEW The first evidence for a physiological change in phospholipid orientation came from Schick and coworkers [9]. Using the nonpermeable probe trinitroben-

Historical Overview

121

zene sulfonic acid (TNBS) to monitor exposure of aminophospholipids, these investigators found that stimulation of blood platelets with thrombin doubled the amount of PE that could be labeled compared with unstimulated platelets. Subsequently, phospholipase-based assays were developed, which could monitor all phospholipid classes. These assays eventually showed that the composition of the lipids in the hydrolyzed (external leaflet) fraction in activated platelets eventually resembled the total phospholipid composition of the platelet, implying randomization of all lipid components over both leaflets of the plasma membrane [10]. Collapse of lipid asymmetry upon platelet activation, with subsequent exposure of PS in the external leaflet of the plasma membrane, is instrumental in promoting the process of blood coagulation [11] because two sequential reactions of the coagulation process, activation of factor X and conversion of prothrombin to thrombin, require a negatively charged phospholipid surface. In addition to its physiological importance, this property of PS-containing membranes is the basis of a noninvasive and sensitive method to monitor changes in transbilayer lipid distribution in platelets as well as other cell types [12–14]. (For more information on the role of lipids, in particular PS, in coagulation, consult Chapter 16.) Soon after the first reports on loss of lipid asymmetry in platelets, evidence was presented that randomization of lipids is also induced in erythrocytes by elevation of the intracellular Ca2+ concentration [15–17]. Strong evidence for the unitary nature of these phospholipid randomizations was provided by an experiment of nature (reviewed in Reference 18). In 1985, we reported that platelets from a patient with the so-called Scott syndrome, named after the propositus, have an impaired scrambling activity. Later studies revealed two other patients, one in France and one in the United Kingdom, with a similar syndrome, and it was confirmed that the defect was heritable and affected lipid scrambling in all cells of the hematologic lineage, including both platelets and erythrocytes. Scott syndrome was also found in a single, inbred colony of dogs with similar clinical symptoms as observed in humans [19]. More details about this disorder and the underlying defect will be discussed in Chapter 16. The development of this field profited from a proliferation of alternative techniques, in addition to phospholipases and chemical modification, to monitor alterations in lipid asymmetry. The impermeant fluorescent dye merocyanine 540 (MC540) shows an increased affinity for membranes when the spacing between phospholipid head groups increases, and its binding was found to be an indirect measure for loss of lipid asymmetry [17, 20]. It was supplanted by an even more convenient and noninvasive probe, Annexin A5 (Annexin V), first introduced by Thiagarajan and Tait to detect PS exposure in activated platelets [21]. In the presence of Ca2+, this protein binds with a high affinity to negatively charged phospholipids; a recombinant protein, tagged with various fluorophores, is now commercially available and is widely applied by many investigators to monitor PS-exposing cells. Recently, MFG-E8 (lactadherin) has been introduced as an alternative to Annexin V that does

122

LOSS OF LIPID ASYMMETRY

not require Ca2+ for binding PS [22]. These techniques that assess the orientation of endogenous phospholipids (largely PS) have been augmented by the use of exogenous probes, beginning with radiolabeled phospholipids [23], but expanding to more versatile spin-labeled [24] or fluorescent-labeled [25] phospholipid analogs. Protocols based on extraction by bovine serum albumin or lipid acceptor vesicles [25, 26], or quenching by reduction (dithionite) [27] or resonance energy transfer [28], provide information on the amount of probe in the outer leaflet. Scramblase activity is measured either as inward movement of phospholipids, and particularly those that are not a substrate for the aminophospholipid translocase, or as outward movement of phospholipids from the inner leaflet. A more detailed discussion on the methodology of measuring lipid flip-flop in biological membranes will be given in Chapters 1, 2, and 6. 7.3â•… PHYSIOLOGICAL IMPORTANCE OF LIPID SCRAMBLING The most dramatic external change in a plasma membrane that has undergone phospholipid scrambling is the exposure of PS, because this lipid is normally restricted entirely to the cytoplasmic leaflet. The critical role of PS in several reactions of the blood coagulation pathway explains why this was one of the first major physiological functions of lipid scrambling to be recognized. The exposure of PS on activated platelets provides a platform for the assembly of highly active protease/cofactor complexes catalyzing two sequential reactions of the coagulation cascade: the formation of activated factor X and thrombin. The significance of phospholipids in these reactions is to (1) increase the local concentration of coagulation factors, (2) induce conformational changes required for optimal functioning, and (3) facilitate the transfer of substrate and product between enzyme complexes. Because PS appears on platelets activated by adhesive interactions with subendothelial matrix molecules, in particular collagen, the protease complexes of the coagulation cascade are physically confined to the site of vessel injury. One important functional requirement for a mechanism of this type results from the fact that the unactivated components of the cascade are part of the protein complement of circulating blood plasma; these components are therefore omnipresent in the circulation. The coagulation system must keep coagulation localized, and confining the reactions to an immobilized surface prevents coagulation from spreading into the circulation away from the wound. A second consequence is that this system means that the cascade may assemble on the surface of any cell on which PS becomes exposed. Nonphysiological exposure of PS in the circulation thus presents a risk of unwanted thrombosis. As noted above, a defective platelet PS exposure is responsible for a bleeding diathesis known as the Scott syndrome [18]. The role of phospholipid scrambling in coagulation is peculiar to vertebrates. A more general role for PS exposure was uncovered when Fadok et al.

Physiological Importance of Lipid Scrambling

123

reported that induction of apoptosis in lymphocytes causes PS exposure [14], a phenomenon that is now generally accepted to be a hallmark of apoptotic cells. PS presented on the surface of apoptotic cells is a key element in their recognition and engulfment, a process that is generally completed in vivo before the apoptotic cells even begin the morphological changes usually associated with cell death [29]. The removal of senescent erythrocytes from the circulation by the reticuloendothelial system is probably a variant on this process of apoptotic cell recognition. When this process is compromised, prolonged circulation of cells with persistent PS exposure and subsequent presentation of self-antigens to the immune system may lead to autoimmune syndromes [30–33]. The mechanisms of recognition and engulfment of apoptotic cells may also play a role in the PS-mediated adhesion of erythrocytes to endothelium or subendothelial matrix structures that has been documented in patients with various hemolytic anemias [34–37]. The PS-dependent recognition and engulfment of apoptotic cells by macrophages suppresses antigen presentation and development of an immune response, as would be expected for a process that removes normal cells during differentiation and tissue homeostasis. These features also present an opportunity to pathogens, and there is evidence that this opportunity has been exploited in several cases. In particular, the invasion of macrophages by Leishmania parasites seems to depend on apoptotic mimicry [38]. There is also evidence that PS exposure plays a role in HIV infection of monocytes [39], and Mercer and Helenius recently demonstrated that infection with vaccinia virus is critically dependent on the presence of exposed PS in the viral membrane [40]. Whether scramblase activity is required for PS exposure in these membranes is not known. Although coagulation and apoptotic cell recognition are the best-studied examples of physiological PS exposure, there are other, less well-characterized instances, where PS exposure, presumably resulting from scramblase activation, has functional implications. In particular, transient exposure of PS has been shown to play an essential role in cell fusion. The best characterized instance of this role is during myotube formation in muscle cell differentiation [41], and PS may play a similar function during formation of the placental syncytiotrophoblast layer from fusing trophoblasts [42]. These differentiation pathways are probably related to the scramblase-mediated collapse of lipid asymmetry in the sperm plasma membrane, which is an essential element in the capacitation reaction [43]. In this case, the reversible exposure of PS renders the sperm plasma membrane both fusogenic and responsive to zona pellucida glycoproteins. As will be described elsewhere in this book, changes in transversal lipid distribution due to the concerted action of scramblase and translocase activity may give rise to membrane bending supporting endo- and exocytosis processes (discussed in Chapter 1). Interestingly, a novel pinocytotic pathway has been described, in which Annexin A5 bound to surface-exposed PS is interÂ� nalized by an endocytotic process with subsequent cytoskeletal-dependent

124

LOSS OF LIPID ASYMMETRY

intracellular trafficking [44], but the physiological significance of this pathway is as yet unclear. 7.4â•… CHARACTERISTICS OF THE PHOSPHOLIPID SCRAMBLING PROCESS 7.4.1â•… A Pivotal Role for Ca2+ Ions From early on, the role of Ca2+ ions in lipid scrambling has been the object of intense study. Resealed erythrocyte ghosts prepared in the absence of Ca2+ retain their normal transbilayer lipid distribution, whereas inclusion of as little as 10â•¯µM Ca2+ can dissipate lipid asymmetry [15–17]. Only Ca2+ and Sr2+ will evoke lipid scrambling; other bivalent cations fail to do so or even inhibit in the cases of Mg2+ and Mn2+ [45]. Moreover, elevation of intracellular Ca2+ by means of selective ionophores causes loss of lipid asymmetry in a variety of hematopoietic cell types, with several interesting exceptions such as Raji cells [46], porcine erythrocytes [47], and blood cells from patients with Scott syndrome [18]. This Ca2+ activation of the scramblase may be specific to cells in the hematopoietic lineage, since the scramblase is difficult or impossible to activate directly with Ca2+ ionophores in cells such as fibroblasts or HeLa cells [48]. The physiological relevance of the Ca2+ activation pathway in hematopoietic cells is supported by the fact that loss of lipid asymmetry resulting from cellular activation via physiological agonists, as was first shown for platelets, is also associated with elevation of intracellular Ca2+, while lowering the Ca2+ levels causes the scrambling process to cease. Using a continuous assay based on fluorescent lipid analogs to monitor lipid scrambling, it was demonstrated that intracellular Ca2+ acts as a switch for the scrambling machinery; ionomycinmediated influx and efflux of Ca2+ in platelets allowed multiple cycles of activation and inactivation, respectively, of the scrambling process [27]. In erythrocytes, once activated by Ca2+, the scrambling pathway remains active for at least 2 hours [49]. When Ca2+ is removed using ethylene glycol tetraacetic acid (EGTA) as chelator, lipid scramblase is blocked and, provided the cell still contains sufficient ATP, aminophospholipid translocase is reactivated [50, 51]. This protocol restores lipid asymmetry only in the remnant cells and not in microvesicles that are produced during the scrambling process, likely because these latter structures lack the capacity to generate ATP [50]. Studies on platelets suggest that a transient elevation of intracellular Ca2+ is not sufficient to evoke lipid scrambling, as illustrated by a comparison of activation of platelets with thrombin or with thrombin plus collagen. Whereas thrombin alone causes intracellular Ca2+ concentration to reach a level close to 1â•¯µM, this agonist does not cause an appreciable loss of lipid asymmetry; on the contrary, aminophospholipid translocase activity is significantly enhanced in thrombinactivated platelets [52]. The elevation of Ca2+ in this protocol is transient because of rapid sequestration of Ca2+ into intracellular stores and pumping

Characteristics of the Phospholipid Scrambling Process

125

by plasma membrane Ca2+-ATPase activity. On the other hand, platelets activated by collagen plus thrombin exhibit a persistent, high intracellular Ca2+ concentration, and this treatment results in externalization of PS [53, 54]. The extracellular Ca2+ concentration required for half maximal scrambling activity as determined from the development of procoagulant activity on ionomycin-activated platelets was estimated to be about 15â•¯µM, although the intracellular Ca2+ concentration is likely to be less [55]. A continuous analysis of transbilayer movement, based on dithionite quenching of externalized 7-nitrobenz-2-oxa- 1,3-diazol-4-yl (NBD)-labeled phospholipids, suggested that the rate of ionomycin-induced PS externalization continued to increase when the Ca2+ concentration was raised from 50â•¯µM to 1â•¯mM [27]. For physiological agonists, the picture is more complicated since, as mentioned above, a persistent elevation of Ca2+ is more important than the maximally attained Ca2+ concentration. Keuren et al. determined that a sustained intracellular Ca2+ concentration of approximately 400â•¯nM is required for the combined action of collagen and thrombin to evoke collapse of lipid asymmetry [56]. Lipid scrambling in red blood cells (RBCs) requires a minimal extracellular Ca2+ concentration between 50 and 100â•¯µM, and perhaps less for resealed ghosts [49, 57]. For unknown reasons, external Ca2+ concentrations above 1â•¯mM appear to be slightly inhibitory [58]. Similarly, in lymphocytes, the rate of lipid scrambling increases in the range from 25 to 100â•¯µM with a slight decrease at 1â•¯mM [59]. The Ca2+ concentration required for activating lipid scrambling is considerably higher than that required for half maximal inhibition of the aminophospholipid translocase (IC50â•¯∼â•¯0.2–2â•¯µM [60]), suggesting that induction of lipid scrambling is different from the inhibition of the translocase. Apoptosis induces PS exposure, and when induced by anti-Fas antibodies or UV irradiation, the process requires the presence of extracellular Ca2+. Hampton et al. demonstrated that removal of extracellular Ca2+ with EGTA inhibited PS exposure in Jurkat cells, while other apoptosis markers, including bleb formation, shrinkage, and DNA laddering, were not affected by extracellular EGTA [61]. The process of PS externalization itself, however, is not itself Ca2+ dependent in apoptotic cells: Chelation of intracellular Ca2+ with 1,2-bis(o-aminophenoxy)-ethane-N,N,N′,N′-tetraacetoxymethyl ester (BAPTAAM) had no effect on PS externalization. This observation is consistent with the fact that the Scott mutation blocks scramblase activation by Ca2+ in a variety of cell types including lymphocytes, but does not block PS exposure in lymphocytes undergoing apoptosis. Bratton et al. found that 100â•¯µM extracellular Ca2+ (approximately 90â•¯nM intracellular Ca2+ concentration) is required for half maximal PS exposure in Jurkats undergoing UV-induced apoptosis [62]. These authors suggested that Ca2+ flux rather than a net change in intracellular Ca2+ is an essential prerequisite in these cells for apoptosis-induced lipid scrambling. This conclusion is consistent with results obtained using the lipid scrambling sensitive dye FM1-43 and the Ca2+ sensor Fura 2 in Jurkat cells. Wurth and Zweifach found that scrambling occurred in >99% of such

126

LOSS OF LIPID ASYMMETRY

cells in which the intracellular Ca2+ rose to 4â•¯µM [63]. However, when these same intracellular Ca2+ levels were induced by treating the cells with thapsigargin, a drug which depletes intracellular Ca2+ stores and induces storeoperated Ca2+ entry, egress of PS did not occur. Together, these results suggest that Ca2+ influx is required for the apoptotic process in these protocols, but unlike the case in platelets and other hematopoietic cells, it is not critical for activating the scramblase itself. 7.4.2â•… Effects of Other Cations on Lipid Scrambling In addition to convincing evidence that Ca2+ ion regulation is pivotal for activating the scrambling mechanism in hematopoietic cells, there is some evidence that monovalent cations or their gradients may have a modulating effect on lipid scrambling. For example, PS exposure in platelets, activated by agents such as collagen in combination with a thrombin receptor agonist, was abolished in Na+-free medium [64], and inhibitors of the Na+/H+ exchanger attenuated the procoagulant response of platelets [64, 65]. High extracellular K+ or specific K+ channel blockers also diminished PS exposure induced by Ca2+ in both platelets and RBCs [66–68], while valinomycin, a K+ ionophore, accelerated Ca2+-induced lipid scrambling in erythrocytes and abolished the inhibition by K+ channel blockers. Although manipulation of the intracellular K+ content might affect lipid scrambling indirectly via changes in morphology and volume of the cells, there is recent evidence that the effect of K+ ions on the rate of scrambling is direct [69]. 7.4.3â•… Rate of Lipid Scrambling in Various Cell Types The rate at which lipids become randomized after scramblase activation differs from cell to cell. Considering its central role in coagulation, it is not surprising that the highest rate of lipid scrambling is observed in platelets, where collapse of lipid asymmetry occurs within a minute after elevation of intracellular Ca2+. The existence of such a rapid rate was initially suggested by the observation that a procoagulant surface develops within a minute after treatment of platelets with ionomycin in the presence of extracellular Ca2+ [10, 70]. It seemed this might be an artifact of the stimulation method, since when platelets were stimulated with physiological agonists, such as collagen plus thrombin, the procoagulant activity increase occurred over a much longer period of 10–15 minutes before reaching an end stage. However, flow cytometric analysis of Annexin A5 binding to platelets showed that after physiological agonists, a fraction of platelets shows PS exposure comparable to the maximum observed after ionomycin treatment, and it is the size of this fraction that increases over the time period of 10–15 minutes [71, 72]. These observations emphasize that bulk measures of PS exposure, such as procoagulant activity or transbilayer movement of lipid analogs in cell populations, conflate the rate of lipid movement with the rate and efficiency of activation of this movement. Even with

Characteristics of the Phospholipid Scrambling Process

127

this caveat, the actual rates of scrambling harbor potentially interesting information. Since the characteristics of the scramblase process such as lipid specificity, directionality, and Ca2+ sensitivity are similar in the various cell types, it is likely that differences in scrambling rate reflect differences in the amount of scramblase protein per cell. With this possibility in mind, it is interesting to consider quantitative measurements of the scramblase rate in different cells [59]. In comparison with the highest rate constant found for platelets (78â•¯·â•¯10−3â•¯s−1), ionophore activation of scramblase in T cells and Epstein–Barr virus (EBV)-transformed B cells gave rate constants over a range from 1.0– 24â•¯·â•¯10−3â•¯s−1. Of all the cells in which scramblase is activated by Ca2+ and ionophore, erythrocytes have the slowest scrambling rate (rate constant, 0.45â•¯·â•¯10−3â•¯s−1). However, there is a large variation in scrambling activity in erythrocytes obtained from different human donors [73], and Ca2+-induced scramblase activity is completely absent in porcine RBCs (but not platelets) [47]. An interesting potential explanation for these observations might be a difference in the efficiency of protein sorting at the enucleation step of erythropoiesis. At this step, the cell remnant containing the nucleus is engulfed by macrophages in a process related to the PS-dependent recognition of apoptotic cells [74, 75]; restricting scramblase to the nucleated remnant cell would promote its PS-dependent recognition and engulfment by macrophages, whereas its exclusion from membrane surrounding the remnant reticulocyte would allow escape from this engulfment and entry into the circulation. The result would be a low level of scramblase in the red cell membrane; how low it finally reaches would reflect the efficiency of the scramblase segregation process at the point of creation of the reticulocyte. As indicated above, lipid scrambling is induced by Ca2+ in hematopoietic cells, and most spectacularly in platelets; it is also induced in a wide variety of cells by apoptosis. Are these the same process? Lymphoid cells provide a place to investigate this question, since both phenomena occur in these cells. Several observations suggest that they are related. For example, the rate of lipid scrambling differs in different lymphoid lines, but in any given line, the scrambling rate induced by apoptosis is roughly similar to that induced by Ca2+ ionophore [59]. In addition, scrambling is bidirectional and insensitive to lipid head group in both cases. Moreover, no increase in the rate of lipid scrambling is observed when apoptotic lymphoid cells were also treated with Ca2+ ionophore [59]. Although hardly definitive, these data suggest that the same membrane scramblase is activated by both apoptosis and Ca2+ ionophore. The processes are not the same, however. Apoptotically induced PS exposure in B cells from a patient with Scott syndrome could not be distinguished from that in normal B cells, whereas ionomycin-induced PS exposure was completely absent. At a minimum, this result implies that at least the activation pathways differ between apoptosis and ionophore treatment, and it is conceivable that the two mechanisms are completely distinct. Curiously, it has been observed that in Scott platelets, the rate and extent of development of procoagulant activity induced by collagen plus thrombin is approximately 30% of that in platelets

128

LOSS OF LIPID ASYMMETRY

from healthy controls [76], even though ionophore-induced PS exposure is completely blocked in Scott platelets. These results raise the possibility that the combination of collagen and thrombin activates some other, non-Ca2+dependent pathway for PS exposure in platelets, and such a pathway might well be related to that activated by apoptosis. 7.4.4â•… Comparison of Scrambling Rates between Various Lipids The earliest experiments investigating transbilayer rearrangement of phospholipids in activated platelets demonstrated that the composition of the hydrolyzed lipid fraction was always very similar to the total phospholipid composition, suggesting that the randomization process does not distinguish between different phospholipids. When spin-labeled and fluorescent (NBD)labeled phospholipid probes became available, in- and outward rates for each of the major lipid classes could be determined, and most studies showed similar in- and outward rates for PS, PC, and PE during lipid scrambling in platelets and erythrocytes. In fact, not only normal structural membrane lipids, but also lipids with unusual polar head groups, such as the D-isomer form of serine [77] or glucosyl- and galactosylceramides [78], and even lipid-like molecules such as platelet-activating factor (PAF) [79], lyso-PC [45], palmitoylcarnitine [45], and trimethylammonium-diphenylhexatriene (TMA-DPH) [80], participate in the scrambling process. There are exceptions to these findings of nonspecificity. For example, one study suggested that in platelets, PS and PE translocate to the external leaflet at a much higher rate than PC translocates to the inner leaflet [81]. In contrast, Dekkers et al. suggested that the outward movement of PE was slower than the outward movement of PC, but the inward rate of scrambling of NBD-PE exceeds that of NBD-PC [78]. Methylation of the amino group of PE made the kinetics of scrambling of this probe indistinguishable from that of PC, suggesting that the presence of an amino group affects the scramblase properties of a lipid molecule, although for NBD-PS, the difference in inward and outward rates was less pronounced. Different kinetics of lipid scrambling was also found for sphingolipids. The internalization rate of SM was reported to be appreciably lower than that of the other phospholipids classes [49, 77, 82], and it was speculated that this may lead to a mass imbalance with an excess of lipid in the outer leaflet, contributing to eversion of the plasma membrane that facilitates microvesicle formation. Interestingly, the inward movement of galactosylceramide was also found to be much slower than that of phosphatidylgalactose, suggesting that lipids with a ceramide backbone are less recognized by the scramblase [78]. However, the rate of Ca2+-induced outward movement of previously internalized NBD-SM and NBD-PC appeared indistinguishable [49]. A decreased rate of “inward” scrambling of ceramide-derived lipids might reflect association with lipid rafts, if dissociation from these rafts is slow or reduces the effective concentration of ceramide-derived lipids that can access the scramblase.

Characteristics of the Phospholipid Scrambling Process

129

The simplest model for a scramblase is one in which a transmembrane proteinaceous pore facilitates migration of the polar head group of the lipids across the hydrophobic core of the bilayer while keeping the acyl chain moieties in the core of the bilayer. If this model is correct, there should be some physical limit to the size of the head group that can access the proteinaceous core. Just such a limit was found by Dekkers et al., who used phospholipase D-catalyzed transphosphatidylation to prepare head-group-modified NBDlipid analogs. Studies of the scrambling of these analogs revealed a relation between size of the polar head group and the rate of inward movement [78], with decreased rates found in the order of phosphatidylgalactose, phosphatidylmaltose, and phosphatidylmaltotriose. Although these findings suggest that there is limit to the size of the opening from one leaflet to the other, they also suggest that the opening is quite large, raising the question of whether activation of the scramblase might also open a nonspecific pathway for ion flow. 7.4.5â•… ATP Dependence While maintenance of lipid asymmetry is an energy-requiring process, loss of lipid asymmetry is not. The simple fact that Ca2+-induced lipid scrambling can even occur in resealed “white” erythrocyte ghosts, devoid of cellular contents, implies that this process is not coupled to the hydrolysis of ATP [16, 17, 83]. However, prolonged incubation in glucose-free medium or in the presence of metabolic inhibitors, both resulting in severe ATP depletion, cause a gradual loss of Ca2+-induced scrambling activity in erythrocytes [25, 84]. Loss of scramblase activity was prevented when the phosphatase inhibitors, vanadate or okadaic acid, were present during ATP depletion [25], and activity could be restored upon ATP repletion [85]. On the other hand, staurosporin, a relatively nonselective kinase inhibitor, did not affect Ca2+-ionophore-induced lipid scrambling in RBCs [86]. These observations suggest that constitutively, phosphorylated components are involved in the scramblase mechanism. Interestingly, Ca2+-induced lipid scrambling in erythrocytes is accompanied by an increased tyrosine phosphorylation, mainly of band 3 protein [87], and this phosphorylation appeared to be defective in erythrocytes from patients with Scott syndrome [86]. A defective tyrosine phosphorylation, albeit less pronounced, was also observed upon activation of platelets from patients with Scott syndrome [86, 88–90]. Although it is tempting to speculate that the apparent phosphorylation defect caused the impaired lipid scrambling in these cells, subsequent experiments revealed that phosphorylation profiles in Scott erythrocyte ghosts after resealing in the presence of ATP were indistinguishable from those obtained from normal ghosts. It was concluded that the impaired tyrosine phosphorylation in Scott cells is an epiphenomenon that could be related to a defective lipid scrambling, but is not causal to this defect [86]. An interesting subject of more recent studies is an apoptosis-like process in erythrocytes termed “eryptosis.” Reviewed recently [91], this process seems

130

LOSS OF LIPID ASYMMETRY

to be related to the membrane events of apoptosis and includes scramblasemediated PS exposure and phagocytosis by macrophages. A large variety of agents induce this process, and Lang and coworkers [92] suggested that one pathway includes protein kinase C (PKC). In particular, glucose depletion causes a sequence of events in which increased intracellular Ca2+ activates a Ca2+-sensitive K+ channel (Gardos channel) leading to K+ efflux and cell shrinkage, followed by activation of a Ca2+-sensitive scramblase with subsequent PS exposure. Staurosporin, a PKC inhibitor, partially inhibited this process, while phorbol esther, a PKC activator, or okadaic acid, a phosphatase inhibitor, mimicked the effect of glucose depletion. How these results relate to earlier observations on ATP depletion is not clear, but they suggest that erythrocytes may be an interesting system for the investigation of the relationship between apoptotic and Ca2+-induced scramblase activation. 7.4.6â•… Critical Role of Free Sulfhydryl (SH) Groups and Disulfide Bonds in Lipid Scrambling A priori, a highly regulated and physiologically critical membrane transport process would seem likely to be protein mediated. Nevertheless, this point has been the focus of considerable attention over the years. One kind of evidence for the involvement of proteins would be sensitivity of the lipid scrambling process to thiol-reactive or reducing agents. In fact, treatment of platelets with the SH-reactive compounds diamide or pyridyldithioethylamine (PDA) is accompanied by an increased PS exposure, which can be reversed upon subsequent addition of permeable (dithiothreitol [DTT]), but not impermeable, (glutathione) reducing agents [71, 93]. This effect requires extracellular Ca2+ and is accompanied by microvesicle formation, suggesting that elevation of cytoplasmic Ca2+ is part of the process. Modification of free thiol groups by the alkylating agent N-ethylmaleimide (NEM) also induces exposure of PS in platelets; the mechanism of this effect seems to be different from the effects of the reversible inhibitors, since it does not require extracellular Ca2+ and is not accompanied by microvesicle formation [72, 94]. However, intracellular Ca2+ was not determined in these studies, and since NEM treatment of Scott platelets fails to expose PS [88], it remains likely that NEM treatment exposes PS by the same mechanism as that which is activated with collagen plus thrombin or Ca2+ ionophore. Interpretation of these experiments is complicated by observations that SH modification can also have an inhibitory effect on the scramblase. Williamson et al. showed that pretreatment of NBD-PS-labeled platelets with PDA results in a dramatic inhibition of Ca2+-ionophore-induced lipid movement [27]. Inhibition is partially abolished by addition of DTT, albeit at high concentration, suggesting that the thiol-reactive site in the scrambling mechanism is located in the core of the membrane where it would be accessible to the more hydrophobic PDA. This inhibitory effect of SH reactants on Ca2+-induced lipid scrambling may have been overlooked in other experiments because high

Characteristics of the Phospholipid Scrambling Process

131

concentrations and prolonged incubation with these reagents can induce PS exposure in the absence of Ca2+ ionophore. An inhibitory effect of SH-modifying reagents suggests that free SH groups are required for efficient PS exposure. However, there is also evidence that disulfides are also required. In RBCs, it was found that reduction of disulfide bonds by treatment with dithioerythritol (DTE) suppressed Ca2+-induced lipid scrambling [95]. On the other hand, other experiments suggest that free SH groups are not critical in erythrocytes since neither diamide nor PDA nor NEM was able to inhibit Ca2+-induced scrambling [46, 95–98]. Indeed, pretreatment of the cells with NEM, diamide, or other SH-modifying compounds exacerbates Ca2+-induced PS exposure. The confusion of results in the literature may have several explanations: (1) SH-reactive reagents block aminophospholipid translocase activity, which will prevent PS that leaks to the surface from being returned to the membrane inner leaflet; (2) SH modification may affect the ionomycin-induced activation mechanism; and (3) the scramblase itself may require either free sulfhydryls or disulfides, or both. Although ineffective on erythrocytes, NEM, PDA, or diamide treatment, in the absence of Ca2+ ionophore, resulted in PS externalization in various nucleated cells, including Raji cells. The latter are an interesting case because they are extremely sensitive to Fas-induced apoptosis without exposing PS, but PS exposure does occur after treatment with NEM or diamide [46]. Anti-Fasinduced PS exposure in Jurkat cells results from apoptosis activation, since it can be inhibited by the pan caspase inhibitor zVAD-fmk. In contrast, externalization of PS induced by SH reagents in the same cells probably results from activation of the Ca2+-dependent pathway. The NEM effect is not blocked by zVAD-fmk [96], both PDA and NEM caused an increase in intracellular Ca2+, and an intracellular Ca2+ chelator (EGTA-AM) inhibited NEM-induced PS exposure in Jurkat cells. SH-reactive reagents are not the only molecules that have been reported to affect lipid scrambling. For example, Ca2+-activated scrambling in erythrocytes is sensitive to the relatively high concentrations of the covalent inhibitor of the anion transport channel, 4,4′-diisothiocyanostilbene-2′,2′-disulfonate (DIDS) as well as 2,4,6-trinitrobenzoate (TNBS) [95]. The erythrocyte membrane is virtually impermeable to both compounds, and these findings therefore suggest that one or more exofacial amino groups of the scramblase are critical for activation of lipid scrambling. Inhibition by these reagents did not occur if they were added after activation of the scramblase [95], suggesting that the activation mechanism is the sensitive step. Finally, screening of a library of organic compounds yielded an irreversible inhibitor of the scramblase called R5421, with an IC50 of 35â•¯µM. High concentrations of this reagent were required—complete inhibition of Ca2+-ionophoreinduced scrambling in both erythrocytes and platelets required an inhibitor concentration of 100â•¯µM. Unfortunately, the mechanism of action of this inhibitor has not been elucidated, and it is not commercially available, making further studies difficult.

132

LOSS OF LIPID ASYMMETRY

7.5â•… TOWARD IDENTIFICATION: PROPOSED CANDIDATE PROTEINS AND MECHANISMS 7.5.1â•… PLSCR Two groups of researchers have attempted to use biochemical methods to isolate a functional scramblase protein. Starting from platelet membranes as a source with the highest scramblase activity, Comfurius et al. isolated a heterogeneous protein fraction that upon reconstitution produced proteoliposomes with Ca2+-inducible scrambling activity, with properties similar to those found in the parent platelets [99]. This activity was not purified further. In the same year, Bassé and coworkers used a similar approach starting from erythrocyte membranes, and identified a 36-kD protein that showed Ca2+-induced lipid scrambling when reconstituted in artificial lipid vesicles [100]. An internal peptide sequence of the purified protein led to cloning of the cDNA from the K562 cell line, and the gene was christened PLSCR1 [101]. Further analysis revealed that hPLSCR1 belongs to a family of four homologous proteins, hPLSCR1-4 [102] with orthologs in mouse, rat, yeast, Drosophila and Caenorhabditis elegans [103]. The protein consists of 318 amino acids with one potential transmembrane domain [101, 104]. A putative Ca2+-binding domain of 12 amino acids, resembling an EF-hand motif, was identified just proximal to the presumed transmembrane segment, and both light scatter and circular dichroism experiments suggested that binding of Ca2+ by this sequence results in large conformational changes, including possible self-association into oligomers [105]. Point mutations in this Ca2+-binding domain were accompanied by loss of scrambling activity [104]. One or more palmitoylated cysteinyl residues were described, suggesting an alternative anchoring point of the protein in the plasma membrane; hydrolysis of thioester bonds resulted in loss of activity in the reconstituted proteoliposomes [106]. A variety of cell biological level studies seemed to support the identification of this protein with the scramblase. Levels of PLSCR1 mRNA correlated reasonably well with scramblase activity in a variety of hematologic and nonhematologic cells and tissues [107]. Investigation of protein levels with PLSCR1-specific antibodies suggested that the number of copies of this protein in platelets is about 10-fold higher than the number in erythrocytes, in agreement with the observed difference in scrambling activity between these cells [101]. The low Ca2+-inducible scrambling activity of Raji cells, which are relatively low in PLSCR1, could be increased after stable transfection with PLSCR1 [107]. One observation of some interest was that the gene for PLSCR1 is normally expressed in patients with Scott syndrome [108, 109]; this result is not entirely unexpected if the primary defect in Scott syndrome is in the activation pathway rather than in the scramblase itself. Altogether, these observations provided circumstantial evidence that the protein product of the PLSCR1 gene is the phospholipid scramblase. However, some features of the protein activity were not consistent with this assignment,

Toward Identification

133

including the fact that the rate of Ca2+-induced lipid movement in the reconstituted vesicles was very slow, but not the same for PS and PC. In time, additional experiments showed that the identification of PLSCR1 as the scramblase was incorrect. An early indication of trouble came with the discovery that expression of PLSCR1 is under cytokine control, but that the upregulation of hPLSCR gene transcription by interferons (IFNs) had no effect on PS exposure [110]. In addition, a lack of correlation was found in six different cell lines between the level of expression of PLSCR1 and the capacity of these cells to externalize PS during apoptosis. Moreover, in contrast to previous findings [107], Fadeel et al. reported that overexpression of PLSCR1 in Raji cells failed to endow these cells with apoptosis-induced lipid scrambling [46]. A fatal result was the observation that platelets from PLSCR−/− mice have no hemostatic defects and expose PS normally upon cell activation, implying that this protein cannot be the scramblase. Similarly, a recent report mentioned that single or double deletion of two PLSCR homologs in Drosophila melanogaster did not compromise PS-mediated phagocytic clearance of apoptotic cells in vivo [111]. In addition, Zullig et al. used RNAi to knock down 10 annotated homologs of hPLSCR1 in C. elegans and found no effect on apoptosis-induced PS exposure [112]. Smrz et al. found that PLSCR1 in rat basophilic leukemia cells was released from the plasma membrane into the cytoplasm during both apoptosis and Ca2+-ionophore-induced PS exposure [113]. In conclusion, it is becoming increasingly evident that the PLSCR proteins are not phospholipid scramblases. Indeed, data have accumulated in recent years suggesting a role in cell signaling; PLSCR1 was found to shuttle between plasma membrane and nucleus, depending on its palmitoylation state and to regulate transcription of an IP3 receptor type 1 gene. Other possible functions of the PLSCR family of proteins have been discussed in recent reviews [103, 149]. These functions in hematopoietic cells may account for results from two recent studies suggesting that a nematode homolog of PLSCR1 plays a role in apoptosis-induced PS exposure and phagocytosis. Inactivation of the scrm-1 gene in C. elegans, the homolog closest in sequence to the hplscr1-4, led to a decreased PS exposure on the surface of apoptotic germ cells, compromising cell corpse engulfment [114]. The same study also showed reduced PS exposure and phagocytosis upon inactivation of the gene, which encodes for WAH1, a mitochondrial apoptotic factor, and the authors suggested that PS exposure is controlled by WAH-1 after its release from the mitochondria during apoptosis. Similarly, Venegas and Zhou selectively depleted nematode homologs of two putative candidate scramblase proteins, PLSCR1 and ABCA1, the latter will be discussed below. Their results suggested that the homolog of PLSCR1 is required for PS exposure on apoptotic germ cells, while CED-7, the homolog of ABCA1, does the same during apoptosis of somatic cells [115]. Whatever the role of these proteins in these two cases, the results do suggest that the mechanism of activation of PS exposure during apoptosis may depend on the type of tissue and developmental stage.

134

LOSS OF LIPID ASYMMETRY

7.5.2â•… ABCA1 ABCA1 is a member of the family of ATP-binding-cassette (ABC) transporter proteins that is required for the conversion of ApoA-1 to high-density lipoprotein (HDL) [116–118]. Mutations in this protein are responsible for Tangier disease, an autosomal recessive disorder causing dyslipidemia, mostly reflected by the virtual absence of HDL and low levels of low-density lipoprotein (LDL). In vivo loss of function and in vitro gain of function experiments suggested a role for ABCA1 in Ca2+-induced PS exposure [119], in line with evidence that engulfment of apoptotic cells is impaired in ABCA1-null mutants. (The significance of ABCA1 in lipid asymmetry and Tangier disease will be discussed in more detail in Chapter 17.) The nature of the ABCA1 activity was suggested by experiments showing that erythrocytes from ABCA1−/− mice have a reduced Ca2+-induced PS exposure, reflected by an impaired formation of procoagulant surface and a decreased externalization of spin-labeled PS. However, Ca2+-induced transbilayer movement of spin-labeled PC was not changed, suggesting that ABCA1 might be a “reversed” ATP-dependent PS translocase as proposed earlier by Bassé et al. to explain PS exposure in platelets [81]. A possible role for ABCA1 in scramblase-mediated PS exposure came from Albrecht et al., who found a single missense mutation in the ABCA1 gene of a patient with Scott syndrome in which lipid scrambling is compromised [120]. Overexpression of ABCA1 in Scott syndrome B lymphocytes complemented the Ca2+-dependent PS exposure at the cell surface. A similar conclusion might be drawn from a report that scramblase activity is affected in fibroblasts from ABCA1−/− mice [119]. Finally, it is interesting to note that generation of ABCA1-null mice was associated with high lethality during the first weeks after birth; autopsies revealed deep perivisceral hemorrhages suggesting that the mutation reduced the effectiveness of the coagulation process [119]. While these observations support a role for ABCA1 in the recognition of apoptotic cells, it is quite clear that this protein, like PLSCR1, is not the scramblase. Although disrupted platelet function and altered hemostatic parameters were found in patients with Tangier disease, PS exposure is normal in platelets from these patients after stimulation with collagen or Ca2+ ionophore [121]. Similarly, Elliott et al. found that PS translocation in ionophore-stimulated murine B cells is independent of ABCA1 [122]. Recently, the various options for a role for ABCA1 either as scramblase itself or as a regulator of scramblase have been (re)addressed experimentally and discussed critically in an extensive study by Williamson et al. [48]. The possibility that ABCA1 is or regulates the activity of the aminophospholipid translocase was experimentally disproved. Studies with ABCA1−/− mouse and human cells (Tangier disease) showed that scramblase is still present and active at normal levels in lymphocytes, fibroblasts, and macrophages. The alternative possibility that ABCA1 modulates the activity of the scramblase was rejected since no appreciable difference was observed in the time required for the onset of scramblase-

Toward Identification

135

mediated phospholipid movement between ABCA1-deficient cells and controls. Finally, the likelihood that ABCA1 operates as a reversed PS translocase next to a Ca2+-inducible scramblase seems remote. As mentioned above, the rates of scramblase-mediated PS and PC movement in various blood cells are quite similar [27, 49], a result difficult to reconcile with the suggestion that the rates of movement are different in normal as well as ABCA1−/− erythrocytes [119]. It is also worth noting that if externalization of PS requires movements mediated by two different activities, ABCA1 and scramblase, both activities must be impaired in cells from Scott syndrome patients. In summary, it is reasonably clear that neither PLSCR1 nor ABCA1 is the scramblase, but there is some reason to believe that a protein is involved in this function, if only from the observation that a defective scrambling mechanism is inherited by Scott syndrome patients. Nevertheless, our understanding of scrambling and its regulation will not progress far until specific proteins are identified that do play a direct role in this mechanism. In the absence of information on such proteins, it is not surprising that the involvement of proteinindependent mechanisms has been postulated, and these will be discussed below. 7.5.3â•… TMEM16F Recently, Suzuki et al. [123] identified the protein TMEM16F as an essential component for Ca2+-dependent exposure of PS. TMEM16F belongs to a family of Ca2+-dependent Cl− channels with eight transmembrane segments. A patient with Scott syndrome harbors a homozygous mutation at a splice-acceptor site that results in a premature termination of this protein in the third transmembrane domain [123]. A second patient with Scott syndrome was found to be a compound heterozygote with two novel TMEM16F mutations, one with a splice-acceptor site mutation resulting in exon 6 skipping and one with a single nucleotide insertion in exon 11 causing premature termination [150]. A separate mutation in TMEM16F enhances the sensitivity of the scrambling process to Ca2+ [123]. These results provide strong evidence that TMEM16F is the protein which is disabled in patients with Scott syndrome. Whether it also provides the pathway for phospholipid movement through the membrane is not yet clear, an obvious focus for future experiments. 7.5.4â•… Formation of Microvesicles It has long been recognized that influx of extracellular Ca2+ causes shedding of microvesicles from the plasma membrane of RBCs, platelets, and endothelial cells [50, 124–127]. The transbilayer lipid distribution is random in the membranes surrounding microvesicles derived from both platelets and erythrocytes, and the remnant cells from which the vesicles derive undergo a progressive loss of lipid asymmetry as the vesicles are shed [50]. Moreover,

136

LOSS OF LIPID ASYMMETRY

microvesicle formation is markedly decreased in platelets and erythrocytes from patients with Scott syndrome [83, 128]. The strong correlation between lipid scrambling and the shedding of membrane microvesicles led Sims and coworkers to propose that vesicle shedding itself might provide a mechanism for the loss of lipid asymmetry [128, 129]. The process of shedding of microvesicles from the plasma membrane requires fusion between opposing segments of cytoplasmic membrane surfaces, and transient formation of non-bilayer structures at the point of fusion might provide a pathway for local collapse of transbilayer lipid asymmetry. Several observations make the above-described model less plausible. Conditions that suppress microvesicle formation in both platelets and erythrocytes did not reduce PS exposure [45, 49, 94, 130–132]. Moreover, microvesicle shedding can be induced by incorporation of dimyristoylphosphatidylcholine (DMPC) into the outer leaflet of the erythrocyte membrane that nevertheless retain their asymmetric lipid distribution [133]. When these microvesicles are treated with Ca2+ ionophore, lipid asymmetry is rapidly lost, but morphology (size) is unchanged. This finding indicates that (1) these vesicles do contain a scramblase mechanism and (2) lipid scrambling can occur without shedding, and thus without membrane fusion events [134]. Indeed, as mentioned above, the rate of outward movement of PS may initially exceed the rate of inward movement of SM during lipid scrambling, and the resulting mass imbalance might cause exfoliation and subsequent shedding of the plasma membrane [81, 134]. The hypothesis that it is lipid scrambling that drives microvesicle formation would be consistent with the decreased microvesicle production observed in blood cells from Scott syndrome. It is not the only mechanism, however, Bucki et al. reported that normal or Scott erythrocyte ghosts, resealed in the presence of the polyamine spermine and then treated with Ca2+, produced microvesicles without concomitant PS exposure [130]. Although the mechanism by which spermine facilitates Ca2+-induced microvesicle formation is not known, these results suggest that phospholipid scrambling and microvesicles production, though closely related, are distinct phenomena. 7.5.5â•… A Role for Phosphatidylinositol Bisphosphate (PIP2) Sulpice and coworkers reported in 1994 that Ca2+-induced lipid scrambling is dependent on the presence of PIP2 in the external leaflet of inside-out vesicles (IOVs) derived from RBCs [84]. They also found a dose-dependent scrambling of spin-labeled phospholipids, when intact erythrocytes, loaded with different amounts of PIP2 in the outer leaflet, were incubated in the presence of extracellular Ca2+. The phenomenon was specific for PIP2, other phosphoinositides being ineffective. In a follow-up study, they showed that Ca2+ triggers a partial redistribution of spin-labeled phospholipids in protein-free large unilamellar vesicles (LUVs) [135]. Based on these observations, the authors suggested that a complex of Ca2+ ions and PIP2 causes destabilization of the bilayer and an accelerated transbilayer movement if lipids. Since increased intracellular Ca2+

Toward Identification

137

activates endogenous phospholipases, it was concluded that only the phospholipase-C-resistant PIP2 pool is involved. This hypothesis explained why severe ATP depletion results in a loss of Ca2+-induced scrambling, since this treatment reduces PIP2 content by 80%. It would also explain why spermine, a polyamine known to interact with PIP2, inhibits scrambling in both PIP2containing LUVs and PIP2-loaded RBC [135]. Shiffer et al. reported that addition of PIP2 to intact erythrocytes, even in the absence of extracellular Ca2+, led to a concentration-dependent increase in cytosolic cations (Ca2+ and or Zn2+) as measured with the Ca2+-fluorophore Fluo-3, and increased PS exposure, detected by increased binding of Annexin A5 [136]. They proposed that PIP2 or its complex with Ca2+ forms a pore through which both cations and phospholipids can cross the bilayer in either direction. However, such a direct and simple role for PIP2 in Ca2+-induced scrambling was ruled out by the finding that human Scott and normal porcine erythrocytes, both defective in Ca2+-induced lipid scrambling, have normal or even slightly increased (for porcine) levels of both total and metabolic resistant pools of PIP2 [47]. Loading Scott red cells with PIP2 could not rescue the impaired lipid scrambling, even in the presence of Ca2+ ionophore, and incorporation of 1% PIP2 in artificial lipid vesicles did not enhance transbilayer movement of NBD-phospholipids. Although there may be a role for PIP2 in scrambling, these results show that this phospholipid is not sufficient to support the scrambling mechanism. 7.5.6â•… A Role for Aminophospholipid Translocase One perennial hypothesis for scramblase activity is that it results from phospholipids moving through an inactivated aminophospholipid translocase. Of course, inhibition of this transporter is required for loss of lipid asymmetry, as the active enzyme will pump PS that reaches the surface back to the plasma membrane inner leaflet. Indeed, surface-exposed PS resulting from activation of scramblase can be returned to the cell interior by reactivation of aminophospholipid translocase activity in activated platelets and ionomycin-treated erythrocytes [50, 93]. However, mere inhibition of the aminophospholipid translocase by ATP depletion or vanadate inhibition is insufficient to cause loss of lipid asymmetry [23, 50, 51]. Kagan’s group suggested that apoptosis-induced PS exposure requires lipid oxidation [137–139] and, in particular, that oxidation of PS may be an essential component of the signaling pathway for PS externalization. A model was proposed in which oxidized PS acts as “suicide substrate” for aminophospholipid translocase or exhibits an intrinsic high spontaneous transbilayer flip-flop [137]. However, such a model does not explain an increased bidirectional transbilayer movement of other phospholipid classes. Intriguingly, a recent study by Zullig et al. suggested that PS exposure in C. elegans depends on one of the putative P-type aminophospholipid transporters [112]. Using a transgenic strain expressing a GFP::Annexin A5 reporter,

138

LOSS OF LIPID ASYMMETRY

they found that RNAi-mediated knockdown of one of the six genes encoding a P4 subfamily of P-type ATPases (transbilayer amphipath transporter, tat-1) in C. elegans abrogated externalization of PS in apoptotic cells. The authors speculated that TAT-1 displays distinct activities, an inward directed transport of aminophospholipids in healthy cells, whereas in apoptotic cells, perhaps due to modification, proteolysis, or interaction with other proteins, the directionality of transport is changed. However, the P4 subfamily of ATPases is involved in vesicular transport pathways, and indirect effects, such as on the delivery of scramblase to the cell surface, cannot be eliminated by such experiments. 7.5.7â•… Membrane Shape and Lipid Packing Since Sheetz and Singer first proposed their bilayer couple theory [140], it is well appreciated and experimentally confirmed that very small changes in lipid mass between the two leaflets of a membrane lead to membrane bending and could thus support processes such as endocytosis, exocytosis, and shedding of membrane-derived microparticles (reviewed in Reference 141). Recently, López-Montero and coworkers demonstrated that in situ conversion of SM to ceramide in artificial lipid bilayers (giant unilamellar vesicles) may lead to the formation of large transition pores [142]. It was suggested that loss of the polar head group by the action of exogenous sphingomyelinase causes the outer leaflet to shrink in surface area, resulting in membrane bending and invagination that may eventually result in the formation of a pore along the edges of which lipids might diffuse between the two leaflets of the membrane. This model would explain previous findings that exogenous added or in situ generated ceramides caused increased scrambling of glycerophospholipids in artificial bilayers [143] as well as the observation that treatment of human platelets with sphingomyelinase caused an increased procoagulant activity, indicating exposure of PS [144]. These observations may also explain the close association between lipid scrambling and ceramide production in apoptosis [145]. Differences in surface tension between the two leaflets of the plasma membrane may also be the driving force for cell-volume-related changes in lipid asymmetry. Both PS exposure and cell shrinkage are hallmarks of apoptosis [146, 147]. Ca2+-induced lipid scrambling in B lymphocytes causes externalization of PS, which appears to be preceded by and dependent on a decrease in lipid packing as detected by an increased binding of merocyanin 540 [122]. In a hypothetical model, Elliott et al. postulated that membrane deformations resulting from cell shrinkage create domains with low and high packing density at the apex and base, respectively, of blebs [122]. They argued that this arrangement would favor outward movement of lipids at the apex and inward movement at the base, in agreement with the observed characteristics of the scramblase process, including a bidirectional movement of all phospholipid classes. It could be speculated that in this model, lipid scrambling may be independent of a protein mediator. In support of this “shrinkage” model, Schneider et al. recently demonstrated a reciprocal correlation between

Abbreviations

139

forward scatter (cell volume) and Annexin A5 binding (PS exposure) in erythrocytes, irrespective of the conditions to manipulate cell volume (changing osmolarity or causing influx or efflux of K+ ions) [148]. The observation that lipid scrambling in resealed erythrocyte ghosts can occur without appreciable changes in cell morphology or cell volume argues against the model described above [69]. On the other hand, membrane morphology does influence the scrambling process. Manipulating the erythrocyte shape with a variety of lipidsoluble compounds demonstrated that lipid scrambling is enhanced in echinocytes, whereas stomatocytes show a reduced rate of lipid scrambling in comparison to normocytes [73]. 7.6â•… CONCLUDING REMARKS Lipid scrambling, with surface exposure of PS as its most prominent consequence, is a major event in several physiological and pathophysiological processes. There seem to be multiple pathways that lead to a collapse of lipid asymmetry, but little concrete information is available on the actual mechanism of lipid scrambling. Many characteristics of the putative scramblase have been documented, but, with the exception of TMEM16F, attempts to identify protein(s) that can directly or indirectly act as scramblase have as yet been unsuccessful. There can be no question that the single most important problem is to identify specific proteins that are involved in the scramblase mechanism. One approach that deserves attention is to make use of modern proteomic tools in experiments comparing cells from Scott syndrome patients and healthy controls to identify proteins associated with TMEM16F, the protein that is inactivated by the disease. In the end, concentrating on the identity of responsible proteins will resolve the mechanism of lipid scrambling. Understanding this mechanism will undoubtedly provide ways to manage pathophysiological consequences of PS exposure in the clinic. ABBREVIATIONS BAPTA-AM 1,2-bis-(o-aminophenoxy)-ethane-N,N,N′,N′tetraacetoxymethyl ester DIDS 4,4′-diisothiocyanostilbene-2′,2′-disulfonate DTE dithioerythritol DTT dithiothreitol EGTA ethylene glycol tetraacetic acid IOV inside-out vesicle LUV large unilamellar vesicle NBD 7-nitrobenz-2-oxa- 1,3-diazol-4-yl NEM N-ethylmaleimide

140

PC PDA PE PIP2 PLSCR PS RBC SH SM TNBS

LOSS OF LIPID ASYMMETRY

phosphatidylcholine pyridyldithioethylamine phosphatidylethanolamine phosphatidylinositol bisphosphate phospholipid scramblase phosphatidylserine red blood cell sulfhydryl sphingomyelin trinitrobenzene sulfonic acid

REFERENCES â•… 1â•… D. L. Daleke, J. Biol. Chem. 2007, 282, 821–825. â•… 2â•… D. W. Dekkers, P. Comfurius, R. G. van Gool, E. M. Bevers, R. F. Zwaal, Biochem. J. 2000, 350, 531–535. â•… 3â•… A. Sohnius, D. Kamp, C. W. Haest, Mol. Membr. Biol. 2003, 20, 299–305. â•… 4â•… R. D. Kornberg, H. M. McConnell, Biochemistry 1971, 10, 1111–1120. â•… 5â•… B. de Kruijff, E. J. van Zoelen, L. L. van Deenen, Biochim. Biophys. Acta 1978, 509, 537–542. â•… 6â•… J. Bai, R. E. Pagano, Biochemistry 1997, 36, 8840–8848. â•… 7â•… A. Papadopulos, S. Vehring, I. López-Montero, L. Kutschenko, M. Stockl, P. F. Devaux, M. Kozlov, T. Pomorski, A. Herrmann, J. Biol. Chem. 2007, 282, 15559–15568. â•… 8â•… M. A. Kol, A. I. de Kroon, J. A. Killian, B. de Kruijff, Biochemistry 2004, 43, 2673–2681. â•… 9â•… P. K. Schick, K. B. Kurica, G. K. Chacko, J. Clin. Invest. 1976, 57, 1221–1226. â•‡ 10â•… E. M. Bevers, P. Comfurius, R. F. Zwaal, Biochim. Biophys. Acta 1983, 736, 57–66. â•‡ 11â•… R. F. Zwaal, P. Comfurius, E. M. Bevers, Biochim. Biophys. Acta 1998, 1376, 433–453. â•‡ 12â•… E. M. Bevers, P. Comfurius, J. L. van Rijn, H. C. Hemker, R. F. Zwaal, Eur. J. Biochem. 1982, 122, 429–436. â•‡ 13â•… J. Connor, C. Bucana, I. J. Fidler, A. J. Schroit, Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 3184–3188. â•‡ 14â•…V. A. Fadok, D. R. Voelker, P. A. Campbell, J. J. Cohen, D. L. Bratton, P. M. Henson, J. Immunol. 1992, 148, 2207–2216. â•‡ 15â•… R. Chandra, P. C. Joshi, V. K. Bajpai, C. M. Gupta, Biochim. Biophys. Acta 1987, 902, 253–262. â•‡ 16â•… J. Connor, K. Gillum, A. J. Schroit, Biochim. Biophys. Acta 1990, 1025, 82–86. â•‡ 17â•… P. Williamson, L. Algarin, J. Bateman, H. R. Choe, R. A. Schlegel, J. Cell. Physiol. 1985, 123, 209–214.

References

141

â•‡ 18â•… R. F. Zwaal, P. Comfurius, E. M. Bevers, Biochim. Biophys. Acta 2004, 1636, 119–128. â•‡ 19â•… M. B. Brooks, J. L. Catalfamo, H. A. Brown, P. Ivanova, J. Lovaglio, Blood 2002, 99, 2434–2441. â•‡ 20â•… P. Williamson, K. Mattocks, R. A. Schlegel, Biochim. Biophys. Acta 1983, 732, 387–393. â•‡ 21â•… P. Thiagarajan, J. F. Tait, J. Biol. Chem. 1990, 265, 17420–17423. â•‡ 22â•… J. Shi, Y. Shi, L. N. Waehrens, J. T. Rasmussen, C. W. Heegaard, G. E. Gilbert, Cytometry A 2006, 69, 1193–1201. â•‡ 23â•… L. Tilley, S. Cribier, B. Roelofsen, J. A. Op den Kamp, L. L. van Deenen, FEBS Lett. 1986, 194, 21–27. â•‡ 24â•… â•‡ 25â•… â•‡ 26â•… â•‡ 27â•… â•‡ 28â•… â•‡ 29â•… â•‡ 30â•… â•‡ 31â•…

M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. O. C. Martin, R. E. Pagano, J. Biol. Chem. 1987, 262, 5890–5898. J. Connor, A. J. Schroit, Biochemistry 1988, 27, 848–851. P. Williamson, E. M. Bevers, E. F. Smeets, P. Comfurius, R. A. Schlegel, R. F. Zwaal, Biochemistry 1995, 34, 10448–10455. J. Connor, A. J. Schroit, Biochemistry 1987, 26, 5099–5105. R. A. Schlegel, P. Williamson, Cell Death Differ. 2001, 8, 551–563. K. Balasubramanian, A. J. Schroit, Annu. Rev. Physiol. 2003, 65, 701–734. E. M. Bevers, P. Comfurius, D. W. Dekkers, M. Harmsma, R. F. Zwaal, Biol. Chem. 1998, 379, 973–986.

â•‡ 32â•… U. S. Gaipl, A. Kuhn, A. Sheriff, L. E. Munoz, S. Franz, R. E. Voll, J. R. Kalden, M. Herrmann, Curr. Dir. Autoimmun. 2006, 9, 173–187. â•‡ 33â•… K. S. Ravichandran, U. Lorenz, Nat. Rev. Immunol. 2007, 7, 964–974. â•‡ 34â•… M. Bonomini, V. Sirolli, F. Gizzi, S. D. Stante, A. Grilli, M. Felaco, Kidney Int. 2002, 62, 1358–1363. â•‡ 35â•… S. Eda, I. W. Sherman, Cell. Physiol. Biochem. 2002, 12, 373–384. â•‡ 36â•… A. B. Manodori, G. A. Barabino, B. H. Lubin, F. A. Kuypers, Blood 2000, 95, 1293–1300. â•‡ 37â•… B. N. Setty, S. Kulkarni, M. J. Stuart, Blood 2002, 99, 1564–1571. â•‡ 38â•… J. L. Wanderley, M. E. Moreira, A. Benjamin, A. C. Bonomo, M. A. Barcinski, J. Immunol. 2006, 176, 1834–1839. â•‡ 39â•… M. K. Callahan, P. M. Popernack, S. Tsutsui, L. Truong, R. A. Schlegel, A. J. Henderson, J. Immunol. 2003, 170, 4840–4845. â•‡ 40â•… J. Mercer, A. Helenius, Science 2008, 320, 531–535. â•‡ 41â•… S. M. van den Eijnde, M. J. van den Hoff, C. P. Reutelingsperger, W. L. van Heerde, M. E. Henfling, C. Vermeij-Keers, B. Schutte, M. Borgers, F. C. Ramaekers, J. Cell. Sci. 2001, 114, 3631–3642. â•‡ 42â•… B. Huppertz, C. Bartz, M. Kokozidou, Micron 2006, 37, 509–517. â•‡ 43â•… R. A. Harrison, B. M. Gadella, Theriogenology 2005, 63, 342–351. â•‡ 44â•… H. Kenis, H. van Genderen, A. Bennaghmouch, H. A. Rinia, P. Frederik, J. Narula, L. Hofstra, C. P. Reutelingsperger, J. Biol. Chem. 2004, 279, 52623–52629. â•‡ 45â•… U. Henseleit, G. Plasa, C. Haest, Biochim. Biophys. Acta 1990, 1029, 127–135.

142

LOSS OF LIPID ASYMMETRY

â•‡ 46â•… B. Fadeel, B. Gleiss, K. Hogstrand, J. Chandra, T. Wiedmer, P. J. Sims, J. I. Henter, S. Orrenius, A. Samali, Biochem. Biophys. Res. Commun. 1999, 266, 504–511. â•‡ 47â•… E. M. Bevers, T. Wiedmer, P. Comfurius, J. Zhao, E. F. Smeets, R. A. Schlegel, A. J. Schroit, H. J. Weiss, P. Williamson, R. F. Zwaal, P. J. Sims, Blood 1995, 86, 1983–1991. â•‡ 48â•… P. Williamson, M. S. Halleck, J. Malowitz, S. Ng, X. Fan, S. Krahling, A. T. Remaley, R. A. Schlegel, PLoS One 2007, 2, e729. â•‡ 49â•… P. Williamson, A. Kulick, A. Zachowski, R. A. Schlegel, P. F. Devaux, Biochemistry 1992, 31, 6355–6360. â•‡ 50â•… P. Comfurius, J. M. Senden, R. H. Tilly, A. J. Schroit, E. M. Bevers, R. F. Zwaal, Biochim. Biophys. Acta 1990, 1026, 153–160. â•‡ 51â•… B. Verhoven, R. A. Schlegel, P. Williamson, Biochim. Biophys. Acta 1992, 1104, 15–23. â•‡ 52â•… R. H. Tilly, J. M. Senden, P. Comfurius, E. M. Bevers, R. F. Zwaal, Biochim. Biophys. Acta 1990, 1029, 188–190. â•‡ 53â•… I. C. Munnix, M. Harmsma, J. C. Giddings, P. W. Collins, M. A. Feijge, P. Comfurius, J. W. Heemskerk, E. M. Bevers, Thromb. Haemost. 2003, 89, 687–695. â•‡ 54â•… E. F. Smeets, J. W. Heemskerk, P. Comfurius, E. M. Bevers, R. F. Zwaal, Thromb. Haemost. 1993, 70, 1024–1029. â•‡ 55â•… P. F. Verhallen, E. M. Bevers, P. Comfurius, R. F. Zwaal, Biochim. Biophys. Acta 1987, 903, 206–217. â•‡ 56â•… J. F. Keuren, S. J. Wielders, H. Ulrichts, T. Hackeng, J. W. Heemskerk, H. Deckmyn, E. M. Bevers, T. Lindhout, Arterioscler. Thromb. Vasc. Biol. 2005, 25, 1499–1505. â•‡ 57â•… L. A. Woon, J. W. Holland, E. P. Kable, B. D. Roufogalis, Cell Calcium 1999, 25, 313–320. â•‡ 58â•… K. de Jong, F. A. Kuypers, Br. J. Haematol. 2006, 133, 427–432. â•‡ 59â•… P. Williamson, A. Christie, T. Kohlin, R. A. Schlegel, P. Comfurius, M. Harmsma, R. F. Zwaal, E. M. Bevers, Biochemistry 2001, 40, 8065–8072. â•‡ 60â•… M. Bitbol, P. Fellmann, A. Zachowski, P. F. Devaux, Biochim. Biophys. Acta 1987, 904, 268–282. â•‡ 61â•… M. B. Hampton, D. M. Vanags, M. I. Porn-Ares, S. Orrenius, FEBS Lett. 1996, 399, 277–282. â•‡ 62â•… D. L. Bratton, V. A. Fadok, D. A. Richter, J. M. Kailey, L. A. Guthrie, P. M. Henson, J. Biol. Chem. 1997, 272, 26159–26165. â•‡ 63â•… G. A. Wurth, A. Zweifach, Biochem. J. 2002, 362, 701–708. â•‡ 64â•… R. Bucki, J. J. Pastore, F. Giraud, P. A. Janmey, J. C. Sulpice, Biochim. Biophys. Acta 2006, 1761, 195–204. â•‡ 65â•… J. Samson, H. Stelmach, M. Tomasiak, Platelets 2001, 12, 436–442. â•‡ 66â•… K. de Jong, S. K. Larkin, L. A. Styles, R. M. Bookchin, F. A. Kuypers, Blood 2001, 98, 860–867. â•‡ 67â•… P. A. Lang, S. Kaiser, S. Myssina, T. Wieder, F. Lang, S. M. Huber, Am. J. Physiol. Cell Physiol. 2003, 285, C1553–C1560. â•‡ 68â•… J. L. Wolfs, S. J. Wielders, P. Comfurius, T. Lindhout, J. C. Giddings, R. F. Zwaal, E. M. Bevers, Blood 2006, 108, 2223–2228.

References

143

â•‡ 69â•… J. L. Wolfs, P. Comfurius, O. Bekers, R. F. Zwaal, K. Balasubramanian, A. J. Schroit, T. Lindhout, E. M. Bevers, Cell. Mol. Life Sci. 2009, 66, 314–323. â•‡ 70â•… J. Rosing, J. L. van Rijn, E. M. Bevers, G. van Dieijen, P. Comfurius, R. F. Zwaal, Blood 1985, 65, 319–332. â•‡ 71â•… J. Dachary-Prigent, J. M. Freyssinet, J. M. Pasquet, J. C. Carron, A. T. Nurden, Blood 1993, 81, 2554–2565. â•‡ 72â•… J. L. Wolfs, P. Comfurius, J. T. Rasmussen, J. F. Keuren, T. Lindhout, R. F. Zwaal, E. M. Bevers, Cell. Mol. Life Sci. 2005, 62, 1514–1525. â•‡ 73â•… J. L. Wolfs, P. Comfurius, E. M. Bevers, R. F. Zwaal, Mol. Membr. Biol. 2003, 20, 83–91. â•‡ 74â•… P. L. Williamson, W. A. Massey, B. M. Phelps, R. A. Schlegel, Mol. Cell. Biol. 1981, 1, 128–135. â•‡ 75â•… H. Yoshida, K. Kawane, M. Koike, Y. Mori, Y. Uchiyama, S. Nagata, Nature 2005, 437, 754–758. â•‡ 76â•… J. Rosing, E. M. Bevers, P. Comfurius, H. C. Hemker, G. van Dieijen, H. J. Weiss, R. F. Zwaal, Blood 1985, 65, 1557–1561. â•‡ 77â•… E. F. Smeets, P. Comfurius, E. M. Bevers, R. F. Zwaal, Biochim. Biophys. Acta 1994, 1195, 281–286. â•‡ 78â•… D. W. Dekkers, P. Comfurius, E. M. Bevers, R. F. Zwaal, Biochem. J. 2002, 362, 741–747. â•‡ 79â•… D. L. Bratton, J. M. Kailey, K. L. Clay, P. M. Henson, Biochim. Biophys. Acta 1991, 1062, 24–34. â•‡ 80â•… E. M. Bevers, P. F. Verhallen, A. J. Visser, P. Comfurius, R. F. Zwaal, Biochemistry 1990, 29, 5132–5137. â•‡ 81â•… F. Bassé, P. Gaffet, F. Rendu, A. Bienvenue, Biochemistry 1993, 32, 2337–2344. â•‡ 82â•… C. W. Haest, D. Kamp, B. Deuticke, Biochim. Biophys. Acta 1997, 1325, 17–33. â•‡ 83â•… E. M. Bevers, T. Wiedmer, P. Comfurius, S. J. Shattil, H. J. Weiss, R. F. Zwaal, P. J. Sims, Blood 1992, 79, 380–388. â•‡ 84â•… J. C. Sulpice, A. Zachowski, P. F. Devaux, F. Giraud, J. Biol. Chem. 1994, 269, 6347–6354. â•‡ 85â•… E. M. Bevers, P. Comfurius, D. W. Dekkers, R. F. Zwaal, Biochim. Biophys. Acta 1999, 1439, 317–330. â•‡ 86â•… D. W. Dekkers, P. Comfurius, W. M. Vuist, J. T. Billheimer, I. Dicker, H. J. Weiss, R. F. Zwaal, E. M. Bevers, Blood 1998, 91, 2133–2138. â•‡ 87â•… G. Minetti, G. Piccinini, C. Balduini, C. Seppi, A. Brovelli, Biochem. J. 1996, 320(Pt 2), 445–450. â•‡ 88â•… J. Dachary-Prigent, J. M. Pasquet, E. Fressinaud, F. Toti, J. M. Freyssinet, A. T. Nurden, Br. J. Haematol. 1997, 99, 959–967. â•‡ 89â•… N. Imam-Sghiouar, I. Laude-Lemaire, V. Labas, D. Pflieger, J. P. L. Caer, M. Caron, D. K. Nabias, R. Joubert-Caron, Proteomics 2002, 2, 828–838. â•‡ 90â•… M. C. Martinez, S. Martin, F. Toti, E. Fressinaud, J. Dachary-Prigent, D. Meyer, J. M. Freyssinet, Biochemistry 1999, 38, 10092–10098. â•‡ 91â•… M. Foller, S. M. Huber, F. Lang, IUBMB Life 2008, 60, 661–668.

144

LOSS OF LIPID ASYMMETRY

â•‡ 92â•… B. A. Klarl, P. A. Lang, D. S. Kempe, O. M. Niemoeller, A. Akel, M. Sobiesiak, K. Eisele, M. Podolski, S. M. Huber, T. Wieder, F. Lang, Am. J. Physiol. Cell Physiol. 2006, 290, C244–C253. â•‡ 93â•… E. M. Bevers, R. H. Tilly, J. M. Senden, P. Comfurius, R. F. Zwaal, Biochemistry 1989, 28, 2382–2387. â•‡ 94â•… J. Dachary-Prigent, J. M. Pasquet, J. M. Freyssinet, A. T. Nurden, Biochemistry 1995, 34, 11625–11634. â•‡ 95â•… D. Kamp, T. Sieberg, C. W. Haest, Biochemistry 2001, 40, 9438–9446. â•‡ 96â•… K. Balasubramanian, B. Mirnikjoo, A. J. Schroit, J. Biol. Chem. 2007, 282, 18357–18364. â•‡ 97â•… K. de Jong, D. Geldwerth, F. A. Kuypers, Biochemistry 1997, 36, 6768–6776. â•‡ 98â•… G. Martin, O. Sabido, P. Durand, R. Levy, Hum. Reprod. 2005, 20, 3459–3468. â•‡ 99â•… P. Comfurius, P. Williamson, E. F. Smeets, R. A. Schlegel, E. M. Bevers, R. F. Zwaal, Biochemistry 1996, 35, 7631–7634. 100â•… F. Bassé, J. G. Stout, P. J. Sims, T. Wiedmer, J. Biol. Chem. 1996, 271, 17205–17210. 101â•… Q. Zhou, J. Zhao, J. G. Stout, R. A. Luhm, T. Wiedmer, P. J. Sims, J. Biol. Chem. 1997, 272, 18240–18244. 102â•… T. Wiedmer, Q. Zhou, D. Y. Kwoh, P. J. Sims, Biochim. Biophys. Acta 2000, 1467, 244–253. 103â•… S. K. Sahu, S. N. Gummadi, N. Manoj, G. K. Aradhyam, Arch. Biochem. Biophys. 2007, 462, 103–114. 104â•… Q. Zhou, P. J. Sims, T. Wiedmer, Biochemistry 1998, 37, 2356–2360. 105â•… J. G. Stout, Q. Zhou, T. Wiedmer, P. J. Sims, Biochemistry 1998, 37, 14860–14866. 106â•… J. Zhao, Q. Zhou, T. Wiedmer, P. J. Sims, Biochemistry 1998, 37, 6361–6366. 107â•… J. Zhao, Q. Zhou, T. Wiedmer, P. J. Sims, J. Biol. Chem. 1998, 273, 6603–6606. 108â•… Q. Zhou, P. J. Sims, T. Wiedmer, Blood 1998, 92, 1707–1712. 109â•… N. Janel, C. Leroy, I. Laude, F. Toti, E. Fressinaud, D. Meyer, J. M. Freyssinet, D. Kerbiriou-Nabias, Thromb. Haemost. 1999, 81, 322–323. 110â•… Q. S. Zhou, J. Zhao, F. A. Zoghaibi, A. M. Zhou, T. Wiedmer, R. H. Silverman, P. J. Sims, Blood 2000, 95, 2593–2599. 111â•… U. Acharya, M. B. Edwards, R. A. Jorquera, H. Silva, K. Nagashima, P. Labarca, J. K. Acharya, J. Cell Biol. 2006, 173, 69–82. 112â•… S. Zullig, L. J. Neukomm, M. Jovanovic, S. J. Charette, N. N. Lyssenko, M. S. Halleck, C. P. Reutelingsperger, R. A. Schlegel, M. O. Hengartner, Curr. Biol. 2007, 17, 994–999. 113â•… D. Smrz, P. Lebduska, L. Draberova, J. Korb, P. Draber, J. Biol. Chem. 2008, 283, 10904–10918. 114â•… X. Wang, J. Wang, K. Gengyo-Ando, L. Gu, C. L. Sun, C. Yang, Y. Shi, T. Kobayashi, Y. Shi, S. Mitani, X. S. Xie, D. Xue, Nat. Cell Biol. 2007, 9, 541–549. 115â•…V. Venegas, Z. Zhou, Mol. Biol. Cell. 2007, 18, 3180–3192. 116â•… M. Bodzioch, E. Orso, J. Klucken, T. Langmann, A. Bottcher, W. Diederich, W. Drobnik, S. Barlage, C. Buchler, M. Porsch-Ozcurumez, W. E. Kaminski, H. W. Hahmann, K. Oette, G. Rothe, C. Aslanidis, K. J. Lackner, G. Schmitz, Nat. Genet. 1999, 22, 347–351.

References

145

117â•… A. Brooks-Wilson, M. Marcil, S. M. Clee, L. H. Zhang, K. Roomp, M. van Dam, L. Yu, C. Brewer, J. A. Collins, H. O. Molhuizen, O. Loubser, B. F. Ouelette, K. Fichter, K. J. Ashbourne-Excoffon, C. W. Sensen, S. Scherer, S. Mott, M. Denis, D. Martindale, J. Frohlich, K. Morgan, B. Koop, S. Pimstone, J. J. Kastelein, J. Genest, Jr., M. R. Hayden, Nat. Genet. 1999, 22, 336–345. 118â•… S. Rust, M. Rosier, H. Funke, J. Real, Z. Amoura, J. C. Piette, J. F. Deleuze, H. B. Brewer, N. Duverger, P. Denefle, G. Assmann, Nat. Genet. 1999, 22, 352–355. 119â•… Y. Hamon, C. Broccardo, O. Chambenoit, M. F. Luciani, F. Toti, S. Chaslin, J. M. Freyssinet, P. F. Devaux, J. McNeish, D. Marguet, G. Chimini, Nat. Cell Biol. 2000, 2, 399–406. 120â•… C. Albrecht, J. H. McVey, J. I. Elliott, A. Sardini, I. Kasza, A. D. Mumford, R. P. Naoumova, E. G. Tuddenham, K. Szabo, C. F. Higgins, Blood 2005, 106, 542–549. 121â•… J. R. Nofer, G. Herminghaus, M. Brodde, E. Morgenstern, S. Rust, T. Engel, U. Seedorf, G. Assmann, H. Bluethmann, B. E. Kehrel, J. Biol. Chem. 2004, 279, 34032–34037. 122â•… J. I. Elliott, A. Sardini, J. C. Cooper, D. R. Alexander, S. Davanture, G. Chimini, C. F. Higgins, Blood 2006, 108, 1611–1617. 123â•… J. Suzuki, M. Umeda, P. J. Sims, S. Nagata, Nature 2010, 468, 834–838. 124â•… D. Allan, R. H. Michell, Nature 1975, 258, 348–349. 125â•… D. Allan, R. H. Michell, Biochem. J. 1977, 166, 495–499. 126â•… H. Sandberg, A. P. Bode, F. A. Dombrose, M. Hoechli, B. R. Lentz, Thromb. Res. 1985, 39, 63–79. 127â•… K. K. Hamilton, R. Hattori, C. T. Esmon, P. J. Sims, J. Biol. Chem. 1990, 265, 3809–3814. 128â•… P. J. Sims, T. Wiedmer, C. T. Esmon, H. J. Weiss, S. J. Shattil, J. Biol. Chem. 1989, 264, 17049–17057. 129â•… C. P. Chang, J. Zhao, T. Wiedmer, P. J. Sims, J. Biol. Chem. 1993, 268, 7171–7178. 130â•… R. Bucki, C. Bachelot-Loza, A. Zachowski, F. Giraud, J. C. Sulpice, Biochemistry 1998, 37, 15383–15391. 131â•… P. Gaffet, N. Bettache, A. Bienvenue, Eur. J. Cell Biol. 1995, 67, 336–345. 132â•… P. Williamson, J. Bateman, K. Kozarsky, K. Mattocks, N. Hermanowicz, H. R. Choe, R. A. Schlegel, Cell 1982, 30, 725–733. 133â•… Z. Beleznay, A. Zachowski, P. F. Devaux, P. Ott, Eur. J. Biochem. 1997, 243, 58–65. 134â•… R. F. Zwaal, P. Comfurius, E. M. Bevers, Biochem. Soc. Trans. 1993, 21, 248–253. 135â•… J. C. Sulpice, C. Moreau, P. F. Devaux, A. Zachowski, F. Giraud, Biochemistry 1996, 35, 13345–13352. 136â•… K. A. Shiffer, L. Rood, R. K. Emerson, F. A. Kuypers, Biochemistry 1998, 37, 3449–3458. 137â•…V. E. Kagan, G. G. Borisenko, B. F. Serinkan, Y. Y. Tyurina, V. A. Tyurin, J. Jiang, S. X. Liu, A. A. Shvedova, J. P. Fabisiak, W. Uthaisang, B. Fadeel, Am. J. Physiol. Lung Cell Mol. Physiol. 2003, 285, L1–L17. 138â•…V. E. Kagan, B. Gleiss, Y. Y. Tyurina, V. A. Tyurin, C. Elenstrom-Magnusson, S. X. Liu, F. B. Serinkan, A. Arroyo, J. Chandra, S. Orrenius, B. Fadeel, J. Immunol. 2002, 169, 487–499.

146

LOSS OF LIPID ASYMMETRY

139â•… Y. Y. Tyurina, F. B. Serinkan, V. A. Tyurin, V. Kini, J. C. Yalowich, A. J. Schroit, B. Fadeel, V. E. Kagan, J. Biol. Chem. 2004, 279, 6056–6064. 140â•… M. P. Sheetz, S. J. Singer, Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 4457–4461. 141â•… P. F. Devaux, A. Herrmann, N. Ohlwein, M. M. Kozlov, Biochim. Biophys. Acta 2008, 1778, 1591–1600. 142â•… I. López-Montero, M. Velez, P. F. Devaux, Biochim. Biophys. Acta 2007, 1768, 553–561. 143â•… F. X. Contreras, G. Basanez, A. Alonso, A. Herrmann, F. M. Goni, Biophys. J. 2005, 88, 348–359. 144â•… E. M. Bevers, P. Comfurius, R. F. Zwaal, Eur. J. Biochem. 1982, 122, 81–85. 145â•… A. D. Tepper, P. Ruurs, T. Wiedmer, P. J. Sims, J. Borst, W. J. van Blitterswijk, J. Cell Biol. 2000, 150, 155–164. 146â•… D. R. Green, J. C. Reed, Science 1998, 281, 1309–1312. 147â•… F. Lang, G. L. Busch, M. Ritter, H. Volkl, S. Waldegger, E. Gulbins, D. Haussinger, Physiol. Rev. 1998, 78, 247–306. 148â•… J. Schneider, J. P. Nicolay, M. Foller, T. Wieder, F. Lang, Cell. Physiol. Biochem. 2007, 20, 35–44. 149â•… Y. Huang, Q. Zhao, C. X. Zhou, Z. M. Gu, D. Li, H. Z. Xu, T. Wiedmer, P. J. Sims, K. W. Zhao, G. Q. Chen, Oncogene 2006, 25, 6618–6627. 150â•… E. Castoldi, P. W. Collins, P. L. Williamson, E. M. Bevers, Blood 2011, 117, 4399–4400.

PART IV ENERGY-DEPENDENT LIPID TRANSPORT ACROSS MEMBRANES

8 FLIP OR FLOP: MECHANISM AND (PATHO)PHYSIOLOGY OF P4-ATPASE-CATALYZED LIPID TRANSPORT Patricia M. Verhulst and Joost C.M. Holthuis Department of Membrane Enzymology, Bijvoet Center and Institute sof Biomembranes, Utrecht University, Utrecht, The Netherlands

Thomas G. Pomorski Department of Plant Biology and Biotechnology, Faculty of Life Sciences, University of Copenhagen, Frederiksberg, Denmark

8.1â•… INTRODUCTION Metabolic processes in eukaryotic cells are compartmentalized into distinct membrane-bound organelles that display striking differences in the composition, texture, and sidedness of their membrane lipids. For instance, sphingolipids and sterols form gradients along the secretory pathway with the highest levels in the plasma membrane and the lowest in the endoplasmic reticulum (ER; Fig. 8.1) [1]. This arrangement has important functional consequences. As sphingolipids and sterols are packed at a higher density than glycerolipids, their accumulation at the plasma membrane is vital for the barrier function of this organelle. On the other hand, their concentration should be kept low in the ER because the continuous insertion of newly synthesized proteins and

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

149

150

P-TYPE LIPID PUMPS AND DISEASE 

sterol/SL gradient



SLs 

rigid and asymmetric

AP

  

TJ

APLs BL Go sterol

ER

APLs   

PC

fluid and symmetric

Figure 8.1.â•‡ Nonrandom lipid distributions in polarized epithelial cells. The plasma membrane (PM) is rich in sterols, sphingolipids, and saturated glycerolipids that promote bilayer thickness and impermeability (rat liver PM composition: 30–40â•¯molâ•›% cholesterol; 10–15â•¯molâ•›% sphingomyelin and glycolipids; 25â•¯molâ•›% phosphatidylcholine; 15â•¯molâ•›% phosphatidylethanolamine; 5â•¯molâ•›% phosphatidylserine; 5â•¯molâ•›% phosphatidylinositol [80]). Apical membranes (APs) in polarized epithelial cells display a two- to fourfold enrichment in sphingolipids (glycosphingolipids or sphingomyelin) relative to basolateral membranes (BLs). These differences reside in the outer leaflet of the PM where they are maintained by the presence of tight junctions (TJs). Moreover, the PM displays an asymmetric lipid arrangement with the sphingolipids (SLs) concentrated in the exoplasmic leaflet and the aminophospholipids (APLs) in the cytoplasmic leaflet; the membrane topology of sterols is not known. The endoplasmic reticulum (ER), on the other hand, shows a symmetric lipid distribution and primarily contains saturated glycerolipids that make the membrane flexible, and therefore facilitate the incorporation of newly synthesized proteins (rat liver ER composition: 5â•¯molâ•›% cholesterol; 50–60â•¯molâ•›% phosphatidylcholine; 25â•¯molâ•›% phosphatidylethanolamine; 10â•¯molâ•›% phosphatidylinositol). Vesicular traffic between the ER and the PM passes through the Golgi (Go), a multicisternal organelle, in which lipid sorting must occur to prevent the randomization of the subcellular lipid distribution.

lipids in this biogenic organelle is best served by a loosely packed lipid bilayer. Superimposed on the sphingolipid and sterol gradients are the asymmetric lipid distributions across the bilayers of the late Golgi, endosomes, and plasma membrane, with the aminophospholipids (APLs) phosphatidylserine (PS) and phosphatidylethanolamine (PE) concentrated in the cytosolic leaflet and the sphingolipids (i.e., sphingomyelin and glycosphingolipids) enriched in the exoplasmic leaflet [2]. This lipid asymmetry serves a multitude of biological functions. A tight packing of sphingolipids and sterols in the exoplasmic leaflet is

Introduction

151

important for membrane stability in circulating blood cells and makes the apical surface of intestinal epithelial cells resistant to the detergent action of bile salts. Conversely, the accumulation of APLs (notably PE) in the cytosolic leaflet of the plasma membrane and on the surface of endocytic and exocytic vesicles may serve to keep these membranes in a fusion-competent state [3]. Following regulated dissipation of lipid asymmetry, PS exposed on the cell surface can act as a susceptibility signal for phagocytosis and as a propagation signal in blood coagulation [4]. Furthermore, by creating imbalances in lipid mass across the bilayer, the dynamic process of lipid translocation has been implicated in membrane bending and in the biogenesis of endocytic and exocytic vesicles [5]. Given that lipid asymmetry influences a host of cellular functions, understanding the mechanisms responsible for its creation has become a major focus of interest. Lipid asymmetry is a consequence of multiple factors. These include the biophysical properties that dictate the ability of a lipid to cross the bilayer spontaneously (e.g., charge and bulkiness of the lipid head group), retentive mechanisms that trap lipids on one side of the bilayer (e.g., lipid packing density), and the presence of protein catalysts, termed flippases, which help to create lipid asymmetry by imposing selectivity and directionality on lipid movement. The term flippase was first coined to refer to the lipid transporters responsible for equilibrating newly synthesized lipids across biogenic membranes such as the ER (reviewed in Reference 6 [see Chapter 6]). ER flippases function independently of metabolic energy and catalyze a fast transverse movement of most phospholipid classes in both directions. As facilitators of nonvectorial lipid transport, they promote a symmetrical phopholipid distribution across the bilayer. The identity of ER flippases is not known. However, the observation that peptides mimicking the α-helices of transmembrane proteins stimulate flip-flop in model membranes suggests that the ability to catalyze flip-flop may not necessarily be restricted to one specific protein [7]. In contrast to the ER, spontaneous flip-flop in the plasma membrane is slow. Hence, the process of catalyzing flip-flop is lost as membrane flows through the Golgi to the plasma membrane, which must be due to a change in lipid and/or protein composition (Fig. 8.2). Slow flip-flop is a prerequisite for preserving asymmetric lipid arrangements that are created by energy-dependent flippases or translocases. These activities use ATP hydrolysis to move specific lipid species against a concentration gradient. The first ATP-fueled flippase described was the red blood cell-associated APL translocase (APLT), postulated by Seigneuret and Devaux in 1984 [8], which catalyzes a fast inward movement of PS and PE across the plasma membrane. APLT activities occur in the plasma membrane of many nucleated cells as well as in the trans-Golgi [9] and in Golgi-derived secretory vesicles [10, 11]. Some cell types show a similar inward transport of phosphatidylcholine (PC) due to the presence of either a PC translocase next to an APLT, or a translocase of different specificity that flips both APLs and PC [12, 13].

152

P-TYPE LIPID PUMPS AND DISEASE

flip-flop APL

TMP

sterols

t1/2 = hours to days

t1/2 = seconds to minutes

synthetic bilayer

ER and cis-Golgi

APLT

t1/2 = hours to days

ATP

t1/2 = hours to days

trans-Golgi and PM

Figure 8.2.â•‡ Transbilayer lipid movement is controlled by membrane proteins and lipid composition. In synthetic lipid bilayers, the spontaneous transbilayer movement of phospholipids or flip-flop is low with a half-time (t1/2) of hours to days. In contrast, flipflop of lipids in the ER and cis-Golgi membranes is much higher (t1/2, sec-min) due to the presence of a bidirectional flippase or the presence of single-membrane-spanning proteins, which cause transient defects in the helix–lipid interface. The gradual increase of sphingolipids and sterols toward the trans-Golgi and plasma membrane (PM) may slow down flip-flop due to an increased packing of the fatty acyl chains in the hydrophobic interior through which a polar head group has to travel (t1/2, hrs-days). Inhibition of spontaneous flip-flop is required to preserve the action of APL translocases (APLTs), which hydrolyze ATP to pump APLs against a concentration gradient across membranes of late secretory and endocytic compartments. TMP, transmembrane protein.

8.2â•… P4-ATPASES ARE PRIME CANDIDATE PHOSPHOLIPID TRANSLOCASES The discovery of an APLT activity in bovine chromaffin granules, termed ATPase II, led to the cloning of a gene currently referred to as ATP8A1 [14]. The corresponding enzyme is homologous to Drs2p, a trans-Golgi resident protein in yeast (see Chapter 9). ATP8A1 and Drs2p are the founding members of a conserved subfamily of P-type ATPases (class IV or P4-ATPases) that includes 5 yeast (Drs2p, Dnf1p, Dnf2p, Dnf3p, and Neo1p; Table 8.1) and 14 human members (ATP8A1-ATP11C; Fig. 8.3). A growing body of evidence indicates that P4-ATPases indeed possess APLT activity. Removal of the yeast plasma membrane-resident P4-ATPases Dnf1p and Dnf2p abolishes ATP-dependent transport of fluorescent 7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD)-labeled analogs of PE, PS, and PC from the exoplasmic to the cytosolic leaflet of the plasma membrane [13] (Fig. 8.4). Although their lipid preference disqualifies these proteins as specific APLTs (see Chapter 10), NBD-labeled analogs of sphingolipids, phosphatidic acid, and phosphatidylglycerol were not recognized. The same P4-ATPases have been implicated in the transport of lyso-PE and lyso-PC across the plasma membrane [15, 16]. Chemical labeling of exoplasmic leaflet phospholipids and staining with APL-specific probes showed an accumulation of PS and PE on the surface of a dnf1dnf2 deletion mutant, a phenotype that is exacerbated when DRS2 is

P4-ATPASES ARE PRIME CANDIDATE PHOSPHOLIPID TRANSLOCASES

153

TABLE 8.1.â•‡ P4-ATPase/CDC50 Complexes and Their Putative Substrate Specificities Organism

P4-ATPase

Cdc50 Subunit

Location

Substrates

Saccharomyces cerevisiae

Dnf1p Dnf2p Dnf3p Drs2p ALA2 ALA3 ATP8B1

Lem3p Lem3p Crf1p Cdc50p ALIS1/3/5 ALIS1/3/5 CDC50A/B

PM PM Golgi, SV Golgi, SV PVC Golgi AP

PC, PE (PS) PC, PE (PS) PC, PE PS, PE PS PC, PE, PS PS

Arabidopsis thaliana Homo sapiens

PM, plasma membrane; SV, secretory vesicle; PVC, prevacuolar compartment; AP, apical membrane; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PS, phosphatidylserine. For references, see text.

ATP9B ATP9A

ATP10A/C

ATP10B ATP10D

Neo1p

ATP11A ATP11C

Dnf3p

ATP11B

ATP8A2

ATP8B1

ATP8A1

ATP8B2

Drs2p ATP8B4 Dnf2p

Dnf1p

ATP8B3

Figure 8.3.â•‡ Phylogenetic tree of human and yeast P4-ATPases. Multiple alignment and unrooted phylogenetic tree of P4-ATPase homologs from Homo sapiens and Saccharomyces cerevisiae were generated with ClustalW and Phylip 3.65 using the DrawTree method with EqualDaylight algorithm.

removed as well [13, 17]. In addition, Drs2p and Dnf3p are required to sustain PE transport and asymmetry in post-Golgi secretory vesicles [11] (Fig. 8.5). Lipid transport assays with purified Golgi membranes containing a temperaturesensitive Drs2 mutant indicated that Drs2p is directly coupled to an NBD-PS translocase activity [9]. Indeed, reconstitution of NBD-PS translocase activity with purified Drs2p indicates that this enzyme functions as a flippase [18].

154

P-TYPE LIPID PUMPS AND DISEASE

(a)

(b)

NBD-PE

DAPI

PHACO

C6-NBD-PE E

wild type

P

O O

O=

O

=O

NH N N NO2

O

∆dnf1 ∆dnf2

Figure 8.4.â•‡ Loss of yeast plasma membrane (PM)-associated P4-ATPases Dnf1p and Dnf2p abolishes nonendocytic uptake of NBD-PE. (a) Myristoyl-(NBD-hexanoyl)-PE (C6-NBD-PE). (b) Wild-type and Δdnf1Δdnf2 mutant cells, prestained with the DNA probe DAPI, were incubated with 100â•›µM C6-NBD-PE at 2°C before imaging by fluorescence (NBD, DAPI) or phase contrast (PHACO) microscopy. Note that wildtype cells accumulate C6-NBD-PE primarily in the mitochondria, as indicated by colocalization with DAPI fluorescence (arrows). Bar, 10â•¯µm. Figure was modified from Reference 13. Color version on the Wiley web site.

Further evidence for a direct relationship between P4-ATPases, flippases, and lipid asymmetry has come from the functional characterization of family members in higher eukaryotes. For instance, the removal of the Caenorhabditis elegans P4-ATPase TAT-1 causes an increased cell surface exposure of PS, resulting in an aberrant phagocytic clearance of living cells [19]. A deficiency of ATP8B1, a P4-ATPase located in the canalicular membrane of liver cells and implicated in familial intrahepatic cholestasis type 1 (FIC1) disease, is accompanied by an enhanced recovery of PS in bile [20]. Moreover, heterologous expression of ATP8B1 restores the nonendocytic uptake of NBD-PS in PS transport-defective Chinese hamster ovary (CHO) mutant cells [21, 22]. One genetic argument for P4-ATPase-catalyzed lipid transport is that two different Arabidopsis P4-ATPase, ALA2 and ALA3, complement a yeast drs2dnf1dnf2 mutant, but generate flipping activities with different lipid specificities. While ALA2 specifically promotes flipping of NBD-PS, ALA3 facilitates the internalization of NBD-PS, -PE, and -PC [23, 24]. This would not have been expected if P4-ATPases indirectly contribute to the function of the same flippase. Finally, ATP8A2, a P4-ATPase present in the disk membranes of rod and cone photoreceptors, displays APLT activity when purified and reconstituted in proteoliposomes [25]. The simplest interpretation of these data is that at least some P4-ATPases are able to catalyze phospholipid transport using the free energy of ATP. Yet

P4-ATPASES ARE PRIME CANDIDATE PHOSPHOLIPID TRANSLOCASES 27°C

(a)

155

TNBS labeling

(d)

V 100

cytosolic

80 60

N

SV

V

% PE

(b)

40 38°C

20 0 20 40 60 80

(c)

100

exoplasmic sec6-4 sec6-4∆drs2∆dnf3

Figure 8.5.â•‡ Loss of yeast Golgi-associated P4-ATPases Drs2p and Dnf3p abolishes PE asymmetry in post-Golgi secretory vesicles (SVs). (a,b) Electron micrographs showing that yeast sec6-4 mutant strains accumulate post-Golgi SVs when shifted to the nonpermissive temperature (38°C). Bar, 2â•¯µm. (c) Electron micrograph of post-Golgi SVs isolated from 38°C-shifted sec6-4 mutant cells by differential centrifugation and visualized by negative staining with uranyl acetate. Bar, 200â•¯nm. (d) SVs isolated from 38°C-shifted sec6-4 and sec6-4Δdrs2Δdnf3 cells were incubated with trinitrobenzene sulfonic acid (TNBS) at 25°C in the presence of ATP. After 30 minutes, the reaction was stopped by addition of glycylglycine, and the membranes were subjected to lipid analysis. Percentages of TNBS-reacted (cytosolic) and unreacted (exoplasmic) PE are shown as the meansâ•¯±â•¯standard deviation (SD) of three independent experiments. Figure was modified from Reference [11]. V, vacuole; N, nucleus.

several complications obscure this simple interpretation. As discussed below, P4-ATPase mutants display pleiotropic phenotypes that often include perturbations in protein trafficking from and to the plasma membrane. Since clathrin and endocytosis mutants exhibit a loss of lipid asymmetry comparable to that of P4-ATPase mutants, lipid randomization may not necessarily be a specific consequence of deleting P4-ATPases [17]. Moreover, P-type ATPases usually pump small cations or metal ions by means of a transport mechanism that appears to be conserved throughout the family. How P4-ATPases adapted this mechanism to flip phospholipids is not understood. Furthermore, P4ATPases form heteromeric complexes with members of the Cdc50 family of

156

P-TYPE LIPID PUMPS AND DISEASE

membrane proteins. Because P4-ATPases copurify with Cdc50 proteins [18, 26], it is not clear whether P4-ATPases alone are sufficient to mediate phospholipid transport or whether they represent only one of the subunits of a flippase complex. 8.3â•… MECHANISM OF P4-ATPASE-CATALYZED LIPID TRANSPORT: ROLE OF ACCESSORY SUBUNITS 8.3.1â•… The P-Type ATPase Transport Cycle Contrary to P4-ATPases, essentially all other P-type ATPases pump small cations or metal ions that are occluded in the interior of the membranespanning domain during transport. These include soft-transitional metaltransporting ATPases (P1), Ca2+-ATPases (P2A/B), Na+/K+-ATPases and H+/ K+-ATPases (P2C), and H+-ATPases (P3) [27]. The designation “P-type” comes from the acid-stable, phosphorylated aspartate (Asp) residue that forms during the pump’s transport cycle (Fig. 8.6). At least two main conformations exist, E1 and E2, with conformational changes being accompanied by translocation of the ligand across a membrane [28]. The ligand binding sites are buried deep inside the M-domain, the region of the pump that crosses the membrane (Fig. 8.7). The transport cycle is strictly controlled so that access to the ligand binding sites alternates between the two sides of the membrane, with the ligands becoming temporarily occluded after each ligand-binding event (i.e., they become inaccessible from either side of the membrane). This mechanism allows P-type ATPases to pump ions against a gradient, while avoiding ion leaks in the opposite direction. In E1, the sites are accessible for ligands from the cytosol, that is, two Ca2+ ions in the case of a P2A-calcium pump (Figs. 8.6 and 8.7). Binding of the

E1•Ca2+

Ca

cyto

E2P

E2 Pi

H2O

ATP ADP E1 PLcyto

P2AATPase

2+

Ca

ATP ADP E1P•Ca2+

2+

E1P P4ATPase

PLexo

exo

E2P•PL

E2•PL Pi

H 2O

Figure 8.6.â•‡ Transport cycle of P2A- and P4-ATPases. Simplified scheme according to the E1/E2 model [32]. Note that P2A-ATPases translocate two Ca2+ ions per hydrolyzed ATP molecule, and pump two to three H+ ions, which bind to the E2P intermediate, in the opposite direction (not shown). Whether P4-ATPase-catalyzed phospholipid (PL) transport is coupled to countertransport of another ligand is not known. Exo, taken up or released from the exoplasmic leaflet; cyto, taken up or released from the cytosolic surface.

