Current Topics in Membranes, Volume 59
Mechanosensitive Ion Channels, Part B
Current Topics in Membranes, Volume 59 Series Editors Dale J. Benos Department of Physiology and Biophysics University of Alabama Birmingham, Alabama
Sidney A. Simon Department of Neurobiology Duke University Medical Centre Durham, North Carolina
Current Topics in Membranes, Volume 59
Mechanosensitive Ion Channels, Part B Edited by Owen P. Hamill Department of Neuroscience and Cell Biology University of Texas Medical Branch Galveston, Texas
AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Academic Press is an imprint of Elsevier
Academic Press is an imprint of Elsevier 525 B Street, Suite 1900, San Diego, California 92101-4495, USA 84 Theobald’s Road, London WC1X 8RR, UK
This book is printed on acid-free paper. Copyright # 2007, Elsevier Inc. All Rights Reserved.
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher’s consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (www.copyright.com), for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-2007 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 1063-5823/2007 $35.00 Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone: (+44) 1865 843830, fax: (+44) 1865 853333, E-mail:
[email protected]. You may also complete your request on-line via the Elsevier homepage (http://elsevier.com), by selecting ‘‘Support & Contact’’ then ‘‘Copyright and Permission’’ and then ‘‘Obtaining Permissions.’’ For information on all Elsevier Academic Press publications visit our Web site at www.books.elsevier.com ISBN-13: 978-0-12-153359-5 ISBN-10: 0-12-153359-X
PRINTED IN THE UNITED STATES OF AMERICA 07 08 09 10 9 8 7 6 5 4 3 2 1
Contents Contributors xiii Foreword xvii Previous Volumes in Series
CHAPTER 1
xix
Mechanosensitive Ion Channels of Spiders: Mechanical Coupling, Electrophysiology, and Synaptic Modulation Andrew S. French and Pa¨ivi H. Torkkeli
I. II. III. IV. V. VI.
Overview 1 Introduction 2 Types of Spider Mechanoreceptors 3 Mechanical Coupling 3 Mechanotransduction in Slit Sensilla 6 Dynamic Properties of Mechanotransduction and Action Potential Encoding 13 VII. Calcium Signaling During Transduction by Spider Mechanoreceptors 14 VIII. Synaptic Modulation of Spider Mechanoreceptors 15 IX. Conclusions 17 References 17
CHAPTER 2
Ion Channels for Mechanotransduction in the Crayfish Stretch Receptor Bo Rydqvist
I. II. III. IV. V.
Overview 21 Introduction 22 Morphology of the SRO 23 Functional Properties 24 Summary and Discussion of Future Research Directions 43 References 45 v
Contents
vi
CHAPTER 3
Mechanosensitive Ion Channels in Caenorhabditis elegans Dafne Bazopoulou and Nektarios Tavernarakis
I. II. III. IV. V. VI.
CHAPTER 4
Overview 49 Introduction 50 C. elegans Mechanosensitive Behaviors C. elegans DEG/ENaCs 55 C. elegans TRP Ion Channels 66 Concluding Remarks 72 References 73
51
Properties and Mechanism of the Mechanosensitive Ion Channel Inhibitor GsMTx4, a Therapeutic Peptide Derived from Tarantula Venom Philip A. Gottlieb, Thomas M. Suchyna, and Frederick Sachs