157

MECHANISM OF P4-ATPASE-CATALYZED LIPID TRANSPORT

(a)

(b) N A P A N H2N P

P

COOH

M

M

(c)

P4-ATPase

P2A-ATPase

N P

A

P

N

P

P

A M

M

N

N A

A M

M

+ + + 2+

E1•Ca

cyto

+

E2P + Ca2+exo

E2P•PLexo

E1 + PLcyto

Figure 8.7.â•‡ Transport cycle-dependent conformational changes in P-type pumps. (a) Structure of the P2A-type calcium pump SERCA in the E1 confirmation, modified from Reference [29]. Indicated are the phosphorylation (P), nucleotide-binding (N), actuator (A), and membrane spanning (M) domains. (b) Membrane topology of P-type ATPases. Transmembrane helices are shown as cylinders and functional domains are marked as in (a). The position of the catalytically important, phosphorylated Asp residue in the large cytoplasmic loop connecting TM4 and TM5 is indicated. (c) Cartoons illustrating transport cycle-dependent movements of the three cytoplasmic domains (P, N, A) and changes in the intramembrane region (M) are shown for the P2A-ATPase (calcium pump SERCA; left) and P4-ATPase (right). E1 is accessible for ligands from the cytosol (two Ca2+ ions in the case of the P2A-ATPase, and unknown for the P4-ATPase) that, upon binding, promote phosphorylation of the P-domain. In E2P, ligands are discharged to the exoplasmic side, and countertransported ligands now can bind (protons or phospholipid, respectively). Then, the enzyme reverts to the E1 state, in which the countertransported ligands are released toward the cytosol. See text for details.

158

P-TYPE LIPID PUMPS AND DISEASE

ligand allows the crucial Asp residue in the phosphorylation (P) domain to become phosphorylated by Mg2+-ATP, which is bound to the nucleotidebinding (N) domain. Formation of the E1P intermediate results in occlusion of the ligands. The pump then releases ADP and relaxes to a lower energy E2P conformation, whereupon a pathway opens to discharge the ligands to the exoplasmic side. The ligand-binding site now has high affinity for the countertransported ligands, that is, two to three H+ ions in the case of a P2A-calcium pump, which bind from the exoplasmic side. Hydrolysis of the phosphorylated Asp residue, catalyzed by the actuator (A) domain, results in another state with occluded ligands, E2. Mg2+ and inorganic phosphate (Pi) dissociate, and the enzyme reverts to the E1 state, in which the countertransported ligands are released into the cytosol.

8.3.2â•… P4-ATPases and the Giant Substrate Problem High-resolution X-ray structures of sarcoplasmic reticulum Ca2+-ATPase [27, 28, 29, 30], Na+/K+-ATPase [31, 32], and H+-ATPase [33] indicate that these pumps have the same architecture. Differences are largely confined to the cation-binding pocket, which consists of a congregation of glutamates and Asps whose anionic carboxyl groups serve to neutralize the charge of the cations. Even though no 3D structural information is available for P4-ATPases, they share a common membrane domain organization and display a clear overall sequence homology with the cation-transporting P-type ATPases [28]. Conserved sequence motifs include the canonical phosphorylation site in the P-domain, the nucleotide-binding site in the N-domain, and a TGES-like sequence in the A-domain [34]. This implies that P4-ATPases and cation pumps essentially share the same mechanism of ATP-fueled ligand transport. P4ATPases are believed to catalyze the translocation of phospholipids from the exoplasmic to the cytosolic leaflet. This would predict that the phospholipid ligand in P4-ATPases binds to the phosphoenzyme intermediate E2P as opposed to the Ca2+ ion in P2A-calcium pumps, which binds to E1 (Fig. 8.6). This model is consistent with the finding that, while Ca2+ triggers a rapid phosphorylation of the P2A-calcium pump, PS induces a fast dephosphorylation of the P4-ATPase ATP8A1 [35]. How did P4-ATPases acquire the ability to translocate phospholipids instead of small cations? Flippases must provide a sizeable hydrophilic pathway for the polar head group to pass through the membrane as well as accommodate the hydrophobic nature of the lipid backbone. In addition, they must be able to discriminate phospholipids based on atoms in the lipid head group and backbone. It is difficult to envision how a binding pocket for sodium, potassium, and calcium ions in the center of a 10-helix bundle can adapt to handle a phospholipid ligand. Consequently, it is unclear how P4-ATPases would execute their presumed flippase activity.

159

MECHANISM OF P4-ATPASE-CATALYZED LIPID TRANSPORT

8.3.3â•… Role of Cdc50 Proteins in P4-ATPase-Catalyzed Phospholipid Transport In view of their overall similarity to cation-transporting P-type ATPases, it has been postulated that P4-ATPases pump cations to generate a concentration gradient that drives phospholipid translocation through a second protein by a symport mechanism [28]. One prediction of this model is that mutation of the putative symporter should also block phospholipid transport and phenocopy mutations in P4-ATPase genes. Mutations in members of the Cdc50 protein family produce indeed such phenotypes. This family includes three yeast proteins (Cdc50p, Crf1p, Lem3p) [36] and three human proteins (CDC50A, CDC50B, CDC50C) [37] that consist of two membrane spans and an N-glycosylated exoplasmic loop stabilized by one or more disulfide bonds (Fig. 8.8). Instead of functioning as transporters themselves, Cdc50 proteins form heteromeric complexes with P4-ATPases. Assembly of these complexes is required for their export from the ER [22, 36, 38]. This suggests that Cdc50 proteins function as chaperones involved in the correct intracellular targeting of P4-ATPases. (b)

(a)

P-type ATPase

subunits

S–S

a P cytosol

g

g

5

9 10

7 b

8

b

g

cytosol

s

6

P2C/a

3 S–S

4

P

b

P4

Cdc50

cytosol

Figure 8.8.â•‡ Composition of the oligomeric P2C- and P4-ATPases. (a) Architecture of the P2C (Na+/K+) ATPase αβγ complex, modified from Reference [31]. (b) Schematic representation of P2C- and P4-ATPases, and their accessory subunits. Transmembrane helices are shown as gray cylinders. The phosphorylation site is indicated in the large, central cytoplasmic loop, connecting TM4 and TM5. Note that the exoplasmic loop of Cdc50 proteins is N-glycosylated (Y) and contains one or more disulfide bridges (S–S), hence analogous to the ectodomain of the P2C-ATPase β-subunit. Color version on the Wiley web site.

160

P-TYPE LIPID PUMPS AND DISEASE

Besides serving as a molecular chaperone, Cdc50 proteins may actively contribute to P4-ATPase-catalyzed phospholipid transport. They could do this in several ways. For instance, creation of a high-affinity phospholipid binding site may require Cdc50-induced conformational changes in the M-domain of P4-ATPases, analogous to the role of the β-subunit in the oligomeric Na+,K+ATPase [39] (Fig. 8.8). Alternatively, by donating two additional helices to the M-domain of the transporter, Cdc50 proteins may help complete a pathway for lipid translocation. Thus, flipping might occur at the interface between the P4-ATPase and its Cdc50-binding partner. Such an arrangement would allow Cdc50 proteins to contribute directly to the transport specificity of the complex. This idea is consistent with the observation that the yeast trans-Golgi P4ATPases Drs2p and Dnf3p, which exhibit different translocation profiles, interact with different Cdc50 homologs (Cdc50p and Crf1p, respectively), whereas the plasma membrane P4-ATPases Dnf1p and Dnf2p, which have the same substrate specificity, both interact with Lem3p (Table 8.1). However, the plant P4-ATPases ALA2 and ALA3 gain functionality when co-expressed with any of three different ALIS Cdc50-like subunits but retain their different lipid substrate specificities independently of the nature of the Cdc50-binding partner [24] (Table 8.1). This indicates that substrate determinants primarily reside in the P4-ATPase and not in the Cdc50 subunit. An intimate role for Cdc50 proteins in P4-ATPase-catalyzed lipid transport can be inferred from the finding that separation of Drs2p from its binding partner Cdc50p disrupts its ability to form a phosphoenzyme intermediate [26]. Moreover, catalytic activity was found to rely on direct and specific interactions between subunit and ATPase. In addition, studies using a genetic reporter system revealed that P4-ATPase–Cdc50 interactions are dynamic and tightly coupled to the ATPase reaction cycle, with the strongest interaction occurring at a point where the enzyme is loaded with phospholipid substrate [26]. Together, these findings indicate that Cdc50 proteins directly participate in the P4-ATPase reaction cycle. Their function bears striking similarity to the role of the β-subunit in Na+/K+-pumping P2C-ATPases, which is believed to stabilize the enzyme complex during loading of K+ ions [39, 40]. So far, a higher number of P4-ATPase family members compared with Cdc50 homologs have been identified in each organism (Table 8.2). This imbal-

TABLE 8.2.â•‡ P4-ATPase and Cdc50 Gene Numbers in Various Organisms Organism Saccharomyces cerevisiae (yeast) Caenorhabditis elegans (nematode) Drosophila melanogaster (fruit fly) Oryza sativa (rice) Arabidopsis thaliana Homo sapiens (human)

P4-ATPase

CDC50

5 6 6 10 12 14

3 2 1 6 5 3

ROLE OF P4-ATPASES IN VESICLE-MEDIATED PROTEIN TRANSPORT

161

anced ratio is particularly striking in humans and in the fruit fly, Drosophila melanogaster, with the latter counting six P4-ATPases and only one Cdc50 protein. This implies that each Cdc50 protein interacts with multiple P4ATPases or, alternatively, that some P4-ATPases function without a Cdc50binding partner. In yeast, Lem3p interacts and sustains functionality of Dnf1p and Dnf2p [34, 36], while in humans, CDC50A and CDC50B bind and facilitate ER export of the same P4-ATPase, ATP8B1 [22]. Likewise, in Arabidopsis, three different ALIS Cdc50-like proteins can interact and facilitate ER export of the P4-ATPases ALA2 and ALA3 [24]. This shows that one Cdc50 protein can interact with multiple P4-ATPases and vice versa. Remarkably, no Cdc50binding partner has thus far been identified for the yeast P4-ATPase Neo1p. Among the five P4-ATPases in yeast, Neo1p is unique in that deletion of its gene is lethal [41]. This suggests that Neo1p possesses a biochemical activity that is different from the other P4-ATPases and for which it may not require a Cdc50-binding partner. Experimental evidence that Neo1p catalyzes phospholipid transport is lacking. Consequently, how Neo1p executes its essential function remains to be established. 8.4â•… ROLE OF P4-ATPASES IN VESICLE-MEDIATED PROTEIN TRANSPORT In addition to their presumed function as flippases, P4-ATPases play a critical role in the biogenesis of transport vesicles during endocytosis and exocytosis [42]. For example, yeast dnf1dnf2 mutant cells show a cold-sensitive defect in the uptake of markers for bulk-phase and receptor-mediated endocytosis [13] (Fig. 8.9). Inactivation of temperature-sensitive Drs2 in vivo rapidly blocks the formation of a clathrin-dependent class of post-Golgi secretory vesicles carrying exocytic cargo [43]. Moreover, conditional alleles of the essential yeast P4-ATPase Neo1p perturbs ADP-ribosylation factor (ARF)-dependent vesicle formation from endosomes [44]. Loss of ALA3, a Golgi-resident P4-ATPase in Arabidopsis thaliana, causes a defect in the production of slime vesicles containing polysaccharides and enzymes for secretion, and impairs the growth of roots and shoots [23]. Finally, the C. elegans P4-ATPase TAT-1 is required for yolk uptake in oocytes and for an early step of fluid-phase endocytosis in the intestine [45]. How P4-ATPases contribute to vesicle formation remains to be established. At least two possibilities can be envisioned. First, by accumulating specific phospholipids in the cytosolic leaflet or providing a scaffold for protein– protein interactions, P4-ATPases may help recruit coat proteins and GTPases that are directly responsible for vesicle formation. This possibility is supported by the observation that Drs2p directly interacts with the ARF activator, Gea2p [46]. The latter protein is a GTP-exchange factor that regulates the recruitment of ARF, adapter protein-1 (AP-1), and clathrin coat proteins to membranes of the trans-Golgi. However, a recent study shows that membrane recruitment

162

P-TYPE LIPID PUMPS AND DISEASE 0 hour

1 hour

4 hours

wild type

∆dnf1 ∆dnf2

∆dnf1 ∆dnf2 ∆drs2

Figure 8.9.â•‡ Yeast P4-ATPase mutant strains are defective in bulk-phase endocytosis. Cells were stained with the bulk-phase endocytosis tracer FM4-64 on ice, and endocytic uptake of the dye was initiated by shifting cells to 15°C. Images were captured at 0, 1, and 4 hours after shift to 15°C. Bar, 10â•¯µm. Figure modified from Reference [13].

of these coat proteins is insufficient to form the clathrin-dependent class of post-Golgi secretory vesicles when Drs2p is absent [47]. Second, by moving lipid mass from the exoplasmic to the cytosolic leaflet, P4-ATPases may create a tension that drives the invagination of the membrane required for vesicle formation [5] (Fig. 8.9) (see Chapter 12). Consistent with this hypothesis, insertion of exogenous APLs in the exoplasmic leaflet of the plasma membrane and their subsequent translocation to the cytosolic leaflet by the APLT provokes the formation of endocytic-like vesicles in red blood cells [48, 49] and accelerates endocytosis in erythroleukemia K562 cells [50]. Direct participation of ATP-driven lipid transport in vesicle budding is further supported by the observation that giant proteoliposomes formed from erythrocytes membrane fragments undergo budding in the presence of ATP [51]. 8.5â•… P4-ATPASE DYSFUNCTION AND DISEASE The P4-ATPase family comprises 14 human members (Fig. 8.3), several of which have been implicated in disease. For instance, ATP8A2 is frequently

P4-ATPASE DYSFUNCTION AND DISEASE

163

deleted in tumorigenic malignancies [52]. The mouse ortholog of ATP10A (Atp10a or pfatp) has been linked to obesity, type 2 diabetes, and nonalcoholic fatty acid liver disease [53]. Similarly, mice carrying a null mutation in Atp10d, which belongs to the same subclass as Atp10a, develop obesity, hyperglycemia, and hypertension when fed a high-fat diet [54]. ATP11C has been mapped to a chromosomal region, Xq27, associated with X-linked disorders that include hypoparathyroidism, albinism–deafness, and thoracoabdominal syndrome [55]. The mouse ortholog of ATP8B3 (Atp8b3) is required for sperm capacitation and may therefore be associated with fertility disorders [56]. At present, only one human disease has been directly linked to a P4-ATPaseencoding gene. Mutations in ATP8B1 cause Byler disease or FIC1 disease [57]. The characteristics of this disease and how ATP8B1 dysfunction may cause its manifestation will be discussed below. 8.5.1â•… ATP8B1 and FIC1 Disease FIC1 disease is a rare autosomal recessive disorder that primarily manifests as a intrahepatic cholestasis (defective bile salt excretion from liver into bile), which can progress to severe, end-stage liver disease before adolescence [58]. Three clinical entities have been described: progressive familial intrahepatic cholestasis type 1 (PFIC1), benign recurrent intrahepatic cholestasis type 1 (BRIC1), and Greenland familial cholestasis (GFC) [59]. In addition, mutations in ATP8B1 can lead to intrahepatic cholestasis of pregnancy (ICP) [60, 61]. It is now understood that these disorders represent a clinical continuum of cholestatic disease with common genetic etiology, with PFIC1 being the most severe manifestation of FIC1 disease [62]. In the liver, ATP8B1 localizes to the canalicular membrane of hepatocytes [63], which also contains the bile salt export pump BSEP (ABCB11). Absence or severe dysfunction of BSEP causes PFIC2, a hereditary cholestasis syndrome showing clinical and biochemical features that are very similar but not identical to FIC1 disease [64–66]. Although ATP8B1 deficiency has a major impact on bile salt transport, it is unclear by which mechanism impaired ATP8B1 function results in cholestasis. Remarkably, interruption of the enterohepatic circulation of bile salts in PFIC1 patients resulted in a normalization of their hepatobiliary output [67]. This indicates that the bile salt transport defect in FIC1 disease is not a direct consequence of ATP8B1 dysfunction. 8.5.2â•… A Potential Mechanism of FIC1 Disease The enhanced recovery of PS in bile from Atp8b1G308V/G308V mutant mice after liver infusion of taurocholate supports a role of ATP8B1 as PS translocase in the canalicular membrane of hepatocytes [20]. Importantly, on infusion with taurocholate, these mice also had an increased biliary output of canalicular cholesterol. This phenotype was independent of the putative cholesterol transporter Abcg5/Abcg8 [68]. Based on these findings, Oude Elferink and

164

P-TYPE LIPID PUMPS AND DISEASE

(a)

(b)

(c)

PS ATP8B1

MDR3

BSEP

atp8b1

MDR3

BSEP

SM/PC cholesterol bile salt

ATP

ATP cytosol

ATP

ATP

ATP

micelle

cytosol

Figure 8.10.â•‡ Potential mechanism of FIC1 disease. (a) Under normal conditions, the canalicular membrane of hepatocytes is kept asymmetric by the continuous removal of phosphatidylserine (PS) from the exoplasmic surface by the APL translocase, ATP8B1. This keeps the sphingolipids and sterols in the exoplasmic leaflet in a tightly packed and highly ordered state to protect the membrane against the detergent action of hydrophobic bile salts, which are exported by the bile salt pump BSEP (ABCB11). Note that MDR3 is an outward-directed transporter of phosphatidylcholine (PC), another important constituent of bile. (b) ATP8B1 dysfunction causes lipid randomization, which reduces the lipid packing density in the exoplasmic leaflet and renders the canalicular membrane more sensitive toward extraction of lipids, including cholesterol and PS, by hydrophobic bile salts. Subsequent reduction in cholesterol levels impairs the activity of the canalicular bile salt transporter BSEP, hence causing cholestasis. (c) Keys to panel (a) and (b). SM, sphingomyelin.

coworkers proposed a model for the link between ATP8B1 and FIC1 disease (Fig. 8.10). They postulate that under normal conditions, the canalicular membrane is highly asymmetric, with the exoplasmic leaflet containing high concentrations of tightly packed sphingolipids (notably sphingomyelin) and sterols that are in a “liquid ordered” state to protect the membrane against the detergent action of hydrophobic bile salts. ATP8B1 activity is crucial in maintaining this detergent resistance by flipping excess of the more loosely packed APLs from the exoplasmic to the cytosolic leaflet. Loss of ATP8B1 function causes lipid randomization, which reduces the lipid ordering in the exoplasmic leaflet of the canalicular membrane and consequently enhances its sensitivity toward hydrophobic bile salts. Increased cholesterol extraction by bile salts reduces the cholesterol content of the bilayer, which in turn may impair the activity of the bile salt export pump BSEP and cause cholestasis [20, 69]. 8.5.3â•… Extrahepatic Symptoms of FIC1 Disease Apart from liver, ATP8B1 is expressed in numerous other tissues, for example, intestine, pancreas, bladder, stomach, and prostate. Remarkably, ATP8B1 is

P4-ATPASE DYSFUNCTION AND DISEASE

165

more abundantly expressed in the epithelial cells of the gastrointestinal tract than in the liver [70]. Indeed, many patients with FIC1 disease suffer from extrahepatic symptoms that include pancreatitis, diarrhea, and hearing loss [71–73]. These symptoms are not corrected and may even be aggravated by liver transplantation [74]. Collectively, this indicates that ATP8B1 function is not exclusively linked to bile salt excretion and that the pathophysiology of FIC1 patients is the result of a more general cellular defect. For example, ATP8B1 may play a role in maintaining APL asymmetry in the apical membrane of different types of epithelial cells, with loss of its activity causing pleiotropic dysfunction of this specialized membrane domain. In addition, analogous to the situation in P4-ATPase mutants in yeast, Arabidopsis and C. elegans, inactivation of ATP8B1 may perturb vesicular trafficking from or to the plasma membrane (see above). To further explore the origin of extrahepatic symptoms associated with FIC1 disease, the impact of blocking ATP8B1 expression on the domainspecific APLT activities and functional organization of polarized human intestinal epithelial Caco-2 cells has been investigated. Although ATP8B1 is abundantly expressed in the apical membrane of these cells, blocking its expression by RNAi affected neither the apical APL transport nor the asymmetric distribution of APLs across the apical bilayer. Nonetheless, ATP8B1depleted Caco-2 cells displayed profound perturbations in apical membrane architecture and composition, including disorganization of the apical actin cytoskeleton, a dramatic loss in microvilli, and a substantial reduction in the expression of apical proteins [75]. These findings point to an essential role of ATP8B1 in apical membrane organization that is unrelated to its presumed flippase activity, yet potentially relevant for the manifestation of FIC1 disease. As microvilli of intestinal epithelial cells are key determinants of the surface area available for adsorption, their disruption in ATP8B1-deficient epithelial cells would provide a rationale for diarrhea being a common symptom associated with FIC1 disease [74]. Similarly, reduction of the apical surface of hepatocytes would likely limit bile salt excretory capacity of the liver in patients with ATP8B1 deficiency. The morphological aberrations in ATP8B1-depleted intestinal epithelial cells show a striking parallel with the loss of stereocilia from the apical surface of the inner hair cells in Atp8B1 mutant mice [76], suggesting that ATP8B1 serves a general role in the formation and/or stabilization of microvillar structures. ATP8B1 is not the only example of a P-type ATPase for which a transportunrelated function has been proposed. Epithelial junction formation and tracheal tube-size control in Drosophila requires a pump-independent function of the Na+/K+-ATPase [77]. The catalytic α-subunit of this pump interacts with ankyrin, a cytoskeletal protein potentially involved in epithelial junction formation [78]. Microvilli formation requires the major actin-binding protein villin. Blocking villin expression in Caco-2 cells results in a shortening and partial loss of microvilli [79], analogous to the situation in ATP8B1-depleted Caco-2 cells. It is tempting to speculate that, in addition to its APL transport

166

P-TYPE LIPID PUMPS AND DISEASE

function, ATP8B1 serves a dual role in microvilli formation/stabilization by acting as a molecular scaffold for the recruitment of villin or other actinmodifying proteins to the apical membrane of polarized epithelial cells. Further characterization of the flippase-independent function of ATP8B1 is warranted to fully uncover the pathophysiology of intrahepatic cholestasis and the extrahepatic symptoms associated with FIC1 disease.

8.6â•… FUTURE CHALLENGES It is evident that phospholipid transport and asymmetry in eukaryotic cells are critically dependent on P4-ATPases and their Cdc50-binding partners. However, whether P4-ATPases alone are sufficient to catalyze phospholipid transport or whether this reaction relies on an active participation of Cdc50 proteins remains to be established. P4-ATPases belong to the P-type ATPase superfamily, whose members usually translocate small cations. It is feasible that Cdc50 proteins help form a pathway for phospholipid translocation and that their acquisition was crucial for the evolution of flippases from a family of cation pumps. Hence, defining the minimal composition of the P4-ATPase-dependent flippase machinery provides one of the most pressing issues to be tackled in the future. Besides forming heteromeric complexes with Cdc50 proteins, P4ATPases physically interact with small GTPases that control membrane trafficking and a whole range of other cellular processes. Consequently, it is not always obvious to what extent phenotypes associated with P4-ATPase dysfunction arise from perturbations in transbilayer lipid organization. Resolving the molecular basis of diseases caused by mutations in P4-ATPase-encoding genes therefore provides another major future challenge.

ACKNOWLEDGMENTS This work was supported by grants from the Dutch Organization of Sciences (NWOCW), the Utrecht University High Potential Program (to J.H.), and the Deutsche Forschungsgemeinschaft (to T.P., PO748/9).

ABBREVIATIONS APL APLT ARF Asp BRIC1 BSEP

aminophospholipid aminophospholipid translocase ADP-ribosylation factor aspartate benign recurrent intrahepatic cholestasis type 1 bile salt export pump

References

167

CHO Chinese hamster ovary ER endoplasmic reticulum FIC1 familial intrahepatic cholestasis type 1 GFC Greenland familial cholestasis ICP intrahepatic cholestasis of pregnancy NBD 7-nitrobenz-2-oxa-1,3-diazol-4-yl PC phosphatidylcholine PE phosphatidylethanolamine PFIC1 progressive familial intrahepatic cholestasis type 1 PS phosphatidylserine

REFERENCES â•‡ 1â•… J. C. Holthuis, T. Pomorski, R. J. Raggers, H. Sprong, G. Van Meer, Physiol. Rev. 2001, 81, 1689–1723. â•‡ 2â•… â•‡ 3â•… â•‡ 4â•… â•‡ 5â•…

J. A. Op den Kamp, Annu. Rev. Biochem. 1979, 48, 47–71. P. K. Kinnunen, J. M. Holopainen, Biosci. Rep. 2000, 20, 465–482. K. Balasubramanian, A. J. Schroit, Annu. Rev. Physiol. 2003, 65, 701–734. P. F. Devaux, A. Herrmann, N. Ohlwein, M. M. Kozlov, Biochim. Biophys. Acta 2008, 1778, 1591–1600.

â•‡ 6â•… T. Pomorski, A. K. Menon, Cell. Mol. Life Sci. 2006, 63, 2908–2921. â•‡ 7â•… M. A. Kol, A. N. van Laak, D. T. Rijkers, J. A. Killian, A. I. de Kroon, B. de Kruijff, Biochemistry 2003, 42, 231–237. â•‡ 8â•… M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. â•‡ 9â•… P. Natarajan, J. Wang, Z. Hua, T. R. Graham, Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 10614–10619. 10â•… A. Zachowski, J. P. Henry, P. F. Devaux, Nature 1989, 340, 75–76. 11â•… N. Alder-Baerens, Q. Lisman, L. Luong, T. Pomorski, J. C. Holthuis, Mol. Biol. Cell 2006, 17, 1632–1642. 12â•… T. Pomorski, A. Herrmann, P. Muller, G. van Meer, K. Burger, Biochemistry 1999, 38, 142–150. 13â•… T. Pomorski, R. Lombardi, H. Riezman, P. F. Devaux, G. van Meer, J. C. Holthuis, Cell 2003, 14, 1240–1254. 14â•… X. Tang, M. S. Halleck, R. A. Schlegel, P. Williamson, Science 1996, 272, 1495–1497. 15â•… W. R. Riekhof, D. R. Voelker, J. Biol. Chem. 2006, 281, 36588–36596. 16â•… W. R. Riekhof, J. Wu, M. A. Gijon, S. Zarini, R. C. Murphy, D. R. Voelker, J. Biol. Chem. 2007, 282, 36853–36861. 17â•… S. Chen, J. Wang, B. P. Muthusamy, K. Liu, S. Zare, R. J. Andersen, T. R. Graham, Traffic 2006, 7, 1503–1517. 18â•… X. Zhou, T. R. Graham, Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 16586–16591.

168

P-TYPE LIPID PUMPS AND DISEASE

19â•… M. Darland-Ransom, X. Wang, C. L. Sun, J. Mapes, K. Gengyo-Ando, S. Mitani, D. Xue, Science 2008, 320, 528–531. 20â•…C. C. Paulusma, A. Groen, C. Kunne, K. S. Ho-Mok, A. L. Spijkerboer, D. Rudi de Waart, F. J. Hoek, H. Vreeling, K. A. Hoeben, J. van Marle, L. Pawlikowska, L. N. Bull, A. F. Hofmann, A. S. Knisely, R. P. Oude Elferink, Hepatology 2006, 44, 195–204. 21â•… P. Ujhazy, D. Ortiz, S. Misra, S. Li, J. Moseley, H. Jones, I. M. Arias, Hepatology 2001, 34, 768–775. 22â•…C. C. Paulusma, D. E. Folmer, K. S. Ho-Mok, D. R. de Waart, P. M. Hilarius, A. J. Verhoeven, R. P. Oude Elferink, Hepatology 2008, 47, 268–278. 23â•… L. R. Poulsen, R. L. Lopez-Marques, S. C. McDowell, J. Okkeri, D. Licht, A. Schulz, T. Pomorski, J. F. Harper, M. G. Palmgren, Plant Cell 2008, 20, 658–676. 24â•… R. L. Lopez-Marques, L. R. Poulsen, S. Hanisch, K. Meffert, M. J. Buch-Pedersen, M. K. Jakobsen, T. G. Pomorski, M. G. Palmgren, Mol. Biol. Cell 2010, 21, 791–801. 25â•… J. A. Coleman, C. M. Kwok, R. S. Molday, J. Biol. Chem. 2009, 284, 32670–32679. 26â•… G. Lenoir, P. Williamson, C. F. Puts, J. C. Holthuis, J. Biol. Chem. 2009, 284, 17956–17967. 27â•… K. B. Axelsen, M. G. Palmgren, J. Mol. Evol. 1998, 46, 84–101. 28â•… W. Kuhlbrandt, Nat. Rev. Mol. Cell Biol. 2004, 5, 282–295. 29â•…C. Toyoshima, G. Inesi, Annu. Rev. Biochem. 2004, 73, 269–292. 30â•…C. Olesen, M. Picard, A. M. Winther, C. Gyrup, J. P. Morth, C. Oxvig, J. V. Moller, P. Nissen, Nature 2007, 450, 1036–1042. 31â•… J. P. Morth, B. P. Pedersen, M. S. Toustrup-Jensen, T. L. Sorensen, J. Petersen, J. P. Andersen, B. Vilsen, P. Nissen, Nature 2007, 450, 1043–1049. 32â•… T. Shinoda, H. Ogawa, F. Cornelius, C. Toyoshima, Nature 2009, 459, 446–450. 33â•… B. P. Pedersen, M. J. Buch-Pedersen, J. P. Morth, M. G. Palmgren, P. Nissen, Nature 2007, 450, 1111–1114. 34â•… G. Lenoir, P. Williamson, J. C. Holthuis, Curr. Opin. Chem. Biol. 2007, 11, 654–661. 35â•… J. Ding, Z. Wu, B. P. Crider, Y. Ma, X. Li, C. Slaughter, L. Gong, X. S. Xie, J. Biol. Chem. 2000, 275, 23378–23386. 36â•… K. Saito, K. Fujimura-Kamada, N. Furuta, U. Kato, M. Umeda, K. Tanaka, Mol. Biol. Cell. 2004, 15, 3418–3432. 37â•… Y. Katoh, M. Katoh, Oncol. Rep. 2004, 12, 939–943. 38â•… N. Furuta, K. Fujimura-Kamada, K. Saito, T. Yamamoto, K. Tanaka, Mol. Biol. Cell 2007, 18, 295–312. 39â•… K. Geering, J. Bioenerg. Biomembr. 2001, 33, 425–438. 40â•…C. F. Puts, J. C. Holthuis, Biochim. Biophys. Acta 2009, 1791, 603–611. 41â•… Z. Hua, P. Fatheddin, T. R. Graham, Mol. Biol. Cell. 2002, 13, 3162–3177. 42â•… T. R. Graham, Trends Cell Biol. 2004, 14, 670–677. 43â•… W. E. Gall, N. C. Geething, Z. Hua, M. F. Ingram, K. Liu, S. I. Chen, T. R. Graham, Curr. Biol. 2002, 12, 1623–1627. 44â•… S. Wicky, H. Schwarz, B. Singer-Kruger, Mol. Cell. Biol. 2004, 24, 7402–7418.

References

169

45â•… A. F. Ruaud, L. Nilsson, F. Richard, M. K. Larsen, J. L. Bessereau, S. Tuck, Traffic 2009, 10, 88–100. 46â•… S. Chantalat, S. K. Park, Z. Hua, K. Liu, R. Gobin, A. Peyroche, A. Rambourg, T. R. Graham, C. L. Jackson, J. Cell. Sci. 2004, 117, 711–722. 47â•… K. Liu, K. Surendhran, S. F. Nothwehr, T. R. Graham, Mol. Biol. Cell 2008, 19, 3526–3535. 48â•… W. Birchmeier, J. H. Lanz, K. H. Winterhalter, M. J. Conrad, J. Biol. Chem. 1979, 254, 9298–9304. 49â•… P. Muller, T. Pomorski, A. Herrmann, Biochem. Biophys. Res. Commun. 1994, 199, 881–887. 50â•… E. Farge, D. M. Ojcius, A. Subtil, A. Dautry-Varsat, Am. J. Physiol. 1999, 276, C725–C733. 51â•… P. Ezanno, S. Cribier, P. F. Devaux, Eur. Biophys. J. 2009, 39, 1277–1280. 52â•… X. L. Sun, D. Li, J. Fang, I. Noyes, B. Casto, K. Theil, C. Shuler, G. E. Milo, Gene Expr. 1999, 8, 129–139. 53â•… M. Dhar, L. Hauser, D. Johnson, Obes. Res. 2002, 10, 695–702. 54â•… S. Collins, T. L. Martin, R. S. Surwit, J. Robidoux, Physiol. Behav. 2004, 81, 243–248. 55â•… M. Andrew Nesbit, M. R. Bowl, B. Harding, D. Schlessinger, M. P. Whyte, R. V. Thakker, Genomics 2004, 84, 1060–1070. 56â•… L. Wang, C. Beserra, D. L. Garbers, Dev. Biol. 2004, 267, 203–215. 57â•… L. N. Bull, M. J. van Eijk, L. Pawlikowska, J. A. DeYoung, J. A. Juijn, M. Liao, L. W. Klomp, N. Lomri, R. Berger, B. F. Scharschmidt, A. S. Knisely, R. H. Houwen, N. B. Freimer, Nat. Genet. 1998, 18, 219–224. 58â•… S. W. van Mil, L. W. Klomp, L. N. Bull, R. H. Houwen, Semin. Liver Dis. 2001, 21, 535–544. 59â•… S. W. van Mil, R. H. Houwen, L. W. Klomp, J. Med. Genet. 2005, 42, 449–463. 60â•… R. Mullenbach, A. Bennett, N. Tetlow, N. Patel, G. Hamilton, F. Cheng, J. Chambers, R. Howard, S. D. Taylor-Robinson, C. Williamson, Gut 2005, 54, 829–834. 61â•… J. N. Painter, M. Savander, A. Ropponen, N. Nupponen, S. Riikonen, O. Ylikorkala, A. E. Lehesjoki, K. Aittomaki, Eur. J. Hum. Genet. 2005, 13, 435–439. 62â•… L. W. Klomp, J. C. Vargas, S. W. van Mil, L. Pawlikowska, S. S. Strautnieks, M. J. van Eijk, J. A. Juijn, C. Pabon-Pena, L. B. Smith, J. A. DeYoung, J. A. Byrne, J. Gombert, G. van der Brugge, R. Berger, I. Jankowska, J. Pawlowska, E. Villa, A. S. Knisely, R. J. Thompson, N. B. Freimer, R. H. Houwen, L. N. Bull, Hepatology 2004, 40, 27–38. 63â•… E. F. Eppens, S. W. van Mil, J. M. de Vree, K. S. Mok, J. A. Juijn, R. P. Oude Elferink, R. Berger, R. H. Houwen, L. W. Klomp, J. Hepatol. 2001, 35, 436–443. 64â•… S. S. Strautnieks, L. N. Bull, A. S. Knisely, S. A. Kocoshis, N. Dahl, H. Arnell, E. Sokal, K. Dahan, S. Childs, V. Ling, M. S. Tanner, A. F. Kagalwalla, A. Nemeth, J. Pawlowska, A. Baker, G. Mieli-Vergani, N. B. Freimer, R. M. Gardiner, R. J. Thompson, Nat. Genet. 1998, 20, 233–238. 65â•… S. W. van Mil, W. L. van der Woerd, G. van der Brugge, E. Sturm, P. L. Jansen, L. N. Bull, I. E. van den Berg, R. Berger, R. H. Houwen, L. W. Klomp, Gastroenterology 2004, 127, 379–384.

170

P-TYPE LIPID PUMPS AND DISEASE

66â•… S. S. Strautnieks, J. A. Byrne, L. Pawlikowska, D. Cebecauerova, A. Rayner, L. Dutton, Y. Meier, A. Antoniou, B. Stieger, H. Arnell, F. Ozcay, H. F. Al-Hussaini, A. F. Bassas, H. J. Verkade, B. Fischler, A. Nemeth, R. Kotalova, B. L. Shneider, J. Cielecka-Kuszyk, P. McClean, P. F. Whitington, E. Sokal, M. Jirsa, S. H. Wali, I. Jankowska, J. Pawlowska, G. Mieli-Vergani, A. S. Knisely, L. N. Bull, R. J. Thompson, Gastroenterology 2008, 134, 1203–1214. 67â•… A. C. Kurbegov, K. D. Setchell, J. E. Haas, G. W. Mierau, M. Narkewicz, J. D. Bancroft, F. Karrer, R. J. Sokol, Gastroenterology 2003, 125, 1227–1234. 68â•… A. Groen, C. Kunne, G. Jongsma, K. van den Oever, K. S. Mok, M. Petruzzelli, C. L. Vrins, L. Bull, C. C. Paulusma, R. P. Oude Elferink, Gastroenterology 2008, 134, 2091–2100. 69â•…C. C. Paulusma, D. R. de Waart, C. Kunne, K. S. Mok, R. P. Oude-Elferink, J. Biol. Chem. 2009, 284, 9947–9954. 70â•… S. W. van Mil, M. M. van Oort, I. E. van den Berg, R. Berger, R. H. Houwen, L. W. Klomp, Pediatr. Res. 2004, 56, 981–987. 71â•… N. Tygstrup, B. A. Steig, J. A. Juijn, L. N. Bull, R. H. Houwen, Hepatology 1999, 29, 506–508. 72â•… P. F. Whitington, D. K. Freese, E. M. Alonso, S. J. Schwarzenberg, H. L. Sharp, J. Pediatr. Gastroenterol. Nutr. 1994, 18, 134–141. 73â•… T. Oshima, K. Ikeda, T. Takasaka, Tohoku J. Exp. Med. 1999, 187, 83–88. 74â•… P. Lykavieris, S. van Mil, D. Cresteil, M. Fabre, M. Hadchouel, L. Klomp, O. Bernard, E. Jacquemin, J. Hepatol. 2003, 39, 447–452. 75â•… P. M. Verhulst, L. M. van der Velden, V. Oorschot, E. E. van Faassen, J. Klumperman, R. H. Houwen, T. G. Pomorski, J. C. Holthuis, L. W. Klomp, Hepatology 2010, 51, 2049–2060. 76â•… J. M. Stapelbroek, T. A. Peters, D. H. van Beurden, J. H. Curfs, A. Joosten, A. J. Beynon, B. M. van Leeuwen, L. M. van der Velden, L. Bull, R. P. Oude Elferink, B. A. van Zanten, L. W. Klomp, R. H. Houwen, Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 9709–9714. 77â•… S. M. Paul, M. J. Palladino, G. J. Beitel, Development 2007, 134, 147–155. 78â•… Z. Zhang, P. Devarajan, A. L. Dorfman, J. S. Morrow, J. Biol. Chem. 1998, 273, 18681–18684. 79â•… M. A. Costa de Beauregard, E. Pringault, S. Robine, D. Louvard, EMBO J. 1995, 14, 409–421. 80â•… J. C. M. Holthuis, T. P. Levine, Nat. Rev. Mol. Cell. Biol. 2005, 6, 209–220.

9 COUPLING DRS2P TO PHOSPHOLIPID TRANSLOCATION, MEMBRANE ASYMMETRY, AND VESICLE BUDDING Xiaoming Zhou, Paramasivam Natarajan, Baby-Periyanayaki Muthusamy, and Todd R. Graham Department of Biological Sciences, Vanderbilt University, Nashville, TN

Ke Liu NIH Chemical Genomics Center, Bethesda, MD

9.1â•… INTRODUCTION Membranes provide a physical boundary to separate a living organism from its environment and divide eukaryotic cells into functionally distinct subcellular compartments. The plasma membrane of the cell has a unique protein and lipid composition that is optimized for establishing an interface between the intracellular and extracellular environments. An important aspect of this interface is the asymmetrical distribution of plasma membrane phospholipids between the two leaflets. For example, the cytosolic leaflet is enriched in phosphatidylserine (PS) and phosphatidylethanolamine (PE), while the extracellular leaflet primarily contains phosphatidylcholine (PC) and sphingolipids. This lipid organization appears to be conserved from yeast to man, but how is membrane asymmetry established and maintained? A growing body of evidence indicates that P4-ATPases contribute significantly to this process by Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

171

172

YEAST P4-ATPASES

flipping specific phospholipid substrates from the exofacial leaflet to the cytosolic leaflet of the membrane bilayer. In addition, studies using the model yeast Saccharomyces cerevisiae first implicated P4-ATPases in the formation of transport vesicles in the secretory and endocytic pathways. Deciphering the mechanism for coupling phospholipid translocation to transport vesicle budding remains an active area of investigation. Here we review the P4ATPases from S. cerevisiae with an emphasis on studies of Drs2p and its chaperone subunit Cdc50p, and describe the contributions of these P4-ATPases to the establishment of membrane asymmetry and vesicle-mediated protein transport. 9.2â•… P4-ATPASES IN BUDDING YEAST 9.2.1â•… Nomenclature Sequencing of the S. cerevisiae genome identified a total of 16 P-type ATPases, which were initially categorized phylogenetically into three classes called P1-, P2-, and P4-ATPases [1]. Drs2p family members, the potential flippases, were originally placed in the P2 class together with H+, Na+, and Ca2+ pumps, whereas the term P4 was proposed for two new P-type ATPases whose substrates are still unknown. In a more recent and detailed phylogenetic classification that included P-type ATPases from bacteria, fungi, plants, and animals, the Drs2p family was placed into an independent new branch designated type IV P-type ATPases [2, 3]. The P-type ATPases with unknown substrates, that is, the original P4-ATPases, were designated type V P-type ATPases. To promote a less cumbersome nomenclature, we and other groups have used the term P4ATPases for the Drs2p family of potential phospholipid translocases. 9.2.2â•… P4-ATPases in Yeast The yeast S. cerevisiae contains five P4-ATPases: Drs2p, Neo1p, Dnf1p, Dnf2p, and Dnf3p [2, 3] (Fig. 9.1). NEO1 is the only essential gene of the P4-ATPase family and yeast carrying a complete disruption of NEO1 fail to grow at any temperature tested [4, 5]. However, phenotypes associated with loss of NEO1 function have been characterized in strains carrying temperature-conditional alleles of NEO1 (neo1-ts) or by placing the NEO1 gene under transcriptional control of a galactose-regulated promoter [6, 7]. In the latter case, expression of NEO1 can be shut off by growth of the strain in glucose. The remaining members of the P4-ATPase family constitute an essential group [4]. Disruption of DRS2 yields a viable strain although these cells display an unusually strong cold-sensitive growth defect. Yeast normally grow over a range of temperatures, from approximately 10 to 40°C, but drs2Δ cells fail to grow at temperatures below 23°C [8]. Individual knockout of DNF genes, or even the triple knockout (dnf1,2,3Δ), does not perturb growth of

173

P4-ATPASES IN BUDDING YEAST

C N Cytosolic

N

C

Lumenal/ Extracellular

PtdSer

PtdEtn

PtdCho Sphingomyelin

Drs2p Dnf1p Dnf2p } Dnf3p Neo1p

Cdc50p/Swa4p Lem3p/Ros3p Crf1p ?

Figure 9.1.â•‡ Membrane asymmetry and yeast P4-ATPases. The catalytic α-subunit is modeled on the Drs2p overlay homology model with SERCA1 shown in Figure 9.2. The connecting lines indicate the noncatalytic β-subunit required to chaperone each P4-ATPase out of the ER. No β-subunit for Neo1p has been identified. Color version on the Wiley web site.

yeast. However, disruption of DNF genes in a drs2Δ background exacerbates the growth defect caused by drs2Δ, in decreasing order of severity: drs2Δdnf1Δ, drs2Δdnf2Δ, and drs2Δdnf3Δ. Growth defects become more severe as more DRS2-DNF genes are deleted and the quadruple mutant (drs2Δdnf1,2,3Δ) is inviable. This indicates some degree of functional redundancy between the DRS2 and DNF genes, with Drs2p primarily providing the essential activity of the group [4]. DRS2 was the first P4-ATPase gene cloned and was originally thought to be a Ca2+-ATPase based on homology searches with sequences available at the time [8]. For example, Drs2p is 14% identical and 41% similar to the sarcoplasmic/endoplasmic reticulum (ER) Ca2+-ATPase 1 (SERCA1). By comparison, Drs2p shares 20%–30% identity and 40%–50% similarity with other yeast P4-ATPases [4]. Sequence analysis of Drs2p identified the key sequence “DKTGTLT,” containing the aspartyl phosphorylation site conserved in all P-type ATPases, as well as other signature motifs for P-type ATPases. Drs2p was also predicted to contain 10 transmembrane segments as does SERCA1. SERCA1 can be considered an archetype of the entire P-type ATPase family because it was the first P-type ATPase whose atomic structure

174

YEAST P4-ATPASES

N

N A

A

P

P

TM

TM

SERCA1;Drs2p

Figure 9.2.â•‡ Homology model of Drs2p based on the crystal structure of the sarcoplasmic/ endoplasmic reticulum Ca2+-ATPase 1 (SERCA1) in the E2 conformation (PDB ID: 1wpg). Green: SERCA1 template; red: Drs2p. Homology modeling was done with SWISS-MODEL server [110–112]. P, phosphorylation domain; A, actuator domain; N, nucleotide-binding domain; TM, transmembrane domain. Color version on the Wiley web site.

was determined by X-ray crystallography [9]. Moreover, the crystal structures of SERCA1 in several different conformational states have been solved, providing the first structure-based mechanistic model for the pumping cycle of a P-type ATPase [10]. An atomic resolution structure of a P4-ATPase would be of great value to gain mechanistic insight into the function of these enzymes, but unfortunately, this has not yet been achieved. However, we have built a homology model of Drs2p by threading its sequence on the atomic structure of SERCA1 (Fig. 9.2). The premise of this method is a high degree of sequence similarity between the target and the template protein, which normally requires 60%–70% similarity or above. However, using the predicted transmembrane segments as anchor points, the intervening loops of Drs2p could be successfully threaded onto the SERCA1 structure. In this homology model, Drs2p has a 10-transmembrane-segment membrane domain, with three well-organized cytosolic domains easily recognized as the actuator domain (A), nucleotidebinding domain (N), and phosphorylation domain (P). In the overlay mode, the majority of Drs2p can be superimposed well with the SERCA1 template,

Evidence That Drs2p Is a Flippase

175

despite their relatively low homology. This homology model of Drs2p is not sufficient to substitute for an experimentally determined atomic model, but rather serves as complementary information that may facilitate mutational studies. 9.2.3â•… P4-ATPase Chaperones Recent studies showed that most yeast P4-ATPases require Cdc50p family members (Cdc50p, Lem3p and Crf1p) for exit from the ER [11–13]. In the absence of Lem3p, Dnf1p and Dnf2p show a characteristic ER pattern that is distinct from their normal localization. Similarly, the absence of Cdc50p causes ER localization of Drs2p, and absence of Crf1p causes ER retention of Dnf3p. The chaperone function between the P4-ATPases and the Cdc50p family is reciprocal. Cdc50p and Lem3p are also retained in the ER when their corresponding P4-ATPase partners are missing. In wild-type cells, Cdc50p colocalizes with Drs2p, and Lem3p with Dnf1p/Dnf2p. Co-immunoprecipitation between P4-ATPases and Cdc50p family members also supports the formation of physical complexes between Drs2p and Cdc50p, and Dnf1p (or Dnf2p) and Lem3p (Fig. 9.1). The relationship between P4-ATPases and Cdc50p family members is reminiscent of Na+/K+ and H+/K+ ATPases, which are among a few oligomeric P-type ATPases possessing a β-subunit in addition to the catalytic α-subunit [14]. Cdc50p family members may be considered the β-subunit for P4-ATPases (α-subunit). An important unanswered question is whether the β-subunits function solely as a chaperone or if they also contribute in a more direct manner to phospholipid translocation and in vivo function of the α-subunits (see Chapter 8). Colocalization and co-immunoprecipitation data suggest that the α/β-subunits remain associated after ER exit [11], but this does not reflect a requirement for the function of β-subunits beyond the ERâ•¯→â•¯Golgi transport step. Since it is not possible to directly evaluate the function of Drs2p in its native Golgi membrane environment in the absence of Cdc50p, alternative strategies are required. For example, a conditional allele of CDC50 that disrupts its association with Drs2p could potentially be used to assess the requirement of the Drs2p-Cdc50p interaction for the flippase activity. 9.3â•… EVIDENCE THAT DRS2P IS A FLIPPASE 9.3.1â•… Flip-Flop of Phospholipids Phospholipids are amphiphilic molecules that can freely diffuse laterally in their own leaflet, but face a substantial barrier to transverse movement across the bilayer, known as flip-flop, due to their hydrophilic head groups. This barrier can be circumvented by involvement of lipid transporters, and several different types of lipid transporters are proposed to exist in eukaryotic cells [15–18]. The ER contains an ATP-independent flippase activity that moves

176

YEAST P4-ATPASES

phospholipids bidirectionally and without apparent head-group specificity [19]. This ER flippase activity is presumably responsible for balanced growth of both leaflets of the ER during phospholipid synthesis, which appears to be catalyzed primarily on the cytosolic leaflet. Therefore, the ER membrane should be symmetric in lipid distribution. Membrane asymmetry appears to be generated in post-ER compartments by ATP-consuming flippases and floppases that translocate specific lipid substrates unidirectionally across the membrane bilayer. Flippases mediate the inward movement of phospholipid from the extracellular/luminal leaflet to the cytosolic leaflet of the membrane, whereas floppases act in the reverse, outward direction. Scramblases, upon activation, collapse the lipid asymmetry of the plasma membrane and result in PS and PE exposure on the outer leaflet. Despite decades of effort to uncover the identities of these lipid transporters, much uncertainty remains. The ER flippase has not been identified yet [20] (see Chapter 6), nor is the identity of the scramblase known even though several candidates have been suggested [21] (see Chapter 7), including the ABC1/CED-7 ABC transporters [22–24], phospholipid scramblase (PLSCR) [24, 25], and the TAT-1 P4-ATPases [26]. A class of ATP-binding-cassette (ABC) transporters involved in multidrug resistance and bile secretion appears to be responsible for the floppase activity. For example, the human MDR3 and the mouse MDR2 P-glycoproteins specifically flop a PC analog, 7-nitrobenz2-oxa-1,3-diazol-4-yl (NBD)-PC [27], while the human MDR1 and the mouse MDR1a flop lipids with a broader substrate specificity including NBD-PC, NBD-PE, and NBD-sphingomyelin [28–30]. Specific phospholipids (such as PS and PE) are rapidly transported inwardly by flippases, while floppases move a much broader spectrum of lipids outwardly. These activities may establish the observed phospholipid asymmetry of the plasma membrane. The primary candidates for the inward directed flippase activity are members of the P4-ATPase family [31–37]. 9.3.2â•… Implication of Drs2p as a Flippase A flippase activity was first observed in the plasma membrane of human red blood cells (RBCs) and was subsequently found to exist in the plasma membrane of many other cell types [38, 39]. Spin-labeled PS and, to a lesser extent, PE, but not PC, were found to undergo rapid transverse movement across the bilayer resulting in their enrichment in the inner, cytosolic leaflet. This flippase activity is Mg2+-ATP-dependent and sensitive to the ATPase inhibitor orthovanadate and N-ethylmaleimide, which are in good accordance with the properties of a novel ATPase, called ATPase II (now designated ATP8A1) purified from bovine chromaffin granules [40]. The connection between the RBC flippase activity and ATPase II was further enhanced by the observation of a PS flippase activity in bovine chromaffin granules with a similar inhibitor spectrum as ATPase II [41]. Tang et al. cloned a cDNA encoding bovine ATPase II and found that it shares 47% amino acid sequence identity and

Evidence That Drs2p Is a Flippase

177

67% similarity to the yeast protein Drs2p [31]. Even though there have been several reports describing the purification of ATPase II [31, 42, 43], this enzyme has not been reconstituted into liposomes to demonstrate that it can directly catalyze phospholipid flippase activity. 9.3.3â•… Yeast Plasma Membrane Flippase Activities A number of studies have tested whether P4-ATPases are necessary for flippase activity using genetic approaches in yeast. As with mammalian cells, yeast also sequester most PS and PE to the inner leaflet of the plasma membrane, and possess a flippase activity there that mediates translocation of PS and PE fluorescent analogs (NBD-PS and NBD-PE) [12, 31, 33, 44–47]. Distinct from the RBC plasma membrane flippase, NBD-PC is also actively internalized by translocation [33, 44, 45], indicating that the yeast plasma membrane flippase(s) has a broader substrate spectrum than the mammalian RBC counterpart. From the early studies after cloning of bovine ATPase II, a major debate that is still unresolved has centered on the contribution of Drs2p to the yeast plasma membrane flippase activity that mediates NBD-PS uptake. Tang et al. exploited a back-exchange approach to monitor the translocation of NBD-PS across the yeast plasma membrane under conditions of partial energy depletion and low temperature (4°C) to block endocytic uptake of the probe [31]. Wild-type yeast cells showed substantial uptake of PS analog while drs2Δ cells were defective in doing so, suggesting a critical role for Drs2p in NBD-PS flip. However, Siegmund et al. found no significant difference between a drs2Δ strain and its isogenic wild-type parent for NBD-PS uptake or distribution at low temperature using fluorescence microscopy or flow cytometry [48]. The controversy continued as Marx et al. used an endocytosis mutant (end4Δ) to prevent endocytic uptake of the probe, rather than low temperature and partial energy depletion, and found only a mild reduction of NBD-PS internalization in drs2Δend4Δ cells compared with the parental end4Δ strain [49]. However, in another study on the ALA1 gene product, a Drs2p homolog in the plant Arabidopsis thaliana, Gomes et al. repeated the original observation made by Tang et al. and showed that heterologous expression of ALA1 in drs2Δ cells complemented the NBD-PS internalization defect [32]. The reasons for these experimental discrepancies are still unclear. However, factors that could have contributed to these contradictory results were the presence of other lipid transporters that may also contribute to NBD-PS uptake, the pleiotropic phenotypes caused by deletion of the DRS2 gene, and lack of information on the subcellular localization of Drs2p. We know now that Drs2p primarily localizes to trans-Golgi network (TGN) rather than the plasma membrane [4, 50]. A relatively small percentage of Drs2p traffics to the plasma membrane, but the presence of multiple endocytosis signals in Drs2p ensures rapid retrieval back to the TGN [51, 52]. The slow egress to the plasma membrane coupled with rapid removal leads to undetectable levels of Drs2p at the plasma membrane of wild-type cells grown under standard

178

YEAST P4-ATPASES

conditions. The influence of partial energy depletion and low temperature on Drs2p localization has not been tested. However, blocking endocytosis with an end4Δ mutant does cause a near complete redistribution of Drs2p to the plasma membrane [51]. Therefore, the work of Marx et al. suggested that Drs2p makes a relatively minor contribution to NBD-PS uptake across the plasma membrane under conditions where Drs2p is primarily localized to the plasma membrane. Overexpression of Drs2p and Cdc50p in an endocytosis mutant, conditions that should dramatically increase the amount of Drs2p on the plasma membrane, only increased NBD-PS uptake across the plasma membrane two- to threefold relative to control strains [11]. As described below, Drs2p is required for a robust NBD-PS flippase activity in Golgi membranes [34] and so it is surprising that redeployment of Drs2p to the plasma membrane leads to a relatively minor enhancement of NBD-PS flip in this location. We suggest that Drs2p is not fully active when it is at the plasma membrane as positive regulators of Drs2p activity are localized to the TGN [53]. Two other members of the P4-ATPase family, Dnf1p and Dnf2p, are easily detected on the plasma membrane and are therefore more likely candidates for the plasma membrane flippase activities [4]. Consistently, the energydependent uptake of NBD-PC and NBD-PE across the yeast plasma membrane is virtually abolished in the dnf1,2Δ double deletion mutant [33]. In addition, the lem3Δ mutant, which accumulates Dnf1p and Dnf2p in the ER, also exhibits a near complete defect in NBD-PC and NBD-PE uptake [11, 47, 54]. Importantly, the internalization of lyso-PE across the yeast plasma membrane is also impaired in dnf1,2Δ or lem3Δ cells, suggesting that Dnf1p-Dnf2p/ Lem3p can also pump a naturally occurring phospholipid across the plasma membrane [55]. Lyso-phospholipids are not transported at an appreciable rate by the RBC flippase, again suggesting a unique mode of substrate recognition for the yeast plasma membrane flippases [56, 57]. Other important potential substrates of Dnf1p-Dnf2p/Lem3p transport activity are edelfosine and miltefosine [54]. These drugs are PC analogs that are toxic to leishmania and trypanosomes. The lem3Δ and dnf1,2Δ yeast strains are resistant to these drugs, as are leishmania harboring mutations in the analogous genes [54, 58]. There remains some controversy on the influence of Dnf1p and Dnf2p on NBD-PS translocation across the plasma membrane. The dnf1,2Δ cells were originally reported to be deficient in internalization of NBD-PS in addition to NBD-PC and NBD-PE [33]. However, several groups have found that in the lem3Δ strain, which should show the same phenotypes as dnf1,2Δ strains, only the NBD-PC and NBD-PE translocation uptake is impaired, leaving NBD-PS internalization unaffected or even increased [11, 47, 54]. A recent study also failed to detect any defect in NBD-PS uptake in dnf1,2Δ cells [46]. Moreover, this group tested a variety of strains carrying mutations in multiple P4-ATPases and found no deficit in NBD-PS uptake. These data suggest that NBD-PS uptake across the plasma membrane is not catalyzed by a P4-ATPase, with the caveat that neo1 mutants were not tested. It seems that the proteins respon-

Evidence That Drs2p Is a Flippase

179

sible for NBD-PS uptake across the yeast plasma membrane are still mysterious. Because the P4-ATPase mutants also perturb protein trafficking [4, 6, 33, 50, 52], these mutants have great potential for displaying pleiotropic phenotypes, for example, by mislocalization of enzymes more directly involved in a transport process, perhaps contributing to the disparate data on the influence of drs2 and dnf mutations on NBD-PS uptake across the plasma membrane. 9.3.4â•… Drs2p-Dependent Flippase Activity in Golgi Membranes Direct evidence supporting the potential flippase activity of Drs2p came from experiments designed to test if Drs2p is required for an NBD-PS flippase activity in yeast TGN membranes, where Drs2p is primarily localized. TGN membranes isolated from wild-type yeast cells contained an ATP-dependent activity that translocates NBD-PS and, to a lesser extent, NBD-PE, from the inner (luminal) leaflet to the outer (cytosolic) leaflet of the membrane. No NBD-PC flippase activity was observed with TGN membranes. This assay required the presence of NBD-phospholipid probes in the inner leaflet of the isolated TGN membrane, which was achieved by initial incorporation of probes in the outer leaflet and incubation of the membrane in the absence of ATP to allow passive diffusion of probe to the inner leaflet. Flipping of NBDphospholipid probes back to the outer leaflet, induced by ATP addition, was monitored by accessibility of the probe to bovine serum albumin (BSA) backextraction [34]. Because of the controversies surrounding the influence of deleting DRS2 on NBD-PS translocation across the plasma membrane, we sought a method to assess the contribution of Drs2p to the TGN flippase activity that could avoid the pleiotropic consequences of deleting DRS2. Ideally, one would like to compare the flippase activity in two TGN membrane preparations that are identical except for the presence or absence of one protein, in this case Drs2p. However, deletion of DRS2 causes mislocalization of TGN resident proteins, and TGN membranes from drs2Δ cells were deficient in multiple proteins relative to wild-type control membranes [4, 50]. To circumvent this problem, a temperature-conditional allele of drs2 was isolated that could be used to assess the immediate consequences of inactivating Drs2p function [59]. Strains carrying Drs2p-ts as the sole source of Drs2p were grown at permissive temperature, where Drs2p-ts is active, so a “normal” TGN membrane preparation could be harvested. This membrane preparation was split in half and assayed for flippase activity at permissive and nonpermissive temperatures [34]. Thus, two membrane samples that are identical except for the presence or absence of the activity of a single enzyme could be compared. A robust ATP-dependent translocation activity for NBD-PS was detected at permissive temperature (27°C) but was lost when the membranes were assayed at a nonpermissive temperature (37°C). As a control, the TGN membrane isolated from an isogenic strain containing wild-type DRS2 showed a similar NBD-PS flippase activity at both 27 and 37°C. These results indicate that Drs2p activity

180

YEAST P4-ATPASES

is necessary for the NBD-PS flippase activity in the TGN membrane, and Drs2p most likely catalyzes this activity directly. No NBD-PE flippase activity was detected with the temperature-conditional mutant form of Drs2p, possiÂ� bly due to insufficient sensitivity of this assay for detection of NBD-PE translocation. Surprisingly, PS is not an obligatory substrate for Drs2p function in vivo even if it appears to be the preferred substrate in vitro. Loss of Drs2p function (drs2Δ) perturbs various protein trafficking pathways [4, 50, 52]. If translocation of PS by Drs2p is required to support the role of Drs2p in protein trafficking, we would expect cells deficient for the PS to display a similar defect. However, cho1Δ yeast cells that lack de novo PS synthesis and are devoid of PS do not phenocopy the trafficking defects of drs2Δ [34]. The cho1Δ cells transport proteins normally via the secretory pathway and still require Drs2p for protein transport as trafficking defects are observed in drs2Δcho1Δ cells. Therefore, Drs2p must have at least one additional substrate other than PS. PE may be such a candidate substrate for Drs2p. Saito et al. showed that in addition to increased NBD-PS uptake, the accumulation of Drs2p-Cdc50p on the plasma membrane of endocytosis mutants also increased NBD-PE internalization [11]. Consistently, Alder-Baerens et al. reported a Drs2pdependent activity that was responsible for NBD-PS and NBD-PE translocation across yeast secretory vesicle membranes when Drs2p was overexpressed [35]. Importantly, this group also showed an ATP-dependent translocation of endogenous PE to the cytosolic leaflet of the vesicle membrane, and this activity was abolished in a drs2Δdnf3Δ strain. Thus, Drs2p may flip both PE and PS across the TGN membrane, and it is possible that translocation of either substrate is sufficient to support Drs2p-dependent protein trafficking pathways. However, it is also possible that Drs2p has additional lipid or nonlipid substrates that contribute to vesicle budding. Although a substantial amount of evidence supported the hypothesis that Drs2p and P4-ATPases in general are flippases, the possibility that they may not be direct phospholipid transporters could not be ruled out. For example, in an alternative model, Drs2p family members may pump an undetermined ion to establish an ion gradient across the membrane, which can be used as an energy source by a second transporter (symporter) that is coupled to phospholipid translocation. This hypothesis has been partially tested in the isolated TGN membranes with a Drs2p-dependent NBD-PS flippase activity [34]. Substitution of other ions for Na+ and Cl−, which were the major ions in the assay, did not affect the NBD-PS translocation. Thus, if Drs2p pumps an ion (or heavy metal) across the membrane, it would have to either do so with a lack of ion specificity or be able to drive the lipid translocation process using a very dilute external ion concentration (potential contaminating ions in the sample). Potentially supporting a secondary transport model, a proton gradient seems to be necessary for the translocation of NBD-PC, NBD-PE, and NBD-PS across the yeast plasma membrane [46, 60, 61]. This gradient is generated by the plasma membrane H+-ATPase (Pma1p) rather than the P4-ATPases

Evidence That Drs2p Is a Flippase

181

Dnf1p/Dnf2p, so how a proton gradient contributes to the Dnf-dependent flippase activity is still unclear. The most definitive approach to determine if P4-ATPase family members are flippases is to biochemically reconstitute a purified P4-ATPase into chemically defined proteoliposomes and assay for flippase activity. 9.3.5â•… Drs2p and Atp8b2 Flippase Activity in Proteoliposomes Two different P4-ATPases have recently been purified, reconstituted into proteoliposomes, and shown to directly catalyze phospholipid flippase activity [36, 37]. Drs2p was purified using either an N-terminal or C-terminal tandem affinity purification (TAP) tag from Saccharomyces strains also overexpressing Cdc50p [37]. Even though both forms of Drs2p are functional in vivo, only the N-terminally tagged protein (TAP-Drs2p) retained ATPase activity in detergent micelles after purification. TAP-Drs2p was sensitive to vanadate (5â•¯µM IC50), displayed a pH optimum at ∼pH 7.5, a Km for ATP of 1.5â•¯mM, and a Vmax of 0.45â•¯µmol ATP hydrolyzed/min/mg. This Vmax is 10- to 50-fold lower than comparable preparations of Atp8A1, and surprisingly, Drs2p ATPase activity was only mildly stimulated by PS [37]. Drs2p was reconstituted into 99% PC, 1% NBD-phospholipid vesicles with the goal of incorporating one Drs2p molecule on average per liposome. However, the resulting proteoliposomes were smaller than anticipated (40-nm average diameter), and only ∼40% of the liposomes were predicted to contain Drs2p. Fortunately, most of the Drs2p was oriented with the ATPase domain facing outward and therefore was capable of being activated by external ATP. Under these conditions, Mg++-ATP induced translocation of ∼4% of the NBD-PS with the TAP-Drs2p proteoliposomes, while no activity toward NBD-PC or NBD-sphingomyelin was detected. As a control, the ATPase-dead C-terminally TAP-tagged form was reconstituted and failed to flip any NBDPS. Thus, it appeared that Drs2p is sufficient in a purified form to catalyze a phospholipid flippase activity. However, the proteoliposomes did contain Cdc50p, estimated to be 1/10 the amount of Drs2p in the samples, and we observed about 10% the expected flippase activity [37]. Therefore, it is possible that the active flippase unit is the Drs2p-Cdc50p heterodimer and further work is required to fully assess the contribution of Cdc50p to the Drs2p flippase activity. The Molday group also succeeded in reconstituting an NBD-PS flippase activity with a purified P4-ATPase [36]. Atp8a2 is highly expressed in the retina and appears to account for ∼50% of the basal ATPase activity in outer segment extracts. Atp8a2 was purified from photoreceptor outer segments by affinity chromatography using a monoclonal antibody. As with Atp8a1, ATPase activity was strongly and specifically stimulated by PS, weakly by PE, and not at all by other lipids tested. The purified protein had a specific activity of ∼50â•¯µmol ATP hydrolyzed/min/mg, a Km of 78â•¯µM for PS, 660â•¯µM for PE, and 0.7â•¯mM for ATP in detergent. Moreover, Atp8a2 reconstituted into proteoliposomes

182

YEAST P4-ATPASES

displayed an NBD-PS-specific flippase activity. Importantly, this group also found that native PS competed for the NBD-PS probe in the flippase assay [36]. These results strongly support the contention that PS is the native substrate for Atp8a2 and related P4-ATPases. 9.3.6â•… P4-ATPases in Yeast Membrane Asymmetry One major function proposed for Drs2p and other P4-ATPases is to generate and maintain membrane phospholipid asymmetry [15, 17, 18] (Fig. 9.1). The fact that the yeast plasma membrane is asymmetrical with PS and PE restricted to the cytosolic leaflet suggests that the membrane asymmetry must be established as membranes flow through the Golgi and/or upon arrival at the plasma membrane. The detection of energy- and P4-ATPase-dependent translocation activities for NBD derivatives of PS, PE, and/or PC analogs in the TGN [34], post-Golgi secretory vesicles [35], and the plasma membrane [33, 46, 55] suggests that P4-ATPases may act at multiple sites to establish and maintain phospholipid asymmetry in yeast. While PS and PE are restricted to the cytosolic leaflet, the distribution of PC between the inner and outer leaflets of the yeast plasma membrane has not been characterized. The presence of an NBD-PC translocase activity suggests that PC might also be restricted to the inner leaflet. Intriguingly, the plasma membrane of yeast has nearly equal parts of glycosphingolipid and glycerophospholipid by mass [62, 63], perhaps suggesting a partitioning of these two lipid classes on either side of the plasma membrane. When the function of one or more P4-ATPases is disrupted, the asymmetric distribution of PS and PE on the plasma membrane is perturbed. The detection of cells exposing PS and PE on the outer leaflet is greatly facilitated by reagents that specifically target, bind, or react with these lipid species, including the use of Ro09-0198 (Ro) peptide for PE detection [12, 33, 47], papuamide B (Pap B) [12, 64] and Annexin V for PS detection [12, 23, 26], and trinitrobenzene sulfonic acid (TNBS) for aminophospholipid detection (PE and PS) [33, 35]. Ro is a tetracyclic peptide antibiotic that specifically binds PE exposed on the cell surface and causes cytolysis [65]. In a genetic approach to identify flippases in yeast that control phospholipid asymmetry of the plasma membrane, yeast cells were screened for mutants that exhibit hypersensitivity to Ro peptide [47]. One of the most sensitive mutants isolated was ros3 (for Rosensitivity), indicating more PE was exposed on the cell surface in this mutant. ROS3 (also known as LEM3) encodes the chaperone and potential β-subunit for Dnf1p and Dnf2p [11]. Moreover, cells losing both Dnf1p and Dnf2p functions (dnf1,2Δ) exposed more endogenous PE to the outer leaflet of the plasma membrane compared with wild-type cells, as judged by increased availability to TNBS labeling and increased Ro sensitivity [33]. Disruption of DRS2 in dnf1,2Δ cells caused even more PE exposure, supporting the roles suggested for Drs2p in regulating the plasma membrane asymmetry by pumping PE to

Drs2p in Protein Transport and Vesicle Budding

183

the cytosolic leaflet of the Golgi prior to export of membrane in vesicles targeted to the plasma membrane. Consistently, isolated post-Golgi secretory vesicles also exhibited less PE to the cytosolic leaflet when the function of Drs2p and/or Dnf3p was disrupted [35]. Pap B, a cyclic lipopeptide, was shown to specifically target PS exposed on the cell surface to induce cytolysis [12, 64]. Hypersensitivity to Pap B and the PS-binding protein Annexin V were used to monitor the exposure of PS in the P4-ATPase and Cdc50p family mutants. drs2Δ, cdc50Δ, and several P4-ATPase mutant combinations, such as dnf1,2Δ, exposed significantly more endogenous PS on the cell surface than did wild-type cells. Particularly, the triple mutant drs2Δdnf1,2Δ exposed much more PS than either drs2Δ or dnf1,2Δ mutants, supporting the hypothesis that Drs2p at the TGN and Dnf1p/Dnf2p on the plasma membrane are working together to maintain PS asymmetry of the plasma membrane. An important caveat of this interpretation is that drs2Δ, cdc50Δ, and other mutants such as dnf1,2Δ also perturb the exocytic and endocytic protein transport pathways [4, 12, 50]. Thus, it is possible that loss of PS asymmetry of the plasma membrane is an indirect consequence of perturbation of protein and membrane trafficking in these mutants. In fact, clathrin and ADP-ribosylation factor (ARF) mutants (chc1Δ, clc1Δ, and arf1Δ), as well as endocytosis mutants (end3Δ and end4Δ), also exposed PS on the cell surface to a level that is comparable to drs2Δ [12]. Surprisingly, a screen of the entire yeast knockout collection for Pap B sensitivity identified several hundred mutants that displayed an increased sensitivity to Pap B relative to wild-type cells [64]. Moreover, acute inactivation of Drs2p or clathrin function by shifting cells bearing drs2-ts or chc1-ts alleles to the nonpermissive temperature failed to cause exposure of PS on the cell surface [12], suggesting that loss of PS asymmetry of the plasma membrane may be a secondary consequence of a chronic defect in protein transport. Because of a lack of specificity for this mutant phenotype, loss of plasma membrane phospholipid asymmetry in P4-ATPase mutants does not prove that these enzymes directly pump endogenous phospholipid across the membrane, nor can these phenotypes be used to infer substrate specificity for the P4-ATPases.

9.4â•… DRS2P IN PROTEIN TRANSPORT AND VESICLE BUDDING 9.4.1â•… Vesicle-Mediated Protein Transport Protein transport in the exocytic and endocytic pathways is primarily mediated by small transport vesicles [66]. For example, vesicles coated with coat protein complex II (COPII) bud from the ER and transport protein cargos to the Golgi, whereas COPI-coated vesicles mediate the retrograde transport from the Golgi back to the ER. Clathrin associates with pathway-specific adapter proteins (APs) and generates clathrin-coated vesicles (CCVs) from the TGN,

184

YEAST P4-ATPASES

endosomes, and the plasma membrane. Vesicle budding involves a sequence of critical steps, including recruitment of coat components and accessory proteins to the proper membrane site, coat assembly, membrane deformation, incorporation of cargo, and finally scission to release the vesicle. ARF is a small GTP-binding protein that plays a key role in vesicle biogenesis by directly binding and recruiting coat proteins onto membranes [67]. ARF-dependent coats include the COPI coatomer complex, clathrin/AP-1, clathrin/Golgi-localized, γ-ear-containing, ARF-binding protein (GGA), AP-3, and AP-4 coat complexes. ARF acts as a molecular switch by cycling between active GTP-bound and inactive GDP-bound forms. The GDP-bound form is largely soluble, whereas the GTP-bound form exposes its myristoylated N-terminus and associates with membranes. The cycling of ARF between its two nucleotide bound states requires catalytic assistance from two classes of proteins. The guanine nucleotide exchange factors (GEFs) promote exchange of GDP for GTP, while the GTPase-activating proteins (GAPs) stimulate the hydrolysis of GTP by ARF to convert the GTP-bound to the GDP-bound form [68]. In S. cerevisiae, ARF is encoded by an essential pair of genes, ARF1 and ARF2, which are 96% identical in protein sequence and redundant in function [69, 70]. In wild-type cells, Arf2p is only expressed at 10% of the level of Arf1p, and strains carrying arf2Δ show a wild-type phenotype [69]. In contrast, arf1Δ cells exhibit partial defects in protein secretion and Golgi-specific glycosylation [70], as well as altered morphology of Golgi and endosomes [71]. To identify other factors involved in vesicle biogenesis with ARF, a genetic screen for mutants that are synthetically lethal with arf1Δ (the swa mutants) was performed [72]. Seven complementation groups were isolated and among them were DRS2 (SWA3) and CDC50 (SWA4) [12, 50]. This screen also yielded mutant alleles of the clathrin heavy chain gene (CHC1/SWA5) and yeast auxilin (SWA2), a protein required for uncoating CCVs [72, 73]. 9.4.2â•… Roles for Drs2p-Cdc50p in Protein Transport The genetic interaction between DRS2 and ARF1 first implied a role for Drs2p in vesicle-mediated protein transport (Fig. 9.3). Drs2p primarily localizes to the TGN, where CCVs are actively formed. Consistently, drs2Δ is also synthetically lethal with a temperature-sensitive allele of clathrin, and exhibits several phenotypes in common with clathrin mutants, such as accumulation of swollen Golgi cisternae. This phenotype normally indicates a defect in vesicle biogenesis from the Golgi. Consistent with this possibility, disruption of DRS2 resulted in a marked reduction of CCVs that can be isolated from these cells [50]. Another indication that Drs2p facilitates vesicle budding came from epistasis analyses where drs2 alleles were combined with mutations that cause an accumulation of transport vesicles [59]. For example, disrupting the actin cytoskeleton (with the sla2Δ mutation or the drug latrunculin A [LatA]) causes the accumulation of dense, post-Golgi vesicles carrying the exocytic protein

185

Drs2p in Protein Transport and Vesicle Budding TGN

PM Pma1p (light) Invertase (dense)

???/clathrin (+Drs2p) Golgi Complex cis medial trans TGN ER

trans

Vacuole

Late Endosome

TGN

Vacuole

ALP

AP-3/clathrin (+Drs2p/Dnf1p)

TGN

EE Chs3p

Drs2p Chs3p Drs2p ?

Early Endosome

Drs2p

AP-1/clathrin (+Drs2p)

TGN

Rcy1p

LE

CPY

GGA/clathrin (+Drs2p/Dnf1p)

Figure 9.3.â•‡ Model for protein transport pathways requiring Drs2p. Drs2p controls protein trafficking between the TGN and the plasma membrane (PM) (1), early endosome (EE, 2 and 3), late endosome (LE, 4), and vacuole (5). See text for details. ALP, alkaline phosphatase. Color version on the Wiley web site.

invertase. Both clathrin and Drs2p are required for the formation of these vesicles as sla2Δ cells harboring temperature-sensitive alleles of DRS2 or clathrin rapidly lose these vesicles upon shift to the nonpermissive temperature (Fig. 9.3, pathway 1). Another similarity to clathrin mutants is that drs2Δ mislocalizes the TGN resident protein Kex2p [34, 50], which normally cycles between the TGN and endosomal compartments to maintain its steady-state TGN localization [74]. Therefore, Drs2p seems to be involved in the clathrin-dependent protein trafficking between the TGN and early endosomes, a pathway associated with AP-1 and clathrin. Indeed, Drs2p could be co-immunoprecipitated with AP-1 and genetic data argued that Drs2p is essential for AP-1 function in yeast [52]. Drs2p itself appears to be a cargo of AP-1/CCVs that bud from the TGN and are presumably targeted to the early endosome (Fig. 9.3, pathway 2). This is suggested by the observation that deletion of AP-1 subunits causes rerouting of Drs2p to the plasma membrane, where it can be trapped behind an endocytosis block. When the endocytosis block is lifted, Drs2p returns back to the

186

YEAST P4-ATPASES

TGN in the absence of AP-1. Thus, Drs2p likely uses an AP-1/clathrin pathway to move from the TGN to the early endosome, but does not use AP-1 for retrieval from the early endosome back to the TGN. Strikingly, inactivation of Drs2p activity (using the temperature-sensitive form of Drs2p) also causes rerouting of Drs2p to the plasma membrane [52]. These data strongly suggest that Drs2p’s activity drives the formation of AP-1/ CCVs from the TGN, and that Drs2p is a cargo of these vesicles. As mentioned above, the most likely destination for these AP-1 vesicles is the early endosome; however, it is also possible that these vesicles carry Drs2p back to earlier Golgi cisternae. Drs2p is also required for the AP-1-dependent retrograde transport of a cargo protein, Chs3p, from the early endosome to the TGN [52]. In the absence of AP-1 or Drs2p, Chs3p is rerouted into the late endosome (Fig. 9.3). Thus, Drs2p is needed at both the TGN and early endosome to support AP-1/clathrin function. Drs2p and Cdc50p are also required for another early endosome to TGN pathway traveled by Snc1p, an exocytic v-SNARE [4, 75]. Snc1p cycles in a TGNâ•¯→â•¯plasma membraneâ•¯→â•¯early endosomeâ•¯→â•¯TGN loop, but maintains a steady-state localization at the plasma membrane of wild-type cells [76]. In contrast, drs2Δ or cdc50Δ cells accumulate Snc1p in punctate, cytoplasmic organelles, shown to be early endosomes in cdc50Δ [4, 13]. The recycling of Snc1p also requires Rcy1p, an F-box protein involved in the early endosome to TGN transport [77]. Cdc50p accumulates in the early endosome of rcy1Δ cells, suggesting that Drs2p also requires Rcy1p for retrieval back to the TGN (Fig. 9.3). Furthermore, Drs2p-Cdc50p physically associates with Rcy1p, suggesting a direct role for Drs2p in budding the “Rcy1p” vesicles [13]. Mutations in a subset of COPI subunits also perturb Snc1p recycling [78], and so it is possible that Drs2p and Rcy1p facilitate budding of COPI vesicles from the early endosome. How Rcy1p contributes to protein transport is unknown. The Rab11 homologs Ypt31p and Ypt32p and the ARF-GAP Gcs1p are also implicated in the Snc1p recycling pathway and display genetic interactions with Drs2p and Cdc50p [13, 75]. 9.4.3â•… Influence of Other P4-ATPases on Protein Transport Importantly, other members of P4-ATPase family are also involved in various protein trafficking pathways. Drs2p and Dnf ATPases constitute an essential group in yeast with functional redundancy, which also occurs at the level of protein transport [4]. For example, Drs2p and Dnf1p show redundant functions for alkaline phosphatase transport from the TGN to the vacuole in a pathway requiring AP-3, but not clathrin. Carboxypeptidase Y (CPY) transport from the TGN to the late endosome, a pathway that appears to use GGA adaptors and clathrin, is kinetically delayed in drs2Δdnf1Δ cells [4] (Fig. 9.3). A similar kinetic delay in CPY transport was observed when drs2Δ single mutant was shifted to a low, nonpermissive growth temperature [50]. Disruption of DNF1 and DNF2 caused a cold-sensitive defect in the internalization step

Drs2p in Protein Transport and Vesicle Budding

187

of endocytosis, which is exacerbated by additional disruption of DRS2, indicating these three proteins may contribute to endocytosis [33]. Moreover, the dnf1,2Δ mutant accumulates Snc1p in internal membranes but has no affect on the trafficking of Ste2p [4], the α-factor receptor that travels the endocytic pathway to the vacuole for degradation, suggesting that Dnf1p and Dnf2p play overlapping roles with Drs2p in the early endosome to the TGN transport pathway as well. The essential Neo1p has also been implicated in vesicle-mediated protein transport [6, 7, 79]. The neo1-ts mutants show several defects at nonpermissive temperatures in common with COPI mutants [6], including the cargo-specific defects in secretion, aberrant glycosylation of cargos in the Golgi and the mislocalization of Rer1p, a protein that normally cycles between the ER and the Golgi, to vacuole. These observations suggest that Neo1p is required for a COPI-dependent retrograde transport pathway from the Golgi to the ER, a pathway essential for yeast viability. On the other hand, neo1-ts mutants were also reported to exhibit fragmented vacuoles and defects in endocytosis [7], accumulation of adaptor protein Gga2p on aberrant membranes [79], and delayed CPY transport [7]. Deletion of GGA2 or ARL1, which encodes an ARF-like protein that functions within the endosomal/Golgi system, suppresses the temperature-sensitive phenotype of neo1-ts mutants [7, 79]. In addition, Neo1p shows both genetic and physical interactions with Ysl2p [7, 79], a potential GEF for Arl1p. Thus, Neo1p appears also to play a role in protein trafficking within the endosomal/Golgi system. 9.4.4â•… Endocytosis of Drs2p Drs2p maintains a steady-state localization to the TGN and cannot be detected on the plasma membrane of wild-type cells [4, 11, 50–52]. However, in several endocytosis-defective mutants such as vrp1, end3, and end4, a substantial amount of Drs2p accumulates on the plasma membrane concomitant with a depletion of the TGN pool of Drs2p [11, 51]. This observation suggests that Drs2p transits the plasma membrane as part of its normal trafficking itinerary. Consistently, Drs2p is also found in post-Golgi secretory vesicles that are targeted to the plasma membrane [35]. To determine the kinetics of Drs2p transport to the plasma membrane, cells expressing GFP-Drs2p were monitored over time after acutely blocking endocytosis by treatment with LatA [51], an inhibitor of actin assembly. GFP-Drs2p was found to accumulate very slowly on the plasma membrane over the course of ∼3 hours. Based on quantitative Western blots, we estimate that wild-type yeast cells have ∼4000 Drs2p molecules per cell distributed among five to six TGN cisternae. If the redistribution of Drs2p to the plasma membrane occurs linearly over the 3-hour incubation with LatA, then only 20 Drs2p molecules per minute (∼0.5% of the total) are transported from the TGN to the plasma membrane. Upon washout of LatA, Drs2p rapidly returns to the TGN, indicating that Drs2p must contain functional endocytosis signals [52].

188

YEAST P4-ATPASES

In yeast, two types of endocytosis signals have been characterized that sort membrane proteins into a clathrin/actin-based endocytic pathway for internalization from the plasma membrane: sequences that mediate phosphorylation and ubiquitination of the cargo [80], such as PEST-like sequences [81], and the NPFXD motif [82]. The NPFXD signal is recognized by the Sla1p subunit of an endocytic coat complex consisting of clathrin, Pan1p, End3p, Sla2p/End4p, and Sla1p [82–84]. Pan1p is a member of the Eps15 family of modular scaffolding proteins and interacts with the clathrin-binding proteins AP180 and epsin, as well as the Arp2/3 complex [85–87]. Therefore, Pan1p may be capable of linking adaptor-bound cargo proteins to clathrin and actin assembly. Pan1p, End3p, and actin assembly are required for both ubiquitin-dependent and NPFXD-dependent endocytosis, while Sla1p is only required for endocytosis of cargos bearing the NPFXD signal [82, 88]. Drs2p contains two NPFXD motifs in its carboxyl terminal cytosolic tail (C-tail) [51]. These two NPFXD motifs interact with the Sla1p homology domain 1 in Sla1p. Surprisingly, disruption of Sla1p or the NPFXD motifs did not cause accumulation of Drs2p on the plasma membrane, suggesting the presence of other endocytic signal(s) in Drs2p. Sequence analysis of Drs2p revealed three PEST-like sequences on its N-terminus, as well as several other weak PEST-like sequences throughout Drs2p. Deletion of all three N-terminal PEST-like sequences from Drs2p did not perturb its function, but caused accumulation of Drs2p on the plasma membrane in sla1Δ cells. These data indicate that Drs2p has multiple endocytosis signals that are functionally redundant. Interestingly, Dnf1p possesses an NPFXD motif near its N-terminus, which contributes significantly to its Sla1p-dependent endocytosis [51]. However, the closely related Dnf2p does not contain such a motif. In addition, Dnf1p, but not Dnf2p, shows redundant protein transport function with Drs2p at the TGN (e.g., in the AP-3 pathway) [4], suggesting that the NPFXD/Sla1p-dependent endocytosis of Dnf1p may be important for its function at the Golgi. Supportively, when the NPFXD motif of Dnf1p was mutated to the sequence found in Dnf2p, the mutant Dnf1p was unable to support protein transport in the AP-3 pathway [51]. Similarly, mutation of Drs2p NPFXD motifs is lethal in pan1-20 cells. These data strongly imply that P4-ATPases cannot exert their essential function from the plasma membrane. 9.4.5â•… The Drs2p C-Tail The C-tail of Drs2p is essential to Drs2p function as a C-tail truncation allele of DRS2 cannot complement the cold-sensitive growth and protein transport defects of drs2Δ [51]. The precise function of the C-tail is not known, although it may serve a regulatory function analogous to the C-terminal tail of the plasma membrane Ca2+-ATPase [89] or the yeast H+-ATPase [90]. Three functionally important motifs have been mapped within the Drs2p C-tail so far: the NPFXDs, a Gea2p interaction motif (GIM) and a highly conserved motif

Drs2p in Protein Transport and Vesicle Budding

F AR F ARF GE

TGN

189

F AR F GE AR F

Figure 9.4.â•‡ Proposed model for how the flippase activity of Drs2p could drive budding of AP-1/clathrin-coated vesicles. Interaction of Drs2p (green) with ArfGEF helps concentrate Arf, AP-1 (blue), and clathrin triskelia (black) at sites of phospholipid translocation. Flippase activity induces membrane curvature that is stabilized and localized by assembly of the clathrin lattice. Drs2p is not required for recruitment of ArfGEF, AP-1, or clathrin to the TGN, but these coat components appear to be incapable of forming vesicles without Drs2p. Color version on the Wiley web site.

[91]. The GIM directly interacts with the Sec7 (GEF) domain of the ARF-GEF Gea2p (Fig. 9.4). Mutations in the GIM only partially perturb Drs2p function in vivo. In addition, Golgi localization of Gea2p is not perturbed in drs2Δ cells, so Drs2p is not required to recruit Gea2p to the Golgi. Adjacent to the GIM is a highly conserved motif found in most close homologs of Drs2p, including mammalian Atp8a1, and deletion of this motif also partially disrupts Drs2p function in vivo. However, a deletion that impinges on both the GIM and the conserved motif eliminates Drs2p function. Thus, the essential function of the C-tail maps to a small region overlapping the GIM and the conserved motif [91]. A recent report identified sequence identity between the conserved motif and a phosphoinositide-binding split pleckstrin homology domain of Vps36p, and went on to show that basic residues in this motif preferentially bind phosphatidylinositol-4-phosphate (PI(4)P) [53]. Moreover, Drs2p flippase activity in purified TGN membranes was stimulated by PI(4)P synthesized by the Pik1p PI 4-kinase. Surprisingly, ArfGEF binding to Drs2p also stimulated

190

YEAST P4-ATPASES

flippase activity and appears to synergize with PI(4)P for this effect. As predicted, mutations that disrupt binding of both PI(4)P and ArfGEF eliminates Drs2p activity in vitro and in vivo, while mutations that only disrupt one of these interactions partially abrogate Drs2p function [53]. These studies imply that the C-tail is an autoinhibitory domain and that binding to PI(4)P and ArfGEF at the TGN relieves the autoinhibition to stimulate Drs2p flippase activity. 9.4.6â•… Flippases and Vesicle Formation It has been proposed for many years that flippases may play a critical role in vesicle biogenesis and protein transport (Fig. 9.4) [17, 18, 92]. Translocation of phospholipids by flippases from the extracellular/luminal leaflet to the cytosolic leaflet of the membrane increases the surface area of the cytosolic leaflet at the expense of the extracellular leaflet. As suggested by the bilayer couple hypothesis [93], the imbalance of phospholipid number and surface area of the membrane bilayer would result in membrane deformation and induction of membrane bending toward the side with more phospholipid and a larger surface area (see Chapters 2 and 13). For instance, incorporation of exogenous aminophospholipids (PS and PE) to the outer leaflet of the human RBCs caused an initial echinocytic morphological change to the cells [38, 56]. Subsequent translocation of the added PS or PE to the inner leaflet, catalyzed by the aminophospholipid translocase, drove the cells back to discocytes or even stomatocytes. Therefore, changes in RBC shape (membrane curvature) correlate well with the imbalance of phospholipid number across the membrane bilayer. Many observations support the concept that unbalanced changes in monolayer surface area can impinge on vesicular transport. For example, sphingomyelinase treatment of cells, which cleaves the head group from sphingomyelin, thereby reducing the surface area of the outer leaflet of the plasma membrane, drives ATP-independent formation of functional endocytic vesicles [94]. Similarly, incorporation of exogenous PS or PE in the outer leaflet of the plasma membrane and subsequent translocation to the inner leaflet significantly enhances endocytosis. In contrast, addition of lyso-PS, which cannot be translocated across the plasma membrane and remains in the outer leaflet, inhibits endocytosis [95]. These observations are in agreement with the role proposed for flippases in vesicle budding. However, the limitation of these approaches is that they make use of unusual perturbations to the cells. The studies on yeast P4-ATPases demonstrate that these enzymes are part of the normal machinery required for vesicle-mediated protein transport, providing the first line of evidence that cells use a bilayer-couple mechanism under normal physiological conditions to support vesicle biogenesis. This function appears to be conserved as plant, Caenorhabditis elegans, and mouse P4ATPases have also been implicated in vesicle-mediated protein transport [96–98].