I. II. III. IV. V. VI.
CHAPTER 5
Overview 81 Introduction 82 Properties and Specificity of GsMTx4 85 Cellular Sites for GsMTx4 95 Potential Therapeutic Uses for GsMTx4 97 Conclusions 103 References 103
Mechanosensitive Channels in Neurite Outgrowth Mario Pellegrino and Monica Pellegrini
I. Overview 111 II. Introduction 112 III. Encoding of Guidance Cues in Axon Pathfinding 112 IV. Requirement of TRP Channels in Calcium-Dependent Axon Pathfinding 114 V. Physical Guidance Cues and Role of Mechanosensitive Ion Channels 116 VI. Ion Channels as Molecular Integrators 119 VII. Concluding Remarks 120 References 122
CHAPTER 6
ENaC Proteins in Vascular Smooth Muscle Mechanotransduction Heather A. Drummond
I. Overview 127 II. Introduction 128
Contents
vii
III. DEG/ENaC/ASIC Proteins are Members of a Diverse Protein Family Involved in Mechanotransduction 129 IV. Involvement of ENaC Proteins in Vascular Smooth Muscle Mechanotransduction 137 V. Summary and Future Directions 145 References 145
CHAPTER 7
Regulation of the Mechano-Gated K2P Channel TREK-1 by Membrane Phospholipids Jean Chemin, Amanda Jane Patel, Patrick Delmas, Fred Sachs, Michel Lazdunski, and Eric Honore
I. Overview 155 II. Introduction 156 III. TREK-1 Stimulation by Membrane Phospholipids 158 IV. TREK-1 Inhibition by Membrane Phospholipids 161 References 168
CHAPTER 8
MechanoTRPs and TRPA1 Andrew J. Castiglioni and Jaime Garcı´a-An˜overos
I. Overview 171 II. MechanoTRP Channels 174 III. Characteristics of TRPA1 Gene and Protein 175 IV. TRPA1 Expression in Mechanosensory Organs 176 V. Function of TRPA1 177 VI. Proposed Biological Roles for TRPA1 185 References 186
CHAPTER 9
TRPCs as MS Channels Owen P. Hamill and Rosario Maroto
I. Overview 191 II. Introduction 192 III. Practical Aspects of Recording MS Channels 193
Contents
viii
IV. Distinguishing Direct vs Indirect MS Channels 195 V. Extrinsic Regulation of Stretch Sensitivity 197 VI. Strategies to Identify MS Channel Proteins 197 VII. General Properties of TRPCs 198 VIII. Evidence for TRPC Mechanosensitivity 203 IX. Conclusions 215 References 218
CHAPTER 10 The Cytoskeletal Connection to Ion Channels as a Potential Mechanosensory Mechanism: Lessons from Polycystin-2 (TRPP2) Horacio F. Cantiello, Nicola´s Montalbetti, Qiang Li, and Xing-Zhen Chen
I. Overview 234 II. Introduction 235 III. Role of Actin Cytoskeletal Dynamics in PC2-Mediated Channel Function 253 IV. Identification of Actin-Binding Protein Interactions with Polycystin-2 261 V. EVect of Hydroosmotic Pressure on PC2 Channel Function: Role of the Cytoskeleton in Osmosensory Function 265 VI. The Channel–Cytoskeleton Interface: Structural–Functional Correlates 272 VII. Perspective and Future Directions 281 References 282
CHAPTER 11 Lipid Stress at Play: Mechanosensitivity of Voltage-Gated Channels Catherine E. Morris and Peter F. Juranka
I. II. III. IV.
Overview 298 The System Components 298 Big Picture Issues 301 Reversible Stretch-Induced Changes in Particular VGCs 319 V. Irreversible Stretch-Induced Gating Changes in VGCs 325 VI. Technical Issues 327 VII. Summary Comments 330 References 330
Contents
ix
CHAPTER 12 Hair Cell Mechanotransduction: The Dynamic Interplay Between Structure and Function Anthony J. Ricci and Bechara Kachar
I. II. III. IV. V. VI. VII. VIII. IX. X. XI. XII. XIII.
Overview 339 Auditory System 340 Hair Bundle Structure 341 MET Involves Mechanically Gated Channels Where are These Channels? 343 The Gating Spring Theory 344 How are the Channels Activated? 347 To Be or Not to Be Tethered 349 Characterizing Channel Properties? 351 MET Channel Pore 352 Adaptation 354 The Dynamic Hair Bundle 361 Summary and Future Directions 365 References 366
341
CHAPTER 13 Insights into the Pore of the Hair Cell Transducer Channel from Experiments with Permeant Blockers Sietse M. van Netten and Corne´ J. Kros
I. II. III. IV.
Overview 376 Introduction 376 Ionic Selectivity of the Transducer Channel 377 Permeation and Block of Mechanoreceptor Channels by FM1-43 378 V. Permeation and Block of the Hair Cell Transducer Channel by Aminoglycoside Antibiotics 382 VI. Transducer Channel Block by Amiloride and Its Derivatives 391 VII. Conclusions 394 References 396
CHAPTER 14 Models of Hair Cell Mechanotransduction Susanne Bechstedt and Jonathon Howard