Concluding Remarks

191

Generation of positive membrane curvature required for vesicle formation is a role that has traditionally been assigned to coat proteins such as clathrin [99, 100]. Structural studies have revealed an intrinsic curvature in the clathrin triskelion [101, 102], and clathrin is capable of self-assembly into polyhedral baskets in the absence of lipids [103]. In an in vitro system, clathrin-coated buds can form from protein-free liposomes with a minimal requirement of clathrin and APs [104]. However, recent theoretical studies estimated that the rigidity of clathrin-AP complex is of the same order of magnitude as the resistance of lipid membranes to bending [105], suggesting that the clathrin coat assembly is unlikely to provide sufficient energy to drive membrane deformation in vivo. Instead, it may serve to stabilize an already curved membrane and prevent the membrane from collapsing back into a planar form. In the “Brownian ratchet” model [106], membranes spontaneously fluctuate, resulting in transient membrane invaginations, which could be captured by the clathrin lattice to form vesicles. Formation of the positive curvature may also be driven directly by accessory proteins such as the Bin/Amphiphysin/Rvs (BAR)-domain-containing proteins [107, 108] and epsins [109] with the capability of deforming membranes. Flippases may facilitate the vesicle formation by several means. First, flippases may help generate the positive curvature and prime the membrane deformation by pumping phospholipids across the membrane bilayer using the energy from ATP hydrolysis [18]. Second, flippases (e.g., Drs2p) may interact directly with coat/accessory proteins (e.g., the AP-1 complex, the ARF-GEF Gea2p and Rcy1p) [13, 52, 91], and recruit and concentrate these proteins to vesicle budding sites. Interestingly, loss of Drs2p perturbs AP-1/clathrin function and CCV formation but not the association of clathrin/AP-1 to the membrane [50, 52, 59], suggesting that the recruitment role of Drs2p is less important. Rather, the ability of Drs2p to generate and/or stabilize positive curvature appears to be more critical to the process of vesicle formation. Another influence that flippases may have on vesicle biogenesis is that, in the process of translocating specific phospholipids, flippases could concentrate certain lipid species and generate a unique local membrane environment (e.g., PE-rich) that is necessary and favorable to vesicle budding. The reconstituted system with purified flippases and composition-defined liposomes should be useful to assess these hypotheses. 9.5â•… CONCLUDING REMARKS Research on Drs2p and other yeast P4-ATPases has supported their proposed flippase function, and greatly expanded our understanding of how membrane asymmetry is generated. From its position in the TGN and the early endosomes, Drs2p controls the trafficking of proteins between the Golgi, plasma membrane, and endocytic pathway, thereby influencing the protein comÂ� position of the plasma membrane as well as the lipid organization in this

192

YEAST P4-ATPASES

membrane. Furthermore, the requirement of Drs2p in protein trafficking pathways is tightly correlated with the proposed phospholipid flippase activity. Conditional mutations in Drs2p inactivate both flippase activity in vitro and vesicle budding in vivo. Thus, Drs2p appears to prime the membrane in the TGN-endosomal system for delivery to the cell surface by imparting curvature to the membrane to facilitate protein sorting into different vesicle-mediated transport pathways, and by establishing an asymmetric concentration of PS and PE in the cytosolic leaflet. However, many uncertainties remain. For example, what role does Cdc50p play in phospholipid translocation? How do these proteins recognize the phospholipid substrate? Do the bulky phospholipid molecules follow a path through Drs2p-Cdc50p analogous to the path Ca2+ moves through the Ca2+ATPase? If not, how are substrate phospholipids flipped across the bilayer? Is PI(4)P and ArfGEF sufficient to regulate Drs2p-Cdc50p activity in proteoliposomes, or do Golgi membranes contain other activators of flippase activity? Does Drs2p-Cdc50p induce curvature in the membrane to facilitate vesicle formation or is it coupled to vesicle biogenesis by another mechanism? Answers to these questions will require a combination of genetic, cell biological, and biochemical approaches. Reconstitution of purified Drs2p or Drs2p-Cdc50p in artificial membranes will provide an opportunity to determine if Drs2p is sufficient to drive phospholipid translocation and test the requirement for Cdc50p. The reconstituted system can also be used to assess the consequences of phospholipid translocation on membrane shape and potentially establish a vesicle budding assay from the proteoliposomes. The last decade has brought a dramatic increase in our understanding of P4-ATPase cellular functions, and the next decade holds great promise for further elucidating the mechanistic basis for these functions.

ABBREVIATIONS AP ARF CCV COP CPY ER GAP GEF GGA GIM Pap B PC

adapter protein ADP-ribosylation factor clathrin-coated vesicle coat protein complex carboxypeptidase Y endoplasmic reticulum GTPase-activating protein guanine nucleotide exchange factor Golgi-localized, γ-ear-containing, ARF-binding protein Gea2p interaction motif papuamide B phosphatidylcholine

References

PE PI(4)P PS RBC Ro SERCA1 TAP TGN TNBS

193

phosphatidylethanolamine phosphatidylinositol-4-phosphate phosphatidylserine red blood cell Ro09-0198 sarcoplasmic/endoplasmic reticulum Ca2+-ATPase 1 tandem affinity purification trans-Golgi network trinitrobenzene sulfonic acid

REFERENCES â•… â•… â•… â•… â•…

1â•… 2â•… 3â•… 4â•… 5â•…

â•… 6â•… â•… 7â•… â•… 8â•… â•… 9â•… â•‡ 10â•… â•‡ 11â•… â•‡ 12â•… â•‡ 13â•… â•‡ 14â•… â•‡ 15â•… â•‡ 16â•… â•‡ 17â•… â•‡ 18â•… â•‡ 19â•… â•‡ 20â•… â•‡ 21â•… â•‡ 22â•… â•‡ 23â•…

P. Catty, A. de Kerchove d’Exaerde, A. Goffeau, FEBS Lett. 1997, 409, 325–332. K. B. Axelsen, M. G. Palmgren, J. Mol. Evol. 1998, 46, 84–101. W. Kuhlbrandt, Nat. Rev. Mol. Cell Biol. 2004, 5, 282–295. Z. Hua, P. Fatheddin, T. R. Graham, Mol. Biol. Cell. 2002, 13, 3162–3177. T. R. Prezant, W. E. Chaltraw, Jr., N. Fischel-Ghodsian, Microbiology 1996, 142 (Pt 12), 3407–3414. Z. Hua, T. R. Graham, Mol. Biol. Cell 2003, 14, 4971–4983. S. Wicky, H. Schwarz, B. Singer-Kruger, Mol. Cell. Biol. 2004, 24, 7402–7418. T. L. Ripmaster, G. P. Vaughn, J. L. Woolford, Jr., Mol. Cell. Biol. 1993, 13, 7901–7912. C. Toyoshima, M. Nakasako, H. Nomura, H. Ogawa, Nature 2000, 405, 647–655. C. Toyoshima, Arch. Biochem. Biophys. 2008, 476, 3–11. K. Saito, K. Fujimura-Kamada, N. Furuta, U. Kato, et al., Mol. Biol. Cell 2004, 15, 3418–3432. S. Chen, J. Wang, B. P. Muthusamy, K. Liu, et al., Traffic 2006, 7, 1503–1517. N. Furuta, K. Fujimura-Kamada, K. Saito, T. Yamamoto, et al., Mol. Biol. Cell 2007, 18, 295–312. K. Geering, J. Bioenerg. Biomembr. 2001, 33, 425–438. P. F. Devaux, Annu. Rev. Biophys. Biomol. Struct. 1992, 21, 417–439. A. K. Menon, Trends Cell Biol. 1995, 5, 355–360. T. Pomorski, J. C. Holthuis, A. Herrmann, G. van Meer, J. Cell Sci. 2004, 117, 805–813. T. R. Graham, Trends Cell Biol. 2004, 14, 670–677. W. R. Bishop, R. M. Bell, Cell 1985, 42, 51–60. S. Sanyal, C. G. Frank, A. K. Menon, Biochemistry 2008, 47, 7937–7946. E. M. Bevers, P. L. Williamson, FEBS Lett. 2010, 584, 2724–2730. Y. C. Wu, H. R. Horvitz, Cell 1998, 93, 951–960. Y. Hamon, C. Broccardo, O. Chambenoit, M. F. Luciani, et al., Nat. Cell Biol. 2000, 2, 399–406.

194

YEAST P4-ATPASES

â•‡ 24â•… V. Venegas, Z. Zhou, Mol. Biol. Cell 2007, 18, 3180–3192. â•‡ 25â•… X. Wang, J. Wang, K. Gengyo-Ando, L. Gu, et al., Nat. Cell Biol. 2007, 9, 541–549. â•‡ 26â•… S. Zullig, L. J. Neukomm, M. Jovanovic, S. J. Charette, et al., Curr. Biol. 2007, 17, 994–999. â•‡ 27â•… S. Ruetz, P. Gros, Cell 1994, 77, 1071–1081. â•‡ 28â•… R. J. Raggers, A. van Helvoort, R. Evers, G. van Meer, J. Cell Sci. 1999, 112(Pt 3), 415–422. â•‡ 29â•… A. van Helvoort, A. J. Smith, H. Sprong, I. Fritzsche, et al., Cell 1996, 87, 507–517. â•‡ 30â•… Y. Romsicki, F. J. Sharom, Biochemistry 2001, 40, 6937–6947. â•‡ 31â•… X. Tang, M. S. Halleck, R. A. Schlegel, P. Williamson, Science 1996, 272, 1495–1497. â•‡ 32â•… E. Gomes, M. K. Jakobsen, K. B. Axelsen, M. Geisler, et al., Plant Cell 2000, 12, 2441–2454. â•‡ 33â•… T. Pomorski, R. Lombardi, H. Riezman, P. F. Devaux, et al., Mol. Biol. Cell 2003, 14, 1240–1254. â•‡ 34â•… P. Natarajan, J. Wang, Z. Hua, T. R. Graham, Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 10614–10619. â•‡ 35â•… N. Alder-Baerens, Q. Lisman, L. Luong, T. Pomorski, et al., Mol. Biol. Cell 2006, 17, 1632–1642. â•‡ 36â•… J. A. Coleman, M. C. Kwok, R. S. Molday, J. Biol. Chem. 2009, 284, 32670–32679. â•‡ 37â•… X. Zhou, T. R. Graham, Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 16586–16591. â•‡ 38â•… M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. â•‡ 39â•… P. F. Devaux, I. Lopez-Montero, S. Bryde, Chem. Phys. Lipids 2006, 141, 119–132. â•‡ 40â•… Y. Moriyama, N. Nelson, J. Biol. Chem. 1988, 263, 8521–8527. â•‡ 41â•… A. Zachowski, J. P. Henry, P. F. Devaux, Nature 1989, 340, 75–76. â•‡ 42â•… J. Ding, Z. Wu, B. P. Crider, Y. Ma, et al., J. Biol. Chem. 2000, 275, 23378–23386. â•‡ 43â•… J. K. Paterson, K. Renkema, L. Burden, M. S. Halleck, et al., Biochemistry 2006, 45, 5367–5376. â•‡ 44â•… L. S. Kean, A. M. Grant, C. Angeletti, Y. Mahe, et al., J. Cell Biol. 1997, 138, 255–270. â•‡ 45â•… A. M. Grant, P. K. Hanson, L. Malone, J. W. Nichols, Traffic 2001, 2, 37–50. â•‡ 46â•… H. C. Stevens, L. Malone, J. W. Nichols, J. Biol. Chem. 2008, 283, 35060–35069. â•‡ 47â•… U. Kato, K. Emoto, C. Fredriksson, H. Nakamura, et al., J. Biol. Chem. 2002, 277, 37855–37862. â•‡ 48â•… A. Siegmund, A. Grant, C. Angeletti, L. Malone, et al., J. Biol. Chem. 1998, 273, 34399–34405. â•‡ 49â•… U. Marx, T. Polakowski, T. Pomorski, C. Lang, et al., Eur. J. Biochem. 1999, 263, 254–263. â•‡ 50â•… C. Y. Chen, M. F. Ingram, P. H. Rosal, T. R. Graham, J. Cell Biol. 1999, 147, 1223–1236. â•‡ 51â•… K. Liu, Z. Hua, J. A. Nepute, T. R. Graham, Mol. Biol. Cell 2007, 18, 487–500.

References

195

â•‡ 52â•… K. Liu, K. Surendhran, S. F. Nothwehr, T. R. Graham, Mol. Biol. Cell 2008, 19, 3526–3535. â•‡ 53â•… P. Natarajan, K. Liu, D. V. Patil, V. A. Sciorra, et al., Nat. Cell Biol. 2009, 11, 1421–1426. â•‡ 54â•… P. K. Hanson, L. Malone, J. L. Birchmore, J. W. Nichols, J. Biol. Chem. 2003, 278, 36041–36050. â•‡ 55â•… W. R. Riekhof, D. R. Voelker, J. Biol. Chem. 2006, 281, 36588–36596. â•‡ 56â•… D. L. Daleke, W. H. Huestis, Biochemistry 1985, 24, 5406–5416. â•‡ 57â•… G. Morrot, P. Herve, A. Zachowski, P. Fellmann, et al., Biochemistry 1989, 28, 3456–3462. â•‡ 58â•… F. J. Perez-Victoria, M. P. Sanchez-Canete, S. Castanys, F. Gamarro, J. Biol. Chem. 2006, 281, 23766–23775. â•‡ 59â•… W. E. Gall, N. C. Geething, Z. Hua, M. F. Ingram, et al., Curr. Biol. 2002, 12, 1623–1627. â•‡ 60â•… P. K. Hanson, J. W. Nichols, J. Biol. Chem. 2001, 276, 9861–9867. â•‡ 61â•… H. C. Stevens, J. W. Nichols, J. Biol. Chem. 2007, 282, 17563–17567. â•‡ 62â•… E. Zinser, C. D. Sperka-Gottlieb, E. V. Fasch, S. D. Kohlwein, et al., J. Bacteriol. 1991, 173, 2026–2034. â•‡ 63â•… P. Hechtberger, E. Zinser, R. Saf, K. Hummel, et al., Eur. J. Biochem. 1994, 225, 641–649. â•‡ 64â•… A. B. Parsons, A. Lopez, I. E. Givoni, D. E. Williams, et al., Cell 2006, 126, 611–625. â•‡ 65â•… Y. Aoki, T. Uenaka, J. Aoki, M. Umeda, et al., J. Biochem. 1994, 116, 291–297. â•‡ 66â•… T. Kirchhausen, Nat. Rev. Mol. Cell Biol. 2000, 1, 187–198. â•‡ 67â•… C. D’Souza-Schorey, P. Chavrier, Nat. Rev. Mol. Cell Biol. 2006, 7, 347–358. â•‡ 68â•… J. G. Donaldson, C. L. Jackson, Curr. Opin. Cell Biol. 2000, 12, 475–482. â•‡ 69â•… T. Stearns, R. A. Kahn, D. Botstein, M. A. Hoyt, Mol. Cell. Biol. 1990, 10, 6690–6699. â•‡ 70â•… T. Stearns, M. C. Willingham, D. Botstein, R. A. Kahn, Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 1238–1242. â•‡ 71â•… E. C. Gaynor, C. Y. Chen, S. D. Emr, T. R. Graham, Mol. Biol. Cell 1998, 9, 653–670. â•‡ 72â•… C. Y. Chen, T. R. Graham, Genetics 1998, 150, 577–589. â•‡ 73â•… W. E. Gall, M. A. Higginbotham, C. Chen, M. F. Ingram, et al., Curr. Biol. 2000, 10, 1349–1358. â•‡ 74â•… N. J. Bryant, T. H. Stevens, J. Cell Biol. 1997, 136, 287–297. â•‡ 75â•… H. Sakane, T. Yamamoto, K. Tanaka, Cell Struct. Funct. 2006, 31, 87–108. â•‡ 76â•… M. J. Lewis, B. J. Nichols, C. Prescianotto-Baschong, H. Riezman, et al., Mol. Biol. Cell 2000, 11, 23–38. â•‡ 77â•… A. Wiederkehr, S. Avaro, C. Prescianotto-Baschong, R. Haguenauer-Tsapis, et al., J. Cell Biol. 2000, 149, 397–410. â•‡ 78â•… M. Robinson, P. P. Poon, C. Schindler, L. E. Murray, et al., Mol. Biol. Cell 2006, 17, 1845–1858.

196

YEAST P4-ATPASES

â•‡ 79â•… B. Singer-Kruger, M. Lasic, A. M. Burger, A. Hausser, et al., EMBO J. 2008, 27, 1423–1435. â•‡ 80â•… L. Hicke, R. Dunn, Annu. Rev. Cell Dev. Biol. 2003, 19, 141–172. â•‡ 81â•… A. F. Roth, D. M. Sullivan, N. G. Davis, J. Cell Biol. 1998, 142, 949–961. â•‡ 82â•… J. P. Howard, J. L. Hutton, J. M. Olson, G. S. Payne, J. Cell Biol. 2002, 157, 315–326. â•‡ 83â•… T. M. Newpher, R. P. Smith, V. Lemmon, S. K. Lemmon, Dev. Cell 2005, 9, 87–98. â•‡ 84â•… M. Kaksonen, C. P. Toret, D. G. Drubin, Nat. Rev. Mol. Cell Biol. 2006, 7, 404–414. â•‡ 85â•… B. Wendland, S. D. Emr, J. Cell Biol. 1998, 141, 71–84. â•‡ 86â•… M. C. Duncan, M. J. Cope, B. L. Goode, B. Wendland, et al., Nat. Cell Biol. 2001, 3, 687–690. â•‡ 87â•… R. C. Aguilar, H. A. Watson, B. Wendland, J. Biol. Chem. 2003, 278, 10737–10743. â•‡ 88â•… N. B. Miliaras, J. H. Park, B. Wendland, Traffic 2004, 5, 963–978. â•‡ 89â•… A. Enyedi, T. Vorherr, P. James, D. J. McCormick, et al., J. Biol. Chem. 1989, 264, 12313–12321. â•‡ 90â•… W. Kuhlbrandt, J. Zeelen, J. Dietrich, Science 2002, 297, 1692–1696. â•‡ 91â•… S. Chantalat, S. K. Park, Z. Hua, K. Liu, et al., J. Cell Sci. 2004, 117, 711–722. â•‡ 92â•… P. F. Devaux, Biochemistry 1991, 30, 1163–1173. â•‡ 93â•… M. P. Sheetz, S. J. Singer, Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 4457–4461. â•‡ 94â•… X. Zha, L. M. Pierini, P. L. Leopold, P. J. Skiba, et al., J. Cell Biol. 1998, 140, 39–47. â•‡ 95â•… E. Farge, D. M. Ojcius, A. Subtil, A. Dautry-Varsat, Am. J. Physiol. 1999, 276, C725–C733. â•‡ 96â•… L. R. Poulsen, R. L. Lopez-Marques, S. C. McDowell, J. Okkeri, et al., Plant Cell 2008, 20, 658–676. â•‡ 97â•… P. Xu, J. Okkeri, S. Hanisch, R. Y. Hu, et al., J. Cell Sci. 2009, 122, 2866–2876. â•‡ 98â•… A. F. Ruaud, L. Nilsson, F. Richard, M. K. Larsen, et al., Traffic 2009, 10, 88–100. â•‡ 99â•… S. L. Schmid, Annu. Rev. Biochem. 1997, 66, 511–548. 100â•… T. Kirchhausen, Annu. Rev. Biochem. 2000, 69, 699–727. 101â•… C. J. Smith, N. Grigorieff, B. M. Pearse, EMBO J. 1998, 17, 4943–4953. 102â•… A. Musacchio, C. J. Smith, A. M. Roseman, S. C. Harrison, et al., Mol. Cells 1999, 3, 761–770. 103â•… M. P. Woodward, T. F. Roth, Proc. Natl. Acad. Sci. U.S.A. 1978, 75, 4394–4398. 104â•… K. Takei, V. Haucke, V. Slepnev, K. Farsad, et al., Cell 1998, 94, 131–141. 105â•… R. Nossal, Traffic 2001, 2, 138–147. 106â•… L. Hinrichsen, A. Meyerholz, S. Groos, E. J. Ungewickell, Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 8715–8720. 107â•… T. Itoh, P. De Camilli, Biochim. Biophys. Acta 2006, 1761, 897–912. 108â•… G. Ren, P. Vajjhala, J. S. Lee, B. Winsor, et al., Microbiol. Mol. Biol. Rev. 2006, 70, 37–120.

References

109â•… 110â•… 111â•… 112â•…

197

M. G. Ford, I. G. Mills, B. J. Peter, Y. Vallis, et al., Nature 2002, 419, 361–366. J. Kopp, T. Schwede, Nucleic Acids Res. 2004, 32, D230–D234. N. Guex, M. C. Peitsch, Electrophoresis 1997, 18, 2714–2723. T. Schwede, J. Kopp, N. Guex, M. C. Peitsch, Nucleic Acids Res. 2003, 31, 3381–3385.

10 SUBSTRATE SPECIFICITY OF THE AMINOPHOSPHOLIPID FLIPPASE Shelley M. Cook and David L. Daleke Department of Biochemistry and Molecular Biology, Medical Sciences, Indiana University School of Medicine, Bloomington, IN

10.1â•… INTRODUCTION Biological membranes are comprised of four major classes of phospholipids. The choline-containing lipids, phosphatidylcholine (PC) and sphingomyelin (SM), reside primarily in the extracytoplasmic monolayer, while the aminophospholipids, phosphatidylserine (PS) and phosphatidylethanolamine (PE), are preferentially located in the cytoplasmic monolayer (review References 1 and 2). Loss of transmembrane lipid asymmetry, and the exposure of PS in the external monolayer, occurs in both normal and pathological conditions [3]. PS externalization is induced early in apoptotic cells [4] and during platelet activation [5]. This perturbation results in a change in cell surface properties, including conversion to a procoagulant state [6], increased adhesion [7], increased aggregation [8], and recognition by phagocytic cells [9]. While these processes are essential for normal cell development and hemostasis, unregulated loss of PS asymmetry has been associated with diseases that have high cardiovascular risk, such as diabetes [8, 10, 11], and may contribute significantly to heart disease and stroke. Transmembrane PS asymmetry is maintained by a combination of seveÂ� ral mechanisms, including slow non-protein-mediated flip-flop, binding to

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

199

200

FLIPPASE SUBSTRATE SPECIFICITY

cytofacial proteins, and protein-mediated transport. However, inwardly directed active transport of PS is likely the primary mechanism for the generation and maintenance of PS asymmetry at the plasma membrane (PM). A PS translocase, or “flippase,” transports PS and, to a lesser extent, PE from the extracytoplasmic to the cytofacial monolayer of the membrane. Flippase activity is vanadate, Ca2+, and N-ethylmaleimide (NEM) sensitive as well as temperature and adenosine triphosphate (ATP) dependent [12–14]. AminoÂ� phospholipid transport activity has been well characterized in human erythrocytes and has also been found in a wide variety of cell types and membranes, including mammalian and yeast PMs as well as some internal organelles. The flippase is one of a growing family of transbilayer lipid transporters, including proteins that catalyze the efflux (floppases) and bidirectional movement (scramblases) of lipids. The regulation of the trafficking of these transporters may serve to generate and maintain an increasingly asymmetric bilayer as newly synthesized membranes traffic from the endoplasmic reticulum (ER) to the Golgi to the PM (for recent reviews, see References 15–19). Although the identification of the protein or proteins involved has been challenging, much is known about the biochemical characteristics of lipid transport activity, including the specificity with which they recognize their substrates. The critical functions that these transporters play are reliant on the degree to which they recognize their lipid substrates. Perhaps the most information regarding substrate specificity has been obtained for the flippase class of lipid transporters, which selectively transport PS across the PM. 10.2â•… SUBSTRATE SPECIFICITY OF THE PM AMINOPHOSPHOLIPID FLIPPASE Transport by the PM aminophospholipid flippase is highly selective and recognizes the head group and glycerol moieties of the substrate. Using spin, fluorescent, and short fatty acyl chain bearing phospholipids, early studies demonstrated that PS, but not PC, was selectively transported across human erythrocyte membranes by an Mg2+-, ATP-, and sulfhydryl-reagent-sensitive process [13, 14, 20]. PE is also a transport substrate and can compete with PS, but is transported at rates 10-fold less than PS [21]. These initial data defined the flippase as a primary amine-containing phospholipid-specific transporter that selectively recognizes and transports PS and PE. Subsequent studies addressed the role of each of the functional groups of the PS molecule in its recognition by the flippase, including the amine, carboxyl, phosphate, and glycerol groups, as well as the composition and mode of attachment of the fatty acyl or alkyl groups and the stereochemical specificity of the glycerol and serine groups (Fig. 10.1). Collectively these studies have revealed that the flippase is exquisitely sensitive to some, but not all, of these structural elements and that this constellation of preferences is unique among PS-binding proteins.

SUBSTRATE SPECIFICITY OF PM AMINOPHOSPHOLIPID FLIPPASE O HO

O OH

P

O–

R1

O

O O

O

H COO–

H3N–

OH

I

R1

O O

O R2

R1

–

P

O O

O

O O

O

O O

R2

P

O O

H COO–

R1

O R2

O

P

O O

H3N+

O O

O R2

O

O– O

H3N+

H COOCH3

VI R1 H COO–

O O

O R2

O

VII R1

P

O

O

–

O

R2

H COO–

IV

H2N+ CH3

O

O– O

–

O O

P

H3N+

V R1

H COO–

O

–

O

R2

H3N–

COO

O

O– O

II

III R1

P

201

P

O– O

H3N+

H COO–

VIII O– O

H3N+ IX

P

R1 H COO–

O O

O R2

O

P

O– O

H3N+

H COO–

X

Figure 10.1.â•‡ Structures of phosphatidylserine analogs. Glycerophosphorylserine (I; GPS); lysophosphatidylserine (II; lyso-PS); phosphatidylhydroxypropionate (III; PP); phosphatidylhomoserine (IV; PhS); N-methylphosphatidylserine (V; N-methylPS); phosphatidylserine-methyl ester (VI; PS-Me); 1,2-sn-phosphatidyl-L-serine (VII; 1,2-sn-P-L-S); 1,2-sn-phosphatidyl-D-serine (VIII; 1,2-sn-P-D-S); 2,3-sn-phosphatidylL-serine (IX; 2,3-sn-P-L-S); 2,3-sn-phosphatidyl-D-serine (X; 2,3-sn-P-D-S). R1 and R2, fatty acyl group.

The distinguishing characteristic of optimal substrate recognition by the flippase is the presence of a primary amine group. The absence of the serine amine group substantially reduces the ability of the flippase to transport the substrate [22, 23], as does acylation of the amine group [24]. However, the transporter tolerates some modifications of the amine group; N-methyl-PS is transported at a rate equivalent to PS [22, 23]. In contrast, progressive methylation of PE reduces transport of the lipid [22, 23], so it may be likely that,

202

FLIPPASE SUBSTRATE SPECIFICITY

although it is also a flippase substrate, this lipid interacts with the protein in a different manner than PS. The phosphate diester and carboxyl groups have similar roles in flippase substrate recognition. The carboxyl group, although not essential, plays a critical role in the recognition of the substrate. The presence of this group substantially stimulates substrate transport and evidence indicates that it must be in the charged state; PS-methyl ester is a weak flippase substrate compared with PS [22, 23]. Regarding the phosphate diester group, an uncharged, isosteric, sulfonate analog of PS, dimyristoylsulfonoserine, was not capable of being transported across erythrocyte membranes (D. Daleke and T. Widlanski, unpublished observations). Thus, the flippase substrate head group likely makes important ionic interactions with its binding site on the protein, which may include carboxylate- and phosphate-interacting positive charges. The glycerol “backbone” of the PS molecule is a key recognition element for the flippase. Substituting ether linkages and alkyl substituents in place of the fatty acyl ester groups reduces transport considerably, especially in the sn-2 position [25], and a PS analog of SM retains only very minimal transport activity [23]. Some variations in the backbone structure are tolerated by the flippase; a butanetriol-based 7-nitro-2-1,3-benzoxadiazol-4-yl (NBD)-PS substrate has been demonstrated to be an effective flippase substrate [26]. These results add to the predications of the molecular characteristics of the putative flippase substrate-binding site and begin to define the important role the glycerol group plays in substrate-flippase interactions. As will be discussed below, the glycerol moiety may be positioned in a unique amphipathic binding site on the protein, to allow the transbilayer movement of both its hydrophobic and hydrophilic substituents. The PS molecule possesses two chiral carbons, one in the C2 position of the glycerol group and another at the α-carbon in the serine head group. These groups reflect the stereochemistry of their naturally occurring precursors. The glycerol group is synthesized primarily by stereospecific reduction of dihydroxyacetone-3-phosphate, which, after acylation and further modification, yields the 1,2-diacyl-sn-glycero-3-phospholipid as the naturally occurring stereoisomer. The PS head group is synthesized in mammalian cells by base exchange with either PC or PE and L-serine, catalyzed by PS synthase I and II [27], respectively, in the ER. The flippase selectively recognizes the stereochemistry of the C2 carbon in the glycerol group; only 1,2-sn-glycero-3-phosphatidylserine, but not 2,3-sn-glycero-3-phosphatidylserine, is maximally transported across fibroblast [28] or erythrocyte [22] PMs. In contrast, both the L-serine and the D-serine analogs of 1,2-sn-glycerolphosphatidylserine are functional flippase substrates [22, 29]. Interestingly, 2,3-sn-glycerophosphatidylL-serine is transported at a slightly faster rate than 2,3-sn-glycerophosphatidylD-serine [22], implying that although the glycerol moiety plays an essential role in stereochemical recognition by the flippase, the stereochemistry of the serine head group may also be recognized and can partially overcome the disadvantage of the 2,3-sn-glyceraol diastereomer.

SUBSTRATE SPECIFICITY OF PM AMINOPHOSPHOLIPID FLIPPASE

203

Subsequent studies have refined the substrate structural elements required for transport and have modified our view of the selectivity of this transporter. A reevaluation of the transport of head-group-modified analogs of PS and PS diastereomers found that, with the exception of PC, many of the lipids previously thought not to be substrates were instead slowly transported across the membrane of intact human erythrocytes by a process that was sensitive to NEM and vanadate [22]. Relative transport rates of these lipids fell into three categories: (1) rapid (t1/2â•¯∼â•¯10 minutes; PS, N-methyl-PS), (2) medium (t1/2â•¯∼â•¯2 hours; PE, 2,3-sn-P-L-S), and (3) slow (t1/2â•¯>â•¯3 hours; phosphatidylhydroxypropionate, phosphatidylhomoserine, PS-methyl ester, 2,3-sn-P-D-S). These results demonstrate that the criteria used to identify flippase-mediated transport, such as sulfhydryl reagent and vanadate sensitivity, also includes the PS analogs described above, with the notable exception of PC. The transport of SM has not been fully examined in this regard, but since it traverses the membrane even more slowly than PC [23], it is unlikely to be a substrate of an inwardly directed flippase. These more recent data have expanded the notions of the specificity of the PS flippase and may indicate that either multiple, previously unknown flippases exist or the PS flippase is capable of transporting a broader range of substrates than previously identified. Insight into the specificity of substrate–transporter interactions may also be gained through the analysis of competitive inhibitors. An early study by Martin and Pagano demonstrated that glycerophosphorylserine ([GPS] Fig. 10.1) and glycerophosphorylethanolamine, but not glycerophosphorylcholine, were capable of inhibiting NBD-PS uptake in cultured fibroblasts [28]. These data indicate that the hydrophobic component of the lipid is not an essential recognition element and are consistent with a binding site encompassing the serine, phosphate, and glycerol groups. Not surprisingly, because of its high water solubility, the concentration of GPS necessary to inhibit the uptake of fluorescent PS (1-acyl-2-(7-nitro-2-1,3-benzoxadiazol-4-yl)-aminocaproyl-PS; NBD-PS) was relatively high (10â•¯mM), thus limiting its general usefulness. Reasoning that substrate analogs of PS might be competitive inhibitors of the flippase, the ability of the PS analogs described above to inhibit NBD-PS uptake in human erythrocytes was tested. Some of the analogs (phosphatidylhydroxypropionate, phosphatidylhomoserine, N-methyl-PS, 1,2-sn-P-D-S) are capable of inhibiting PS uptake, while others are ineffective (Smriti and D. Daleke, unpublished observations). Further exploitation of these flippase inhibitors may lead to the development of useful tools for the study of the flippase as well as potential therapeutics. In contrast to the high selectivity for the hydrophilic components of the lipid substrate, transport activity is relatively insensitive to the composition of the acyl groups attached to the glycerol group. The flippase is capable of transporting short (C6-C12), medium (C14), and long (C18:1) chain-containing PS molecules, as well as PS with fatty acids containing spin and fluorescent (NBD) labels. The rates of transport of saturated PS (diC8PS–diC14PS) and dioleoylphosphatidylserine (DOPS) are similar [13, 30], but the rates of lipid

204

FLIPPASE SUBSTRATE SPECIFICITY

transport differ between these unlabeled lipids and the fluorescent or spinlabeled lipids with short-chain linker groups. However, increasing the length of the alkyl linker between, for example, the carboxyl group and the NBD label increases the rate of transport to rates that are comparable to those of unlabeled PS [31]. The nature of the effect of acyl chain length and hydrophobicity on the transport of labeled lipids is unclear, but since the labels tend to be relatively hydrophilic, they may be positioned closer to the membrane interface, resulting in configurational changes in the glycerol backbone that interfere with substrate–transporter interactions. Further evidence for this hypothesis is that lyso-PS is not a substrate for the flippase, and although it exhibits appreciable rates of spontaneous transbilayer flip-flop (t1/2â•¯∼â•¯3 hours in erythrocytes), its movement across the erythrocyte membrane is not sensitive to sulfhydryl reagents [22]. Acylation of the sn-2 hydroxyl with a group as small as acetate is sufficient to restore sulfhydryl-reagent-sensitive transport to lyso-PS [23]. These observations are consistent with the requirement that the glycerol moiety must be correctly positioned in the membrane to allow transport. Without a hydrophobic anchor, the free sn-2 hydroxyl group is likely positioned near the polar interface in a conformation that is unproductive for transport. Taken together, these data identify the essential characteristics of a potential PS flippase substrate binding site. Data from the effect of lipid head-group modification on lipid transport indicate that both the PS amine and carboxyl groups are critical for substrate binding, perhaps as either hydrogen or ionic bonding participants. The reduction in transport by modifications that eliminate the charge on either the amine or the carboxyl group suggests that ionic pairing with the flippase binding site is involved, although one cannot eliminate the possibility of an internal salt or metal counterion. Regardless of the structure of this binding site, even though it likely retains multiple points of binding to the substrate, the lack of stereospecificity for the serine head group suggests a binding site with significant conformational flexibility or one that possesses numerous alternative sites that could pair with the critical elements of the hydrophilic head group. A speculative model for the mechanism of PS transport, one consistent with the substrate specificity data and with a configuration that could overcome the thermodynamic barrier to phospholipid transport, might include at its core an arrangement of transbilayer helices with appropriately positioned hydrogen bonding or ionic groups that would accommodate the hydrophilic head group (Fig. 10.2). The flippase model must account for the high degree of specificity for the orientation of the glycerol backbone of PS, yet retain flexibility in the recognition of the phosphoserine head group. In order for fast transport to occur, the hydrophilic head group must be isolated from the hydrophobic interior of the lipid bilayer. A possible pathway for protein-mediated PS transport might involve placement of the head group into a channel formed by transmembrane helices. (Four are shown for illustration purposes, but a greater number is also possible.) The glycerol backbone of the lipid would traverse the space between two of the transmembrane helices, allowing the fatty acid

Specificity of Candidate Aminophospholipid Flippases

+ O

–

O O

P

+ O

O

+ O–

O –

O

205

NH3+

O O

Figure 10.2.â•‡ Speculative pathway for selective transport of 1,2-sn-phosphatidylserine by the flippase. Four transmembrane helices of a putative flippase are shown, although more may be possible. Hypothetical electrostatic interactions between the channel interior and PS head group are indicated with “+” and “–.”Specificity for the 1,2-snglycerol (marked with a dashed line) isomer of PS might be explained if transport is established through the binding of the hydrophilic head group within a channel with the acyl chains protruding into the bilayer interior. Accommodation of the glycerol backbone between two transmembrane helices may define a stereospecific interaction or gate. Conformational changes attendant to energy transduction (not shown) would expose the head group to the opposite surface of the membrane, completing the transport cycle. Color version on the Wiley web site.

hydrocarbon moieties to remain buried in the lipid bilayer. The interaction between the glycerol backbone and the two transmembrane helices that sandwich it to enclose the head group in a protective channel provides a possible explanation of the glycerol backbone stereospecificity exhibited by the PS flippase. Though speculative, this model may provide an explanation for the requirement for stereospecific interaction with the glycerol and the lack of specificity for fatty acyl chain length and composition. 10.3â•… IDENTIFICATION AND SUBSTRATE SPECIFICITY OF CANDIDATE AMINOPHOSPHOLIPID FLIPPASES The PS flippase is associated with both ATP-dependent PS translocase activity and PS-dependent ATPase activity. Since the measurement of transport

206

FLIPPASE SUBSTRATE SPECIFICITY

activity is technically complicated, and indeed has only been demonstrated for a small number of lipid transporters in purified systems, the biochemical properties of the ATPase activity of the flippase have been used for the identification of putative transporters, reasoning that the subset of ATPases that possess these characteristics may contain the PS flippase. Using this strategy, a PSstimulated, NEM-, vanadate-, and Ca2+-sensitive Mg2+-ATPase has been purified from human erythrocytes [32–34] and bovine adrenal chromaffin granules [35] as candidate flippases, using protocols developed previously for the purification of ATPases with similar characteristics from bovine brain clathrincoated vesicles [36] and bovine chromaffin granules [37]. In one study, the partially purified human erythrocyte Mg2+-ATPase was shown to be capable of aminophospholipid transport activity when reconstituted into proteoliposomes [34], indicating that the aminophospholipid flippase is likely associated with this vanadate-sensitive Mg2+-ATPase. The biochemical requirements of the aminophospholipid flippase are relatively broad and nonspecific and are of limited usefulness in defining potential flippases. Transport data collected from intact cells reveal that the flippase must be a sulfhydryl-containing Mg2+-ATPase, but these criteria only partially limit the likely candidates from a pool of potential transporters. However, the unique lipid substrate specificity, particularly the stereospecificity of the glycerol moiety, of the flippase can be used to substantially narrow the number of likely candidates. The substrate specificity of the enzymatic activity of aminophospholipid flippase candidates provides perhaps the best biochemical indication of their involvement in PS transport. 10.3.1â•… PS-Stimulated Erythrocyte Mg2+-ATPase A candidate aminophospholipid flippase was purified from human erythrocytes using PS-stimulated, vanadate-sensitive ATPase activity as an enzyme marker [33, 38, 39]. Polyacrylamide gel electrophoretic analysis of the purified erythrocyte ATPase indicated that the preparation did not contain a predominant protein in the 120-kDa range, the molecular weight of other candidate erythrocyte flippase/Mg2+-ATPases [32, 34, 35]. Rather, the molecular weight of the predominant protein was approximately 80–88â•¯kDa [38]. Subsequent work showed that Mg2+-ATPase activity was associated with this 88-kDa protein [33, 38], and further analysis demonstrated that the protein could be cross-linked to ATP and could bind to anti-ATPase antibodies [40]. The purified erythrocyte ATPase demonstrates a unique requirement for lipid [2, 33, 39]. The detergent-solubilized enzyme is inactive in the absence of added phospholipid. PC and other net neutral lipids do not stimulate the ATPase. However, negatively charged lipids, such as PS, phosphatidylinositol (PI), phosphatidylglycerol (PG), and phosphatidic acid (PA), activate the purified, detergent-soluble enzyme [32, 33, 41]. Of these lipids, PS maximally activates the ATPase. Stimulation by PS is unique and is determined by multiple structural elements of the lipid. Modification of the amine group (N-methyl-

Specificity of Candidate Aminophospholipid Flippases

207

PS, phosphatidylhydroxypropionate), the carboxyl group (PS-methyl ester), or head-group size (phosphatidylhomoserine) reduces enzyme activity significantly. Most importantly, recognition of the lipid by the enzyme is stereospecific. Though serine head-group stereochemistry (P-L-S or P-D-S) has little effect on ATPase activity, the non-natural 2,3-sn-glycerol-PS isomers show little activation of the enzyme in detergentâ•›:â•›lipid mixed micelles. When reconstituted into proteoliposomes using a hydrophobic bead detergent removal technique [42], the enzyme displays an increased specificity for PS; the reconstituted enzyme is not activated by PI and is maximally activated only by PS, and activation is specific for the 1,2-sn-glycerol-PS isomer, but insensitive to the stereochemistry of the serine head group [43]. These data support the hypothesis that the PS-stimulated Mg2+-ATPase is a strong candidate flippase. The catalog of putative lipid transporters in the erythrocyte membrane is limited, reflecting the specialized nature of this cell, and is a good target for the identification of the flippase. Transport activity measurements have revealed the presence of putative inwardly directed PS flippases (ATP8A1), outwardly directed floppases that may transport cholesterol and glycerophospholipids (ABCA1, ABCB6, ABCC1), and a putative bidirectional lipid scramblase (PLSCR1) (reviewed in Reference 44) (see Chapter 7). Whole membrane proteomic analyses have confirmed the presence of these proteins in erythrocyte membranes [45]. The presence of Atp8a1 in mouse erythrocyte membranes was confirmed by both mRNA analysis of reticulocytes and Western blotting of mature erythrocytes with antibodies raised by Xie et al. [46] against bovine brain isoforms of ATP8A1 [47]. Furthermore, analysis of the purified PS-stimulated erythrocyte ATPase by matrix assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry revealed the presence of at least two ATPases: ABCB6 and ATP8A (D. Daleke, unpublished results). 10.3.2â•… The P4 Family of ATPases Evidence that bovine chromaffin granules possess flippase activity [48] prompted the purification of the only unassigned ATPase in these organelles (ATPase II, ATP8A1) [35]. The enzyme was found to be vanadate sensitive, NEM sensitive, and selectively stimulated by PS [37]. The gene encoding this protein was cloned [35] and sequence analysis indicated that the ATPase belongs to a new subfamily of P-type ATPases, including the previously reported yeast DRS2 gene product [49] (see Chapters 8 and 9). A yeast strain possessing a drs2 mutant was found to be defective in PS transport [35], further implicating this protein in flippase activity. These initial studies spawned a number of investigations that ultimately led to the identification of a new class of P-type ATPases, the P4-ATPases, which may possess phospholipid or amphipath transport activity. Fourteen mammalian and five yeast P4-ATPases have been identified, and orthologs have been discovered in worms, flies, and plants, but not prokaryotic cells. A growing body of biochemical and structural

208

FLIPPASE SUBSTRATE SPECIFICITY

data regarding these proteins has been collected, some of which supports the hypothesis that they are involved in transbilayer lipid transport or at the least in vesicular trafficking [17, 50]. 10.3.3â•… Yeast P4-ATPases Five P4-ATPases have been identified in yeast (Dnf1p, Dnf2p, Dnf3p, Drs2p, and Neo1p) [51], and these constitute an essential class of enzymes [52]. The yeast P4-ATPases have been associated primarily with vesicle trafficking and phospholipid transport activities. The founding member of this class of enzymes, Drs2p, is localized to the late Golgi and, to a smaller extent, the PM. Accumulating evidence indicates that Drs2p is a selective PS transporter. Golgi vesicles containing Drs2p transport PS, but not PC, but vesicles containing a mutant form of Drs2p are incapable of PS transport [53]. More significantly, proteoliposomes containing reconstituted purified Drs2p support the transport of NBD-PS, but not NBD-PC [54]. Not only are Dnf1p and Dnf2p localized to the PM and facilitate the transport of PS, PE, and PC [55], but they also mediate the uptake of lyso-PE [56] and alkylphosphocholine drugs [57, 58], indicating that they are transporters of relatively low specificity. Dnf3p localizes to the late Golgi and secretory vesicles much like Drs2p, but likely transports PC and not PS (or PE) [55]. All of the yeast P4-ATPases have been associated genetically and physically with a lower-molecular-weight partner protein of the CDC50 family (Cdc50p with Drs2p, Lem3p with Dnf1p/Dnf2p, Crf1p with Dnf3p) that may function to route the P4-ATPase to its proper intracellular location [59] or regulate its catalytic cycle [60]. Of the yeast P4ATPases, NEO1 is the only essential gene, and unlike the other P4-ATPases, Neo1p does not interact with a CDC50 partner protein. The lipid specificity, if any, of Neo1p is unknown, but it is likely not a lipid transporter. 10.3.4â•… ATP8 The closest mammalian ortholog to Drs2p is ATP8, and four isoforms of this protein (ATP8A1, ATP8A2, ATP8B1, and ATP8B2) have been identified [61]. The bovine chromaffin granule ATPase, ATP8A1, and the murine isoform of this enzyme have been particularly well studied with respect to lipid substrate specificity. ATP8A1 isolated from bovine chromaffin granules is selectively stimulated by PS, with maximal activation obtained in the presence of bovine brain PS, dimyristoylphosphatidylserine (DMPS), or dipalmitoylphosphatidylserine (DPPS) [62]. Other lipids and detergents (cardiolipin, lyso-PS) have little or no effect on the ATPase activity of ATP8A1 [62]. Four isoforms of ATP8A1 have been identified in bovine brain, and when expressed in insect cells with recombinant baculovirus and purified, the enzymes exhibited a degree of total activity and specificity for PS (compared with PE), that was dependent on the presence or absence of a 15-amino acid sequence and the presence of one of two alternative 15-amino sequences, respectively [46]. The

Specificity of Candidate Aminophospholipid Flippases

209

presence of the sequence 433HVPEPEDYGCSPDEW conferred greater PSstimulated activity on the enzyme, but had no effect on PE-stimulated activity, which was 2.5- to 9-fold less than that of PS. However, the presence of one of two alternative sequences, 152NVGDIVIIKGKEYI, conferred twofold higher PE-stimulated ATPase activity than enzymes containing the alternative sequence, 152AVGEIVKVTNGEHL. PS-stimulated activity was similar when either of these sequences was present. Although the reasons for these differences are unknown, it is tempting to speculate that the first sequence mediates PS binding to the enzyme, while the latter two sequences regulate PE binding. The murine isoform of Atp8a1 has been expressed in insect cells, using a recombinant baculovirus, and purified by affinity chromatography [63]. Similar to flippase activity in intact cells, the ATPase activity of Atp8a1 exhibited a preference for PS, including a dependence on the 1,2-sn-glycerol backbone configuration, and a lack of dependence on the serine head-group stereochemistry [63]. The purified detergent-solubilized enzyme exhibited no activity in the absence of added lipid and was not activated by other aminophospholipids or other anionic lipids. Mutations in human ATP8B1 result in the genetic disorder progressive familial intrahepatic cholestasis [64], which is characterized by a decrease in hepatocyte bile salt excretion [65]. Rats lacking Atp8b1, which localizes to the canalicular membrane [66], exhibited an increased sensitivity to bile salts [65, 67], and an increased amount of PS, but not PC, PE, or SM in the bile [65], suggesting that ATP8B1 is responsible for PS transport, at least at the canalicular membrane of hepatocytes. Recent studies have confirmed the lipid specificity of these transporters in reconstituted systems. The yeast flippase, Drs2p, has been purified and reconstituted into liposomes [54]. The reconstituted protein transports NBD-PS, but not NBD-PC, across the liposome membrane in an Mg2+-ATP-dependent manner. The rod photoreceptor outer segment contains well-organized disk membranes that are enriched in Atp8a2 [68]. Similar to murine Atp8a1 [63], detergent-solubilized, immunoaffinity-purified Atp8a2 exhibits PS-selective ATPase activity, weak activation by PE, and no activation by negatively charged (PI, PA, PG) or zwitterionic (PC, SM) lipids [68]. Atp8a2 reconstituted into proteoliposomes supports transport of NBD-PS, but not NBD-PE or NBD-PC, and is suppressed by the competitor DOPS [68]. Other recent studies have also shown that P4-ATPases may be regulated by lipids at a site distinct from the substrate binding site. Drs2p has been shown to be a downstream effector of phosphatidylinositol-4-phosphate (PIP) and contains an activity regulating PIP-binding site in its C-terminal tail that is homologous to a split PH domain [50]. A study of protein interactors of Drs2p yielded several enzymes involved in phosphoinositide metabolism, confirming the link between the flippase and this class of lipids. The mechanism by which phosphoinositides regulate flippase activity and whether this regulation is a general feature of all ATP8 family members provides an interesting new avenue of study for these enzymes.

210

FLIPPASE SUBSTRATE SPECIFICITY

10.3.5â•… ATP9 The closest ortholog to the mammalian ATP9 protein, Neo1p, is the only essential P4-ATPase [52]. Neo1p is localized to the ER [69], Golgi [69, 70], and late endosomes [70]. While lipid transport activity has not been associated with Neo1p, conditional knockdown of neo1 expression indicates a role for the protein in retrograde vesicle transport from the Golgi to the ER [69]. Neo1p interacts physically with Ysl2p, the guanine nucleotide exchange factor for the GTPase Arl1p, further supporting a role for Neo1p involvement in vesicle trafficking [70]. The mammalian ATP9 proteins, ATP9A and ATP9B, have not been investigated in detail, and although ATP9A and ATP9B are ubiquitously expressed in mouse tissue [71], the functions of these proteins and the lipid specificity, if any, have yet to be identified. 10.3.6â•… ATP10 and ATP11 ATP10 and ATP11 comprise three submembers each and are not orthologous to any of the yeast P4-ATPases. Mutations in human ATP10C have been linked to Angelman syndrome [72], a developmental central nervous system disorder. In mice, Atp10c deficiency results in obesity, possibly the result of inefficient insulin signaling and exocytosis of GLUT4-containing vesicles transporting the glucose transporter to the PM [73]. A splice isoform of rabbit ATP11B has been localized to the inner nuclear membrane, where it interacts with the RING motifs of RUSH transcription factors [74]; however, this is a speciesspecific splice event and occurs in rabbits but not humans [75, 76]. Whether ATP10 and ATP11 are involved with vesicle trafficking or whether they interact specifically with phospholipids has not yet been addressed. 10.4â•… IS THE LIPID SPECIFICITY OF CANDIDATE AMINOPHOSPHOLIPID FLIPPASES UNIQUE? A significant body of previously published studies exists regarding the regulation of other P-type ATPases by lipids. Evidence that the P4 family of P-type ATPases may encode putative lipid transporters has raised the question of whether all P-type ATPases share a common mode of interaction with activating lipids. However, the lipid–protein interaction of most of the members of the broader P-type ATPase superfamily do not demonstrate the unique specificity for PS that is characteristic of the P4-ATPases, further supporting the unique interactions of PS as a transport substrate with the P4-ATPases. 10.4.1â•… PM Ca2+-ATPase (P2b-ATPase) The PM Ca2+-ATPase is responsible for transporting Ca2+ out of the cytoplasm (one Ca2+ in exchange for one H+) and plays a crucial role in maintaining a low level of intracellular Ca2+. The Ca2+-ATPase is regulated by interaction

IS LIPID SPECIFICITY OF AMINOPHOSPHOLIPID FLIPPASES UNIQUE?

211

with other proteins (calmodulin), phosphorylation (by Ser/Thr kinases and Tyr kinases), and phospholipids. The lipid activators include not only abundant membrane components like PS and fatty acids, but also lipids that act as second messengers, such as phosphatidylinositol-4,5-bisphosphate (PIP2). Similar to calmodulin [77], the anionic phospholipids PS [77–83], PI [77, 78, 83, 84], PIP2 [80, 85, 86], PA [78, 86, 87], and cardiolipin [77, 78] are capable of activating the Ca2+-ATPase. The phospholipid derivatives, phosphatidylethanol and phosphatidylbutanol, are also able to activate the Ca2+-ATPase [81]. However, the neutral lipids PC [78–80], PE [78, 84], and SM [78, 79] are not. Two anionic phospholipid-binding domains exist on the Ca2+-ATPase, one of them shared with the calmodulin-binding domain and binds PIP2 [88, 89], while the other acts as an inhibitory domain, until bound by phospholipid [87]. The interaction of the anionic phospholipid-binding domains of the enzyme with the calmodulin-binding domain, and its apparent specificity for PIP2, suggests that the Ca2+-ATPase interaction with phospholipids plays a role in Ca2+ signaling. This is in contrast to the environmental lipid requirements for Ca2+ATPase activity, in which the lipids participate in charge stabilization, hydrophobic matching, and membrane fluidity. Regardless, although this enzyme is activated by PS, other negatively charged lipids support ATPase activity, indicating that the mode of activation by lipids is qualitatively different from the P4-ATPases. 10.4.2â•… SERCA Pump (P2a-ATPase) Similar to the PM Ca2+-ATPase, the SERCA pump is also regulated by lipid. Delipidating the sarcoplasmic reticulum (SR) Ca2+-ATPase causes a decrease in ATPase activity [90] and PIP associates particularly strongly with the protein [91]. ATPase activity is maintained at maximal levels in reconstituted proteoliposomes containing PC vesicles and low mole fractions of anionic phospholipids, such as PS [92–94], PA [92], cardiolipin [94], and PIP [93, 95]. Ca2+ uptake is greatest in PC vesicles containing PS [93, 96], PA [95], PI [95], PIP [95, 96], or cardiolipin [95] compared with PC alone. Increasing the PE content in vesicles containing the SR Ca2+-ATPase results in Ca2+ uptake to almost the same level as the enzyme reconstituted into SR phospholipids [97], while including lyso-PC in the phospholipid mixture inhibits activity through restricting movement in the nucleotide-binding domain [98]. PE decreases the rate of hydrolysis of the phosphoenzyme intermediate compared with PC [99]. In sum, these data indicate that, like the PM Ca2+-ATPase, the SR Ca2+ATPase does not display the same unique selectivity for PS that is characteristic of some of the P4-ATPases. 10.4.3â•… Na+/K+-ATPase (P2c-ATPase) The Na+/K+-ATPase is one of the most well-studied P-type ATPases responsible for generating the electrochemical gradient across the PM, allowing neurons to fire and cells to maintain ionic homeostasis. Early studies on the

212

FLIPPASE SUBSTRATE SPECIFICITY

lipid regulation of the Na+/K+-ATPase utilized detergents to solubilize the enzyme and remove associated phospholipids, and found that the re-addition of PS [100, 101], PI [100], or PA [101] could activate the enzyme, while PC [100, 101] or PE [100] could not. Detergents, however, may not remove phospholipids associated with the enzyme with high affinity. Later studies identified that residual PI remained associated with the Na+/K+-ATPase after delipidation [102, 103] and may explain the persistence of ATPase activity. Phospholipids that associate closely with membrane proteins can be removed under more stringent conditions, including phospholipase digestion, extraction into solvents, or enzymatic modification of the protein-bound phospholipids to completely inactivate the enzyme prior to testing for required membrane components. With these methods, the anionic phospholipid requirement for ATPase activity was confirmed for PS [102, 104], PA [104], PI [102], and PG [102], while SM [104], PE [102, 104], and PC [102, 104] did not activate the Na+/K+-ATPase. By enzymatically converting PS to PE, it was found that the enzyme was inactivated only after the last 13â•¯molâ•›% of PS had been converted to PE [104]. Taniguchi and Tonomura [105] found that adding back a combination of PI, PS, and PE to delipidated enzyme resulted in the greatest activation of Na+/K+-ATPase activity, although complete reactivation could be achieved with PI and PS alone. This suggests that while anionic phospholipids can hyperactivate the Na+/K+-ATPase, a more natural membrane environment is required for full activity. Phospholipid head-group specificity was examined through chemical modification of the head-group moieties, confirming activation of the ATPase by PS, PA, and PG, but not PC [106]. The extent of activation is somewhat dependent on the acyl chain composition, as PA and PG derived from different parent phospholipids had differing extents of Na+/K+ATPase activation [106]. Lipids may play a role at specific points in the reaction cycle of the Na+/ + K -ATPase. Wheeler [107] noted that PS did not affect Na+-activated phosphorylation, but did stimulate the release of phosphate. Fluorescent labeling of amino acid residues has been used to examine the role of anionic phospholipids over the Na+/K+-ATPase reaction cycle. Anionic phospholipids influence the stages of the Na+/K+-ATPase reaction cycle, facilitating the formation of stable intermediate enzyme conformations; however, these do not seem to be specific for a particular phospholipid. Another possibility is that negatively charged phospholipids, PS, PG, and PI, stabilize the oligomerization of the Na+/ K+-ATPase as a functional unit and not necessarily stabilizing ATPase reaction intermediates [108]. Unlike the P4-ATPase, ATP8, other members of the P-type ATPase superfamily do not interact with PS with a high degree of specificity but, in general, are activated by anionic lipids. It will be enlightening to study the effect on other members of the P-type superfamily of the catalog of PS structural and stereoisomers that have been developed for the study of the P4-ATPases (Fig. 10.1). Similarly, these lipids could also be used to reveal during which steps of the ATP8 reaction cycle the protein interacts selectively with PS, likely the

Lipid Specificity of Other PS-Binding Proteins

213

E1P intermediate [17], as has been explored with substrates of other P-type ATPases. 10.5â•… LIPID SPECIFICITY OF OTHER PS-BINDING PROTEINS Selective interaction with PS is not a unique property of the flippase. A number of well-studied proteins bind to, and are activated by, PS. These interactions vary in degree of selectivity for PS, but all exhibit a distinct type of protein–lipid interaction than that observed with the flippase. Examples of this class of PS-binding proteins include protein kinase C (PKC), blood clotting factors, and the macrophage PS receptor (Table 10.1). PKC is a lipid-regulated serine/threonine kinase that is activated by diacylglycerol and PS. Cell activation stimulates the production of diacylglycerol that recruits PKC to the membrane interface, where binding to PS induces a conformational change that releases an autoinhibitory pseudosubstrate domain from the active site. Activation by PS is dependent on the serine headgroup structure and stereochemistry; modifications to the amine group, carboxyl group, stereochemistry, and size abolish activation of the enzyme [109]. In contrast to the flippase, PKC is not activated by the D-serine isomer of PS [110]. Binding of PKC to bilayers shows a similar pattern of specificity, except that, like other negatively charged lipids, the enzyme will bind to 1,2-sn-P-D-S [110]. Plasma clotting factors are activated by membrane surfaces that contain exposed PS. Factors Va and Xa assemble on PS-containing membranes in the presence of Ca2+ to activate the conversion of prothrombin to thrombin. The interaction of these factors is dependent on the stereochemistry and structure of the serine head group; the L-serine analog, but not the D-serine or Nmethylserine analog, is capable of productively activating the clotting factors [111, 112]. Similarly, Factor VIII activity, deficiencies of which result in hemophilia A, is also dependent on the stereochemistry of the L-serine head group [113, 114]. The recognition and destruction of PS-exposing cells by macrophages is mediated in part by a receptor that selectively recognizes PS [115]. Binding of PS to the PS receptor, and subsequent endocytosis of the bound cell, requires the L-serine isomer. The PS receptor does not recognize the D-serine isomer of PS. The essential distinction between the activation or binding of these proteins to PS and that of the PS flippase, the erythrocyte PS-stimulated Mg2+-ATPase, and ATP8A1 is the recognition of the chiral centers in the PS molecule. Whereas the flippase and flippase candidates do not discriminate between the L- and D-serine head groups, other PS-specific proteins bind, or are activated by, only the L-serine isomer of PS (Fig. 10.3). In contrast, the flippase and the two ATPases described above are dependent on the stereochemistry of the sn-2 position of the glycerol moiety, but the other PS-interacting proteins are not.

214

Flippase Atp8a1 Erythrocyte ATPase Prothrombinase complex Factor VIII Factor Xa Factor Va PKC PKC PS receptor PS receptor

Protein

Transport ATPase ATPase Prothrombinase Binding Binding Binding Binding Kinase Binding Phagocytosis

Activity P-L-S

P-D-S +++ ++ ++ − − − − + − − −

P-L-S +++ ++ ++ + + + + + + + +

+ − −

−

+ − −

+ +

P-D-S

2,3-sn-glycerol

1,2-sn-glycerol

− −

+ + −

PP + − −

− − −

− − − −

PS-O-Me

+++ − +

PS-N-Me

1,2-sn-glycerol

TABLE 10.1.â•‡ Lipid Structural Specificity of PS-Activated Proteins (Adapted from Reference 121)

+ − + − − + − − −

PhS [63] [33, 122] [112] [113, 114] [111] [112] [110, 123, 124] [110, 123, 124] [115, 125] [115, 125]

Reference

215

Sequence Elements That Bind to PS

L-Serine Required Protein Kinase C Clotting Factors (Va, VIII, Xa) Macrophage PS Receptor

1,2-sn-Glycerol Required Flippase RBC ATPase; P4-ATPases O

O O

O O O

O–

P

O H

H3N

+

O

O– Serine Stereochemistry Independent Flippase RBC ATPase; P4-ATPases

Figure 10.3.â•‡ Summary of the lipid specificity of phosphatidylserine-dependent proteins. Soluble membrane-binding proteins that show a specificity for PS recognize the stereochemistry of the serine head group and prefer the L-isoform, whereas enzymes associated with PS flippase activity do not, but rather prefer the 1,2-sn configuration of the glycerol moiety of the lipid. RBC, red blood cell.

These data suggest that at least two different motifs for selectively binding PS have evolved, depending on whether the interacting protein is a soluble or a transmembrane protein. Soluble proteins involved in interfacial interactions with PS-containing membranes, such as PKC, clotting factors, or the PS receptor, approach a target membrane with a different geometry than the flippase, the erythrocyte Mg2+-ATPase, and ATP8A1, which interact with PS laterally within the membrane. Although it is possible that transmembrane proteins, such as the flippase, may contain loops or segments that approach the membrane from the aqueous phase in a manner similar to interfacial binding proteins, the sequences of amino acids that mediate these two very distinct modes of interaction with PS will likely have significantly different structures. 10.6â•… SEQUENCE ELEMENTS THAT BIND TO PS The variety of proteins that recognize PS selectively implies that specific amino acid sequences may have evolved to recognize this lipid. Unlike other well-known lipid-binding sequences, such as the pleckstrin homology (PH) domain [116, 117], a characteristic PS-binding domain has not been identified. However, clues to potential PS-binding sites may be uncovered by comparing the sequence of known PS-binding proteins. In an elegant series of experiments, Igarashi et al. discovered a potential PS-binding domain by sequencing the hypervariable regions of an anti-PS antibody [118]. Homology searches revealed that similar sequences were

216

FLIPPASE SUBSTRATE SPECIFICITY

TABLE 10.2.â•‡ Consensus PS-Binding Sequences Protein

Accession

Species

Sequence

Consensus PKC alpha PKC beta PKC gamma PS decarboxylase ATP8A1

NP_002728 NP_002729 NP_002730 NP_055153 NP_033857

Human Human Human Human Mouse

FxFxLKxxxKxR 227 FTFKLKPSDKDR 227 FRFQLKESDKDR 227 FVFNLKPGDVER 356 FNFQLKTGGKIR 520 FVFTGRTPDSVI 956 FWFPLKALQYGT

present in PS-binding proteins that contain C2 domains (PKC, phospholipase C) and PS-metabolizing enzymes (PS decarboxylase, PS synthase; Table 10.2). The consensus sequence (FxFxxRKxxxL) contains elements (aromatic and positively charged amino acids) that reasonably indicate that it might interact with negatively charged lipid membranes, and a search of the sequence of one putative flippase (ATP8A1) reveals the presence of two instances of the N-terminal half of this sequence (Table 10.2). Given that the C-terminal half of the Igarashi consensus sequence is relatively nonspecific, one might expect that the required elements could be provided by neighboring, out-of-sequence amino acids. Alternatively, as has been observed with PH domains, the remaining half of this sequence might show weak primary homology but significant secondary structural homology [119]. However, conclusive evidence that this sequence selectively recognizes PS is lacking. Indeed, the hypothesis that this domain is involved in PS binding was tested directly by mutating each of the positively charged residues in the C2 domain sequence of PKC. None of these mutations, including mutation of the full set of basic residues, decreased binding to PS [120]. Whether these sequences directly interact with PS or are involved in some other common structural requirement for PS-binding proteins remains to be fully tested. 10.7â•… CONCLUSIONS The unique substrate specificity of the PS flippase gives clues to the structure of the lipid substrate binding site as well as provides essential criteria for the identification of putative flippases. Substrate transport requires the interaction of almost all of the structural elements of PS with the transporter, broadly identifying possible residues necessary for lipid binding and transport. The requirements of a substantial aqueous channel to accommodate the polar lipid head group, with appropriate residues to bind to the amine, carboxyl, and phosphate groups, and a structure that allows the fatty acyl groups to be readily accommodated in the bilayer, place the glycerol backbone in a position between transmembrane helices that could provide an explanation for the

217

Acknowledgments

(b)

(a)

out

in

(c)

ATP

Figure 10.4.â•‡ The flippase may contain two specific lipid-binding sites. Soluble PSbinding proteins may have a binding site (a) that is structurally distinct from the substrate-binding site of the PS flippase (b) that accommodates the different geometries of the approach of the binding sites to the membrane. The flippase may also contain nonsubstrate lipid-binding sites (c) for PS, or other lipids, in which cytosolic domains of the flippase approach the membrane in a manner similar to soluble membrane-binding proteins. Color version on the Wiley web site.

stereospecific recognition of this lipid (Fig. 10.2). This unique requirement is a good criterion for the search for potential flippases, and selective interaction with PS is shared by Mg2+-ATPase purified from human erythrocytes and Atp8a1. This selectivity is distinct from the relatively nonspecific requirement that other P-type ATPases possess for anionic lipids. In addition, the glycerol stereospecificity also distinguishes the flippase from other PS-binding proteins that are sensitive to the stereochemistry of the serine head group, rather than the glycerol backbone, and indicates that interfacial PS-binding proteins have evolved a binding site that is different from the PS-flippase-binding site, which may involve interactions with the lipid laterally within the membrane (Fig. 10.4). Finally, comparison of the primary sequence of putative flippases with other PS-binding proteins identifies common sequences that may prove to be PS-binding sites.

ACKNOWLEDGMENTS This work was supported by grants from the National Institutes of Health (GM47230), the American Heart Association, and the Indiana University School of Medicine.

218

FLIPPASE SUBSTRATE SPECIFICITY

ABBREVIATIONS ATP DMPS DOPS DPPS GPS NBD NEM PA PC PE PG PH PI PIP PIP2 PS RBC SM SR

adenosine triphosphate dimyristoylphosphatidylserine dioleoylphosphatidylserine dipalmitoylphosphatidylserine glycerophosphorylserine 7-nitro-2-1,3-benzoxadiazol-4-yl N-ethylmaleimide phosphatidic acid phosphatidylcholine phosphatidylethanolamine phosphatidylglycerol pleckstrin homology phosphatidylinositol phosphatidylinositol-4-phosphate phosphatidylinositol-4,5-bisphosphate phosphatidyserine red blood cell sphingomyelin sarcoplasmic reticulum

REFERENCES â•… â•… â•… â•…

1â•… 2â•… 3â•… 4â•…

â•… 5â•… â•… 6â•… â•… 7â•… â•… 8â•… â•… 9â•… â•‡ 10â•… â•‡ 11â•… â•‡ 12â•…

P. F. Devaux, Biochemistry 1991, 30, 1163–1173. D. Daleke, J. Lipid Res. 2003, 44, 233–242. R. F. Zwaal, P. Comfurius, E. M. Bevers, Cell. Mol. Life Sci. 2005, 62, 971–988. V. A. Fadok, D. R. Voelker, P. A. Campbell, J. J. Cohen, D. L. Bratton, P. M. Henson, J. Immunol. 1992, 148, 2207–2216. E. M. Bevers, P. Comfurius, J. L. van Rijn, H. C. Hemker, R. F. Zwaal, Eur. J. Biochem. 1982, 122, 429–436. B. Lubin, D. Chiu, J. Bastacky, B. Roelofsen, L. L. Van Deenen, J. Clin. Invest. 1981, 67, 1643–1649. R. A. Schlegel, L. McEvoy, P. Williamson, Bibl. Haematol. 1985, 51, 150–156. R. K. Wali, S. Jaffe, D. Kumar, V. K. Kalra, Diabetes 1988, 37, 104–111. Y. Tanaka, A. J. Schroit, J. Biol. Chem. 1983, 258, 11335–11343. M. J. Wilson, K. Richter-Lowney, D. L. Daleke, Biochemistry 1993, 32, 11302–11310. D. M. Niedowicz, D. L. Daleke, Cell Biochem. Biophys. 2005, 43, 289–330. M. Bitbol, P. Fellmann, A. Zachowski, P. F. Devaux, Biochim. Biophys. Acta 1987, 904, 268–282.

References

â•‡ 13â•… â•‡ 14â•… â•‡ 15â•… â•‡ 16â•… â•‡ 17â•… â•‡ 18â•… â•‡ 19â•… â•‡ 20â•… â•‡ 21â•… â•‡ 22â•… â•‡ 23â•… â•‡ 24â•… â•‡ 25â•… â•‡ 26â•… â•‡ 27â•… â•‡ 28â•… â•‡ 29â•… â•‡ 30â•… â•‡ 31â•… â•‡ 32â•… â•‡ 33â•… â•‡ 34â•… â•‡ 35â•… â•‡ 36â•… â•‡ 37â•… â•‡ 38â•… â•‡ 39â•… â•‡ 40â•…

â•‡ 41â•… â•‡ 42â•…

219

D. L. Daleke, W. H. Huestis, Biochemistry 1985, 24, 5406–5416. M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. D. L. Daleke, J. Biol. Chem. 2007, 282, 821–825. P. F. Devaux, I. Lopez-Montero, S. Bryde, Chem. Phys. Lipids 2006, 141, 119–132. G. Lenoir, P. Williamson, J. C. Holthuis, Curr. Opin. Chem. Biol. 2007, 11, 654–661. G. van Meer, D. Halter, H. Sprong, P. Somerharju, M. R. Egmond, FEBS Lett. 2006, 580, 1171–1177. C. F. Puts, J. C. Holthuis, Biochim. Biophys. Acta 2009, 1791, 603–611. J. Connor, A. J. Schroit, Biochemistry 1987, 26, 5099–5105. A. Zachowski, E. Favre, S. Cribier, P. Herve, P. F. Devaux, Biochemistry 1986, 25, 2585–2590. T. Smriti, E. C. Nemergut, D. L. Daleke, Biochemistry 2007, 46, 2249–2259. G. Morrot, P. Herve, A. Zachowski, P. Fellmann, P. F. Devaux, Biochemistry 1989, 28, 3456–3462. D. C. Drummond, D. L. Daleke, Chem. Phys. Lipids 1995, 75, 27–41. P. Fellmann, P. Herve, T. Pomorski, P. Muller, D. Geldwerth, A. Herrmann, P. F. Devaux, Biochemistry 2000, 39, 4994–5003. V. Puri, C. M. Gupta, Biochim. Biophys. Acta 1998, 1373, 59–66. J. E. Vance, J. Lipid Res. 2008, 49, 1377–1387. O. C. Martin, R. E. Pagano, J. Biol. Chem. 1987, 262, 5890–5898. M. P. Hall, W. H. Huestis, Biochim. Biophys. Acta 1994, 1190, 243–247. R. K. Loh, W. H. Huestis, Biochemistry 1993, 32, 11722–11726. M. Colleau, P. Herve, P. Fellmann, P. F. Devaux, Chem. Phys. Lipids 1991, 57, 29–37. G. Morrot, A. Zachowski, P. F. Devaux, FEBS Lett. 1990, 266, 29–32. M. L. Zimmerman, D. L. Daleke, Biochemistry 1993, 32, 12257–12263. M. E. Auland, B. D. Roufogalis, P. E. Devaux, A. Zachowski, Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 10938–10942. X. Tang, M. S. Halleck, R. A. Schlegel, P. Williamson, Science 1996, 272, 1495–1497. X.-S. Xie, D. K. Stone, E. Racker, J. Biol. Chem. 1989, 264, 1710–1714. Y. Moriyama, N. Nelson, J. Biol. Chem. 1988, 263, 8521–8527. D. L. Daleke, J. V. Lyles, Biochim. Biophys. Acta 2000, 1486, 108–127. J. V. Lyles, K. Cornely-Moss, C. M. Smith, D. L. Daleke, Methods Mol. Biol. 2003, 228, 257–269. J. V. Lyles, Purification and nucleotide specificity of the human erythrocyte aminophospholipid transporter. PhD Dissertation, Bloomington, Indiana University, 1998. D. L. Daleke, J. V. Lyles, E. Nemergut, M. L. Zimmerman, NATO ASI Ser. 1995, H 91, 49–59. J.-L. Rigaud, M.-T. Paternostre, A. Bluzat, Biochemistry 1988, 27, 2677–2688.

220

FLIPPASE SUBSTRATE SPECIFICITY

â•‡ 43â•… M. Zimmerman, D. L. Daleke, Biophys. J. 1995, 68, A443. â•‡ 44â•… D. L. Daleke, Curr. Opin. Hematol. 2008, 15, 191–195. â•‡ 45â•… S. R. Goodman, A. Kurdia, L. Ammann, D. Kakhniashvili, O. Daescu, Exp. Biol. Med. (Maywood) 2007, 232, 1391–1408. â•‡ 46â•… J. Ding, Z. Wu, B. P. Crider, Y. Ma, X. Li, C. Slaughter, L. Gong, X. S. Xie, J. Biol. Chem. 2000, 275, 23378–23386. â•‡ 47â•… E. Soupene, F. A. Kuypers, Br. J. Haematol. 2006, 133, 436–438. â•‡ 48â•… A. Zachowski, J. P. Henry, P. F. Devaux, Nature 1989, 340, 75–76. â•‡ 49â•… T. L. Ripmaster, G. P. Vaughn, J. L. Woolford, Jr., Mol. Cell. Biol. 1993, 13, 7901–7912. â•‡ 50â•… B. P. Muthusamy, P. Natarajan, X. Zhou, T. R. Graham, Biochim. Biophys. Acta 2009, 1791, 612–619. â•‡ 51â•… P. Catty, A. de Kerchove d’Exaerde, A. Goffeau, FEBS Lett. 1997, 409, 325–332. â•‡ 52â•… Z. Hua, P. Fatheddin, T. R. Graham, Mol. Biol. Cell 2002, 13, 3162–3177. â•‡ 53â•… P. Natarajan, J. Wang, Z. Hua, T. R. Graham, Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 10614–10619. â•‡ 54â•… X. Zhou, T. R. Graham, Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 16586–16591. â•‡ 55â•… N. Alder-Baerens, Q. Lisman, L. Luong, T. Pomorski, J. C. Holthuis, Mol. Biol. Cell 2006, 17, 1632–1642. â•‡ 56â•… W. R. Riekhof, J. Wu, M. A. Gijon, S. Zarini, R. C. Murphy, D. R. Voelker, J. Biol. Chem. 2007, 282, 36853–36861. â•‡ 57â•… P. K. Hanson, L. Malone, J. L. Birchmore, J. W. Nichols, J. Biol. Chem. 2003, 278, 36041–36050. â•‡ 58â•… F. J. Perez-Victoria, M. P. Sanchez-Canete, S. Castanys, F. Gamarro, J. Biol. Chem. 2006, 281, 23766–23775. â•‡ 59â•… K. Saito, K. Fujimura-Kamada, N. Furuta, U. Kato, M. Umeda, K. Tanaka, Mol. Biol. Cell 2004, 15, 3418–3432. â•‡ 60â•… G. Lenoir, P. Williamson, C. F. Puts, J. C. Holthuis, J. Biol. Chem. 2009, 284, 17956–17967. â•‡ 61â•… M. S. Halleck, D. Pradhan, C. Blackman, C. Berkes, P. Williamson, R. A. Schlegel, Genome Res. 1998, 8, 354–361. â•‡ 62â•… Y. Moriyama, N. Nelson, M. Maeda, M. Futai, Arch. Biochem. Biophys. 1991, 286, 252–256. â•‡ 63â•… J. K. Paterson, K. Renkema, L. Burden, M. S. Halleck, R. A. Schlegel, P. Williamson, D. L. Daleke, Biochemistry 2006, 45, 5367–5376. â•‡ 64â•… L. N. Bull, M. J. van Eijk, L. Pawlikowska, J. A. DeYoung, J. A. Juijn, M. Liao, L. W. Klomp, N. Lomri, R. Berger, B. F. Scharschmidt, A. S. Knisely, R. H. Houwen, N. B. Freimer, Nat. Genet. 1998, 18, 219–224. â•‡ 65â•… C. C. Paulusma, A. Groen, C. Kunne, K. S. Ho-Mok, A. L. Spijkerboer, D. Rudi de Waart, F. J. Hoek, H. Vreeling, K. A. Hoeben, J. van Marle, L. Pawlikowska, L. N. Bull, A. F. Hofmann, A. S. Knisely, R. P. Oude Elferink, Hepatology 2006, 44, 195–204. â•‡ 66â•… P. Ujhazy, D. Ortiz, S. Misra, S. Li, J. Moseley, H. Jones, I. M. Arias, Hepatology 2001, 34, 768–775.

References

221

â•‡ 67â•… S.Y. Cai, S. Gautam,T. Nguyen, C. J. Soroka, C. Rahner, J. L. Boyer, Gastroenterology 2009, 136, 1060–1069. â•‡ 68â•… J. A. Coleman, M. C. M. Kwok, R. S. Molday, J. Biol. Chem. 2009, 284, 32670–32679. â•‡ 69â•… Z. Hua, T. R. Graham, Mol. Biol. Cell 2003, 14, 4971–4983. â•‡ 70â•… S. Wicky, H. Schwarz, B. Singer-Kruger, Mol. Cell. Biol. 2004, 24, 7402–7418. â•‡ 71â•… M. S. Halleck, J. J. Lawler, S. Blackshaw, L. Gao, P. Nagarajan, C. Hacker, S. Pyle, J. T. Newman, Y. Nakanishi, H. Ando, D. Weinstock, P. Williamson, R. A. Schlegel, Physiol. Genomics 1999, 1, 139–150. â•‡ 72â•… M. Meguro, A. Kashiwagi, K. Mitsuya, M. Nakao, I. Kondo, S. Saitoh, M. Oshimura, Nat. Genet. 2001, 28, 19–20. â•‡ 73â•… M. S. Dhar, J. S. Yuan, S. B. Elliott, C. Sommardahl, J. Nutr. Biochem. 2006, 17, 811–820. â•‡ 74â•… M. Mansharamani, A. Hewetson, B. S. Chilton, J. Biol. Chem. 2001, 276, 3641–3649. â•‡ 75â•… M. S. Halleck, R. A. Schlegel, P. L. Williamson, J. Biol. Chem. 2002, 277, 9736–9740. â•‡ 76â•… A. Hewetson, A. E. Wright-Pastusek, R. A. Helmer, K. A. Wesley, B. S. Chilton, Mol. Cell. Endocrinol. 2008, 292, 79–86. â•‡ 77â•… T. A. Ansah, A. Molla, S. Katz, J. Biol. Chem. 1984, 259, 13442–13450. â•‡ 78â•… V. Niggli, E. S. Adunyah, E. Carafoli, J. Biol. Chem. 1981, 256, 8588–8592. â•‡ 79â•… M. Hermoni-Levine, H. Rahamimoff, Biochemistry 1990, 29, 4940–4950. â•‡ 80â•… J. Lehotsky, L. Raeymaekers, L. Missiaen, F. Wuytack, H. De Smedt, R. Casteels, Biochim. Biophys. Acta 1992, 1105, 118–124. â•‡ 81â•… M. Suju, M. Davila, G. Poleo, R. Docampo, G. Benaim, Biochem. J. 1996, 317(Pt 3), 933–938. â•‡ 82â•… H. P. Adamo, T. Pinto Fde, L. M. Bredeston, G. R. Corradi, Ann. N.Y. Acad. Sci. 2003, 986, 552–553. â•‡ 83â•… C. V. Filomatori, A. F. Rega, J. Biol. Chem. 2003, 278, 22265–22271. â•‡ 84â•… L. M. Bredeston, A. F. Rega, Biochem. J. 2002, 361, 355–361. â•‡ 85â•… A. Enyedi, M. Flura, B. Sarkadi, G. Gardos, E. Carafoli, J. Biol. Chem. 1987, 262, 6425–6430. â•‡ 86â•… D. Choquette, G. Hakim, A. G. Filoteo, G. A. Plishker, J. R. Bostwick, J. T. Penniston, Biochem. Biophys. Res. Commun. 1984, 125, 908–915. â•‡ 87â•… T. Pinto Fde, H. P. Adamo, J. Biol. Chem. 2002, 277, 12784–12789. â•‡ 88â•… A. G. Filoteo, A. Enyedi, J. T. Penniston, J. Biol. Chem. 1992, 267, 11800–11805. â•‡ 89â•… P. Brodin, R. Falchetto, T. Vorherr, E. Carafoli, Eur. J. Biochem. 1992, 204, 939–946. â•‡ 90â•… C. Hidalgo, M. de la Fuente, M. E. Gonzalez, Arch. Biochem. Biophys. 1986, 247, 365–371. â•‡ 91â•… M. Varsanyi, H. G. Tolle, M. G. Heilmeyer, Jr., R. M. Dawson, R. F. Irvine, EMBO J. 1983, 2, 1543–1548. â•‡ 92â•… K. A. Dalton, S. Mall, J. D. Pilot, J. M. East, A. G. Lee, Biochem. Soc. Trans. 1998, 26, S234.

222

FLIPPASE SUBSTRATE SPECIFICITY

â•‡ 93â•… G. Szymanska, H. W. Kim, E. G. Kranias, Biochim. Biophys. Acta 1991, 1091, 127–134. â•‡ 94â•… R. Vemuri, K. D. Philipson, J. Biol. Chem. 1989, 264, 8680–8685. â•‡ 95â•… K. A. Dalton, J. D. Pilot, S. Mall, J. M. East, A. G. Lee, Biochem. J. 1999, 342(Pt 2), 431–438. â•‡ 96â•… G. Szymanska, H. W. Kim, J. Cuppoletti, E. G. Kranias, Mol. Cell. Biochem. 1992, 114, 65–71. â•‡ 97â•… G. W. Hunter, S. Negash, T. C. Squier, Biochemistry 1999, 38, 1356–1364. â•‡ 98â•… G. W. Hunter, D. J. Bigelow, T. C. Squier, Biochemistry 1999, 38, 4604–4612. â•‡ 99â•… A. P. Starling, K. A. Dalton, J. M. East, S. Oliver, A. G. Lee, Biochem. J. 1996, 320(Pt 1), 309–314. 100â•… K. P. Wheeler, R. Whittam, J. Physiol. 1970, 207, 303–328. 101â•… L. J. Fenster, J. H. Copenhaver, Jr., Biochim. Biophys. Acta 1967, 137, 406–408. 102â•… J. G. Mandersloot, B. Roelofsen, J. de Gier, Biochim. Biophys. Acta 1978, 508, 478–485. 103â•… B. Roelofsen, M. Van Linde-Sibenius Trip, Biochim. Biophys. Acta 1981, 647, 302–306. 104â•… B. Roelofsen, L. L. van Deenen, Eur. J. Biochem. 1973, 40, 245–257. 105â•… K. Taniguchi, Y. Tonomura, J. Biochem. 1971, 69, 543–557. 106â•… J. A. Walker, K. P. Wheeler, Biochim. Biophys. Acta 1975, 394, 135–144. 107â•… K. P. Wheeler, Biochem. J. 1975, 146, 729–738. 108â•… K. Mimura, H. Matsui, T. Takagi, Y. Hayashi, Biochim. Biophys. Acta 1993, 1145, 63–74. 109â•… M.-H. Lee, R. M. Bell, J. Biol. Chem. 1989, 264, 14797–14805. 110â•… A. C. Newton, L. M. Keranen, Biochemistry 1994, 33, 6651–6658. 111â•… M. Banerjee, D. C. Drummond, A. Srivastava, D. Daleke, B. R. Lentz, Biochemistry 2002, 41, 7751–7762. 112â•… X. Zhai, A. Srivastava, D. C. Drummond, D. Daleke, B. R. Lentz, Biochemistry 2002, 41, 5675–5684. 113â•… G. E. Gilbert, A. A. Arena, J. Biol. Chem. 1995, 270, 18500–18505. 114â•… G. E. Gilbert, D. Drinkwater, Biochemistry 1993, 32, 9577–9585. 115â•… V. A. Fadok, D. L. Bratton, D. M. Rose, A. Pearson, R. A. Ezekewitz, P. M. Henson, Nature 2000, 405, 85–90. 116â•… B. J. Mayer, R. Ren, K. L. Clark, D. Baltimore, Cell 1993, 73, 629–630. 117â•… R. J. Haslam, H. B. Koide, B. A. Hemmings, Nature 1993, 363, 309–310. 118â•… K. Igarashi, M. Kaneda, A. Yamaji, T. C. Saido, U. Kikkawa, Y. Ono, K. Inoue, M. Umeda, J. Biol. Chem. 1995, 270, 29075–29078. 119â•… M. J. Rebecchi, S. Scarlata, Annu. Rev. Biophys. Biomol. Struct. 1998, 27, 503–528. 120â•… J. E. Johnson, A. S. Edwards, A. C. Newton, J. Biol. Chem. 1997, 272, 30787–30792. 121â•… J. K. Paterson, Lipid specificity and transport activity of the murine ATPase II, a candidate aminophospholipid flippase. Dissertation, Bloomington, Indiana University, 2003.

References

223

122â•… M. L. Zimmerman, Ph.D. dissertation, Department of Chemistry, Bloomington, Indiana University, 210, 1996. 123â•… M. H. Lee, R. M. Bell, J. Biol. Chem. 1989, 264, 14797–14805. 124â•… J. E. Johnson, M. L. Zimmerman, D. L. Daleke, A. C. Newton, Biochemistry 1998, 37, 12020–12025. 125â•… V. A. Fadok, A. de Cathelineau, D. L. Daleke, P. M. Henson, D. L. Bratton, J. Biol. Chem. 2001, 276, 1071–1077.

11 THE FLIPPASE DELUSION? Naomi L. Pollock, Petra H.M. Niesten, and Richard Callaghan Nuffield Department of Clinical Laboratory Sciences, John Radcliffe Hospital, University of Oxford, Oxford, UK

11.1â•… ATP-BINDING CASSETTE (ABC) TRANSPORTERS AND LIPID FLIP-FLOP Delusion may be defined as a belief or an impression that is not real. What does that have to do with phospholipid (PL) translocation? Is the process considered not real, or does the delusion apply to the existence of specific proteins mediating the flippase activity? Exploring this issue is the focus of this chapter, and its contents will be restricted largely to the process of PL translocation (aka flippase or flip-flop activity). As discussed earlier in this book, cellular lipid bilayers display considerable asymmetry with respect to lipid composition between the two leaflets (see Chapter 3). How do lipid bilayers create distinct pools of PLs between the hemi-leaflets and how is the dynamic exchange mediated? The latter process is not a simple diffusion-like event as observed for drugs and fatty acids, due primarily to the large polar head group in PLs. Biochemical investigations have suggested that the rates of PL translocation (t1/2 ∼ hours or days) are several orders of magnitude slower than fatty acids. However, recent studies [1] suggest that the rates for “natural PLs” are more rapid than previously observed using PLs with bulky reporter groups (e.g., 7-nitrobenz-2-oxa-1,3diazol-4-yl (NBD) or tetramethyl-1-piperidinyloxyl (TEMPO)), although they

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

225

226

THE FLIPPASE DELUSION OF ABC PROTEINS

remain different to fatty acids and drugs. Bioinformatic investigations suggest that the molecular mechanism of non-protein-mediated PL flip-flop requires spontaneous formation of a water pore to accommodate movement of the polar moiety. It is also well established that PL translocation in natural membranes occurs at considerably faster rates than in synthetic, or non-proteincontaining, bilayers. Indeed, a number of studies have demonstrated that directed flip-flop activity in biomembranes has a protein component. Clearly, there are limitations to the trans-bilayer movement of PLs, but the process can occur. Therefore, lipid translocation per se cannot be a delusion. Is there a biological need for PL translocation? The idea that lipid bilayers simply serve as a barrier organelle and a “workbench” for proteins has long since been banished. Lipid bilayers are involved in a plethora of biological activities and all aspects of biophysical properties including movement restricted microdomains and asymmetries in composition are involved. For example, the translocation of phosphatidylserine (PS) to the outer leaflet of the bilayer occurs during senescence and apoptosis, ultimately instigating engulfment of cells by macrophages [2, 3]. PL flip-flop has also been intimately associated with the process of vesicle formation, which is essential for cellular processes such as endo- and exocytosis as well as endosomal transport pathways. The translocation of PLs between the leaflets of a bilayer may contribute to vesicle formation by either recruitment or mechanics. For example, a localized switch in PL distribution may promote the recruitment of proteins involved in the formation of vesicles (e.g., clathrin). Lipid microdomains such as “rafts” are well known to effectively concentrate interacting proteins, thereby facilitating functional pathways such as signaling. Alternatively, the localized redistribution of PLs between leaflets of a bilayer could cause deformation and curvature, which is an essential mechanical disruption during vesicle biogenesis. The advent of the human genome and the advances in gene technology have resulted in an increased awareness of membrane lipid transport processes as etiologic factors in a number of diseases. Early studies with knockout mice revealed that disruption of the ABCB4 (mdr2/MDR3) gene abolished biliary PL secretion. The accumulation of PLs in the liver resulted in inflammation and fibrosis. A similar chain of events has been implicated with a subset of human patients afflicted with familial intrahepatic cholestasis [4, 5]. This serious disease frequently progresses to cirrhosis and liver failure, with mutations in the ABCB4 gene an underlying cause. Polymorphic variations and mutations in ABCB4 and several other canalicular ABC transporters have been implicated as risk factors for cholestatic liver diseases [6] including the rare disease cholelithiasis, also associated with low biliary PL [7]. Other canalicular ABC transporters involved in bile formation include ABCB11 (BSEP) and ABCC2 (mrp2). Polymorphisms and altered function of these proteins have also been associated with high risk of developing hepatic and intestinal disorders linked to altered PL or bile acid secretion [6, 8] as well as increased susceptibility to drug-induced liver injury [9].

ATP-Binding Cassette (ABC) Transporters and Lipid Flip-Flop

227

The ABCA1 gene has been implicated in the final stages of apoptosis via mediation of PS translocation across the plasma membrane to enable clearing of the dying cell. Further evidence for a potential role in lipid movement was the finding that mutations in the ABCA1 gene cause a rare inherited disorder known as Tangier disease [10–13]. The function of ABCA1 appears to be efflux/translocation of membrane PLs (and potentially cholesterol) to lipiddepleted apoprotein acceptor particles. This activity is essential in the formation of high-density lipoprotein (HDL) particles and, therefore, the reverse cholesterol pathway. Mutations in the ABCG5 and ABCG8 proteins also affect lipid homeostasis, causing a rare inherited disease known as sitosterolemia [14]. The disease is characterized by elevated plasma levels of a plant cholesterol derivative known as sitosterol, in addition to a number of other phytosterols. The defect appears to be due to significantly reduced biliary secretion of sitosterol, suggesting that the ABC transporters are associated with outward secretion of the plant sterols. Patients with sitosterolemia display decreased cholesterol secretion and hypercholesterolemia, which agrees with the purported role for ABCG5/G8. Another member of the A subclass of ABC transporters, ABCA4, is also associated with a genetic disorder. Mutations in ABCA4 lead to development of Stargardt disease (STGD), which is characterized by loss of central vision that ultimately leads to legal blindness and is typically observed during childhood. Pathologically, the disease is associated with gradual deposition of the pigment lipofuscin caused by increased levels of A2-E, which is a toxic byproduct of the retinal regeneration cycle. A2-E is formed by the condensation of two molecules of all-trans-retinal with phosphatidylethanolamine (PE). The precise biochemical malfunctions underlying STGD have not been unequivocally determined. The working hypothesis is that mutations in ABCA4 prevent its ability to translocate N-retinylidene-PE ([NrPE] a single molecule of alltrans-retinal conjugated with PE) in the disk membranes of rod and cone cells. The X-linked adrenoleukodystrophy (X-ALD) is caused by perturbed lipid movement and is also associated with mutations in the gene for an ABC transporter, in this case ABCD1. The protein is localized to the peroxisomal membrane and is thought to mediate the movement of very long-chain fatty acids (VLCFA), or their CoA derivatives into peroxisomes. Loss of this activity in X-ALD prevents the peroxisomal oxidation of VLCFA, leading to increased cellular accumulation. Ultimately, this causes adrenal insufficiency and demyelination of neurones and, as a consequence, severe neurodegeneration. The preceding few paragraphs have demonstrated that perturbation or inhibition of lipid translocation pathways has severe and often fatal consequences. The movement of PLs, sterols, and even fatty acids across bilayers is thought to occur via a flip-flop mechanism. Clearly, there is no delusion for the serious consequences of disrupting flippase activity. The genetic basis for the disorders implicates a number of ABC transporters and indeed the function of these proteins is regarded as mediating lipid flip-flop. However, the

228

THE FLIPPASE DELUSION OF ABC PROTEINS

evidence supporting a flip-flop mechanism for these ABC proteins has not yet been shown using classical biochemical techniques. This lack of direct evidence lies at the heart of what may be deemed the “flippase delusion,” or does it? 11.2â•… ABCA4 AND LIPID TRANSLOCATION: EXPLAINING A PHENOTYPE? ABCA4, also known as ABCR or rim protein, is a 250-kDa glycoprotein that is localized to the disk membranes of both rod and cone photoreceptor cells in the eye [15–20]. As with many other ABC transporters, the importance of ABCA4 was discovered by the consequences of its deficiency. Mutations in the gene encoding for ABCA4 can lead to a variety of autosomal recessive retinal degenerative diseases including STGD [16], cone–rod dystrophy [21], retinitis pigmentosa [22], and age-related macular degeneration [23]. STGD is the most common form of autosomal recessive macular degeneration with an incidence of 1 in 10,000 [24]. The disease is characterized by a juvenile onset with a loss of central vision, progressive atrophy of the retinal pigment epithelium (RPE), and accumulation of lipofuscin deposits in the RPE that can be seen as yellow or white flecks by ophthalmoscopy [25, 26]. To understand more about the biochemical malfunction underlying STGD, one needs to know more about the visual cycle (Fig. 11.1) [27]. It is believed that in the visual cycle, the chromophore of rhodopsin, 11-cisretinal, is isomerized into all-trans-retinal upon exposure to light [28]. Most of

Light

all-t-rol

all-t-rol

RDH all-t-ral all-t-ral

11-c-ral

PE Rhodopsin

PE

all-t-rol

Ret-PE

+

ATP

Ret-PE ABCA4

11-c-ral

Rhodopsin

11-c-ral

11-c-ral

Opsin

Rod outer segment

RPE Cell

Figure 11.1.â•‡ Biochemical aspects of the visual cycle in rod cells. For details of the process, see the text [27]. Color version on the Wiley web site.

ABCA4 and Lipid Translocation: Explaining a Phenotype

229

the all-trans-retinal remains noncovalently bound to the “exit site” of rhodopsin, where it can be reduced to all-trans-retinol by retinol dehydrogenase (RDH) and subsequently diffuse out of the photoreceptor cell [29–31]. Alternatively, all-trans-retinal can be released to the lipid membrane before being reduced, presumably to the inner leaflet of the disk membrane where all-trans-retinal is not accessible for RDH-mediated reduction [31]. There it can react with PE, forming the Schiff-base NrPE that cannot freely cross the membrane [32]. Subsequently, a second all-trans-retinal can react, forming A2-PE [33]. This is converted into the condensation product N-retinylideneN-retinylethanolamine (A2-E), which is the major component of lipofuscin [34]. There are no enzymes that can degrade A2-PE or A2-E, and they accumulate in the RPE. This leads to perturbation of cell membranes and, ultimately, death of the RPE cells. Insights into the pathogenesis of STGD have come from studies in knockout mice lacking a functional ABCA4 gene. Like ABCA4-deficient humans, ABCA4 knockout mice showed a light-dependent increase of all-trans-retinal, NrPE, and PE in photoreceptor outer segments and accumulation of A2-E in the RPE [35–37]. Biochemical studies on immunopurified and reconstituted ABCA4 demonstrated that the retinoid all-trans-retinal, 11-cis-retinal, 13-cisretinal, and NrPE were able to stimulate the ATPase activity of the protein [38, 39]. In fact, stimulation of the ATPase activity of ABCA4 depended on the presence of PE in the system [39]. From these data, Weng et al. [35] proposed the working hypothesis that the role of ABCA4 is to translocate or “flip” NrPE out of the disk, thereby preventing the formation of A2-PE and A2-E, and ultimately the accumulation of lipofuscin in the RPE. Further support for this theory has come from a study by Beharry et al. [40] in which the binding of various retinoids to immunoaffinity-purified ABCA4 was examined using both high-performance liquid chromatography (HPLC) and radiolabeling techniques. It was found that ABCA4 could bind NrPE upon addition of alltrans-retinal in the presence of PE. ATP binding and hydrolysis promoted the dissociation of NrPE from ABCA4, indicating a role for ABCA4 in the transport of NrPE, although transport itself was not measured in this study. This brings us to the crux of the “flippase delusion”: All the aforementioned evidence for ABCA4 as a lipid flippase is indirect. So far, no direct transport assay for ABCA4 has been developed. Measuring lipid translocation is not straightforward, but in the case of ABCA4, it is even more complicated. The protein is only found in the disk membranes of the very specialized rod and cone photoreceptor cells for which there is currently no cell line available that could potentially be used for a transport assay. In transfected cells, ABCA4 does not localize to the plasma membrane but to intracellular vesicles and the endoplasmic reticulum (ER)–Golgi network [41]. Another, more general problem associated with measuring lipid translocation is, of course, the delivery of the substrate. NrPE is very hydrophobic and therefore insoluble in water. Even if one could use a synthetic analog of the lipid designed to enhance its water solubility, this may no longer behave the same as the natural lipid

230

THE FLIPPASE DELUSION OF ABC PROTEINS

[42]. In other words, if ABCA4 is shown to transport a synthetic analog of NrPE, that does not necessarily mean it also transports the natural lipid. Another important question we should ask is whether we need a flippase to move the NrPE to the outer leaflet of the disk membrane. As stated above, it is only presumed that a proportion of the all-trans-retinal released from rhodopsin ends up in the inner (luminal) leaflet of the disk membrane rather than the outer (cytosolic) leaflet. There is no proof for this, but it does explain how all-trans-retinal evades reduction by RDH in the first place. If there is another explanation, and the all-trans-retinal does react with PE in the outer leaflet, transport of the formed NrPE would no longer be needed because it would already be on the correct side of the membrane. If we assume that the current hypothesis for flippase activity of ABCA4 in the visual cycle is correct, we are faced with another conundrum: The direction of transport is from the lumen to the cytoplasmic side of the disk membrane. This suggests that ABCA4 is an inwardly directed “flippase” or importer [27], whereas all other eukaryotic ABC transporters studied so far, including the closely related ABCA1 (52% homologous) [43], are suggested to be exporters. It is increasingly clear that there are significant structural and mechanistic differences between importers and exporters of the ABC superfamily [44–46]. If it proves to be unique among the eukaryotic ABC transporters in functioning as an import protein, it is of greatest importance to improve our understanding of ABCA4. Clearly, a great deal more mechanistic information on the activity, molecular mechanism, and role of ABCA4 in the visual cycle is required. Currently, with even its physiological substrate a matter of uncertainty, assigning ABCA4 as a lipid flippase seems to be a hasty decision. 11.3â•… MsbA AND LIPID TRANSLOCATION: A KEY TO SURVIVAL The protein MsbA was first identified as an essential gene in Escherichia coli, and bacteria with mutated MsbA have major defects in lipid trafficking. Consequently, they lack the ability to build and maintain the cellular membranes, which leads to aberrant morphology and restricted growth. The cellular role of MsbA is linked to the assembly of the cell wall of gram-negative bacteria. In gram-negative bacteria, the cell wall consists of an inner membrane (IM) and an outer membrane (OM) separated by a periplasmic region (Fig. 11.2). The IM is a classical PL-based membrane and the location of many vital bioenergetic processes. The OM regulates the passage of substrates between the cell and an often-hostile environment. This membrane is heavily asymmetric, with lipopolysaccharide (LPS) constituting the majority of its outer leaflet. LPS forms a permeability barrier for the cell and protects gramnegative bacteria against the mammalian immune system [47, 48].

231

MsbA AND LIPID TRANSLOCATION: A KEY TO SURVIVAL Polysaccharide Outer membrane

Lipid A

Porin

LPS

Phospholipids

Periplasm

Peptidoglycan

Inner membrane

Phospholipids Membrane proteins

Figure 11.2.â•‡ Structure of the gram-negative bacterial cell membrane. Color version on the Wiley web site.

Biosynthesis of LPS is a complex process, and the molecule consists of three main components: lipid A, a core oligosaccharide, and the O-antigen (Fig. 11.3a). Components are synthesized at different locations, followed by stepwise assembly that is punctuated by transport across both the IM and OM (Fig. 11.3b). Several gene mutations have been associated with, or implicated in causing defects in the LPS biosynthetic pathway. For example, the absence of a gene called htrB has been associated with inadequate formation of the cell wall, leading to morphological abnormalities, changes in individual cells (Fig. 11.4), and temperature-sensitive growth. At 32°C, htrB-null mutant bacteria show normal growth patterns, but above this, temperature growth ceases, bulges appear in the membranes, and incompletely processed LPS accumulates in the IM. A plasmid-containing wild-type msbA can restore normal growth by complementing the htrB-null mutant phenotype [50]. Analysis of the putative protein-encoding section of this gene correctly predicted a 64.5-kDa membranespanning protein. Expression of radioactively labeled MsbA in E. coli, followed by subcellular fractionation, demonstrated that the protein localized to the IM [51]. Further scrutiny of the protein sequence revealed that MsbA contains the signature sequence of the ABC transporter superfamily, and has homology to other members of the family: MsbA has 30% identity and 46% similarity to ABCB1, a human multidrug transporter [8]. In other words, 30% of the amino acid residues in MsbA are identical to those in ABCB1, and a further 16% of the residues have similar side-chain properties. Genetic studies had already linked MsbA to LPS metabolism, but its homology to the ABC transporters necessitated revision of the hypothesis to a role in translocation rather than processing of LPS. In turn, this spawned the idea that MsbA could act as a flippase of LPS or its precursor lipids [51]. The function of MsbA has now been examined in considerable detail. Early studies used radioactively labeled lipids to demonstrate the trafficking of LPS between the IM and OM [52]. htrB-deficient cells grown on radioactive medium were

232

THE FLIPPASE DELUSION OF ABC PROTEINS

(a)

Precursors including GlcNAc, hydroxymyristoyl-ACP

Disaccharide-1-phosphate

Lipid IVa

HtrB (lauroyltransferase) MsbB (myristoyltransferase) Kdo2-lipid A

Lipid A eg Ra-lipid A; Re-lipid A

N-acetyl glucosamine (GlcNAc)

Kdo2-IVa

C12–C14 acyl chain

Outer core oligosaccharide

2-keto-3-deoxyoctonic acid (Kdo)

Phosphate group

(b) LPS

OM

RfaL ligase

MsbA? Lipid A Kdo O-antigen

IM

Kdo-lipid A

Outer core

Figure 11.3.â•‡ Assembly of lipopolysaccharide (LPS). An overview of the assembly and trafficking of LPS showing (a) the intermediates and some enzymes involved in lipid A biosynthesis and (b) the assembly of LPS from its components parts, lipid A, core oligosaccharides, and O-antigen. Color version on the Wiley web site.

MsbA AND LIPID TRANSLOCATION: A KEY TO SURVIVAL

233

WT

htrB

Figure 11.4.â•‡ Morphological differences between wild-type (WT) and htrB-null cells (from Reference 49).

transfected with an msbA plasmid. Following a period of growth, the IM and OM were separated using sucrose density gradient ultracentrifugation. The lipid composition of each membrane was analyzed by thin-layer chromatography, and by measuring the radioactivity in each lipid fraction, the distribution of lipids between the two membranes was quantified. The results demonstrated that the presence of MsbA did not alter the acylation state of the lipids but increased the amount of lipid IVa (Fig. 11.3) that reached the OM. Conversely, decreasing the msbA expression led to the accumulation of lipid IVa at the IM. Another study investigated the differences between wild-type MsbA and the temperature-sensitive mutant WD2. This mutant contains an A270T substitution in the fifth transmembrane helix (TM5) of MsbA and displays defective lipid-transportation properties [53]. Cells with the WD2 mutant form of MsbA grow normally at 30°C, but at 44°C (the “nonpermissive” temperature), MsbA function was rapidly depleted and growth of cells ceased. Doerrler et al. [53] used radioactive labeling of lipids to show that ∼90% of lipid transport ceased in WD2 cells at the non permissive temperature. Folds and invaginations of the IM as a result of lipid accumulation at the IM were visible by electron microscopy [54]. Both studies provided further evidence for the role of MsbA as a lipid transporter, rather than an intermediate in the biosynthetic pathway. A number of subsequent investigations have focused on molecular aspects of MsbA function. For example, the ATPase activity of MsbA was characterized in detail by Doerrler and Raetz [55], using MsbA purified and reconstituted into proteoliposomes. This study showed a basal rate of ATP hydrolysis (i.e., in the absence of the transport substrate) of 37â•›nmol ATP/min/mg protein, which is within the broad range reported for other ABC transporters including ABCB1. This study also examined the stimulation of ATPase activity by a range of potential lipid-based substrates. 2-Keto-3-deoxyoctonic acid

234

THE FLIPPASE DELUSION OF ABC PROTEINS

(Kdo)-lipid A (Fig. 11.3) stimulated the highest rate of ATPase activity at 154â•›nmol/min/mg protein, a fourfold increase from the basal rate. The stimulation of ATPase activity by a number of compounds implies that MsbA does not have strict selectivity for its transport substrate, but may transport a range of LPS precursors. A direct assay has recently been reported that explores the potential for MsbA to mediate LPS/precursor translocation [56]. The assay used purified, reconstituted MsbA to measure the transbilayer movement of NBD-labeled analogs of the lipid species. The major findings from this study were that lipid was transported in an ATP-dependent manner. Moreover, translocation was sensitive to vanadate, which has been observed in other ABC transporters and reinforces the conclusion that the observed translocation of lipids was MsbA dependent. A higher rate of translocation was observed for lipids with the NBD group attached to the acyl chain rather than the head group, which suggested that head-group recognition is likely to be vital for substrate recognition. This would agree with the observation that MsbA can transport lipid IVa in bacterial cells [52]. Furthermore, Eckford and Sharom [56] suggested that the binding cavity is of sufficient size to bind the large head group of lipid A, while the acyl tails may be dragged through the lipid bilayer, rendering the acylation state of minimal importance in substrate recognition. However, NBD-labeled LPS is not available commercially, and consequently, these studies were carried out using NBD-PE as the lipid substrate for transport. Though it has previously been shown that E. coli MsbA can transport PLs, this ability is not conserved between species [57], and as discussed, most of the evidence is for LPS or a derivative as the physiological substrate of MsbA. To temper this criticism, the authors showed that the addition of LPS to the system reduced the rate of NBD-PE transport, hinting at competition between NBDphosphatidylcholine (PC) and LPS for transport by MsbA. This is a plausible explanation for these observations, but it cannot be ruled out that LPS interacts directly with the lipid vesicle or micelle to perturb the activity of MsbA independently of its interaction with the protein itself. It is well established that the composition of the membrane, liposome, or detergent micelle in which a protein resides has frequently been shown to have an effect on its activity [58, 59], and this is a major technical difficulty in investigating flippase activity. However, at present, it is difficult to see how this problem can be circumvented. The suggestion of polyspecificity in the binding site of MsbA, intimated by its ability to transport multiple derivatives of LPS as well as LPS, has been elaborated with the observation that it may also bind and transport drugs. This observation was partly supported by the sequence similarity of MsbA to the multidrug efflux pump ABCB1. The drug-binding properties of MsbA have been investigated, although few firm conclusions have emerged: One study showed that daunorubicin, vinblastine, and verapamil had no effect on the activity of MsbA [60], and another demonstrated that daunorubicin and vin-

235

MsbA AND LIPID TRANSLOCATION: A KEY TO SURVIVAL

blastine stimulated its activity [61]. In gram-positive cells, which lack LPS and an OM, heterologous expression of MsbA can substitute for LmrA, another homolog of ABCB1 and a known multidrug ABC transporter. These partly conflicting findings mean that the issue remains unclear [62]. Drug, lipid, and nucleotide binding have been further characterized by tryptophan fluorescence quenching experiments [62]. Tryptophan fluorescence quenching relies on detecting the intrinsic fluorescence of Trp residues in the protein. Substrate binding and conformational changes alter or “quench” this fluorescence by direct or allosteric effects. These spectral changes can be used to estimate the affinity of the substrate or nucleotide for the protein. This study reported an apparent binding affinity of 6.4â•›µM for both LPS and lipid A, though the data suggested that relatively minor conformational change occurs upon substrate binding. MsbA has also been extensively characterized by electron paramagnetic resonance (EPR) spectroscopy to follow conformational changes in the protein during its ATP hydrolysis cycle [63–65]. Addition of LPS to the system altered the degree of conformational motion in MsbA during ATP hydrolysis. These conformational changes instigated during ATP hydrolysis caused alterations in the accessibility of residues to the aqueous and lipidic phases. The EPR approach enables the investigation of the dynamic structure of protein that is reconstituted in a membrane-like environment. By contrast, X-ray crystallographic data provide a snapshot of the protein in a defined conformation. There are now three published X-ray crystallographic structures of MsbA [46]. Two represent the protein in an inward-facing nucleotide-free state, while the third is trapped in an outward-facing posthydrolytic state with nucleotide bound (Fig. 11.5). Unfortunately, there has been considerable controversy surrounding the structures of MsbA. Initially, errors in data processing led to the publication (a)

EL3

(b)

EL1

EL2

EL2 TM1 TM5 TM4

IH2

TM6

TM3 Elbow Helix

IH1

(c)

EL3 EL1

TM2

TM5 TM4

TM3

EL3 EL2

TM2

TM1

TM6

TM3 TM5 TM4 TM6

Elbow Helix

IH2

IH1

EL1 TM2

TM1

Elbow Helix

IH2

IH1

Figure 11.5.â•‡ Structures of MsbA based on crystallographic data. There are two inward facing, nucleotide-free states: “open apo” (a) and “closed apo” (b), and one outwardfacing nucleotide-bound state (c). Color version on the Wiley web site.

236

THE FLIPPASE DELUSION OF ABC PROTEINS

of physiologically improbable and, as it transpired, incorrect structures [66, 67]. With these problems resolved, the structures were rereleased [46] and have been examined in some detail [63–65, 68]. However, concerns remained regarding both the physiological relevance of the conformations that are represented in these structures and their resolution; ranging from 3.7 to 5.2â•›Å, this must be called medium, rather than high, resolution. With these caveats in mind, what conclusions can be drawn from the structural information available for MsbA? Of relevance here are two questions: what can these structures tell us about the mechanism of MsbA and can they clarify its status as a lipid flippase? A striking feature of these structures is the size of the central chamber, the putative substrate binding site, which is particularly apparent in the “open apo” state. This large chamber may account for the tendency of MsbA, like its homolog ABCB1, toward polyspecificty. MsbA appears to be able to transport lipid A at several stages of its synthesis, which manifests in its ability to complement HtrB deficiency. It may also explain its ability to transport a variety of drugs: A large binding chamber is necessary to transport a large substrate; however, the increased dimensions of the LPS-binding site may have the added effect of providing more potential sites for interaction of the protein with endo- or xenobiotics. The abundance of ATP in the cytoplasm suggests that the nucleotide-free conformations are unlikely to exist in cells. However, EPR data do support the existence of these conformations of MsbA. The “open apo” state shows the nucleotide-binding domains widely spaced and the transmembrane domains (TMDs) in a wide V shape and is thus inferred to represent the inward facing conformation. This hints that the substrate-binding chamber is available to both the cytoplasm and the inner leaflet of the IM. The alternative inward-facing conformation is referred to as the “closed apo” state, and this also displays sufficient flexibility to allow substrate entry, despite close contact between the nucleotide-binding domains. As discussed above, MsbA has now been extensively characterized using the powerful combination of EPR spectroscopy and X-ray crystallography. Zou et al. [64, 65] interpreted the crystallographic and EPR data to suggest a mechanism of action that supports the alternating access hypothesis that has been advanced for other ABC transporters. In short, ATP binding leads to closure of the transmembrane (TM) chamber, manifested as reduced accessibility of these residues to water. The change in polarity of the chamber promotes translocation of the substrate toward more hydrated regions, presumably on the periplasmic side of the protein. This results in movement of the bound substrate across the membrane. Unfortunately, this does not help us to classify the mechanism as either translocation or flipping of lipids, as substrate may access the protein from either the cytoplasm or the inner leaflet of the membrane. In summary, several lines of evidence including extensive genetic characterization of MsbA-mutant cell lines, lipid transport measurements, and increasingly detailed analyses of conformational change are converging to

Drug and Lipid Movement by ABCB1

237

support the hypothesized flippase mechanism of MsbA. It remains unclear what, if any, is the preferred substrate for MsbA, due to its stimulation by many different lipids, and some drugs. Nonetheless, the combination of genetic, biochemical, and structural data make this perhaps the most compelling case for flippase activity in the ABC transporter family. 11.4â•… DRUG AND LIPID MOVEMENT BY ABCB1: IS THE MECHANISM A FLIP-FLOP? ABCB1, also known as P-glycoprotein or MDR1, is a 170-kDa glycoprotein linked to multidrug resistance (MDR) in cancer. Unlike other ABC transporters, which tend to be relatively specific for their substrates, ABCB1 has a surprisingly broad spectrum of amphiphilic substrates. It was suggested by Higgins and Gottesman that this broad substrate specificity of ABCB1 could be explained by the flippase model [69]. In this model, the primary determinant of specificity is the ability of a substrate to intercalate into the bilayer, where ABCB1 can then act as a flippase, not only for amphiphilic drugs but also for various endogenous PLs or sphingolipids [69]. It was soon discovered that the closely related (over 75% sequence homology) liver transporter ABCB4 was required for PC secretion in mouse bile [70] and that it enhanced transport of newly synthesized radiolabeled PC to the surface of transgenic mouse fibroblasts [71]. This was in agreement with a study by Ruetz and Gros [72] on secretory vesicles (SVs) from yeast that express Mdr2, the mouse homolog of ABCB4. They showed that Mdr2 expression enhanced the translocation of C6-NBD-PC, a short-chain fluorescent PC analog containing NBD, to the inner leaflet of the membrane in an ATP- and Mg2+-dependent fashion. Both studies implicate ABCB4 in lipid translocation. This is discussed further in the subsequent section and also raises the possibility that its close relative ABCB1 may play a role in membrane lipid movement. To address the issue of lipid translocation, a number of studies have used short-chain NBD-labeled PL analogs to study lipid translocation [73–78]. In studies using SVs, crude membrane vesicles, or proteoliposomes, the fluorescence of the NBD group can be chemically quenched by dithionite. Dithionite is membrane impermeable and will reduce only the NBD moiety of lipids present in the outer leaflet. Upon detergent disruption of the membranes, the fluorescent emission of the remaining NBD lipids will be destroyed, and this further reduction in fluorescence reflects the amount of NBD lipid associated with the inner leaflet of the bilayer. This method is called the “dithionite quenching” technique [79, 80] (see Chapters 1 and 6). In studies using whole cells, the NBD lipids can be extracted from the outer leaflet of the membrane using albumin (bovine serum albumin [BSA]) or empty lipid vesicles as acceptors for the probe lipid. This procedure is called “back-extraction” [81, 82], and in combination with “dithionite quenching,” both have been used to examine lipid flippase and/or translocation.

238

THE FLIPPASE DELUSION OF ABC PROTEINS

Van Helvoort et al. were the first to demonstrate translocation of NBD lipids by ABCB1 [83]. They used polarized pig kidney epithelial cells transfected with ABCB1 and recovered newly synthesized short-chain analogs of various membrane lipids in the apical medium using “back-extraction.” Cells expressing ABCB1 showed greater translocation of C6-NBD-PC, PE, glucosylceramide (GlcCer), and sphingomyelin (SM). The authors also demonstrated translocation of the lipid analogs C8C8-[3H]PC, C8C8-[14C]PE, and C8C8-[3H]GlcCer. Although these radiolabeled analogs lack the NBD moiety and are, in that respect, more natural versions of these lipids, they also have two short acyl chains, making them more water soluble than endogenous C16 lipids [83]. In a later study, the same group also demonstrated translocation of the lipid analogs C6-NBD-SM, [14C]C6-GlcCer, and C6-NBD-GlcCer in a variety of cell lines [73]. Lipid translocation was blocked by cyclosporine A and PSC 833, both inhibitors of ABCB1, suggesting that this protein was indeed responsible for the translocation. Pohl et al. [77] used human gastric carcinoma cells overexpressing ABCB1 to demonstrate outward translocation of newly synthesized C6-NBD-PC/PE/SM in addition to inward transport of C6-NBD-PS. Transport was reduced in the presence of the ABCB1 inhibitors PSC 833, cyclosporine A, and dexniguldipine. The aforementioned studies all implicate ABCB1 in lipid transport in intact cells; however, the translocation cannot be directly attributed to ABCB1. Reconstituted systems on the other hand should provide a system for directly examining lipid transport by a specific protein. Unfortunately, studies with purified and reconstituted ABCB1 have yielded contradicting results. Romsicki and Sharom [76] used the “dithionite quenching” technique to show that ABCB1 reconstituted into proteoliposomes could translocate a wide variety of long, short, saturated, and unsaturated NBD lipids from the outer to the inner leaflet of the bilayer. They found an ATP-dependent increase in the reorientation of analogs of PC, PE, PS, and SM that could be inhibited by vanadate, verapamil, cyclosporine A, and vinblastine. However, the amount of lipid translocated represented a very small fraction of the total. Rothnie et al. [59] used dithionite quenching of NBD lipids as well as EPR of spin-labeled lipids to assess translocation of various lipids by ABCB1. Although they found a low increase in transport of PC, PE, and ceramide in the presence of ABCB1, this was not dependent on ATP hydrolysis by the protein. The authors suggested that there are technical limitations when using proteoliposomes with an average diameter of 200â•›nm. Inward transport of lipid results in a buildup of lipids in the inner leaflet of the membrane, which creates high surface tension. This tension would likely preclude further translocation of lipids. Unlike cells or giant liposomes, these relatively small proteoliposomes cannot produce effects such as membrane bending or endocytosis to compensate for altered surface tension [84]. Despite the studies above, it remains unclear whether ABCB1 can transport natural membrane lipids. van Helvoort et al. [83] pointed out that, although the ABCB1 mouse homolog Mdr1a could translocate C6-NBD-PC, its inabil-

Drug and Lipid Movement by ABCB1

239

ity to restore the transport of PC into the bile of mice lacking Mdr2 (homologous to human ABCB4) [70] suggests that natural long-chain PC is not an ABCB1 substrate. In support, Kalin et al. [85] demonstrated that in contrast to C6-NBD-PC, translocation of [14C]PC in human erythrocytes was insensitive to the specific ABCB1 inhibitor PSC 833. However, to further confuse the issue, they also showed reduced translocation of natural PC in erythrocytes from ABCB1 knockout mice, although perhaps this discrepancy reflects differences between human and mouse erythrocytes. What about other lipid types such as long-chain PE and PS? The aminophospholipid translocase ensures that PE and PS remain in the inner leaflet of the plasma membrane. Outward translocation of these lipids by ABCB1 would appear to produce a futile ATP-utilizing cycle. However, it has been demonstrated by Pohl et al. [77] that the presence of ABCB1 correlated with enhanced exposure of endogenous PS on the cell surface of ABCB1 overexpressing cells and that this could be reduced by specific inhibitors of ABCB1. This major issue also remains unexplained. Possibly, the inhibitors of ABCB1 are not fully selective to this protein or the ABC transporter only displays activity under specific cellular conditions. As mentioned earlier, one limitation of using synthetic lipid derivatives is the possibility that they do not have the same behavior as natural lipids. A good example of this comes from studies on GlcCer transport by ABCB1. Whereas ABCB1 translocated several derivatives of GlcCer in transfected cells [73, 75, 83] as well as proteoliposomes [78], it was recently shown by Halter et al. [42] that it could not translocate natural GlcCer. The study used mouse fibroblast cell lines derived from the ABCB1a/ABCB1b/ABCC1 triple knockout mouse (TKO) [86] and compared them with TKO cells stably transfected with human ABCB1. Natural GlcCer was metabolically radiolabeled and extracted from the cell surface using the glycolipid transfer protein. Whereas the transport of C6-NBD-GlcCer was strongly reduced in TKO cells and could be partially restored by transfection with ABCB1, knockout of the multidrug transporters had no effect on the transport of natural GlcCer. In addition, transfection of TKO cells with ABCB1 did not increase the rate of natural GlcCer translocation. Even though short-chain lipid derivatives may not always reflect the physiological behavior of natural lipids, does that render studies with them irrelevant? Endogenous short-chain lipids occur in living cells and could be substrates for ABCB1. One example is the lipid platelet-activating factor (PAF), a naturally occurring short-chain PC (1-O-alkyl-2-acetyl-sn-glycero-3PC), which is synthesized in many cell types. In a study by Ernest and BelloReuss [87], it was shown that PAF inhibited ABCB1-mediated transport of rhodamine 123 in multidrug-resistant KBV-1 cells and in human mesangial cells. Similarly, inhibitors of ABCB1 (verapamil, cyclosporine A, PSC 833) blocked the secretion of endogenous PAF from cultured human mesangial cells and downregulation of ABCB1 expression also inhibited PAF secretion. The authors concluded that ABCB1 transports endogenously produced PAF

240

THE FLIPPASE DELUSION OF ABC PROTEINS

in cultured human mesangial cells. However, transport was measured as the absolute amount of PAF secreted into the medium but did not take into account the rate of PAF synthesis. It is possible that reduced synthesis may have contributed to the decrease in PAF secretion and that ABCB1 may play a role in the transport of precursors or cofactors. In support of a role for ABCB1 in the transport of PAF is the study by Raggers et al. [88]. The study assessed the transport of endogenous radiolabeled [14C]PAF from polarized kidney epithelial cells using the “back-extraction” technique. ABCB1 accelerated PAF transport, which was independent of vesicular traffic and could be inhibited by the ABCB1 inhibitors PSC 833 and cyclosporine A. Since ABCB1 is able to translocate the natural short-chain lipid PAF, as well as a variety of short-chain PL analogs, it may also recognize naturally occurring short-chain PLs as well. Potential candidates could be oxidatively fragmented PLs, lysophospholipids, or short-chain glycosphingolipids. This hypothesis needs further examination and provides a way to recognize differences in the ability of ABCB1 to translocate short- and long-chain lipids. Is the mechanism of drug and lipid translocation by ABCB1 really akin to flippase activity? Does it move hydrophobic substrates between hemi-leaflets of a membrane without desorption (i.e., extraction to the aqueous phase)? Short-chain lipid analogs require less activation energy for desorption [89] than leaflet-to-leaflet movement, thereby arguing against the need for a flippase mechanism. Another interesting issue is whether ABCB1 is primarily a drug transporter that happens to move some lipids too, perhaps even as a consequence of bilayer disruption caused by drug movement. Although progress has been made in identifying the lipid substrates of ABCB1, the transport process itself is undefined. Until then, how transport occurs in mechanistic terms remains an open question, and assigning a flippase activity is surely within the realms of delusion. 11.5â•… ABCB4: THE FORGOTTEN AND LIKELY LIPID FLIPPASE? The human gene MDR3 was first cloned from a human cDNA library in 1987 [90]. The ABCB4 protein it encodes is expressed primarily in hepatocytes and has a homology of 70% to ABCB1. Based on its sequence homology to ABCB1, ABCB4 was initially assumed to play a role in MDR phenotypes and was regarded as a functional homolog of ABCB1. However, early experiments failed to demonstrate drug transport [91–94], which led to a more detailed examination of the protein. As with many other ABC transporters, insight into the function of this protein has come from understanding its role in disease. A number of mutations in ABCB4, including nonsense and deletion mutations, have been associated with a rare disease of the liver characterized by impaired bile production [4, 95]. Known as progressive familial intrahepatic cholestasis (PFIC), the disease is classified into three subtypes caused by defects at different stages

241

ABCB4: The Forgotten and Likely Lipid Flippase

of bile production (for a review, see Reference 96). The PFIC3 subtype (or MDR3 disease) has been associated with mutations in the ABCB4 protein. The genetics of PFIC3 are complex and have been investigated in some detail [97, 98]: Different mutations give rise to a broad spectrum of disease, but symptoms commonly include pruritus (itching), gallstones, jaundice, and hepatomegaly, leading ultimately to liver failure in children [97]. In one study, the mean age for the occurrence of liver failure was found to be 7.5 years, although the onset of liver failure is thought to be linked to the severity of the mutations in ABCB4. In many cases, liver transplantation is the only successful treatment [96, 97]. Less serious mutations to ABCB4 are thought to contribute to a condition called intrahepatic cholestasis of pregnancy, which, though temporary, is linked to increased incidence of prematurity and stillbirth [99]. PFIC is a disease of bile production, and in healthy individuals, bile consists of a mixture of bile acids, PLs, organic ions, and cholesterol (Fig. 11.6). Bile acids are biological surfactants and form mixed micelles with lipids. In individuals affected by PFIC, bile lacks both PLs and cholesterol. The resulting high proportion of free bile acids (in the bile) are thought to disrupt the membranes of cells lining the canaliculus [97, 100], giving rise to the clinical symptoms described above.

Bile micelle

Canalicular membrane ABCB4

ABCB11

Bile acid

Cholesterol

ABCG5/G8

PC

Other PL

Figure 11.6.â•‡ Bile production in the bile canaliculus (adapted from Reference 101). Phosphatidylcholine (PC) is transported to the outer leaflet of the canalicular membrane by ABCB4, bile salts by ABCB11, and cholesterol by ABCG5/G8. Bile acids extract the other components from the membrane to form mixed micelles. Defects in any of the transporters impair bile formation and may cause PFIC. Color version on the Wiley web site.

242

THE FLIPPASE DELUSION OF ABC PROTEINS

Experiments using knockout mice shed further light on the role of ABCB4. Mice possess an ortholog of MDR3, named Mdr2 [92]. Knocking out this gene produced mice with liver disease displaying histological and biological parameters characteristic of PFIC3 [4, 5, 70]. Bile samples from Mdr2(–/–) mice lack PC, leading to the conclusion that Mdr2 had a role in the secretion of PLs from hepatocytes to produce bile. Moreover, transfection of Mdr2(–/–) mice with human MDR3 prevented the development of liver disease [102] and suggested functional equivalence between mouse Mdr2 and human ABCB4. This functional equivalence has enabled the Mdr2-deficient mouse to be used as a model for human ABCB4 deficiencies. The association between ABCB4 deficiency and PFIC, and the experiments with knockout mice led to the hypothesis that ABCB4 may actively transport PC. Evidence from genetic studies was supported by biochemical investigations demonstrating the ability of ABCB4 to bind and translocate PLs. A yeast cell line with a defect in vesicular fusion was used to overexpress Mdr2 [72, 103, 104]. The defective yeast cells accumulated large numbers of yeast secretory vesicles (YSVs) enriched with Mdr2. The orientation of the protein in the YSV membrane was such that any transported substrate accumulated in the lumen of the vesicle. The short-chain lipid analog C6-NBD-PC was used as the substrate in these ABCB4-rich YSVs. The time-dependent partitioning of lipid between the IM and OM leaflets was measured using the “dithionite quenching” assay. The data indicate that ABCB4, like ABCB1, could increase the trans-bilayer translocation of the short-chain PC-lipid analog. As detailed in the previous section, experiments with short-chain lipids may provide ambiguous findings. With this in mind, a second in vitro assay measured long-chain PC transport using transgenic murine fibroblasts expressing Mdr2 (Fig. 11.7) [71]. Fibroblasts were incubated with [14C]choline, resulting in the synthesis of [14C]PC inside the cell. Lipid trafficking from the cytoplasm led to the insertion of [14C]PC into the IM, and its subsequent

(a)

(b)

S

MOR

[14C]choline

MO

RS

ER cytoplasm 14

PM

PC-TP Liposome in medium

PM

Figure 11.7.â•‡ [ C]choline labeling assay [105]. PM, plasma membrane; MORS, muopioid receptor system; PC-TP, phosphatidyl choline transfer protein.

ABCB4: The Forgotten and Likely Lipid Flippase

243

accumulation in the outer leaflet was measured (Fig. 11.7). The difference in accumulation of [14C]PC in the outer leaflet of wild-type fibroblasts, which do not express a PC flippase, and transgenic fibroblasts expressing Mdr2, was used to quantify flipping of PC between the hemi-leaflets. Transgenic cells did indeed display greater transfer of [14C]PC to the acceptor liposomes, presumably reflecting greater translocation of the lipid to the outer leaflet of the bilayer by Mdr2. Despite early indications that it did not function as a drug transporter, more recent assays using polarized cells transfected with MDR3 revealed directional transport of digoxin, vinblastine, and paclitaxel, well-known substrates for ABCB1 [106]. These drugs were shown to directly interact with ABCB4 and inhibit vanadate-dependent nucleotide trapping of the protein. Some of these drugs, including paclitaxel and vinblastine, as well as known inhibitors of ABCB1, prevented the transport of PC analogs by ABCB4. The transport of drugs, even at low levels, presents a more complicated picture of the physiological role of ABCB4 than first envisaged and raises the possibility that it may contribute to MDR [106], or at least retain the ability to transport some compounds other than PC. However, the low rates of drug transport suggest that PC is the preferred substrate. ABCB4 is unlikely to make a significant contribution to the MDR phenotype and the high levels of PC in membranes in vivo would outcompete drugs for binding to ABCB4. A more important issue is resolving the question of the mechanism by which ABCB4 transports its allocrites. Flippase activity of ABCB4 was first postulated by Smit et al. [70], based on their observations of Mdr2-knockout mice and the hypothesized flippase activity of ABCB1 [69]. Evidence from the in vitro models of ABCB4 activity lends weight to this hypothesis, and ABCB4 is now widely regarded as a PC flippase [71, 107]. However, proving the socalled flippase activity requires more than just proof of lipid movement across the bilayer [108–110]. This is an issue at the heart of the flippase delusion: Whether we define lipid flipping and lipid translocation as different processes, and if so, can we overcome the technical difficulties in experimentally distinguishing them? Recent research efforts into ABCB4 have concentrated on identifying the genetic basis of PFIC in humans [95, 96, 111]. Though this information is useful for the diagnosis of disease, overlooking the mechanistic aspects of the protein has led to the neglect of what should be a promising lead in the search for an ABC flippase. Detailed structural data showing lipids bound at different stages of the transport cycle may allow the substrates, binding sites, and transport mechanism of ABCB4 to be described with more confidence. Recent advance in the structural biology of ABC transporters [45, 46, 112, 113] renders this an ambitious but achievable goal. What can be said with certainty is that ABCB4 is highly specific for the transport of PC, and this compound is translocated unidirectionally from the inner to the outer leaflet of the cell membrane. Having met these basic criteria, if a true ABC lipid flippase exists, ABCB4 must be a likely candidate.

244

THE FLIPPASE DELUSION OF ABC PROTEINS

11.6â•… CONCLUSIONS AND PERSPECTIVES There is a vast amount of data demonstrating that ABC transporters transport most classes of biological molecules, including lipids. This chapter has detailed the intimate involvement of four ABC proteins in the translocation of a variety of lipids. The latter include membrane PLs in addition to specialized lipid types such as PAF, lipid A, and a retinoid-PE conjugate. Data supporting this role for ABC transporters has been obtained from both genetic and direct biochemical studies. The main point that still requires clarification is the precise molecular mechanism of translocation. One potential mechanism involves a classical transport process, which requires (1) lipid desorption from the bilayer, (2) association with protein and movement across the bilayer, and (3) intercalation back into the opposite leaflet. The alternative molecular process may be a flippase mechanism that requires (1) access to the protein from within the bilayer environment, (2) “dragging” the polar head group to the opposite leaflet of the bilayer, and (3) dissociation from the protein into the lipid milieu. Unfortunately, discriminating between these two processes is not currently possible. One reason is the lack of sufficient fundamental mechanistic information on the transporters involved. However, even where these data are available (e.g., ABCB1), the lack of a suitable experimental setup is the limiting factor. For example, reconstituted liposomes containing purified transporters typically require the use of short-chain lipid analogs, the vagaries of which have been outlined in the chapter. In addition, the elevated surface tension produced by movement of a small proportion of the lipids will rapidly halt the process. The use of giant unilamellar vesicles (GUVs) and measuring shape changes can overcome these deficiencies. Unfortunately, the process of protein insertion into GUVs is harsh and usually leads to inactivation of the transporter. The potential advantages of this investigative system do justify continued refinement and development. That transporters mediate lipid translocation is almost certainly not a delusion. Perhaps our continued desire to discriminate between classical transport and lipid flip-flop is the flippase delusion, at least until the challenge of devising an appropriate experimental procedure can be overcome.

ABBREVIATIONS ABC GlcCer GUV HDL IM LPS

ATP-binding cassette glucoceramide giant unilamellar vesicle high-density lipoprotein inner membrane lipopolysaccharide

References

MDR NBD OM PAF STGD VLCFA X-ALD YSV

245

multidrug resistance 7-nitro-2,1,3-benzoxadiazol-4-yl outer membrane platelet-activating factor Stargardt disease very long-chain fatty acid X-linked adrenoleukodystrophy yeast secretory vesicle

REFERENCES â•… 1â•… J. Liu, J. C. Conboy, Biophys. J. 2005, 89, 2522–2532. â•… 2â•… V. A. Fadok, D. L. Bratton, D. M. Rose, A. Pearson, et al., Nature 2000, 405, 85–90. â•… 3â•… X. Wang, Y. C. Wu, V. A. Fadok, M. C. Lee, et al., Science 2003, 302, 1563–1566. â•… 4â•… J. M. de Vree, E. Jacquemin, E. Sturm, D. Cresteil, et al., Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 282–287. â•… 5â•… E. Jacquemin, J. M. D. Vree, D. Cresteil, E. M. Sokal, et al., Gastroenterology 2001, 120, 1448–1458. â•… 6â•…Y. Meier, C. Pauli-Magnus, U. M. Zanger, K. Klein, et al., Hepatology 2006, 44, 62–74. â•… 7â•… O. Rosmorduc, R. Poupon, Orphanet J. Rare Dis. 2007, 2, 29. â•… 8â•… W. A. Alrefai, R. K. Gill, Pharm. Res. 2007, 24, 1803–1823. â•… 9â•… C. Lang, Y. Meier, B. Stieger, U. Beuers, et al., Pharmacogenet. Genomics 2007, 17, 47–60. â•‡ 10â•… M. Bodzioch, E. Orso, J. Klucken, T. Langmann, et al., Nat. Genet. 1999, 22, 347–351. â•‡ 11â•… A. Brooks-Wilson, M. Marcil, S. M. Clee, L. H. Zhang, et al., Nat. Genet. 1999, 22, 336–345. â•‡ 12â•… W. Drobnik, G. Liebisch, C. Biederer, B. Tr mbach, et al., Arterioscler. Thromb. Vasc. Biol. 1999, 19, 28–38. â•‡ 13â•… S. Rust, M. Rosier, H. Funke, J. Real, et al., Nat. Genet. 1999, 22, 352–355. â•‡ 14â•… K. E. Berge, H. Tian, G. A. Graf, L. Yu, et al., Science 2000, 290, 1771–1775. â•‡ 15â•… D. S. Papermaster, B. G. Schneider, M. A. Zorn, J. P. Kraehenbuhl, J. Cell Biol. 1978, 78, 415–425. â•‡ 16â•… R. Allikmets, N. Singh, H. Sun, N. F. Shroyer, et al., Nat. Genet. 1997, 15, 236–246. â•‡ 17â•… H. Sun, J. Nathans, Nat. Genet. 1997, 17, 15–16. â•‡ 18â•… M. Illing, L. L. Molday, R. S. Molday, J. Biol. Chem. 1997, 272, 10303–10310. â•‡ 19â•… L. L. Molday, A. R. Rabin, R. S. Molday, Nat. Genet. 2000, 25, 257–258. â•‡ 20â•… R. S. Molday, L. L. Molday, J. Cell Biol. 1987, 105, 2589–2601.

246

THE FLIPPASE DELUSION OF ABC PROTEINS

â•‡ 21â•… F. P. Cremers, D. J. van de Pol, M. van Driel, A. I. den Hollander, et al., Hum. Mol. Genet. 1998, 7, 355–362. â•‡ 22â•… A. Martinez-Mir, E. Paloma, R. Allikmets, C. Ayuso, et al., Nat. Genet. 1998, 18, 11–12. â•‡ 23â•… R. Allikmets, N. F. Shroyer, N. Singh, J. M. Seddon, et al., Science 1997, 277, 1805–1807. â•‡ 24â•… P. A. Blacharski, Fundus flavimaculatus. In Retinal Dystrophies and Degenerations, D. A. Newsome, ed. Raven Press, New York, 1988, 135–159. â•‡ 25â•… K. Stargardt, Albrecht Von Graefes Arch. Klin. Exp. Ophthalmol. 1909, 71, 534–550. â•‡ 26â•… D. M. Paskowitz, M. M. LaVail, J. L. Duncan, Br. J. Ophthalmol. 2006, 90, 1060–1066. â•‡ 27â•… R. S. Molday, J. Bioenerg. Biomembr. 2007, 39, 507–517. â•‡ 28â•… T. D. Lamb, E. N. Pugh, Jr., Prog. Retin. Eye Res. 2004, 23, 307–380. â•‡ 29â•… F. Lion, J. P. Rotmans, F. J. Daemen, S. L. Bonting, Biochim. Biophys. Acta 1975, 384, 283–292. â•‡ 30â•… S. Ishiguro, Y. Suzuki, M. Tamai, K. Mizuno, J. Biol. Chem. 1991, 266, 15520–15524. â•‡ 31â•… M. Rozanowska, T. Sarna, Photochem. Photobiol. 2005, 81, 1305–1330. â•‡ 32â•… R. E. Anderson, M. B. Maude, Biochemistry 1970, 9, 3624–3628. â•‡ 33â•… S. Ben-Shabat, C. A. Parish, H. R. Vollmer, Y. Itagaki, et al., J. Biol. Chem. 2002, 277, 7183–7190. â•‡ 34â•…G. E. Eldred, Nature 1993, 364, 396. â•‡ 35â•… J. Weng, N. L. Mata, S. M. Azarian, R. T. Tzekov, et al., Cell 1999, 98, 13–23. â•‡ 36â•… N. L. Mata, J. Weng, G. H. Travis, Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 7154–7159. â•‡ 37â•… N. L. Mata, R. T. Tzekov, X. Liu, J. Weng, et al., Invest. Ophthalmol. Vis. Sci. 2001, 42, 1685–1690. â•‡ 38â•… H. Sun, R. S. Molday, J. Nathans, J. Biol. Chem. 1999, 274, 8269–8281. â•‡ 39â•… J. Ahn, J. T. Wong, R. S. Molday, J. Biol. Chem. 2000, 275, 20399–20405. â•‡ 40â•… S. Beharry, M. Zhong, R. S. Molday, J. Biol. Chem. 2004, 279, 53972–53979. â•‡ 41â•… J. Ahn, S. Beharry, L. L. Molday, R. S. Molday, J. Biol. Chem. 2003, 278, 39600–39608. â•‡ 42â•… D. Halter, S. Neumann, S. M. van Dijk, J. Wolthoorn, et al., J. Cell Biol. 2007, 179, 101–115. â•‡ 43â•… W. E. Kaminski, A. Piehler, J. J. Wenzel, Biochim. Biophys. Acta 2006, 1762, 510–524. â•‡ 44â•… P. M. Jones, M. L. O’Mara, A. M. George, Trends Biochem. Sci. 2009, 34, 520–531. â•‡ 45â•… M. L. Oldham, A. L. Davidson, J. Chen, Curr. Opin. Struct. Biol. 2008, 18, 726–733. â•‡ 46â•… A. Ward, C. L. Reyes, J. Yu, C. B. Roth, G. Chang, Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 19005–19010. â•‡ 47â•… H. Nikaido, M. Vaara, Microbiol. Rev. 1985, 49, 1–32.

References

247

â•‡ 48â•… C. R. Raetz, C. Whitfield, Annu. Rev. Biochem. 2002, 71, 635–700. â•‡ 49â•… M. Karow, O. Fayet, A. Cegielska, T. Ziegelhoffer, C. Georgopoulos, J. Bacteriol. 1991, 173, 741–750. â•‡ 50â•… M. Karow, C. Georgopoulos, Mol. Microbiol. 1993, 7, 69–79. â•‡ 51â•… A. Polissi, C. Georgopoulos, Mol. Microbiol. 1996, 20, 1221–1233. â•‡ 52â•… Z. Zhou, K. A. White, A. Polissi, C. Georgopoulos, C. R. H. Raetz, J. Biol. Chem. 1998, 273, 12466–12475. â•‡ 53â•… W. T. Doerrler, M. C. Reedy, C. R. H. Raetz, J. Biol. Chem. 2001, 276, 11461–11464. â•‡ 54â•… W. T. Doerrler, C. R. H. Raetz, J. Biol. Chem. 2002, 277, 36697–36705. â•‡ 55â•… W. T. Doerrler, C. R. Raetz, J. Biol. Chem. 2002, 277, 36697–36705. â•‡ 56â•… P. D. Eckford, F. J. Sharom, Biochem. J. 2010, 429, 195–203. â•‡ 57â•… B. Tefsen, M. P. Bos, F. Beckers, J. Tommassen, H. de Cock, J. Biol. Chem. 2005, 280, 35961–35966. â•‡ 58â•… S. Modok, C. Heyward, R. Callaghan, J. Lipid Res. 2004, 45, 1910–1918. â•‡ 59â•… A. Rothnie, D. Theron, L. Soceneantu, C. Martin, et al., Eur. Biophys. J. 2001, 30, 430–442. â•‡ 60â•… W. T. Doerrler, H. S. Gibbons, C. R. H. Raetz, J. Biol. Chem. 2004, 279, 45102–45109. â•‡ 61â•…G. Reuter, T. Janvilisri, H. Venter, S. Shahi, et al., J. Biol. Chem. 2003, 278, 35193–35198. â•‡ 62â•… P. D. W. Eckford, F. J. Sharom, J. Biol. Chem. 2008, 283, 12840–12850. â•‡ 63â•… P. P. Borbat, K. Surendhran, M. Bortolus, P. Zou, et al., PLoS Biol. 2007, 5, e271. â•‡ 64â•… P. Zou, M. Bortolus, H. S. McHaourab, J. Mol. Biol. 2009, 393, 586–597. â•‡ 65â•… P. Zou, H. S. McHaourab, J. Mol. Biol. 2009, 393, 574–585. â•‡ 66â•…G. Chang, C. B. Roth, C. L. Reyes, O. Pornillos, et al., Science 2006, 314, 1875b. â•‡ 67â•… P. D. Jeffrey, Acta Crystallogr. D Biol. Crystallogr. 2009, 65, 193–199. â•‡ 68â•… I. D. Kerr, P. M. Jones, A. M. George, FEBS J. 2010, 277, 550–563. â•‡ 69â•… C. F. Higgins, M. M. Gottesman, Trends Biochem. Sci. 1992, 17, 18–21. â•‡ 70â•… J. J. Smit, A. H. Schinkel, R. P. Oude Elferink, A. K. Groen, et al., Cell 1993, 75, 451–462. â•‡ 71â•… A. J. Smith, J. L. Timmermans-Hereijgers, B. Roelofsen, K. W. Wirtz, et al., FEBS Lett. 1994, 354, 263–266. â•‡ 72â•… S. Ruetz, P. Gros, Cell 1994, 77, 1071–1081. â•‡ 73â•… A. van Helvoort, M. L. Giudici, M. Thielemans, G. van Meer, J. Cell. Sci. 1997, 110(Pt 1), 75–83. â•‡ 74â•… I. Bosch, K. Dunussi-Joannopoulos, R. L. Wu, S. T. Furlong, J. Croop, Biochemistry 1997, 36, 5685–5694. â•‡ 75â•…G. van Meer, D. Sillence, H. Sprong, N. Kalin, R. Raggers, Biosci. Rep. 1999, 19, 327–333. â•‡ 76â•…Y. Romsicki, F. J. Sharom, Biochemistry 2001, 40, 6937–6947. â•‡ 77â•… A. Pohl, H. Lage, P. Muller, T. Pomorski, A. Herrmann, Biochem. J. 2002, 365, 259–268.

248

THE FLIPPASE DELUSION OF ABC PROTEINS

â•‡ 78â•… P. D. Eckford, F. J. Sharom, Biochem. J. 2005, 389, 517–526. â•‡ 79â•… J. C. McIntyre, R. G. Sleight, Biochemistry 1991, 30, 11819–11827. â•‡ 80â•… T. Pomorski, A. Herrmann, A. Zachowski, P. F. Devaux, P. Muller, Mol. Membr. Biol. 1994, 11, 39–44. â•‡ 81â•… W. van’t Hof, G. van Meer, J. Cell Biol. 1990, 111, 977–986. â•‡ 82â•… T. Pomorski, A. Herrmann, B. Zimmermann, A. Zachowski, P. Muller, Chem. Phys. Lipids 1995, 77, 139–146. â•‡ 83â•… A. van Helvoort, A. J. Smith, H. Sprong, I. Fritzsche, et al., Cell 1996, 87, 507–517. â•‡ 84â•… P. F. Devaux, I. Lopez-Montero, S. Bryde, Chem. Phys. Lipids 2006, 141, 119–132. â•‡ 85â•… N. Kalin, J. Fernandes, S. Hrafnsdottir, G. van Meer, J. Biol. Chem. 2004, 279, 33228–33236. â•‡ 86â•… J. Wijnholds, E. C. de Lange, G. L. Scheffer, D. J. van den Berg, et al., J. Clin. Invest. 2000, 105, 279–285. â•‡ 87â•… S. Ernest, E. Bello-Reuss, J. Am. Soc. Nephrol. 1999, 10, 2306–2313. â•‡ 88â•… R. J. Raggers, I. Vogels, G. van Meer, Biochem. J. 2001, 357, 859–865. â•‡ 89â•… J. W. Nichols, Biochemistry 1985, 24, 6390–6398. â•‡ 90â•… A. M. van der Bliek, EMBO J. 1987, 6, 7. â•‡ 91â•… E. Buschman, P. Gros, Mol. Cell. Biol. 1991, 11, 595–603. â•‡ 92â•… P. Gros, M. Raymond, J. Bell, D. Housman, Mol. Cell. Biol. 1988, 8, 2770–2778. â•‡ 93â•… M. Raymond, E. Rose, D. E. Housman, P. Gros, Mol. Cell. Biol. 1990, 10, 1642–1651. â•‡ 94â•… A. H. Schinkel, M. E. M. Roelofs, P. Borst, Cancer Res. 1991, 51, 2628–2635. â•‡ 95â•… D. Degiorgio, C. Colombo, M. Seia, L. Porcaro, et al., Eur. J. Hum. Genet. 2007, 15, 1230–1238. â•‡ 96â•… F. T. Alissa, R. Jaffe, B. L. Shneider, J. Pediatr. Gastroenterol. Nutr. 2008, 46, 241–252. â•‡ 97â•… J. F. Deleuze, E. Jacquemin, C. Dubuisson, D. Cresteil, et al., Hepatology 1996, 23, 904–908. â•‡ 98â•… R. Oude Elferink, C. Paulusma, Pflügers Arch. Eur. J. Physiol. 2007, 453, 601–610. â•‡ 99â•… D. Gotthardt, H. Runz, V. Keitel, C. Fischer, et al., Hepatology 2008, 48, 1157–1166. 100â•… P. Borst, N. Zelcer, A. van Helvoort, Biochim. Biophys. Acta 2000, 1486, 128–144. 101â•… D. M. Small, Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 4–6. 102â•… A. J. Smith, J. M. de Vree, R. Ottenhoff, R. P. Oude Elferink, et al., Hepatology 1998, 28, 530–536. 103â•… M. Raymond, P. Gros, M. Whiteway, D. Y. Thomas, Science 1992, 256, 232–234. 104â•… S. Ruetz, M. Raymond, P. Gros, Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 11588–11592. 105â•… A. J. Smith, J. L. P. M. Timmermans-Hereijgers, B. Roelofsen, K. W. A. Wirtz, et al., FEBS Lett. 1994, 354, 263–266.

References

249

106â•… A. J. Smith, A. van Helvoort, G. van Meer, K. Szabo, et al., J. Biol. Chem. 2000, 275, 23530–23539. 107â•… A. Pohl, P. F. Devaux, A. Herrmann, Biochim. Biophys. Acta 2005, 1733, 29–52. 108â•… T. Pomorski, J. C. M. Holthuis, A. Herrmann, G. van Meer, J. Cell. Sci. 2004, 117, 805–813. 109â•… T. Pomorski, A. Menon, Cell. Mol. Life Sci. 2006, 63, 2908–2921. 110â•…G. van Meer, D. Halter, H. Sprong, P. Somerharju, M. R. Egmond, FEBS Lett. 2006, 580, 1171–1177. 111â•… M. Ziol, V. Barbu, O. Rosmorduc, A. Frassati-Biaggi, et al., Gastroenterology 2008, 135, 131–141. 112â•… R. J. P. Dawson, K. P. Locher, Nature 2006, 443, 180–185. 113â•… S. G. Aller, J. Yu, A. Ward, Y. Weng, et al., Science 2009, 323, 1718–1722.

PART V RELEVANCE OF LIPID TRANSMEMBRANE DISTRIBUTION FOR MEMBRANE PROPERTIES AND PROCESSES

12 MEMBRANE LIPID ASYMMETRY AND PERMEABILITY TO DRUGS: A MATTER OF SIZE Adam Blanchard and Cyril Rauch School of Veterinary Medicine and Science, University of Nottingham, Sutton Bonington, Leicestershire, UK

12.1â•… INTRODUCTION The time from drug discovery to launch is currently ∼12 years and costs ∼$750 million/drug. Naturally, the pharmaceutical industry is determined to reduce both the cost and timescale of this process. A significant bottleneck in this pipeline is determining the properties of a drug that facilitate its delivery to, and uptake by, target tissues/cells. To this end, Lipinski and collaborators from Pfizer produced a set of rules that attempt to identify the physicochemical properties required for an oral compound to achieve maximum bioavailability, that is, to cross all biological barriers before reaching its target. The first of Lipinski’s rules is based on the lipophilic index of the drug, the second is on the drug’s molecular weight (abbreviated “MW” in the remaining text), and the third and fourth rules concern the drug’s charge properties. These rules are established drug discovery paradigms and have been largely embraced by the pharmaceutical industry. However, a full and systematic scientific exploration of the way drugs interact with cells or tissues to generate these rules has yet to be carried out. Testing the validity of these rules and thereby devising robust design models that maximize drug bioavailability, would transform

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

253

254

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

these rules into a set of laws that are of central importance to the pharmaceutical industry. Furthermore, such research could enlarge the libraries of active compounds beyond Lipinski’s limits, which in turn should (1) reduce the time and cost in delivering product and (2) inform therapeutic strategies that increase drug bioavailability. Of the four rules, the second (MWâ•¯<â•¯500) stands out by way of its apparent simplicity, being unrelated to complex physicochemical properties of a drug (as is its state of charge or lipophilic index) but governed solely a drug’s size or volume. This simplicity infers that basic mechanics apply when drugs cross cells or tissues. Thus, it is critical to understand the mechanical properties of biological barriers and how they may be manipulated to maximize both drugs’ bioavailability and efficacy. Therefore, prior to establishing a fundamental link between Lipinski’s second rule and the ubiquitous membrane lipid asymmetry mediated by a flippase, it is important to understand the context in which Lipinski’s strategy has emerged. This, in turn, will emphasize the constant need for research at the interface between the sciences, often leading to new concepts spanning across disciplines, in this case among biology, physics, and biophysics. Lipid membrane asymmetry exemplifies this need for interdisciplinary research, merging soft matter physics to membrane biophysics and biology, with important implications in the field of drug delivery and bioavailability. 12.2â•… THE ORIGIN OF LIPINSKI’S SECOND RULE FROM THE POINT OF VIEW OF THE PHARMACEUTICAL INDUSTRY 12.2.1â•… The Pharmaceutical Industry: A Business in Need of Blockbusters The 1990s were gloomy years for the pharmaceutical industry with productivity falling below expectations. Indeed, the 10 leading companies’ newly marketed compounds increased their revenues by only ∼10%, and the average innovation deficit was ∼1.3–1.8 new chemical entities per year [1]. Evidently, the pharmaceutical industry has adopted a “megabrand” marketing concept that advocates a strong focus on single products yielding significant returns at peak sales (∼$1 billion/annum) [1, 2]. Accordingly, this vision has emulated the notion of “blockbusters” that, as we shall see, has the toxic side effect of narrowing the scope of investigations by limiting the emergence of new therapeutics. Following this notion, it is therefore understandable that the strategies adopted by these companies are those that provide information, in advance of costly clinical trials, about which chemicals are most likely to become blockbusters. Thus, the properties that make a chemical a “likely” drug predefine and are central to the “blockbuster” product. Two approaches are used to determine how likely a chemical will become a successful drug. Given that most successful drugs will be administered orally, the first way is to focus scientifically on the specific biological tissues that

The Origin of Lipinski’s Second Rule

255

impede the action of a specified chemical and, of course, to find a way to maximize drug action. This first focus has been initially developed and is summarized by the absorption, distribution, metabolism, and excretion (ADME) (see below). The second way to solve the same problem is to focus statistically on the physicochemical properties of marketed drugs to date and to screen out from clinical development those molecules with inappropriate properties. This point has been developed by Lipinski. 12.2.2â•… ADME The bioavailability of drugs can be summarized by four notions grouped under the acronym ADME, which stands for absorption, distribution, metabolism, and excretion. Each of these notions is involved in a particular aspect of the physiological interactions between the body tissues and drugs, which explain drugs bioavailability. Orally administrated drugs must be absorbed through the gastrointestinal tract and “diffuse” into capillaries or lymphatic vessels to be distributed in the body tissues. Note that the term used “diffuse” is only a simplistic way of representing how drugs move from one place to another in the body and thus does not correspond to passive diffusion defined physically. Although a good biodistribution of drugs often ensures their efficacy, nonetheless the accumulation of drugs in the body is toxic. Thus, the biological barriers in place in our body control the distribution of drugs and their toxicity. This control is performed via two principal, specific biological mechanisms. The first one involves enzymes known as cytochrome P450s. Cytochrome P450s metabolize drugs in the intestine or liver and impede their action by conjugating the drugs with small chemical groups that have high water affinity. As a result, the drug loses its initial affinity with lipid bilayers and is excreted in the urine. The second type of enzyme involve drug membrane transporters, which line the intestinal epithelium (but also other specific body compartments, e.g., the blood–brain barrier) and oppose drug entry into the body tissues [3, 4]. It is remarkable that these transporters that are known to be involved today in the bioavailability of drugs, were initially discovered from studies focused on drug resistance [5]. It is believed that P450s and drug transporters act synergistically to impair the bioavailability of drugs. It is assumed that transporters expel drugs, thus avoiding P450 enzyme saturation by drugs. It follows that as drug accumulation is toxic, the body mechanisms involved in drug metabolism and excretion are thus vital. Therefore, drug activity against its target is balanced by these four parameters (called first-pass mechanism). Accordingly, inappropriate ADME means inefficiency (too many drugs being excreted) or toxicity (drugs not efficiently excreted and thus body accumulation). At first sight, these four notions seem relatively simple; however, this is without taking into consideration the complex heterogeneity of the body. To

256

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

give an example, drug adsorption would depend not only on drug’s membrane permeability, water solubility, and its charge state properties, all of which are important when drugs interact with lipid membranes, but also on the local properties of the gastrointestinal tract (e.g., luminal pH and surface area) [6]. Therefore, physicochemical and biological interactions between chemicals and body tissues generate extra dimensions with complex and difficult appreciations. Completely understanding the physicochemical properties a drug should have will indeed provide the adequate drug likeliness properties for best bioavailability. The downside is that this strategy takes naturally, has already taken, a long time. The reluctance of pharma companies to undertake this “academic research pathway” is more understandable given their drive for high and rapid returns. 12.2.3â•… Lipinski’s Rules Lipinski’s strategy answered a real demand in an innovative way. Although the 1990s were gloomy for the pharmaceutical industry, this period has also seen the explosion of what is known today as “modern medicine,” that is, potential new targets deduced from genomic, proteomic, and metabolomic studies. Accordingly and given the constantly increasing number of clinically relevant targets, there was a need to identify more potent chemicals to act on these predefined targets. In turn, this approach has led, or enhanced, the need for combinatorial chemistry leading to pharmaceutical companies holding large chemical libraries. The expectation with this approach was that there would be an increase in good quality drug-like molecules to “hit” new targets [7]. In this environment of discovery, Lipinski and his collaborators from Pfizer analyzed the physicochemical properties of marketed oral drugs. Lipinski found that marketed drugs follow four rules. The first rule is based on the lipophilic index of drugs (octanol–water partitioning: LogPâ•¯<â•¯5); the second rule is based on the drugs’ MW, which needs to be <500; and the third and fourth rules are based on the drugs’ state of charge (number of hydrogen-bond donors, i.e., number of OHâ•¯+â•¯NH bondsâ•¯>â•¯5, and number of hydrogen bonds acceptor, i.e., number of Oâ•¯+â•¯N atomsâ•¯>â•¯10). Together, these rules define the ninetieth percentile of physicochemical properties drugs should have to achieve the greatest bioavailability [8]. As these rules concern synthetic chemicals, they were initially criticized as many pharmaceutical compounds are also natural compounds; however, it was later found that natural compounds, unsurprisingly, also follow Lipinski’s rules [9]. Given the potentially huge return from Lipinski’s rules, which predicts inefficient compounds prior to them reaching the development stage, this specific route regarding the determination of drug likeliness properties has been taken by others [10, 11], extending these initial rules to other physicochemical properties compounds should have [11, 12].

Solving Lipinski’s Second Rule

257

Today, the application of these guidelines to determine the drug likeliness properties of potential lead compounds has been largely embraced by the pharmaceutical industry, and accordingly, Lipinski-type rules are now an integral part of the decision-making process in this industry. Accordingly, these rules have overtaken the initial basic drug discovery process initiated by Erlich (1920s) and Fleming (1930s) (reviewed in References 13 and 14) and are now considered drug discovery paradigms: Any compound that does not follow Lipinski’s rules will be disregarded. 12.2.4â•… Side Effects of Lipinski’s Rules and Perspectives Using Lipinski’s rules in a decision process signifies simply that it will be preferable to change the physicochemical properties of compounds to fit Lipinski’s rules than to understand why such rules exist. By definition, rules are not laws, and there is a real need to transform these rules into laws. Pharmacologically speaking, this means that new biological targets might be determined to bend and promote the control of these rules. As stated in the introduction, one rule among the others stands out by its simplicity. It is the second rule that is linked to drug MW. This chapter aims to demonstrate that this rule may be transformed into a law, so long as the mechanical properties of the membrane driven by the lipid asymmetry are taken into consideration. Of course, understanding how these mechanical properties are intimately linked to membrane lipid asymmetry is central to this purpose. 12.3â•… SOLVING LIPINSKI’S SECOND RULE 12.3.1â•… Act One: Definition of Relevant Biophysical Parameters To be (bio)available, drugs must traverse cellular barriers—usually epithelia (e.g., of the gastrointestinal tract, renal tubules, or the blood–brain barrier). To traverse cellular barriers, drugs must cross lipid membranes, and for this, Lipinski’s second rule postulates that drugs must have an MW lower than 500. Compounds that are small in size have an MW that is proportional to their volume; this rule suggests then that drug volume or size is a limiting parameter when crossing biomembranes. In turn, this suggests a basic mechanical interaction between a drug and the cell membrane.1

1â•›

Thermodynamically speaking, the physical parameters that are related to spatial dimensions (namely, volume [V], cross-section area [a], or line [r]) are the pressure “P”: δEâ•¯=â•¯−P·δV, the surface tension “σ”: δEâ•¯=â•¯σ·δa, and the tension line “γ ”: δEâ•¯=â•¯γ ·δr. “δ ” is the differential operator and “E” is the energy. As far as a membrane is considered, it is the surface tension (and thus the cross-section area of the drug) that best describes the mechanical (i.e., physical) interaction, and Equation 12.1 is deduced by posing δEâ•¯∼â•¯kBT.

258

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

Therefore, in the sum of energies making up the total activation energy required for a drug to cross cellular membranes, there must exist an energy term that is a specific function of the drug’s dimension so that the drug/ membrane interaction yields an energy ≥kBT (kB is Boltzmann’s constant and T is the temperature in Kelvin). In this case, that is, when the plasma membrane is considered, the physical parameter that best fits such an interaction is the leaflets’ surface pressure (σ and unit [σ]â•¯=â•¯N/m).2 Dimensionally speaking, it follows that a critical cross section (ac) can be defined simply by

ac ∼ kBT / σ .

(12.1)

Of course, these parameters need a proper definition. In bilayer membranes, two types of membrane surface tension can be distinguished: the mean surface tension noted σ0, which corresponds to the sum of individual leaflet’s surface tension, and the difference in surface tensions Δσ, which corresponds to the difference between individual leaflet’s surface tension. Using optical techniques, M. Sheetz and his collaborators demonstrated that cells have a large reservoir of membrane [15] and an average membrane tension that is remarkably low (σ0∼0.003â•›mN/m) [16], similar to the mean surface tension measured from in vitro systems (i.e., large liposome where the thermal undulations are dominant and dictate the mean membrane tension [17]). On the other hand, the difference in surface tensions between leaflets has been demonstrated to be much higher Δσ∼0.9â•¯mN/m [18]. Accordingly and given the magnitude of this parameter, it is more likely to be involved in impairing the transverse movement of chemicals. The previous equation can thus be refined as follows:3

ac ∼ kBT / ∆σ .

(12.2)

Dealing with a parameter as Δσ is not intuitive physiologically and thus difficult to apprehend. Thus, the last equation needs to be resolved physiologically. A fundamental aspect of the difference in surface tension corresponds to its role in fluid-phase endocytosis. 12.3.2â•… Act Two: The Role of Fluid-Phase Endocytosis The physical reason behind the difference in surface tension is associated with the role of lipid flippase, which maintain membrane lipid asymmetry [19]. In 2â•› Note that in the following text, surface pressure or tension will be used without conceptual difference. In both cases, they refer to the mechanical packing of lipids in membrane leaflets. 3â•› Liposome studies have demonstrated that intercalation of chemicals in the outer leaflet trigger an outward membrane budding. At first sight, we could expect a similar effect in cells. However, an important difference between the cell membrane and a liposome is that cells contain an important reservoir of intracellular membrane stored as organelles or intracellular vesicles. This means that in the presence of lipophilic agents, membrane exocytosis is triggered to balance the tension induced by the chemical upon intercalation [15]. It is the reason why the lipid asymmetry is conserved upon of chemical intercalation in the cell membrane.

Solving Lipinski’s Second Rule

259

particular, flippases actively relocate phosphatidylserine (PS) and phosphatidylethanolamine (PE) from the outer into the inner leaflet (see Chapters 8– 10). A direct consequence associated with this inward pumping is a more highly packed inner leaflet as it contains more phospholipids than the outer leaflet. As a result, there exists a difference in surface tensions (Δσâ•¯=â•¯ σoutâ•¯−â•¯σinâ•¯∼â•¯0.9â•›mN/m) between the inner (cytosolic) and outer leaflets of the plasma membrane. Naturally, bilayer membranes are soft objects and as such, will attempt to release this energy stored as a differential lipid packing. Accordingly, it has been demonstrated that lipid asymmetry corresponds to the physiological motor force that triggers membrane budding, leading to fluid-phase endocytosis (Fig. 12.1) [18, 20, 21]. Whatever the difference in leaflets’ packing, there is a cost to bend the membrane that is directly associated with the membrane bending modulus (noted kc). Studies have demonstrated that the magnitude of the bending modulus in living cells is kcâ•¯∼â•¯2.7â•¯×â•¯10−19J [16]. The latter value is similar to the bending modulus measured in liposomes composed of phosphatidylcholine (PC) [17]. From the competition between the difference in surface tensions (Δσâ•¯=â•¯σoutâ•¯−â•¯ σinâ•¯∼â•¯0.9â•›mN/m), which encourage membrane budding, and the membrane bending modulus (kcâ•¯∼â•¯2.7â•¯×â•¯10−19J), which opposes membrane curvature, it is possible to theoretically predict the size of instable membrane buds. As said above, the energy of a membrane patch budding of radius R, thickness h, and of neutral surface area S can be described by the sum of two terms. The first term describes the force driving membrane budding that is associated with the endogenous difference in surface tensions between the plasma membrane leaflets, and that is linked to phospholipid number asymmetry given by hΔσS/2RV, where Δσâ•¯=â•¯−2KδN/N0 is the difference in surface tensions, K is the elastic modulus of leaflets, δN is the number of phospholipids in excess in the inner leaflet (compared with the outer one), and N0 is the average phospholipid number in each leaflet. The second term is the bending energy, which corresponds to the resistance to membrane curvature and is given by 2kc S / RV2 , where kc is the membrane bending modulus. The competition between these two energies provides an optimal budding radius:

R = 8kc / h∆σ .

(12.3)

Finally, inserting Equation 12.3 into the sum of the difference in surface tensions and bending energies allows the classical determination of the energy released by the membrane after completion of the formation of a vesicle that is given by −8πkc, which is naturally independent of the vesicle size. From Equation 12.3, it is also possible to predict the numerical value of the fluidphase vesicle radius Râ•¯=â•¯8kc/hΔσ∼35â•¯nm [18]. Although it would be very easy to replace the latter formula into Equation 12.2 to determine the critical crosssection area we are looking for, however, doing so would not solve a paradox linked to the fact that the membrane has always been the exclusive central point for membrane budding mechanisms.

260

Phospholipid Pumping

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

Endocytosis

Exocytosis

Dilation

Dilation

Compression

Compression

Bending

Vesiculation

Figure 12.1.â•‡ The lipid number asymmetry-induced fluid-phase endocytosis model. Sketch representing the current model that has been applied in living cells links fluidphase endocytosis and the membrane phospholipid number asymmetry maintained by a flippase. In the left figure, the translocation of dark-head lipids into the inner leaflet induces a differential lipid packing between leaflets (namely a difference in surface tensions) leading to membrane bending and vesiculation [18, 20]. Note that it is assumed that the membrane recycling that occurs in cells, that is, the exocytosis of vesicles of a size similar to endocytic vesicles, also allows the maintenance of the lipid asymmetry at the level of the plasmalemma. The relationship existing between the lipid number asymmetry and the vesicle radius is given by Râ•¯=â•¯8kc/hΔσ or equivalently Râ•¯=â•¯4kc/hK ·1(δN/N0), where kc, K, h, Δσ, and δN/N0 are the membrane bending modulus, membrane elastic modulus, membrane thickness, difference in surface tensions, and the lipid number asymmetry between leaflets. Accordingly, the lipid number asymmetry has been experimentally deduced from studies on cells that δN/N0â•¯=â•¯2â•¯±â•¯0.5% providing a ∼35-nm vesicle radius [18]. Color version on the Wiley web site.

Earlier on, we have seen that cells have a very low membrane tension and that vesiculation is performed thanks to the membrane lipid number asymmetry. Accordingly, the budding mechanism that initiates fluid-phase endocytosis can be described as a membrane instability mechanism because the membrane releases energy for each cytosolic vesicle created equivalent to −8πkc per vesicle (note that there is a minus sign because the membrane

Solving Lipinski’s Second Rule

261

releases the energy). As a result, this model suggests that membrane budding occurs to cancel the lipid asymmetry of the plasma membrane. However, an apparent paradox arises from this model that exclusively focused on the membrane: As it is favorable for the membrane to release vesicles, should not the cell aim to totally eliminate lipid asymmetry through the formation for many vesicles? Accordingly, as more vesicles are formed, the mean membrane tension would increase up to the point where the lateral mean membrane tension could inhibit membrane budding. In turn, this would contradict the initial hypothesis and associated an experimental observation; that is, the mean surface tension is low enough to allow membrane vesicles. Therefore, we are presented with the question of why is the plasma membrane lipid asymmetry constant? It follows that membrane lipid asymmetry is not conceptually sufficient to explain membrane vesiculation, that is, fluid-phase endocytosis. The paradox is solved by invoking a coherent assumption, that is, as the amount of endocytosed membrane is equivalent to the amount of exocytosed membrane and as the size of endocytic and exocytic vesicles are similar, the plasma membrane has a constant lipid number asymmetry and remains at low mean membrane tension. Overall, this biological assumption ensures that there is a conservation of membrane pools between the cytoplasm and the plasmalemma. Although this assumption is consistent, the reason for the constant membrane lipid asymmetry and low membrane tension can be clarified in a much simpler way, by considering the hydrostatic cytosolic pressure and the energies exchanged between the cell membrane and the cytosol. 12.3.3â•… Act Three: The Role of Cytosolic Pressure on Both Maintenance of Plasma Membrane Lipid Number Asymmetry and Membrane Recycling It is well known that the cell membrane is poorly permeable to a large range of solutes. Therefore, when an intracellular vesicle is created from the plasma membrane, the content of the new vesicle is not in contact with the cytosol. Thus, every newly created vesicle of volume, V, introduces a new volume into 0 the cytosol that must equilibrate the cytoplasmic pressure, Pcell , due to the 4 excluded-volume effect. In turn, this argument suggests that the membrane must provide sufficient energy to compensate the excluded-volume effect.

4â•›

We acknowledge that the terms “volume exclusion” or “excluded volume” are used outside the context of statistical physics (as this terminology is from van der Waals’ description of gases). Nonetheless, this physical notion represents quite well what we mean by the energy cost needed to generate a cytosolic vesicle from the membrane. Indeed, cytosolic volume must be freed to allow membrane vesiculation because the vesicular membrane is impermeable to water and most solutes, meaning that the vesicular volume is initially excluded from the cytosol. In essence, the vesicle has to “force its way” into the cytosol, which results in the notion of cytosolic “volume exclusion.”

262

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

Considering the system composed of the membrane and cytoplasm, the conservation of energies implies that

δ Em + δ Ecyto = 0,

(12.4)

where δEm is the membrane energy and δEcyto is the cytoplasmic energy. When a vesicle is formed, the membrane has released an energy equivalent to δEmâ•¯=â•¯−8πkc. However, the introduction of one vesicle into the cytosol withdraws a cytoplasmic volume δVcellâ•¯=â•¯−V that corresponds to the size of the vesicle to be introduced in the cytosol. As a result, the cytosolic energy linked 0 0 to volume exclusion increases by a factor δ Ecyto = − Pcell δ Vcell = Pcell V . Finally, the conservation of energies (Eq. 12.4) implies that

0 −8π kc + Pcell V = 0.

(12.5)

Equation 12.5, which defines the conservation of energies, can be represented as a pressure equilibrium: the membrane exerting a pressure 8πkc/V on 0 the cytosol and the cytosol exerting a pressure Pcell on the membrane. Given Equation 12.5, the numerical value of the vesicle radius can be determined. In culture conditions, the pressure applied to cells is similar to the atmospheric 0 pressure, Pcell ∼ 10 5 Pa [22], using kcâ•¯∼â•¯2.7â•¯×â•¯10−19J it follows R∼31â•¯nm, which agrees well with previous findings (i.e., R∼35â•¯nm, see Fig. 12.1) [18], and strongly suggests that the cytosolic pressure does indeed play a role in membrane vesiculation. Under these conditions, it follows that there is a direct connection between the ability of the membrane to generate vesicles and the cytosolic pressure. In fact, as volume exclusion effects are now taken into consideration, the membrane recycling process is thus intimately dependent on the intracellular pressure. Accordingly, the total energy of the system, “cytosolâ•¯+â•¯membrane,” exchanged during the membrane recycling remains constant so long as the sum of inward (Jin) and outward (Jout) flows of vesicular volumes compensate for one another (Jinâ•¯+â•¯Joutâ•¯=â•¯0). Thus, under these conditions, the systems pressure equilibrium implies that

( NV / t )in = −( NV / t )out = cte,

(12.6)

where N and t are, respectively, the vesicle number and the characteristic time for vesicle generation (endocytosis) or fusion (exocytosis). Note that given this equation, it follows that the fraction of plasma membrane undergoing vesiculation NS/t (where S is the vesicle surface area) should be inversely proportional to the vesicle’s radius. Finally, as the vesicle radius is inversely proportional to the lipid number asymmetry (see legend in Fig. 12.1), as a result, the fraction of the plasma membrane undergoing vesiculation should be directly proportional to the membrane lipid number asymmetry:

NV / t = cte ⇒ NS / t ∼ 1 / R ⇒ NS / t ∼ δ N / N 0.

(12.7)

263

Fluorescence quenching dynamics (% / min)

Solving Lipinski’s Second Rule 8 7 6 5 4 3 2 1 0 0

1

2

3

4

Added C6-PS (%)

Figure 12.2.â•‡ Endocytosis measurements. Original figure from Reference [20] where the increase in fluid-phase endocytosis kinetic rates were measured as a function of the amount of PS translocated to the inner leaflet of cells. The outer membrane of cells was labeled using a biotin–streptavidin–fluorescein isothiocyanate (FITC) treatment, FITC being sensitive to (i.e., bleached by) the luminal acidic pH of early intracellular compartments. The data agree qualitatively and quantitatively (not shown) with Equation 12.7.

A method of measuring the kinetic rate of endocytosis is performed by monitoring the fraction of plasma membrane undergoing vesiculation. This experiment represented in Figure 3.2 has effectively demonstrated that relocation of PS from the outer to the inner leaflet increases the kinetic rate of fluid-phase endocytosis in a linear way (Fig. 12.2). Thus, the pressure equilibrium would remain unchanged; that is, the system would remain at thermodynamic equilibrium, so long as the relationships defined by Equations 12.5 and 12.7 are verified. It follows therefore that fluid-phase endocytosis is controlled by cytosolic pressure, which means that the cytosol is not a reservoir of free space available for vesicles. Consequently, the cellular plasma membrane is indeed a reservoir of lipid asymmetry as postulated initially [18], and this reservoir is not freely poured or released into the cytosol by the membrane, but instead controlled by the energies exchanged between the membrane and the cytosol. 12.3.4â•… Final Act: Lipinski’s Second Rule and Perspectives So far, we have seen that given the flippase activity at the plasma membrane, a lipid asymmetry between leaflets will be generated and that this lipid asymmetry will favor membrane bending. However, the bending of the membrane is tightly controlled physically by the cytosolic pressure, which, in turn, allows the membrane to maintain constant the lipid asymmetry. Accordingly, we can affirm that as fluid-phase endocytosis is ubiquitous, it is very likely that there

264

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

exists a constant lipid asymmetry maintained by the plasmalemma of every cell. Given this, it is now possible to go back to Lipinski’s second rule and to express the scientific law that is associated to it. As the difference in surface tensions is dominant and that for drugs small enough their MW is proportional 3 ), that is, MWâ•¯∼â•¯Vâ•¯∼â•¯a3/2, using to their van der Waals volume (expressed in A the theoretical expression of the vesicle radius (Eq. 12.3), a critical MW (MWc) can be determined given by

MWc = (4 / 3 π )(hRkBT / 8kc )3 / 2 .

(12.8)

The latter equation plotted in Figure 12.3a provides a law with regard to the drugs’ size (or MW) selectivity on their permeation across cellular membranes. Using the numerical values of physical constants or biological parameters seen until know, it comes that MWcâ•¯≅â•¯240 at 37°C [23]. As stated before, the MW cutoff as defined by Lipinski’s second rule, that is, MWcâ•¯=â•¯500, describes the ninetieth percentile. The latter values are remarkably close. Under these conditions, Equation 12.8 would be Lipinski’s second law. Finally, if the drug/membrane interaction yields an energy ≥kBT, this means that the drug’s membrane transverse movement is impaired (which corresponds to the upper area in Fig. 12.3a). Conversely, smaller drugs will traverse the membrane more easily (which corresponds to the lower area in Fig. 12.3a). Thus, it is noteworthy that changing the fluid-phase vesicle radius, namely the lipid number asymmetry, may lead to facilitated or impaired membrane transverse movement (depending on the nature of the changes). More specifically and using Arrhenius’ law, the influx into the cytosol of a given MW drug should be

J ∼e

 MW  −  MWc 

2/3

.

(12.9)

As a result, and as stated initially, transforming rules to laws brings potentially new targets, and in this case, the lipid metabolism or membrane lipid asymmetry may well be an important target regarding drug delivery. To understand how Lipinski’s second law stands in the field of drug delivery, it is important to place it in its context, and we will focus on the relationship between drug sizes, membrane transporters, and mechanical properties. This specific scenario is encountered when drugs try to bypass the intestinal barrier. 12.4â•… LIPINSKI’S SECOND LAW AND POTENTIAL APPLICATION 12.4.1â•… Interaction between Membrane and Drug Transporters The first membrane barrier that oral drugs encounter is the membrane of the intestinal barrier. This barrier is lined by transporters, which, like those involved in drug resistance, expel drugs from the membrane [5]. These types of trans-

265

Lipinski’s Second Law and Potential Application

(a)

1000 800

Membrane transverse movement of the drug impeded

600

~ R 3/2

MWc 400 Membrane transverse movement of the drug facilitated

200 0 0

20

40

60

80

100

R (nm) 35,000

(b) 30,000

Count

25,000 20,000 15,000 10,000 5000 0 0

300

600

900

1200

1500

Molecular weight

Figure 12.3.â•‡ Membrane mechanical properties involved in drug transverse movement across the membrane (a): Relationship between the drugs’ MW and their ability to bypass the membrane barrier as a function of vesicles radius R(nm) expressed in nanometers, scaling as MWcâ•¯∼â•¯R3/2 (exactly: MWcâ•¯≅â•¯1.1R3/2 using constants seen in the text). (b) The histogram represent the number of natural bioavailable compounds as a function of their MW [9]. The red arrow points the theoretical prediction as determined by Equation 12.8. Note that with a rational vesicle radius of 45â•›nm, the red arrow would point an MW of 350, which corresponds to the highest count. This histogram confirms that Lipinski’s second rule can be applied to synthetic or natural products.

266

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

porter also have many other body locations; for example, they are found on the apical surface of endothelial cells lining the brain capillaries and form a major component of the blood–brain barrier. Their functional importance has been clearly demonstrated. Transgenic mice lacking P-glycoprotein (Pgp), a supposedly drug transporter, do not display any particular abnormalities; nonetheless, their blood–brain barrier is strongly disrupted, and as a result, Pgp knockout mice are known to be excessively sensitive to drugs [24]. Although the exact mechanism by which these transporters expel drugs is still an active field of research, it is clear that the interaction between the drug and the membrane plays a key role. This is particularly well demonstrated experimentally. For example, all Pgp transporter substrates are highly lipophilic and have a slow diffusion time when they traverse lipid bilayer membranes, and the higher the affinity between a drug and the membrane, the better its Pgp efficiency (reviewed in References 25 and 26). In addition, Pgplike transporters recognize their substrates from the lipid phase, and Pgpdependent pumping kinetic studies can be clearly interpreted in favor of the hypothesis that Pgp handle these substrates in the inner leaflet before they reach the cytosol [27–33], which suggests that the inner membrane leaflet is coupled to Pgp-mediated drug extrusion. Finally, it is now also acknowledged that physical parameters related to the size of the drug (e.g., MW, volume, or van der Waals surface area) are important characteristics regarding chemicals extrusion via Pgp-like transporters [34–37]. Therefore, given the importance of membrane transporters, bilayer membrane physical properties, and chemical sizes to the concept of bioavailability, there is a strong suggestion that the mechanical properties of the membrane and membrane transporter activity are coupled together. 12.4.2â•… On the Relationship between Membrane Lipid Asymmetry and Drug Transporters Given the involvement of Pgp-like transporters in drug bioavailability, an important aspect that has remained elusive is the reason how a drug incorporated into the membrane meets a transporter. Instead, the elusive “vacuum cleaner” hypothesis was invoked [38] to explain how a drug would meet a transporter and be expelled. As far as the membrane’s mechanical properties and Pgp-like transporters functionality are concerned, it is possible to provide another hypothesis that correlates basic principles from membrane physics with transporter biology. In very simple terms, membrane-embedded drugs laterally diffuse randomly toward transporters. In addition, the reason why the meeting between a drug and a transporter, (i.e., prior to drugs being expelled) can occur within the inner leaflet would be because the lipids in the inner leaflet are highly packed as they are more numerous. This high inner leaflet packing would mechanically block or “squeeze” drugs in this leaflet (Fig. 12.4a). Given this working hypothesis and Equation 12.8 (i.e., Lipinski’s second

(a)

(b)

h

Out

In

Drugs Flow Membrane

U(x)

Transporter

x

Drug Diffusive path

Figure 12.4.â•‡ Activation energies and their role in the two-dimensional random walk model. (a) Representation of the different energy barriers involved when a drug traverses the bilayer of the cellular membrane. Two leaflets have been represented with an inner leaflet containing more phospholipids, which is related to the increase in the difference in surface tensions (upper graph). The activation energy profile U(x), that is, the energy that a drug is expected to encounter when crossing the membrane and that needs to be bypassed, is plotted against the membrane thickness. The activation energy is related to the surface tensions in the leaflets (plain curve) and dehydration (dashed curve-middle graph); both provide a penalty energy with regard to the drug transbilayer movement. The surface tension of one leaflet corresponds to the average value of the profile (plain curve). Given the lipid asymmetry between leaflets, the energy profile of the surface tension in the inner leaflet is higher, which allows drugs to remain in this leaflet prior to meeting, and being expelled by, a transporter. (b) Twodimensional random walk hypothesis: During its residency time in the membrane, we assume that the drug diffuses laterally over a length that is related to the membrane barrier energy to bypass. As a result, a drug may encounter a membrane transporter [38]. Equation 12.10 is determined as follows. Assuming that the number of membrane transporters on the cellular surface of cells remains globally constant and that the number of drugs in the membrane and the outward pumping kinetics are respectively low and fast enough, such that the probability that two drugs meet a given transporter at the same time is negligible, as a result, transporters can be considered as static with a probability of presence on the cellular surface given by ρPgp. During its residency time in the cellular membrane, a drug is expected to follow a two-dimensional random walk, with a formal condition restricted to timescales greater than the single transverse diffusion time, that is, h2/4D. Consequently, the number of steps performed is Kâ•¯=â•¯t0/ (h2/4D)â•¯=â•¯exp(U/kBT), where t0 is the membrane residency time of the drug and U is the activation energy. In the present case, U corresponds to the sum of two terms, the drug dehydration energy, ΔG, and the mechanical interaction between the drug and the membrane difference in surface tensions, (MW/MWc)2/3. Therefore, the probability of a drug and a transporter meeting will depend on nonrecurring walks, and for K large (i.e., t0â•¯>â•¯h2/4D), the number of nonrecurring walks is given by πK/ln(K) [41]. Finally, the probability of the event “drug meeting a Pgp” in the membrane, namely Equation 12.10, is equivalent to the probability of meeting one transporter, ρPgp, multiplied by the number of nonrecurring steps, πK/ln(K), which provides Equation 12.10.

268

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

law), it has been possible to formalize the meeting probability of interaction between a membrane-embedded drug and a transporter. Noting ρPgp, the surface density of transporters is defined as ρPgpâ•¯=â•¯NPgpSPgp/Scell, where NPgp, SPgp, and Scell represent the number of transporters, the crosssection area of transporters, and the cellular surface, respectively. Using Equation 12.8, and noting ΔG, the penalty energy of a drug crossing the membrane linked to its hydration property, the probability of a drug and transporter meeting in the inner leaflet is determined by [23] p Pgp ≅ ρPgpπ

e

2/3  MW   MWc 

∆G  MW  +  kBT  MWc 

2/3

.

(12.10)

Each terms of Equation 12.10 are described in Figure 12.4. From Equation 12.10, the condition p Pgp = 1 provides the conditions required for any membrane-embedded drugs to be extruded. Thus, it is possible to define the critical surface density of transporter for total removal of every membraneembedded drug:

ρ

c Pgp

1  ∆G  MW  ≅  +  π  kBT  MWc 

2/3

MW   −  MW  c e 

2/3

.

(12.11)

Given the two-dimensional random walk hypothesis (Fig. 12.4b), Equation 12.11 corresponds to the “state equation” of drug bioavailability or resistance (as transporters are involved). Thus, Equation 12.11 shows how fundamental parameters, related to drugs and transporters, have to be linked together to trigger drug extrusion from the membrane. It follows that changing any of Equation 12.11’s parameter would increase, or decrease, the ability of transporters to extrude drugs. This means therefore that the expression level of membrane transporters has to specifically “match” a particular drug and a particular lipid asymmetry. It follows then that when two or more chemicals are involved, for example, when an intestinal bolus that may contain thousands of chemicals is considered, the total extrusion of all these chemicals from the membrane is unlikely and thus, some chemicals will penetrate cells but at lower rates. This latter point is in agreement with the recognized physiological role of transporters, namely avoiding CYP450s saturation. Therefore, the parameter that needs to be used and that best conjugates both the ability of the membrane to restrain the transverse movement of chemicals and the active pumping mediated by membrane transporters, is the escape rate (or influx) of membrane-embedded drugs into the cytosol (Jdrug), written as [23]

∆G

J drug

MW 2 / 3

  4D − −  = 2 e kBT  MWc  h

ρPgp    1 − ρ c  . Pgp

(12.12)

269

Lipinski’s Second Law and Potential Application

c In Equation 12.12, ρPgp , D, and h represents, respectively, the critical surface density of transporter needed to block the transverse movement of drugs given by Equation 12.11, the membrane diffusion coefficient, and the membrane thickness. Finally, Equation 12.12 plotted in Figure 12.5a demonstrates that both the transporter and the MW matter when drugs have to bypass barriers. There are two biological parameters that can be used to modulate the intracellular accumulation of drugs: the expression level of transporters and the difference in surface tensions (lipid packing asymmetry between leaflets). The latter parameter is biologically determined by cells via their own lipid metabolism. Given that the critical MW “MWc” in Equation 12.12 is defined by Equation 12.8 and is intimately dependent on lipid asymmetry, it may well be

rPgp

(a)

c rPgp

(b)

= 0.1

Control

0.4

Jdrug 4D / h 2

∆G e kBT

0.3

+PS

0.2 rPgp 0.1 r c = 0.9 Pgp

×1/2

×1/2 0

0.5

1

1.5

2

2.5

+PC

3

MW MWc

Figure 12.5.â•‡ Influence of membrane lipid composition in multidrug resistance (MDR). (a) Representation of Equation 12.11 in which two extreme cases are shown, low (blue curve) and high (red curve) surface density of transporters. At low surface density of transporters, the MW strongly affects the transverse movement of drugs across the membrane. At high surface density of transporters, the drug MW only plays a minor role. The dashed curves represent the predicted drug flow into cells if the lipid asymmetry was divided by a factor 2 (“×1/2” in the figure). In each case, a decrease in the lipid number asymmetry would facilitate the transverse movement of drugs across the membrane even in the presence of transporters. (b) MCF-7R cells (breast-derived cancer cells) resistant to doxorubicin were incubated with phosphatidylcholine (+PC) or phosphatidylserine (+PS). The cells were then challenged using doxorubicin at a concentration of 7â•›µM for 30 minutes prior to microscopy observation (0.7â•›µM of doxorubicin was used to maintain the resistance level in culture condition). The results demonstrate that the intracellular accumulation of drugs is affected by membrane lipid asymmetry in qualitative agreement with (a). In this case, PS is expected to increase the inner leaflet packing of lipids, thus increasing the resistance to drugs, whereas PC should generate the opposite effect. Color version on the Wiley web site.