I. II. III. IV. V. VI.
Overview 399 Introduction 400 Transduction Channel Properties 401 Gating 408 Active Hair Bundle Motility 415 Conclusions 418 References 418
Contents
x
CHAPTER 15 Touch Liam J. Drew, Francois Rugiero, and John N. Wood
I. II. III. IV. V. VI. VII. VIII. IX. X. XI. XII.
Overview 426 Introduction 426 Structure of Skin and Touch Receptors 427 Physiology of Mechanoreceptive Nerve Fibers 432 Quantitating Mechanical Responses in Animal Models 435 Electrophysiological Approaches to Mechanosensation in Rodents 436 Mechanosensitive Ion Channels in Cultured Sensory Neurons 437 Gating MS Ion Channels in DRG Neurons 446 Candidate Ion Channels 447 Voltage-Gated Channels and Mechanosensation 454 Indirect Signaling Between Sensory Neurons and Nonneuronal Cells 456 Conclusions 457 References 457
CHAPTER 16 Mechanosensitive Ion Channels in Dystrophic Muscle Jeffry B. Lansman
I. II. III. IV.
Overview 467 Introduction 468 MS Channel Expression During Myogenesis Permeabilty Properties of MS Channels in Skeletal Muscle 470 V. Gating 471 VI. Pharmacology 478 VII. Conclusions 481 References 482
CHAPTER 17 MscCa Regulation of Tumor Cell Migration and Metastasis Rosario Maroto and Owen P. Hamill
I. Overview 485 II. Introduction 486
469
Contents
xi
III. IV. V. VI.
DiVerent Modes of Migration 487 Ca2þ Dependence of Cell Migration 490 The Role of MscCa in Cell Migration 499 Can Extrinsic Mechanical Forces Acting on MscCa Switch on Cell Migration? 501 References 502
CHAPTER 18 Stretch-Activated Conductances in Smooth Muscles Kenton M. Sanders and Sang Don Koh
I. Overview 511 II. Introduction 512 III. Mechanosensitive Conductances that Generate Inward Currents 514 IV. Mechanosensitive Conductances that Generate Outward Currents 527 References 535
CHAPTER 19 Mechanosensitive Ion Channels in Blood Pressure-Sensing Baroreceptor Neurons Mark W. Chapleau, Yongjun Lu, and Francois M. Abboud
I. II. III. IV. V. VI.
Index 569
Overview 541 Introduction 542 BR Sensory Transduction 544 Mechanosensitive Channels in BR Neurons 548 Methodological Limitations and Challenges 558 Summary and Future Directions 560 References 561
This page intentionally left blank
Contributors Numbers in parentheses indicate the pages on which the author’s contributions begin.
Francois M. Abboud (541), The Cardiovascular Center, Department of Internal Medicine, and Department of Molecular Physiology & Biophysics, The University of Iowa Carver College of Medicine, Iowa City, Iowa 52242 Dafne Bazopoulou (49), Institute of Molecular Biology and Biotechnology, Foundation for Research and Technology, Heraklion 71110, Crete, Greece Susanne Bechstedt (399), Max‐Planck‐Institute of Molecular Cell Biology and Genetics (MPI‐CBG), 01307 Dresden, Germany Horacio F. Cantiello (233), Renal Unit, Massachusetts General Hospital East, Charlestown, Massachusetts 02129; Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115; Laboratorio de Canales Io´nicos, Departamento de Fisicoquı´mica y Quı´mica Analı´tica, Facultad de Farmacia y Bioquı´mica, Buenos Aires 1113, Argentina Andrew J. Castiglioni (171), Departments of Anesthesiology, Physiology, and Neurology, Northwestern University Institute for Neuroscience, Feinberg School of Medicine, Northwestern University, Chicago, Illinois 60611 Mark W. Chapleau (541), The Cardiovascular Center, Department of Internal Medicine, and Department of Molecular Physiology & Biophysics, The University of Iowa Carver College of Medicine, Iowa City, Iowa 52242; Veterans Affairs Medical Center, Iowa City, Iowa 52246 Jean Chemin (155), Institut de Genomique Fonctionnelle, UPR 2580 CNRS, F‐34094 Montpellier cedex 05, France Xing‐Zhen Chen (233), Department of Physiology, University of Alberta, Edmonton T6G2H7, Canada
xiii
xiv
Contributors
Patrick Delmas (155), Laboratoire de Neurophysiologie Cellulaire, Faculte de Medecine, UMR 6150 CNRS, 13916 Marseille Cedex 20, France Liam J. Drew (425), Molecular Nociception Group, Biology Department, University College London, London WC1E 6BT, United Kingdom Heather A. Drummond (127), Department of Physiology, The Center for Excellence in Cardiovascular–Renal Research, University of Mississippi Medical Center, Jackson, Mississippi 39216 Andrew S. French (1), Department of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia B3H 1X5, Canada Jaime Garcı´a‐An˜overos (171), Departments of Anesthesiology, Physiology, and Neurology, Northwestern University Institute for Neuroscience, Feinberg School of Medicine, Northwestern University, Chicago, Illinois 60611 Philip A. Gottlieb (81), The Department of Physiology and Biophysics, Center for Single Molecule Biophysics, SUNY at Buffalo, Buffalo, New York 14214 Owen P. Hamill (191, 485), Department of Neuroscience and Cell Biology, University of Texas Medical Branch, Galveston, Texas 77555 Eric Honore (155), Institut de Pharmacologie Mole´culaire et cellulaire, UMR 6097 CNRS, 06560 Valbonne, France Jonathon Howard (399), Max‐Planck‐Institute of Molecular Cell Biology and Genetics (MPI‐CBG), 01307 Dresden, Germany Peter F. Juranka (297), Neuroscience, Ottawa Health Research Institute, Ottawa Hospital, Ottawa, Ontario K1Y 4E9, Canada Bechara Kachar (339), Section of Structural Biology, National Institutes of Deafness and Communicative Disorders, Bethesda, Maryland 20892 Sang Don Koh (511), Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, Nevada 89557
Contributors
Corne´ J. Kros (375), School of Life Sciences, University of Sussex, Falmer, Brighton BN1 9QG, United Kingdom Jeffry B. Lansman (467), Department of Cellular and Molecular Pharmacology, School of Medicine, University of California, San Francisco, California 94143 Michel Lazdunski (155), Institut de Pharmacologie Mole´culaire et cellulaire, UMR 6097 CNRS, 06560 Valbonne, France Qiang Li (233), Department of Physiology, University of Alberta, Edmonton T6G 2H7, Canada Yongjun Lu (541), The Cardiovascular Center and Department of Internal Medicine, The University of Iowa Carver College of Medicine, Iowa City, Iowa 52242 Rosario Maroto (191, 485), Department of Neuroscience and Cell Biology, University of Texas Medical Branch, Galveston, Texas 77555 Nicola´s Montalbetti (233), Laboratorio de Canales Io´nicos, Departamento de Fisicoquı´mica y Quı´mica Analı´tica, Facultad de Farmacia y Bioquı´mica, Buenos Aires 1113, Argentina Catherine E. Morris (297), Neuroscience, Ottawa Health Research Institute, Ottawa Hospital, Ottawa, Ontario K1Y 4E9, Canada Amanda Jane Patel (155), Institut de Pharmacologie Mole´culaire et cellulaire, UMR 6097 CNRS, 06560 Valbonne, France Monica Pellegrini (111), Scuola Normale Superiore, Pisa, Italy Mario Pellegrino (111), Dipartimento di Fisiologia Umana ‘‘G. Moruzzi,’’ Universita` di Pisa, Pisa, Italy Anthony J. Ricci (339), Department of Otolaryngology, Stanford University, Stanford, California 94305 Francois Rugiero (425), Molecular Nociception Group, Biology Department, University College London, London WC1E 6BT, United Kingdom Bo Rydqvist (21), Department of Physiology and Pharmacology, Karolinska Institutet, SE‐171 77 Stockholm, Sweden Frederick Sachs (81, 155), The Department of Physiology and Biophysics, Center for Single Molecule Biophysics, SUNY at Buffalo, Buffalo, New York 14214
xv
xvi
Contributors