270

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

that changing this asymmetry, or at best, altering the lipid metabolism, would increase the flux of drugs and thus their bioavailability (Fig. 12.5b). 12.4.3â•… On the Relationship between Cross-Resistance to Drugs and Lipinski’s Second Law As discussed previously, drug bioavailability and drug resistance display very strong similarities. It is therefore interesting to determine whether the set of equations from the previous paragraph can be applied to describe drug resistance, at least in well-defined and specific cases. There are different ways of generating drug resistance in cells. The first involves the transfection of cells with transporters, whereas the second way uses a drug as selective pressure. Accordingly, it is demonstrated that drug resistance levels increase when cells are incubated with an increasing concentration of a drug [34]. This result is related to the expression of membrane transporters. In particular, a very interesting case arises when drug resistance is selected via incubation of a large MW drug. In these instances, as the drug is large, the plasma membrane mechanical properties are expected to play a significant role by impairing the drug’s transverse movement. Thus, Equation 12.11 predicts that drug resistance extrusion should only require a relatively small amount of membrane transporters. For example, assuming that actinomycin D (MWâ•¯=â•¯1240) is used to generate the resistant phenotype, then the surface density of transporters needed to trigger drug resistance should be ∼3% (using Eq. 12.11 and assuming that the dehydration energy of actinomycin D is typically ΔG/kBTâ•¯∼â•¯10). If cross-resistance to drugs is now considered and that drugs with smaller MW (MWâ•¯<â•¯1240) are used, it follows that these new compounds should cross the membrane more easily. Accordingly, the meeting probability between small drugs and transporters should be low. In due course, only the drug dehydration energy and MW would matter. In this case, using Equation 12.12, the predicted influx of drugs should be

Jdrug ∼ e

−

∆G  MW  2 / 3 −  kBT  MWc 

.

(12.13)

As drugs’ effectiveness is dependent on their ability to cross the membrane, Equation 12.13 should also predict drug resistance levels given the MW and dehydration energy of chemicals (Fig. 12.6 and Table 12.1). Therefore, Lipinski’s second law applications seem to span beyond the drug/bioavailability field, which suggests that there is a real need for a conceptual unification of drug bioavailability, delivery, and resistance involving the membrane. 12.5â•… CONCLUSION The first study that has highlighted the membrane as a biomechanical object dates from the 1970s [39, 40]. Since then, a huge effort has been made by many

271

Conclusion

Ln(Ln(RRL))

2 1.5

~0.59 ¥ Ln(MW)

1

(R2 = 0.84)

0.5 0 5.5

6

6.5

7

7.5

Ln(MW) Figure 12.6.â•‡ Influence of membrane mechanical properties in multidrug resistance (MDR). As early as the 1970s, it was suggested that the MW of a drug is a central parameter when chemicals cross membrane bilayers. The leading study done by Biedler and Riehm in multidrug-resistant cells demonstrated that, statistically speaking, the MW of chemicals can explain up to 80% (i.e., R2â•¯≅â•¯0.80) of cross-resistance levels to drugs. Their data is reported in Table 12.1. In this study, the authors selected the initial level of resistance using actinomycin D, a large MW chemical [34]. Subsequently, they determined the relative cross-resistance using drugs with a smaller MW than actinomycin D. Experimentally, they determined the ratio of ED50 between drug-resistant and -sensitive cells (ED50 stands for the effective dose to kill 50% of a cell population). The results they obtained are plotted as the double natural logarithm (“Ln(Ln(.â•¯.â•¯.))”) of relative (to the control) resistance levels (RRLs) against the single natural logarithm of MW of drugs they used (see Table 12.1). Although the authors initially used a straight line to fit their data, their experimental results are best fitted by a power law under the form ∼MW0.59±0.11 (R2â•¯=â•¯0.846). This result confirms that, statistically, ∼85% of crossresistance levels to drugs are associated with the MW of drugs. The above power law is similar to the one given by Lipinski’s second law (see Eq. 12.8: 3/2â•¯∼â•¯0.66). Application of Equation 12.8 for Biedler and Riehm’s results can be justified as follows. The ED50 is an indicator of drug efficiency as it provides the drug concentration needed to kill 50% of cells in a given population. Accordingly, the higher the ED50, the lower the sensitivity to drugs and thus the lower the drugs’ influx. Therefore, there exists an inverse relationship between the ability of drugs to cross the membrane of cells and the ED50s given by Ln(ED50)â•¯∼â•¯Ln(1/J)â•¯∼â•¯(MW/MWc)2/3 (see Eq. 12.13 and assuming also that the drugs’ dehydration energy has a minimal effect on their transbilayer movement compared with their MW).

to focus on the role of these biomechanical properties in basic biology. This chapter had two main goals. The first was to demonstrate that, thanks to the research in the field relating to lipid asymmetry mediated by a lipid flippase, it is possible to determine a law regarding the MW cutoff as initially discovered by Lipinski and collaborators, which, it is hoped, will enable a better understanding of drug bioavailability, potentially providing new drug targets. The second aim of this chapter was to demonstrate that as biology is a challenging and complex area, there is a constant need for studies at the interface

272

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

TABLE 12.1.â•‡ Influence of Membrane Mechanical Properties in Multidrug Resistance (MDR) Drugs Actinomycin D Mythramycin Vinblastine Vincristine Daunomycin Puromycin

Ln(MW)

Ln(Ln(RRL))

7.13 6.99 6.81 6.82 6.27 6.15

1.78 1.87 1.70 1.66 1.49 1.21

Ln(Ln(RRL)), double natural logarithm of relative (to the control) resistance levels (RRLs); Ln(MW), natural logarithm of drug MW. For references, see text.

between sciences and that, together, scientists from different backgrounds can defy important scientific challenges. ACKNOWLEDGMENT This work has been supported by the University of Nottingham (NRF4305).

ABBREVIATIONS a D Δσ ΔG δN

cross-section area of drug in the membrane diffusion coefficient difference in surface pressures/tensions drug dehydration property number of phospholipids in excess in the inner leaflet (compared with the outer one) δN/N0 lipid asymmetry fraction between the inner and outer leaflet h membrane thickness J flux K elastic modulus of leaflet kB Boltzmann’s constant kc membrane bending modulus MW molecular weight N0 average number phospholipids in each leaflet NPgp number of transporters R vesicle radius S neutral surface of the membrane

References

Scell σ σ0 SPgp T ρPgp

273

cellular surface surface pressure/tension of a leaflet (monolayer) mean surface pressure/tension of the bilayer cell membrane membrane cross-section area of transporters temperature in Kelvin surface density of transporters

REFERENCES J. Drews, Drug Discov. Today 2003, 8, 411–420. T. I. Oprea, Mol. Divers. 2002, 5, 199–208. U. Brinkmann, M. Eichelbaum, Pharmacogenomics J. 2001, 1, 59–64. P. H. Marathe, A. D. Rodrigues, Curr. Drug Metab. 2006, 7, 687–704. G. A. Kullak-Ublick, M. B. Becker, Drug Metab. Rev. 2003, 35, 305–317. A. Avdeef, Curr. Top. Med. Chem. 2001, 1, 277–351. A. Golebiowski, S. R. Klopfenstein, D. E. Portlock, Curr. Opin. Chem. Biol. 2001, 5, 273–284. â•‡ 8â•… C. A. Lipinski, F. Lombardo, B. W. Dominy, P. J. Feeney, Adv. Drug Deliv. Rev. 2001, 46, 3–26. â•‡ 9â•… R. J. Quinn, A. R. Carroll, N. B. Pham, P. Baron, et al., J. Nat. Prod. 2008, 71, 464–468. 10â•… J. R. Proudfoot, Bioorg. Med. Chem. Lett. 2002, 12, 1647–1650. 11â•… D. F. Veber, S. R. Johnson, H. Y. Cheng, B. R. Smith, et al., J. Med. Chem. 2002, 45, 2615–2623. 12â•…K. Palm, P. Stenberg, K. Luthman, P. Artursson, Pharm. Res. 1997, 14, 568–571. 13â•… F. Winau, O. Westphal, R. Winau, Microbes Infect. 2004, 6, 786–789. 14â•… J. W. Bennett, K. T. Chung, Adv. Appl. Microbiol. 2001, 49, 163–184. 15â•… D. Raucher, M. P. Sheetz, Biophys. J. 1999, 77, 1992–2002. 16â•… F. M. Hochmuth, J. Y. Shao, J. Dai, M. P. Sheetz, Biophys. J. 1996, 70, 358–369. 17â•… W. Rawicz, K. C. Olbrich, T. McIntosh, D. Needham, E. Evans, Biophys. J. 2000, 79, 328–339. 18â•… C. Rauch, E. Farge, Biophys. J. 2000, 78, 3036–3047. 19â•… M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. 20â•… E. Farge, D. M. Ojcius, A. Subtil, A. Dautry-Varsat, Am. J. Physiol. 1999, 276, C725–C733. 21â•… E. Farge, Biophys. J. 1995, 69, 2501–2506. 22â•… J. Dai, M. P. Sheetz, Biophys. J. 1999, 77, 3363–3370. 23â•… C. Rauch, A. Pluen, Eur. Biophys. J. 2007, 36, 121–131. 24â•… A. H. Schinkel, Adv. Drug Deliv. Rev. 1999, 36, 179–194. 25â•… F. J. Sharom, Biochem. Cell Biol. 2006, 84, 979–992. 26â•… J. Ferte, Eur. J. Biochem. 2000, 267, 277–294. â•‡ 1â•… â•‡ 2â•… â•‡ 3â•… â•‡ 4â•… â•‡ 5â•… â•‡ 6â•… â•‡ 7â•…

274

PHYSICAL BIOLOGY OF DRUG TRANSVERSE MOVEMENT

27â•… A. B. Shapiro, V. Ling, Eur. J. Biochem. 1997, 250, 122–129. 28â•… F. M. Sirotnak, C. H. Yang, L. S. Mines, E. Oribe, J. L. Biedler, J. Cell. Physiol. 1986, 126, 266–274. 29â•… A. Ramu, H. B. Pollard, L. M. Rosario, Int. J. Cancer 1989, 44, 539–547. 30â•… W. D. Stein, C. Cardarelli, I. Pastan, M. M. Gottesman, Mol. Pharmacol. 1994, 45, 763–772. 31â•… D. Nielsen, C. Maare, T. Skovsgaard, Biochem. Pharmacol. 1995, 50, 443–450. 32â•… A. B. Shapiro, A. B. Corder, V. Ling, Eur. J. Biochem. 1997, 250, 115–121. 33â•… A. B. Shapiro, V. Ling, Eur. J. Biochem. 1997, 250, 130–137. 34â•… J. L. Biedler, H. Riehm, Cancer Res. 1970, 30, 1174–1184. 35â•… T. Litman, T. Zeuthen, T. Skovsgaard, W. D. Stein, Biochim. Biophys. Acta 1997, 1361, 159–168. 36â•… J. M. Zamora, H. L. Pearce, W. T. Beck, Mol. Pharmacol. 1988, 33, 454–462. 37â•… H. L. Pearce, M. A. Winter, W. T. Beck, Adv. Enzyme Regul. 1990, 30, 357–373. 38â•…K. Dano, Biochim. Biophys. Acta 1973, 323, 466–483. 39â•… M. P. Sheetz, R. G. Painter, S. J. Singer, J. Cell Biol. 1976, 70, 193–203. 40â•… M. P. Sheetz, S. J. Singer, Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 4457–4461. 41â•… J. Rudnick, G. Gaspari, Elements of the Random Walk, Cambridge University Press, Cambridge, 2004.

13 ENDOCYTOSIS AND LIPID ASYMMETRY Nina Ohlwein Institut de Biologie Physico-Chimique, Paris, France Department of Biology, Humboldt University of Berlin, Berlin, Germany

Andreas Herrmann Department of Biology, Humboldt University of Berlin, Berlin, Germany

Philippe F. Devaux Institut de Biologie Physico-Chimique, Paris, France

13.1â•… INTRODUCTION Vesicular traffic to and from the plasma membrane plays an important role in the exchange of molecules within a eukaryotic cell and with its environment. During endocytosis, the formation of vesicles out of a comparably flat plasma membrane requires a high local membrane curvature. Although many proteins involved have been identified, the molecular mechanism remains unclear. Therefore, understanding the mechanisms of generating membrane curvature became one of the major subjects of intracellular trafficking research [1, 2]. Endocytic engulfment causes significant area changes of the membrane leaflets with respect to each other. Due to limited lateral elasticity as well as negligible spontaneous lipid translocation, specific mechanisms are required to induce membrane curvature [3]. In the course of protein-related pathways, several individual proteins such as dynamin, Bin/Amphiphysin/Rvs

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

275

276

ENDOCYTOSIS AND LIPID ASYMMETRY

(BAR)-domain-containing proteins, and epsin can induce the required local curvature [4]. Alternatively, membrane curvature can be caused by differences between the surface areas of the two monolayers related to leaflet-specific modulation of lipid composition. Thus, it was proposed that ATP-dependent lipid translocators pumping phospholipids from one membrane leaflet to the opposite leaflet account for generation of altered area relation between both leaflets as an early step in endocytosis [5] (see also Chapter 9). In agreement with such a hypothesis, it was shown that endocytosis can be enhanced or reduced by increasing the amount of phospholipids and, hence, the surface area of the inner and outer leaflet, respectively, with respect to the opposite leaflet [6]. Moreover, endocytosis in yeast cells was inhibited by knocking out putative ATP-dependent lipid translocators [7]. By now, this hypothesis has been supported by several studies establishing a relationship between the activity of lipid translocators and vesicle formation, and it seems commonly accepted that, in particular, aminophospholipid translocases (APLTs) are involved in endocytosis. Here, we will briefly review various mechanisms leading to membrane bending and will subsequently summarize how membrane curvature can be induced in particular by the transversal lipid asymmetry of the plasma membrane. Furthermore, the studies supporting an involvement of lipid translocators in endocytosis and two models of how flippases could be connected with the generation of endocytic invaginations will be discussed. 13.2â•… BENDING A MEMBRANE A crucial moment in the initial phase of endocytosis is the bending of a comparatively flat plasma membrane to generate the high curvature necessary for endocytic invaginations. There are several mechanisms to induce temporary high curvature, which can be divided into two categories: Curvature is achieved either by the interplay between proteins and lipids or by leaflet-specific changes in the lipid composition of the membrane. Proteins can interact in various ways with membranes to stabilize or generate curvature, such as scaffolding by peripheral membrane proteins having an intrinsic static geometry or active helix insertion into membranes (for reviews, see References 1 and 4). For the latter mechanism, the effectiveness of spontaneous membrane curvature caused by hydrophobic insertion of different helices was characterized in theoretical predictions, and it was determined that inclusions shaped like shallow rods are very efficacious in membrane shaping [8]. The interaction of proteins bending membranes by insertion of their amphiphatic helices is dependent on the lipid composition of the respective membrane area [9]. Thus, particular curvature-generating proteins can be recruited to specific endocytic pathways due to different lipid compositions of endocytic buds or vesicles. Many curvature-generating or -sensing proteins are known to be involved in endocytic pathways (for a review, see Reference 10). For instance, epsin as

Bending a Membrane

277

well as BAR, N-BAR, and F-BAR (Fes/Cip4 homology-BAR or EFC for extended FCH homology) protein families are well-known examples of proteins that are involved in the formation of clathrin-coated pits during clathrinmediated endocytosis and can generate, sense, and stabilize membrane bending. The amphiphatic helix of epsin inserts into the proximal leaflet of a lipid bilayer, which causes a displacement of lipid head groups and a reorientation of acyl chains. Upon binding directly to inositol lipids and the simultaneous insertion of the helix, epsin generates membrane curvature by increasing the protein-proximal monolayer area relatively to the other layer. In this process, formation of the amphipathic helix in epsin is thought to be coupled to inositol lipid binding [11, 12]. Proteins containing a concavely shaped BAR domain bind to negatively charged membrane surfaces mainly due to electrostatic interactions. In addition to a BAR domain, some proteins contain an aminoterminal amphipathic helix (N-BAR domain), which inserts into the membrane like the helix of epsin. F-BAR proteins have a more shallow curvature than BAR proteins and seem to deform membranes into structures with a wider diameter. Several F-BAR proteins have an established role in binding phospholipids and are known to be involved in actin dynamics at the plasma membrane. In this way, the actin polymerization apparatus and the machineryregulating membrane dynamics are linked by these proteins [13]. The subsequent fission of clathrin-coated invaginations and their release into the cytoplasm is mediated by dynamin. The membrane scission protein dynamin, a large, modular guanosine triphosphate (GTP)ase, is recruited by curvaturesensing proteins and has domains that support binding to phosphatidylinositol 4,5-bisphosphate. By creating a GTP-dependent conformational change into helical rings, it induces fission by tracing the small neck of the coated bud like a scaffold and ties up the membrane into tubular structures [4, 14]. Additionally, the scission of the coated vesicle could be supported by a redistribution of membrane lipids. Boundary forces at model membranes were shown to result in a phase separation of lipids, creating a breaking point that severs when longitudinal forces are applied [15, 16]. For clathrin-independent pathways, the individual roles of proteins that are noted for their ability to induce and stabilize membrane bending and curvature are only beginning to be known [17]. The membrane scission protein dynamin was shown to be involved in caveolae- and ras homolog gene family, member A (RhoA)-regulated pathways, but it is not crucial for cell division control protein 42 homolog (Cdc42)- and ADP-ribosylation factor (ARF)6regulated endocytosis. In most cases, this has been determined by overexpression of individual dynamin-1 or dynamin-2 mutants. Other methods are needed to assess and confirm the effects of the various dynamin isoforms on clathrinindependent internalization, such as the use of more specific inhibitors of dynamin [18]. Thus, it could turn out in future studies that the role of dynamin has to be reconsidered. Likewise, the GTPase-regulated clathrin-independent pathways were thought to depend on actin. These GTPases are known to interact with the actin machinery, and it seems likely that further investigations

278

ENDOCYTOSIS AND LIPID ASYMMETRY

will show the role of actin for all clathrin-independent pathways [19, 20]. Furthermore, the activation of the ARF family GTP-binding proteins ARF1 and ARF6, where the latter one is the key regulator for a clathrin-independent endocytic pathway, was suggested to be dependent on membrane curvature [21]. ARF family GTP-binding proteins serve as regulatory proteins for numerous cellular processes, such as membrane-trafficking events and maintenance of organelle structure. Thus, these proteins are recruited to curved membranes and feedback to further enhance this curvature, thereby linking sensing and generation of membrane curvature as well as initiating curvature that is further enhanced by endocytic coat proteins during vesicle formation [22]. Apart from this role of membrane-bending proteins, it was proposed that chiral and tilted membrane lipid molecules alone are sufficient to generate membrane curvature during clathrin-independent endocytosis [23]. Microdomains seemed to be involved in most, probably even all, clathrin-independent pathways, and the molecular chirality and tilt of the raft-forming constituents could favor a bent formation of the membrane. This induction of curvature could then allow the recruitment of proteins that sense and stabilize curvature, such as BARdomain-containing proteins [24]. Alternatively, membrane curvature can be induced by a difference between the surface areas of the two monolayers caused by leaflet-specific changes of lipid composition [3], which constitutes the mechanism of the well-established bilayer couple model [25]. Upon insertion of additional molecules in one of the bilayer leaflets, spectacular shape changes in membranes of red blood cells were observed (see Chapter 2). Further experimental verification of this model was performed in biological as well as in model systems. Addition of phospholipids led to a considerable change in the initial shape of platelets from a discoid form to a smaller body with very long pseudopods [26]. In giant unilamellar vesicles (GUVs) of pure lipid membranes, the creation of small area differences between the vesicle monolayers was sufficient to drive vesicle shape changes, which had a character of budding invaginations [27–29] (see Chapter 2). The structure difference of the resulting shape changes can be explained by the existence of intracellular structures, such as the cytoskeleton, which provide an increase of the bilayer lateral tension [30]. 13.3â•… SHAPE CHANGES OF GUVs INDUCED BY LIPID ASYMMETRY Theoretical foundations for shape changes in vesicles or cells were already suggested in 1970: It was demonstrated that the biconcave shape of human red blood cells coincides with a minimum value of membrane-bending energy [31]. Correspondingly, two different models for the generation of curvature were proposed: the bilayer couple model as mentioned above [25] and, 2 years later, the spontaneous curvature model [32]. In 1991, these two models were compared as follows: Shape transformations of pure lipid bilayers were theoreti-

SHAPE CHANGES OF GUVs INDUCED BY LIPID ASYMMETRY

279

cally predicted by determining the phase diagram for each of the two models [33]. For the bilayer couple model, they found an extreme sensitivity to an asymmetry in the monolayer area. This sensitivity could be supported in various experimental studies by using different ways to alter the monolayer area in giant liposomes. For example, by varying the temperature, a sequence of shape changes was observed and explained by a small asymmetry in the thermal expansivities of the two monolayers [34]. Alternatively, area differences between the two leaflets can be achieved by the manipulation of phospholipids in each monolayer. Two different procedures were used for this purpose: either a small quantity of egg phosphatidylglycerol was redistributed through the bilayer by means of a pH gradient or lysophosphatidylcholine was added or removed from the external leaflet of GUV [27]. The experiments confirmed the general nature of the bilayer couple model by generating important shape changes of nonspherical GUV due to a redistribution of lipids as well as an addition or depletion of lipids from one monolayer. In another study, the external monolayer area of GUV was altered by the addition of ceramides with different acyl chains, which triggered shape changes from prolate to pearshaped vesicles [28]. In addition to the generation of curvature by the creation of area differences between vesicle monolayers, it could be shown that triggered shape transitions are reversible in the presence of translocator proteins that allow lipids to redistribute between the two membrane leaflets [29]. In this study, the formation of a bud-like structure, which was induced by a very small excess of lipid in the outer monolayer, was stable under conditions of negligible flipflop. Upon reconstitution of the energy-independent flippase activity of the yeast endoplasmic reticulum into GUV, the initial bud formation was shown to be reversible, and the prolate shape was recovered (see Chapter 6). This was ascribed to a rapid flip-flop leading to relaxation of the monolayer area difference. These studies with artificial systems, in which the generation of monolayer area asymmetry in unilamellar vesicles is controllable, are very helpful in understanding the physical properties of a membrane necessary for effective membrane bending. However, the size and number of the obtained budding invaginations in GUV is rather different from vesicles in living cells, for instance, endocytic vesicles. In all of the above-mentioned studies, only one bud per liposome could be obtained, and the bud size was not more than one order of magnitude smaller than the size of the initial vesicle. Recently, it was shown theoretically that lateral tension is required to cause vesiculation comparable to the building of intracellular transport vesicles [30]. Considering a closed membrane subjected to lateral tension and a difference in the monolayer spontaneous areas, the results of the analysis predicted the existence of three different vesiculation regimes: one regime with no vesicle formation, a single vesicle regime, and a multiple vesicle regime. The respective phase diagram demonstrated that the larger the tension, the larger the monolayer area asymmetry needed to generate a vesicle. The model also predicted that

280

ENDOCYTOSIS AND LIPID ASYMMETRY

in the absence of lateral tension, the monolayer area asymmetry can only generate a single vesicle, while application of sufficient asymmetry in the presence of a low lateral tension results practically always in the generation of multiple vesicles. Thus, this theory supports the hypothesis that a difference between the two monolayer areas composing a membrane, which is under lateral tension as it is the case for biological membranes, is sufficient for the formation of multiple, small vesicles. 13.4â•… HOW ENDOCYTOSIS IS LINKED TO LIPID ASYMMETRY On the basis of the above-described mechanism to induce membrane curvature, it was proposed that APLTs account for generation of altered area relation as an early step in endocytosis. Upon the translocation of lipids from the exoplasmic leaflet to the cytoplasmic layer of the plasma membrane, APLT can mediate an increase of the inner surface area and thereby support the generation of membrane curvature during endocytosis. First mentioned in 1991 [5], this hypothesis was supported in several studies, and by now, it seems commonly accepted that APLT are implicated in endocytosis [35–37]. In the following, we will discuss studies supporting this hypothesis. Already in 1994, it was shown that the incorporation of phospholipid analogs into the plasma membrane of human erythrocyte ghosts affected the ATP-induced vesiculation [38]. Erythrocyte ghosts were incubated in the presence of ATP at 37°C, and the percentage of the total membrane taken in as vesicles was quantified by determining the loss of acetylcholinesterase activity after definite times. The addition of spin-labeled phosphatidylcholine or sphingomyelin, which remain preferentially in the outer leaflet of the membrane, to ATP-containing erythrocyte ghosts inhibited the formation of the so-called endocytic vesiculation immediately after the addition, and vesiculation remained reduced in further incubation for about 40 minutes at 37°C. Interestingly, the addition of spin-labeled phosphatidylserine (PS), a substrate of APLT, also reduced the vesiculation directly after addition, but subsequent incubation for about 20 minutes at 37°C abolished the suppression of vesiculation, reaching approximately the extent of the control after 60 minutes. The decrease of vesiculation immediately after the addition as well as the increasing kinetics after further incubations were dependent on the amount of analog added. Further investigations of kinetics of redistribution of phospholipid analogs in the plasma membrane by using a back-exchange assay showed that the enhancement of vesiculation after PS addition was comparable to the kinetics of inward translocation of PS mediated by APLT. Approximately 40% of the analog was translocated to the inner layer within 30 minutes at 37°C. Thus, the incorporation of phospholipids in the outer layer of the plasma membrane inhibited endocytic-like processes, while the redistribution of phospholipid analogs from the outer to the inner leaflet restored the activity of

How Endocytosis Is Linked to Lipid Asymmetry

281

vesiculation, which supports the hypothesis of the implication of APLT in endocytosis. More recently, it was shown in several studies that APLT are required for vesicle formation in yeast cells (see Chapters 8 and 9). Yeast lacking the plasma membrane-associated P4-ATPases Dnf1p and Dnf2p displayed an abolished ATP-dependent transport of (7-nitro-2-1,3-benzoxadiazol-4-yl)amino- (NBD-) labeled phospholipids from the outer to the inner plasma membrane leaflet [7]. The loss of Dnf1p and Dnf2p also caused an exposure of endogenous phosphatidylethanolamine (PE) to the outer leaflet, which was twofold higher than in wild-type cells. Furthermore, the activity of bulk flow and receptormediated endocytosis in yeast cells with genomic deletions of Dnf1p, Dnf2p, and Drs2p (ATPase II homolog in the yeast Golgi) was investigated in this study. When cells were stained with a marker for bulk flow endocytosis, the internalization and delivery to the vacuole of the marker was essentially inhibited in the triple mutant. Likewise, the uptake of a specific factor to follow receptor-mediated endocytosis was greatly reduced in cells lacking all three translocators. Collectively, these results indicated that a loss of Dnf1p, Dnf2p, and Drs2p caused a general defect in the internalization step of endocytosis. In agreement with these findings, in other studies with yeast cells, it was determined that the inactivation of P4-ATPase Drs2p rapidly inhibited translocation of fluorescently labeled PS and lowered the formation of vesicles [35], and the loss of the P4-ATPases Drs2p and Dnf3p disrupted aminophospholipid translocation and asymmetry in post-Golgi secretory vesicles [39]. Furthermore, the Caenorhabditis elegans Drs2p homolog TAT-1 is required for yolk uptake in oocytes and an early step of fluid-phase endocytosis in the intestine [40]. Also, endocytic activity of eukaryotic cells was shown to be accelerated by the addition of exogenous lipids actively transported from the exoplasmic to the cytoplasmic leaflet [6]. Erythroleukemia K562 cells are derived from human erythroblasts, a precursor of erythrocytes, and have a high activity of the APLT [41]. To study the influence of altered area relation of the plasma membrane on endocytic activity, various exogenous lipids were added to K562 cells. By the direct addition of short-chain lipids that have sufficient water solubility to the cells in suspension, the lipids were incorporated into and, depending on the lipid species, subsequently translocated across the plasma membrane. When short-chain aminophospholipid phosphatidylethanolamine (C6-PE) or serine (C6-PS) was added, an enhancement of endocytic activity was obtained. For the latter lipid, this stimulation increased with the amount of added lipids. In contrast, the addition of lysophosphatidylserine or poly(ethylene)50-glycolcholesterol, which remain on the outer layer, caused a partial inhibition of endocytic activity. All experiments were performed after an incubation of cells with exogenous lipids for 30 minutes at 37°C, and the results were explained by the changes of surface area relation in the plasma membrane due to the addition of exogenous lipids. However, great care should be taken when cells are incubated with exogenous lipids for such a long time.

282

0 minute

ENDOCYTOSIS AND LIPID ASYMMETRY

30 minutes at 4°C

30 minutes at 37°C

Figure 13.1.â•‡ Distribution of C6-NBD-PS in K562 cells after incubation at different temperatures. Four percent of C6-NBD-PS was added to K562 cells, and confocal microscopic pictures were taken immediately after the addition of lipids and after 30 minutes when cells were incubated with lipids either at 4°C or at 37°C. As shown in the left image, immediately after the addition C6-NBD-lipids were only present in the plasma membrane. After 30 minutes at 4°C, the main part was still in the plasma membrane and only a slight dying of intracellular structures could be observed (middle). After 30 minutes at 37°C, the lipids were dispersed in the whole cell and seemed to accumulate in inner cell compartments (right). Staining of the plasma membrane was strongly reduced in comparison to 30 minutes at 4°C. This shows that, after an incubation of 30 minutes at 37°C, the main part of added exogenous lipids was not present in the plasma membrane anymore and could therefore not account for a difference in the monolayer area asymmetry. Color version on the Wiley web site.

After 30 minutes at 37°C, the majority of exogenously added phospholipids is most likely no longer present in the plasma membrane but released into other compartments within the cell (Fig. 13.1). This means that the majority of exogenous added lipids, or at least a large amount, cannot account for an increased surface area of one of the two layers, because they are most likely already released into the cell after such an incubation. Thus, the strong influence of exogenous phospholipids on endocytic activity, as obtained in the study mentioned above [6], could only partly be explained by an alteration of surface areas of the two monolayers composing the plasma membrane bilayer. Another possible explanation for a stimulating effect of exogenous aminophospholipids on endocytosis, which may supplementarily enhance endocytosis, is an activation of endocytosis due to the modified lipid composition of the plasma membrane. The plasma membrane of mammalian cells consists of about 15% PS in relation to the total phospholipid amount of the plasma membrane (see Chapter 3), and an addition of 4% exogenous C6-PS increases this amount to nearly 20%. Cells are able to regulate the activity of endocytic pathways for adaptation to specific conditions. Thus, it is most likely that the increased amount of PS in the plasma membrane leads to an upregulation of endocytic activity in order to retrieve the normal lipid composition of the plasma membrane. By the internalization of plasma membrane parts and the subsequent

ROLE OF P4-ATPases IN THE FORMATION OF ENDOCYTIC INVAGINATIONS

283

sorting and/or degradation of PS, the cell can gain back the normal lipid composition. 13.5â•… ROLE OF P4-ATPases IN THE FORMATION OF ENDOCYTIC INVAGINATIONS Currently, two models have been suggested to explain the implication of APLT in endocytosis [37]. In the first model, P4-ATPases are thought to support the recruitment of the vesicle-budding machinery to the plasma membrane through protein–protein interactions and/or by increasing the concentration of specific lipids in the cytoplasmic leaflet (see Chapters 8 and 9). The hypothesis of recruitment by protein–protein interactions is supported by the following observation: the yeast P4-ATPase Drs2p directly interacts with Gea2p, an ARF activator, that regulates recruitment of ARF, adaptor protein-1, and clathrin-network proteins to trans-Golgi membranes [42]. More recently, it was shown that Drs2p acts independently of coat recruitment to facilitate clathrin-dependent vesicle budding from the trans-Golgi network [43]. However, these data indicate that, in the absence of Drs2p, the recruitment of accessory proteins to the membrane would not result in the formation of clathrin-coated vesicles. This underlines the general assumption of interaction of P4-ATPases with endocytic regulators. The second model is based on a more direct and mechanical role in the formation of invaginations: the generation of membrane area asymmetry as outlined above. Upon translocating lipids from the exoplasmic to the cytoplasmic leaflet of the plasma membrane, the translocators create an imbalance in the surface areas of the two monolayers, which leads to bending of the membrane (see above for examples). The specific activation or boost of activation of APLT in membrane regions, where endocytic invaginations need to be induced, could be provided by protein–protein interactions of APLT with regulatory proteins as part of endocytic pathways (Fig. 13.2). For instance, it was shown that RhoA, a small GTPase holding a central role in a clathrinindependent pathway, is involved in the regulation of the localization of Na+/ K+-ATPase in the plasma membrane of epithelial cells [44]. By microinjection of the constitutively active mutant of RhoA, the Na+/K+-ATPase was translocated to the spike-like protrusions over the apical surfaces. Supporting these findings, in a later study, it was determined that activation of RhoA is necessary for the endocytosis of Na+/K+-ATPases in cells exposed to hypoxia [45]. Similar to Na+/K+-ATPases, P4-ATPases are activated by specific subunits [37]. Interestingly, Cdc50 proteins, chaperones of P4-ATPases, are strikingly reminiscent of the β-subunit of the Na+/K+-ATPase. Therefore, it is conceivable that P4-ATPases interact in a comparable way with regulatory proteins of endocytic pathways. This hypothesis is supported by the recent findings that Cdc50p is required for the APLT activity of purified and reconstituted Drs2p [46] and that three potential binding partners of Drs2p are involved

284

ENDOCYTOSIS AND LIPID ASYMMETRY Cdc50

P4-ATPase

Cytoplasm ATP ARF-activator/GEF ARF/GTPase

Protein coat

Figure 13.2.â•‡ Suggested role of P4-ATPases in the formation of endocytic invaginations. The model is based on the mechanical role of P4-ATPases in the formation of vesicles. By the translocation of lipids from the exoplasmic to the cytoplasmic leaflet of the plasma membrane, the proteins generate a difference of the monolayer area asymmetry, which induces membrane curvature. Additionally or alternatively, P4-ATPases are thought to support the recruitment of the vesicle-budding machinery to the plasma membrane by interacting directly and/or in an indirect way via subunits with endocytic regulators (Adapted from Reference 37, Fig. 3.).

in phosphoinositide metabolism [47]. Thus, interaction between Drs2p and a phosphatidylinositol-4-phosphatase, which is required for efficient recruitment of the vesicle budding machinery, could link APLT activity to coat recruitment for the formation of secretory vesicles. Furthermore, family members of the P4-ATPases were found to interact with cytoplasmic proteins, such as guanine nucleotide exchange factors (GEFs) and small GTPases [48, 49], which are crucial for the recruitment of coat proteins in endocytic pathways. This may lead to the recruitment of endocytic accessory proteins to sites of ATPase-dependent phospholipid translocation. 13.6â•… CONCLUDING REMARKS Summarizing the role of lipid asymmetry in the generation of membrane invaginations during endocytosis, it can be concluded that a difference between the areas of the two monolayers composing the bilayer can induce significant shape changes in biological as well as in pure lipid membranes. Furthermore, ATP-dependent specific flippases can alter the monolayer area differences of the plasma membrane. Additionally, there are several studies that support the

References

285

hypothesis of an involvement of flippases in endocytosis, but this awaits further experimental confirmation. Further experiments with, for instance, downregulated flippases or inactive mutants, could verify one of the described models, or even a combination of both. To date, most efforts to elucidate endocytic mechanisms have focused on involved proteins; the generation of membrane curvature, necessary for endocytic invaginations, has mainly been explained by proteins that can induce and stabilize curvature. The role of lipids and the properties of the membrane, which needs to be bent, have been largely pursued with less attention. In this regard, the role of lipids composing the plasma membrane may have been underestimated. Due to the physical properties of the plasma membrane, it seems unlikely that membranes are curved only by the force of proteins when other mechanisms could facilitate this procedure. ACKNOWLEDGMENTS The work of the authors was generously supported by grants from Marie Curie Early Stage Training (EST BioMem) and from the Deutsche Forschungsgemeinschaft (SFB 740).

ABBREVIATIONS APLT ARF BAR Cdc42 F/EFC- GEFs GTP GUVs NBD- PE PS RhoA

aminophospholipid translocase ADP-ribosylation factor Bin/Amphiphysin/Rvs cell division control protein 42 homolog Fes/Cip4 homology- or EFC for extended FCH homologyguanine nucleotide exchange factors guanosine triphosphate giant unilamellar vesicles (7-nitro-2-1,3-benzoxadiazol-4-yl)aminophosphatidylethanolamine phosphatidylserine ras homolog gene family, member A

REFERENCES â•‡ 1â•… J. Zimmerberg, M. M. Kozlov, Nat. Rev. Mol. Cell Biol. 2006, 7, 9–19. â•‡ 2â•…K. Powell, Nature 2009, 460, 318–320. â•‡ 3â•… P. F. Devaux, Biochimie 2000, 82, 497–509. â•‡ 4â•… H. T. McMahon, J. L. Gallop, Nature 2005, 438, 590–596.

286

ENDOCYTOSIS AND LIPID ASYMMETRY

â•‡ 5â•… P. F. Devaux, Biochemistry 1991, 30, 1163–1173. â•‡ 6â•… E. Farge, D. M. Ojcius, A. Subtil, A. Dautry-Varsat, Am. J. Physiol. 1999, 276, C725–C733. â•‡ 7â•… T. Pomorski, R. Lombardi, H. Riezman, P. F. Devaux, G. van Meer, J. C. Holthuis, Mol. Biol. Cell 2003, 14, 1240–1254. â•‡ 8â•… F. Campelo, H. T. McMahon, M. M. Kozlov, Biophys. J. 2008, 95, 2325–2339. â•‡ 9â•…G. Drin, B. Antonny, FEBS Lett. 2009, 584, 1840–1847. 10â•… R. Lundmark, S. R. Carlsson, Semin. Cell Dev. Biol. 2010, 21, 363–370. 11â•… M. G. Ford, I. G. Mills, B. J. Peter, Y. Vallis, G. J. Praefcke, P. R. Evans, H. T. McMahon, Nature 2002, 419, 361–366. 12â•… Y. Yoon, J. Tong, P. J. Lee, A. Albanese, N. Bhardwaj, M. Kallberg, M. A. Digman, H. Lu, E. Gratton, Y. K. Shin, W. Cho, J. Biol. Chem. 2009, 285, 531–540. 13â•… P. Aspenstrom, Int. Rev. Cell Mol. Biol. 2009, 272, 1–31. 14â•… Y. J. Chen, P. Zhang, E. H. Egelman, J. E. Hinshaw, Nat. Struct. Mol. Biol. 2004, 11, 574–575. 15â•… A. Roux, D. Cuvelier, P. Nassoy, J. Prost, P. Bassereau, B. Goud, EMBO J. 2005, 24, 1537–1545. 16â•… J. Liu, M. Kaksonen, D. G. Drubin, G. Oster, Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 10277–10282. 17â•… C. G. Hansen, B. J. Nichols, J. Cell Sci. 2009, 122, 1713–1721. 18â•… E. Macia, M. Ehrlich, R. Massol, E. Boucrot, C. Brunner, T. Kirchhausen, Dev. Cell 2006, 10, 839–850. 19â•… H. Girao, M. I. Geli, F. Z. Idrissi, FEBS Lett. 2008, 582, 2112–2119. 20â•…K. Sandvig, M. L. Torgersen, H. A. Raa, B. van Deurs, Histochem. Cell Biol. 2008, 129, 267–276. 21â•… R. Lundmark, G. J. Doherty, Y. Vallis, B. J. Peter, H. T. McMahon, Biochem. J. 2008, 414, 189–194. 22â•… J. G. Donaldson, Biochem. J. 2008, 414, e1–e2. 23â•… R. C. Sarasij, S. Mayor, M. Rao, Biophys. J. 2007, 92, 3140–3158. 24â•…G. J. Doherty, H. T. McMahon, Annu. Rev. Biochem. 2009, 78, 857–902. 25â•… M. P. Sheetz, S. J. Singer, Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 4457–4461. 26â•… A. Sune, A. Bienvenue, Biochemistry 1988, 27, 6794–6800. 27â•… E. Farge, P. F. Devaux, Biophys. J. 1992, 61, 347–357. 28â•… I. Lopez-Montero, N. Rodriguez, S. Cribier, A. Pohl, M. Velez, P. F. Devaux, J. Biol. Chem. 2005, 280, 25811–25819. 29â•… A. Papadopulos, S. Vehring, I. Lopez-Montero, L. Kutschenko, M. Stockl, P. F. Devaux, M. Kozlov, T. Pomorski, A. Herrmann, J. Biol. Chem. 2007, 282, 15559–15568. 30â•… P. F. Devaux, A. Herrmann, N. Ohlwein, M. M. Kozlov, Biochim. Biophys. Acta 2008, 1778, 1591–1600. 31â•… P. B. Canham, J. Theor. Biol. 1970, 26, 61–81. 32â•… H. J. Deuling, W. Helfrich, Biophys. J. 1976, 16, 861–868. 33â•… U. Seifert, K. Berndl, R. Lipowsky, Phys. Rev. A 1991, 44, 1182–1202.

References

287

34â•… E. Sackmann, H. P. Duwe, H. Engelhardt, Faraday Discuss. Chem. Soc. 1986, 81, 281–290. 35â•… T. R. Graham, Trends Cell Biol. 2004, 14, 670–677. 36â•… T. Pomorski, A. K. Menon, Cell. Mol. Life Sci. 2006, 63, 2908–2921. 37â•… C. F. Puts, J. C. Holthuis, Biochim. Biophys. Acta 2009, 1791, 603–611. 38â•… P. Muller, T. Pomorski, A. Herrmann, Biochem. Biophys. Res. Commun. 1994, 199, 881–887. 39â•… N. Alder-Baerens, Q. Lisman, L. Luong, T. Pomorski, J. C. Holthuis, Mol. Biol. Cell 2006, 17, 1632–1642. 40â•… A. F. Ruaud, L. Nilsson, F. Richard, M. K. Larsen, J. L. Bessereau, S. Tuck, Traffic 2009, 10, 88–100. 41â•… S. Cribier, J. Sainte-Marie, P. F. Devaux, Biochim. Biophys. Acta 1993, 1148, 85–90. 42â•… S. Chantalat, S. K. Park, Z. Hua, K. Liu, R. Gobin, A. Peyroche, A. Rambourg, T. R. Graham, C. L. Jackson, J. Cell Sci. 2004, 117, 711–722. 43â•…K. Liu, K. Surendhran, S. F. Nothwehr, T. R. Graham, Mol. Biol. Cell 2008, 19, 3526–3535. 44â•… A. Maeda, M. Amano, Y. Fukata, K. Kaibuchi, Biochem. Biophys. Res. Commun. 2002, 297, 1231–1237. 45â•… L. A. Dada, E. Novoa, E. Lecuona, H. Sun, J. I. Sznajder, J. Cell Sci. 2007, 120, 2214–2222. 46â•… X. Zhou, T. R. Graham, Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 16586–16591. 47â•… C. F. Puts, G. Lenoir, J. Krijgsveld, P. Williamson, J. C. Holthuis, J. Proteome Res. 2010, 9, 833–842. 48â•… C. Y. Chen, M. F. Ingram, P. H. Rosal, T. R. Graham, J. Cell Biol. 1999, 147, 1223–1236. 49â•… S. Wicky, H. Schwarz, B. Singer-Kruger, Mol. Cell. Biol. 2004, 24, 7402–7418.

PART VI APOPTOSIS AND DISEASES: CONSEQUENCES OF DISRUPTION TO LIPID TRANSMEMBRANE ASYMMETRY

14 MEMBRANE LIPID ASYMMETRY IN AGING AND APOPTOSIS Krishnakumar Balasubramanian University of Pittsburgh, Pittsburgh, PA

Alan J. Schroit The Department of Pharmacology, The University of Texas Southwestern Medical Center, Dallas, TX

14.1â•… INTRODUCTION Membrane lipids play key roles in cell physiology by participating in processes that are critical to cell survival and tissue homeostasis. Structurally, membrane lipids constitute a semipermeable barrier that separates the extracellular milieu from the cytosol and the cytosol from the lumen of intracellular organelles. The activities of many membrane-bound enzymes, receptors, and ion channels are tightly regulated by the transbilayer distribution of specific lipid species. For example, kinases require the presence of anionic phospholipids at the plasma membrane (PM) inner leaflet for activation of downstream signal transduction pathways. Sphingolipids and cholesterol, on the other hand, participate in the assembly of lipid rafts at the outer leaflet. Many studies have established the unifying concept that phospholipids are asymmetrically distributed across the PM bilayer (see Chapter 3). This distribution can be reversibly altered in response to a variety of external and internal stimuli that direct specific biological functions. For example, the movement

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

291

292

MECHANISMS OF PS EXTERNALIZATION IN APOPTOSIS AND AGING

of phosphatidylserine (PS) to the cell surface is required for activation of the coagulation cascade and the recognition of dying cells by phagocytes (see Chapter 7). In some cases, the redistributions are transient, with lipid asymmetry being restored through the actions of specific lipid transporters. Phosphatidylethanolamine (PE), for example, is transiently externalized during cell division at the cleavage furrow [1], and PS is expressed in granulocytes [2], B lymphocytes [3, 4], mast cells [5], and macrophages [6, 7]. One of the identifying characteristics of dying cells is the presence of PS at the cell’s outer leaflet. This results in their recognition and clearance by phagocytes through a noninflammatory pathway [8–10]. The importance of PS to this process is underscored by observations that apoptotic cells can mount an inflammatory and autoimmune response should their externalization or recognition mechanisms turn defective [11, 12]. While many lipids may undergo lateral or transbilayer redistributions during apoptosis, the spatiotemporal distribution of PS is the most dramatic alteration and most studied. It is important to note that while the concept of membrane lipid asymmetry was established almost half a century ago, there is still no consensus on the identity of the proteins that mediate and regulate transbilayer lipid movements in apoptosis and in various pathologies. Nevertheless, it is now recognized that the ratio of Ca2+ to K+ in the cytosol is a primary determinant of membrane lipid sidedness [13, 14]. This chapter will primarily focus on mechanisms that likely regulate loss of PS asymmetry and its appearance at the cell surface during aging and apoptosis. While we will discuss putative classical mechanisms, we will also introduce heretofore unappreciated mechanisms that might also play a role in these processes. 14.2â•… PHOSPHOLIPID TRANSPORTERS Early work with red blood cell (RBC) has provided a large body of evidence indicating that mammalian cells maintain normal lipid asymmetry through the concerted action of three activities. The ATP-dependent aminophospholipid translocase (flippase) transports the aminophospholipids PS and PE from the cell’s outer-to-inner membrane leaflet (see Chapters 8–10). Outward movement, on the other hand, is mediated by two distinct activities: a slow, regulated ATP-dependent “floppase” activity (see Chapters 11 and 17) and a catastrophic Ca2+-dependent “scramblase” (see Chapter 7) that equilibrates lipids across both bilayer leaflets. 14.2.1â•… Inward Movement: The Aminophospholipid Translocase Using spin-labeled phospholipid analogs, Seigneuret and Devaux observed that, in contrast to phosphatidylcholine (PC), the PS analog was protected from reduction by ascorbate after a short incubation with RBC [15, 16]. Because this activity was inhibited by fluoride and vanadate, they postulated

Phospholipid Transporters

293

that the lipid was protected as a result of ATPase-dependent sequestration to the cell’s inner membrane leaflet. These results were later confirmed using both short-chain [17] and fluorescent [18] phospholipid analogs. Various ATPases, including the 80- to 120-kDa Mg2+-ATPases isolated from red cell membranes [19–22] and the bovine chromaffin granule ATPase [23], have been implicated as transporters that regulate the movement of PS from the cell’s inner-to-outer membrane leaflet. Cloning of the bovine chromaffin granule ATPase indicated that it belonged to a subfamily of P-type ATPases [24]. In yeast, disruption of the DRS2 gene that encodes the mammalian P-type ATPase homolog inhibited transport of exogenously supplied fluorescent PS, a result that supports its function as an aminophospholipid translocase [24]. Indeed, a plant gene involved in cold tolerance that shares homology with DRS2 reconstituted PS transport activity in transport-deficient DRS2 mutants [25]. Similarly, knockdown of the Caenorhabditis elegans P-type ATPase homolog, TAT-1, facilitated constitutive reorientation of PS from the inner-toouter membrane leaflet in germ cells [26]. Despite these promising leads, similar deletion experiments carried out in other laboratories failed to corroborate a role for both DRS2 [27, 28] and TAT-1 [29] in PS transport. While it is difficult to reconcile these differences, it is clear that inward lipid movement is dependent on an ATPase activity. It is important to emphasize that none of these candidates have been shown to directly bind phospholipids. This leaves open the possibility that, together with other proteins, these ATPase(s) play a regulatory role in PS externalization as part of a multicomponent aminophospholipid transport complex. 14.2.2â•… Outward Movement: Floppase Experiments using fluorescent and spin-labeled phospholipid analogs provided evidence for the existence of an ATP- and protein-dependent floppase that transports lipid from the inner-to-outer membrane leaflet irrespective of the lipid’s head-group structure [30–34]. Proteins with potential floppase activity include members of the ATP-binding cassette (ABC) transporter superfamily including P-glycoprotein (multidrug resistance [MDR]), multidrug resistance-associated protein (MRP), and mitoxantrone resistance protein (MXR) [28, 35]. More recent data, however, has revealed that the export of the lipid analogs in these experiments was most likely due to glutathione disulfide-dependent exclusion of the reporter tags to the cell surface by MRP1 [36, 37], a function reminiscent of many drug resistance proteins. This leaves open the question whether MRP play any physiological role in PS movement. In contrast, using cells isolated from ABC1 null mice, Hamon et al. [38], provided the first unequivocal evidence that this protein plays a primary role in the movement of endogenous PS from the inner-to-outer membrane leaflet. These data are further supported by observations that a missense mutation in ABC1 correlated with the scramblase-deficient phenotype of cells from Scott syndrome patients [39]. These data, however, are not consistent with a

294

MECHANISMS OF PS EXTERNALIZATION IN APOPTOSIS AND AGING

follow-up study that showed that the inward and outward movement of PS was normal in cells obtained from Tangier disease patients who express severely truncated forms of ABC1 [40]. 14.2.3â•… Bidirectional Movement: Scramblase Studies in platelets and RBC revealed the existence of a Ca2+-dependent scramblase that facilitates complete intermixing of all lipids between bilayer leaflets [41]. PLSCR1 was initially identified as the scramblase; however, later studies in null animals failed to confirm its role in bidirectional lipid movement [42]. Notwithstanding these observations, it was proposed that C. elegans proteins such as the human apoptosis-inducing factor (AIF) homolog WAH-1 [43], the human PLSCR1 homolog SCRM-1 [43], and more recently, the yeast DRS2 homolog TAT-1 [26, 29], function as regulators of outward PS movement. Knockout of these genes prevented the externalization of PS in response to apoptotic stimuli [29], though other studies have implicated the TAT1 gene product as a regulator of inward (translocase) lipid movement [26]. Clearly, more detailed analysis of the role of TAT1 in lipid transport will require unequivocal demonstration of lipid binding.

14.3â•… LIPID ASYMMETRY IN ERYTHROCYTES Mammalian erythrocytes (RBCs) are formed by terminal differentiation of reticulocytes. This process is characterized by enucleation and loss of intracellular organelles through regulated mechanisms that leave behind an essentially Ca2+-free (<100â•¯nM) cytosol encapsulated within a remnant PM bilayer. As a result, mammalian erythrocytes lack the machinery that drives intrinsic and extrinsic apoptotic pathways. Nonetheless, red cells lose their asymmetric lipid distribution upon aging and in various pathologies indicating that specific PS transport mechanisms are retained in mature RBC. Indeed, transbilayer PS distribution in RBC can be reversibly regulated by ionophore-mediated changes to cytosolic Ca2+ [41]. 14.3.1â•… Regulation of PS Externalization in RBCs Due to the lack of intracellular organelles and apoptosis regulatory mechanisms, RBCs do not undergo classical apoptosis. They do, however, undergo a process of aging and are cleared from the peripheral circulation by the liver and spleen. While there have been reports that red cells harbor several apoptosis regulators such as Bcl-XL, Bak, and caspases [44–46], there is no evidence to indicate that these proteins actively participate in red cell aging. It is more likely that these proteins are simply left over remnants from precursor reticulocytes. One of the most studied features of RBC aging is the appearance of

Red Cell Aging

295

PS at the cell surface. While the critical players in this process remain unknown, it is believed to result from a culmination of several events including, ATP depletion [47] and oxidative damage [48, 49] that lead to increased osmotic fragility [50] and a progressive increase in cell density [51]. All these changes can be induced experimentally by enforced elevation of cytosolic Ca2+ that results in loss of intracellular K+ through the Gardos channel. It is important to note that while loss of K+ alone results in increased cell density (similar to aged cells) as a result of dehydration, these cells do not express PS [52]. Similarly, ionophore-enforced elevation of Ca2+ in the presence of Gardos channel inhibitors failed to yield PS-positive cells [52, 53]. Furthermore, the inhibitory effects of Gardos channel antagonists on Ca2+-dependent PS externalization can be mitigated by co-incubation with the K+ ionophore, valinomycin [54]. Taken together, these observations indicate that PS externalization in aged RBC requires both elevation of cytosolic Ca2+ and depletion of cytosolic K+ and suggests that altered [Ca2+]i-to-[K+]i ratios, rather than unilateral changes to either ion, are important determinants for PS externalization during RBC aging. While it is assumed that Ca2+ directly binds to and activates a phospholipid “scramblase,” recent studies have raised the possibility that PS externalization might require the activation of Ca2+-dependent signaling pathways. This is underscored by observations that treatment of RBC with the protein kinase C (PKC) activator, phorbol myristate acetate (PMA), results in both Ca2+ influx and PS externalization [55]. This suggests that phosphorylation of yet unidentified downstream targets are critical for Ca2+-dependent transbilayer PS movement. Thus, PKC could either phosphorylate and directly activate a putative “scramblase,” or activate intermediate regulatory proteins that control scramblase activity [56]. While such a mechanism presumes protein-mediated lipid transport, it is also possible that lipid redistribution could be achieved in a protein-independent manner through the generation of non-bilayer structures. Indeed, several model membrane studies have provided evidence indicating that the in situ generation of lysophospholipids induce localized micellar structures [57–59] with concomitant intermixing of lipids across bilayer leaflets [60, 61]. While there is no evidence for a role of lysolipids in dynamic alterations of lipid asymmetry in cells, it is of interest to note that the activity of phospholipases are regulated by both phosphorylation and Ca2+ [62–64]. In principle therefore, activation of kinases could promote transbilayer movement through the activation of phospholipases and the production of lysophospholipids (Fig. 14.1). Cumulative oxidative damage [65, 66] can also contribute to PS exposure by directly inhibiting lipid transporters [49, 67] or by oxidation of the PS acyl chains [68]. 14.3.2â•… Red Cell Aging Several studies have indicated that aged red cells undergo membrane-related alterations that contribute to the senescent phenotype. These include the

PLG

C

PK PKC Intracellular Calcium IP3R Stores

PLA2

Nucleus PLA 2

KC

P

Phosphatidylcholine (PC)

Diacylglycerol (DAG)

IP3 receptor (IP3R)

Phosphatidylethanolamine (PE)

Calciumions

Calcium ATPases

Phosphatidylinositol (PI)

Potassiumions

Gardos channel

Phosphatidylserine (PS)

Inositoltriphosphate (IP3)

Calcium channels

Lysophosphatidylcholine (LPC)

Phosphorylation

Phospholipidtransporters

Figure 14.1.â•‡ Potential Ca2+-dependent mechanisms that regulate the outward movement of PS in cells. This figure shows several critical events that participate in the regulation of PS externalization. (1) In normal cells, cytosolic Ca2+ levels are maintained in the nanomolar range through the concerted activities of PM Ca2+ channels and PM Ca2+-ATPases. (2) In principle, increases in resting cytosolic Ca2+ levels can be achieved either by a reduction in the activity of the Ca2+-ATPases or by activation of inward Ca2+ channels. (3) This can lead to activation of phospholipase C, which generates diacylglycerol and IP3. (4) IP3 promotes the release of intracellular Ca2+ through IP3 receptors, thereby triggering massive Ca2+ release from intracellular stores. (5) This could then lead to the activation of protein kinases that (a) directly regulate lipid transporters by their phosphorylation or (b) activate phospholipase A2 (PLA2), leading to lysolipid formation that induces localized non-bilayer structures that promote lipid intermixing and hence PS externalization through a protein-independent mechanism. Color version on the Wiley web site.

Lipid Asymmetry during Apoptosis

297

appearance of the eat-me signal, PS [7], and/or the loss of the don’t eat-me signals, CD47 [69] and sialic acid residues [70]. Although not mutually exclusive, several studies have shown that PS serves as a critical determinant for the recognition and clearance of red cells in vivo. PS-expressing RBCs and PS vesicles are cleared from the peripheral circulation at rates that are significantly faster than normal red cells and vesicles composed exclusively of PC [51, 71, 72]. In vivo studies on aged red cells are not possible since the PS-expressing population is efficiently removed from the peripheral blood. Thus, most studies on the relationship between lipid asymmetry and RBC aging rely on in vitro surrogates. These studies have shown that a direct relationship exists between red cell age [72], density [73], and expression of PS at the cell surface [51]. Typically, density-separated cells exhibit decreased cell volume [74, 75], cell size [76], and deformability [77, 78], as well as increased osmotic fragility [50] due to loss of electrolytes and microvesiculation [79]. Considering that PS externalization is regulated by the ratio of intracellular [Ca2+] to [K+], the increased density of aged cells could be related to loss of water as a result of Ca2+ influx-induced K+ efflux [80]. In principle, increased Ca2+ influx in these cells could occur through loss of membrane integrity and/or reduced Ca2+-ATPase activity as a result of decreased levels of intracellular ATP [73, 81]. 14.3.3â•… Pathological Red Cells A heterogeneous group of red cell pathologies, including red cells from sicklecell disease [82–85], β-thalassemia [86, 87], and diabetic [88, 89] patients, have been shown to be associated with increased elevations in cell surface PS. Although each of these diseases is distinct, they are characterized by anemia that can be caused by cell lysis [72] following assembly of complement proteins at the cell surface [90, 91] or by accelerated cell clearance through PSdependent phagocytosis [51, 92]. Thus, disturbances in membrane phospholipid asymmetry might, at least partially, account for many of the associated pathologies in these diseases. Although never studied together, many of these pathologies are also characterized by increased levels of intracellular Ca2+ [93, 94]. These consistent and uniform observations among diverse pathologies raise the possibility of a physiologically relevant relationship between [Ca2+]i and loss of PS asymmetry. Taken together with the data available on aged red cells, these observations help establish a unifying model linking Ca2+, K+, and PKC activation to PS externalization in vivo. 14.4â•… LIPID ASYMMETRY DURING APOPTOSIS Apoptosis is a physiological process that is critical for tissue remodeling, and the development and resolution of inflammation. Depending on the nature of

298

MECHANISMS OF PS EXTERNALIZATION IN APOPTOSIS AND AGING

the stimulus, execution of apoptosis occurs along either the extrinsic death receptor pathway or the intrinsic mitochondrial pathway. The extrinsic mechanism is triggered by engagement of receptors that belong to the tumor necrosis factor (TNF) receptor family by their corresponding ligands. This results in activation of caspase 8 that leads to direct cleavage and activation of caspase 3 (type I cells) and/or the cleavage of Bid (type I and II cells). The later event promotes the disengagement of Bcl-2 and consequent insertion of Bax and Bak into the mitochondrial membrane that leads to the release of cytochrome c (cyt c) into the cytosol. The intrinsic pathway is activated by several agents including, but not limited to, nutrient withdrawal, radiation, and cytotoxic drugs. These initiate a series of events that lead to permeabilization of the mitochondrial outer membrane and the release of several mitochondrial proteins, including cyt c and AIF into the cytoplasm. Once in the cytosol, cyt c promotes the assembly and activation of the apoptosome that leads to procaspase 9 cleavage and subsequent activation of effector caspases 3 and 7. Thus, while both extrinsic and intrinsic routes have distinct points of origin, they ultimately converge at the mitochondria. To distinguish apoptotic from normal, cells must acquire specific surface characteristics that facilitate their recognition as a phenotype destined for clearance from the host. While several cell surface markers of apoptosis have been identified, there is a large body of evidence that indicates that PS externalization is critical for the cells recognition by phagocytes. It is now appreciated that engulfment of dying cells through PS-mediated recognition mechanisms occurs through a silent noninflammatory pathway [8]. In contrast, necrotic cells that fail to express PS are cleared through a pathological inflammatory pathway that can trigger autoimmune responses [95, 96]. Despite tremendous progress in deciphering various apoptotic pathways, the mechanism responsible for the appearance of PS at the cell surface remains elusive. Although a number of theories relating to its mechanism(s) of exposure have emerged, the critical event that is common to all is an obligate requirement for sustained increases in cytosolic Ca2+ concentrations. 14.5â•… CA2+ HOMEOSTASIS DURING APOPTOSIS The levels of intracellular Ca2+ play critical regulatory roles in many cellular processes. While crucial to cell survival, the sustained presence of high levels of Ca2+ in the cytosol is toxic. Typically, the concentration of free Ca2+ in the cytosol of resting cells varies between 20 and 100â•¯nM. This is maintained by the buffering action of Ca2+-binding proteins combined with its sequestration into the endoplasmic reticulum (ER) [97–101], mitochondria [102–105], Golgi [106, 107], and lysosomes [108]. Of these stores, the ER lumen accounts for more than 95% of intracellular Ca2+. The concentration of total ER Ca2+ (freeâ•¯+â•¯bound) ranges between 1 and 3â•¯mM [109, 110] with the free [Ca2+]ER ranging from 60 to 400â•¯µM when stores are full, and from 1 to 50â•¯µM when

Membrane Phospholipid Asymmetry: Static or Dynamic?

299

stores are empty [111]. ER Ca2+ is regulated by the buffering actions of Ca2+binding chaperones [109, 111], in concert with the ER Ca2+-ATPase (SERCA) that regulates influx into the lumen and the inositol-(1,4,5)-triphosphate (IP3) and ryanodine (Ry) receptors that mediate efflux into the cytosol. Due to the relatively high concentrations of Ca2+ in the ER, cytosolic Ca2+ overload can be achieved by altered activities of these channels/transporters without the contribution of extracellular Ca2+ [103]. These changes, however, are generally short lived because of influx into the mitochondria and/or efflux to the extracellular space by the activities of the mitochondrial uniporter and PM Ca2+ATPases, respectively [102, 104, 105, 112]. Activation of apoptosis, however, results in the progressive release of Ca2+ from these stores, which if left uncorrected, results in sustained increases in cytosolic Ca2+ to toxic levels that trigger apoptosis. Although the role of ER and mitochondrial Ca2+ stores in apoptosis has been well established, the potential contribution of endosomal and lysosomal Ca2+ to apoptosis has not been appreciated. Since pinocytotic vesicles sequester the extracellular milieu, the intravesicular Ca2+ concentrations and ultimately lysosomal Ca2+ levels should reflect that of the medium and be in the millimolar range. Indeed, recent observations suggest that the Ca2+ concentration of the lysosomes is in the order of 0.5â•¯mM [113]. Moreover, since endosomes and lysosomes have Ca2+ transporters [114, 115], it is possible that these organelles also sequester cytosolic Ca2+ in their lumen. This raises the possibility that, similar to other organelles, lysosomes also regulate intracellular Ca2+ homeostasis and contribute to apoptosis-triggered increases in cytosolic Ca2+ levels. 14.6â•… MEMBRANE PHOSPHOLIPID ASYMMETRY: STATIC OR DYNAMIC? Normal cell physiology is characterized by incessant inward and outward vesicle trafficking pathways that result in continuous perturbations in the distribution of lipids across the PM bilayer. If left uncorrected, the exposure of PS at the cell surface could lead to catastrophic thromboembolic events and initiate unwarranted phagocytic responses. Since this does not normally occur, cells are likely protected from these events by the activity of the aminophospholipid translocase that reestablishes normal lipid asymmetry. This implies that membrane lipid asymmetry is maintained through a correction-ondemand process that ensures that PS only remains exposed in the cell’s outer leaflet when physiologically required. In contrast to trafficking-dependent transbilayer lipid movements, cells can also manifest continuous bidirectional movement. In this case, steady-state membrane lipid asymmetry will be determined by the relative activities of the transporters that regulate the outward and inward movements. Indeed, recent data by Darland-Ransom et al. [26] indicated that deletion of TAT1 in

300

MECHANISMS OF PS EXTERNALIZATION IN APOPTOSIS AND AGING

C. elegans resulted in the constitutive presence of PS at the cell’s outer leaflet. This suggests that outward PS movement is a continuous process that is corrected through TAT1 activity. However, the authors did not consider the possibility of altered basal cytosolic Ca2+ levels between wild-type and TAT1deficient cells. This is of particular importance since any sustained increase in cytosolic Ca2+ (as a result of the deletion) would also trigger the externalization of PS [116]. Unlike “dynamic” maintenance of lipid asymmetry, it is also possible that lipid asymmetry is inherently static unless perturbed by an appropriate stimulus. This possibility is supported by observations from our laboratory, which showed that, in the absence of cytosolic Ca2+ (that initiates scramblase activity), inhibition of inward PS movement with cell impermeable thiol reagents failed to trigger PS externalization (unpublished observations). 14.7â•… REGULATION OF LIPID ASYMMETRY DURING APOPTOSIS Although many hypotheses concerning the regulation of membrane phospholipid asymmetry in general, and PS externalization in particular, have been put forth, there is no unified consensus on the identity of the transporters involved in these processes. The only certainty in the field is that PS externalization requires sustained increases in the concentration of intracellular Ca2+. Mechanistically, Ca2+ is assumed to play opposing roles by promoting outward regulating “scramblase” and/or “floppase” activity and, at the same time, inhibiting inward rectifying translocase activity. While direct Ca2+ interaction with these transporters is possible, there is no evidence in support of such a conclusion. Hence, one must also consider the possibility that sustained Ca2+ levels activate downstream regulatory pathways that ultimately control bidirectional lipid movement. Indeed, inhibitors of kinases [55] and phospholipases effectively block the externalization of PS in red cell (K. Balasubramanian, unpublished observations) and apoptotic cell systems [117]. Studies in many model membrane systems have shown that Ca2+ promotes anionic phospholipid-dependent membrane fusion events. Because of the luminal localization of PS in eukaryotic cells, it is conceivable that the increase in cytosolic Ca2+ levels achieved during apoptosis also promote intra-organelle/ membrane fusion events. Indeed, there are several lines of evidence indicating that organelle proteins and lipids appear at the cell surface during apoptosis. Although never considered before, fusion during vesicle trafficking could also abolish lipid asymmetry by promoting intermixing of lipids between bilayer leaflets as a result of transient bilayer to non-bilayer lipid polymorphisms. 14.7.1â•… Role of K+ in Regulating Outward Movement of PS Several studies have indicated that the aged red cell phenotype is characterized by a progressive increase in cell density that is associated with concomi-

Regulation of Lipid Asymmetry during Apoptosis

301

tant increase in cell surface PS. The increased density of aged RBC is a result of K+ efflux and consequent loss of water. It has recently been shown that the Gardos channel, a Ca2+-activated K+ efflux pump, regulates PS externalization in platelets [54] and red cells [13, 14]. Studies on the mechanism of Gardos channel control of PS transport revealed an inverse relationship between [K+]i and [Ca2+]i, suggesting an inhibitory role for K+ on scramblase/floppase activity. Moreover, the inverse correlation between [K+] and outward PS transport in erythrocytes is in agreement with recent data demonstrating inhibition of agonist (collagen and thrombin)-induced PS exposure at the platelet surface by K+ channel blockers and their subsequent reversal with valinomycin [54]. While it is unclear whether loss of intracellular K+ is critical for apoptosisregulated activation of scramblase in nucleated cells, there are reports indicating an inhibitory role for intracellular K+ on caspase, nuclease, and apoptosome activity [118–120]. This raises the possibility that, similar to red cells, intracellular K+ is a negative regulator of several key events in apoptosis including PS externalization.

14.7.2â•… Role of Kinases and Phospholipases in Outward PS Movement Studies in RBC model systems revealed that PKC activators such as PMA promoted PS externalization, while PKC inhibitors such as calphostin and chelerythrine chloride blocked its appearance at the cell surface [55]. Since normal RBC actively maintain very low intracellular Ca2+ levels through the activity of Ca2+-ATPases, it is reasonable to assume that the initial trigger for PKC activity is because of passive low-level influx of extracellular Ca2+. Since PKC activation also requires diacylglycerol (DAG), increased activity of phospholipase C (PLC) activity is likely upstream to PKC activation. Thus, a modest increase in intracellular Ca2+ as a result of reduced Ca2+-ATPase activity in aging cells could result in PLC-mediated generation of DAG. This leads to further activation of PKC that results in an additional robust increase in intracellular Ca2+ [55, 121, 122], K+ efflux, phosphorylation of a critical transporter, and subsequent PS externalization (Fig. 14.1). In contrast to RBC that do not contain intracellular Ca2+ stores, the Ca2+demand in nucleated cells can be supplied from intracellular stores including the ER and Golgi. In this case, initiation of apoptosis would lead to Ca2+dependent activation of phosphatidylinositol-specific PLC that generates both DAG and IP3. Although IP3 promotes Ca2+ release through IP3 receptors, most of the Ca2+ is immediately sequestered into the mitochondria [123]. As apoptosis progresses, however, loss of mitochondrial membrane potential results in reduced Ca2+-buffering capacity and subsequent flooding of the cytoplasm with Ca2+ (Fig. 14.1). Data obtained from several experimental systems with inhibitors of PLA2 have indicated that cPLA2α activity is likely required for PS externalization in both RBC (K. Balasubramanian, unpublished results) and nucleated cells

302

MECHANISMS OF PS EXTERNALIZATION IN APOPTOSIS AND AGING

[117]. Although the mechanism underlying this inhibition is unclear, it is likely related to the generation of the second messengers, lysophospholipids and arachidonate. 14.7.3â•… Role of Lipid Peroxidation in Outward PS Movement Activation of the intrinsic mitochondrial phase of apoptosis leads to opening of the permeability transition pore and the release of several proteins including cyt c into the cytoplasm. Here, cyt c binds to and promotes assembly of the apoptosome that catalyzes caspase 9 cleavage and activation, leading to activation of effector caspases 3 and 7 that participate in the final execution phase of apoptosis. In addition to its role as a cofactor in caspase activation, cyt c can also oxidize lipids. Cyt c plays an important role in the transfer of electrons during oxidative phosphorylation [124, 125]. Dysregulation of electron transport during apoptosis leads to accumulation of free radicals, which when left uncorrected, can lead to cyt c-dependent oxidation and subsequent cell death by apoptosis [126, 127]. Indeed, free radical scavengers and overexpression of antioxidant proteins inhibit H2O2-mediated peroxidation of cyt c [128] and abrogate drug-induced apoptosis [129]. Moreover, several reports have suggested a link among free radical generation, Ca+2 homeostasis [130– 132], and oxidation of cyt c. There is also evidence indicating that PS is a major target for cyt c-dependent peroxidation during apoptosis [49, 68, 133, 134]. This is supported by observations that cyt c binds negatively charged lipids like CL and PS [135–137] and that this binding leads to disruption of the Met80Fe(heme) coordination bond and partial unfolding of the protein that facilitates reorientation of the heme moiety along the membrane surface. This conformation renders Fe more catalytically redox active and positions the protein’s heme catalytic site closer to phospholipid acyl chains. It is therefore conceivable that once released from mitochondria, oxidized cyt c binds PS on the inner leaflet of the PM, unfolds to expose its redox-active heme moiety and oxidizes the lipid’s acyl chains. This, in turn, results in localized disturbance in lipid packing that energetically favors diffusion of all lipids between bilayer leaflets [138]. 14.7.4â•… Role of Membrane Fusion in Outward PS Movement Morphologically, apoptotic cells are differentiated from their nonapoptotic counterparts by their distinctive PM blebbing pattern. Membrane blebbing results in the progressive reduction in PM surface area and cytoplasmic volume that result in shrinkage of the remnant cell into “apoptotic bodies.” This prepares the dying cell for engulfment and catabolism by phagocytes. While the mechanism that leads to PM blebbing and shedding during apoptosis is unclear, it likely requires cytoskeletal reorganization in concert with dramatic and unrestrained increases in PM surface area. The requirement for an increase in PM surface area for blebbing can be seen from the hypothetical relationship

303

Regulation of Lipid Asymmetry during Apoptosis

Fold increase in surface area required

3

2

1 1

5

9

12

16

20

Number of particles formed

Figure 14.2.â•‡ Hypothetical model for the increase in PM surface area required to accommodate the generation of n particles. The data were generated by calculating the surface area of a perfect sphere of diameter. The volume of the sphere was then divided by the number of assumed blebbed particles (abscissa). Again, assuming perfect spheres, the surface area required to accommodate the blebbed sphere volume was calculated and multiplied by the number of particles divided by the surface area of the parent sphere (ordinate).

between the number of blebs excised from the parent apoptotic cell and the fractional increase in PM surface area required for their generation (Fig. 14.2). In principle, the requirement for increased surface area can be achieved by movement of internal membranes toward the subplasmalemmal region of the PM followed by their fusion. Indeed, recent studies have revealed that proteins and lipids from intracellular organelles redistribute to the cell’s outer membrane leaflet as a consequence of apoptosis. These include calreticulin from the ER [139], cardiolipin [140, 141] from the mitochondria, and several nuclear antigens [142]. Since membrane fusion is characterized by the formation of transient bilayer to non-bilayer structures that facilitates free intermixing of lipids between bilayer leaflets [58, 143], fusion could, in addition to bleb formation, influence loss of PS sidedness during apoptosis. In addition to the contribution of ER, mitochondria, and nuclear membranes, recent data from our laboratory indicate that lysosomes also fuse to the PM during apoptosis [144]. This is particularly interesting, since lysosome to PM fusion is a major player in membrane repair responses [145–148]. Data from several laboratories have established that lysosomal damage repair

304

MECHANISMS OF PS EXTERNALIZATION IN APOPTOSIS AND AGING

responses are primarily initiated by the elevation of cytosolic Ca2+. While extracellular Ca2+ entry through the damage site is the presumed mechanism for increased cytosolic Ca2+, it is not known whether this mode of entry can be complemented by efflux from intracellular stores. Since endosomes and lysosomes also have Ca2+ transporters [114, 115], it is possible that these organelles also function as Ca2+ depots. Irrespective of the source of Ca2+, it is clear that elevation in cytosolic Ca2+ leads to Ca2+-dependent docking of many intracellular organelles including lysosomes at the subplasmalemmal surface of the PM and subsequent membrane fusion [149–153]. Since fusion can trigger intramembrane lipid mixing through the transient formation of nonbilayer structures, it is possible that lipid asymmetry is lost during the formation of the endosomes before they fuse with lysosomes. If this is indeed the case, then endosome-derived lysosomes should express PS at their luminal leaflet. Thus, subsequent fusion of the lysosomes to the PM during apoptosis or during membrane damage repair would result in the redistribution of PS to the cell’s outer membrane leaflet (Fig. 14.3). The critical difference between these two processes is that in membrane repair, basal cytosolic Ca2+ levels are rapidly restored by the actions of the Ca2+ pumps [108, 154]. This is followed by reactivation of the Ca2+-inhibited aminophospholipid translocase resulting in restoration of PS asymmetry. During apoptosis, on the other hand, persistent high cytosolic Ca2+ precludes reactivation of the translocase resulting in the sustained and uncorrectable presence of PS at the cell surface. 14.8â•… SIGNIFICANCE Apoptosis is a programmed cell death mechanism that prepares the dying cell for elimination without pathological immune or inflammatory consequences [8]. Necrosis, on the other hand, is characterized by loss of PM integrity and release of intracellular components that activate the immune system [95], resulting in inflammatory responses and, in certain instances, autoimmunity [96]. Thus, sequestration of the cytosol within the PM barrier is crucial to prevention of pathological responses [155]. This is achieved during apoptosis either by recognition and elimination of early apoptotic cells that express PS at their surface [156–158], or by phagocytosis of PS-expressing PM blebs that are released by late apoptotic cells. The importance of PM integrity is underscored by observations that similar to necrotic cells, apoptotic cells can mount an inflammatory and autoimmune response should their blebbing mechanisms turn defective or phagocytosis be delayed [159, 160]. The observations that synaptotagmin VII-deficient mice display both defective membrane repair and an autoimmune pathology reinforces the relevance of membrane blebbing to the prevention of autoimmune responses [161]. It is interesting to note that, unlike (anti-inflammatory) apoptotic cells, necrotic cells are distinguished by their absence of blebbing and increased cell volume [162], events that can lead

305

Ea

om

e

Restoration of lipid asymmetry

Lysosome–PM fusion

(b)

Early endosome Late endosome

Phosphatidylethanolamine (PE) Phosphatidylinositol (PI) Phosphatidylserine (PS) Ganglioside

Aminophospholipid translocase Calciumions

Phosphatidylcholine (PC)

Exocytosis

Endocytosis

Invagination

Luminal membrane protein

Lysosome

Lysosome solution phase

Hybrid organelle

Figure 14.3.â•‡ Regulation of lipid asymmetry by plasma membrane repair: Membrane damage results in loss of PM integrity that causes the unregulated influx of Ca2+ into the cytosol (a). This results in the inactivation of the aminophospholipid translocase and initiates a lysosomedependent repair response [108, 148] that seals the damaged PM through lysosome–PM fusion. Fusion results in the formation of non-bilayer structures that result in transbilayer lipid scrambling and perturbation in the asymmetric distribution of PS. In nonapoptotic cells, this is immediately followed by restoration of intracellular Ca2+ to basal levels and restoration of membrane lipid asymmetry by reactivation of the Ca2+inhibited aminophospholipid translocase. In apoptotic cells, on the other hand, elevation in intracellular Ca2+ activates a relentless pseudomembrane “repair response” due to the sustained presence of Ca2+ in the cytosol. This leads to the irreparable presence of PS at the cell surface. Color version on the Wiley web site.

Hybrid organelle

e

Lys oso m

Lysosomal exocytosis

Inhibition of aminophospholipid translocase

Damage site

Late endosome

rly

(a)

os

en d

306

MECHANISMS OF PS EXTERNALIZATION IN APOPTOSIS AND AGING

to PM rupture, and release of cellular contents. Thus, understanding the key players in lysosomal exocytosis during apoptosis and necrosis could have important implications for the treatment of autoimmune and inflammatory diseases. 14.9â•… CONCLUDING REMARKS Physiological cell death (apoptosis) is regulated by a series of highly synchronized processes that prepare the cell for engulfment and disposal by phagocytes. Morphologically, apoptotic cells are distinguished from their normal nonapoptotic counterparts by characteristic (PM) blebbing and the specific redistribution of the PM lipid, PS, from the cell’s inner-to-outer PM leaflet. These “hallmarks” of apoptosis are regulated by temporal and spatial intracellular Ca2+ transients that result in increases in cytosolic [Ca2+]. While the regulatory mechanisms that control physiological cell repair are beginning to emerge [146–148, 163, 164], there is a significant gap in our understanding of how Ca2+ turns from a signal critical to many physiological processes to an apoptotic death response signal. An intriguing possibility is that apoptotic cells mount a futile pseudorepair response mechanism that, instead of protecting against cell death, promotes an irreparable suicidal death response that results in PM expansion and the formation of PS-expressing blebs that are excised from the remnant cell. Understanding the molecular events that regulate this process might enable us to alter the “silent,” nonimmune physiological fate of these cells, to a response that will stimulate innate immunity. This could be important for the development of novel therapeutic regimens that could trigger specific antitumor responses following chemotherapy.

ABBREVIATIONS cyt c DAG ER IP3 PC PE PKC PLC PM PS RBC TNF

cytochrome c diacylglycerol endoplasmic reticulum inositol-(1,4,5)-triphosphate phosphatidylcholine phosphatidylethanolamine protein kinase C phospholipase c plasma membrane phosphatidylserine red blood cell tumor necrosis factor

References

307

REFERENCES â•… 1â•… K. Emoto, T. Kobayashi, A. Yamaji, H. Aizawa, I. Yahara, K. Inoue, M. Umeda, Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 12867–12872. â•… 2â•… M. Y. Yang, H. Chuang, R. F. Chen, K. D. Yang, J. Leukoc. Biol. 2002, 71, 231–237. â•… 3â•… S. R. Dillon, M. Mancini, A. Rosen, M. S. Schlissel, J. Immunol. 2000, 164, 1322–1332. â•… 4â•… S. R. Dillon, A. Constantinescu, M. S. Schlissel, J. Immunol. 2001, 166, 58–71. â•… 5â•… S. Martin, I. Pombo, P. Poncet, B. David, M. Arock, U. Blank, Int. Arch. Allergy Immunol. 2000, 123, 249–258. â•… 6â•… M. K. Callahan, P. Williamson, R. A. Schlegel, Cell Death Differ. 2000, 7, 645–653. â•… 7â•… V. A. Fadok, A. de Cathelineau, D. L. Daleke, P. M. Henson, D. L. Bratton, J. Biol. Chem. 2001, 276, 1071–1077. â•… 8â•… P. M. Henson, D. L. Bratton, V. A. Fadok, Curr. Biol. 2001, 11, R795–R805. â•… 9â•… J. Savill, J. Leukoc. Biol. 1997, 61, 375–380. â•‡ 10â•… M. Gray, K. Miles, D. Salter, D. Gray, J. Savill, Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 14080–14085. â•‡ 11â•… L. Casciola-Rosen, A. Rosen, M. Petri, M. Schlissel, Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 1624–1629. â•‡ 12â•… A. Rosen, R. Casciola, Cell Death Differ. 1999, 6, 6–12. â•‡ 13â•… F. Lang, K. S. Lang, P. A. Lang, S. M. Huber, T. Wieder, Antioxid. Redox Signal. 2006, 8, 1183–1192. â•‡ 14â•… F. Lang, S. M. Huber, I. Szabo, E. Gulbins, Arch. Biochem. Biophys. 2007, 462, 189–194. â•‡ 15â•… M. Seigneuret, A. Zachowski, A. Hermann, P. F. Devaux, Biochemistry 1984, 23, 4271–4275. â•‡ 16â•… M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. â•‡ 17â•… D. L. Daleke, W. H. Huestis, Biochemistry 1985, 24, 5406–5416. â•‡ 18â•… J. Connor, A. J. Schroit, Biochemistry 1987, 26, 5099–5105. â•‡ 19â•… G. Morrot, A. Zachowski, P. F. Devaux, FEBS Lett. 1990, 266, 29–32. â•‡ 20â•… M. E. Auland, B. D. Roufogalis, P. F. Devaux, A. Zachowski, Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 10938–10942. â•‡ 21â•… M. E. Auland, M. B. Morris, B. D. Roufogalis, Arch. Biochem. Biophys. 1994, 312, 272–277. â•‡ 22â•… D. L. Daleke, K. Cornely-Moss, J. Lyles, C. M. Smith, M. Zimmerman, Ann. N.Y. Acad. Sci. 1992, 671, 468–470. â•‡ 23â•… A. Zachowski, J. P. Henry, P. F. Devaux, Nature 1989, 340, 75–76. â•‡ 24â•… X. Tang, M. S. Halleck, R. A. Schlegel, P. Williamson, Science 1996, 272, 1495–1497. â•‡ 25â•… E. Gomes, M. K. Jakobsen, K. B. Axelsen, M. Geisler, M. G. Palmgren, Plant Cell 2000, 12, 2441–2454.

308

MECHANISMS OF PS EXTERNALIZATION IN APOPTOSIS AND AGING

â•‡ 26â•… M. Darland-Ransom, X. Wang, C. L. Sun, J. Mapes, K. Gengyo-Ando, S. Mitani, D. Xue, Science 2008, 320, 528–531. â•‡ 27â•… A. Siegmund, A. Grant, C. Angeletti, L. Malone, J. W. Nichols, H. K. Rudolph, J. Biol. Chem. 1998, 273, 34399–34405. â•‡ 28â•… U. Marx, T. Polakowski, T. Pomorski, C. Lang, H. Nelson, N. Nelson, A. Herrmann, Eur. J. Biochem. 1999, 263, 254–263. â•‡ 29â•… S. Zullig, L. J. Neukomm, M. Jovanovic, S. J. Charette, N. N. Lyssenko, M. S. Halleck, C. P. Reutelingsperger, R. A. Schlegel, M. O. Hengartner, Curr. Biol. 2007, 17, 994–999. â•‡ 30â•… M. Bitbol, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 6783–6787. â•‡ 31â•… J. Connor, C. H. Pak, R. F. Zwaal, A. J. Schroit, J. Biol. Chem. 1992, 267, 19412–19417. â•‡ 32â•… J. Connor, M. Bar-Eli, K. D. Gillum, A. J. Schroit, J. Biol. Chem. 1992, 267, 26050–26055. â•‡ 33â•… S. Ruetz, P. Gros, Cell 1994, 77, 1071–1081. â•‡ 34â•… A. J. Smith, J. L. Timmermans-Hereijgers, B. Roelofsen, K. W. Wirtz, W. J. van Blitterswijk, J. J. Smit, A. H. Schinkel, P. Borst, FEBS Lett. 1994, 354, 263–266. â•‡ 35â•… P. Borst, N. Zelcer, A. van Helvoort, Biochim. Biophys. Acta 2000, 1486, 128–144. â•‡ 36â•… D. W. Dekkers, P. Comfurius, A. J. Schroit, E. M. Bevers, R. F. Zwaal, Biochemistry 1998, 37, 14833–14837. â•‡ 37â•… D. Kamp, C. W. Haest, Biochim. Biophys. Acta 1998, 1372, 91–101. â•‡ 38â•…Y. Hamon, C. Broccardo, O. Chambenoit, M. F. Luciani, F. Toti, S. Chaslin, J. M. Freyssinet, P. F. Devaux, J. McNeish, D. Marguet, G. Chimini, Nat. Cell Biol. 2000, 2, 399–406. â•‡ 39â•… C. Albrecht, J. H. McVey, J. I. Elliott, A. Sardini, I. Kasza, A. D. Mumford, R. P. Naoumova, E. G. Tuddenham, K. Szabo, C. F. Higgins, Blood 2005, 106, 542–549. â•‡ 40â•… J. R. Nofer, G. Herminghaus, M. Brodde, E. Morgenstern, S. Rust, T. Engel, U. Seedorf, G. Assmann, H. Bluethmann, B. E. Kehrel, J. Biol. Chem. 2004, 279, 34032–34037. â•‡ 41â•… P. Williamson, A. Kulick, A. Zachowski, R. A. Schlegel, P. F. Devaux, Biochemistry 1992, 31, 6355–6360. â•‡ 42â•… Q. Zhou, J. Zhao, T. Wiedmer, P. J. Sims, Blood 2002, 99, 4030–4038. â•‡ 43â•… X. Wang, J. Wang, K. Gengyo-Ando, L. Gu, C. L. Sun, C. Yang, Y. Shi, T. Kobayashi, Y. Shi, S. Mitani, X. S. Xie, D. Xue, Nat. Cell Biol. 2007, 9, 541–549. â•‡ 44â•… K. Bosslet, H. D. Mennel, F. Rodden, B. L. Bauer, F. Wagner, A. Altmannsberger, H. H. Sedlacek, H. Wiegandt, Cancer Immunol. Immunother. 1989, 29, 171–188. â•‡ 45â•… D. Mandal, P. K. Moitra, S. Saha, J. Basu, FEBS Lett. 2002, 513, 184–188. â•‡ 46â•… M. Walsh, R. J. Lutz, T. G. Cotter, R. O’Connor, Blood 2002, 99, 3439–3448. â•‡ 47â•… B. A. Klarl, P. A. Lang, D. S. Kempe, O. M. Niemoeller, A. Akel, M. Sobiesiak, K. Eisele, M. Podolski, S. M. Huber, T. Wieder, F. Lang, Am. J. Physiol. Cell Physiol. 2006, 290, C244–C253. â•‡ 48â•… S. K. Jain, J. Biol. Chem. 1984, 259, 3391–3394.

References

309

â•‡ 49â•… V. E. Kagan, J. P. Fabisiak, A. A. Shvedova, Y. Y. Tyurina, V. A. Tyurin, N. F. Schor, K. Kawai, FEBS Lett. 2000, 477, 1–7. â•‡ 50â•… J. M. Rifkind, K. Araki, E. C. Hadley, Arch. Biochem. Biophys. 1983, 222, 582–589. â•‡ 51â•… J. Connor, C. C. Pak, A. J. Schroit, J. Biol. Chem. 1994, 269, 2399–2404. â•‡ 52â•… J. L. Wolfs, P. Comfurius, O. Bekers, R. F. Zwaal, K. Balasubramanian, A. J. Schroit, T. Lindhout, E. M. Bevers, Cell. Mol. Life Sci. 2008, 66, 319–323. â•‡ 53â•… P. A. Lang, S. Kaiser, S. Myssina, T. Wieder, F. Lang, S. M. Huber, Am. J. Physiol. Cell Physiol. 2003, 285, C1553–C1560. â•‡ 54â•… J. L. Wolfs, S. J. Wielders, P. Comfurius, T. Lindhout, J. C. Giddings, R. F. Zwaal, E. M. Bevers, Blood 2006, 108, 2223–2228. â•‡ 55â•… K. de Jong, M. P. Rettig, P. S. Low, F. A. Kuypers, Biochemistry 2002, 41, 12562–12567. â•‡ 56â•… S. C. Frasch, P. M. Henson, J. M. Kailey, D. A. Richter, M. S. Janes, V. A. Fadok, D. L. Bratton, J. Biol. Chem. 2000, 275, 23065–23073. â•‡ 57â•… R. P. Rand, W. A. Pangborn, A. D. Purdon, D. O. Tinker, Can. J. Biochem. 1975, 53, 189–195. â•‡ 58â•… P. R. Cullis, B. de Kruijff, Biochim. Biophys. Acta 1979, 559, 399–420. â•‡ 59â•… W. J. Brown, K. Chambers, A. Doody, Traffic 2003, 4, 214–221. â•‡ 60â•… A. R. Poole, J. I. Howell, J. A. Lucy, Nature 1970, 227, 810–814. â•‡ 61â•… H. Wu, L. Zheng, B. R. Lentz, Biochemistry 1996, 35, 12602–12611. â•‡ 62â•… J. L. Knopf, M. H. Lee, L. A. Sultzman, R. W. Kriz, C. R. Loomis, R. M. Hewick, R. M. Bell, Cell 1986, 46, 491–502. â•‡ 63â•… L. L. Lin, M. Wartmann, A. Y. Lin, J. L. Knopf, A. Seth, R. J. Davis, Cell 1993, 72, 269–278. â•‡ 64â•… A. J. Verkleij, J. A. Post, J. Membr. Biol. 2000, 178, 1–10. â•‡ 65â•… C. R. Kiefer, L. M. Snyder, Curr. Opin. Hematol. 2000, 7, 113–116. â•‡ 66â•… T. K. Tang, J. Formos. Med. Assoc. 1997, 96, 779–783. â•‡ 67â•… A. Herrmann, P. F. Devaux, Biochim. Biophys. Acta 1990, 1027, 41–46. â•‡ 68â•… J. P. Fabisiak, Y. Y. Tyurina, V. A. Tyurin, J. S. Lazo, V. E. Kagan, Biochemistry 1998, 37, 13781–13790. â•‡ 69â•… P. A. Oldenborg, A. Zheleznyak, Y. F. Fang, C. F. Lagenaur, H. D. Gresham, F. P. Lindberg, Science 2000, 288, 2051–2054. â•‡ 70â•… G. Bartosz, E. Grzelinska, A. Bartkowiak, Mech. Ageing Dev. 1984, 24, 1–7. â•‡ 71â•… A. Chonn, S. C. Semple, P. R. Cullis, J. Biol. Chem. 1995, 270, 25845–25849. â•‡ 72â•… F. E. Boas, L. Forman, E. Beutler, Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 3077–3081. â•‡ 73â•… S. Piomelli, C. Seaman, Am. J. Hematol. 1993, 42, 46–52. â•‡ 74â•… G. B. Nash, S. J. Wyard, Biorheology 1980, 17, 479–484. â•‡ 75â•… O. Linderkamp, H. J. Meiselman, Blood 1982, 59, 1121–1127. â•‡ 76â•… C. J. van Oss, Biorheology 1982, 19, 725. â•‡ 77â•… M. R. Clark, N. Mohandas, S. B. Shohet, Blood 1983, 61, 899–910.