Kenton M. Sanders (511), Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, Nevada 89557 Thomas M. Suchyna (81), The Department of Physiology and Biophysics, Center for Single Molecule Biophysics, SUNY at Buffalo, Buffalo, New York 14214 Nektarios Tavernarakis (49), Institute of Molecular Biology and Biotechnology, Foundation for Research and Technology, Heraklion 71110, Crete, Greece Pa¨ivi H. Torkkeli (1), Department of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia B3H 1X5, Canada Sietse M. van Netten (375), Department of Neurobiophysics, University of Groningen, 9747AG, Groningen, The Netherlands John N. Wood (425), Molecular Nociception Group, Biology Department, University College London, London WC1E 6BT, United Kingdom
Foreword Mechanosensitive Ion Channels, Part B Owen P. Hamill Department of Neuroscience and Cell Biology, University of Texas Medical Branch, Galveston, Texas
One of the great challenges in studying mechanotransduction (MT) has been to identify the mechanisms that underlie the exquisite sensitivity and high‐frequency response of specific animal mechanotransducers—a spider detects substrate vibrations within thermal noise limits, whereas a bat generates and detects ultrasounds of frequencies up to 100 kHz in echolocating flying prey. Part B of this volume on mechanosensitive (MS) channels covers the diversity of MS channels and MT mechanisms evident in diVerent invertebrate and vertebrate mechanotransducers. The combined chapters highlight the integration of MS channels into signaling complexes that interact with ancillary structures and other channels that are critical in shaping the specific input–output relations of mechanotransducers. The opening chapters describe MT in the slit sensilla in the spider’s leg, the stretch receptor organ in crayfish muscle, and specific touch receptors in the nematode worm, Caenorhabditis elegans. The studies indicate at least two major channel families, the epithelial Naþ channel (ENaC) and the transient receptor potential (TRP) channels, are involved in MT in lower invertebrates. Subsequent chapters review the roles for ENaC and various TRP channels, and also the MS two‐ pore‐domain Kþ channels and MS voltage‐gated channels in mediating MT in mammalian cells. One of the major hurdles in studying MT has been the absence of specific agents that selectively target MS channels—the potential for the tarantula spider venom peptide GsMTx4 to serve this role is discussed in one chapter. Perhaps one of the interesting actions of GsMTx4 is that it strongly potentiates neurite outgrowth presumably via block of an MS channel that acts as a negative regulator of neurite outgrowth first demonstrated in the leech and reviewed in another chapter. Several chapters highlight diVerent aspects of the most intensely studied of all biological mechanotransducers, namely those mediating vertebrate hearing and touch. These two forms of MT have presented xvii
xviii
Foreword
the greatest challenge in identifying the membrane proteins forming MS channels, and each chapter provides new information and diVerent approaches that should help in completing this goal. The last part of the volume includes chapters that address the properties of MS channels in cell types where abnormalities in MT contribute to significant human pathologies, including the elevated stretch‐induced Ca2þ influx that contributes to muscle fiber degeneration in muscular dystrophy, abnormalities in the regulation of smooth muscle tone, and baroreception that lead to hypertension, and the alterations in MS channel functional expression that may contribute to increased tumor cell motility and invasion during cancer progression. As indicated in Part A of this volume, I would like to thank Dale Benos for his original invitation to submit the proposal to Elsevier. I would also like to thank all those involved in the production of the volume and, in particular, Phil Carpenter for his continual and patient eVorts during the compilation phase. Finally, I would like to thank all the scientists for presenting their discoveries regarding MS channels.
Previous Volumes in Series Current Topics in Membranes and Transport Volume 23 Genes and Membranes: Transport Proteins and Receptors* (1985) Edited by Edward A. Adelberg and Carolyn W. Slayman Volume 24 Membrane Protein Biosynthesis and Turnover (1985) Edited by Philip A. Knauf and John S. Cook Volume 25 Regulation of Calcium Transport across Muscle Membranes (1985) Edited by Adil E. Shamoo Volume 26 Naþ Hþ Exchange, Intracellular pH, and Cell Function* (1986) Edited by Peter S. Aronson and Walter F. Boron Volume 27 The Role of Membranes in Cell Growth and Differentiation (1986) Edited by Lazaro J. Mandel and Dale J. Benos Volume 28 Potassium Transport: Physiology and Pathophysiology* (1987) Edited by Gerhard Giebisch Volume 29 Membrane Structure and Function (1987) Edited by Richard D. Klausner, Christoph Kempf, and Jos van Renswoude Volume 30 Cell Volume Control: Fundamental and Comparative Aspects in Animal Cells (1987) Edited by R. Gilles, Arnost Kleinzeller, and L. Bolis Volume 31 Molecular Neurobiology: Endocrine Approaches (1987) Edited by Jerome F. Strauss, III, and Donald W. Pfaff
*Part of the series from the Yale Department of Cellular and Molecular Physiology. xix
xx
Previous Volumes in Series
Volume 32 Membrane Fusion in Fertilization, Cellular Transport, and Viral Infection (1988) Edited by Nejat Du¨zgu¨nes and Felix Bronner Volume 33 Molecular Biology of Ionic Channels* (1988) Edited by William S. Agnew, Toni Claudio, and Frederick J. Sigworth Volume 34 Cellular and Molecular Biology of Sodium Transport* (1989) Edited by Stanley G. Schultz Volume 35 Mechanisms of Leukocyte Activation (1990) Edited by Sergio Grinstein and Ori D. Rotstein Volume 36 Protein–Membrane Interactions* (1990) Edited by Toni Claudio Volume 37 Channels and Noise in Epithelial Tissues (1990) Edited by Sandy I. Helman and Willy Van Driessche
Current Topics in Membranes Volume 38 Ordering the Membrane Cytoskeleton Trilayer* (1991) Edited by Mark S. Mooseker and Jon S. Morrow Volume 39 Developmental Biology of Membrane Transport Systems (1991) Edited by Dale J. Benos Volume 40 Cell Lipids (1994) Edited by Dick Hoekstra Volume 41 Cell Biology and Membrane Transport Processes* (1994) Edited by Michael Caplan Volume 42 Chloride Channels (1994) Edited by William B. Guggino Volume 43 Membrane Protein–Cytoskeleton Interactions (1996) Edited by W. James Nelson Volume 44 Lipid Polymorphism and Membrane Properties (1997) Edited by Richard Epand Volume 45 The Eye’s Aqueous Humor: From Secretion to Glaucoma (1998) Edited by Mortimer M. Civan
Previous Volumes in Series
xxi
Volume 46 Potassium Ion Channels: Molecular Structure, Function, and Diseases (1999) Edited by Yoshihisa Kurachi, Lily Yeh Jan, and Michel Lazdunski Volume 47 AmilorideSensitive Sodium Channels: Physiology and Functional Diversity (1999) Edited by Dale J. Benos Volume 48 Membrane Permeability: 100 Years since Ernest Overton (1999) Edited by David W. Deamer, Arnost Kleinzeller, and Douglas M. Fambrough Volume 49 Gap Junctions: Molecular Basis of Cell Communication in Health and Disease Edited by Camillo Peracchia Volume 50 Gastrointestinal Transport: Molecular Physiology Edited by Kim E. Barrett and Mark Donowitz Volume 51 Aquaporins Edited by Stefan Hohmann, Søren Nielsen and Peter Agre Volume 52 Peptide–Lipid Interactions Edited by Sidney A. Simon and Thomas J. McIntosh Volume 53 CalciumActivated Chloride Channels Edited by Catherine Mary Fuller Volume 54 Extracellular Nucleotides and Nucleosides: Release, Receptors, and Physiological and Pathophysiological Effects Edited by Erik M. Schwiebert Volume 55 Chemokines, Chemokine Receptors, and Disease Edited by Lisa M. Schwiebert Volume 56 Basement Membrances: Cell and Molecular Biology Edited by Nicholas A. Kefalides and Jacques P. Borel Volume 57 The Nociceptive Membrane Edited by Uhtaek Oh Volume 58 Mechanosensitive Ion Channels, Part A Edited by Owen P. Hamill
CHAPTER 1 Mechanosensitive Ion Channels of Spiders: Mechanical Coupling, Electrophysiology, and Synaptic Modulation Andrew S. French and Pa¨ivi H. Torkkeli Department of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia B3H 1X5, Canada