310

MECHANISMS OF PS EXTERNALIZATION IN APOPTOSIS AND AGING

G. B. Nash, H. J. Meiselman, Biophys. J. 1983, 43, 63–73. U. J. Dumaswala, T. J. Greenwalt, Transfusion 1984, 24, 490–492. A. D. Maher, P. W. Kuchel, Int. J. Biochem. Cell Biol. 2003, 35, 1182–1197. G. Bartosz, Gerontology 1991, 37, 33–67. E. Middelkoop, B. H. Lubin, E. M. Bevers, J. A. Op den Kamp, P. Comfurius, D. T. Chiu, R. F. Zwaal, L. L. van Deenen, B. Roelofsen, Biochim. Biophys. Acta 1988, 937, 281–288. â•‡ 83â•… J. F. Tait, D. Gibson, J. Lab. Clin. Med. 1994, 123, 741–748. â•‡ 84â•… F. A. Kuypers, R. A. Lewis, M. Hua, M. A. Schott, D. Discher, J. D. Ernst, B. H. Lubin, Blood 1996, 87, 1179–1187. â•‡ 85â•… K. de Jong, S. K. Larkin, L. A. Styles, R. M. Bookchin, F. A. Kuypers, Blood 2001, 98, 860–867. â•‡ 86â•… V. Borenstain-Ben Yashar, Y. Barenholz, E. Hy-Am, E. A. Rachmilewitz, A. Eldor, Am. J. Hematol. 1993, 44, 63–65. â•‡ 87â•… F. A. Kuypers, J. Yuan, R. A. Lewis, L. M. Snyder, C. R. Kiefer, A. Bunyaratvej, S. Fucharoen, L. Ma, L. Styles, K. de Jong, S. L. Schrier, Blood 1998, 91, 3044–3051. â•‡ 88â•… R. K. Wali, S. Jaffe, D. Kumar, V. K. Kalra, Diabetes 1988, 37, 104–111. â•‡ 89â•… M. J. Wilson, K. Richter-Lowney, D. L. Daleke, Biochemistry 1993, 32, 11302–11310. â•‡ 90â•… R. H. Wang, G. Phillips, Jr., M. E. Medof, C. Mold, J. Clin. Invest. 1993, 92, 1326–1335. â•‡ 91â•… C. Liu, P. Marshall, I. Schreibman, A. Vu, W. Gai, M. Whitlow, Blood 1999, 93, 2297–2301. â•‡ 92â•… R. S. Schwartz, Y. Tanaka, I. J. Fidler, D. T. Chiu, B. Lubin, A. J. Schroit, J. Clin. Invest. 1985, 75, 1965–1972. â•‡ 93â•… V. L. Lew, A. Hockaday, M. I. Sepulveda, A. P. Somlyo, A. V. Somlyo, O. E. Ortiz, R. M. Bookchin, Nature 1985, 315, 586–589. â•‡ 94â•… K. U. Suthipark, A. Likidlilid, S. Fucharoen, P. Pootrakul, D. Shumnumsirivath, S. Ong-ajyooth, D. Plaskett, J. Webb, Southeast Asian J. Trop. Med. Public Health 1991, 22, 171–175. â•‡ 95â•… B. Sauter, M. L. Albert, L. Francisco, M. Larsson, S. Somersan, N. Bhardwaj, J. Exp. Med. 2000, 191, 423–434. â•‡ 96â•… S. Gallucci, M. Lolkema, P. Matzinger, Nat. Med. 1999, 5, 1249–1255. â•‡ 97â•… D. H. MacLennan, J. Biol. Chem. 1970, 245, 4508–4518. â•‡ 98â•… C. Toyoshima, H. Nomura, Nature 2002, 418, 605–611. â•‡ 99â•… C. Toyoshima, M. Nakasako, H. Nomura, H. Ogawa, Nature 2000, 405, 647–655. 100â•… H. Streb, R. F. Irvine, M. J. Berridge, I. Schulz, Nature 1983, 306, 67–69. 101â•… H. Takeshima, S. Nishimura, T. Matsumoto, H. Ishida, K. Kangawa, N. Minamino, H. Matsuo, M. Ueda, M. Hanaoka, T. Hirose, Nature 1989, 339, 439–445. 102â•… E. Carafoli, Trends Biochem. Sci. 2003, 28, 175–181. 103â•… E. Carafoli, Trends Biochem. Sci. 2004, 29, 371–379. 104â•…Y. Kirichok, G. Krapivinsky, D. E. Clapham, Nature 2004, 427, 360–364. 105â•… N. E. Saris, E. Carafoli, Biochem (Russia) 2005, 70, 187–194. â•‡ 78â•… â•‡ 79â•… â•‡ 80â•… â•‡ 81â•… â•‡ 82â•…

References

311

106â•… P. Pinton, T. Pozzan, R. Rizzuto, EMBO J. 1998, 17, 5298–5308. 107â•… N. J. Dolman, A. V. Tepikin, Cell Calcium 2006, 40, 505–512. 108â•… P. R. Pryor, B. M. Mullock, N. A. Bright, S. R. Gray, J. P. Luzio, J. Cell Biol. 2000, 149, 1053–1062. 109â•… J. Meldolesi, T. Pozzan, Trends Biochem. Sci. 1998, 23, 10–14. 110â•… E. F. Corbett, M. Michalak, Trends Biochem. Sci. 2000, 25, 307–311. 111â•… M. Michalak, J. M. Robert Parker, M. Opas, Cell Calcium 2002, 32, 269–278. 112â•… M. J. Berridge, Cell Calcium 2002, 32, 235–249. 113â•… K. A. Christensen, J. T. Myers, J. A. Swanson, J. Cell Sci. 2002, 115, 599–607. 114â•… S. A. Hilden, N. E. Madias, J. Membr. Biol. 1989, 112, 131–138. 115â•… R. M. Lemons, J. G. Thoene, J. Biol. Chem. 1991, 266, 14378–14382. 116â•… K. Balasubramanian, B. Mirnikjoo, A. J. Schroit, J. Biol. Chem. 2007, 282, 18357–18364. 117â•… B. Mirnikjoo, K. Balasubramanian, A. J. Schroit, J. Biol. Chem. 2009, 284, 6918–6923. 118â•… K. D. E. L. Jozsef, T. Khreiss, J. G. Filep, Cell. Signal. 2006, 18, 2302–2313. 119â•… P. Karki, C. Seong, J. E. Kim, K. Hur, S. Y. Shin, J. S. Lee, B. Cho, I. S. Park, Cell Death Differ. 2007, 14, 2068–2075. 120â•… K. Cain, C. Langlais, X. M. Sun, D. G. Brown, G. M. Cohen, J. Biol. Chem. 2001, 276, 41985–41990. 121â•… H. Fathallah, M. Sauvage, J. R. Romero, M. Canessa, F. Giraud, Am. J. Physiol. 1997, 273, C1206–C1214. 122â•… D. A. Andrews, L. Yang, P. S. Low, Blood 2002, 100, 3392–3399. 123â•… M. C. Bassik, L. Scorrano, S. A. Oakes, T. Pozzan, S. J. Korsmeyer, EMBO J. 2004, 23, 1207–1216. 124â•… E. Cadenas, Mol. Aspects Med. 2004, 25, 17–26. 125â•… E. Cadenas, K. J. Davies, Free Radic. Biol. Med. 2000, 29, 222–230. 126â•… C. S. Boyd, E. Cadenas, Biol. Chem. 2002, 383, 411–423. 127â•… P. S. Brookes, A. L. Levonen, S. Shiva, P. Sarti, V. M. Darley-Usmar, Free Radic. Biol. Med. 2002, 33, 755–764. 128â•… T. J. Pinheiro, G. A. Elove, A. Watts, H. Roder, Biochemistry 1997, 36, 13122–13132. 129â•… A. J. Forsberg, V. E. Kagan, A. J. Schroit, Antioxid. Redox Signal. 2004, 6, 203–208. 130â•… C. Richter, Biosci. Rep. 1997, 17, 53–66. 131â•… I. I. Kruman, M. P. Mattson, J. Neurochem. 1999, 72, 529–540. 132â•… A. A. Starkov, C. Chinopoulos, G. Fiskum, Cell Calcium 2004, 36, 257–264. 133â•… N. F. Schor, Y. Y. Tyurina, J. P. Fabisiak, V. A. Tyurin, J. S. Lazo, V. E. Kagan, Brain Res. 1999, 831, 125–130. 134â•…Y. Y. Tyurina, A. A. Shvedova, K. Kawai, V. A. Tyurin, C. Kommineni, P. J. Quinn, N. F. Schor, J. P. Fabisiak, V. E. Kagan, Toxicology 2000, 148, 93–101. 135â•… A. Walter, D. Margolis, R. Mohan, R. Blumenthal, Membr. Biochem. 1986, 6, 217–237.

312

MECHANISMS OF PS EXTERNALIZATION IN APOPTOSIS AND AGING

136â•… A. Rietveld, W. Jordi, B. de Kruijff, J. Biol. Chem. 1986, 261, 3846–3856. 137â•… A. Rietveld, T. A. Berkhout, A. Roenhorst, D. Marsh, B. de Kruijff, Biochim. Biophys. Acta 1986, 858, 38–46. 138â•… A. E. Drobnies, E. A. Venczel, R. B. Cornell, Biochim. Biophys. Acta 1998, 1393, 90–98. 139â•… S. J. Gardai, K. A. McPhillips, S. C. Frasch, W. J. Janssen, A. Starefeldt, J. E. Murphy-Ullrich, D. L. Bratton, P. A. Oldenborg, M. Michalak, P. M. Henson, Cell. 2005, 123, 321–334. 140â•… M. Sorice, A. Circella, R. Misasi, V. Pittoni, T. Garofalo, A. Cirelli, A. Pavan, G. M. Pontieri, G. Valesini, Clin. Exp. Immunol. 2000, 122, 277–284. 141â•… M. Sorice, A. Circella, I. M. Cristea, T. Garofalo, L. Di Renzo, C. Alessandri, G. Valesini, M. D. Esposti, Cell Death Differ. 2004, 11, 1133–1145. 142â•… A. Shiratsuchi, T. Mori, Y. Takahashi, K. Sakai, Y. Nakanishi, J. Biochem. 2003, 133, 211–218. 143â•… D. P. Siegel, J. L. Burns, M. H. Chestnut, Y. Talmon, Biophys. J. 1989, 56, 161–169. 144â•… B. Mirnikjoo, K. Balasubramanian, A. J. Schroit, J. Biol. Chem. 2009, 284, 22512–22516. 145â•… A. Rodriguez, P. Webster, J. Ortego, N. W. Andrews, J. Cell Biol. 1997, 137, 93–104. 146â•… P. L. McNeil, T. Kirchhausen, Nat. Rev. Mol. Cell Biol. 2005, 6, 499–505. 147â•… P. L. McNeil, M. Terasaki, Nat. Cell Biol. 2001, 3, E124–E129. 148â•… P. L. McNeil, J. Cell Sci. 2002, 115, 873–879. 149â•… I. Martinez, S. Chakrabarti, T. Hellevik, J. Morehead, K. Fowler, N. W. Andrews, J. Cell Biol. 2000, 148, 1141–1149. 150â•… E. R. Chapman, Nat. Rev. Mol. Cell Biol. 2002, 3, 498–508. 151â•… E. R. Detrait, S. Yoo, C. S. Eddleman, M. Fukuda, G. D. Bittner, H. M. Fishman, J. Neurosci. Res. 2000, 62, 566–573. 152â•… Z. P. Pang, E. Melicoff, D. Padgett, Y. Liu, A. F. Teich, B. F. Dickey, W. Lin, R. Adachi, T. C. Sudhof, J. Neurosci. 2006, 26, 13493–13504. 153â•… M. Geppert, T. C. Sudhof, Annu. Rev. Neurosci. 1998, 21, 75–95. 154â•… N. Demaurex, C. Distelhorst, Science 2003, 300, 65–67. 155â•…Y. Shi, W. Zheng, K. L. Rock, Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 14590–14595. 156â•… V. A. Fadok, P. M. Henson, Curr. Biol. 1998, 8, R795–R805. 157â•… V. A. Fadok, D. L. Bratton, S. C. Frasch, M. L. Warner, P. M. Henson, Cell Death Differ. 1998, 5, 551–562. 158â•… V. A. Fadok, D. R. Voelker, P. A. Campbell, J. J. Cohen, D. L. Bratton, P. M. Henson, J. Immunol. 1992, 148, 2207–2216. 159â•… D. S. Pisetsky, Autoimmun. Rev. 2004, 3, 500–504. 160â•… K. L. Rock, A. Hearn, C. J. Chen, Y. Shi, Springer Semin. Immunopathol. 2005, 26, 231–246. 161â•… S. Chakrabarti, K. S. Kobayashi, R. A. Flavell, C. B. Marks, K. Miyake, D. R. Liston, K. T. Fowler, F. S. Gorelick, N. W. Andrews, J. Cell Biol. 2003, 162, 543–549.

References

313

162â•… D. Kanduc, A. Mittelman, R. Serpico, E. Sinigaglia, A. A. Sinha, C. Natale, R. Santacroce, M. G. D. Corcia, A. Lucchese, L. Dini, P. Pani, S. Santacroce, S. Simone, R. Bucci, E. Farber, Int. J. Oncol. 2002, 21, 165–170. 163â•… J. P. Luzio, P. R. Pryor, N. A. Bright, Nat. Rev. Mol. Cell Biol. 2007, 8, 622–632. 164â•… A. Terman, T. Kurz, B. Gustafsson, U. T. Brunk, IUBMB Life 2006, 58, 531–539.

15 PHOSPHATIDYLSERINE EXPOSURE IN HEMOGLOBINOPATHIES Frans A. Kuypers and Eric Soupene Children’s Hospital Oakland Research Institute, Oakland, CA

15.1â•… INTRODUCTION Hemoglobinopathies affect millions of individuals worldwide [1]. The hemoglobin mutations lead to alterations in the function of the red blood cell (RBC) membrane, which in turn plays an important role in the pathology in this patient population. Maintaining the lipid composition and organization of its plasma membrane is essential for RBC function. Mechanisms involving enzymatic reactions, membrane transport, and signal transduction pathways maintain membrane lipid homeostasis. Dysfunction of one or more of these mechanisms in hemoglobinopathies, such as sickle cell disease and thalassemia, leads to pathophysiologic problems. The structural backbone of the RBC plasma membrane comprises more than 250 molecular species of glycerophospholipid and sphingomyelin (SM) [2], synthesized de novo during the early stages of erythropoiesis. These phospholipids, together with cholesterol, make up most of plasma membrane lipids. The major phospholipid classes move dynamically in the plane and across the lipid bilayer of the RBC, but this movement is highly organized and leads to areas that are enriched in specific phospholipid molecular species. The mature mammalian RBC lacks internal organelles and is incapable of de novo lipid synthesis. Nevertheless, a very dynamic system maintains the overall composition and organization of the membrane during the life of the Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

315

316

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

adult RBC. Lipid molecules are continuously renewed and rapidly move in the plane of each membrane layer, and are transported across the bilayer by proteins. Fatty acids are rapidly taken up from plasma and incorporated into phospholipids. Despite this rapid phospholipid turnover, and the continuous changes in the substrate pool (plasma fatty acids, determined by diet), the molecular species composition of RBC is remarkably constant. This indicates a selective system in which several components act in sync, and it seems appropriate to hypothesize that the activities of these proteins are regulated by lipid–protein and/or protein–protein interactions in the membrane that “sense” the need for the generation of certain molecular species or the need for lipid (re-)distribution in the plane and across the bilayer. The mechanisms for these interactions are largely unknown, in part because many of the membrane proteins involved are poorly characterized. Similarly, whereas the exposure of phosphatidylserine (PS) on the surface of both adult RBC and RBC precursors is well recognized as an important contributor to the pathology in sicklecell disease (SCD) or thalassemia, the mechanisms that lead to this exposure are still poorly defined. 15.2â•… RBC PHOSPHOLIPID ORGANIZATION The difference in phospholipid composition between the two leaflets of the mammalian RBC plasma membrane bilayer was first reported more than three decades ago [3, 4] (see Chapter 3). Studies using chemical labeling [3] or phospholipase degradation [5] to differentiate the lipids located in the inner and outer leaflet established that the outer monolayer of the red cell membrane contains mainly the choline-containing phospholipids SM and phosphatidylcholine (PC), and that the inner leaflet contains most of the phosphatidylethanolamine (PE) and all of the PS. This asymmetric transbilayer organization in the membrane of the RBC was originally regarded as a static distribution across the bilayer, generated during erythropoiesis and maintained during the life of the cell, in part by the interaction between PS and spectrin [6–9]. Phospholipids not only move rapidly in the plane of the bilayer but can also move across the bilayer. The transbilayer movement of phospholipids is slow in pure lipid bilayers but can be accelerated in the presence of proteins. In 1984, it was shown that the asymmetric transbilayer distribution is the result of an ATP-dependent transport process [10]. The aminophospholipids PS and PE are actively transported from the outer to the inner monolayer by an aminophospholipid translocase or flippase [11] (see Chapters 8–10). The removal of virtually all the PS from the surface of the plasma membrane seems typical for normal mammalian cells [12]. Measurements of phospholipid transport first developed for RBC are variants on the same basic principle. A labeled phospholipid analog is introduced into the outer leaflet of the plasma membrane. After a defined incubation time, the fraction of that analog in the

RBC Phospholipid Organization

317

inner monolayer is measured. This is accomplished by removal, or chemically altering the fraction in the outer monolayer. This approach allows the assessment of the equilibrium state at that specific point in time. In RBC, this equilibrium is established across one bilayer, the plasma membrane. When the phospholipid analog can distribute to intracellular membranes, as is the case in cells with internal membranes, the results of plasma membrane transbilayer movement and equilibrium distribution are much more difficult to interpret (see, e.g., Reference 13). The ATP-driven transbilayer movement in the RBC is specific for the aminophospholipid head group (PE and PS), but seems not affected by the stereo isoforms (D or L) of the head group (see Chapter 10). On the other hand, transport was reported to be stereo-specific for the C2 stereo center on the C2 carbon of the glycerol backbone [14]. The mode of binding of the fatty acyl chains to the glycerol backbone (ester, ether, or vinyl ether) affects the movement of nitroxide and fluorescent phospholipid analogs [15]. Whereas aminophospholipids are transported regardless of length and unsaturation of the fatty acid side chains, the kinetics of transbilayer movement may be affected [16, 17]. Most of the reported results on transbilayer movement are acquired by the use of phospholipid analogs as probes. In most cases, these phospholipid molecules are distinctly different from the native molecular phospholipid species. While measurements of spin-labeled or fluorescent phospholipid probes have rendered important information [15], they can only describe the movement of these particular molecules across the bilayer. The kinetics of movement of individual native molecules in the phospholipid molecular species mixture [2] are difficult to assess and are limited to a small number of available radioactive molecules [16, 17]. Based on the results obtained with radioactive species, which are in structure the closest to the native phospholipids in the membrane, it seems that different molecules exhibit different rates of movement across the bilayer. Hence, generalizations with respect to the kinetics of transbilayer movement should be considered with caution. Despite these concerns, it has been established that transbilayer movement of phospholipid analogs reaches an equilibrium distribution across the bilayer that correlates very well with reports that use the chemical labeling and phospholipase degradation of native phospholipid species [15]. Importantly, it should be noted that this approach reports on transbilayer movement in both directions of the labeled phospholipid. A distinction between inward and outward movement is often difficult to make, in particular when transbilayer movement is fast, relative to the quenching or removal reaction (see Chapter 6). In addition, the movement in one direction across the bilayer is coordinated with a movement in the reverse direction. It was reported that membrane surface area asymmetry could be a driving force for vesicle formation during endocytosis, and that transient asymmetries in lipid concentration might contribute to the formation of endocytic vesicles [18] (see Chapters 2, 12, and 13). Conversely, the transfer of PS from the outer to inner monolayer requires movement of lipid molecular species in the other direction to maintain a proper balance across the bilayer, and with it the shape of the RBC. Moreover,

318

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

in addition to a mass balance, each phospholipid molecular species has a specific shape or geometry, based on the head group and the fatty acyl chains. Replacement of native molecular species of PC in the RBC membrane with PC species that have a different geometry can lead to shape changes, that can be predicted based on the shape of the molecules exchanged [19] (see Chapter 2). Therefore, a redistribution of phospholipids across the bilayer must be a specific process to maintain the balance across the bilayer, governed by the sum of the geometries of the species in both leaflet to maintain the overall shape of the RBC. The mechanisms that govern this redistribution of phospholipid molecular species across the bilayer are poorly understood. The “flip” of aminophospholipid in one direction and the “flop” of another molecule in the reverse to compensate may involve a “floppase” or proteins such as the multidrug resistance class of transporters [20, 21]. While a different sensitivity to inhibitors for flip and flop has been used to argue a distinction between flippase and floppase activity [22] (see Chapter 11), it cannot be excluded that the flippase plays a role in the transport of phospholipids in the reverse direction to maintain a proper balance across the bilayer. It is also important to consider that this redistribution is affected by the rapid turnover of phospholipids in the RBC membrane. The incorporation of new fatty acids at the inner monolayer and movement of these species to the outer monolayer links transbilayer movement with the network of proteins involved in the maintenance of the RBC phospholipid composition. The turnover of membrane fatty acyl groups and redistribution of phospholipids across the bilayer are, in part, the results of the basic function of the RBC: transport of oxygen. The binding to and release of oxygen from hemoglobin lead to the generation of high levels of reactive oxygen species (ROS). Despite the presence of a network of antioxidant systems, oxidant stress will lead to damage of the polyunsaturated acyl groups in the different phospholipid classes. This is in particular important either when antioxidants are compromised or, as is the case in hemoglobinopathies, when more ROS are generated. The polyunsaturated fatty acids in the phospholipids are prime targets for oxygen radical alterations. When oxygen reacts with the double bonds, it introduces a polar entity in the apolar chains. This breach in the normal membrane structure is recognized by phospholipases [23], which hydrolyze the ester bond and generate lysophospholipids starting a repair process [24]. Several proteins involved in the deacylation/reacylation mechanism (the so-called Lands pathway [25]) have been identified, including acylcoenzyme A (CoA) synthetases, acyl-CoA-binding proteins, and acyl-CoA transferases [26–29]. The acyl-CoA synthetase ACSL6 activates fatty acids taken up from plasma by ligating the fatty acid to CoA at the expense of ATP [29]. Subsequently, the acyl-CoA is transferred to lysophospholipid by members of the elusive lysophospholipid acyl-CoA acyltransferase (LPLAT) family. Different members of the LPLAT family have preferences for the different phospholipid head groups, and different proteins may be responsible for the reacylation of lysophosphatidylcholine (LPC), lysophosphatidylethanolamine

The RBC Flippase

319

(LPE), and lysophosphatidylserine (LPS). The protein that reacylates LPC to PC in the RBC membrane was recently identified [30]. Oxidant damage is random in nature, and the phospholipid repair system should be able to replace the damaged phospholipid with the same molecular species to maintain phospholipid composition. This requires the ability to reacylate lysophospholipids using a specific fatty acid from a substrate pool that changes depending on diet. The mechanisms that govern this well-regulated repair system are not known. Transbilayer movement is coupled with this reacylation process. The reacylation takes place in the inner monolayer, and proper redistribution of the molecular species is required, involving the proteins that transport phospholipids across the bilayer. Therefore, it is logical to assume that the redistribution of these phospholipid species across the bilayer, when they are generated at the inner leaflet, is also a specific process. A fast translocation from inner to outer monolayer of newly synthesized PC but not PE is observed, suggesting a selective, protein-mediated outward translocation of newly synthesized PC [31]. The proteins involved in this repair process are themselves vulnerable to oxidant damage. Since the adult RBC is not able to replace damaged proteins, this may lead to the inability to properly repair damaged phospholipids, a dysfunction of the membrane, and demise of the cell. Together, a very complex picture emerges of phospholipid movement, kinetics, and distribution for the different molecular species across the RBC bilayer. Nevertheless, the phospholipid head-group distribution at equilibrium is well maintained during the life of the RBC. 15.3â•… THE RBC FLIPPASE The RBC flippase has been described to be involved in aminophospholipid transport activity for the last 25 years, but the identification of this protein, needed as a first step toward determining its structure, has proven to be very difficult. In 1989, Zachowski et al. in the laboratory of Devaux [32] showed that chromaffin granules contain a vanadate sensitive, ATP-dependent phospholipid transporter with similar substrate specificity as was observed in the RBC membrane. Moriyama and Nelson [33, 34] purified a membrane ATPase called ATPase II from the chromaffin granule membrane, which was identified by the laboratory of Williamson as ATP8A1, a member of a new group of P-type ATPases [13]. This protein presents all of the biochemical characteristics of the erythroid flippase. Like the RBC flippase, it requires Mg2+-ATP for activity, is stimulated by PS, and inhibited by calcium and vanadate. The activity is affected by sulfhydryl-modifying agents, such as N-ethylmaleimide (NEM), and the protein is highly specific for glycerol phospholipids with an amino head group [33, 35, 36]. Moreover, overexpression in a neuronal cell line results in a large increase of PS transport. The family of P-type ATPase transport proteins contains well-known enzymes that transport Ca2+ or Na+

320

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

and K+ ions [37]. Whereas members of this family of proteins differ significantly in sequence and the compound they transport, domains conserved in all members of the flippase subclass, as well as motif specifics, can be identified (see Chapters 8 and 9). The large family of P-type ATPases is divided into subfamilies or types (1–5), further subdivided in classes, based on sequence similarity and with apparently distinct substrate specificity. Not all types are present in every organism, and although the cellular functions of many types are still unknown, several have been associated with disorders. Type 3 is absent in animals, including mammals, while types 4 and 5 are exclusively found in eukaryotes. Table 15.1 summarizes several (human) forms of these proteins (types 1, 2, 4, and 5). Types 1, 2, and 3 are proteins that couple ATP hydrolysis to transport various ions; the function of type 5 proteins is largely unknown. Type 4 P4-ATPases differ from the cation-ATPase pumps as they do not appear to transport ions. In fact, several have been shown to flip phospholipid across membranes and constitute good candidates for the elusive plasma membrane flippase. The members of this diverse lipidtransporting subfamily are relatively poorly defined to date [38, 39], and unfortunately, there are many reports in the literature referring to flipase or flippase or phospholipid translocase even though the function of the proteins reported is different. This has created confusion regarding the definition of a flippase, as it applies to the RBC. Moreover, in several systems, “flippase activity” is often used for a characteristic of a membrane rather than a specific protein. Based on the original description of the RBC flippase, the protein should be able to transport PS across the RBC bilayer, to reside in the plasma membrane, and to process a hydrolyzable form of Mg2+-ATP, with specificity for aminophospholipid. The protein is capable of unidirectional transport from the outer to cytofacial layer and requires PS (or PE) for dephosphorylation of the obligatory aspartyl-phosphate intermediate. This protein should be sensitive to vanadate and sulfhydryl modification with compounds such as NEM. Studies that define lipid movement in membranes of internal organelles, such as ER–Golgi vesicles, instead of the plasma membrane often describe flippase activity as the product of several different genes and, equally important, the presence of different spliced isoforms of the same gene. In addition, various proteins and/or mutants, particularly in yeast (see Chapters 8 and 9), were described as belonging to the flippase family but were shown to be involved in various unrelated cellular processes such as ribosome assembly, transport of compounds not related to aminophospholipids (such as PC, amino glucoside), and proteins that are bidirectionally active without a need for ATP. Thus, the current literature is rather confusing as many of the socalled flippases act as “floppases” or “scramblases.” The group of P4 or type IV P-type ATPases [39] includes 14 genes in mammals and 5 genes in yeast. In humans, 14 of the 36 known P-ATPases belong to this subfamily, and as many as 80 different spliced isoforms are known (SWISS-PROT). Three members of this type have been studied more extensively, as they are linked to disease. In humans, an inherited mutation in ATP8B1 (FIC1), shown to

321

ATP7A

ATP7B

ATP2A1

ATP2A2

ATP2A3 ATP2C1

ATP2C2

ATP2B1

ATP2B2

ATP2B3

ATP2B4

ATP1A1

ATP1A2

ATP1A3

Cu+

Cu+

Ca2+

Ca2+

Ca2+ Ca2+/Mn2+

Ca2+/Mn2+

Ca2+

Ca2+

Ca2+

Ca2+

Na+/K+

Na+/K+

Na+/K+

1B

1B

2A

2A

2A 2A

2A

2B

2B

2B

2B

2C

2C

2C

Gene

Substrate (Known or Predicted)

Type

TABLE 15.1.â•‡ The Human P-Type ATPases

MHP2

PMCA4

PMCA3

PMCA2

PMCA1

SPCA2

SERCA3 BCPM

SERCA2

SERCA1

WND

MNK

Previous Symbols ATPase, Cu++ transporting, alpha polypeptide (Menkes syndrome) ATPase, Cu++ transporting, beta polypeptide ATPase, Ca++ transporting, cardiac muscle, fast twitch 1 ATPase, Ca++ transporting, cardiac muscle, slow twitch 2 ATPase, Ca++ transporting, ubiquitous ATPase, Ca++ transporting, type 2C, member 1 ATPase, Ca++ transporting, type 2C, member 2 ATPase, Ca++ transporting, plasma membrane 1 ATPase, Ca++ transporting, plasma membrane 2 ATPase, Ca++ transporting, plasma membrane 3 ATPase, Ca++ transporting, plasma membrane 4 ATPase, Na+/K+ transporting, alpha 1 polypeptide ATPase, Na+/K+ transporting, alpha 2 (+) polypeptide ATPase, Na+/K+ transporting, alpha 3 polypeptide

Approved Name

(Continued)

NM_152296

NM_000702

NM_000701

NM_001001396

NM_021949

NM_001001331

NM_001682

NM_014861

NM_174953 NM_001001486

NM_001681

NM_004320

NM_000053

NM_000052

Accession Numbers

322

ATP12A

ATP4A

H+/K+

H+/K+

PL

PL

PL PL PL PL PL PL PL PL PL PL PL PL ? ? ? ? ?

2C

2C

4

4

4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5

FIC1, BRIC, PFIC1 ATPID, KIAA1137 ATPIK ATPIM, KIAA1939 KIAA0611, ATPIIA ATPIIB ATP10C ATPVB, KIAA0715 ATPVD, KIAA1487 ATPIH, ATPIS, KIAA1021 ATPIF, ATPIR, KIAA0956 ATPIG, ATPIQ KIAA1825, CGI-152 HSA9947 AFURS1 DKFZp761I1011 FLJ16025

ATPIB, ML-1

ATPase II, ATPIA

ATP6A

ATP1AL1

ATP1AL2

Previous Symbols ATPase, Na+/K+ transporting, alpha 4 polypeptide ATPase, H+/K+ transporting, nongastric, alpha polypeptide ATPase, H+/K+ exchanging, alpha polypeptide ATPase, aminophospholipid transporter (APLT), class I, type 8A, member 1 ATPase, APLT-like, class I, type 8A, member 2 ATPase, class I, type 8B, member 1 ATPase, class I, type 8B, member 2 ATPase, class I, type 8B, member 3 ATPase, class I, type 8B, member 4 ATPase, class II, type 9A ATPase, class II, type 9B ATPase, class V, type 10A ATPase, class V, type 10B ATPase, class V, type 10D ATPase, class VI, type 11A ATPase, class VI, type 11B ATPase, class VI, type 11C ATPase type 13A1 ATPase type 13A2 ATPase type 13A3 ATPase type 13A4 ATPase type 13A5

Approved Name

The table indicates the type and the substrate for the ATPase, either known or predicted, as well as the accession numbers.

ATP8B1 ATP8B2 ATP8B3 ATP8B4 ATP9A ATP9B ATP10A ATP10B ATP10D ATP11A ATP11B ATP11C ATP13A1 ATP13A2 ATP13A3 ATP13A4 ATP13A5

ATP8A2

ATP8A1

ATP1A4

Na+/K+

2C

Gene

Substrate (Known or Predicted)

Type

TABLE 15.1.â•‡ (Continued )

NM_005603 NM_020452 NM_138813 NM_024837 NM_006045 NM_198531 NM_024490 NM_025153 NM_020453 NM_015205 NM_014616 NM_173694 NM_020410 NM_022089 NM_024524 NM_032279 NM_198505

NM_016529

NM_006095

NM_000704

NM_001676

NM_144699

Accession Numbers

The RBC Flippase

323

transport conjugated bile acids, is responsible of two types of autosomalrecessive liver disorders, Summerskill syndrome and Byler disease, which are characterized by a defect in bile secretion [40]. Mutations in ATP10A (previously indicated as ATP10C, see Table 15.1) are associated with Angelman syndrome and obesity [41]. In mice, disruption of another member exclusively expressed in testis, Atp8B3, leads to premature PS exposure of spermatozoa (before sperm capacitation), altering their ability to penetrate the zona pellucida and to fertilize eggs. The murine version of ATP8A1 (the chromaffin granule enzyme) was expressed in, and purified from, cultured insect cells. This purified enzyme was then used to study the activation of the enzyme by phospholipids and phospholipid analogs [35, 36]. The presence of ATP8A1 was reported in adult RBC and the mRNA in RBC precursors [42], and confirmed by two recently published full-scale RBC proteome studies [43, 44]. Whereas the chromaffin granule enzyme was purified from detergent extracts on the basis of ATPase activity, transport of phospholipids can only be assayed in an intact membrane system. When ATP8A1 is expressed in yeast vesicles [45], it acts as an aminophospholipid-activated P-ATPase that can be inhibited by similar factors that affect transbilayer movement in RBC and is able to transport 2-(6-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)hexanoyl-1-hexadecanoyl-snglycero-3-phosphoserine (NBD-PS) across a bilayer [45]. The structures of three different P-type ATPases, including multiple conformations of the CaATPase [46], as well as studies on the mechanism of Ca2+ transport [47], have provided evidence that the basic mechanisms of ATP hydrolysis and energy transduction are likely applicable to the entire family [48, 49]. The ability to move large, complex molecules with a polar head group across the apolar phase of the lipid bilayer distinguishes the phospholipid transporters from the ion (Ca2+, Na+, K+) transporting ATPases. Whereas this ability to transport phospholipids is very different from the ATP-driven movement of small, spherical metal ions, the mechanism to transport large, amphipathic molecules may have evolved from the ability to transport cations across a lipid bilayer, and based on this assumption, a model was proposed for the steps of phospholipid movement in the lipid transporters [50]. Similarly, to the calcium ATPase, the hydrolysis of Mg2+-ATP, the phosphorylation of the flippase, the binding of the phospholipid, and dephosphorylation describe several aspects of its action. However, many questions remain to be answered. In addition to the unknown mechanisms how an aminophospholipid is actually transported from the outer to inner monolayer, it is unclear how the balance across the bilayer is maintained. As indicated above, the shape of the bilayer is regulated by a fine balance of the total amount of phospholipid molecular species and their individual packing characteristics across the bilayer. The role of the flippase in these movements and counter movements is currently poorly understood. It has been suggested that the P4-ATPases might require a subunit. This is intriguing since most P-type ATPases operate as self-sufficient monomers.

324

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

Na/K-ATPases form a complex with two smaller proteins to form a heterotrimer [49]. The function of these subunits, and particularly the larger subunit that confers ouabain sensitivity on the Na/K-ATPase, is still unclear [51, 52]. Drs2 (defined as “deficiency in ribosomal subunits”) is the closest relative of ATP8A1 in yeast and was reported to be associated with Cdc50, a protein previously studied for its role in polarization of cell morphology during bud formation [53]. Recently, it was shown that Cdc50 may play a vital role in the DRS2p protein [54]. Yeast vesicles containing ATP8A1 exhibit PSstimulated ATPase and transport activity without the co-expression of another mammalian protein [45]. Similar vesicles generated with an empty vector have very low ATPase activity and do not transport PS upon the addition of ATP [45]. Since no subunit was introduced in this expression system, these results suggest that ATP8A1 can act without the presence of a mammalianderived subunit. It cannot be excluded that such a cofactor would affect flippase activity in the intact RBC or that, not yet identified yeast proteins, would support the action of the mammalian flippase when expressed in yeast. Figure 15.1 indicates the putative action of a P-ATPase to transfer aminophospholipids across the bilayer and the compensatory movement of phospholipids in the reverse direction. The structure of the RBC flippase is not known and the depicted molecular structure is based on the calcium transporter [46]. 15.4â•… PS EXPOSURE IN RBCS The loss of the asymmetric distribution of phospholipids across the bilayer, and in particular the exposure of PS on the surface of the cell, is an important starting point in many physiological processes including the regulation of blood coagulation [55], and the recognition and removal of apoptotic cells [56] (see Chapter 14). Dying cells, with few exceptions [12, 57], trigger cellular events that result in PS appearing on the cell surface. This PS exposure typically occurs during the last stage of the life of a cell, and triggers phagocyte docking and subsequent engulfment and degradation of the apoptotic cell [12, 58–60]. PS externalization is critical to this process as its absence results in impaired recognition and clearance of apoptotic debris, which can lead to inflammatory and autoimmune responses [61, 62]. The appearance of PS at the apoptotic cell surface has been reported to be regulated by specific intracellular signaling events including mitochondrial stress, changes in cytosolic calcium [63], oxygen radicals, cytochrome C release, caspase activation, and DNA fragmentation [64]. Similar to apoptosis, PS is exposed in RBC when phospholipids are scrambled across the bilayer in both directions (see Chapter 7). Several in vitro conditions can lead to RBC PS exposure [65], and PS-exposing RBCs in vivo are found in RBCs from sickle-cell and thalassemia patients [66]. Given the importance of PS exposure, several techniques have been developed that allow

325

PS Exposure in RBCs

Phospholipid (%)

Outer monolayer

SM PC PS PE Inner monolayer

Flop

Flip +

– ATP Flippase

Scramblase

Ca++

Figure 15.1.â•‡ The red cell membrane bilayer. A large variety of phospholipid molecular species moves rapidly in the plane of the RBC plasma membrane in a dynamic but highly organized fashion. Areas are enriched in certain lipids essential for proper protein function, and phospholipids are asymmetrically distributed across the lipid bilayer. An aminophospholipid translocase or flippase uses ATP to transport amino phospholipids (PS and PE) from the outer to inner monolayer. This results in an asymmetric bilayer distribution with choline-containing phospholipids (SM and PC) predominantly in the outer leaflet and PE and PS in the inner leaflet. The inward movement of aminophospholipids must be compensated by an outward movement to maintain balance across the bilayer and the shape of the RBCs. PS is exposed by a calcium- and protein-facilitated scrambling process. At present, little structural information is available on the red cell flippase or scramblase. Based on sequence analysis and comparison with other P-ATPases such as SERCA1 [46], it is expected that only a small portion of the protein extends beyond the 10 transbilayer stretches into the medium outside the cell. Similarly, a structural model of the putative scramblase TMEM16F is indicated based on the reported structure of TMEM16A, which shares a momodimeric architecture with voltage-gated chloride channels [155]. Color version on the Wiley web site.

the assessment of PS on the surface of the RBC, in particular the use of Annexin V analogs [67] and components of the prothrombinase complex [55]. Fluorescent labeling of these proteins allows the identification of PS-exposing cells in microscopy or flow cytometry [68, 69], and allows the identification of cells that are unable to maintain their normal phospholipid asymmetry. The underlying mechanism for this scrambling process and exposure of PS is not clear. It involves both a dysfunction of the flippase and increased transbilayer movement defined as phospholipid scrambling. The transbilayer movement in pure lipid bilayers is relatively slow, but many compounds that intercalate in

326

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

the bilayer increase phospholipid bilayer movement. The channel-forming antibiotic gramicidin [70], amphotericin [71], or alcohols [72] will increase transbilayer movement, often correlated with changes in membrane packing [73] or lipid-phase characteristics [74]. Dielectric breakdown of the RBC membrane [75, 76] will increase transbilayer movement, and the incorporation of the calcium ionophore A23187 in the presence of Ca2+ [77] is well established to scramble membrane phospholipids. Exposure of PS on the surface induced by the increase in cytosolic calcium could involve a complex of phosphatidylinositol bisphosphate (PIP2) and Ca2+ [78], but other factors were shown to play a role [79]. The multidrug resistance protein (MRP) has been reported to be involved in the movement from inner to outer leaflet [80, 81], but also ubiquitous RBC membrane proteins such as the anion exchanger Band 3 (AE1) are reported to be involved in the movement of phospholipids across the membrane [22, 39, 72, 82, 83]. Together, many factors may affect phospholipid movement across the RBC bilayer, and it has been difficult to clearly define a specific floppase or scramblase that is activated under conditions that lead to the exposure of PS. The purification of the human phospholipid scramblase 1 (PLSCR1) [84], and its presence in the RBC suggested that this protein facilitated the calcium-induced phospholipid scrambling (see Chapter 7). PLSCR1 is a protein of 318 amino acids that is found in many cell types [84–86]. It is one of the four members of the scramblase family annotated in the human genome. This family of cytoplasmic membrane-associated proteins was identified based on their capacity to mediate a Ca2+-dependent bidirectional movement of phospholipids across membrane bilayers [86–88], and to share between 46% and 59% amino acid identity in their primary sequences [85]. However, the exact function and mechanism(s) of these proteins remain obscure. Scott syndrome, a rare congenital bleeding disorder, shows a defective scrambling of membrane phospholipids of activated blood platelets and platelet-derived microvesicles [89], and led to the assessment of PLSCR1 function in these patients. The lack of response to calcium-induced scrambling in blood of these patients is also found in lymphoid cells, but PS exposure is induced normally during apoptosis [90]. The deduced sequence of PLSCR1 in these patients is identical to that of normal controls [91], and the protein isolated from detergent-solubilized RBC exhibits normal function when incorporated into proteoliposomes. These data suggested that the defect in Scott syndrome is related either to aberrant posttranslational processing or to a defect or deficiency in an unknown cofactor that is required for normal expression of plasma membrane PLSCR function in situ [92], or that the PLSCR1 is not a scramblase. The impaired procoagulant response was partially restored by pretreatment with valinomycin, providing evidence for the involvement of efflux of K+ ions through Ca2+-activated K+ channels in the procoagulant response [93]. Mouse knockouts of PLSCR1 affects myelopoiesis in response to granulocyte colony-stimulating factor (G-CSF), causing a defect in emergency

PS Exposure in RBCs

327

granulopoiesis, but RBCs that lack PLSCR1 show a normal phospholipid scrambling response to elevated calcium. It was concluded that in RBCs, another scramblase may be present compensating for the loss of PLSCR1. Proteomic studies showed the presence of PLSCR4 [43] in human RBC, a protein of which little is known regarding the subcellular localization or specific function (see Reference 94 for a review). The role of either PLSCR1 or PLSCR4 in RBC is unclear and is complicated by recent findings that indicate that these proteins may serve functions other than their proposed activity as PLSCRs [95]. Data from several laboratories suggest that in addition to their putative role in mediating transbilayer movement of membrane lipids, the PLSCRs may also function to regulate diverse processes including signaling, apoptosis, cell proliferation, and transcription. When palmitoylated, PLSCR1 was shown to be raft-associated and interacts with ligand-activated epidermal growth factor (EGF) receptors, the protein serves as an adapter to promote interaction between the adapter protein Shc and the EGF receptor kinase [96]. PLSCR1 has been shown to be involved in mast cell activation [97], and both PLSCR1 and PLSCR4 interact directly with the CD4 receptor and are cellular receptors for the secretory leukocyte protease inhibitor [98]. Sequence analysis indicates that these proteins belong to a new superfamily of membranetethered transcription factors [99]. Palmitoylation is essential for membrane binding, and the predicted three-dimensional structure suggests a closed, symmetric beta-barrel of 12 beta-strands wrapped around a very hydrophobic C-terminal helix [99] unlikely to be embedded in the membrane, and therefore, it seems puzzling how these proteins themselves could induce lipid scrambling in a lipid bilayer. Recently, it was reported that a cell line with a D to G point mutation at position 409 in TMEM16F, a protein with eight transmembrane segments, rendered these cells highly sensitive to calcium-induced phospholipid scrambling [100]. The TMEM16 family, to which TMEM16F belongs, consists of 10 members in humans and mice. The founding member of the family, human TMEM16A, is a calcium-dependent chloride channel [101–103]. Even in resting cells, in which the cytosolic calcium concentration was below 100â•¯nM, phospholipid scrambling was noted in cells with the D409G mutant. These cells showed an increased labeling with Annexin V or MFG-E8, as well as Ro09-0198, indicating increased exposure of both amionophospholipids PS and PE [104–106]. The role of TMEM16F in phospholipid scrambling was further indicated by the analysis of a cell line derived from a patient with Scott syndrome [89, 107]. While A23187 resulted in PS exposure in cell lines derived from controls, B-cell lines established from a patient with Scott syndrome cells did not expose PS in response to a calcium ionophore agreement with previous reports [90]. In contrast to controls, cells from the patient were found to carry a mutation at a splice-acceptor site of the gene encoding TMEM16F, causing the premature termination of the protein. Together, these data strongly suggest that TMEM16F plays a role in calcium-induced plasma membrane scrambling. Figure 15.1 indicates the presence of TMEM16F as an RBC scramblase, but experimental evidence that this protein is responsible for PS exposure in RBC

328

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

is currently lacking. It is expected that the putative role of TMEM16F in phospholipid scrambling in the RBC membrane, either by itself or in combination with other factors, will be elucidated soon. 15.5â•… PS EXPOSURE IN HEMOGLOBINOPATHIES In SCD and thalassemia, the exposure of PS on the RBC surface is highly amplified. These hemoglobinopathies affect many millions of individuals worldwide. The mutation in the globin genes may affect the structure of hemoglobin as seen in SCD, or the imbalanced generation of the globin chains (thalassemia syndromes). The changes in hemoglobin can have a profound effect on the RBC membrane, which in turn will lead to the altered behavior of the RBCs in the circulation and the vasculopathy typical for these genetic disorders. In early studies, phospholipase treatment of RBC samples from sickle-cell patients indicated a loss of phospholipid asymmetry related to the sickling process [108, 109]. It was thought that the separation of the membrane skeleton from the lipid bilayer in areas as the result of hemoglobin polymerization led to a rapid transbilayer movement and the exposure of PS [110] in all sickled cells. Using flow cytometry and PS labeling techniques, it has become clear that both in thalassemia and SCD subpopulations of red cells are present in the peripheral blood that expose PS [68, 111–113]. Most of the RBCs do not expose PS and seem to be able to maintain their phospholipid asymmetry. Moreover, the size of the PS-exposing subpopulation varies largely from patient to patient. While some patients do not seem to exhibit PS-exposing cells higher than controls, the presence of 10% of PS-exposing cells has been reported in the circulation of hemoglobinopathy patients. The presence of these PS-exposing RBCs in the circulation may be a hallmark of “premature aging” of the RBCs and be related to the anemia in these patients. The lifespan of RBCs in hemoglobinopathies is significantly shorter than the RBC survival in normal humans. Exposure of PS leads to recognition and removal of apoptotic cells in other tissues, and similarly, PSexposing RBC will be removed from the circulation. Interestingly, as these cells can be observed in the peripheral blood, this indicates that PS-exposing RBC are being generated in such high numbers that the normal processes to remove these PS-exposing cells is compromised, or simply are unable to keep up. The compromised spleen function in sickle-cell patients [114] may play an important role in the inability of the body to remove these cells from the circulation. This seems underscored by the fact that PS-exposing cells are specifically found in thalassemia patients that are splenectomized [112, 115]. In hemoglobin E-beta-thalassemia, a relation was reported between hemoglobin E levels, the loss of sialylated glycoconjugates, and PS exposure on the RBC surface, indicating a correlation with the severity of the disease [116]. In addition to PS-exposing RBCs in the peripheral blood, RBC precursors are present in bone marrow that, similar to nonviable cells in other

Consequences of PS Exposure

329

tissues, trigger apoptotic processes. This leads to the exposure of PS and removal by macrophages. In particular, in thalassemia, erythropoiesis in bone marrow is marked by high levels of apoptosis or “ineffective erythropoiesis” [117, 118]. The presence of PS-exposing RBC is indicative of the inability to maintain proper membrane organization, related to the network of phospholipid maintenance mechanisms as indicated above. Several of these processes occur simultaneously to maintain the proper phospholipid organization. Dysfunction of one or several of these mechanisms is therefore linked and can be related to pathology observed in hemoglobinopathies. When RBCs from sickle-cell patients are analyzed, a large diversity in RBC density is observed, and sickle RBCs exhibit a wide range of PS externalization depending on the age and density of the cells [119], and density correlates with a decreased rate of PS movement [120]. The difference in density in subpopulations of sickle RBCs points at the inability of cells to maintain a proper ion balance across the membrane. Calcium–calmodulin was reported to be involved in the PS exposure in sickle RBCs [121]. Membrane-ion transporters will be affected by the lipid organization of the membrane bilayer, which in turn may be linked to the altered phospholipid molecular species organization in these density-separated fractions. Since all RBCs are exposed to the same fatty acid substrates, a difference in phospholipid composition suggests a different phospholipid turnover and in turn suggests alterations in the enzymes involved in the Lands pathway as described above. Together, the maintenance of composition and organization of the plasma membrane phospholipids are closely connected, and a loss of function in any of these components can be expected to lead to a loss of membrane viability and can be related to pathology in SCD. 15.6â•… CONSEQUENCES OF PS EXPOSURE The presence of PS-exposing RBCs in the circulation or during erythropoiesis has many pathological consequences. The rapid removal of PS-exposing erythroid cells (either in marrow or peripheral blood) plays a role in anemia, as macrophages recognize and remove these cells [111, 122], a process in part responsible for the reduced RBC lifespan in both sickle-cell patients and transgenic mice. This increased cell–cell interaction is not limited to RBC–white blood cell contacts. Interactions between (PS-exposing) RBCs and vascular endothelium results in endothelial dysfunction and is related to vaso-occlusive crisis [113]. Hydroxyurea treatment, used in a significant number of sickle-cell patients, appears to attenuate activated neutrophil-mediated PS exposure in sickle RBCs and adhesion to pulmonary vascular endothelium [123]. The exposure of PS is essential in platelet activation, but a docking surface for hemostatic factors is unwarranted in RBCs and leads to an imbalance in hemostasis [124]. Prothrombotic levels of coagulation factors in sickle-cell patients may be

330

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

related to PS exposure and play an important role in pathology [125, 126]. The ineffective removal of PS-exposing cells can have many other consequences for vascular function. Subpopulations of PS-exposing RBCs in SCD are exposed to secretory phospholipase A2, an inflammatory lipid mediator elevated in sickle-cell plasma [127]. Whereas this enzyme does not break down normal RBCs, it will hydrolyze phospholipids in PS-exposing RBCs [67, 128] and hemolyze the PS-exposing RBCs. The delayed hemolytic transfusion reaction in SCD patients can occur in the absence of detectable antibody and may be related to PS-exposing RBCs [129], and PS exposure could be triggered by plasma from these patients. The combination of elevated secretory phospholipase A2 and increasing amounts of PS-exposing RBCs could be clinically relevant under these conditions. Hemolysis has been correlated with a dysregulated arginine and nitric oxide metabolism, and vascular dysfunction [130–132]. In addition to hemolysis, phospholipid hydrolysis by phospholipase A2 will generate phospholipid breakdown products that will affect endothelial function and, in turn, may be related to acute chest syndrome [127]. In RBCs that have lost their ability to maintain phospholipid asymmetry and expose PS, phospholipase D is activated [133] to generate phosphatidic acid from PC. The relatively high levels of phosphatidic acid in PS-exposing cells will lead to the formation of lysophosphatidic acid (LPA), a powerful lipid mediator involved in a myriad of signal transduction pathways including vascular dysfunction [128]. 15.7â•… PHOSPHOLIPID TRANSBILAYER MOVEMENT IN HEMOGLOBINOPATHIES The exposure of PS on RBC subpopulations is well established, and the pathological consequences of these cells are well recognized [66], but the underlying cause remains unclear. Initiation of scrambling by loading RBCs with calcium leads ultimately to the exposure of PS, a process that is greatly accelerated when the cells are pretreated with sulfhydryl reagents such as NEM. On the other hand, NEM treatment or oxidant damage will inhibit flippase activity but does not lead to rapid PS exposure in vitro unless scrambling is also initiated [134]. This suggests an efficient competition between the inward movement of PS and calcium-induced phospholipid scrambling despite the rapid ATP decrease in the RBCs under these conditions and the inhibitory effect of calcium on the Mg-ATP-dependent aminophospholipid translocase. These data suggest that PS exposure becomes apparent in a short period of time only when both the inward movement is blocked and scrambling is initiated. The common denominator in PS-exposing sickle RBCs seems the inability to move fluorescent PS from outer to inner monolayer [134]. This apparent “slower” inward movement of fluorescently labeled PS can result from both a decreased inward movement of PS (lower activity of the aminophospholipd translocase) as well as a rapid scrambling (a return movement of PS from inner

Phospholipid Transbilayer Movement in Hemoglobinopathies

331

to outer monolayer). Measurements using intact RBCs report on the transbilayer movement and distribution at steady state of the entire system, and not on the activity of a particular protein. In other words, these measurements describe the ability of the cell to maintain phospholipid asymmetry and do not describe the activity of a specific protein in the red cell membrane. Fluorescently labeled PC is not transported by an ATP-dependent mechanism. Hence, a rapid redistribution of PC across the bilayer suggests increased phospholipid scrambling in PS-exposing cells. The measurement of movement of both fluorescently labeled PS and PC suggests that inward movement in sickle RBCs is inhibited and that this, together with an increase in scrambling, leads to PS exposure. As indicated earlier, the transport of oxygen by an iron-containing protein (hemoglobin) results in high levels of oxidant stress in all RBCs. Most of the ROS that are formed are neutralized by an efficient antioxidant system, which includes enzymes such as superoxide dismutase and catalase, cytosolic compounds such as glutathione, and peroxyredoxin as well as lipid antioxidants such as vitamin E. Glycolysis efficiently generates ATP in the RBCs and contributes toward the maintenance of the proper reducing environment by generating NADPH and fueling the activities of proteins such as glutathione reductase. Despite these antioxidant activities, both proteins and lipids are altered by oxidant damage during the life of the RBCs. The extremely high level of oxidant stress in sickle RBCs [135] or thalassemic RBCs [117, 118, 136] leads to higher levels of damage and premature aging of these cells. Footprints of increased oxidant damage such as methemoglobin, lipid peroxidation products, or altered protein thiol status are found in both thalassemic and sickle RBCs. In addition, RBCs are characterized by altered ion transport, resulting in a shift in density and alterations of cytosolic calcium [137–139]. In normal RBCs, thiol modification can alter transbilayer movement of aminophospholipid across the RBC bilayer [140–142], and similarly, movement is affected by oxidant stress. Normal or sickle murine RBCs that are temporarily depleted of ATP show a significant difference in their ability to restore PS asymmetry when ATP is restored, which is correlated with oxidant stress [140], suggesting that the flippase is irreversibly oxidatively damaged in subpopulations of sickle RBCs. The RBCs cannot replace oxidatively damaged membrane proteins, and these proteins will either be inactivated, be cleaved by proteases, or assume an altered function. Oxidant stress is random in nature, and many avenues may lead to an inhibition or alteration of the RBC flippase, ultimately leading to the inability of the RBCs to maintain its phospholipid asymmetry. The mechanisms by which sulfhydryl-reactive compounds and oxidant modifications act on the different isoforms of the ATP8A1 are poorly understood. The subpopulation of RBCs that exposes PS in SCD shows an inhibited transbilayer movement of PS [134], and the inhibition by NEM points at one or more cysteines vulnerable for oxidative damage, but other oxidative modifications cannot be excluded to play a role in the inhibition of the flippase. ATP8A1 has 21 cysteines, and based on sequence homology with

332

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

other P-ATPases, all of them are likely to be located in the RBC cytosol. It is currently not known which of these residues are involved in sulfhydryl bridges or are potentially vulnerable to modifications. In an intact RBC, NEM affects many different proteins, and a direct relation between ATPase activity transport and NEM modification is difficult to make. At this point, no direct proof is available on the oxidative modification of the flippase in sickle RBCs, in part due to the lack of detailed information on this protein, or the involvement (and dysregulation) of other components in the maintenance of phospholipid asymmetry. In a yeast vesicle system, the phosphorylation of the aspartate residue, essential to the phosphate cycle of P-ATPases, seems to be related to NEM modification of ATP8A1 [45]. While both PS and PE are actively transported by ATP-dependent processes, RBC storage studies have shown that PE transbilayer movement in RBCs can be compromised while PS is still actively transported [143]. These data suggest that inactivation of the flippase may not be an all-or-nothing process, and since two isoforms have been identified in RBCs [42], these forms may be differently affected. Sickle-cell patients are very diverse in their clinical pathology despite the fact that all have the same mutation in the beta locus of hemoglobin. Similarly, the number of PS-exposing RBCs in the circulation varies significantly. The highly increased erythropoiesis may affect the expression of the different isoforms of ATP8A1. Several efforts are under way to link the clinical data in this patient population to differences in genetic background. These studies may also shed some light on the expression of ATP8A1 or the relevance of single-nucleotide polymorphisms (SNPs) in the gene for ATP8A1. In sickle-cell patients, specific SNPs have been identified that are related to the expression of hemoglobin F [144– 147], in particular the repressor BCL11A [148–152], and/or to a positive response to hydroxyurea treatment [153]. It cannot be excluded that abnormal regulation of erythroid-specific genes observed in SCD including the modulation by treatment regimen aimed at gamma globin production could also affect the expression of the isoforms of ATP8A1 and affect flippase activity or specificity. 15.8â•… CONCLUSION In 1949, Linus Pauling identified sickle-cell anemia as the first molecular disease [154]. The single nucleotide change of the β-globin gene results in glutamate being substituted by valine at position 6. However, while the pathophysiology in SCD may be uniquely related to the polymerization of sickle hemoglobin under low oxygen conditions, it has become apparently clear that many factors are involved in the vasculopathy that characterizes this disorder that affects millions of individuals worldwide. The altered RBC membrane plays an important role in the dysfunctional interactions of the sickle RBCs with other blood cells and vascular endothelium, and leads to premature rec-

Abbreviations

333

ognition and removal, an imbalance in hemostasis, vaso-occlusive crisis, intravascular hemolysis, and may be involved in acute chest syndrome [127]. Similarly, in thalassemia, the imbalance in globin production has profound effects on the RBC membrane and plays a role in the pathology of this disease. In hemoglobinopathies, the complex, and well-orchestrated RBC membrane phospholipid organization is apparently lost in subpopulations of RBCs during erythropoiesis as well as in the circulation. Increased oxidant stress may play an important role in the inability of the RBCs to maintain phospholipid asymmetry, but the mechanisms that lead to PS exposure are poorly understood. This lack of knowledge is, in part, due to the incomplete characterization of the proteins involved in the maintenance of phospholipid asymmetry in the RBCs, as well as the complexity of studying a complete plasma membrane where several protein entities act in sync with each other and are governed by protein–protein and protein–lipid interactions. The purification and/or expression of the proteins that are thought to be involved in membrane organization in well-defined lipid bilayers, their three-dimensional structural modeling, and detailed functional characterization may lead to a better understanding on their individual functions as well as their interaction with other entities in the bilayer. This will also lead to a better understanding how their function is impaired or altered, leading to PS exposure in hemoglobinopathies, and define the molecular underpinnings for a complex pathophysiology.

ABBREVIATIONS CoA EGF LPA LPC LPE LPLAT LPS MRP NEM PC PE PIP2 PLSCR1 PS RBC SM

coenzyme A epidermal growth factor lysophosphatidic acid lysophosphatidylcholine lysophosphatidylethanolamine lysophospholipid acyl-CoA acyltransferase lysophosphatidylserine multidrug resistance protein N-ethylmaleimide phosphatidylcholine phosphatidylethanolamine phosphatidylinositol 4,5-bisphosphate phospholipid scramblase 1 phosphatidylserine red blood cell sphingomyelin

334

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

REFERENCES J. M. Old, Blood Rev. 2003, 17, 43–53. J. J. Myher, A. Kuksis, S. Pind, Lipids 1989, 24, 396–407. M. S. Bretscher, J. Mol. Biol. 1972, 71, 523–528. M. S. Bretscher, Nat. New Biol. 1972, 236, 11–12. R. F. Zwaal, B. Roelofsen, P. Comfurius, L. L. van Deenen, Biochim. Biophys. Acta 1975, 406, 83–96. â•… 6â•… C. W. Haest, B. Deuticke, Biochim. Biophys. Acta 1976, 436, 353–365. â•… 7â•… C. W. Haest, G. Plasa, D. Kamp, B. Deuticke, Biochim. Biophys. Acta 1978, 509, 21–32. â•… 8â•… C. W. Haest, J. Erusalimsky, V. Dressler, I. Kunze, B. Deuticke, Biomed. Biochim. Acta 1983, 42, S17–S21. â•… 9â•… V. Dressler, C. W. Haest, G. Plasa, B. Deuticke, J. D. Erusalimsky, Biochim. Biophys. Acta 1984, 775, 189–196. â•‡ 10â•… M. Seigneuret, P. F. Devaux, Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 3751–3755. â•‡ 11â•… D. L. Daleke, J. Biol. Chem. 2007, 282, 821–825. â•‡ 12â•… K. Balasubramanian, A. J. Schroit, Annu. Rev. Physiol. 2003, 65, 701–734. â•‡ 13â•…X. Tang, M. S. Halleck, R. A. Schlegel, P. Williamson, Science 1996, 272, 1495–1497. â•‡ 14â•… M. P. Hall, W. H. Huestis, Biochim. Biophys. Acta 1994, 1190, 243–247. â•‡ 15â•… P. F. Devaux, P. Fellmann, P. Herve, Chem. Phys. Lipids 2002, 116, 115–134. â•‡ 16â•… E. Middelkoop, B. H. Lubin, J. A. Op den Kamp, B. Roelofsen, Biochim. Biophys. Acta 1986, 855, 421–424. â•‡ 17â•… L. Tilley, S. Cribier, B. Roelofsen, J. A. Op den Kamp, L. L. van Deenen, FEBS Lett. 1986, 194, 21–27. â•‡ 18â•… E. Farge, D. M. Ojcius, A. Subtil, A. Dautry-Varsat, Am. J. Physiol. 1999, 276, C725–C733. â•‡ 19â•… F. A. Kuypers, B. Roelofsen, W. Berendsen, J. A. Op den Kamp, L. L. van Deenen, J. Cell Biol. 1984, 99, 2260–2267. â•‡ 20â•… D. W. Dekkers, P. Comfurius, A. J. Schroit, E. M. Bevers, R. F. Zwaal, Biochemistry 1998, 37, 14833–14837. â•‡ 21â•… D. W. Dekkers, P. Comfurius, R. G. van Gool, E. M. Bevers, R. F. Zwaal, Biochem. J. 2000, 350(Pt 2), 531–535. â•‡ 22â•… M. V. Serra, D. Kamp, C. W. Haest, Biochim. Biophys. Acta 1996, 1282, 263–273. â•‡ 23â•… J. J. M. van den Berg, J. A. F. Op den Kamp, B. H. Lubin, F. A. Kuypers, Biochemistry 1993, 32, 4962–4967. â•‡ 24â•… B. H. Lubin, F. A. Kuypers, Phospholipid Repair in Human Erythrocytes, Pergamon Press, New York, 1991. â•‡ 25â•… W. E. Lands, Annu. Rev. Biochem. 1965, 34, 313–346. â•‡ 26â•… H. Fyrst, J. Knudsen, B. H. Lubin, F. A. Kuypers, Biochem. J. 1995, 306, 793–799. â•‡ 27â•… K. Malhotra, K. Malhotra, B. Lubin, F. Kuypers, Biochem. J. 1999, 344, 135–141. â•‡ 28â•… E. Soupene, V. Serikov, F. A. Kuypers, J. Lipid Res. 2008, 49, 1103–1112. â•… â•… â•… â•… â•…

1â•… 2â•… 3â•… 4â•… 5â•…

References

335

â•‡ 29â•… E. Soupene, F. A. Kuypers, BMC Mol. Biol. 2006, 7, 21. â•‡ 30â•… E. Soupene, H. Fyrst, F. A. Kuypers, Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 88–93. â•‡ 31â•… C. Andrick, K. Broring, B. Deuticke, C. W. Haest, Biochim. Biophys. Acta 1991, 1064, 235–241. â•‡ 32â•… A. Zachowski, J. P. Henry, P. F. Devaux, Nature 1989, 340, 75–76. â•‡ 33â•… Y. Moriyama, N. Nelson, J. Biol. Chem. 1988, 263, 8521–8527. â•‡ 34â•… Y. Moriyama, N. Nelson, M. Maeda, M. Futai, Arch. Biochem. Biophys. 1991, 286, 252–256. â•‡ 35â•… J. Ding, Z. Wu, B. P. Crider, Y. Ma, X. Li, C. Slaughter, L. Gong, X.-S. Xie, J. Biol. Chem. 2000, 275, 23378–23386. â•‡ 36â•… J. K. Paterson, K. Renkema, L. Burden, M. S. Halleck, R. A. Schlegel, P. Williamson, D. L. Daleke, Biochemistry 2006, 45, 5367–5376. â•‡ 37â•… M. J. Fagan, M. H. Saier, Jr., J. Mol. Evol. 1994, 38, 57–99. â•‡ 38â•… M. S. Halleck, D. Pradhan, C. Blackman, C. Berkes, P. Williamson, R. A. Schlegel, Genome Res. 1998, 8, 354–361. â•‡ 39â•… K. B. Axelsen, M. G. Palmgren, Plant Physiol. 2001, 126, 696–706. â•‡ 40â•… M. J. Harris, D. G. L. Couteur, I. M. Arias, J. Gastroenterol. Hepatol. 2005, 20, 807–817. â•‡ 41â•… M. Meguro, A. Kashiwagi, K. Mitsuya, M. Nakao, I. Kondo, S. Saitoh, M. Oshimura, Nat. Genet. 2001, 28, 19–20. â•‡ 42â•… E. Soupene, F. A. Kuypers, Br. J. Haematol. 2006, 133, 436–438. â•‡ 43â•… E. M. Pasini, M. Kirkegaard, P. Mortensen, H. U. Lutz, A. W. Thomas, M. Mann, Blood 2006, 108, 791–801. â•‡ 44â•… E. M. Pasini, M. Kirkegaard, D. Salerno, P. Mortensen, M. Mann, A. W. Thomas, Mol. Cell Proteomics 2008, 7, 1317–1330. â•‡ 45â•… E. Soupene, D. Utami Kemaladewi, F. A. Kuypers, J. Recept. Ligand Channel Res. 2008, I, 1–10. â•‡ 46â•… C. Toyoshima, M. Nakasako, H. Nomura, H. Ogawa, Nature 2000, 405, 647–655. â•‡ 47â•… C. Olesen, M. Picard, A. M. Winther, C. Gyrup, J. P. Morth, C. Oxvig, J. V. Moller, P. Nissen, Nature 2007, 450, 1036–1042. â•‡ 48â•… B. P. Pedersen, M. J. Buch-Pedersen, J. P. Morth, M. G. Palmgren, P. Nissen, Nature 2007, 450, 1111–1114. â•‡ 49â•… J. P. Morth, B. P. Pedersen, M. S. Toustrup-Jensen, T. L. Sorensen, J. Petersen, J. P. Andersen, B. Vilsen, P. Nissen, Nature 2007, 450, 1043–1049. â•‡ 50â•… G. Lenoir, P. Williamson, J. C. Holthuis, Curr. Opin. Chem. Biol. 2007, 11, 654–661. â•‡ 51â•… K. Geering, J. Bioenerg. Biomembr. 2001, 33, 425–438. â•‡ 52â•… K. Geering, G. Crambert, C. Yu, T. V. Korneenko, N. B. Pestov, N. N. Modyanov, Biochemistry 2000, 39, 12688–12698. â•‡ 53â•… K. Saito, K. Fujimura-Kamada, N. Furuta, U. Kato, M. Umeda, K. Tanaka, Mol. Biol. Cell 2004, 15, 3418–3432. â•‡ 54â•… G. Lenoir, P. Williamson, C. F. Puts, J. C. Holthuis, J. Biol. Chem. 2009, 284, 17956–17967.

336

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

â•‡ 55â•… R. F. Zwaal, P. Comfurius, E. M. Bevers, Biochim. Biophys. Acta 1998, 1376, 433–453. â•‡ 56â•… P. Williamson, R. A. Schlegel, Biochim. Biophys. Acta 2002, 1585, 53–63. â•‡ 57â•… J. I. Elliott, A. Surprenant, F. M. Marelli-Berg, J. C. Cooper, R. L. Cassady-Cain, C. Wooding, K. Linton, D. R. Alexander, C. F. Higgins, Nat. Cell Biol. 2005, 7, 808–816. â•‡ 58â•… V. A. Fadok, D. L. Bratton, D. M. Rose, A. Pearson, R. A. Ezekewitz, P. M. Henson, Nature 2000, 405, 85–90. â•‡ 59â•… V. A. Fadok, A. de Cathelineau, D. L. Daleke, P. M. Henson, D. L. Bratton, J. Biol. Chem. 2001, 276, 1071–1077. â•‡ 60â•… J. Savill, V. Fadok, Nature 2000, 407, 784–788. â•‡ 61â•… U. S. Gaipl, R. E. Voll, A. Sheriff, S. Franz, J. R. Kalden, M. Herrmann, Autoimmun. Rev. 2005, 4, 189–194. â•‡ 62â•… W. M. Kuhtreiber, T. Hayashi, E. A. Dale, D. L. Faustman, J. Mol. Endocrinol. 2003, 31, 373–399. â•‡ 63â•… B. Mirnikjoo, K. Balasubramanian, A. J. Schroit, J. Biol. Chem. 2009, 284, 6918–6923. â•‡ 64â•… V. E. Kagan, Y. Y. Tyurina, H. Bayir, C. T. Chu, A. A. Kapralov, I. I. Vlasova, N. A. Belikova, V. A. Tyurin, A. Amoscato, M. Epperly, J. Greenberger, S. Dekosky, A. A. Shvedova, J. Jiang, Chem. Biol. Interact. 2006, 163, 15–28. â•‡ 65â•… F. Lang, E. Gulbins, H. Lerche, S. M. Huber, D. S. Kempe, M. Foller, Cell. Physiol. Biochem. 2008, 22, 373–380. â•‡ 66â•… F. A. Kuypers, Hematology Am. Soc. Hematol. Educ. Program 2007, 2007, 68–73. â•‡ 67â•… F. A. Kuypers, S. K. Larkin, J. J. Emeis, A. C. Allison, Thromb. Haemost. 2007, 97, 478–486. â•‡ 68â•… F. A. Kuypers, R. A. Lewis, J. D. Ernst, D. Discher, B. H. Lubin, Blood 1996, 87, 1179–1187. â•‡ 69â•… B. L. Wood, D. F. Gibson, J. F. Tait, Blood 1996, 88, 1873–1880. â•‡ 70â•… J. Classen, C. W. Haest, H. Tournois, B. Deuticke, Biochemistry 1987, 26, 6604–6612. â•‡ 71â•… C. W. Haest, J. Classen, Biomed. Biochim. Acta 1987, 46, S16–S20. â•‡ 72â•… C. Schwichtenhovel, B. Deuticke, C. W. Haest, Biochim. Biophys. Acta 1992, 1111, 35–44. â•‡ 73â•… A. Bootsveld, R. Degenhardt, D. Kamp, C. W. Haest, Mol. Membr. Biol. 2004, 21, 315–322. â•‡ 74â•… H. Tournois, J. Leunissen-Bijvelt, C. W. Haest, J. de Gier, B. de Kruijff, Biochemistry 1987, 26, 6613–6621. â•‡ 75â•… V. Dressler, K. Schwister, C. W. Haest, B. Deuticke, Biochim. Biophys. Acta 1983, 732, 304–307. â•‡ 76â•… C. W. Haest, A. Oslender, D. Kamp, Biochemistry 1997, 36, 10885–10891. â•‡ 77â•… U. Henseleit, G. Plasa, C. Haest, Biochim. Biophys. Acta 1990, 1029, 127–135. â•‡ 78â•… K. A. Shiffer, L. Rood, R. K. Emerson, F. A. Kuypers, Biochemistry 1998, 37, 3449–3458.

References

337

â•‡ 79â•… E. M. Bevers, T. Wiedmer, P. Comfurius, J. Zhao, E. F. Smeets, R. A. Schlegel, A. J. Schroit, H. J. Weiss, P. Williamson, R. F. Zwaal, P. J. Sims, Blood 1995, 86, 1983–1991. â•‡ 80â•… D. Kamp, C. W. Haest, Biochim. Biophys. Acta 1998, 1372, 91–101. â•‡ 81â•… A. Sohnius, D. Kamp, C. W. Haest, Mol. Membr. Biol. 2003, 20, 299–305. â•‡ 82â•… A. Vondenhof, A. Oslender, B. Deuticke, C. W. Haest, Biochemistry 1994, 33, 4517–4520. â•‡ 83â•… A. Kleinhorst, A. Oslender, C. W. Haest, B. Deuticke, J. Membr. Biol. 1998, 165, 111–124. â•‡ 84â•… Q. Zhou, J. Zhao, J. G. Stout, R. A. Luhm, T. Wiedmer, P. J. Sims, J. Biol. Chem. 1997, 272, 18240–18244. â•‡ 85â•… T. Wiedmer, Q. Zhou, D. Y. Kwoh, P. J. Sims, Biochim. Biophys. Acta 2000, 1467, 244–253. â•‡ 86â•… J. Zhao, Q. Zhou, T. Wiedmer, P. J. Sims, J. Biol. Chem. 1998, 273, 6603–6606. â•‡ 87â•… E. M. Bevers, P. Comfurius, D. W. Dekkers, R. F. Zwaal, Biochim. Biophys. Acta 1999, 1439, 317–330. â•‡ 88â•… P. F. Devaux, Annu. Rev. Biophys. Biomol. Struct. 1992, 21, 417–439. â•‡ 89â•… R. F. Zwaal, P. Comfurius, E. M. Bevers, Biochim. Biophys. Acta 2004, 1636, 119–128. â•‡ 90â•… P. Williamson, A. Christie, T. Kohlin, R. A. Schlegel, P. Comfurius, M. Harmsma, R. F. Zwaal, E. M. Bevers, Biochemistry 2001, 40, 8065–8072. â•‡ 91â•… J. G. Stout, F. Basse, R. A. Luhm, H. J. Weiss, T. Wiedmer, P. J. Sims, J. Clin. Invest. 1997, 99, 2232–2238. â•‡ 92â•… Q. Zhou, P. J. Sims, T. Wiedmer, Blood 1998, 92, 1707–1712. â•‡ 93â•… J. L. Wolfs, S. J. Wielders, P. Comfurius, T. Lindhout, J. C. Giddings, R. F. Zwaal, E. M. Bevers, Blood 2006, 108, 2223–2228. â•‡ 94â•… S. K. Sahu, S. N. Gummadi, N. Manoj, G. K. Aradhyam, Arch. Biochem. Biophys. 2007, 462, 103–114. â•‡ 95â•… P. J. Sims, T. Wiedmer, Thromb. Haemost. 2001, 86, 266–275. â•‡ 96â•… J. Sun, M. Nanjundan, L. J. Pike, T. Wiedmer, P. J. Sims, Biochemistry 2002, 41, 6338–6345. â•‡ 97â•… O. Amir-Moazami, C. Alexia, N. Charles, P. Launay, R. C. Monteiro, M. Benhamou, J. Biol. Chem. 2008, 283, 25514–25523. â•‡ 98â•… B. Py, S. Basmaciogullari, J. Bouchet, M. Zarka, I. C. Moura, M. Benhamou, R. C. Monteiro, H. Hocini, R. Madrid, S. Benichou, PLoS One 2009, 4, e5006. â•‡ 99â•… A. Bateman, R. D. Finn, P. J. Sims, T. Wiedmer, A. Biegert, J. Söding, Bioinformatics 2009, 25, 159–162. 100â•… J. Suzuki, M. Umeda, P. J. Sims, S. Nagata, Nature 2010, 468, 834–838. 101â•… A. Caputo, E. Caci, L. Ferrera, N. Pedemonte, C. Barsanti, E. Sondo, U. Pfeffer, R. Ravazzolo, O. Zegarra-Moran, L. J. Galietta, Science 2008, 322, 590–594. 102â•… B. C. Schroeder, T. Cheng, Y. N. Jan, L. Y. Jan, Cell 2008, 134, 1019–1029. 103â•… Y. D. Yang, H. Cho, J. Y. Koo, M. H. Tak, Y. Cho, W. S. Shim, S. P. Park, J. Lee, B. Lee, B. M. Kim, R. Raouf, Y. K. Shin, U. Oh, Nature 2008, 455, 1210–1215. 104â•… J. Shi, G. E. Gilbert, Blood 2003, 101, 2628–2636.

338

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

105â•… R. Hanayama, M. Tanaka, K. Miwa, S. Nagata, J. Immunol. 2004, 172, 3876–3882. 106â•… K. Emoto, N. Toyama-Sorimachi, H. Karasuyama, K. Inoue, M. Umeda, Exp. Cell Res. 1997, 232, 430–434. 107â•… H. J. Weiss, B. Lages, Blood 1997, 90, 475–476. 108â•… P. F. Franck, D. T. Chiu, J. A. Op den Kamp, B. Lubin, L. L. van Deenen, B. Roelofsen, J. Biol. Chem. 1983, 258, 8436–8442. 109â•… B. Lubin, D. Chiu, J. Bastacky, B. Roelofsen, L. L. Van Deenen, J. Clin. Invest. 1981, 67, 1643–1649. 110â•… P. F. Franck, E. M. Bevers, B. H. Lubin, P. Comfurius, D. T. Chiu, J. A. Op den Kamp, R. F. Zwaal, L. L. van Deenen, B. Roelofsen, J. Clin. Invest. 1985, 75, 183–190. 111â•… K. de Jong, R. Emerson, H. Butler, M. Narla, F. A. Kuypers, Blood 2001, 98, 1577–1584. 112â•… F. A. Kuypers, J. Yuan, R. A. Lewis, L. M. Snyder, C. R. Kiefer, A. Bunyaratvej, S. Fucharoen, L. Ma, L. Styles, K. de Jong, S. L. Schrier, Blood 1998, 91, 3044–3051. 113â•… B. N. Setty, S. Kulkarni, M. J. Stuart, Blood 2002, 99, 1564–1571. 114â•… P. A. Lane, Curr. Opin. Pediatr. 1995, 7, 36–41. 115â•… S. T. Singer, F. Kuypers, N. Olivieri, D. Weatherall, R. Mignacca, T. Coates, S. Davies, N. Sweeters, E. P. Vichinsky, Br. J. Haematol. 2005, 131, 378–388. 116â•… S. Basu, D. Banerjee, S. Chandra, A. Chakrabarti, Br. J. Haematol. 2008, 141, 92–99. 117â•… S. L. Schrier, Annu. Rev. Med. 1994, 45, 211–218. 118â•… S. L. Schrier, Curr. Opin. Hematol. 2002, 9, 123–126. 119â•…Z. Yasin, S. Witting, M. B. Palascak, C. H. Joiner, D. L. Rucknagel, R. S. Franco, Blood 2003, 102, 365–370. 120â•… N. Blumenfeld, A. Zachowski, F. Galacteros, Y. Beuzard, P. F. Devaux, Blood 1991, 77, 849–854. 121â•… R. L. Sabina, N. J. Wandersee, C. A. Hillery, Br. J. Haematol. 2009, 144, 434–445. 122â•… F. A. Kuypers, K. de Jong, Cell. Mol. Biol. (Noisy-le-grand) 2004, 50, 147–158. 123â•… J. Haynes, Jr., B. Obiako, R. B. Hester, B. S. Baliga, T. Stevens, Am. J. Physiol. Heart Circ. Physiol. 2008, 294, H379–H385. 124â•… R. F. Zwaal, P. Comfurius, E. M. Bevers, Cell. Mol. Life Sci. 2005, 62, 971–988. 125â•… M. O. Adedeji, J. Cespedes, K. Allen, C. Subramony, M. D. Hughson, Arch. Pathol. Lab. Med. 2001, 125, 1436–1441. 126â•… A. Marfaing-Koka, C. Boyer-Neumann, M. Wolf, C. Leroy-Matheron, T. Cynober, G. Tchernia, Nouv. Rev. Fr. Hematol. 1993, 35, 425–430. 127â•… F. A. Kuypers, L. A. Styles, Cell. Mol. Biol. (Noisy-le-grand) 2004, 50, 87–94. 128â•… N. A. Neidlinger, S. K. Larkin, A. Bhagat, G. P. Victorino, F. A. Kuypers, J. Biol. Chem. 2006, 281, 775–781. 129â•… P. Chadebech, A. Habibi, R. Nzouakou, D. Bachir, N. Meunier-Costes, P. Bonin, M. Rodet, B. Chami, F. Galacteros, P. Bierling, F. Noizat-Pirenne, Transfusion 2009, 49, 1785–1792.

References

339

130â•… C. R. Morris, F. A. Kuypers, G. J. Kato, L. Lavrisha, S. Larkin, T. Singer, E. P. Vichinsky, Ann. N.Y. Acad. Sci. 2005, 1054, 481–485. 131â•… M. T. Gladwin, G. J. Kato, Haematologica 2008, 93, 1–3. 132â•… G. J. Kato, M. T. Gladwin, M. H. Steinberg, Blood Rev. 2007, 21, 37–47. 133â•… P. Bütikofer, M. C. Yee, M. A. Schott, B. H. Lubin, F. A. Kuypers, Eur. J. Biochem. 1993, 213, 367–375. 134â•… K. de Jong, S. K. Larkin, L. A. Styles, R. M. Bookchin, F. A. Kuypers, Blood 2001, 98, 860–867. 135â•… R. P. Hebbel, Semin. Hematol. 1990, 27, 51–69. 136â•… E. Fibach, E. Rachmilewitz, Curr. Mol. Med. 2008, 8, 609–619. 137â•… T. Tiffert, N. Daw, Z. Etzion, R. M. Bookchin, V. L. Lew, J. Gen. Physiol. 2007, 129, 429–436. 138â•… R. M. Bookchin, Z. Etzion, M. Sorette, N. Mohandas, J. N. Skepper, V. L. Lew, Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 8045–8050. 139â•… V. L. Lew, Z. Etzion, R. M. Bookchin, Blood 2002, 99, 2578–2585. 140â•… T. Banerjee, F. A. Kuypers, Br. J. Haematol. 2004, 124, 391–402. 141â•… K. de Jong, F. A. Kuypers, Br. J. Haematol. 2006, 133, 427–432. 142â•… K. De Jong, D. Geldwerth, F. A. Kuypers, Biochemistry 1997, 36, 6768–6776. 143â•… D. Geldwerth, F. A. Kuypers, P. Bütikofer, M. Allary, B. H. Lubin, P. F. Devaux, J. Clin. Invest. 1993, 92, 308–314. 144â•…Z. Chen, H. Y. Luo, R. K. Basran, T. H. Hsu, D. W. Mang, L. Nuntakarn, C. G. Rosenfield, G. P. Patrinos, R. C. Hardison, M. H. Steinberg, D. H. Chui, Mol. Cell. Biol. 2008, 28, 4386–4393. 145â•… G. Lettre, V. G. Sankaran, M. A. Bezerra, A. S. Araujo, M. Uda, S. Sanna, A. Cao, D. Schlessinger, F. F. Costa, J. N. Hirschhorn, S. H. Orkin, Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 11869–11874. 146â•… P. Sebastiani, L. Wang, V. G. Nolan, E. Melista, Q. Ma, C. T. Baldwin, M. H. Steinberg, Am. J. Hematol. 2008, 83, 189–195. 147â•… M. Uda, R. Galanello, S. Sanna, G. Lettre, V. G. Sankaran, W. Chen, G. Usala, F. Busonero, A. Maschio, G. Albai, M. G. Piras, N. Sestu, S. Lai, M. Dei, A. Mulas, L. Crisponi, S. Naitza, I. Asunis, M. Deiana, R. Nagaraja, L. Perseu, S. Satta, M. D. Cipollina, C. Sollaino, P. Moi, J. N. Hirschhorn, S. H. Orkin, G. R. Abecasis, D. Schlessinger, A. Cao, Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 1620–1625. 148â•… S. Menzel, C. Garner, I. Gut, F. Matsuda, M. Yamaguchi, S. Heath, M. Foglio, D. Zelenika, A. Boland, H. Rooks, S. Best, T. D. Spector, M. Farrall, M. Lathrop, S. L. Thein, Nat. Genet. 2007, 39, 1197–1199. 149â•… Y. Saiki, Y. Yamazaki, M. Yoshida, O. Katoh, T. Nakamura, Genomics 2000, 70, 387–391. 150â•… V. G. Sankaran, T. F. Menne, J. Xu, T. E. Akie, G. Lettre, B. Van Handel, H. K. Mikkola, J. N. Hirschhorn, A. B. Cantor, S. H. Orkin, Science 2008, 322, 1839–1842. 151â•… A. E. Sedgewick, N. Timofeev, P. Sebastiani, J. C. So, E. S. Ma, L. C. Chan, G. Fucharoen, S. Fucharoen, C. G. Barbosa, B. N. Vardarajan, L. A. Farrer, C. T. Baldwin, M. H. Steinberg, D. H. Chui, Blood Cells Mol. Dis. 2008, 41, 255–258.

340

THE RED CELL MEMBRANE IN HEMOGLOBINOPATHIES

152â•… T. Senawong, V. J. Peterson, M. Leid, Arch. Biochem. Biophys. 2005, 434, 316–325. 153â•… Q. Ma, D. F. Wyszynski, J. J. Farrell, A. Kutlar, L. A. Farrer, C. T. Baldwin, M. H. Steinberg, Pharmacogenomics J. 2007, 7, 386–394. 154â•… L. Pauling, H. A. Itano, S. J. Singer, I. C. Wells, Science 1949, 110, 543–548. 155â•… G. Fallah, T. Roemer, S. Detro-Dassen, U. Braam, F. Markwardt, G. Schmalzing, Mol. Cell Proteomics 2011, 10, M110.004697-1–M110.004697-10.

16 SCOTT SYNDROME: MORE THAN A HEREDITARY DEFECT OF PLASMA MEMBRANE REMODELING Florence Toti and Jean-Marie Freyssinet U. 770 INSERM, Hôpital de Bicêtre, Le Kremlin-Bicêtre, France Faculté de Médecine, Université Paris-Sud, Le Kremlin-Bicêtre, France Institut d’Hématologie & Immunologie, Faculté de Médecine, Université de Strasbourg, Strasbourg, France

16.1â•… INTRODUCTION Like many other basic cellular constituents, phosphatidylserine (PS) is a highly conserved molecule. In the eukaryotic world, it is present in ancestral unicellular organisms, and the related phospholipid biosynthetic pathways have been retained in mammals [1]. Because of its particular membrane distribution, in the inner leaflet of the plasma membrane more specifically (see Chapter 3), and its mode of redistribution during cell activation or apoptosis, PS has gained several key functions during evolution. For instance, PS is externalized in apoptotic yeast [2] and becomes a determinant for phagocytic clearance of senescent cells in multicellular organisms [3, 4], indicating that phagocytes have learned to interpret such a signal. The same is probably true regarding the evolution of coagulation reactions [5] in which some clotting factors interact with PS made accessible at the external surface of the plasma membrane of activated cells and derived fragments, the so-called microparticles (MPs) or microvesicles (see Chapter 7). Despite the physiological importance of

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

341

342

SCOTT SYNDROME AND MEMBRANE REMODELING

externalized PS, the mechanisms governing its transmembrane migration remain obscure. In this context, Scott syndrome, an inherited hemorrhagic disorder characterized by a lack of PS exposure in stimulated hematopoietic cells [6, 7], offers a valuable opportunity for a better understanding of this form of membrane remodeling followed by the release of membrane MPs. Because Scott syndrome has been reviewed in detail rather recently, and no major breakthrough has been reported since [8], this chapter will focus more on the significance of PS and MPs that Scott syndrome can teach us in basic biological functions related to the plasma membrane remodeling other than blood coagulation addressed in Chapter 7. 16.2â•… SCOTT SYNDROME FEATURES AND PHENOTYPE Anionic (amino)phospholipids, chiefly PS, exert a key catalytic role by promoting the assembly of the vitamin K-dependent enzyme complexes of the coagulation cascade, first not only at the surface of stimulated platelets aggregated at the site of the wound, but also at the surface of other activated blood and vascular cells and shed MPs. After the loss of membrane phospholipid asymmetric distribution, vitamin K-dependent factors (VII, IX, X, and II) bind to accessible PS through their γ-carboxyglutamyl residues, together with activated cofactors VIIIa and Va. The concentration of enzymes, cofactors, and zymogens at the membrane surface enables fulfillment of the kinetics constraints for optimal thrombin generation that will in turn stimulate platelets to increase PS exposure in an amplification loop at the basis of the rapidity of the hemostatic response (see Chapter 7). In activated Scott syndrome hematopoietic cells, PS remains sequestered in the inner leaflet of the plasma membrane, even after drastic stimulation, by Ca2+ ionophores for instance. As a consequence, the degree of membrane vesiculation is considerably reduced with a dramatic decrease of MPs. Severe to moderate bleeding episodes, generally provoked and very rarely spontaneous, constitute the clinical phenotype, with no other apparent disorders in homozygous-like patients. The first case, giving her name to the syndrome, was reported in 1979 to be associated with an isolated deficiency of platelet procoagulant activity [9], and soon after observed to present a defect of factor Xa-factor Va platelet binding sites [10]. Low prothrombin consumption in the serum was the sole abnormal hemostasis parameter. This was confirmed a few years later through a major decrease of platelet membrane-dependent activation of factor X and prothrombin in purified systems [11] and through a defect of procoagulant MP shedding [12]. The anomaly was further evidenced in red blood cells [13] and lymphocytic cells [14], indicating that it occurs in multiple hematologic lineages. The possibility of an inherited disorder was discussed but not confirmed [10]. Fusion experiments between transformed B lymphocytes expressing the membrane defect of Scott syndrome and normal lymÂ� phoblasts led to correction of the Scott abnormality, suggesting possible

Cell Biology of Scott Syndrome

343

complementation of (an) aberrant gene(s) accounting for this disorder [14]. In 1996, we identified and characterized the first family where Scott syndrome was shown to be indeed transmitted as an autosomal recessive trait. Given her biological phenotype, the symptomatic patient, born from first cousins, was likely to be of homozygous status, while her asymptomatic children presented an intermediate biological phenotype, compatible with a heterozygous status [7]. More recently, another isolated case was reported as Scott syndrome after having been revisited [15] from an earlier study of familial hemostatic defects associated with reduced serum prothrombin consumption [16]. Finally, a hereditary bleeding disorder related to a deficiency of platelet procoagulant activity was identified in dogs with features of Scott syndrome [17]. As yet, no other case than the American, French, and Welsh patients has been firmly confirmed, making this relatively moderate hemorrhagic syndrome an actual clue regarding extreme rarity that precludes the identification of the inheritance pattern. From these documented cases, the diagnosis of Scott syndrome appears rather simple to perform in patients presenting bleeding episodes under provoked conditions. High residual prothrombin in the serum is the first mandatory biological anomaly [6]. Platelets stimulated by either the thrombinâ•¯+â•¯ collagen combination or the Ca2+ ionophores should develop minimal prothrombinase activity and exhibit considerably reduced membrane vesiculation, the latter being monitored by flow cytometry. Like erythrocytes, other blood cells should not expose PS after Ca2+ ionophore stimulation, as evidenced by prothrombinase assay, and should undergo minimum membrane vesiculation, again monitored by flow cytometry. In all cases, flow cytometry may confirm the defect of PS exposure and that of MP shedding when using Annexin V as a PS probe [6, 7, 13]. 16.3â•… CELL BIOLOGY OF SCOTT SYNDROME Owing to ethical considerations, investigations could only be performed in circulating cells from Scott patients’ peripheral blood, mainly platelets, erythrocytes, and lymphocytes. From the latter, B-lymphoblastic cell lines presenting the Scott membrane characteristics have been derived by infection by the Epstein–Barr virus (EBV) [7, 14, 18], which allowed the establishment of some links between the defect of PS egress and alterations of intracellular signaling. It has to be emphasized that in the three cases reviewed above, blood cells had normal phospholipid composition and showed normal ability to internalize plasma membrane aminophospholipids [8]. 16.3.1â•… Intracellular Signaling The most striking requirement for triggering PS transmembrane redistribution in activated platelets and other mammalian cells is a robust and sustained

344

SCOTT SYNDROME AND MEMBRANE REMODELING

cytosolic Ca2+ elevation after release from intracellular organelles that in turn signals extracellular Ca2+ entry. This mode of refilling of empty internal stores, referred to as store-operated Ca2+ entry (SOCE) [19], has been observed diminished in EBV B lymphoblasts derived from the French patient’s B cells [20], whereas it was apparently normal in counterpart cells from the American and Welsh patients [15, 18]. Hence, it could not be concluded that defective SOCE is part of the Scott syndrome features. The most likely explanation of this discrepancy is the heterogeneity of EBV transformation. In our experience, some B-lymphoblastic cell lines presented lower SOCE, so that if such cells have been used as controls, no noticeable difference could be evidenced. It also has to be kept in mind that Ca2+ signaling may depend on the cell type, with different links with associated phosphorylation pathways [21–23]. Tyrosine kinase-dependent phosphorylation alterations have also been reported in Scott platelets stimulated by physiological agonists [24] as well as distinctive Ca2+-induced shape change [25], and were even more pronounced in Scott erythrocytes challenged by Ca2+ ionophore (for a review, see Reference 8). Interestingly, nonphysiological amphiphilic membrane drugs such as tetracaine or propanolol promoted the collapse of membrane phospholipid asymmetry and MP shedding in Scott platelets [24], suggesting this form of plasma membrane remodeling can occur in response to other agonists, not specifically procoagulants. Subproteomic analysis of the EBV B lymphoblasts from the French patient revealed only minor differences of tyrosine phosphorylation [26]. Whatever the case, it appeared that the defect of tyrosine phosphorylation is most likely a consequence rather than a cause of deficiency of aminophospholipid transverse redistribution and does not account for Scott syndrome primary abnormality. As emphasized by Zwaal et al. [8], erythrocytes provide valuable and perhaps the most objective information on the relationship between Ca2+ and transmembrane lipid redistribution, gained without possible interference of other cellular functions such as in platelets or lymphocytes. Scott erythrocytes do not undergo membrane remodeling or vesiculation, even after drastic stimulation by Ca2+ ionophores, and resealed ghosts retain membrane asymmetry and behave as true Scott red blood cells [13]. These studies are suggestive of a direct interaction between Ca2+ and membrane protein(s) involved in the loss of membrane asymmetry. More recently, the manipulation of the Gardos channel in platelets was reported to produce a reversible inhibition of the PSdependent procoagulant response [27], and K+ ions were found to directly inhibit phospholipid scrambling in erythrocytes [28] (see also Chapter 7). Recent advances in the characterization of the TMEM16 transmembrane protein family, initially identified in silico as possible calcium activated chloride channels, brings new insights on other ion fluxes accompanying cell dehydration and possibly affecting plasma membrane remodelling. For instance the a-hemolysin from E. coli triggers erythrocyte shrinkage and PS egress through the dual activation of Gardos K+ channels and of the TMEM16A Cl− channel [28a] (see also Section 16.4).