I. II. III. IV. V.
VI. VII. VIII. IX.
Overview Introduction Types of Spider Mechanoreceptors Mechanical Coupling Mechanotransduction in Slit Sensilla A. The Ionic Selectivity of Spider Mechanosensitive Channels B. The Location of VS‐3 Mechanosensitive Channels C. Mechanosensitive Channel Conductance, Density, and pH Sensitivity D. Temperature Sensitivity of Mechanosensitive Channels E. Molecular Characterization of Spider Mechanosensitive Channels Dynamic Properties of Mechanotransduction and Action Potential Encoding Calcium Signaling During Transduction by Spider Mechanoreceptors Synaptic Modulation of Spider Mechanoreceptors Conclusions References
I. OVERVIEW Arthropods have provided several important mechanoreceptor models because of the relatively large size and accessibility of their primary sensory neurons. Three types of spider receptors: tactile hairs, trichobothria, and slit sensilla have given important information about the coupling of external Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59001-5
2
French and Torkkeli
mechanical stimuli to the neuronal membrane, transduction of mechanical force into receptor current, encoding of aVerent action potentials, and eVerent modulation of peripheral sensory receptors. Slit sensilla, found only in spiders, have been particularly important because they allow intracellular recording from sensory neurons during mechanical stimulation. Experiments on slit sensilla have shown that their mechanosensitive ion channels are sodium selective, blocked by amiloride, and open more at low pH. This evidence suggests that the channels are members of the same molecular family as degenerins, acid‐ sensitive ion channels, and epithelial sodium channels. Slit sensilla have also yielded evidence about the location, density, single‐channel conductance, and dynamic properties of the mechanosensitive channels. Spider mechanoreceptors are modulated in the periphery by eVerent neurons, and possibly by circulating chemicals. Mechanisms of modulation, intracellular signaling, and the role of intracellular calcium are areas of active investigation.
II. INTRODUCTION Humans inhabit a sensory world dominated by vision, but we also use mechanotransduction to provide the senses of hearing, vestibular sensation, touch, and vibration, as well as chemotransduction for the senses of taste and smell. In contrast to our visual world, a spider’s life is dominated by vibration and other mechanical inputs, even in those spider species that have relatively good vision. Waiting for prey to land on a web, hunting along the ground or on a plant, and negotiating a vibratory mating ritual—in all their daily activities the mechanical senses are vitally important. In addition, both humans and spiders detect a variety of internally generated mechanical signals from their musculoskeletal systems and internal organs that allow feedback regulation of movement and many internal physiological processes. Although mechanotransduction is such an important sense for humans, spiders, and most other animals, its fundamental mechanisms have been diYcult to unravel, mainly due to the small size and complex morphology of most mechanoreceptor endings. Arthropods (insects, arachnids, and crustaceans) not only possess large arrays of diVerent mechanoreceptors, but the relatively large sizes of some of their sensory neurons, and the close association of many mechanosensory neurons to the external cuticle have provided several model systems for investigating fundamental mechanisms of mechanotransduction. The most crucial step in mechanotransduction is a change in cell membrane potential, the receptor potential, produced by the application of a mechanical stimulus to the cell. To study this phenomenon ideally requires a preparation where the electrical event can be directly observed during
1. Mechanotransduction in Spiders
3
accurately controlled mechanical stimulation. This is possible in several spider preparations, and the information thus obtained will be the major subject here.