CANDIDATE PROTEINS IN TRANSMEMBRANE REDISTRIBUTION OF PS

345

16.3.2â•… Apoptosis Because PS egress at the exoplasmic leaflet of the plasma membrane is a general characteristic of apoptotic cells [29, 30] (see Chapter 14), as well as MP release [31], and PS becomes a determinant for phagocytic clearance of senescent or activated cells [3, 4, 32], the question of the impact of the Scott defect on apoptosis is indeed obvious. In fact, Scott lymphocytic cells present normal [33] or even somewhat enhanced [34] susceptibility to apoptosis without impairment of PS externalization. This is consistent with the lack of clinical phenotype(s) other than bleeding in Scott syndrome patients; otherwise, one might have expected autoimmune disorders as a result of a lack of elimination of autoreactive B cells for instance. Although occurring on different timescales, PS contributed by apoptotic Scott cells and released MPs can also participate in procoagulant reactions [35, 36], which may partially compensate Scott platelet deficiency. Because it has recently been shown that apoptotic-like processes can lead to PS externalization in platelets, independently of physiological activation pathways [37], it would be of value to know whether this can also happen in human Scott platelets. In canine Scott syndrome, the autosomal recessive trait solely affects platelets pointing at their prime importance in PS egress and microparticle shedding. The defect was found downstream the formation of permeability transition pores, a key step in apoptosis and an integral event in PS egress. Only observed in a subset of platelets submitted to thrombin and convulxin, the physiological contribution of each pathway to the hemostatic response is still being questioned [37a].

16.4â•… CANDIDATE PROTEINS IN THE TRANSMEMBRANE REDISTRIBUTION OF PS Although the machinery governing the asymmetric distribution of the constitutive phospholipids of the plasma membrane remains partly elusive, it is certainly complex (see relevant Chapters 3 and 8–10). The observation that PS externalization occurs in apoptotic Scott cells suggests that at least two distinct pathways are involved in phospholipid translocation. In platelets, the coexistence of one pathway governed by a rapid Ca2+ influx, and the other under the dependence of apoptotic mitochondrial changes, also brings into question the number and fuction of translocators dedicated to PS egress in a single lineage [37, 37b]. When considering PS externalization, two groups of proteins have been more particularly investigated, the so-called phospholipid scramblases (PLSCRs), with four product members of the corresponding PLSCR gene family [38], and members of the ATP-binding cassette (ABC) transporters as possible floppases. The biological functions associated to these candidates are detailed in Chapter 7; none of them could be firmly established as a key element in the transmembrane migration of PS. For instance, PLSCR1, the most studied member of scramblases, is normally expressed in Scott cells

346

SCOTT SYNDROME AND MEMBRANE REMODELING

[39, 40], and PLSCR1(–/–) mice show normal hemostasis but defective hematopoietic response to growth factors [41]. In fact, PLSCR have been recently suggested to belong to a new superfamily of membrane-tethered transcription factors [42]. Concerning the ABC family of transporters, the P-glycoproteins encoded by the multidrug resistance genes (MDR1 and MDR2) were not expressed in the EBV B lymphoblasts from the French patient [43]. Despite the fact that MDR1 and MDR2 can act as phospholipid translocases [44], they were not confirmed as candidate floppases because no deletion or significant homozygosity of the patient’s MDR region on chromosome 7 (bands q21.1 and q21.2) could be seen [43], and because of normal expression of MDR1 and MDR2 genes in the counterpart cells from the American patient [39]. The knockout of a second member of this family, ABCA1, led to ∼75% reduction of PS exposure and consecutive MP release in mouse erythrocyte that showed impaired PS redistribution ability. In addition, Ca2+ ionophore-stimulated embryonic cells presented significantly decreased binding of Annexin V [45]. Although a missense mutation (ABCA1 R1925Q) was claimed to account in part for Scott syndrome in the British patient [46], ABCA1 has been firmly excluded as a candidate floppase defective in canine Scott syndrome [47]. Finally, in addition to the above groups of proteins, regulators of cytosolic Ca2+ fluxes should not be ignored. They play a key role in the maintenance of cytoskeleton integrity and reorganization during the loss of membrane asymmetry. Studies in apoptotic and stimulated Scott cells confirmed that PS egress is under the dependence of coordinate interactions between cytosolic proteases, ion channel regulators, and floppase-like phospholipid transporters [48]. Most recently, the TMEM16 transmembrane protein family attracted interest because some members were up-regulated in tumor and developmental deficiencies (see Section 16.3.1). Although their sequence bears characteristics of Ca2+activated Cl− channels, the 10 mammalian TMEM16 genes have multiple splice variants, raising the question of the actual function of these proteins, some of them being soluble [48a]. Most recently, TMEM16F, located in the chromosome 12 (12q12), was identified as a candidate floppase using a gain of function approach in apoptotic murine B cells [48b]. In two patients with Scott syndrome, mutations in TMEM16F were evidenced that predicted a premature truncated protein, devoid of its calcium sensing domain [48b, 48c]. Functional assessment of TMEM16F in human hematopoietic lineages and Scott cells should contribute to the elucidation of the molecular actors tuning scrambling pathways in haemostasis and apoptosis. 16.5â•… THE SIGNIFICANCE OF MEMBRANE VESICULATION AND OF DERIVED MPs Because MP release from the budding plasma membrane after cell stimulation or apotosis is consecutive to aminophospholipid exposure, MPs, more particu-

What Can Be Learned from Scott Syndrome?

347

larly those of platelet origin, could be viewed as a means to increase the catalytic procoagulant surface for optimal hemostatic response. The ability of other cells to shed MPs harboring PS can appear questionable when considering cell stimulation other than procoagulant. MPs may however constitute a broad primitive, beneficial, or noxious response to stimuli or stress of various nature [49]. MPs are plasma membrane fragments, highly heterogeneous in size, with dimensions in the range 0.1–1â•¯µm, and in composition depending on the stimulus and cells at their origin [50]. In these respects, MPs have to be clearly discriminated from exosomes, another category of membrane vesicles but of smaller size and more homogeneous composition, acting as conveyors of immune responses [51]. MPs are detectable in healthy subjects where they are thought to reflect a balance between cell proliferation, stimulation, and death, surviving longer than the activated cells they are stemming from, probably owing to their smaller size and therefore ability to diffuse more rapidly, transiently escaping phagocytosis [31]. There is increasing evidence that quantitative or qualitative variations of MPs are interpreted by proximal or distal cells, making them actors in intercellular exchanges of biological information, with true pathophysiological significance [49, 50, 52, 53]. In pathology, MPs appear as makers and conveyors in the nexus between inflammation, immunity, and thrombosis [53a]. Cardiovascular disorders constitute the most explored field [54–56], but cancer [57] and neurological diseases [58] are also actively investigated, among others. In addition to a part in intercellular communication and blood coagulation reactions, membrane remodeling and MP shedding could have a clearing function by allowing stimulated cells to expulse toxic metabolites or noxious modified cellular components [50], and to participate in the secretion of mediators devoid of signal sequence such as IL-1β [59]. As already stated above, Scott syndrome cells do not undergo plasma membrane vesiculation when stimulated by platelet agonists for platelets or Ca2+ ionophores for platelets, erythrocytes, and B-lymphocytic cells. Does this imply that Scott patients are devoid of circulating MPs? Although this has not been systematically assessed, from our experience, the answer is no. This is in compliance with the absence of evidence of abnormal apoptosis and, more generally, abnormal cellular turnover.

16.6â•… WHAT CAN BE LEARNED FROM SCOTT SYNDROME? Because the Scott phenotype is borne by platelets, erythrocytes, and B cells [7, 13, 14], the primary defect can probably be traced back at least to the pluripotent stem cells at the origin of myeloid and lymphoid lineages. Unfortunately, ethical considerations have precluded investigation of other cell types, from solid tissues more particularly.

348

SCOTT SYNDROME AND MEMBRANE REMODELING

16.6.1â•… Developmental Considerations Given its high degree of conservation throughout phylogeny and particular mode of expression at the cell surface and that of derived MPs, PS could be considered a potential membrane-anchored ligand, provided (counter)receptors exist. Indeed, several of them have been proposed to mediate engulfment of apoptotic cells and corpses [60, 61], even though they have not been fully qualified as such, probably owing to a lack of robust specific functional assays [62]. In metazoan development, phagocytosis is a fundamental process [4] and the ubiquitous character of PS can explain that such receptors have been independently selected during evolution in order that cell lineages having to fulfill a clearance function possess more than one kind. Hence, when totipotent stem cells are affected by a lack of PS exposure, it is reasonable to assume that at this early stage of embryonic development, the Scott syndrome defect may well account for lethality. This would be in accordance with the extremely low incidence of this paradoxically moderate syndrome, survival being possibly due to rescue by (an)other opportune mutation(s). If this hypothesis is relevant, a more complex genetic pattern than a single gene mutation can be expected, making essential the availability of larger families for the identification of mutated genes. 16.6.2â•… Pharmacological Impact Given its ubiquitous character and basic functions, targeting PS may be counterproductive and, in fact, life threatening. However, the absence of any bleeding symptom in heterozygous-type Scott subjects [7] provides a strong indication of the possible safety of a treatment that could achieve a reduction of the degree of procoagulant PS exposure and MP shedding of at least 50% in stimulated platelets. Hence, if the knowledge of the mutated element(s) at the origin of Scott syndrome could allow manipulation of the levels of PS exposure and procoagulant MP shedding specifically in stimulated platelets, this would allow the assessment of whether a 50% reduction, or even more, is protective against thrombosis without associated bleeding risk. This would indeed represent a new pharmacological approach to the thrombotic risk. The control of MP release could also be protective against cerebral malaria and probably associated inflammation, since MPs have been shown to be pathogenic in this neurological syndrome [63] and platelet MPs mediate parasite cytoadherence to human brain endothelium [64]. 16.7â•… CONCLUSION Understanding the molecular basis of Scott syndrome remains a true challenge, with an impact far beyond coagulation reactions, since it could yield information on as yet unraveled PS-associated signaling pathways of basic

349

Abbreviations (a)

? Totipotent cells

1 Scott mutation

+

Rescue

2

(<<0.1% of cases)

+

Myeloid 4

+

?

+

Death when no rescue (>>99.9%

3

+

+

of cases)

Pluripotent cells

+

Differentiated cells

Lymphoid 4

+

(b) Normal platelets, pharmacological targets for modulation of procoagulant phosphatidylserine exposure, and microparticle shedding

Figure 16.1.â•‡ Hypothesis on how Scott syndrome can be so extremely rare, and indications on pharmacological targets. (a) It is not known whether the Scott mutation is already present in totipotent cells, but it could occur during symmetric stem cell division (1), and be transmitted at the stages of asymmetric stem cell (2) and progenitor (3) division, and terminal differentiation (4). In the absence of rescue mutation, for example, in more than 99.9% of the cases, the Scott mutation is lethal owing to the absence of phosphatidylserine externalization and associated signaling. (b) Despite extreme rarity, Scott syndrome, however, indicates that normal platelets, part of the ultimate products of the myeloid differentiation cascade, could be considered as pharmacological targets with respect to their potential phosphatidylserine-dependent procoagulant activity and ability to release procoagulant membrane microparticles. Color version on the Wiley web site.

significance in metazoan development. Figure 16.1 summarizes the hypothesis explaining how Scott syndrome can be so rare. ABBREVIATIONS ABC EBV MDR MPs PLSCR PS SOCE

ATP-binding cassette Eptein–Barr virus multidrug resistance microparticles phospholipid scramblase phosphatidylserine store-operated Ca2+ channels

350

SCOTT SYNDROME AND MEMBRANE REMODELING

REFERENCES â•‡ 1â•… A. Lykidis, Prog. Lipid Res. 2007, 46, 171–199. â•‡ 2â•… F. Madeo, E. Frohlich, K. U. Frohlich, J. Cell Biol. 1997, 139, 729–734. â•‡ 3â•… V. A. Fadok, D. R. Voelker, P. A. Campbell, J. J. Cohen, et al., J. Immunol. 1992, 148, 2207–2216. â•‡ 4â•… R. B. Birge, D. S. Ucker, Cell Death Differ. 2008, 15, 1096–1102. â•‡ 5â•… C. J. Davidson, R. P. Hirt, K. Lal, P. Snell, et al., Thromb. Haemost. 2003, 89, 420–428. â•‡ 6â•… H. J. Weiss, Semin. Hematol. 1994, 31, 312–319. â•‡ 7â•… F. Toti, N. Satta, E. Fressinaud, D. Meyer, et al., Blood 1996, 87, 1409–1415. â•‡ 8â•… R. F. Zwaal, P. Comfurius, E. M. Bevers, Biochim. Biophys. Acta 2004, 1636, 119–128. â•‡ 9â•… H. J. Weiss, W. J. Vicic, B. A. Lages, J. Rogers, Am. J. Med. 1979, 67, 206–213. 10â•… J. P. Miletich, W. H. Kane, S. L. Hofmann, N. Stanford, et al., Blood 1979, 54, 1015–1022. 11â•… J. Rosing, E. M. Bevers, P. Comfurius, H. C. Hemker, et al., Blood 1985, 65, 1557–1561. 12â•… P. J. Sims, T. Wiedmer, C. T. Esmon, H. J. Weiss, et al., J. Biol. Chem. 1989, 264, 17049–17057. 13â•… E. M. Bevers, T. Wiedmer, P. Comfurius, S. J. Shattil, et al., Blood 1992, 79, 380–388. 14â•… H. Kojima, D. Newton-Nash, H. J. Weiss, J. Zhao, et al., J. Clin. Invest. 1994, 94, 2237–2244. 15â•… I. C. Munnix, M. Harmsma, J. C. Giddings, P. W. Collins, et al., Thromb. Haemost. 2003, 89, 687–695. 16â•… D. H. Parry, J. C. Giddings, A. L. Bloom, Br. J. Haematol. 1980, 44, 323–334. 17â•… M. B. Brooks, J. L. Catalfamo, H. A. Brown, P. Ivanova, et al., Blood 2002, 99, 2434–2441. 18â•… J. I. Elliott, A. D. Mumford, C. Albrecht, P. W. Collins, et al., Thromb. Haemost. 2004, 91, 412–415. 19â•… A. B. Parekh, J. W. Putney, Jr., Physiol. Rev. 2005, 85, 757–810. 20â•… M. C. Martinez, S. Martin, F. Toti, E. Fressinaud, et al., Biochemistry 1999, 38, 10092–10098. 21â•… C. Kunzelmann-Marche, J. M. Freyssinet, M. C. Martinez, J. Biol. Chem. 2002, 277, 19876–19881. 22â•… A. Arachiche, I. Badirou, J. Dachary-Prigent, I. Garcin, et al., Cell. Mol. Life Sci. 2008, 65, 3861–3871. 23â•… A. Arachiche, D. Kerbiriou-Nabias, I. Garcin, T. Letellier, et al., Arterioscler. Thromb. Vasc. Biol. 2009, 29, 1883–1889. 24â•… J. Dachary-Prigent, J. M. Pasquet, E. Fressinaud, F. Toti, et al., Br. J. Haematol. 1997, 99, 959–967. 25â•… N. Bettache, P. Gaffet, N. Allegre, L. Maurin, et al., Br. J. Haematol. 1998, 101, 50–58.

References

351

26â•… N. Imam-Sghiouar, I. Laude-Lemaire, V. Labas, D. Pflieger, et al., Proteomics 2002, 2, 828–838. 27â•… J. L. Wolfs, S. J. Wielders, P. Comfurius, T. Lindhout, et al., Blood 2006, 108, 2223–2228. 28â•… J. L. Wolfs, P. Comfurius, O. Bekers, R. F. Zwaal, et al., Cell. Mol. Life Sci. 2009, 66, 314–323. 28aâ•… M. Skals, N. R. Jorgensen, J. Leipziger, H. A. Praetorius, Proc. Natl. Acad. Sci. U S A. 2009, 106, 4030–4035. 29â•… S. J. Martin, C. P. Reutelingsperger, A. J. McGahon, J. A. Rader, et al., J. Exp. Med. 1995, 182, 1545–1556. 30â•… B. Verhoven, R. A. Schlegel, P. Williamson, J. Exp. Med. 1995, 182, 1597–1601. 31â•… K. Aupeix, B. Hugel, T. Martin, P. Bischoff, et al., J. Clin. Invest. 1997, 99, 1546–1554. 32â•… R. A. Schlegel, P. Williamson, Cell Death Differ. 2001, 8, 551–563. 33â•… P. Williamson, A. Christie, T. Kohlin, R. A. Schlegel, et al., Biochemistry 2001, 40, 8065–8072. 34â•… M. C. Martinez, J. M. Freyssinet, BMC Cell Biol. 2001, 2, 20. 35â•… K. Aupeix, F. Toti, N. Satta, P. Bischoff, et al., Biochem. J. 1996, 314(Pt 3), 1027–1033. 36â•… W. Pickering, E. Gray, A. H. Goodall, S. Ran, et al., J. Thromb. Haemost. 2004, 2, 459–467. 37â•… S. M. Schoenwaelder, Y. Yuan, E. C. Josefsson, M. J. White, et al., Blood 2009, 114, 663–666. 37aâ•… M. B. Brooks, J. L. Catalfamo, P. Friese, G. L. Dale, J. Thromb. Haemost. 2007, 5, 1972–1974. 37bâ•… E. M. Bevers, P. L. Williamson, FEBS Lett. 2010, 584, 2724–2730. 38â•… T. Wiedmer, Q. Zhou, D. Y. Kwoh, P. J. Sims, Biochim. Biophys. Acta 2000, 1467, 244–253. 39â•… Q. Zhou, P. J. Sims, T. Wiedmer, Blood 1998, 92, 1707–1712. 40â•… N. Janel, C. Leroy, I. Laude, F. Toti, et al., Thromb. Haemost. 1999, 81, 322–323. 41â•… Q. Zhou, J. Zhao, T. Wiedmer, P. J. Sims, Blood 2002, 99, 4030–4038. 42â•… A. Bateman, R. D. Finn, P. J. Sims, T. Wiedmer, et al., Bioinformatics 2009, 25, 159–162. 43â•… F. Toti, V. Schindler, J. F. Riou, G. Lombard-Platet, et al., Biochem. Biophys. Res. Commun. 1997, 241, 548–552. 44â•… A. van Helvoort, A. J. Smith, H. Sprong, I. Fritzsche, et al., Cell 1996, 87, 507–517. 45â•… Y. Hamon, C. Broccardo, O. Chambenoit, M. F. Luciani, et al., Nat. Cell Biol. 2000, 2, 399–406. 46â•… C. Albrecht, J. H. McVey, J. I. Elliott, A. Sardini, et al., Blood 2005, 106, 542–549. 47â•… M. B. Brooks, J. L. Catalfamo, K. Etter, A. Brisbin, et al., J. Thromb. Haemost. 2008, 6, 1608–1610. 48â•… O. Morel, N. Morel, J. M. Freyssinet, F. Toti, Platelets 2008, 19, 9–23. 48aâ•… H. C. Hartzell, K. Yu, O. Xiao, L. T. Chien, Z. Qu, J. Physiol. 2009, 587(Pt10), 2127–2139.

352

SCOTT SYNDROME AND MEMBRANE REMODELING

48bâ•… J. Suzuki, M. Umeda, P. J. Sims, S. Nagata, Nature 2010, 468, 834–838. 48câ•… E. Castoldi, P. W. Collins, P. L. Williamson, E. M. Bevers, Blood 2011, 117, 4399–4400. 49â•… J. M. Freyssinet, J. Thromb. Haemost. 2003, 1, 1655–1662. 50â•… B. Hugel, M. C. Martinez, C. Kunzelmann, J. M. Freyssinet, Physiology (Bethesda) 2005, 20, 22–27. 51â•… C. Thery, M. Ostrowski, E. Segura, Nat. Rev. Immunol. 2009, 9, 581–593. 52â•… O. Morel, F. Toti, B. Hugel, J. M. Freyssinet, Curr. Opin. Hematol. 2004, 11, 156–164. 53â•… O. Morel, F. Toti, N. Morel, J. M. Freyssinet, Haematologica 2009, 94, 313–317. 53aâ•… O. Morel, N. Morel, L. Jesel, J.-M. Freyssinet, F. Toti, Semin. Immunopathol. DOI 10.1007/s00281-010-0239-3. 54â•… O. Morel, F. Toti, B. Hugel, B. Bakouboula, et al., Arterioscler. Thromb. Vasc. Biol. 2006, 26, 2594–2604. 55â•… M. J. VanWijk, E. VanBavel, A. Sturk, R. Nieuwland, Cardiovasc. Res. 2003, 59, 277–287. 56â•… A. S. Leroyer, A. Tedgui, C. M. Boulanger, J. Intern. Med. 2008, 263, 528–537. 57â•… D. Castellana, C. Kunzelmann, J. M. Freyssinet, Hamostaseologie 2009, 29, 51–57. 58â•… L. Doeuvre, L. Plawinski, F. Toti, E. Angles-Cano, J. Neurochem. 2009, 110, 457–468. 59â•… A. MacKenzie, H. L. Wilson, E. Kiss-Toth, S. K. Dower, et al., Immunity 2001, 15, 825–835. 60â•… Z. Zhou, Dev. Cell 2007, 13, 759–760. 61â•… D. L. Bratton, P. M. Henson, Curr. Biol. 2008, 18, R76–R79. 62â•… R. A. Schlegel, P. Williamson, Sci. STKE 2007, 2007, pe57. 63â•… V. Combes, N. Coltel, M. Alibert, M. van Eck, et al., Am. J. Pathol. 2005, 166, 295–302. 64â•… D. Faille, V. Combes, A. J. Mitchell, A. Fontaine, et al., FASEB J. 2009, 23, 3449–3458.

17 ABCA1, TANGIER DISEASE, AND LIPID FLOPPING Ana Zarubica and Giovanna Chimini Centre d’Immunologie de Marseille-Luminy, Parc Scientifique de Luminy, INSERMCNRS-Université de La Méditerranée, Marseille, France

17.1â•… HISTORICAL NOTES: TANGIER DISEASE (TD) AND ATP-BINDING CASSETTE TRANSPORTER 1 (ABCA1) In the early 1960s, an unusual disease was reported in a 5-year-old boy living on Tangier Island in Chesapeake Bay in the United States. This patient had large lemon-colored tonsils in which foam cells laden with cholesteryl esters (CEs) had accumulated. The patient also displayed an extremely low plasma level of high-density lipoprotein cholesterol (HDL-C). The tonsils of his sister had a similar appearance and content in CEs. This novel syndrome was thus named Tangier disease (TD). Nowadays, approximately 60 cases have been diagnosed worldwide [1]. TD is defined as a genetic disorder mainly characterized by abnormal plasma lipoprotein profiles and a variety of distinct clinical manifestations. These include enlarged tonsils, splenomegaly, hepatomegaly, thrombocytopenia, ocular abnormalities, peripheral neuropathy, and atherosclerosis [2]. Biochemical hallmarks of the dislipidemia are the virtual absence of plasma HDL-C and apolipoprotein A-I (Apo A-I), and low level of plasma low-density lipoprotein cholesterol (LDL-C) (approximately 50% of normal) with a moderate elevation in triglycerides [3]. In addition, there is a pronounced accumulation of CEs in the reticuloendothelial cells of different tissues, such as tonsils, thymus, lymph node, bone marrow, spleen, liver, gall Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

353

354

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

bladder, and intestinal mucosa. Lipid deposits can also be found in neuronal Schwann cells, smooth muscle cells, and fibroblasts [4]. Numerous epidemiological and clinical studies demonstrated that an inverse and dependent association between the level of HDL and incidence of cardiovascular diseases exists [5]. More than 40% of patients with myocardial infarction have low HDL-C [6]. In line with this knowledge, cardiovascular disease is observed in 20% of Tangier patients compared with 5% in control, and in the age group between 35 and 65 years, the incidence of cardiovascular disease is 44% among Tangier patients versus less than 10% in the matched control group [7]. Twenty years after the clinical definition of TD, it was shown that the disease originated from impaired cellular ability to efflux phospholipids and cholesterol to apolipoprotein. Indeed, nascent Apo A-I in Tangier patients fails to acquire lipids and, being unable to mature into spherical lipid-rich HDL particles, is rapidly removed from the plasma. As a result, Tangier patients display very low plasma levels of both HDL and Apo A-I. The rapid degradation of Apo A-I is unrelated to intrinsic defects in Apo A-I metabolism since its structure and its synthetic rate are perfectly normal [8]. The clinical sign of low levels of circulating Apo A-I in Tangier patients is, thus, determined by the original failure of lipid transport outside the cell. A historical breakthrough took place in 1999. Indeed, the simultaneous reports by three independent groups identified mutations in the ABCA1 gene in Tangier patients [9–12]. The ABCA1 transporter had been previously cloned by Luciani et al. in 1994 [13], but its function as a transporter was still undefined. The evidence that ABCA1 loss of function was sufficient to give rise to TD was further corroborated by the study of animal models of genetic invalidation of the transporter [14]. This finding raised the possibility that ABCA1 may act as the receptor for Apo A-I in line with the idea of a specific molecularmediated interaction with cell membranes [15]. Though it is now clear that Apo A-I docking to cell membranes is ABCA1 dependent, the details of interaction between the two proteins are still controversial. Nowadays, it is established that ABCA1 controls the loading of cellular phospholipids and secondarily of cholesterol onto Apo A-I. This event is rate limiting in the formation of HDL particles and initiates the whole reverse cholesterol transport pathway (Fig. 17.1) [16]. 17.2â•… THE ABCA1 GENE AND THE REGULATION OF ITS EXPRESSION The ABCA1 gene maps to human and mouse chromosome at positions 9q31.1 and 4B2 (23.1â•¯cm), respectively, and its complete gene sequence and organization including promoter and regulatory elements has been reported (NCBI reference sequence accession numbers: NM_013454.3; GI: 90568037; [17, 18]). The genomic organization of ABCA1 gene is similar to that of the ABCA4

The ABCA1 Gene and the Regulation of Its Expression Macrophages

Cholesterol efflux

ABCA1 CE

X

355

Tangier disease Apo A-I

FC PL Pre-b HDL Cholesterol pool

1 Blood

LDLR Pre-b-HDL

CE, FC

2 LCAT

HL

Bile

Mature a-HDL

SR-BI

5

Liver

3

4 LDL Macrophages

VLDL

Blood

Figure 17.1.â•‡ Schematic model of the reverse cholesterol transport pathway (adapted from References 16 and 135). Reverse cholesterol transport controls the removal of excess free cholesterol (FC) from peripheral tissues and its return to the liver for excretion in the bile. Five steps are involved in the reverse cholesterol transport pathway. (1) Uptake of cholesterol from cells by specific acceptors (cholesterol efflux). Nascent pre-β HDL, secreted by the intestine or liver, is a potent acceptor of effluxed cholesterol from peripheral tissues. The ABCA1 protein is crucial for this initial step, since it controls the lipidation of apolipoprotein A-I (Apo A-I), the major apolipoprotein of HDL. The process is impaired in Tangier disease. (2) Esterification of cholesterol within pre-β HDL by lecithin–cholesterol acyltransferase (LCAT). Upon esterification, cholesterol progresses toward the center of the HDL particle, and nascent HDL particles are converted into mature α-HDL. Hepatic lipase (HL) can convert mature α-HDL into nascent pre-β HDL. (3) Remodeling of HDL. Cholesteryl esters (CEs) from HDL are transferred onto very low-density lipoproteins (VLDLs) or other lipoproproteins (like low-density lipoproteins [LDLs] not shown) by phospholipase transfer protein (PLTP). (4) These particles are subsequently taken up by the liver or, in case of LDL, by macrophage LDL receptor (LDLR). (5) Uptake of HDL cholesterol by the liver (cholesterol uptake). CEs in the HDL particles are taken up by the liver via the scavenger receptor of class B1 (SR-BI). SR-BI performs the uptake of cholesterol ester without internalization or degradation of the HDL particle. PL, phospholipid. Color version on the Wiley web site.

356

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

(ABCR) transporter, a member of ABCA family encoding a rod photoreceptorspecific membrane protein [19]. All these genes contain 50 exons interrupted by 49 introns. The analysis of the mouse and human ABCA1 promoters highlighted 90% sequence similarity over the whole genomic region and multiple motifs that are strongly conserved between the two species, suggesting important biological functions. These include potential binding sites for transcription factors (i.e., specificity protein 1 [SP1], hepatocyte nuclear factor 1 [HNF-3b], activator protein 1 [AP-1], nuclear factor kappa-light-chain-enhancer of activated B cells [NF-kBs]) also present in the promoters of molecules involved in lipid metabolism, such as the low-density lipoprotein (LDL) receptor, the LDL-receptorrelated protein, cluster of differentiation 36 (CD36), scavenger receptor class B member 1 (SR-BI), and the scavenger receptor A (SR-A) [20–22]. Moreover, the ABCA1 gene promoter contains several consensus binding sequences for transcription factors driving monocyte/macrophage differentiation and for the liver-enriched transcription factor HNF-3b, which activates genes important for liver development and function as well as lipid metabolism. The identification of ABCA1 as a key player in cellular cholesterol efflux has prompted intensive investigation on the control of its expression both at the transcriptional and posttranscriptional levels. Initial studies established the existence of a highly complex regulatory network, including secondary messengers like cyclic adenosine monophosphate (cAMP); the nuclear orphan receptors liver X receptor (LXR), peroxisome proliferator-activated receptor (PPAR), and retinoid X receptor (RXR) [23]; and cytokines such as interferon gamma (IFN-γ) [24] or oncostatin M [25]. A thorough overview of this regulation has been reported in Reference 26. The treatment of mouse macrophage cell lines (RAW 264.7 and J774) with cAMP analogs causes a huge increase in both ABCA1 messenger RNA (mRNA) and protein levels (estimated at 50- to 70-fold) [15]. The identification of a cAMP-responsive element has been reported recently. Its absence in the human ABCA1 gene explains the lack of the cAMP stimulation on the human ABCA1 gene [27]. Recently, it has been reported that ABCA1 mRNA and protein decline after hepatic expression of microRNA-33 (miR-33) encoded by sterol regulatory element-binding protein SREBP-2, whereas they increase after hepatic miR-33 silencing [28–30]. Of particular importance is the regulation of ABCA1 expression as a lipid transporter by LXRs. These are crucial actors in modulating intracellular cholesterol metabolism, displaying anti-inflammatory activities, and promoting macrophage survival in bacterial infection settings. Cytokines such as tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), IFN-γ, and transforming growth factor-β (TGF-β) also exert effects on the ABCA1 transcription [24, 31, 32]. 17.3â•… THE ABCA1 PROTEIN AND ITS INTERACTIONS ABCA1 is a large polytopic membrane protein consisting of 2261 amino acids (NCBI reference sequence accession numbers: NM_013454.3; GI: 90568037;

357

The ABCA1 Protein and Its Interactions ECL-2

ECL-1

out PM

T M D

-

1

T M D

-

A er

er A

2

in NH2

lk

Wa

COOH

lk

Wa

NBD-1 Walker

B

NBD-2

Walker B

Figure 17.2.â•‡ Topological model of the structure of an ABCA1 monomer. Monomeric ABCA1 is composed of two hydrophobic transmembrane domains (TMD-1 and TMD2) each containing six α-helices (shown in blue) and two nucleotide-binding domains (NBD-1 and NBD-2). They are arranged in tandem succession. Each NBD is characterized by the presence of the Walker A and B motifs (shown in pink) and the ABC signature consensus sequence (shown in green). ATP binds to NBDs to provide the energy required to transport substrates across the membrane. ABCA1 also contains two large extracellular loops (ECL-1 and ECL-2) with multiple sites for N-linked glycosylation (shown in red). The intracellular positions of the NH2 (N) and COOH (C) terminus are indicated as well as the membrane orientation (in and out) and plasma membrane (PM) localization. Color version on the Wiley web site.

[17, 18]). The membrane topology and domain organization of ABCA1 have been proposed to be similar to that of ABCA4 transporter based on sequence similarities. ABCA1 is a full-length ABC transporter consisting of a single polypeptide chain arranged in two symmetrical halves each bearing a transmembrane domain (TMD) composed of six predicted α-helices, and followed by a nucleotide-binding domain (NBD) (Fig. 17.2). A structural particularity of ABCA family members is the presence of two large extracellular loops (ECLs) located between TMD-1 and TMD-2 and between TMD-7 and TMD8. Each ECL contains multiple potential glycosylation sites and may be covalently bound by disulfide bonds [33]. Intriguingly, a missense mutation in one of the conserved cysteine residues in ECL-2 (C1477R) causes TD. This suggests its participation in disulfide bonds crucial for protein folding and stability [10]. Moreover, it has been reported that palmitoylation as lipid modification may also regulate ABCA1 localization and function. Indeed, ABCA1 is robustly palmitoylated at cysteines 3, −23, −1110, and −1111 [34]. In addition, molecular assembly has been proposed as a major parameter essential for ABCA1 function. Indeed, tetrameric ABCA1 complexes are required for efficient lipidation of Apo A-I particles [35]. These transiently assemble from the predominant dimeric forms during the ATP catalytic cycle [36]. Altogether,

358

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

these dynamic variations in molecular complexes strongly support the concept of modulation of ABCA1 function via molecular interactions. Emerging evidence indicates that protein–protein interactions other than homodimerization are important in posttranslational regulation of ABCA1 stability and trafficking. Direct interaction of ABCA1 and β1-syntrophin through the postsynaptic density, disk large, and ZO-1 protein (PDZ)-binding motif at the C-terminus of the transporter has been reported to stimulate ABCA1 activity and decrease its degradation [37]. In addition, Okuhira et al. identified a novel physical and functional interaction between ABCA1 and PDZ-RhoGEF/LARG, which activates ras homolog gene family, member A (RhoA), resulting in ABCA1 protein stabilization and cholesterol efflux activity [38]. More recently, a subunit of the enzyme involved in sphingomyelin synthesis, serine palmitoyltransferase, SPTLC1, has been found to copurify with ABCA1 and to negatively regulate its function [39]. Quite recently, it has also been shown that LXRα/RXR dimers bind directly to ABCA1 at the plasma membrane of macrophages and thereby modulates cholesterol excretion [40]. In parallel, it has been reported that Rab8 reduces foam cell formation by facilitating ABCA1 surface expression and stimulating endosomal cholesterol efflux to Apo A-I in primary human macrophages [41]. Moreover, newly discovered molecular partner of ABCA1, calmodulin protects ABCA1 from calpain-mediated degradation and upregulates HDL generation [42]. 17.4â•… ABCA1: MUTATIONS AND CLINICAL SIGNS Mutations at the ABCA1 gene locus cause TD or a form of the less severe familial HDL deficiency (FHD) with clinical signs analogous to the Tangier phenotype [43]. Interestingly, the clinical signs in FHD patients are related to heterozygous mutation in ABCA1 genes, whereas in Tangier patients, homozygosity or compound heterozygozity is required for the clinical manifestations. Lately, it has been reported that compound heterozygosity for the nonsense R282X and the missense mutation Y1532C in the ABCA1 gene causes TD. R282X has a detrimental effect on the function of ABCA1 since a premature stop codon is introduced. Mutation Y1532C disrupts the normal function of ABCA1 and its cell membrane localization as determined by in vitro analyses [44]. This implies that a specific set of mutations in the ABCA1 gene can transmit pathological phenotypes as a dominant trait. From a molecular standpoint, oligomerization of the transporter perfectly explains these divergent genetic behaviors [35, 36]. Indeed, mutations impairing the oligomerization exert a trans-dominant negative effect and gravely alter the cellular behavior even on a heterozygous genetic context [36]. Wide heterogeneity in clinical and biochemical lipid phenotypes is observed in patients that harbor mutations in ABCA1 gene [45, 46]. These may stem directly from the discrete and diverse functional effects of individual ABCA1 mutations. Accordingly, a spectrum of ABCA1 functional defects that corre-

359

ABCA1: Mutations and Clinical Signs

late with the observed clinical variability of both TD and FHD have been identified [47]. Several studies have shown that some rare ABCA1 mutations have a major effect on HDL levels, collectively accounting for up to 10% of low HDL subjects [48]. In contrast, it has been reported that common single-nucleotide polymorphisms (SNPs) causing amino acid changes in ABCA1 coding sequence have only modest effects on HDL-C levels [45]. However, certain SNPs in ABCA1 promoter region have been reported to influence HDL-C levels and susceptibility to cardiovascular disease [49]. Three point mutations detected in ABCA1 gene in the region 580–600 in Tangier pedigrees (namely R587W, W590S, and Q597R) show inadequate phospholipid effluxes [50] (Fig. 17.3). It was demonstrated that HDL-C phenotypes were associated with mutation-defined heterozygotes [46]. Recently, two heterozygous mutations c.5398Aâ•¯>â•¯C and c.2369Gâ•¯>â•¯A in the ABCA1 gene associated with HDL cholesterol deficiency in the serum have been reported [51]. Undoubtedly, mutations in ABCA1 predispose to, and are associated with, an increased risk of

R587W W590SQ597R C1477R Y1532C

A255T

COOH NH2

∆L693 T929I

R1680W

N935S A1046D

M1091T

N1800H R2081W P2150L

Figure 17.3.â•‡ ABCA1 mutations and associated phenotypes (adapted from References 46 and 136). A schematic diagram mapping specific mutations along the sequence and predicted topology of ABCA1 is presented. The mutations were characterized according to their capacity to elicit not only surface binding of Apo A-I and lipid efflux but also intracellular trafficking. Seven mutants, R587W, Q597R, L693, N935S, N1800H, R2081W, and Y1532C, do not traffic properly to the plasma membrane. They show reduced ability to elicit Apo A-I binding, indicating that a proper localization is mandatory for function. Conversely, plasma membrane localization is essential but not sufficient to drive effluxes as demonstrated by C1477R. M1091T is a dominant negative mutant that shows the most severe phenotype in heterozygous context. W590S, A255T, and T929I mutants exhibit normal Apo A-I binding while displaying impaired phospholipid and cholesterol efflux, thus reinforcing the concept that effluxes require lipid transport function. Color version on the Wiley web site.

360

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

TABLE 17.1.â•‡ Mutation/Phenotype Correlation for Heterozygous and Homozygous Mutations in ABCA1 (Adapted from References 46 and 136) ABCA1 Mutations and Associated Phenotypes FHD (Heterozygous Mutations) HDL-C level â•… <50% HDL-C

50% HDL-C

ABCA1 mutations â•… A255T P85L â•… W590S R587W â•… T929I ΔL693 â•… A937V R909X â•… R1680W N935S A1046D D1099Y C1477R ΔE1893 ΔD1894 R2081W 2145X 2203X

Tangier Disease (Homozygous Mutations)

>50% HDL-C

10%–15% HDL-C

1%–4% HDL-C

M1091T

A255T Q597R R1680W

635X N935S N1800H 1851X 2203X C-term Δ

Localization of major missense mutations, their respective phenotype, and associated level of HDL-C are indicated. FHD, familial hypoalphalipoproteinemia; Δ, deletion.

cardiovascular diseases and premature atherosclerosis. In all cases, low level of HDL-C and decreased cholesterol effluxes from peripheral cells are evidenced. Whether the increased risk of pathology results from the lower than normal level of HDL-C or from the compromised function of ABCA1 in macrophages has still to be established [52] (Table 17.1). Most of ABCA1 mutants with a complete loss of function do not traffic correctly to the plasma membrane. Impaired trafficking, which results, in this case, in a complete loss of Apo A-I binding and of lipid efflux [47], is a frequently encountered molecular defect in diseases related to ABC transporters. Indeed, a correct subcellular localization of these molecules has obvious and dramatic consequences on the physiology of the transport that they drive [53–55]. 17.5â•… TARGETED INACTIVATION AND OVEREXPRESSION OF ABCA1 IN ANIMAL MODELS The generation of animal models with targeted inactivation of abca1 has contributed to the accumulating evidence that ABCA1 is intimately involved

Targeted Inactivation and Overexpression of ABCA1

361

in lipoprotein metabolism and the biogenesis of HDL-C. Generalized deletion of abca1 in mice produces a pathophysiological phenotype similar to human TD, with low levels or loss of HDL-C, and lipid deposition in various tissues. Knockout mouse models have been generated on two distinct genetic backgrounds: the C57BL/6SJL and DBA/1LacJ strains. Both models show virtual absence of HDL in homozygous mice and sterility of homozygous females due to fetal loss during pregnancy [56]. Conversely, the extent and severity of lipid accumulation in various tissues differed in the two backgrounds. In the C57BL/6SJL strain, there is accumulation in the testis, thymus, liver, and placenta, whereas in the DBA/1LacJ strain, a major site of lipid accumulation was the lungs [14]. The reasons for these differences remain largely unexplained, although the most likely rely on the presence of strain-specific modifier genes. The Wisconsin hypo-alpha mutant (WHAM) in chicken is a natural animal model of ABCA1 deficiency. It shows high phenotypic similarities with the murine knockout models or human TD [57], notably an almost complete lack of circulating HDL particles and accumulation of the lipids in the liver and intestine. Studies in transgenic mice have identified distinctive roles for ABCA1 in plasma lipoprotein metabolism, macrophage cholesterol homeostasis, and atherogenic risk. Selective overexpression of ABCA1 in the liver increases plasma HDL-C levels, highlighting that besides its role in the biogenesis of Apo A-I and nascent HDL particles, the liver is also involved in ABCA1-mediated hepatic cholesterol efflux. Singaraja et al. showed that ABCA1 overexpression in the liver slows the catabolism of HDL in plasma and reduces selective highdensity lipoprotein cholesteryl esters (HDL-CE) uptake by the liver [58, 59]. Targeted inactivation of hepatic abca1 confirmed the critical contribution of the liver in maintaining the levels of circulating mature HDL by direct lipidation of hepatic lipid poor Apo A-I, in slowing Apo A-I catabolism by the kidney and in prolonging its plasma residence time [8, 60]. In vivo studies by Basso and coworkers have suggested that ABCA1 modulates intracellular cholesterol concentration in both hepatocytes and in peripheral cells [61]. This is consistent with the increase of intrahepatic cholesterol in Tangier patients and WHAM chicken [57]. The studies on various abca1 transgenic mice models generated conflicting results on the role of ABCA1 in atherosclerosis. While targeted overexpression of human ABCA1 in macrophages of ldlr-deficient mice has clearly been shown to be atheroprotective [62], its untargeted overexpression resulted in either atheroprotective or atherogenic effects [63, 64]. Though surprisingly these divergent results can originate from differences in genetic backgrounds, tissue specificity of abca1 expression in the transgenic lines and strength of the promoter are used to drive expression of the transgene. In support of these explanations, important variations in HDL handling have been reported among mouse strains [65–67].

362

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

17.6â•… LIVER AND MACROPHAGE ABCA1: LIPID EFFLUX AND HDL FORMATION While it is universally accepted that HDL-C concentration is an indicator of the risk for development of cardiovascular disease, the relative contribuÂ� tion of specific tissues to the maintenance of plasma HDL-C level is still unclear. In spite of an almost ubiquitous expression of ABCA1, the accumulation of cholesterol in TD or abca1 knockout mice is almost exclusively restricted to macrophages. This has been considered as indirect evidence that macrophages serve as a major source for HDL-C particle formation in the body. However, studies on animal models with monocyte/macrophage-specific abca1 expression or inactivation demonstrated that the contribution of macrophages to the absolute amount of the plasma HDL-C is minor [68, 69]. These studies provide a fair quantification of the relative contribution of various tissues to the circulating levels of HDL-C. Quantitatively, the liver represents the major source of HDL, and accounts for 70%–80% of plasma HDL-C in mice, the remaining fraction comes largely from the intestine, the second contributor to plasma HDL levels [70, 71]. This is not unexpected since the liver is one of the largest organs in the body and has the highest metabolic activity. Hepatocytes compose 70%–80% of liver mass [72], and they are involved in the synthesis of Apo A-I, with which they assemble and export lipoproteins such as very low-density lipoproteins (VLDLs) and HDL. In addition, the liver contains the largest macrophage population in the body (20% of total body macrophages). As compared with hepatocytes, the total number of macrophages in the mouse is small (approximately 108 cells) [73]. Hence, quantitatively, the mass amount of circulating cholesterol contributed by macrophages must be minor. It logically follows that the liver contribution to overall HDL-C is dominant and that the cell mass/cholesterol flux ratio of liver versus extrahepatic tissue macrophages is largely in favor of the liver. While hepatic ABCA1 appears crucial for phospholipid transport, extrahepatic tissues play an important role in cholesterol transport to HDL-C. This identifies specific and discrete roles of both liver and extrahepatic ABCA1 in HDL-C biogenesis in vivo and shows that ABCA1 has lipid cargo selectivity depending on its expression site [59]. ABCA1 is expressed on extrahepatic tissue macrophages and plays an essential role in controlling their lipid hemostasis [74]. ABCA1-impaired lipid efflux from these cells contributes to the formation of lipid-loaded macrophagederived foam cells and to the development of atherosclerotic lesions in mice [74]. Indeed, monocyte-derived macrophages are central cells that accumulate cholesterol in early atherosclerotic lesions, a manifestation of their scavenging function [31]. Cholesterol overload triggers, in these cells, compensatory pathways orchestrated by the LXR nuclear receptors that ultimately lead to activation of lipid efflux mediated by ABCA1 [75].

ABCA1 and Membrane Function

363

Collectively, these data highlight that ABCA1 expressed in liver and in macrophage exert complementary roles in the overall cholesterol trafficking in the body [76]. 17.7â•… ABCA1 AND MEMBRANE FUNCTION ABCA1 is localized at the plasma membrane, and several functions involvÂ� ing membrane lipid dynamic have been attributed to ABCA1 since its discovery. By using in vivo loss-of-function and in vitro gain-of-function models, it has been demonstrated that ABCA1 generates enhanced Annexin V binding at the cell surface; this reflects outward exposure of phosphatidylserine [77]. Exofacial phosphatidylserine exposure is crucial for many cell functions: the formation of plasma membrane protrusions enabling the engulfment of apoptotic cells [78, 79] and receptor-mediated or fluid-phase endocytosis [80]. Indeed in macrophages, ABCA1 is required for efficient clearance of corpses generated by programmed cell death [81], and ABCA1 overexpressing cells exhibit a reduced rate of endocytosis [82–84]. Besides this scavenger function, it has been suggested that ABCA1-controlled phosphatidylserine outward flopping modifies the biophysical properties of the membrane environment. This, by itself, may favor the release of phospholipids and cholesterol to Apo A-I [85]. Additionally, Wang et al. demonstrated that cells overexpressing ABCA1 assume a peculiar morphology due to the accumulation of phospholipids or other amphipathic molecules in the outer leaflet of plasma membrane [86, 87]. All these processes are likely to result from an ABCA1-dependent redistribution of lipid species across the membrane bilayer, preceding the phenomenon of ABCA1-driven lipid effluxes. In line with these findings, a number of clinical phenotypes unrelated to lipid efflux have been described in mouse models carrying a genetic invalidation of abca1. These phenotypes are distinct, but they are invariably related to the physicochemical behavior of membranes. They range from impaired immune responses during malarial infection [88] to increased development of Alzheimer’s plaque lesions [89]. Complete resistance to fatal cerebral malaria upon infection with Plasmodium berghei ANKA (PbA) has been demonstrated in abca1−/− mice [88]. This is accompanied by an impairment of cellular responses, as shown by lower plasma TNF-α levels and lower number of plasma microparticles (MPs). In this study, it was argued that protection from cerebral malaria results from a decrease in microparticle production late during the course of infection. This is consistent with the reported influence of ABCA1 on phosphatidylserine exposure, a hallmark of production of MPs [88]. The impaired ability to expose phosphatidylserine in abca1−/− mice may thus be at the origin of the phenotype, independently of lipid effluxes to HDL.

364

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

Abca1−/− mice, when crossed to Alzheimer’s disease prone mice, have decreased lipid associated to Apo E and show increased amyloid deposition [90]. Since formation of amyloid is dependent on lipid rafts [89], the phenotype may not stem from the function of ABCA1 as a lipid provider for the HDL particles in the brain, but rather from ABCA1-induced changes in membrane lipid organization. 17.8â•… ABCA1: LIPID FLOP AND LIPID EFFLUX As previously mentioned, the contact between Apo A-I and ABCA1-expressing cells is critical in mediating lipid transport. However, the molecular basis of the interaction between ABCA1 and Apo A-I is yet to be elucidated in spite of the immense efforts gathered so far. On the one side, chemical cross-linking studies suggested that Apo A-I may directly interact with ABCA1 during the lipid transfer process in cells stimulated with cAMP [15, 91]. Using cells overexpressing mutated forms of ABCA1, Fitzgerald et al. demonstrated that the ECLs of ABCA1 could be responsible for binding Apo A-I [92]. Moreover, several naturally occurring mutations in the ECLs of ABCA1 display defective Apo A-I binding and cholesterol efflux activity, suggesting the need for a direct receptor/ligand interaction. However, since selected ABCA1 variants can display defective cholesterol efflux but normal Apo A-I binding [92], it seems plausible to envision that other elements in addition to the binding of Apo A-I are required to promote cholesterol efflux. Assessments by fluorescent recovery after photobleaching (FRAP) of mobility parameters of Apo A-I bound to the surface of lipid efflux-competent cells reinforced the idea of an interaction (direct or indirect) with an integral membrane protein acting as a specific receptor [93]. Conversely, numerous reports indicate that the Apo A-I interaction with the membrane does not require a specific receptor-mediated interaction. Indeed, it has been shown that the majority of apolipoproteins with lipophilic A class amphipathic helices including Apo A-I, Apo A-II, Apo E, and likely Apo A-IV, can provoke lipid efflux [94] and interact with ABCA1-expressing cells. ABCA1 thus could recognize structural features such as cluster of amphipathic helices shared by all apolipoproteins [95]. These uncertainties on the nature of interactions were reinforced by Chambenoit and coworkers, who, on the basis of diffusion measurements by fluorescence correlation spectroscopy (FCS), proposed an interaction of Apo A-I with membrane lipid rather than proteins [85]. The importance of an intact ABCA1 ATPase activity for the Apo A-I binding at the cell surface also supports this hypothesis. In the attempt to accommodate most of the reported data, a multistep model has been proposed. In this view, Apo A-I would initially bind to ABCA1generated lipid domains at the cell surface, in close physical proximity to

ABCA1: Lipid Flop and Lipid Efflux

365

ABCA1. The “tethered” Apo A-I would then diffuse within the plane of the membrane until it contacts ABCA1, and protein–protein interaction would trigger the lipidation of Apo A-I [94]. Interestingly, Vaughan et al. proposed that the ABCA1 ability to promote cholesterol effluxes is independent from and precedes its actual binding to Apo A-I. This corroborates the hypothesis that ABCA1 primary function is to redistribute cholesterol to membrane domains readily accessible to apolipoproteins rather than directly pump cholesterol out of the cell [96]. Other studies have demonstrated that the retroendocytosis pathway is important for the acquisition of lipids by Apo A-I. Whether this is the predominant mechanism or accounts only for a contingent of the effluxed cholesterol is still subject of debate [97, 98]. Although ABCA1 promotes the efflux of both cholesterol and phospholipids, it is still unclear whether their transport occurs simultaneously or independently. Fielding and colleagues could dissociate phospholipid and cholesterol efflux to Apo A-I in the presence of vanadate and okadaic acid, both of which exclusively inhibited the latter [99]. Their data suggested that ABCA1dependent phospholipid efflux preceded that of free cholesterol to Apo A-I, and that phospholipid containing Apo A-I is a better acceptor of free cholesterol than Apo A-I itself. Similarly, Wang and colleagues came to the conclusion that the binding of Apo A-I to ABCA1 causes the formation of Apo A-I/ phospholipid complexes that subsequently promote free cholesterol efflux. Indeed, cyclodextrin preincubation abolished ABCA1-mediated cholesterol efflux but left unaffected phospholipid efflux and Apo A-I binding [100]. Yamauchi et al. compared Apo A-I-mediated cholesterol and phospholipid efflux from several fibroblast cell lines expressing ABCA1 [101]. They concluded that ABCA1 was required for the assembly of HDL from phospholipids. Finally, Sun et al. found that when fibroblasts are cotransfected with ABCA1 and stearoyl-CoAA desaturase 1 or 2 (SCD1 or SCD2), an inhibition of ABCA1-mediated cholesterol efflux was found. This could be correlated to the desaturase-dependent reduction in size of Triton X-100-resistant domains at the membrane [102]. Taken together, these data show that ABCA1-mediated phosphatidylcholine and free cholesterol efflux are separable events, and that ABCA1 is essential for the generation of Apo A-I/phospholipid complexes. Once generated, these complexes are able to take up cholesterol efficiently, even in the absence of ABCA1. In spite of the experimental dissociability of effluxes, phosphatidylcholine and free cholesterol could in real life be cotransported by ABCA1. Indeed, several reports suggest that ABCA1 can drive cotransport across cell membranes of a range of nonphospholipid molecules such as α-tocophenol [103], Apo E [104], and IL-1β [105]. By analogy, it could be envisioned that ABCA1 functions by transporting lipophilic molecules, including cholesterol, once complexed with phospholipids [75]. Considering that structural studies have proposed that ABC transporters, such as P-glycoprotein (Pgp), possess a

366

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

Out

PM

In 1

Substrate

Lipid-free Apo A-I

2

3

HDL

ABCA1

Figure 17.4.â•‡ Mechanistic model for the function of ABCA1 as a lipid floppase (adapted from Reference 126). Upon binding of the substrate to ABCA1, the ATP-derived energy induces a conformational change (1). This may bring the lipid into a state that favors extrusion to Apo A-I, if the acceptor is in proximity (2), or lipid flop to the external membrane leaflet (3). The membrane orientation is indicated (in and out). PM, plasma membrane. Color version on the Wiley web site.

central pore or pocket of 3â•¯nm in diameter, the hypothesis of simultaneous cotransport of large lipid complexes cannot be excluded a priori [106] (see also Chapter 11). Again by analogy with other members of the ABC family, we can extrapolate a schematic cycle for the function of ABCA1 as a lipid pump. The substrate (or substrates) will be acquired in the open conformation, when the binding site at the cytosolic side of the transporter is accessible. Then, upon ATP binding, NDBs will dimerize, close the conformation of the transporter, and lead to outward flop of the bound lipid. Once in the outer membrane environment, the lipid may either insert into the exofacial leaflet or, in the presence of specific acceptors, be released to the external environment (Fig. 17.4). 17.9â•… ABCA1 AND THE LIPID MICROENVIRONMENT AT THE MEMBRANE ABCA1 ability to facilitate desorption of membrane phospholipids and cholesterol is expected to have consequences on the lateral arrangement of membrane lipids and on the sorting of proteins in membrane domains.

ABCA1 and the Lipid Microenvironment at the Membrane

367

Recently, we provided several lines of evidence that ABCA1 expression leads to the alteration of the plasma membrane lipid microenvironment and suggested that the lipid rearrangement dependent on ABCA1 primes cells for lipid efflux [96]. Our studies indicated that lipid packing is the principal target of ABCA1 function. Taking advantage of biophysical techniques applied to living cells, we could determine that ABCA1 modifies the attributes of both the inner and outer membrane leaflet. Thus, we proposed that ABCA1 destabilizes the preexisting membrane lipid arrangement in domains and, as a consequence, modulates the membrane sorting of surface proteins such as transferrin receptor. Our data corroborate previous suggestions indicating that ABCA1, through its ATPase activity, reorganizes lipid microenvironment of the plasma membrane. Such reorganization destabilizes the lipid-phase order at the membrane, expands the membrane nonraft fractions, and hence primes cells for cholesterol efflux [107]. We originally proposed that an ABCA1-mediated phosphatidylserine flopping would act as a trigger for lipid rearrangement at the plasma membrane [77]. This active translocation would in practice lead to lateral redistribution of cholesterol across membrane domains and increased cholesterol chemical activity [96]. A major determinant of cholesterol escape tendency from membranes is its interaction with membrane phospholipids. Indeed, a decrease of plasma membrane sphingomyelin in CHO cell line promotes cholesterol export by ABCA1 to Apo A-I [108]. Similarly, synthetic ceramides, believed to displace cholesterol from phospholipid complexes, promote the export of cholesterol mediated by ABCA1 [109, 110]. Finally and in agreement with our suggestion, a chemically induced increase in the content of phosphatidylserine in the outer leaflet would also increase the chemical activity of cell surface cholesterol [111]. The relationship between ABCA1-dependent lipid efflux and membrane microdomains, defined as cholesterol-enriched areas, has prompted several investigations. Chronologically, Drobnik et al. first showed that ABCA1 is partially localized in Lubrol WX but not in Triton X-resistant microdomains in monocytederived macrophages, and that the cholesterol and choline phospholipids released to Apo A-I derived from Lubrol WX microdomains [112]. Later, it has been proposed that, while the interaction of Apo A-I with lipid rafts is necessary to stimulate cholesterol efflux, the exported cholesterol molecules rather originate from nonraft membrane regions [113]. Conversely, the finding that ABCA1 enhanced the release of cholesterol-enriched membrane vesicles, (microparticles) from macrophages and human fibroblasts provided indirect support to the concept of involvement of cholesterol-rich membrane domains in ABCA1-mediated lipid release onto HDL [114]. Liu et al. monitored the efflux of phosphatidylcholine, sphingomyelin and free cholesterol from J774 macrophages overexpressing ABCA1, and investigated the nature of released lipid particles. They found a large heterogeneity both in size and lipid content

368

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

of released particles consistent with an equally heterogeneous lipid origin from distinct membrane domains [115]. In agreement with the idea of flexible and multiple origins of lipid particles, it has been assessed that the rate of formation, the lipid composition, and the three-dimensional structure of cholesterol-rich HDL-C are dictated primarily by lipid-phase properties and cholesterol content of the original cell membrane [116]. Recently, it has been shown that ABCA1-expressing cells produced measurable amount of cholesterol/phospholipid microparticles, independently of Apo A-I; this implies that ABCA1, even in the absence of Apo A-I, can remodel the plasma membrane and favor the cholesterol membrane escape activity [117]. Collectively, though it is likely that the effluxes originate from specific membrane districts abounding in cholesterol molecules, it is still not clear what mechanisms drive the process. ABCA1, acting as a lipid floppase, affects the membrane lipid organization, but it is conceivable that modifications of membrane lipid composition may, in turn, modulate the activities of the transporter. Indeed, it has been suggested that the stimulation of ABCA1 ATPase activity is lipid specific [118]. In line with this, other ABC transporters such as Pgp and MRP1 are known to depend on the lipid environment for proper functioning [119, 120]. In membrane model systems, the ATPase activity of both proteins depends on the close proximity of specific phospholipids, especially phosphatidylethanolamine [121, 122]. In the case of Pgp, a higher affinity for substrates is measured when the embedding environment consists of lipid arranged in lo phase, biochemically defined as insolubility in cold nonionic detergents, such as Triton X-100 or Lubrol [123]. Finally, also the ATPase activity of ABCA4, the closest homolog of ABCA1 can be stimulated by lipids [124]. These findings collectively imply a bidirectional relationship between ABCA1 and the lipid environment at the membrane. On the one hand, ABCA1 affects dynamically the organization of lipid domains, and on the other hand, the proper functioning of the transporter is highly dependent on the arrangement and composition of the surrounding lipids. 17.10â•… CONCLUSIONS In spite of the physiological evidence that ABCA1 acts as a lipid transporter, a number of questions remain obscure. In fact, the nature of the substrate or the mechanistic of the transport itself are not yet ascertained and hitherto represent experimental challenges. Additional technical caveats exist in the field. The most relevant concerns the use of short-chain fluorescent lipids as potential substrates when addressing the problem of lipid translocation at the membrane. Considering the non-negligible differences with respect to natural lipids, these may be actually recognized as foreign elements or “drugs” rather than effectively transported as natural membrane lipids. In line with this, recently, it has been reported that ABCB1 and ABCC1 translocate short-chain

Abbreviations

369

glucosylceramide (GlcCer) [125, 126], but the translocation of natural GlcCer across the endoplasmic reticulum (ER)–Golgi membrane was independent of these multidrug transporters [127]. Therefore, reconstitution in in vitro systems as well as the use of natural lipid substrates is required to reach unequivocal identification of ABCA1-specific lipid substrates. In addition, the development of sophisticated biophysical approaches is required to improve both the temporal and spatial resolutions of lipid dynamics in the plane of a membrane; only those would allow to properly address the problem of the topographical cholesterol distribution and of microdomain reorganization under the influence of ABCA1. New methodologies, such as FCS [128–130], single particle (fluorophore) tracking (SPT) [131], multiphoton laser-scanning microscopy (MPLSM) imaging [132], atomic force microscopy (AFM) [133], and stimulated emission depletion (STED) far-field fluorescence nanoscopy [134], might represent adequate tools for future investigation in this field. However, only an intensive cross talk between biologists, chemists, and physicists will ultimately enable to draw a coherent picture emerging from the various experimental assessments and to properly evaluate the contribution of lipid translocation across membranes to the physicochemical attributes of biological bilayers. ACKNOWLEDGMENTS The authors wish to thank all the community of scientists working in the field of lipid transport and ABCA1 transporter, whose work may not have been quoted directly in this chapter due to space limitation. A.Z. was supported by a training grant from the European community (FLIPPASE network) and by a fellowship from Fondation de la Recherche Medical. The authors wish to thank all the partners in the FLIPPASE network for fruitful exchanges and collaborations.

ABBREVIATIONS ABCA1 AFM AP-1 Apo A-I cAMP CD36 CE ECL ER FC

ATP-binding cassette transporter A1 atomic force microscopy activator protein 1 apolipoprotein A-I cyclic adenosine monophosphate cluster of differentiation 36 cholesteryl ester extracellular loop endoplasmic reticulum free cholesterol

370

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

FCS FHD FRAP GlcCer HDL HDL-C HDL-CE HL HNF-3b IFN-γ IL-1β LCAT LDL-C LDLR LXR miR-33 MPs MPLSM mRNA NBD NF-kBs PbA PDZ Pgp PL PLTP PPAR RXR SCD1 or 2 SNPs SP1 SPT SR-BI STED TD TGF-β TMD TNF-α

fluorescence correlation spectroscopy familial HDL deficiency fluorescent recovery after photobleaching glucosylceramide high-density lipoprotein high-density lipoprotein cholesterol high-density lipoprotein cholesteryl ester hepatic lipase hepatocyte nuclear factor 1 interferon gamma Interleukin-1β lecithin–cholesterol acyltransferase low-density lipoprotein cholesterol low-density lipoprotein receptor liver X receptor microRNA-33 plasma microparticles multiphoton laser-scanning microscopy messenger RNA nucleotide-binding domain nuclear factor kappa-light-chain-enhancer of activated B cells Plasmodium berghei ANKA postsynaptic density, disk large, and ZO-1 protein P-glycoprotein phospholipid phospholipase transfer protein peroxisome proliferator-activated receptor retinoid X receptor stearoyl-CoAA desaturase 1 or 2 single-nucleotide polymorphisms specificity protein 1 single particle tracking scavenger receptor class B member 1 stimulated emission depletion Tangier disease transforming growth factor-β transmembrane domain tumor necrosis factor-α

References

VLDL WHAM

371

very low-density lipoprotein Wisconsin hypo-alpha mutant

REFERENCES â•… 1â•… J. R. Nofer, A. T. Remaley, Cell. Mol. Life Sci. 2005, 62, 2150–2160. â•… 2â•… A. D. Attie, Trends Biochem. Sci. 2007, 32, 172–179. â•… 3â•… G. D. Kolovou, D. P. Mikhailidis, K. K. Anagnostopoulou, S. S. Daskalopoulou, D. V. Cokkinos, Curr. Med. Chem. 2006, 13, 771–782. â•… 4â•… J. F. Oram, Trends Mol. Med. 2002, 8, 168–173. â•… 5â•… D. J. Gordon, B. M. Rifkind, N. Engl. J. Med. 1989, 321, 1311–1316. â•… 6â•… J. J. Genest, Jr., J. R. McNamara, B. Upson, D. N. Salem, J. M. Ordovas, E. J. Schaefer, M. R. Malinow, Arterioscler. Thromb. 1991, 11, 1129–1136. â•… 7â•… C. Serfaty-Lacrosniere, F. Civeira, A. Lanzberg, P. Isaia, J. Berg, E. D. Janus, M. P. Smith, Jr., P. H. Pritchard, J. Frohlich, R. S. Lees, G. F. Barnard, J. M. Ordovas, E. J. Schaefer, Atherosclerosis 1994, 107, 85–98. â•… 8â•… J. Y. Lee, J. M. Timmins, A. Mulya, T. L. Smith, Y. Zhu, E. M. Rubin, J. W. Chisholm, P. L. Colvin, J. S. Parks, J. Lipid Res. 2005, 46, 2233–2245. â•… 9â•… M. Bodzioch, E. Orso, J. Klucken, T. Langmann, A. Bottcher, W. Diederich, W. Drobnik, S. Barlage, C. Buchler, M. Porsch-Ozcurumez, W. E. Kaminski, H. W. Hahmann, K. Oette, G. Rothe, C. Aslanidis, K. J. Lackner, G. Schmitz, Nat. Genet. 1999, 22, 347–351. â•‡ 10â•… A. Brooks-Wilson, M. Marcil, S. M. Clee, L. H. Zhang, K. Roomp, M. van Dam, L. Yu, C. Brewer, J. A. Collins, H. O. Molhuizen, O. Loubser, B. F. Ouelette, K. Fichter, K. J. Ashbourne-Excoffon, C. W. Sensen, S. Scherer, S. Mott, M. Denis, D. Martindale, J. Frohlich, K. Morgan, B. Koop, S. Pimstone, J. J. Kastelein, J. Genest, Jr., M. R. Hayden, Nat. Genet. 1999, 22, 336–345. â•‡ 11â•… R. M. Lawn, D. P. Wade, M. R. Garvin, X. Wang, K. Schwartz, J. G. Porter, J. J. Seilhamer, A. M. Vaughan, J. F. Oram, J. Clin. Invest. 1999, 104, R25–R31. â•‡ 12â•… S. Rust, M. Rosier, H. Funke, J. Real, Z. Amoura, J. C. Piette, J. F. Deleuze, H. B. Brewer, N. Duverger, P. Denefle, G. Assmann, Nat. Genet. 1999, 22, 352–355. â•‡ 13â•… M. F. Luciani, F. Denizot, S. Savary, M. G. Mattei, G. Chimini, Genomics 1994, 21, 150–159. â•‡ 14â•… J. McNeish, R. J. Aiello, D. Guyot, T. Turi, C. Gabel, C. Aldinger, K. L. Hoppe, M. L. Roach, L. J. Royer, J. de Wet, C. Broccardo, G. Chimini, O. L. Francone, Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 4245–4250. â•‡ 15â•… J. F. Oram, R. M. Lawn, M. R. Garvin, D. P. Wade, J. Biol. Chem. 2000, 275, 34508–34511. â•‡ 16â•… A. R. Tall, Eur. Heart J. 1998, 19(Suppl. A),A31–A35. â•‡ 17â•… M. Dean, R. Allikmets, J. Bioenerg. Biomembr. 2001, 33, 475–479. â•‡ 18â•… S. Santamarina-Fojo, K. Peterson, C. Knapper, Y. Qiu, L. Freeman, J. F. Cheng, J. Osorio, A. Remaley, X. P. Yang, C. Haudenschild, C. Prades, G. Chimini, E. Blackmon, T. Francois, N. Duverger, E. M. Rubin, M. Rosier, P. Denefle, D. S. Fredrickson, H. B. Brewer, Jr., Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 7987–7992.

372

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

â•‡ 19â•… C. Albrecht, E. Viturro, Pflugers Arch. 2007, 453, 581–589. â•‡ 20â•… A. L. Armesilla, M. A. Vega, J. Biol. Chem. 1994, 269, 18985–18991. â•‡ 21â•… G. Cao, C. K. Garcia, K. L. Wyne, R. A. Schultz, K. L. Parker, H. H. Hobbs, J. Biol. Chem. 1997, 272, 33068–33076. â•‡ 22â•… A. F. Valledor, F. E. Borras, M. Cullell-Young, A. Celada, J. Leukoc. Biol. 1998, 63, 405–417. â•‡ 23â•… A. Chawla, J. J. Repa, R. M. Evans, D. J. Mangelsdorf, Science 2001, 294, 1866–1870. â•‡ 24â•… C. G. Panousis, S. H. Zuckerman, Arterioscler. Thromb. Vasc. Biol. 2000, 20, 1565–1571. â•‡ 25â•… T. Langmann, M. Porsch-Ozcurumez, S. Heimerl, M. Probst, C. Moehle, M. Taher, H. Borsukova, D. Kielar, W. E. Kaminski, E. Dittrich-Wengenroth, G. Schmitz, J. Biol. Chem. 2002, 277, 14443–14450. â•‡ 26â•… G. Schmitz, T. Langmann, Biochim. Biophys. Acta 2005, 1735, 1–19. â•‡ 27â•… W. L. Goff, P. Zheng, G. Brubaker, J. D. Smith, Arterioscler. Thromb. Vasc. Biol. 2006, 26, 527–533. â•‡ 28â•… T. J. Marquart, R. M. Allen, D. S. Ory, A. Baldan, Proc. Natl. Acad. Sci. U.S.A. 2010, 107(27), 12228–12232. â•‡ 29â•… S. H. Najafi-Shoushtari, F. Kristo, Y. Li, T. Shioda, D. E. Cohen, R. E. Gerszten, A. M. Naar, Science 2010, 328, 1566–1569. â•‡ 30â•… K. J. Rayner, Y. Suarez, A. Davalos, S. Parathath, M. L. Fitzgerald, N. Tamehiro, E. A. Fisher, K. J. Moore, C. Fernandez-Hernando, Science 2010, 328, 1570–1573. â•‡ 31â•… A. J. Lusis, Nature 2000, 407, 233–241. â•‡ 32â•… C. G. Panousis, G. Evans, S. H. Zuckerman, J. Lipid Res. 2001, 42, 856–863. â•‡ 33â•… S. Bungert, L. L. Molday, R. S. Molday, J. Biol. Chem. 2001, 276, 23539–23546. â•‡ 34â•… R. R. Singaraja, M. H. Kang, K. Vaid, S. S. Sanders, G. L. Vilas, P. Arstikaitis, J. Coutinho, R. C. Drisdel, D. El-Husseini Ael, W. N. Green, L. Berthiaume, M. R. Hayden, Circ. Res. 2009, 105, 138–147. â•‡ 35â•… M. Denis, B. Haidar, M. Marcil, M. Bouvier, L. Krimbou, J. Genest, J. Biol. Chem. 2004, 279, 41529–41536. â•‡ 36â•… D. Trompier, M. Alibert, S. Davanture, Y. Hamon, M. Pierres, G. Chimini, J. Biol. Chem. 2006, 281, 20283–20290. â•‡ 37â•… K. Okuhira, M. L. Fitzgerald, D. A. Sarracino, J. J. Manning, S. A. Bell, J. L. Goss, M. W. Freeman, J. Biol. Chem. 2005, 280, 39653–39664. â•‡ 38â•… K. Okuhira, M. L. Fitzgerald, N. Tamehiro, N. Ohoka, K. Suzuki, J. Sawada, M. Naito, T. Nishimaki-Mogami, J. Biol. Chem. 2010, 285, 16369–16377. â•‡ 39â•… N. Tamehiro, S. Zhou, K. Okuhira, Y. Benita, C. E. Brown, D. Z. Zhuang, E. Latz, T. Hornemann, A. von Eckardstein, R. J. Xavier, M. W. Freeman, M. L. Fitzgerald, Biochemistry 2008, 47, 6138–6147. â•‡ 40â•… M. Hozoji, Y. Munehira, Y. Ikeda, M. Makishima, M. Matsuo, N. Kioka, K. Ueda, J. Biol. Chem. 2008, 283, 30057–30063. â•‡ 41â•… M. D. Linder, M. I. Mayranpaa, J. Peranen, T. E. Pietila, V. M. Pietiainen, R. L. Uronen, V. M. Olkkonen, P. T. Kovanen, E. Ikonen, Arterioscler. Thromb. Vasc. Biol. 2009, 29, 883–888.

References

373

â•‡ 42â•… N. Iwamoto, R. Lu, N. Tanaka, S. Abe-Dohmae, S. Yokoyama, Arterioscler. Thromb. Vasc. Biol. 2010, 30, 1446–1452. â•‡ 43â•… A. Soro-Paavonen, J. Naukkarinen, M. Lee-Rueckert, H. Watanabe, E. Rantala, S. Soderlund, A. Hiukka, P. T. Kovanen, M. Jauhiainen, L. Peltonen, M. R. Taskinen, J. Lipid Res. 2007, 48, 1409–1416. â•‡ 44â•… J. Cameron, T. Ranheim, B. Halvorsen, M. A. Kulseth, T. P. Leren, K. E. Berge, Atherosclerosis 2010, 209, 163–166. â•‡ 45â•… R. Frikke-Schmidt, B. G. Nordestgaard, G. B. Jensen, A. Tybjaerg-Hansen, J. Clin. Invest. 2004, 114, 1343–1353. â•‡ 46â•… R. R. Singaraja, H. Visscher, E. R. James, A. Chroni, J. M. Coutinho, L. R. Brunham, M. H. Kang, V. I. Zannis, G. Chimini, M. R. Hayden, Circ. Res. 2006, 99, 389–397. â•‡ 47â•… L. R. Brunham, R. R. Singaraja, M. R. Hayden, Annu. Rev. Nutr. 2006, 26, 105–129. â•‡ 48â•… P. Pajukanta, J. Clin. Invest. 2004, 114, 1244–1247. â•‡ 49â•… T. L. Slatter, M. J. Williams, R. Frikke-Schmidt, A. Tybjaerg-Hansen, I. M. Morison, S. P. McCormick, Atherosclerosis 2006, 187, 393–400. â•‡ 50â•… V. Rigot, Y. Hamon, O. Chambenoit, M. Alibert, N. Duverger, G. Chimini, J. Lipid Res. 2002, 43, 2077–2086. â•‡ 51â•… G. U. Denk, C. Aslanidis, G. Schmitz, K. G. Parhofer, T. Pusl, Exp. Clin. Endocrinol. Diabetes 2011, 119, 53–55. â•‡ 52â•… R. J. Aiello, D. Brees, O. L. Francone, Arterioscler. Thromb. Vasc. Biol. 2003, 23, 972–980. â•‡ 53â•… M. D. Coltrera, S. M. Mathison, T. A. Goodpaster, A. M. Gown, Ann. Otol. Rhinol. Laryngol. 1999, 108, 576–581. â•‡ 54â•… A. Gilbert, M. Jadot, E. Leontieva, S. Wattiaux-De Coninck, R. Wattiaux, Exp. Cell Res. 1998, 242, 144–152. â•‡ 55â•… E. Puchelle, D. Gaillard, D. Ploton, J. Hinnrasky, C. Fuchey, M. C. Boutterin, J. Jacquot, D. Dreyer, A. Pavirani, W. Dalemans, Am. J. Respir. Cell Mol. Biol. 1992, 7, 485–491. â•‡ 56â•… T. A. Christiansen-Weber, J. R. Voland, Y. Wu, K. Ngo, B. L. Roland, S. Nguyen, P. A. Peterson, W. P. Fung-Leung, Am. J. Pathol. 2000, 157, 1017–1029. â•‡ 57â•… A. D. Attie, Y. Hamon, A. R. Brooks-Wilson, M. P. Gray-Keller, M. L. MacDonald, V. Rigot, A. Tebon, L. H. Zhang, J. D. Mulligan, R. R. Singaraja, J. J. Bitgood, M. E. Cook, J. J. Kastelein, G. Chimini, M. R. Hayden, J. Lipid Res. 2002, 43, 1610–1617. â•‡ 58â•… R. R. Singaraja, B. Stahmer, M. Brundert, M. Merkel, J. Heeren, N. Bissada, M. Kang, J. M. Timmins, R. Ramakrishnan, J. S. Parks, M. R. Hayden, F. Rinninger, Arterioscler. Thromb. Vasc. Biol. 2006, 26, 1821–1827. â•‡ 59â•… R. R. Singaraja, M. Van Eck, N. Bissada, F. Zimetti, H. L. Collins, R. B. Hildebrand, A. Hayden, L. R. Brunham, M. H. Kang, J. C. Fruchart, T. J. Van Berkel, J. S. Parks, B. Staels, G. H. Rothblat, C. Fievet, M. R. Hayden, Circulation 2006, 114, 1301–1309. â•‡ 60â•… J. M. Timmins, J. Y. Lee, E. Boudyguina, K. D. Kluckman, L. R. Brunham, A. Mulya, A. K. Gebre, J. M. Coutinho, P. L. Colvin, T. L. Smith, M. R. Hayden, N. Maeda, J. S. Parks, J. Clin. Invest. 2005, 115, 1333–1342.

374

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

â•‡ 61â•… F. Basso, L. Freeman, C. L. Knapper, A. Remaley, J. Stonik, E. B. Neufeld, T. Tansey, M. J. Amar, J. Fruchart-Najib, N. Duverger, S. Santamarina-Fojo, H. B. Brewer, Jr., J. Lipid Res. 2003, 44, 296–302. â•‡ 62â•… M. Van Eck, R. R. Singaraja, D. Ye, R. B. Hildebrand, E. R. James, M. R. Hayden, T. J. Van Berkel, Arterioscler. Thromb. Vasc. Biol. 2006, 26, 929–934. â•‡ 63â•… C. W. Joyce, M. J. Amar, G. Lambert, B. L. Vaisman, B. Paigen, J. Najib-Fruchart, R. F. Hoyt, Jr., E. D. Neufeld, A. T. Remaley, D. S. Fredrickson, H. B. Brewer, Jr., S. Santamarina-Fojo, Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 407–412. â•‡ 64â•… R. R. Singaraja, C. Fievet, G. Castro, E. R. James, N. Hennuyer, S. M. Clee, N. Bissada, J. C. Choy, J. C. Fruchart, B. M. McManus, B. Staels, M. R. Hayden, J. Clin. Invest. 2002, 110, 35–42. â•‡ 65â•… M. Mehrabian, L. W. Castellani, P. Z. Wen, J. Wong, T. Rithaporn, S. Y. Hama, G. P. Hough, D. Johnson, J. J. Albers, G. A. Mottino, J. S. Frank, M. Navab, A. M. Fogelman, A. J. Lusis, J. Lipid Res. 2000, 41, 1936–1946. â•‡ 66â•… B. Paigen, P. A. Holmes, D. Mitchell, D. Albee, Atherosclerosis 1987, 64, 215–221. â•‡ 67â•… B. Paigen, D. Mitchell, P. A. Holmes, D. Albee, Biochem. Genet. 1987, 25, 881–892. â•‡ 68â•… R. S. Kiss, D. C. McManus, V. Franklin, W. L. Tan, A. McKenzie, G. Chimini, Y. L. Marcel, J. Biol. Chem. 2003, 278, 10119–10127. â•‡ 69â•… M. Haghpassand, P. A. Bourassa, O. L. Francone, R. J. Aiello, J. Clin. Invest. 2001, 108, 1315–1320. â•‡ 70â•… L. R. Brunham, J. K. Kruit, J. Iqbal, C. Fievet, J. M. Timmins, T. D. Pape, B. A. Coburn, N. Bissada, B. Staels, A. K. Groen, M. M. Hussain, J. S. Parks, F. Kuipers, M. R. Hayden, J. Clin. Invest. 2006, 116, 1052–1062. â•‡ 71â•… L. R. Brunham, J. K. Kruit, T. D. Pape, J. S. Parks, F. Kuipers, M. R. Hayden, Circ. Res. 2006, 99, 672–674. â•‡ 72â•… G. Ramadori, F. Moriconi, I. Malik, J. Dudas, J. Physiol. Pharmacol. 2008, 59 (Suppl. 1), 107–117. â•‡ 73â•… S. H. Lee, P. M. Starkey, S. Gordon, J. Exp. Med. 1985, 161, 475–489. â•‡ 74â•… M. D. Wang, V. Franklin, Y. L. Marcel, Arterioscler. Thromb. Vasc. Biol. 2007, 27, 1837–1842. â•‡ 75â•… J. F. Oram, Curr. Opin. Lipidol. 2002, 13, 373–381. â•‡ 76â•… O. L. Francone, L. Royer, G. Boucher, M. Haghpassand, A. Freeman, D. Brees, R. J. Aiello, Arterioscler. Thromb. Vasc. Biol. 2005, 25, 1198–1205. â•‡ 77â•… Y. Hamon, C. Broccardo, O. Chambenoit, M. F. Luciani, F. Toti, S. Chaslin, J. M. Freyssinet, P. F. Devaux, J. McNeish, D. Marguet, G. Chimini, Nat. Cell Biol. 2000, 2, 399–406. â•‡ 78â•… B. Fadeel, Antioxid. Redox Signal. 2004, 6, 269–275. â•‡ 79â•… V. A. Fadok, A. de Cathelineau, D. L. Daleke, P. M. Henson, D. L. Bratton, J. Biol. Chem. 2001, 276, 1071–1077. â•‡ 80â•… E. Farge, D. M. Ojcius, A. Subtil, A. Dautry-Varsat, Am. J. Physiol. 1999, 276, C725–C733. â•‡ 81â•… M. F. Luciani, G. Chimini, EMBO J. 1996, 15, 226–235.

References

375

â•‡ 82â•… N. Alder-Baerens, P. Muller, A. Pohl, T. Korte, Y. Hamon, G. Chimini, T. Pomorski, A. Herrmann, J. Biol. Chem. 2005, 280, 26321–26329. â•‡ 83â•… E. B. Neufeld, J. A. Stonik, S. J. Demosky, Jr., C. L. Knapper, C. A. Combs, A. Cooney, M. Comly, N. Dwyer, J. Blanchette-Mackie, A. T. Remaley, S. Santamarina-Fojo, H. B. Brewer, Jr., J. Biol. Chem. 2004, 279, 15571–15578. â•‡ 84â•…X. Zha, J. Genest, Jr., R. McPherson, J. Biol. Chem. 2001, 276, 39476–39483. â•‡ 85â•… O. Chambenoit, Y. Hamon, D. Marguet, H. Rigneault, M. Rosseneu, G. Chimini, J. Biol. Chem. 2001, 276, 9955–9960. â•‡ 86â•… M. P. Sheetz, S. J. Singer, Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 4457–4461. â•‡ 87â•… N. Wang, D. L. Silver, P. Costet, A. R. Tall, J. Biol. Chem. 2000, 275, 33053–33058. â•‡ 88â•… V. Combes, N. Coltel, M. Alibert, M. van Eck, C. Raymond, I. Juhan-Vague, G. E. Grau, G. Chimini, Am. J. Pathol. 2005, 166, 295–302. â•‡ 89â•… Y. Sun, J. Yao, T. W. Kim, A. R. Tall, J. Biol. Chem. 2003, 278, 27688–27694. â•‡ 90â•… S. E. Wahrle, H. Jiang, M. Parsadanian, J. Kim, A. Li, A. Knoten, S. Jain, V. Hirsch-Reinshagen, C. L. Wellington, K. R. Bales, S. M. Paul, D. M. Holtzman, J. Clin. Invest. 2008, 118, 671–682. â•‡ 91â•… J. Wang, J. R. Burnett, S. Near, K. Young, B. Zinman, A. J. Hanley, P. W. Connelly, S. B. Harris, R. A. Hegele, Arterioscler. Thromb. Vasc. Biol. 2000, 20, 1983–1989. â•‡ 92â•… M. L. Fitzgerald, A. L. Morris, J. S. Rhee, L. P. Andersson, A. J. Mendez, M. W. Freeman, J. Biol. Chem. 2002, 277, 33178–33187. â•‡ 93â•… J. D. Smith, C. Waelde, A. Horwitz, P. Zheng, J. Biol. Chem. 2002, 277, 17797–17803. â•‡ 94â•… S. E. Panagotopulos, S. R. Witting, E. M. Horace, D. Y. Hui, J. N. Maiorano, W. S. Davidson, J. Biol. Chem. 2002, 277, 39477–39484. â•‡ 95â•… A. T. Remaley, J. A. Stonik, S. J. Demosky, E. B. Neufeld, A. V. Bocharov, T. G. Vishnyakova, T. L. Eggerman, A. P. Patterson, N. J. Duverger, S. SantamarinaFojo, H. B. Brewer, Jr., Biochem. Biophys. Res. Commun. 2001, 280, 818–823. â•‡ 96â•… A. Zarubica, A. P. Plazzo, M. Stockl, T. Trombik, Y. Hamon, P. Muller, T. Pomorski, A. Herrmann, G. Chimini, FASEB J. 2009, 23, 1775–1785. â•‡ 97â•… L. E. Faulkner, S. E. Panagotopulos, J. D. Johnson, L. A. Woollett, D. Y. Hui, S. R. Witting, J. N. Maiorano, W. S. Davidson, J. Lipid Res. 2008, 49, 1322–1332. â•‡ 98â•… I. Lorenzi, A. von Eckardstein, C. Cavelier, S. Radosavljevic, L. Rohrer, J. Mol. Med. 2008, 86, 171–183. â•‡ 99â•… P. E. Fielding, K. Nagao, H. Hakamata, G. Chimini, C. J. Fielding, Biochemistry 2000, 39, 14113–14120. 100â•… N. Wang, D. L. Silver, C. Thiele, A. R. Tall, J. Biol. Chem. 2001, 276, 23742–23747. 101â•… Y. Yamauchi, S. Abe-Dohmae, S. Yokoyama, Biochim. Biophys. Acta 2002, 1585, 1–10. 102â•… Y. Sun, M. Hao, Y. Luo, C. P. Liang, D. L. Silver, C. Cheng, F. R. Maxfield, A. R. Tall, J. Biol. Chem. 2003, 278, 5813–5820. 103â•… J. F. Oram, A. M. Vaughan, R. Stocker, J. Biol. Chem. 2001, 276, 39898–39902.

376

ABCA1 AFFECTS THE LIPID MICROENVIRONMENT AT THE MEMBRANE

104â•… A. Von Eckardstein, C. Langer, T. Engel, I. Schaukal, A. Cignarella, J. Reinhardt, S. Lorkowski, Z. Li, X. Zhou, P. Cullen, G. Assmann, FASEB J. 2001, 15, 1555–1561. 105â•…X. Zhou, T. Engel, C. Goepfert, M. Erren, G. Assmann, A. von Eckardstein, Biochem. Biophys. Res. Commun. 2002, 291, 598–604. 106â•… M. F. Rosenberg, R. Callaghan, R. C. Ford, C. F. Higgins, J. Biol. Chem. 1997, 272, 10685–10694. 107â•… Y. D. Landry, M. Denis, S. Nandi, S. Bell, A. M. Vaughan, X. Zha, J. Biol. Chem. 2006, 281, 36091–36101. 108â•… K. Nagao, K. Takahashi, K. Hanada, N. Kioka, M. Matsuo, K. Ueda, J. Biol. Chem. 2007, 282, 14868–14874. 109â•… Megha, E. London, J. Biol. Chem. 2004, 279, 9997–10004. 110â•… S. R. Witting, J. N. Maiorano, W. S. Davidson, J. Biol. Chem. 2003, 278, 40121–40127. 111â•… Y. Lange, J. Ye, T. L. Steck, Biochemistry 2007, 46, 2233–2238. 112â•… W. Drobnik, H. Borsukova, A. Bottcher, A. Pfeiffer, G. Liebisch, G. J. Schutz, H. Schindler, G. Schmitz, Traffic 2002, 3, 268–278. 113â•… K. Gaus, L. Kritharides, G. Schmitz, A. Boettcher, W. Drobnik, T. Langmann, C. M. Quinn, A. Death, R. T. Dean, W. Jessup, FASEB J. 2004, 18, 574–576. 114â•… P. T. Duong, H. L. Collins, M. Nickel, S. Lund-Katz, G. H. Rothblat, M. C. Phillips, J. Lipid Res. 2006, 47, 832–843. 115â•… L. Liu, A. E. Bortnick, M. Nickel, P. Dhanasekaran, P. V. Subbaiah, S. Lund-Katz, G. H. Rothblat, M. C. Phillips, J. Biol. Chem. 2003, 278, 42976–42984. 116â•… J. B. Massey, H. J. Pownall, Biochim. Biophys. Acta 2008, 1781, 245–253. 117â•… S. Nandi, L. Ma, M. Denis, J. Karwatsky, Z. Li, X. C. Jiang, X. Zha, J. Lipid Res. 2009, 50, 456–466. 118â•… K. Takahashi, Y. Kimura, N. Kioka, M. Matsuo, K. Ueda, J. Biol. Chem. 2006, 281, 10760–10768. 119â•… P. K. Dudeja, K. M. Anderson, J. S. Harris, L. Buckingham, J. S. Coon, Arch. Biochem. Biophys. 1995, 319, 309–315. 120â•… F. A. Sinicrope, P. K. Dudeja, B. M. Bissonnette, A. R. Safa, T. A. Brasitus, J. Biol. Chem. 1992, 267, 24995–25002. 121â•…X. B. Chang, Y. X. Hou, J. R. Riordan, J. Biol. Chem. 1997, 272, 30962–30968. 122â•… C. A. Doige, X. Yu, F. J. Sharom, Biochim. Biophys. Acta 1993, 1146, 65–72. 123â•… Y. Romsicki, F. J. Sharom, Biochemistry 1999, 38, 6887–6896. 124â•… J. Ahn, J. T. Wong, R. S. Molday, J. Biol. Chem. 2000, 275, 20399–20405. 125â•… A. van Helvoort, A. J. Smith, H. Sprong, I. Fritzsche, A. H. Schinkel, P. Borst, G. van Meer, Cell 1996, 87, 507–517. 126â•… G. van Meer, D. Halter, H. Sprong, P. Somerharju, M. R. Egmond, FEBS Lett. 2006, 580, 1171–1177. 127â•… D. Halter, S. Neumann, S. M. van Dijk, J. Wolthoorn, A. M. de Maziere, O. V. Vieira, P. Mattjus, J. Klumperman, G. van Meer, H. Sprong, J. Cell Biol. 2007, 179, 101–115. 128â•… P. F. Lenne, L. Wawrezinieck, F. Conchonaud, O. Wurtz, A. Boned, X. J. Guo, H. Rigneault, H. T. He, D. Marguet, EMBO J. 2006, 25, 3245–3256.

References

377

129â•… J. Lippincott-Schwartz, E. Snapp, A. Kenworthy, Nat. Rev. Mol. Cell Biol. 2001, 2, 444–456. 130â•… D. Marguet, P. F. Lenne, H. Rigneault, H. T. He, EMBO J. 2006, 25, 3446–3457. 131â•… A. Serge, N. Bertaux, H. Rigneault, D. Marguet, Nat. Methods 2008, 5, 687–694. 132â•… A. L. McIntosh, B. P. Atshaves, H. Huang, A. M. Gallegos, A. B. Kier, F. Schroeder, H. Xu, W. Zhang, S. Wang, J. C. Liu, Methods Mol. Biol. 2007, 398, 85–105. 133â•… D. J. Muller, Biochemistry 2008, 47, 7986–7998. 134â•… C. Eggeling, C. Ringemann, R. Medda, G. Schwarzmann, K. Sandhoff, S. Polyakova, V. N. Belov, B. Hein, C. von Middendorff, A. Schonle, S. W. Hell, Nature 2009, 457, 1159–1162. 135â•… A. R. Tall, J. Intern. Med. 2008, 263, 256–273. 136â•… R. R. Singaraja, L. R. Brunham, H. Visscher, J. J. Kastelein, M. R. Hayden, Arterioscler. Thromb. Vasc. Biol. 2003, 23, 1322–1332.

INDEX

Note: Page numbers in italics indicate figures; t denotes tables. ABCA1, xix, 133, 134–135 in animal models, targeted inactivation and overexpression of, 360–361 description of and interactions with, 356–358 generation of animal models with targeted inactivation of, 360, 361 genomic organization of, 354, 356 HDL levels and, 359 lipid flop and lipid efflux, 364–366 as lipid floppase, mechanistic model for, 366 lipid microenvironment at membrane and, 366–368 liver and macrophage: lipid efflux and HDL formation, 362–363 membrane function and, 363–364 monomer, topological model of structure of, 357 mutation/phenotype correlation for heterozygous and homozygous mutations in, 360t

mutations and associated phenotypes, 359 R1925Q, 345 regulation of expression for, 356 reverse cholesterol transport pathway and, 355 Tangier disease and, 89 historical notes, 353–354 mutations and clinical signs, 358–360 mutations in, 227 ABCA4, 354, 357 lipid translocation and: explaining a phenotype?, 228–230 Stargardt disease and mutations in, 227–230 ABCB1, 69, 368 cholesterol transport/efflux and, 89 drug and lipid movement by, 237–240 GlcCer transport by, 239 inhibitors of, 243

Transmembrane Dynamics of Lipids, First Edition. Edited by Philippe F. Devaux, Andreas Herrmann. © 2012 John Wiley & Sons, Inc. Published 2012 by John Wiley & Sons, Inc.

379

380 ABCB4 allocrite transport and, 243 familial intrahepatic cholestasis and, 226, 240 Mdr2 and, 242–243 ABCB11, bile formation and, 226 ABCC1, 68, 120, 368 ABCC2, bile formation and, 226 ABCD1, X-linked adrenoleukodystrophy and, 227 ABCG1, cholesterol transport and, 89 ABCG2, 68 ABCG4, cholesterol transport and, 89 ABCG5 progress in reconstitution of, 90 sitosterolemia and mutations in, 227 steroid secretion and, 89 ABCG8 progress in reconstitution of, 90 sitosterolemia and mutations in, 227 steroid secretion and, 89 Abcg5/Abcg8, 163 ABCG family, cholesterol transport/ efflux and proteins in, 89 ABC proteins cholesterol transbilayer movement and, 89 function of, 90 role of, xvii translocation of cholesterol across the membrane and, 90 ABCR (or rim protein), 228 ABC transporters. See ATP-binding cassette (ABC) transporters Absorption, distribution, metabolism, and excretion. See ADME Acetyl coenzyme A, cholesterol synthesis and, 75 Actinomycin D, 270, 271 influence of membrane mechanical properties in MDR and, 271, 272t Activated factor X, PS and formation of, 122 Activation energy, random walk model and, 267 Activation protein 1 ABCA1 gene promoter and, 356 Gea2p and recruitment to trans-Golgi membranes, 161 Acute chest syndrome, 330, 333

INDEX

Acyl-CoA-binding proteins, Lands pathway and, 318 Acyl-CoA transferases, Lands pathway and, 318 Acyl-coenzyme A, Lands pathway and, 318 Acyl groups attached to glycerol group, lipid transport activity and, 203–204 ADE. See Area difference elasticity ADE model, phase diagram of, 29 Adenosine triphosphate, flippase activity and, 200 ADME, bioavailability of drugs and, 255–256 ADP-ribosylation factor Gea2p and recruitment of, to membranes of trans-Golgi, 161 P4-ATPases in yeast membrane asymmetry and, 183 vesicle biogenesis and, 184 vesicle-mediated protein transport and, 161 Adrenoleukodystrophy, xix AE1. See Anion exchanger Band 3 AFM. See Atomic force microscopy AFM microscope, qualities of, 37 Aging, of red cells, 294, 295, 297 ALA1 in Arabidopsis thaliana, 55 properties of, found in Arabidopsis thaliana genome, 56t protein, aminophospholipid translocase(s) in plant cells and, 54 reduced expression of, in Arabidopsis thaliana, 58 ALA2, 160 isoform, plant cells and, 55 percentages of identity between isoforms of putative aminophospholipid translocases found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 56t ALA3, 160 isoform, plant cells and, 55 percentages of identity between isoforms of putative

INDEX

aminophospholipid translocases found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 56t ALA4 percentages of identity between isoforms of putative aminophospholipid translocases found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 56t ALA5 percentages of identity between isoforms of putative aminophospholipid translocases found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 56t ALA6 percentages of identity between isoforms of putative aminophospholipid translocases found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 57t ALA7 percentages of identity between isoforms of putative aminophospholipid translocases found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 57t ALA8 percentages of identity between isoforms of putative aminophospholipid translocases found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 57t ALA9 percentages of identity between isoforms of putative aminophospholipid translocases

381 found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 57t ALA10, 58 percentages of identity between isoforms of putative aminophospholipid translocases found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 57t ALA11 percentages of identity between isoforms of putative aminophospholipid translocases found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 57t ALA12 in Arabidopsis thaliana, 55 percentages of identity between isoforms of putative aminophospholipid translocases found in Arabidopsis thaliana genome, 56t properties of, found in Arabidopsis thaliana genome, 57t ALA expression, in Arabidopsis thaliana, during cold stress, 58, 59 Albinism-deafness, ATP11C linked to, 163 Albumin back-exchange procedure, 15–17 Albumin-extraction assays, for fluorescent and spin-labeled lipid analogs, 8, 9 Alcohols, transbilayer movement and, 326 Algae, aminophospholipid translocases in, 55 ALIS1, 55 All-trans-retinal, STGD and, 229, 230 Alzheimer’s disease ABCA1 and, 363, 364 lipid asymmetry and, xix Amine group, primary, optimal substrate recognition by flippase and, 201

382 Aminophospholipids, 121, 150, 150 accumulation of, in cytosolic leaflet of plasma membrane, 151 anionic, coagulation cascade, 342 flippases, biochemical requirements of, 206 flippases, lipid specificity for, question of uniqueness, 210-213 spin-labeled, lipid asymmetry and, 6 transport of, 88, 317 Aminophospholipid translocase(s), 49 asymmetric lipid distribution and, 120 clustering of 12 genes coding for, in Arabidopsis thaliana, 60, 61 endocytosis and, 276 linked to lipid asymmetry, 280 models explaining implication of, 283 inhibition of spontaneous flip-flop and, 152 inward lipid movement and, 292–292 percentages of identity between 12 isoforms of, found in Arabidopsis thaliana genome, 56t, 57t phylogeny of, from Arabidopsis thaliana, 56 in plant cells, 54–55, 58–59 properties of, found in Arabidopsis thaliana genome, 56t–57t red blood cell-associated, 151 scramblase activity and role for, 137–138 vesicle formation in yeast cells and, 281 Aminophospholipid transport activity, RBC flippase and, 319 Amphotericin, transbilayer movement and, 326 Anemias, hemolytic, PS-dependent recognition/engulfment of apoptotic cells and, 123 Angelman syndrome, ATP10A mutations and, 210, 323 Animal membranes, lipid asymmetry in, 47–49 Anion exchanger Band 3, 326 Anionic phospholipids, Ca2+-ATPase activation and, 211

INDEX

Annexin V, 51, 121, 327, 346 assessment of PS on surface of RBCs and, 325 diagnosis of Scott syndrome and, 343 enhanced binding of, ABCA1 and, 363 fluorescent, phosphatidylserine asymmetry and, 48 P4-ATPases in yeast membrane asymmetry and, 182, 183 protein-binding assay for endogenous lipids and, 11 Antibody-binding assay, for endogenous lipids, 8, 9 AP-1. See Activation protein 1 AP-1/clathrin-coated vesicles, flippase activity of Drs2p and budding of, 189 AP180, 188 Apical membranes, in polarized epithelial cells, 150 APLs. See Aminophospholipids APL translocase(s) (APLTs). See Aminophospholipid translocase(s) Apolipoprotein A-1 (Apo A-1) ABCA1 and, 359, 364–366 hepatocytes and synthesis of, 362 reverse cholesterol transport pathway and, 355 Tangier disease and absence of, 353, 354 Apoptosis. See also Dying cells Ca2+ homeostasis during, 298–299 cell biology of Scott syndrome and, 344–345 defined, 304 eryptosis and, 129–130 lipid asymmetry during, 297–298 lipid scrambling and ceramide production in, 138 PM blebbing and shedding during, 302 PS exposure, extracellular Ca2+ and, 125 PS exposure on surface of cell and, 324 regulation of, 306 regulation of lipid asymmetry during, 300 Scott defect and, 345 scrambling rates and, 127

INDEX

spatiotemporal distribution of PS and, 292 tissue remodeling and, 297 Apoptotic bodies, 302 Apoptotic cells ABCA1 and, 134 PS exposure and recognition of, 123 PS exposure in red blood cells and, 324 APs. See Apical membranes Arabidopsis thaliana ABC transporters in, 59 aminophospholipid ATPase and, 54 aminophospholipid translocases in, 55 clustering of 12 genes coding for aminophospholipid translocase in, expression levels, 60, 61 loss of ALA3 in, 161 microarray analysis of ALA expression in, during cold stress, 58, 59 percentages of identity between 12 isoforms of putative aminophospholipid translocase found in genome, 56t–57t percentages of identity between some (putative) aminophospholipid translocases from, 57t P4-ATPase and CDC50 gene numbers in, 160t phylogeny of 12 potential aminophospholipid translocases from, 56t properties of 12 putative aminophospholipid translocases found in genome, 56t–57t reduced expression of ALA1 in, 58 Arachidonate, 302 Area difference elasticity, shape changes of vesicles and, 28 ARF. See ADP-ribosylation factor ARF1, membrane curvature and, 278 ARF1 gene, 184 Arf1p, 184 ARF2 gene, 184 Arf2p, 184 ARF6, membrane curvature and, 278 ARF-dependent coats, types of, 184 ArfGEF, Drs2p C-tail and, 190

383 Arp2/3 complex, 188 Arrhenius’ law, influx into cytosol of given MW drug and, 264 Ascorbate reduction assay and, 15 tracing transmembrane migration of lipid analogs and, 105 Ascorbate assays, for fluorescent and spin-labeled lipid analogs, 8, 9 Aspartate (Asp), phosphorylated, P-type ATPase transport cycle and, 156 Assays back-exchange, 102–103 cholesterol movement and rationale of, 77 choline labeling assay, 242 for detection of lipid transbilayer distribution, 8 development of, distribution of lipids across membranes and, 4–7 intracellular trafficking of analogs for, consequences, 18–19 for measuring transbilayer cholesterol movement, problems related to, 80–81 for measuring transbilayer distribution of endogenous phospholipids, 100–102 chemical labeling, 101–102 phospholipase assay, 100–101 for measuring transbilayer distribution of phospholipid analogs, 102–105 bovine serum albumin, 102–103 exchange of lipid analogs between vesicles, 106 selective chemical reduction, 105 stopped-flow assays, 103, 105 water-soluble analogs, 106 overview, measuring distribution/ translocation of lipids across membranes, 7, 9 phospholipase-based, 121 prothrombinase, 48 Asymmetrical bilayers, vesicular shapes and, 27 Asymmetry loss of, ATP dependence and, 129 membrane lipid asymmetry/ permeability to drugs, 253–272

384 Asymmetry (cont’d) of membranes, shape changes and: examples, 28, 30 phospholipid, factors contributing to, 119 transmembrane PS, maintenance of, 199–200 Atherogenic risk, distinctive roles for ABCA1 in, 361 Atherosclerosis, ABCA1 and, 361 Atomic force microscopy, 369 defined, 36 direct observation of nanoscale defects produced by SM to ceramide conversion, 40 measurement of transmembrane flip-flop of unlabeled lipids with, 36–40 ATP8, substrate specificity of candidate aminophospholipid flippases and, 208–209 ATP8A1, 153, 176, 208, 319 cloning of, 152 consensus PS-binding sequences, 216t insect cells, and murine isoform of, 209 lipid structural specificity of, 214t location of cysteines in, 331 murine version of, 323 NEM modification of, 332 ATP8A2, 153, 162–163 ATP8B1, 153 deficiency of, 154 FIC1 disease and, 163–166 extrahepatic symptoms of, 164–166 mutations in, 209 potential mechanism of, 163–164, 164 Summerskill syndrome, Byler disease and, 320 ATP8B2, 153 flippase activity, in proteoliposomes, 181–182 ATP8B3, 153, 163 ATP8B4, 153 ATP9, substrate specificity of candidate aminophospholipid flippases and, 210 ATP9A, 153 ATP9B, 153

INDEX

ATP10, substrate specificity of candidate aminophospholipid flippases and, 210 ATP10A, Angelman syndrome, obesity and mutations in, 320 ATP10B, 153 ATP10D, 153 ATP11, substrate specificity of candidate aminophospholipid flippases and, 210 ATP11A, 153 ATP11B, 153 ATP11C, 153, 163 ATPase activity, of MsbA, 233–234 ATPase(s) inward lipid movement and, 293 P-type, xvii ATPase II, xvi, 49, 152, 176, 319 ATP-binding cassette (ABC) transporters, 49–50 ABCA1 and, 134 cholesterol transport across plasma membrane and, 89 lipid flip-flop and, 225–228 lipid transport and, 6–7 liver diseases and, 226 in plant cells, 59 PS externalization and, 345 ATP depletion, red blood cell aging and, 295 ATP hydrolysis, conformational motion in MsbA during, 235 A23187, 326, 327 A2-E, 229 A2-PE, 229 Autoimmune responses, membrane blebbing and, 304 Bacillus megaterium, flip-flop of PE in, TNBS assay and, 101 Back-exchange assay, 102–103 assessing transbilayer distribution/ movement of spin-labeled and fluorescent lipid analogs with, 15–18 back-extraction technique ABCB1 in transport of PAF study and, 240 lipid flippase and/or translocation and, 237, 238

INDEX

Bacteria, early investigations on transbilayer movement of phospholipids in, 111 Bacterial membranes inner phospholipid biosynthetic enzymes and, 99 transbilayer movement of phospholipids in, 110–112 transverse distribution of phospholipids in, 111 Bak, 294, 298 BAPTA-AM, chelation of intracellular Ca2+ with, 125 BAR. See Bin/Amphiphysin/Rvs Barley, transmembrane galactolipid asymmetry and, 53 Basolateral membranes, 150 Bax, 298 Bayer’s bridges, intermembrane lipid transport and, 111 Bcl-2, 298 BCL11A, 332 Bcl-XL, 294 Bending a membrane. See Membrane bending Bending modulus, magnitude of, in living cells, 259 Bending theory of vesicle shapes, 27–28 Benign recurrent intrahepatic cholestasis type 1, ATP8B1 and, 163 Beta-sitosterol, 89 β-thalassemia, pathological red cells and, 297 Bidirectional lipid movement, scramblase and, 294 Bid (type I and II cells), 298 Bilayer phospholipid movement across, 316 preserving stability of, 82 roles of lipid translocation through, 58 Bilayer couple hypothesis, flippases, vesicle formation and, 190 Bilayer couple model, 138 cholesterol flip-flop and, 91 shape changes, generation of curvature and, 278, 279

385 Bile formation, canalicular ABC transporters and, 226 PFIC and, 241 Bile acids, 80, 241 Bile canaliculus, bile production in, 241 Bile salt biosynthesis, cholesterol and, 75 Bile salt export pump, 163, 164 Bile salt transport, ATP8B1 deficiency and, 163 Bin/Amphiphysin/Rvs, membrane curvature and, 191, 275–276, 277 Bioavailability of drugs, 271 ADME and, 255–256 defining relevant biophysical parameters and, 257–258 drug resistance and, 270 maximizing, 253–254 Pgp-like transporters and, 266 Biogenic membranes flippases of, 50 rapid lipid flip-flop in, 112–113 Biological membranes lipid bilayer concept and, xv lipids as building blocks of, xiv phospholipid classes in, 199 transbilayer distribution of cholesterol in lipid membranes and, 82, 87 Biophysical parameters, Lipinsky’s second rule and definition of, 257–258 Biosynthetic labeling, for determining transbilayer distribution of lipid analogs, 20 Biotin-streptavidin-FITC treatment, endocytosis measurements and, 263 Black nightshade, transmembrane galactolipid asymmetry and, 53 Blebbing. See Membrane blebbing Blockbusters, pharmaceutical industry’s need for, 254–255 Blood-brain barrier, drug transporters and, 266 BLs. See Basolateral membranes B lymphocytes, PS expressed in, 292 Boltzmann’s constant (kB), energy required for drug crossing cellular membranes and, 258

386 Bovine chromaffin granules Drs2p as flippase and, 176 flippase activity possessed by, 207 Bovine serum albumin, 122 back-exchange assay and, 15 dithionite quenching technique and, 237 line shape and extraction of analogs from membranes to, 103 measuring transbilayer dynamics of lipid analogs by extraction to, 102–103 rapid extraction of lipid analogs from monolayer and, 104, 105 shape changes of membranes and, 30 Brefeldin A, sphingomyelin and, 70 BRIC1. See Benign recurrent intrahepatic cholestasis type 1 BSA. See Bovine serum albumin BSEP. See Bile salt export pump Budding transition C6-ceramide and externally added to egg-PC giant vesicle, 34 inducing, 33, 33 “shape change” approach and: perspectives, 35 shape changes in GUVs and, 31, 32 flip-flop of phospholipids assessed by, 107, 107–108 Bulk-phase endocytosis, yeast P4-ATPase mutant strain defects in, 162 Byler disease, ATP8B1 mutations and, 163, 323 Caco-2 cells, ATP8B1-depleted, FIC1 disease and, 165 Caenorhabditis elegans (nematode) inactivation of scrm-1 gene in, 133 P4-ATPase and CDC50 gene numbers in, 160t P-type aminophospholipid transporters, and PS exposure in, 137 TAT-1 deletion in, 300 Caenorhabditis elegans P4-ATPase TAT-1, removal of, 154 Calcium-calmodulin, PS exposure in sickle RBCs and, 329

INDEX

Calcium-induced phospholipid scrambling, PLSCR1 and, 326 Calcium-induced plasma membrane scrambling, TMEM16F and, 327 Calmodulin, 211 Calreticulin, apoptosis and, 303 cAMP. See Cyclic adenosine monophosphate Cancer ABCB1 and multidrug resistance in, 237 lipid asymmetry and, xix microparticles and, 347 Carboxyl groups, flippase substrate recognition and, 202 Carboxyl terminal cytosolic tail (C-tail), NPFXD motifs and, 188 Carboxypeptidase Y, protein trafficking pathways and, 186 Cardiolipin, 211 transport of, to inner membrane of Salmonella typhimurium, 110 Cardiovascular disease, in Tangier patients, 354 Caspase 3, 298 Caspase 8, 298 Caspases, 294 Castor bean (Ricinus communis), lipid asymmetry in, 51 Catalase, 331 Cations, lipid scrambling and effects of, 124–126 Ca2+-dependent mechanisms, outward movement of PS in cells and, 296 Ca2+ homeostasis, during apoptosis, 298–299 Ca2+ ions, phospholipid scrambling and role of, 124–126 CCVs. See Clathrin-coated vesicles CD36. See Cluster of differentiation 36 Cdc50, DRS2p protein and, 324 CDC50A, 159 CDC50B, 159 CDC50C, 159 CDC50 complexes, putative substrate specificities of, 153t CDC50 gene numbers, in various organisms, 160t

INDEX

Cdc50p, 159, 160, 283 family, chaperone function between P4-ATPases and, 175 proteins, P4-ATPase-catalyzed phospholipid transport and role of, 159–161 C8C8-GlcCer, 68, 238 C8C8-PE, 238 Cell fusion, PS exposure and, 123 Cell lysis, partial redistribution of lipids within bilayer and, 48 Cell physiology, membrane lipids and, 291 Cellular barriers, traversal of, bioavailability of drugs and, 257–258 Cellular membranes, assays for measuring distribution of lipids across, overview, 7, 9 Ceramidase, 68 Ceramide, xvi, 68 Cerebral malaria, MP release and, 348 CEs. See Cholesterol esters C5-BIODIPY analogs, 14 C57BL/6SJL strain, overexpression of ABCA1 and, 361 C-4-hydroxylase, 68 Charge properties, of drugs, 253 Chemical labeling measuring transbilayer distribution of endogenous phospholipids and, 101–102 transbilayer distribution of endogenous lipids and, 9–10 Chemical libraries, 256 Chemical modification assays, for endogenous lipids, 8 Chemical reduction, selective, measuring transbilayer dynamics of lipid analogs by, 105 Chinese hamster ovary, ATP8B1 and nonendocytic uptake of NBD-PS in, 154 Chiral carbons, PS molecule and, 202 Chiral membrane lipid molecules, membrane curvature and, 278 Chlamydomonas reinhardtii, percentages of identity between some (putative)

387 aminophospholipid translocases from, 57t Chloroplast, defined, 52 CHO. See Chinese hamster ovary Cholestatic liver diseases AT8B1 mutations and, 163 canalicular ABC transporters and, 226 Cholesterol ABCA1 and efflux of, 365 assaying transbilayer movement of, principles, 79 asymmetric distribution of, cause of, 87 flip-flop: fast or slow?, 87–88 functions of, 75 physicochemical features of, 76–77 proteins and transport across membranes, 88–90 rapid spontaneous diffusion of, xvi in red blood cell plasma membrane, 315 transbilayer movement and distribution of, 75–92 chemical structure of cholesterol analogs used for measuring, 78 lateral and transbilayer dynamics in membrane, liquid-ordered and liquid-disordered domains, 92 in lipid and biological membranes, 82, 87 in lipid membranes, summary of data published on, 83t–86t methods for measuring, 77–78, 80–81 in model membranes, 81–82 Cholesterol analogs endogenous cholesterol and reliability of, 81 MβCD-mediated removal of, 18 Cholesterol effluxes, ABCA1 and, 356, 365 Cholesterol-enriched areas, 5, 367 Cholesterol esters, Tangier disease and, 353, 355 Cholesterol flip-flop passive, unilamellar liposomes and data on, 81 quantifying rate of, 77–78, 80 Cholesterol oxidase, 11, 80

388 Cholesterol/steroid molecules, modification of, in one membrane leaflet, 77–78, 80 Choline labeling assay, 242 Choline phototransferase, 99 Chromaffin granule enzyme (murine version of ATP8A1), 323 Cinnamycin Ro09-0198, 11 Cirrhosis, 226 cis-Golgi membranes, flip-flop of lipids in, 152 CL. See Cardiolipin Clathrin-coated vesicles, vesiclemediated protein transport and, 183 Clathrin coat proteins, Gea2p and recruitment of, to membranes of trans-Golgi, 161 Clathrin-mediated endocytosis, membrane bending and, 277 Clathrin triskelion, intrinsic curvature in, 191 Cluster of differentiation 36, ABCA1 gene promoter and, 356 CoA. See Acyl-coenzyme A Coagulation critical role of PS and, 122–123 lipid flip-flop and, xviii phosphatidylserine and role of, 341 PS exposure on surface of cell and, 324 role of lipids in, 121 Coagulation cascade anionic aminophospholipids and, 342 phosphatidylserine and activation of, 292 Coat protein complex II, vesicles coated with, 183 Cold stress, ALA expression in Arabidopsis thaliana during, 58, 59 Collagen, PS, coagulation cascade and, 122 Collagen plus thrombin, platelets activated by, Ca2+ concentration and, 125 Cone-rod dystrophy, ABCA4 mutations and, 228 COPII. See Coat protein complex II

INDEX

CPY. See Carboxypeptidase Y Crflp, 159, 160 Critical cross section, defined, 258 Cross-resistance to drugs, Lipinski’s second law and, 270 C6 ceramides experimental time evolution of asymmetry function c for, 33 shape changes induced by, externally added to egg-PC giant vesicle, 34 unlabeled, use of “shape change” methodology and, 33–34 C6-NBD analogs, 14 C6-NBD-GlcCer, 68 C6-NBD-PC, 237, 238 intracellular uptake of, 19 C6-NBD-PS, distribution of, in K562 cells, after incubation at different temperatures, 281–283, 282 C16 ceramides, unlabeled, use of “shape change” methodology and, 33 C10 ceramides, unlabeled, use of “shape change” methodology and, 33 Curvature-generating or -sensing proteins, endocytic pathways and, 276–277 Cyclic adenosine monophosphate, ABCA1 and, 356 Cyclodextrins extraction of cholesterol from membranes and, 80 measuring cholesterol efflux from mammalian cell lines onto, 82 Cytochrome c (cyt c), 298, 302 Cytochrome P450s, ADME and, 255 Cytoplasmic leaflet of biogenic membranes, newly synthesized phospholipids in, 99 Cytosol, escape rate (or influx) of membrane-embedded drugs into, 268 Cytosolic Ca2+ fluxes, Scott cells and, 346 DAB/1LacJ strain, overexpression of ABCA1 and, 361 DAG. See Diacylglycerol Daunomycin, membrane mechanical properties in MDR and, 272

INDEX

Daunorubicin, MsbA studies and, 234–235 D-catalyzed transphosphatidylation, head-group-modified NBD-lipid analogs and, 128 Deficiency in ribosomal subunits. See Drs2 Dehydroergosterol, fluorescent, 81 Delta 4-desaturase, 68 DGDG. See Digalactosyldiacylglycerol DHE. See Dehydroergosterol Diabetes pathological red cells and, 297 type 2, mouse ortholog of ATP10A linked to, 163 Diacylglycerol outward PS movement and, 301 rapid spontaneous diffusion of, xvi Diamide, PS exposure and, 130 Dibutyroyl phosphatidylcholine (diC4PC), water solubility of, 106 DIDS, lipid scrambling and, 131 Dielectric breakdown of RBC membrane, transbilayer movement and, 326 Digalactosyldiacylglycerol structure of, 53 thylakoid membrane lipids and, 52 transmembrane galactolipid asymmetry and, 54 Digoxin, ABCB4 and, 243 Dihydroceramide, production of, 68 Dimyristoylphosphatidylcholine, microvesicle shedding induced by, 136 Dimyristoylsulfonoserine, 202 Di (N-acetylgluosaminyl) pyrophosphorylnerol (GlcNAcPP-nerol), 106 Disease, membrane lipid transport processes and, 226 Dismutase, 331 Disocytes, shape change to echinocytes, 30 Disulfide bonds, lipid scrambling, free sulfhydryl groups and, 130–131 Dithioerythritol, reduction of disulfide bonds and, 131

389 Dithionite fluorescence of NBD group chemically quenched by, 237 reduction assay and, 15 tracing transmembrane migration of lipid analogs and, 105 Dithionite assays, for fluorescent and spin-labeled lipid analogs, 8, 9 Dithionite quenching assay, 242 Dithiothreitol, 130 DLPC AFM, study of flip-flop and, 38 supported lipid bilayers made of, three different configurations with, 39 DMPC. See Dimyristoylphosphatidylcholine DMPS, 208 Dnf ATPases, protein trafficking pathways and, 186 Dnf1p, xvii, 153, 160, 208 defect in internalization step of endocytosis and loss of, 281 NBD-PS translocations across plasma membrane and, 178 plasma membrane flippase activities and, 178 in yeast Saccharomyces cerevisiae, 172 Dnf2p, xvii, 153, 160, 208 defect in internalization step of endocytosis and loss of, 281 NBD-PS translocations across plasma membrane and, 178 plasma membrane flippase activities and, 178 in yeast Saccharomyces cerevisiae, 172 Dnf3p, xvii, 153, 153, 160, 208 loss of, and abolition of PE asymmetry in post-Golgi secretory vesicles, 155 in yeast Saccharomyces cerevisiae, 172 Dolichyl phosphate, Glc3Man2GlcNac2PP-dolichol synthesized to, 115 DOPS, 203, 209 DPPS, 208 Drosophila melanogaster (fruit fly), 161 epithelial junction formation and tracheal tube-size control in, 165 P4-ATPase and CDC50 gene numbers in, 160t

390 Drs2, inward lipid movement and, 293 DRS2 genes, yeast and disruption of, 172–173 Drs2p, xvii, 152, 153, 153, 160, 161, 162, 208 Atp8b2 flippase activity and, in proteoliposomes, 181–182 defect in internalization step of endocytosis and loss of, 281 as downstream effector of PIP, 209 endocytosis of, 187–188 flip-flop of phospholipids and, 175–176 flippase activity of budding of AP-1/clathrin-coated vesicles and, 189 in Golgi membranes, 179–181 homology model of, based on crystal structure of sarcoplasmic/ endoplasmic reticulum Ca2+ATPase 1 in E2 conformation, 174, 174–175 implication of, as a flippase, 176–177 loss of, and abolition of PE asymmetry in post-Golgi secretory vesicles, 155 membrane phospholipid asymmetry and, 182 model for protein transport pathways requiring, 185 P4-ATPases in yeast membrane asymmetry and, 182–183 phosphoinositide metabolism and potential binding partners of, 283–284 protein trafficking pathways and, 186 in protein transport and vesicle budding, 183–191 Drs2p C-Tail, 188–190 endocytosis of Drs2p, 187–188 flippases and vesicle formation, 190–191 influence of other P4-ATPases on protein transport, 186–187 roles for Drs2p-Cdc50p in protein transport, 184–186 vesicle-mediated protein transport, 183–184 purified, reconstitution of, in artificial membranes, 192

INDEX

yeast plasma membrane activity and, debate over, 177 yeast plasma membrane flippase activities, 177–179 in yeast Saccharomyces cerevisiae, 172 Drs2p-Cdc50p, protein transport and roles for, 184–186 Drs2p C-tail, 188–190 Drs2p-dependent flippase activity, in Golgi membranes, 179–181 Drs2p flippase activity, in proteoliposomes, 181–182 DRS2p protein, Cdc50 and, 324 Drug bioavailability. See Bioavailability of drugs Drug extrusion from membrane, triggering, 268 Drug membrane transporters, bioavailability of drugs and, 255 Drug movement, by ABCB1, 237–240 Drug resistance, in cells, generating, 270 Drugs cross-resistance to, 270 Lipinski’s rules and physicochemical properties of, 256 membrane lipid asymmetry and permeability to, 253–272 modulating intracellular accumulation of, parameters related to, 269 successful, approaches for, 254–255 Drug transporters interaction between membrane and, 264, 266 membrane lipid asymmetry and, 266, 268–270 surface density of, defined, 268 Drug transverse movement, across membrane, mechanical properties involved in, 265 DTE. See Dithioerythritol DTT. See Dithiothreitol Dying cells. See also Apoptosis identifying characteristics of, 292 PS exposure on surface of cell and, 325 Dynamin membrane bending and, 277 membrane curvature and, 275 Dyslipidemia, Tangier disease and, 134

INDEX

Eat-me signal, red cell aging and, 297 EBV. See Epstein-Barr virus Echinocytes, disocytes changing shape to, 30 ECL-1, monomeric ABCA1 and, 357 ECL-2, monomeric ABCA1 and, 357 Edelfosine, 178 eggPC, Triton X-100-solubilized ER membrane proteins combined with NBD-labeled phospholipid and, 114 eggPC-GUV, time series of, 107 EGTA. See Ethylene glycol tetraacetic acid Electroformation, fabricating giant vesicles and, 31 Electron paramagnetic resonance (EPR) spectroscopy, MsbA characterized by, 235, 236 Electron spin resonance, labeled lipids developed for, 25 Electroswelling method, GUVs and, 106 Embryonic development, cholesterol and, 75 end3, endocytosis of Drs2p and, 187 end4, endocytosis of Drs2p and, 187 Endocytic invaginations membrane bending and, 276 P4-ATPases and, 283–284, 284 Endocytosis bulk-phase, yeast P4-ATPase mutant strains defective in, 162 lipid asymmetry and, 275–285 bending a membrane, 276–278 linkage between, 280–283 role of P4-ATPases in forming endocytic invaginations, 283–284, 284 shape changes of GUVs induced by, 278–280 lipid number asymmetry-induced fluid-phase endocytosis model and, 260 measurements of, 263 measuring kinetic rate of, 263 membrane bending and, 138 membrane surface asymmetry and vesicle formation during, 317

391 Endogenous lipids chemical modification assays for, 8 enzymatic assays for, 8 relevant assays for assessing transbilayer movement/ distribution of lipid analogs, 7, 8, 9 reporter moiety and behavior of, 4 shape changes and redistribution of, 30 techniques for determining transbilayer distribution of, in cell membranes, 9–11 chemical labeling, 9–10 enzymatic treatment, 10–11 protein-binding assay, 11 Endoplasmic reticulum, 119, 149, 369 ceramide synthesis in, 67 cholesterol and, 75 flip-flop of lipids in, 152 phospholipid biosynthetic enzymes and, 99 symmetric lipid distribution shown in, 150 transbilayer movement of phospholipids in, 108–110 translocation of ceramide across, 68 of yeast, shape change approximation and, 34 Endoplasmic reticulum flippase, in plant cells, 59, 61 Enzymatic assays, for endogenous lipids, 8 Enzymatic treatment, transbilayer distribution of endogenous lipids in cell membranes and, 10–11 E1 conformation P-Type ATPase transport cycle and, 156, 156, 157, 158 structure of P2A-type calcium pump SERCA in, 157 E1/E2 model, transport cycle of P2A- and P4-ATPases according to, 156 Epithelial cells, polarized, nonrandom lipid distributions in, 150 Epsin, 188, 276–277 Epstein-Barr virus, Scott membrane characteristics and, 343–344 Epstein-Barr virus (EBV)-transformed B cells, scramblase rate, 127

392 ER. See Endoplasmic reticulum ER Ca2+, regulation of, 299 ER flippases, independent function of, 151 Erythrocyte ATPase, lipid structural specificity of, 214t Erythrocyte ghosts (human), ATPinduced vesiculation and phospholipid analogs incorporated into, 280 Erythrocyte membrane extract, giant vesicle made from, 36 Erythrocyte morphology, incorporation/ distribution of lipids on, 28 Erythrocytes asymmetrical distribution in, 47 ATP depletion and loss of Ca2+induced scrambling activity in, 129 K+ ions and inhibiting phospholipid scrambling in, 344 lipid asymmetry in, 294 scanning electron micrographs of, 30 scrambling rate for, 127 Erythrocyte shape, lipid scrambling and, 139 Erythropoiesis ineffective, in thalassemia, 329 presence of PS-exposing RBCs in circulation during, 329 Escape rate (or influx), of membraneembedded drugs into cytosol, 268 Escherichia coli bidirectional phospholipid transport between inner and outer membrane in, 111 lipid extract, regions with different lipid-packing density in supported lipid bilayer made of, 37 measuring flip -flop of phospholipids in inner membrane of, 101 Msba as essential gene in, 230 Ethylene glycol tetraacetic acid, Ca2+ removal with, 124 E2 conformation homology model of Drs2p, based on crystal structure of sarcoplasmic/ endoplasmic reticulum Ca2+ATPase 1 in, 174

INDEX

P-Type ATPase transport cycle and, 156, 156, 157, 158 Eukaryotic cell membranes, definition of various lipid transporters in, xv Eukaryotic cells exogenous lipids and accelerated endocytic activity of, 281 sphingolipids as typical features of, 65 Eukaryotic ER, flip-flop half-times and transbilayer movement in, 112 Exocytosis lipid number asymmetry-induced fluid-phase endocytosis model and, 260 membrane bending and, 138 Exosomes, MPs discriminated from, 347 Factor Va, lipid structural specificity of, 214t Factor VIII, lipid structural specificity of, 214t Factor Xa, lipid structural specificity of, 214t Familial HDL deficiency, heterozygous mutations in ABCA1 and, 360t Familial intrahepatic cholestasis type 1 disease. See FIC1 disease FAPP2, 69 Farber’s disease, ceramidase and, 68 Fast scanning probe equipment, AFM and, 37 Fatty acid chain, lipid translocation and, 14 Fatty acid liver disease, nonalcoholic, mouse ortholog of ATP10A linked to, 163 F-BAR, membrane curvature and, 277 FCS. See Fluorescence correlation spectroscopy Fertility disorders, ATP8B3 and, 163 FHD. See Familial HDL deficiency FIC1 disease, 163 ABCB4 and, 226 ATP8B1 and, 154, 163–166 description of, 163 extrahepatic symptoms of, 164–166 potential mechanism of, 163–164

INDEX

Flip-flop AFM and study of, 38 cholesterol fast or slow?, 87–88, 90–91 recent data on, 90 of cholesterol, 77 of lipids, mediating in biogenic membranes, 7 measuring, shape changes of GUVs as tool for, 106–108 of phospholipids assessed by shape changes of GUVs, 107 in biogenic membranes, 99–116 Drs2p and, 175–176 SFVS and measurement of, 21 rapid, of spin-labeled phospholipids assessed by, 104, 105 rapid lipid, mechanism of, in biogenic membranes, 112–113 shape change, reconstituted ER flippase activity and, 35 Flip-flop half-time of phospholipids biogenic membranes and, 101 comparative values of, 26 rapid lipid flip-flop in biogenic membranes and, 112 Flip-flop rates, of nonlabeled lipids, shape change approach and, 31, 34–35 Flippase, viii, xv, xvi, 7 ATP-dependent, xvii, xviii of biogenic membranes, 50 containing two specific lipid-binding sites, 217 fast transbilayer movement of lipids in ER and, 112 implication of Drs2p as, 176–177 lipid structural specificity of, 214t phospholipid, identifying, 113 RBC flippase, 319–320, 323–325 substrate specificity of PM aminophospholipid flippase, 200–205 vesicle formation and, 190–191 yeast plasma membrane flippase activities, 177–179 Floppase, xv, xvi, xvii outward lipid movement and, 293–294

393 “Fluid mosaic membrane” model, xv Fluid-phase endocytosis cytosolic pressure and control of, 263 Lipinski’s second rule and role of, 258–261 Fluid-phase endocytosis model, lipid number asymmetry-induced, 260 Fluorescamine, 9, 101 endogenous lipid modification and, 8 vacuole treated with, 52 Fluorescence correlation spectroscopy, 364, 369 Fluorescence recovery after photobleaching, 38, 364 Fluorescence spectroscopy, labeled lipids developed for, 12, 25 Fluorescence techniques, AFM and, 39–40 Fluorescent-labeled phospholipid analogs, 122 Fluorescent lipid analogs assessing transbilayer distribution and movement of, 14–19 back-exchange assay, 15–18 consequences for intracellular trafficking of analogs for assays, 18–19 reduction assay, 15 fluorescently labeled PS and PC, phospholipid transbilayer movement in hemoglobinopathies and, 331 lipid asymmetry and, 6 with short fatty acid chain, 12–14 stopped-flow assays and, 103 structure of, 13 uptake of, into osteoblasts, 19 Förster resonance energy transfer, quenching of fluorescent analogs by, 78, 80 4,4’-diisothiocyanostilbene-2,2’disulfonate. See DIDS FRAP. See Fluorescence recovery after photobleaching Free cholesterol, ABCA1 and, 365, 367 Free fatty acids (or esters), rapid spontaneous diffusion of, xvi Free sulfhydryl (SH) groups, lipid scrambling, role of disulfide bonds and, 130–131

394 FRET. See Förster resonance energy transfer Galactolipids, 52, 53 Galactosylceramides (GalCer), 68, 69, 128 Gardos channel, 130, 301, 344 Gea2p, 161 Gea2p interaction motif, Drs2p C-tail and, 188–189 GEFs. See Guanine nucleotide exchange factors Genomic studies, “modern medicine” and, 256 GFC. See Greenland familial cholestasis Giant unilamellar vesicles, 26, 99, 244 measurement of transmembrane flip-flop of unlabeled lipids by shape changes of, 27–35 asymmetry and shape changes of membranes: examples, 28, 30 bending theory of vesicle shapes, 27–28 examples, 33–35 flip-flop detection, 30–31 perspectives, 35 theoretical model, 31–32 of pure lipid membranes, vesicle shape changes and, 278 reconstitution of energy-independent flippase activity of yeast ER into, 110 shape changes of induced by lipid asymmetry, 278–280 as tool to measure flip-flop, 106–108 Giant vesicle, made from erythrocyte membrane extract, 36 GIM. See Gea2p interaction motif GlcCer. See Glucosylceramide GlcCer transport, by ABCB1, 239 Glc3Man2GlcNac2-PP-dolichol, 115 Glucoceramide. See GlcCer Glucose depletion, 130 Glucose-phosphate-citronellol (Glc-PCit), 106 Glucosylceramide, 68, 69, 128, 369 Glutathione, 331

INDEX

Glycerol backbone of PS, flippase model and specificity for orientation of, 204 stereospecificity of PS flippase and, 205 Glycerol group, substrate-flippase interactions and, 202 Glycerolipids, saturated, plasma membrane rich in, 150 Glycerophospholipids, 100, 315 Glycerophosphorylethanolamine, specificity of substrate-transporter interactions and, 202 Glycerophosphorylserine specificity of substrate-transporter interactions and, 202 structure of, 201 Glycolipids flip-flop of, in biogenic membranes, summary of, 115 major, structure of, in plastidial membranes, 53 Glycosphingolipids, 11, 68–69, 150 Glyoxysomal membrane, lipid asymmetry in, 51–52 Golgi (Go), lipid sorting and, 150 Golgi membranes, Drs2p-dependent flippase activity in, 179–181 Gottesman, M. M., xvii GPS. See Glycerophosphorylserine Gramicidin, transbilayer movement and, 326 Gram-negative bacteria MsbA and, 230 structure of membrane, 231 Gram-positive bacteria, early investigations on transbilayer movement of phospholipids in, 111 Granulocytes, PS expressed in, 292 Grape (Vitis vinifera) aminophospholipid translocases in, 55 percentages of identity between some (putative) aminophospholipid translocases from, 57t Greenland familial cholestasis, 163 GTP. See Guanosine triphosphate GTPase-activating proteins, 184 GTPases, vesicle formation and, 161

INDEX

Guanine nucleotide exchange factors, 184, 284 Guanosine triphosphate, membrane bending and, 277 GUVs. See Giant unilamellar vesicles HDL-C. See High-density lipoprotein cholesterol HDL-CEs. See High-density lipoprotein cholesteryl esters HDLs. See High-density lipoproteins Helix insertion into membranes, curvature and, 276 Hemoglobin F, sickle-cell disease and, 332 Hemoglobinopathies lifespan of RBCs in, 328 phospholipid transbilayer movement in, 330–332 PS exposure in, 328–329 reactive oxygen species and, 318 Hemostatic response, vitamin K-dependent factors and, 342 Hepatic lipase, reverse cholesterol transport pathway and, 355 Hepatocyte nuclear factor 1, ABCA1 gene promoter and, 356 Hepatocytes, ATP8B1 as PS translocase in canalicular membrane of, 163–164 High-density lipoprotein cholesterol ABCA1 and, 360, 368 liver and macrophage ABCA1 and, 362 Tangier disease and level of, 354 High-density lipoprotein cholesteryl esters, ABCA1 overexpression and, 361 High-density lipoproteins, 227 ABCA1, lipid efflux and formation of, 362–363 ABCA1 and levels of, 359 particle formation, 89 HIV infection, of monocytes, PS exposure and, 123 HL. See Hepatic lipase HNF-3b. See Hepatocyte nuclear factor 1 hPLSCR1, 132

395 hPLSCR1-4 family of proteins, 132 htrB-null cells lipopolysaccharide and, 231 wild-type cells vs., 233 Human P-type ATPases, substrate, gene, previous symbols, approved name, and accession numbers, 321–322 Hydroxyurea treatment, sickle-cell disease and, 329, 332 Hypoparathyroidism, ATP11C linked to, 163 ICP. See Intrahepatic cholestasis of pregnancy IFNs. See Interferons IM. See Inner membrane Inflammation, apoptosis and development/resolution of, 297 Inner membrane in gram-negative bacteria, 230, 231 MsbA, lipid translocation and, 231, 233 Inner membrane vesicles, inverted, investigations on, from E. coli, 111 Inositol lipid binding, formation of amphipathic helix in epsin and, 277 Inositol-(1,4,5)-triphosphate, Ca2+ homeostasis during apoptosis and, 299 Inside-out vesicles, PIP2 in external leaflet of, 136 Interferon gamma, ABCA1 and, 356 Interferons, PLSCR and, 133 Interleukin-1β, ABCA1 transcription and, 356 Intestinal barrier, oral drugs and, 264, 266 Intestine, TAT-1 and fluid-phase endocytosis in, 161 Intracellular membranes of plants, lipid asymmetry in, 51–54 Intracellular signaling, cell biology of Scott syndrome and, 343–344 Intrahepatic cholestasis, progressive, xix, 163 Intrahepatic cholestasis of pregnancy ABCB4 mutations and, 241 ATP8B1 mutations and, 163

396 Intrinsic mitochondrial pathway, apoptosis and, 298 Inward lipid movement, aminophospholipid translocase and, 292 Ion transporting ATPases, phospholipid transporters vs., 323 IOVs. See Inside-out vesicles IP3. See Inositol-(1,4,5)-triphosphate Isoprenoid-based glycolipids, flipping of, 115 Isoprenoid-based lipids, phospholipid flipping and, 113 Jurkat cells, anti-Fas-induced PS exposure in, 131 K+, regulating outward movement of PS and role of, 300–301 KBV-1 cells, MDR, PAF and inhibition of ABCB1-mediated transport of rhodamine 123 in, 239 K+ channel blockers, PS exposure in platelets and, 126 Keratinocytes, GlcCer transport and, 69 Keto-deoxyloctulosome (Kdo)-lipid A, ATPase activity of MsbA and, 233–234 K562 cells, distribution of C6-NBD-PS in, after incubation at different temperatures, 281–283, 282 Kinases, outward PS movement and, 301–302 K+ ions, inhibiting phospholipid scrambling in erythrocytes and, 344 Lactadherin, 11, 121 Lands pathway, 318, 329 Large unilamellar vesicles, 25, 30 Lateral cholesterol dynamics, in membrane with liquid-ordered and liquid-disordered domains, 92 Lateral lipid membrane domains, cholesterol and formation of, 91 Lateral tension, different vesiculation regimes and, 279–280 Latrunculin A (LatA), 184

INDEX

LCAT. See Lecithin-cholesterol acyltransferase LDL-C. See Low-density lipoprotein cholesterol LDLs. See Low-density lipoproteins Lecithin-cholesterol acyltransferase, reverse cholesterol transport pathway and, 355 Leishmania parasites, apoptotic mimicry and invasion of macrophages by, 123 LEM3, 182 Lem3p, 159 Lipid A, lipopolysaccharide and, 231, 232 Lipid analogs exchange of, between vesicles, 106 fluorescent, lipid asymmetry and, 6 measurement of transverse diffusion of lipids and, 25 measuring transbilayer dynamics of, by selective chemical reduction, 105l techniques for determining transbilayer distribution of, 12–20 biosynthetic labeling, 20 SFVS, 20–21 spin-labeled and fluorescent lipid analogs, 12–19 Lipid asymmetry in animal membranes, 47–49 during apoptosis, 297–298 biological advantages of, xviii–xix biological functions and multiple factors relative to, 150–151 endocytosis and, 275–285 bending a membrane, 276–278 linkage between, 280–283 role of P4-ATPases in forming endocytic invaginations, 283–284, 284 shape changes of GUVs induced by, 278–280 in erythrocytes, 294 fluid-phase endocytosis and, 259 fluorescent lipid analogs and, 6 in plant membranes, 50–55, 58–59, 61 ABC transporters, 59 aminophospholipid translocase(s) in plant cells, 54–55, 58–59

INDEX

endoplasmic reticulum flippase, 59, 61 intracellular membranes, 51–54 plasma membrane (plasmalemma), 50–51 plasma membrane repair and regulation of, 305 regulation of, during apoptosis, 300 spin-labeled aminophospholipids and, 6 spontaneous generation of, factors related to, 5 Lipid bilayers, stretching vs. bending, 27 Lipid-binding sites, flippase containing two specific sites, 217 Lipid efflux, ABCA1 and, 364–366 Lipid flip-flop, xvi ATP-binding cassette (ABC) transporters and, 225–228 physiological problems related to, xvii–xix Lipid flippase, surface tension and role of, 258 Lipid membranes, summary of data on cholesterol translocation and distribution in, 83t–86t Lipid microenvironment, at membrane, ABCA1 and, 366–368 Lipid movement, by ABCB1, 237–240 Lipid packing, membrane shape and, 138–139 Lipid peroxidation, outward PS movement and, 302 Lipid polymorphism, viii Lipid probes, use of, 25 Lipid randomization, ATP8B1 dysfunction and, 164, 164 Lipids. See also Endogenous lipids; Unlabeled lipid transmembrane movement biosynthetic labeling of, 20 coagulation and role of, 121 lateral segregation vs. transversal segregation of, 5 sophisticated behaviors and multiple roles of, viii transbilayer distribution of, milestones in understanding dynamics of, 4

397 Lipid scrambling, xviii Ca2+ ions and, 124–126 free sulfhydryl groups and disulfide bonds in, 130–131 physiological importance of, 122–124 resolving mechanism of, 139 role of, in various cell types, 126–128 Lipid specificity of candidate aminophospholipid flippases, 210–213 Na+/K+-ATPase (P2c-ATPase), 211–213 PM Ca2+-ATPase (P2b-ATPase), 210–211 SERCA pump (P2a-ATPase), 211 of other PS-binding proteins, 213, 215 of phosphatidylserine-dependent proteins, summary, 215 of PS-activated proteins, 214t Lipid transbilayer distribution, detecting, assays for, 8 Lipid translocation ABCA4 and: explaining a phenotype, 228–230 fatty acid chain and, 14 MsbA and, 230–231, 233–237 understanding mechanism of, xx Lipid translocation pathways, perturbation/inhibition of, 227 Lipinski’s rules, 256 basis of, 253, 254 side effects of, 257 Lipinski’s second rule origin of, from pharmaceutical industry’s point of view, 254–257 potential application of, 264–272 cross-resistance to drugs, 270 interaction between membrane and drug transporters, 264, 266 membrane lipid asymmetry and drug transporters, 266, 268–270 solving, 257–264 definition of relevant biophysical parameters (act one), 257–258 perspectives (final act), 263–264 role of cytosolic pressure (act three), 261–263 role of fluid-phase endocytosis (act two), 258–261

398 Lipofuscin, deposits in retinal pigment epithelium, STGD and, 227, 229 Lipophilic index of drugs, 253, 256 Lipopolysaccharides assembly of, 232 biosynthesis of, gene mutations and defects in, 231 in outer membrane of gram-negative bacteria, 230 Lipoproteins, extraction of cholesterol from membranes and, 78 Liquid-ordered/disordered domains, scheme of lateral and transbilayer cholesterol dynamics in membrane with, 92 Liver ABCA1 expression and cholesterol trafficking in body, 362–363 selective overexpression of ABCA1 in, 361 transplantation of, 165 Liver disease, 163 Liver failure, 226 progressive familial intrahepatic cholestasis and, 241 Liver X receptor, ABCA1 and, 356 Long-chain radiolabeled phospholipids, measuring transbilayer redistribution of, 16–17 Low-density lipoprotein cholesterol, Tangier disease and low level of, 353 Low-density lipoproteins, Tangier disease and, 134 Low surface density of drug transporters, MW and transverse movement of drugs across membrane, 269 LPA. See Lysophosphatidic acid LPC. See Lysophosphatidylcholine LPE. See Lysophosphatidylethanolamine LPLAT family. See Lysophospholipid acyl-CoA acyltransferase family LPS/precursor translocation, MsbA and, 234 LPSs. See Lipopolysaccharides L693, ABCA1 and, 359 Lubrol WX microdomains, Apo A-1 derived from, 367

INDEX

LUVs. See Large unilamellar vesicles LXR. See Liver X receptor Lymphoid lines, scrambling rates and, 127 Lysophosphatidic acid, 330 Lysophosphatidylcholine, 91, 319 scrambling process and, 128 Lysophosphatidylethanolamine, 318 Lysophosphatidylserine, 201, 318 Lysophospholipid acyl-CoA acyltransferase family, 318 Lysophospholipids, 302, 319 Lysosomes, fusion of, during apoptosis, 303 Macrophage cholesterol homeostasis, ABCA1 and, 361 Macrophages ABCA1 expression and cholesterol trafficking in body, 362–363 PS expressed in, 292 Macular degeneration, ABCA4 mutations and, 228 Magainin, 113 Mammalian cells sphingolipid asymmetry and transmembrane translocation in, 65–71 ceramide, 68 future perspectives, 71 glycosphingolipids, 68–69 sphingomyelin, 70 sphingosine, sphingosine-1phosphate, and ceramide, 67–68 Mammalian plasma membranes, cholesterol and, 75–76 Mannose-phosphate-citronellol (Man-PCit), 106 Mast cells PLSCR1 and activation of, 327 PS expressed in, 292 MCF-7R cells, resistant to doxorubicin, influence of membrane lipid composition and, 269 M-domain, P-Type ATPase transport cycle and, 156 MDR. See Multidrug resistance MDR1, xvii, 69, 176, 237, 346

INDEX

MDR2, 237, 346 ABCB4 and, 242–243 MDR3, 49, 164 first cloning of, 240 Mean surface tension, drugs crossing cellular membranes and types of, 258 Membrane asymmetry, yeast P4-ATPases and, 173 Membrane bending, 138 cytosolic pressure and, 263 endocytosis and, 276–278 fluid-phase endocytosis and, 259 lipid scrambling and, 123 spontaneous segregation of lipids and, 5 Membrane biogenesis, cholesterol and, 75 Membrane blebbing lipid packing and, 138 prevention of autoimmune responses and, 304 Membrane budding mechanisms, fluidphase endocytosis and, 259 Membrane curvature, intracellular trafficking research on mechanisms of, 275 Membrane fusion, outward PS movement and, 302–304 Membrane lipid asymmetry in aging and apoptosis, 291–307 Ca2+ during apoptosis, 298–299 lipid asymmetry during apoptosis, 297–298 lipid asymmetry in erythrocytes, 294 pathological red cells, 297 phospholipid transporters, 292–294 red cell aging, 295, 297 regulation of lipid asymmetry during apoptosis, 300 regulation of PS externalization in RBCs, 294–295 role of kinases and phospholipases in outward PS movement, 301–302 role of K+ in regulating outward movement of PS, 300–301 role of lipid peroxidation in outward PS movement, 302

399 role of membrane fusion in outward PS movement, 302–304 significance, 304, 306 Membrane lipid asymmetry/permeability, to drugs, 253–272 Membrane lipids, cell physiology and role of, 291 Membrane mechanical properties, multidrug resistance and, 271, 272t Membrane proteins, transbilayer lipid movement controlled by, 152 Membrane recycling, cytosolic pressure and, 261–263 Membrane repair responses, lysosome to PM fusion and, 304 Membranes, physical boundaries provided by, 171 Membrane shape, lipid packing and, 138–139 Membrane sphingolipids, organization of, 65 Membrane transporters, drug resistance and, 270 Merocyanine 540 (MC540), 121, 138 Mesangial cells (human), PAF and inhibition of ABCB1-mediated transport of rhodamine 123 in, 239 Methyl-β-cyclodextrin cholesterol flip-flop and, 87 removal of cholesterol analogs and, 18 transbilayer cholesterol movement and, 81, 90 MFG-E8 (lactadherin), 121, 327 MGDG. See Monogalactosyldiacylglycerol Micrococcus luteus, PG and CL in, 111 Microparticles, 341 activated Scott syndrome hematopoietic cells and, 342 Scott syndrome, significance of membrane vesiculation and, 347 Microvesicles, 135–136, 341 Milk fat globule-EGF factor 8 (MFGE8), 11 Miltefosine, 178 Mitoxantrone resistance protein, floppase activity and, 293

400 Molecular weight of drugs, 253, 256 cross-resistance levels and, 271 multidrug resistance, membrane lipid composition and, 269 relevant biophysical parameters and, 257–258 drug transverse movement and, 265 Monogalactosyldiacylglycerol structure of, 53 thylakoid membrane lipids and, 52 transmembrane galactolipid asymmetry and, 53 Monolayer, outer to inner, transfer of PS from, 317 Monolayer area asymmetry lateral tension, different vesiculation regimes and, 279–280 shape changes in GUVs and time behavior of, 32 Monolayer surface area, vesicular transport and unbalanced changes in, 190 Monovalent cations, lipid scrambling and effects of, 126 Mosaicity, xv MPLSM. See Multiphoton laser-scanning microscopy MPs. See Microparticles MRP. See Multidrug resistance protein MRP1, 68, 368 MsbA “closed apo” state, 235, 236 lipid translocation and, 230–231, 233–237 LPS metabolism and, 231 “open apo” state, 235, 236 polyspecificity of, in binding site of, 234–235 X-ray crystallographic data on structures of, 235–236, 236 MsbA function, molecular aspects of, 233–234 Multidrug resistance in cancer, ABCB1 and, 237 membrane lipid composition in, 269 membrane mechanical properties in, 271, 272t Multidrug resistance protein, 293, 326

INDEX

Multiphoton laser-scanning microscopy, 369 Mung bean (Vigna radiata) protoplasts, phospholipid distribution in, 50–51 MurJ/MviN, 115 Muscle cell differentiation, PS exposure and, 123 MW. See Molecular weight MW cutoff, defining, Lipinski’s second rule and, 264 MXR. See Mitoxantrone resistance protein Myocardial infarction, low HDL-C and, 354 Myristoyl-(NBD-hexanoyl)-PE, 154 Mythramycin, membrane mechanical properties in MDR and, 272 NADPH, 331 Na+-free medium, PS exposure in platelets and effects of, 126 Na+/K+-ATPase epithelial junction formation, tracheal tube-size control in Drosophila and, 165 lipid specificity of candidate aminophospholipid flippases and, 211–213 N-BAR, membrane curvature and, 277 NBD measuring transbilayer dynamics and, 102 spin-labeled and fluorescent analogs with short fatty acid chain and, 14 NBD-1, monomeric ABCA1 and, 357 NBD-2, monomeric ABCA1 and, 357 NBD-labeled analogs, tracing transmembrane migration of analogs and reduction of, 105 NBD-labeled LPS, MsbA and, 234 NBD-labeled phosphatidylcholine analogs, chemical structures of, 13 NBD-labeled phospholipids, 49 Ca2+ concentration and, 125 Triton X-100-solubilized ER membrane proteins combined with eggPC and, 114 NBD moiety, cholesterol analogs and, 81

INDEX

NBD-PC translocation of, across yeast plasma membrane, 180 yeast plasma membrane flippase activities and, 177 NBD-PE loss of yeast PM-associated P4-ATPases Dnf1p and Dnf2p abolishing nonendocytic uptake of, 154 translocation of, across yeast plasma membrane, 180 yeast plasma membrane flippase activities and, 177 NBD-PS translocation of, across yeast plasma membrane, 180 yeast plasma membrane flippase activities and, 177 NBD-short-chain analogs, stopped-flow assays and, 103 Necrosis, defined, 304 Necrotic cells, absence of blebbing and increased cell volume in, 304 NEM. See N-ethylmaleimide NEO1 gene, P4-ATPases in yeast and, 172 Neo1p, xvii, 153, 161, 208, 210 vesicle-mediated protein transport and, 187 in yeast Saccharomyces cerevisiae, 172 N-ethylmaleimide, 319, 330 flippase activity and, 200 PS exposure in platelets and, 130 substrate-transporter interactions and, 202 Neurological diseases, microparticles and, 347 NF-kBs. See Nuclear factor kappa-lightchain-enhancer of activated B cells N-methylphosphatidylserine, structure of, 201 N935S, ABCA1 and, 359 N1800H, ABCA1 and, 359 Nonfluorescent derivatives, tracing transmembrane migration of lipid analogs and, 105

401 Nonlabeled lipids, “shape change” approach to inferring flip-flop rates of, 31 Non-protein-mediated PL flip-flop, molecular mechanism of, 226 NPFXDs, Drs2p C-tail and, 188 NPFXD signal, endocytosis signals in yeast and, 188 N-retinylidene-N-retinylethanolamine, 229 NrPE, Stargardt disease and, 229, 230 Nuclear factor kappa-light-chainenhancer of activated B cells, ABCA1 gene promoter and, 356 O-antigens isoprenoid-based glycolipids and assembly of, 115 lipopolysaccharide and, 231, 232 Oats, transmembrane galactolipid asymmetry and, 53 Obesity ATP10A mutations and, 323 Atp10c deficiency and, 210 mouse ortholog of ATP10A linked to, 163 Okadaic acid, 129, 130 Oligosaccharides, core, lipopolysaccharides and, 231, 232 OM. See Outer membrane Oncostatin M, ABCA1 and, 356 1,2-bis-(o-aminophenoxy)-ethaneN,N,N’,N’-tetraacetoxymethyl ester. See BAPTA-AM 1,2-dilauroyl-sn-glycero-3phosphocholine. See DLPC 1,2-distearoyl-sn-glycero-3phosphocholine. See DSPC 1,2-sn-glycero-3-phosphatidylserine, substrate-transporter interactions and, 202 1,2-sn-phosphatidyl-D-serine, structure of, 201 1,2-sn-phosphatidyl-L-serine, structure of, 201 1,2-sn-phosphatidylserine, speculative pathway for selective transport of, 205 Optimal budding radius, 259

402 Organelle membranes determination of lipid distribution in, studies of, 48 organization of lipids in, 51 Oryza sativa (rice), P4-ATPase and CDC50 gene numbers in, 160t Osteoblasts, uptake of fluorescent lipid analogs into, 19 Outer aqueous volume, shape changes, GUVs and, 32 Outer membrane in gram-negative bacteria, 230, 231 MsbA, lipid translocation and, 231, 233 Outward lipid movement, floppase and, 293–294 Outward PS movement Ca2+-dependent mechanisms and regulation of, 296 role of kinases and phospholipases in, 301–302 role of K+ in regulation of, 300–301 role of lipid peroxidation in, 302 role of membrane fusion in, 302–304 Oxidant stress, in sickle and thalassemic RBCs, 331 Oxidative damage, red blood cell aging and, 295 Oxidative phosphorylation, cyt c and transfer of electrons during, 302 Oxygen transport, by hemoglobin, oxidant stress in RBCs and, 331 Paclitaxel, ABCB4 and, 243 PAF. See Platelet-activating factor Palmitoylation, membrane binding and, 327 Palmitoylcarnitine, scrambling process and, 128 Pan1p, endocytosis signals in yeast and, 188 Papuamide B (Pap B), P4-ATPases in yeast membrane asymmetry and, 182, 183 Passive flip-flop of cholesterol, data on, unilamellar liposomes and, 81 Pathological red cells, 297 PbA. See Plasmodium berghei ANKA PC. See Phosphatidylcholine

INDEX

PDA. See Pyridyldithioethylamine PDZ-RhoGEF/LARG, ABCA1 and, 358 PE. See Phosphatidylethanolamine Peas, transmembrane galactolipid asymmetry and, 53 Peptide-binding assay, for endogenous lipids, 8, 9 Peptidoglycan, isoprenoid-based glycolipids and assembly of, 115 Peroxisome proliferator-activated receptor, ABCA1 and, 356 Peroxyredoxin, 331 PEST-like sequences, endocytosis signals in yeast and, 188 PFIC. See Progressive familial intrahepatic cholestasis PFIC1. See Progressive familial intrahepatic cholestasis type 1 Pfizer, 253, 256 P4ATPase-catalyzed lipid transport mechanism and (patho)physiology of, 149–166 mechanism of: role of accessory subunits, 156–161 giant substrate problem, 158 P-type ATPase transport cycle, 156, 158 role of Cdc50 proteins in, 159–161 P4-ATPase-catalyzed phospholipid transport, role of Cdc50 proteins in, 159–161 P4-ATPase chaperones, yeast and, 175 P4-ATPase complexes, putative substrate specificities of, 153t P4-ATPase dysfunction, future challenges related to, 166 P4-ATPase mutants defective in bulk-phase endocytosis, 162 pleiotropic phenotypes displayed by, 155 P4-ATPases, xvii, 88 in budding yeast, 172–175 nomenclature, 172 P4-ATPase chaperones, 175 characteristics of, 49 dysfunction and disease, 162–163 formation of endocytic invaginations and role of, 283–284, 284

INDEX

gene numbers in various organisms, 160t human and yeast, phylogenetic tree of, 153 oligomeric, composition of, 159 as prime candidate phospholipid translocases, 152–156 progress in molecular biology and purification of, xx role of, in vesicle-mediated protein transport, 161–162 subunits and, 323–324 transport cycle of, 156 in yeast membrane asymmetry, 182–183 PG. See Phosphatidylglycerol P-glycoprotein (Pgp), 237, 365, 368 floppase activity and, 293 transgenic mice without, blood-brain barrier disruptions and, 266 Phenolic compounds, in plant cells, 50 Phorbol esther, 130 Phosphate diester, flippase substrate recognition and, 202 Phosphatidic acid lipid asymmetry in plasma membrane and, 51 rapid spontaneous diffusion of, xvi Phosphatidylcholine, 151, 164, 259 ABCA1 and, 365, 367 asymmetrical distribution in animal membranes, 47 bidirectional movement of, in ER membrane, 50 chemical labeling and, 101 in extracellular leaflet, 171 lipid asymmetry in plasma membrane and, 51 MCF-7R cells resistant to doxorubicin, incubated with, 269 phospholipid asymmetry and, 120 RBC phospholipid organization and, 316 shape changes of red blood cells and, 28 spin-labeled, inhibition of endocytic vesiculation and addition of, 280 transbilayer movement of phospholipids in ER and flip-flop half-time of, 108

403 transport of, to inner membrane of Salmonella typhimurium, 110 vacuole and, 52 Phosphatidylethanolamine, xv, xix, 101, 150, 227, 259 aminophospholipid translocase and ATP-dependent transport of, 120 asymmetrical distribution in animal membranes, 47 chemical labeling of aminophospholipids and, 9 cinnamycin Ro09-1098 and, 11 cytosolic leaflet enriched with, 171 lipid asymmetry in plasma membrane and, 51 P4-ATPases in yeast membrane asymmetry and, 182 phospholipid asymmetry and, 119 phospholipids, chemical labeling and, 101 progressive methylation of, and reduced transport of lipids, 201 RBC phospholipid organization and, 316 studies on transbilayer distribution of endogenous lipids in plasma membrane of RBCs and, 4 transbilayer movement of phospholipids in ER and flip-flop half-time of, 108 transient externalization of, at cell division, 292 transport of, to inner membrane of Salmonella typhimurium, 110 vacuole and, 52 vesicle formation in yeast cells and, 281 yeast and sequestering of, to inner leaflet of plasma membrane, 177 Phosphatidylgalactose, polar head group size, scrambling and, 129 Phosphatidylglycerol, 10 shape changes and redistribution of endogenous lipids and, 30 transport of, to inner membrane of Salmonella typhimurium, 110 Phosphatidylhomoserine, structure of, 201 Phosphatidylhydroxypropionate, structure of, 201

404 Phosphatidylinositol transbilayer movement of phospholipids in ER and flip-flop half-time of, 108 vacuole and, 52 Phosphatidylinositol bisphosphate lipid scrambling and role for, 136–137 PS exposure on surface and, 326 Phosphatidylinositol 4,5-bisphosphate, 211, 277 Phosphatidylinositol-4-phosphate, Drs2p as downstream effector of, 209 Phosphatidylmaltose, polar head group size, scrambling and, 129 Phosphatidylmaltotriose, polar head group size, scrambling and, 129 Phosphatidylserine, xv, xix, 48, 150, 259 in activated Scott syndrome hematopoietic cells, 342 activation of coagulation cascade and, 292 aminophospholipid translocase and ATP-dependent transport of, 120 Annexin V binding and, 11 asymmetrical distribution of, in animal membranes, 47, 48 chemical labeling of aminophospholipids and, 9 cytosolic leaflet enriched with, 171 dying cells and presence of, at outer leaflet, 292 head-group-modified analogs of, substrate-transporter interactions and, 202 K+ and regulation of outward movement of, 300–301 key functions of, 341 lipid asymmetry in plasma membrane and, 51 on outer membrane leaflet, procoagulant cell surface and, 55 P4-ATPases in yeast membrane asymmetry and, 182 phospholipid asymmetry and, 119 phospholipid scrambling and critical role of, 122 Scott syndrome and candidate proteins in transmembrane redistribution of, 345–346

INDEX

sequence elements binding to, 215–216, 216t shape changes of red blood cells and, 28 studies on transbilayer distribution of endogenous lipids in plasma membrane of red blood cells and, 4 time of inward transport for, 88 transbilayer movement of phospholipids in ER and flip-flop half-time of, 108 yeast and sequestering of, to inner leaflet of plasma membrane, 177 Phosphatidylserine analogs, structures of, 201 Phosphatidylserine decarboxylase, 101 Phosphatidylserine-dependent proteins, lipid specificity of, summary, 215 Phosphatidylserine exposure consequences of, 329–330 exofacial, cell functioning and, 363 in hemoglobinopathies, 328–329 in red blood cells, 324 Phosphatidylserine-methyl ester, structure of, 201 Phospholipase A1, 100 Phospholipase A2 (enzyme) treatment, phospholipid conversion and, 8 Phospholipase assay, 100–101 Phospholipase A, 100 Phospholipase A2, 108 subpopulations of PS-exposing RBCs in SCD and, 330–331 Phospholipase-based assays, 121 Phospholipase C, 100, 108 Phospholipase D, 100 Phospholipases outward PS movement and, 301–302 Phospholipase transfer protein, reverse cholesterol transport pathway and, 355 Phospholipid asymmetry, factors contributing to, 119 Phospholipid flip-flop in biogenic membranes, 99–116 assays for measuring transbilayer distribution of endogenous phospholipids, 100–102

INDEX

chemical labeling, 101–102 phospholipase assay, 100–101 assays for measuring transbilayer distribution of phospholipid analogs, 102–106 bovine serum albumin, 102–103 exchange of lipid analogs between vesicles, 106 selective chemical reduction, 105 stopped-flow assays, 103, 105 water-soluble analogs, 106 conclusion, 115–116 flipping of isoprenoid-based glycolipids, 115 identifying phospholipid flippases, 113 rapid lipid flip-flop in biogenic membranes, 112–113 shape changes of GUVs as a tool to measure flip-flop, 106–108 transbilayer movement of phospholipids in bacterial inner membrane, 110–112 in the ER, 108–110 Phospholipid flippases discovery of, xvi identifying, efforts in, 113 Phospholipid molecular species, red cell membrane bilayer and movement of, 324 Phospholipid movement across RBC layer, factors related to, 325–326 Phospholipid randomizations, unitary nature of, evidence for, 121 Phospholipids ABCA1 and efflux of, 365 asymmetrical distribution of, across PM bilayer, 291 classes of, 199 defined, 175 Drs2p and flip-flop of, 175–176 mechanisms of cholesterol transbilayer movement and, 77 RBC phospholipid organization, 316–319 reverse cholesterol transport pathway and, 355 transport of, 88, 89

405 Phospholipid scramblase, 119–139, 326, 345 ABCA1 and, 134–135 characteristics of scrambling process, 124–131 ATP dependence, 129–130 comparison of rates between various lipids, 128–129 critical role of free sulfhydryl groups and disulfide bonds, 130–131 effects of other cations on, 126 pivotal role for Ca2+ ions, 124–126 rate of, in various cell types, 126–128 historical review, 120–122 K+ ions and inhibition of, in erythrocytes, 344 physiological importance of, 122–124 PLSCR and, 132–133 proposed candidate proteins and mechanisms, 132–139 TMEM16F and, 135–139 formation of microvesicles, 135–136 membrane shape and lipid packing, 138–139 role for aminophospholipid translocase, 137–138 role for phosphatidylinositol bisphosphate, 136–137 TMTM16F and, 327 Phospholipid transbilayer movement, in hemoglobinopathies, 330–332 Phospholipid transporters, 49–50, 292–294 aminophospholipid translocase; see Aminophospholipid translocase(s) ATB-binding cassette (ABC) transporters, 49–50 bidirectional movement: scramblase, 294 flippases of biogenic membranes, 50 inward movement: aminiphospholipid translocase, 292–293 ion transporting ATPases vs., 323 outward movement: floppase, 293–294 scramblase, 50 Phosphorylated components, scramblase mechanism and, 129 Photosynthesis, chloroplast and, 52

406 Physicochemical properties of drugs, Lipinski’s rules and, 256 Phytoceramide, 68 Phytosterols, 227 PI. See Phosphatidylinositol PIP. See Phosphatidylinositol-4-phosphate PIP2. See Phosphatidylinositol bisphosphate PI(4)P, Drs2p C-tail and, 190 PKC. See Protein kinase C PKC alpha, consensus PS-binding sequences, 216t PKC beta, consensus PS-binding sequences, 216t PKC gamma, consensus PS-binding sequences, 216t Plant cell membranes, lipid asymmetry and translocation in, 50–55, 58–59, 61 Plant sterols, ABC transporters and outward secretion of, 227 Plasma clotting factors, activating, 213 Plasma HDL-C, Tangier disease and absence of, 353 Plasmalemma, lipid asymmetry in, 50–51 Plasma lipoprotein metabolism, distinctive roles for ABCA1 in, 361 Plasma membrane ABCA1 and function of, 363–364 speed of spontaneous flip-flop in, 151 unique protein and lipid composition of, 171 Plasma membrane lipid number asymmetry, cytosolic pressure and, 261–263 Plasma membrane repair, regulation of lipid asymmetry by, 305 Plasma microparticles, ABCA1 and, 363 Plasmodium berghei ANKA, abca1 and, 363 Plastidial membranes, structure of major glycolipids found in, 53 Platelet-activating factor ABC1-mediated transport of rhodamine 123 and, 239 scrambling process and, 128 Platelets, lipid scrambling rate in, 126

INDEX

Pleckstrin homology (PH) domain, 215 PLs. See Phospholipids PLSCR. See Phospholipid scramblase PLSCR1, 132, 133, 135, 294, 327 mast cell activation and, 327 Scott cells and expression of, 346 Scott syndrome and, 326 PLSCR4, 327 PLTP. See Phospholipase transfer protein PM. See Plasma membrane PM bilayer, asymmetrical distribution of phospholipids across, 291 PM blebbing and shedding, during apoptosis, 302 PM Ca2+-ATPase (P2b-ATPase), lipid specificity of candidate aminophospholipid flippases and, 210–211 PM surface area, increase in, to accommodate generation of n particles, 303, 303 Polar head group of lipids membrane bending and size of, 5 scrambling rates and, 128, 129 Polarized epithelial cells, nonrandom lipid distributions in, 150 Polysaccharides, in plant cells, 50 Polyunsaturated fatty acids, in phospholipids, oxygen radical alterations and, 318 Pore model, rapid lipid flip-flop and, 112 PPAR. See Peroxisome proliferatoractivated receptor Progressive familial intrahepatic cholestasis ABCB4 mutations and, 240 ATP8B1 mutations and, 209 genetics and symptoms of, 241 Progressive familial intrahepatic cholestasis type 1, 163 Prokaryotic plasma membranes, flip-flop half-times and transbilayer movement in, 112 Protein-binding assay for endogenous lipids, 8, 9 transbilayer distribution of endogenous lipids in cell membranes and, 11

INDEX

Protein kinase C eryptosis and, 130 lipid specificity and, 213 lipid structural specificity of, 214t PS externalization in RBCs and, 295 Protein-mediated PS transport, possible pathway for, 204–205 Proteins cholesterol transport across membranes and role of, 88–90 fast flip-flop across the ER and role of, 108–109 Lands pathway and, 318 in plant cells, 50 transport of critical role of flippases in, 190 influence of other P4-ATPases on, 186–187 roles for Drs2p-Cdc50p in, 184–186 vesicle-mediated, 183–184 vesicle-mediated, role of P4-ATPases in, 161–162 Proteoliposomes, Drs2p and Atp8b2 flippase activity in, 181–182 Prothrombinase assay, 48 Prothrombinase complex assessment of PS on surface of RBCs and, 325 lipid structural specificity of, 214t Protoplasts, plant cell membranes and, 50 PS. See Phosphatidylserine PS-activated proteins, lipid structural specificity of, 214t PS decarboxylase, consensus PS-binding sequences, 216t PS diastereomers, head-group-modified analogs of, substrate-transporter interactions and, 202 Pseudoxanthoma elasticum, xix PS externalization critical events related to regulation of, 296 in red blood cells, regulation of, 294–295 PS flippase substrate binding site, potential, identifying essential characteristics of, 204 PS-methyl ester, 202

407 PS molecule, flippase substrate recognition and glycerol “backbone” of, 202 PS receptor, lipid structural specificity of, 214t PS-stimulated erythrocyte Mg2+-ATPase, substrate specificity of candidate aminophospholipid flippases and, 206–207 PS transport, speculative model for mechanism of, 204–205 P2A-ATPases, transport cycle of, 156 P2c-ATPases, oligomeric, composition of, 159 P-type ATPases, 152 first structure-based mechanistic model for pumping cycle of, 174 human, 321–322 membrane topology of, 157 types and classes of, 320 P-type ATPase transport cycle of, 156, 158 designation of “P-Type” and, 156 P-type pumps, transport cycle-dependent conformational changes in, 157 Puromycin, influence of membrane mechanical properties in MDR and, 272 Pyridyldithioethylamine, PS exposure and, 130 Rab8, ABCA1 and, 358 Radioactively labeled lipids, measurements of transmembrane motion and, 26 Radiolabeled phospholipids, 122 long-chain, measuring transbilayer redistribution of, 16–17 Rafts, 77, 226 lateral segregation of lipids in, 5 Raji cells, 131, 132 Random walk model two-dimensional, activation energies and their role in, 267 two-dimensional, state equation of drug bioavailability or resistance and, 268 RBC flippase, 319–320, 323–325 Rcy1p, 186

408 RDH. See Retinol dehydrogenase Reactive oxygen species, 318, 331 Reconstituted systems, lipid specificity of ATP8 family members and, 209 Red blood cell plasma membrane, components of, 315 Red blood cells, 292 aging of, 294, 295, 297 asymmetry and shape changes of membranes and, 28 Ca2+ concentration and lipid scrambling in, 125 cholesterol extraction from membranes and, 78 cholesterol flip-flop speed and, 87, 88 flippase activity first observed in, 176 lifespan of, in hemoglobinopathies, 328 mammalian, formation of, 294 membrane curvature, imbalance of phospholipid number across membrane bilayer and, 190 oxidant stress in sickle and thalassemic RBCs, 331 pathological, 297 phospholipid organization, 316–319 PS exposure in, 325 regulation of PS externalization in, 294–295 understanding lipid asymmetry in plasma membrane of, 5–6 Reduction assay, assessing transbilayer distribution/movement of spinlabeled and fluorescent lipid analogs with, 15 Reporter moiety back-exchange assay and, 16 behavior of endogenous lipids and, 4 kinetics of redistribution of spinlabeled phospholipid analogs in membranes and, 110 reduction assay and, 15 size of, avoiding/minimizing steric perturbations and, 14 Retinal pigment epithelium, STGD and lipofuscin deposits in, 228 Retinitis pigmentosa, ABCA4 mutations and, 228 Retinoid X receptor, ABCA1 and, 356 Retinol dehydrogenase, 229

INDEX

Retroendocytosis pathway, acquisition of lipids by Apo A-1 and, 365 Reverse cholesterol transport pathway, model of, 355 R5421, 131 R587W, ABCA1 and, 359, 359 RhoA, endocytosis of Na+/K+-ATPase and activation of, 283 Rhodamine 123, PAF and inhibition of ABCB1-mediated transport of, 239 Rhodopsin, 228, 229, 230 Rice (Oryza sativa) aminophospholipid translocases in, 55 percentages of identity between some (putative) aminophospholipid translocases from, 57t Ro09-0198, 182, 327 Rod cells, biochemical aspects of visual cycle in, 228 ROS. See Reactive oxygen species RPE. See Retinal pigment epithelium RRLs. See Relative resistance levels RUSH transcription factor, rabbit ATP11B and RING motifs of, 210 RXR. See Retinoid X receptor Ryanodine receptors, Ca2+ homeostasis during apoptosis and, 299 Saccharomyces cerevisiae (yeast) ARF encoding in, 184 five P4-ATPases in, 172 multiple alignment and unrooted phylogenetic tree of P4-ATPases homologs from, 153 percentages of identity between some (putative) aminophospholipid translocases from, 57t P4-ATPase and CDC50 gene numbers in, 160t sequencing of genome, P-type ATPases and, 172 Salmonella typhimurium, phospholipid transport in, 110 Sarcoplasmic reticulum, SERCA pump and, 211 Scaffolding by peripheral membrane proteins, curvature and, 276

INDEX

Scavenger receptor A, ABCA1 gene promoter and, 356 Scavenger receptor class B1 ABCA1 gene promoter and, 356 reverse cholesterol transport pathway and, 355 Scott platelets, NEM treatment of, 130 Scott syndrome, xix, 121, 124, 127, 139, 327, 341–349 ABCA1 and, 134, 135 candidate proteins in transmembrane redistribution of PS, 345–346 cell biology of, 343–345 apoptosis, 345 intracellular signaling, 343–344 defective platelet PS and, 122 diagnosis of, 343 features and phenotype, 342–343 learning from, 347–348 developmental considerations, 348 pharmacological impact, 348 microvesicle formation and, 136 normal expression of gene for PLSCR1 in, 132 overview, 341–342 PLSCR1 function and, 326 rarity of, hypothesis on, 349 significance of membrane vesiculation and derived MPs, 347 TMEM16F and, 135 transmission of, as autosomal recessive trait, 343 tyrosine phosphorylation and, 129 Scramblase, xix, 50, 120. See also Phospholipid scramblase bidirectional lipid movement and, 294 Ca2+ activation of, 124 transbilayer motion of lipids and, 7 Scramblase rates comparing, between various lipids, 128–129 quantitative measurement of, in different cells, 127 in various cell types, 126–128 SCRM-1, regulation of outward PS movement and, 294 Secretory vesicles Mdr2 and, 237

409 post-Golgi, loss of Drs2p and Dnf3p and abolition of PE asymmetry in, 155 SERCA1, P4-ATPases in yeast and, 173 SERCA pump (P2a-ATPase), lipid specificity of candidate aminophospholipid flippases and, 211 Serine, scrambling process and D-isomer form of, 128 Serine palmitoyltransferase, ABCA1 and, 358 7-nitrobenz-2-oxa-1,3-diazol-4-yl. See NBD SFVS. See Sum frequency vibrational spectroscopy SH. See Sulfhydryl “Shape change” approach advantages and shortcomings with, 35 to infer flip-flop rates of nonlabeled lipids, basis of, 31 Shape change experiments, with unlabeled ceramides, 33–35 Shape changes of GUVs flip-flop measurement and, 106–108 flip-flop of phospholipids assessed by, 107 lipid asymmetry and, 278–280 of membranes asymmetry and: examples, 28, 30 flip-flop detection by, in giant vesicles, 30–31 scanning electron micrographs of erythrocytes, 30 redistribution of phospholipids across bilayer and, 318 Short-chain lipids, ambiguous findings in experiments with, 237–239, 242 Short-chain NBD-labeled PL analogs, lipid translocation and, 237 Short-chain spin-labeled phospholipid analogs, structure of, 13 Short fatty acid chain, spin-labeled and fluorescent analogs with, 12–14 SH-reactive reagents, lipid scrambling and, 131 “Shrinkage” model, membrane shape, lipid packing and, 138–139

410 Sialic acid residues, red cell aging and, 297 Sickle-cell anemia as first molecular disease, 332 PS exposure in RBCs and, 328 Sickle-cell disease hemoglobin F and, 332 pathological red cells and, 297 PC exposure in hemoglobinopathies and, 328, 329, 330 PS exposure in RBCs and, 316 Single particle tracking, 369 Sitosterol, 227 Sitosterolemia, xix cause of, 89 mutations in ABCG5 and ABCG8 related to, 227 Slip-pop model, flippase function and, 112 SLs. See Sphingolipids SM. See Sphingomyelin Small unilamellar vesicles, xv SM to ceramide conversion, direct observation of nanoscale defects produced by, 40 Snc1p, 186 SOCE. See Store-operated Ca2+ entry Specificity protein 1 (SP1), ABCA1 gene promoter and, 356 Sperm cells, time of inward transport for, 88 Spermine, scrambling, PIP2 and, 137 Sperm plasma membrane, PS, differentiation pathways and, 123 Sphingolipid-enriched domains, lateral segregation of lipids in, 5 Sphingolipids, 49, 149, 164 kinetics of lipid scrambling and, 128 plasma membrane rich in, 150 translocation, measuring, 20 transmembrane translocation of, 66 Sphingolipid asymmetry and transmembrane translocation in mammalian cells, 65–71, 66 ceramide, 68 future perspectives, 71 glycosphingolipids, 68–69 sphingomyelin, 70 sphingosine, sphingosine-1phosphate, and ceramide, 67–68

INDEX

Sphingomyelin, xv, 70, 150, 164, 315 ABCA1 and, 367 asymmetrical distribution of, in animal membranes, 47 phospholipid asymmetry and, 120 spin-labeled, inhibition of endocytic vesiculation and addition of, 280 studies on transbilayer distribution of endogenous lipids in plasma membrane of red blood cells and, 4 transbilayer movement of phospholipids in ER and flip-flop half-time of, 108 Sphingomyelinase, 8, 138 Sphingomyelinase C, 70 Sphingomyelin/cholesterol mixtures, “shape change” approach and, 35 Sphingosine, 67 Sphingosine-1-phosphate, 67–68 Sphingosine synthesis, first step in, 67 SphK1, 68 SphK2, 68 Spinach, transmembrane galactolipid asymmetry and, 53 Spin-labeled aminophospholipids, lipid asymmetry and, 6 Spin-labeled analogs assessing transbilayer distribution and, 14–19 back-exchange assay, 15–18 consequences of intracellular trafficking of analogs for assays, 18–19 reduction assay, 15 early studies using, 12 principal structure of, 13 with short fatty acids, 12–14 Spin-labeled lipid, EPR signal of, 104 Spin-labeled phosphatidylcholine analogs, chemical structures of, 13 Spin-labeled phospholipid analogs, 122 kinetics of redistribution of, in human red blood cells at 37°C, 17 transbilayer movement of phospholipids in ER and, 109 Spin-labeled phospholipids, rapid flipflop of, assessed by stopped-flow, 104, 105

INDEX

Spin-labeled short-chain analogs, stopped-flow assays and, 103 Spin-labeling technique, invention of, 12 Spontaneous curvature model, 278 SPT. See Single particle tracking SR. See Sarcoplasmic reticulum Sr2+, lipid scrambling and, 124 SR-B1. See Scavenger receptor class B1 SREBP-2. See Sterol regulatory elementbinding protein Stargardt disease, xix ABCA4 mutations and, 227–230 State of charge, for drugs, 253, 256 Staurosporin, 129, 130 STED. See Stimulated emission depletion Steroid hormone biosynthesis, cholesterol and, 75 Sterol regulatory element-binding protein, 356 Sterols, 149, 164, 164 lipid transport and, 7 plasma membrane rich in, 150 STGD. See Stargardt disease Stimulated emission depletion, 369 Stomatocyte configurations, shape change to, 30 Stopped-flow assays, 103, 105 Store-operated Ca2+ entry, 344 Streptoverticillium griseoverticillatum, cinnamycin Ro09-0198 and, 11 Substrate specificity of candidate aminophosphalipid flippases, 205–209 ATP8, 208–209 ATP9, 210 ATP10 and ATP11, 210 P4 family of ATPases, 207–208 PS-stimulated erythrocyte Mg2+ATPase, 206–207 yeast P4-ATPases, 208 of PM aminophospholipid flippase, 200–205 Substrate transport, requirements for, 216–217 Subunits, P4-ATPases and, 323–324 Sulfhydryl, free, critical role of, in lipid scrambling, 130–131 Sulfoquinovosyldiacylglycerol, in plastidial membranes, 53

411 Sum frequency vibrational spectroscopy, 12 measuring transbilayer lipid movement with, 20–21 measuring translocation of unlabeled lipids on lipid bilayers with, 26 Summerskill syndrome, inherited mutation in ATP8B1 and, 323 Surface tensions, modulating intracellular accumulation of drugs and, 269 SUVs. See Small unilamellar vesicles SVs. See Secretory vesicles Symmetrical bilayers, vesicular shapes and, 27 Tangier disease ABCA1 and: historical notes, 353–354 cause of, 89 characterization of, xix defined and clinical manifestations of, 353–354 homozygous mutations in ABCA1 and, 360t increased intrahepatic cholesterol and, 361 model of reverse cholesterol transport pathway in, 355 mutations in ABCA1 gene and, 134, 227 TAT-1, 138 deletion in C. elegans and constitutive presence of PS in outer leaflet, 300 inward lipid movement and, 293 regulation of outward PS movement and, 294 yolk uptake in oocytes and, 161, 281 TEMPO, transmembrane diffusion of lipid analog and, 12 TEMPO-DPPC flip-flop, spin-labeled, 12 TGN. See Trans-Golgi network Thalassemia, 333 PC exposure in hemoglobinopathies and, 328 PS exposure in RBCs and, 316, 324 Thin-layer chromatography, PE detected/ separated from nonreacted PE by, 101

412 Thiol modification, transbilayer movement of aminophospholipid across RBC bilayer and, 331 Thoracoabdominal syndrome, ATP11C linked to, 163 3-ketosphinganine, 67 Thrombin, PS and formation of, 122 Thylakoid membrane, of chloroplast, 52 Thylakoids, lipid asymmetry and, 53 Tight junctions, 150 Tilted membrane lipid molecules, membrane curvature and, 278 Time-resolved, small-angle neutron scattering technique, 26 Tissue homeostasis, membrane lipids and, 291 TJs. See Tight junctions TMA-DPH. See Trimethylammoniumdiphenylhexatriene TMD-1, monomeric ABCA1 and, 357 TMD-2, monomeric ABCA1 and, 357 TMDs. See Transmembrane domains TMEM16A, 327 TMEM16F, 135, 328 TMEM16 family, members of, 327 TNBS. See Trinitrobenzene sulfonic acid TNF. See Tumor necrosis factor Tonoplast, 52 Topological asymmetry with lipids, membrane curvature and, 5 Toroidal pores, 113 Transbilayer amphipath transporter (tat-1), 138 Transbilayer cholesterol dynamics, in membrane with liquid-ordered and liquid-disordered domains, 92 Transbilayer diffusion, of unlabeled molecules, methods related to, 26 Transbilayer distribution of endogenous lipids in cell membranes, techniques, 9–11 chemical labeling, 9–10 enzymatic treatment, 10–11 protein-binding assay for endogenous lipids, 11 Transbilayer distribution of lipid analogs techniques for determining, in cell membranes, 12–20

INDEX

biosynthetic labeling, 20 SFVS, 20–21 spin-labeled and fluorescent lipid analogs, 12–19 Transbilayer lipid movement, control of, 152 Transbilayer lipid organization, new methods, and discovering essential aspects of, 4 Transbilayer movement of phospholipids, RBC phospholipid organization and, 316 Transforming growth factor-β, ABCA1 transcription and, 356 Trans-Golgi and plasma membrane, transbilayer lipid movement toward, 152 Trans-Golgi network, Drs2p localization to, 177 Translocation, clarifying precise molecular mechanism of, 244 Translocation kinetics, transbilayer distribution of cholesterol deduced from, 80 Translocon, 112 Transmembrane domains, “open apo” state in MsbA and, 236 Transmembrane PS asymmetry, maintenance of, 199–200 Transversal segregation of lipids, 5 Trimethylammonium-diphenylhexatriene, scrambling process and, 128 Trinitrobenzene sulfonic acid, 9, 101 endogenous lipid modification and, 8 P4-ATPases in yeast membrane asymmetry and, 182 phospholipid orientation and, 120–121 vacuole treated with, 52 Triton X-100-solubilized ER membrane proteins, combining with eggPC and NBD-labeled phospholipid, 114 Tryptophan fluorescence quenching, drug, lipid, and nucleotide binding and, 235 Tumor necrosis factor, apoptosis and, 298 Tumor necrosis factor alpha, ABCA1 transcription and, 356

INDEX

Two-dimensional random walk hypothesis, 267, 268 2,3-sn-glycerolphosphatidyl-D-serine, substrate-transporter interactions and, 202 2,3-sn-glycerolphosphatidyl-L-serine, substrate-transporter interactions and, 202 2,3-sn-phosphatidyl-D-serine, structure of, 201 2,3-sn-phosphatidyl-L-serine, structure of, 201 2,4,6-trinitrobenzoate, 131 Tyrosine kinase-dependent phosphorylation alterations, in Scott platelets, 344 Tyrosine phosphorylation, Ca2+-induced lipid scrambling in erythrocytes and, 129 “Umbrella model,” cholesterol and, 76 Unilamellar liposomes, data on passive flip-flop of cholesterol in membranes from, 81 Unlabeled lipids, AFM and measurement of transmembrane flip-flop of, 36–40 Unlabeled lipid transmembrane movement detection and measurement of, 25–41 flip-flop measurement by shape change of GUVs and, 27–35 asymmetry and shape changes of membranes: examples, 28, 30 bending theory of vesicle shapes, 27–28 detecting flip-flop by shape changes in giant vesicles, 30–31 examples, 33–35 perspectives, 35 theoretical model, 31–32 measurement of transmembrane flip-flop by shape change of GUVs, 27–35 using AFM, 26–40 Unlabeled phospholipids, time-resolved, small-angle neutron scattering technique and flip-flop of, 26

413 Vaccinia virus, PS exposure and, 123 Vacuole, lipid distribution in membranes and role of, 52 “Vacuum cleaner” hypothesis, Pgp-like transporters in drug availability and, 266 Valinomycin, 326 Vanadate ATP depletion, prevention of scrambling activity and, 129 flippase activity and, 200 substrate-transporter interactions and, 202 Verapamil, MsbA studies and, 234–235 Very long-chain fatty acids, X-linked adrenoleukodystrophy and, 227 Very low-density lipoproteins hepatocytes and export of, 362 reverse cholesterol transport pathway and, 355 Vesicle biogenesis, critical role of flippases in, 190 Vesicle budding critical steps in, 184 Drs2p in protein transport and, 183–191 Vesicle formation during endocytosis, membrane surface asymmetry and, 317 flippases and, 190–191 lipid translocators and, 275–276 PL flip-flop and, 226 Vesicle-mediated protein transport, 161– 162, 183–184 Vesicle shapes bending theory of, 27–28 relative area between two monolayers and, 27 Vesicle shedding, lipid asymmetry and, 136 Villin expression, blocking in Caco-2 cells, microvilli and, 165 Vinblastine ABCB4 and, 243 influence of membrane mechanical properties in MDR and, 272 MsbA studies and, 234–235 Vincristine, influence of membrane mechanical properties in MDR and, 272

414 Visual cycle, STGD and, 228–230 Vitamin E, 331 Vitamin K-dependent factors, hemostatic response and, 342 VLCFA. See Very long-chain fatty acids VLDLs. See Very low-density lipoproteins vrp1, endocytosis of Drs2p and, 187

INDEX

Xq27 region, 163 X-ray crystallographic data, structures of MsbA based on, 235, 236

WAH-1 PS exposure and, 133 regulation of outward PS movement and, 294 Water-soluble analogs, 106 Water-soluble probes, study of transmembrane diffusion and, 26 WD2, temperature-sensitive mutant, wild-type MsbA vs., 233 W590S, ABCA1 and, 359 WHAM. See Wisconsin hypo-alpha mutant Wild-type MsbA, temperature-sensitive mutant WD2 vs., 233 Wild-type (WT) cells, htrB-null cells vs., 233 Wisconsin hypo-alpha mutant, in chicken, ABCA1 deficiency and, 361 WzxE protein, 106, 115

Yeast budding, P4-ATPases in, 172–175 types of endocytosis signals in, 188 Yeast cells APLT and vesicle formation in, 281 TNBS assay and assessing ER of, 101 Yeast ER, reconstitution of energyindependent flippase activity from into GUVs, 110 Yeast membrane asymmetry, P4-ATPases in, 182–183 Yeast P4-ATPases membrane asymmetry and, 173 substrate specificity of candidate aminophospholipid flippases and, 208 Yeast plasma membrane flippase activities, 177–179 Yeast vesicles ATP8A1 expression in, 324 secretory, ABCB4 and, 242 Yolk uptake in oocytes, P4-ATPase TAT-1 required for, 161 Ysl2p, Neo1p interactions with, 210 YSVs. See Yeast secretory vesicles

X-linked adrenoleukodystrophy (X-ALD), cause of, 227

ZO-1 protein (PDZ)-binding motif, ABCA1 and, 358