III. TYPES OF SPIDER MECHANORECEPTORS The hairiness of spiders is well known, but what are the functions of the thousands of hairs covering a typical spider? Many provide nonsensory functions. These include adhesion to the substrate via surface tension, combing of silk threads from spinnerets, supporting the air bubbles of water spiders, providing attachment sites for spiderlings clinging to a female, and deterring predators by intense skin irritation (reviewed by Foelix, 1996). However, most of the surface hairs are sensory structures. Two major types of sensory hairs are the trichobothria, or filiform hairs, and the shorter tactile hairs (Fig. 1). Each of these hair structures is innervated by multiple neurons, typically four in Cupiennius salei, although it is not clear that all these neurons are mechanically sensitive. This situation contrasts somewhat with insects, which typically have only one sensory neuron per hair, but the general structures are otherwise similar. In addition to hairs that extend beyond the cuticle, embedded in spider cuticle are numerous mechanoreceptors of a type that is not found in other arthropods, the slit sensilla (Figs. 1 and 2). These are widely distributed in the exoskeleton, including the legs, pedipalps, and body (Barth and Libera, 1970; Barth, 1985, 2001; Patil et al., 2006). They detect mechanical events in the cuticle, primarily strains imposed by normal movements of the animal and vibrations due to predators, prey, and mates. Spiders also possess a range of mechanoreceptors deeper within the animal, particularly the joint receptors and muscle receptors, but spiders apparently lack the chordotonal structures that are widespread in insects and crustaceans, serving particularly as vibration and auditory receptors (Seyfarth, 1985; Barth, 2001).
IV. MECHANICAL COUPLING The first functional stage of any mechanoreceptor is mechanical coupling from the initial stimulus to the mechanically sensitive membrane of the sensory neuron. A large contribution to overall function is suggested, although not yet proven, by the wide range of accessory structures found in mechanoreceptors of both vertebrates and invertebrates, which are assumed
4
French and Torkkeli
50 mm
Supporting cells
10 mm Cuticle
To cell bodies
Sensory dendrites Slit
Lymph space
FIGURE 1 Major types of spider cuticular mechanoreceptors. Top left: hair sensilla at the joint between the tibia (left) and the femur of a leg of Cupiennius salei. Longer, vertical hairs are trichobothria, typically about 1‐mm long, surrounded by numerous shorter tactile hairs. Top right: scanning electron micrograph of a lyriform organ consisting of approximately parallel slit sensilla from a leg of C. salei. Dark circles are the sockets of broken hair sensilla. Lower drawing shows the arrangement of sensory neurons and surrounding tissues at a typical slit sensillum. Pairs of sensory dendrites, up to 200‐m long terminate in a ciliary enlargement that leads to a tubular body surrounded by a dense dendritic sheath. Supporting cells produce a lymph space surrounding the terminal dendrites that has a diVerent ionic composition than the normal extracellular fluid. One of the two sensory dendrites proceeds further into the slit structure, but the functional reason for this diVerence is unknown. On the basis of data from Barth, 2001, 2004; Widmer et al. (2005).
to serve a mechanical coupling role. Detailed quantitative understanding of this coupling function is limited by the relatively small sizes of most receptors and the unknown mechanical properties of the materials used to construct the structures surrounding the sensory endings. The dynamic properties of coupling structures are particularly diYcult to elucidate because it is hard to
5
1. Mechanotransduction in Spiders
100 mm
Patella cuticle
Slits
Stimulator probe 500 pA
1 mm 200 ms FIGURE 2 Intracellular recording from VS‐3 neurons. The approximately tubular patella is split in two along its length and the muscle tissues removed to reveal the mechanosensory neurons lying in the hypodermal membrane. A glass microelectrode is used to penetrate the soma of a neuron while a mechanical probe is raised from below to indent the appropriate slit from the outside. Step indentations under voltage clamp produce inward receptor currents that saturate at a few micrometers. The receptor currents have an adapting component, but most of the current adapts relatively slowly and incompletely. On the basis of data from Ho¨ger et al. (1997).
measure the individual movements of each component as the sensillum is mechanically stimulated. Barth (2001, 2004) has discussed in depth the available evidence about mechanical coupling of spider trichobothria, hair sensilla, and slit sensilla. This work also builds on a substantial base of comparable studies in insect cuticular sensilla. Tactile hairs, as the name implies, are thought to serve as touch detectors. They can bend, as well as rotate within their sockets, providing a reduction of movement estimated to be about 1:750, so that relatively
6
French and Torkkeli
large external movements can be detected without damaging the hair. The longer trichobothria are specialized to detect air movements, and their varying lengths appear to be tuned to the fluid dynamics of air flow over the spider surface, especially considering the boundary layer eVect. Estimates of their sensitivity indicate that they can detect movements carrying energy equivalent to a single photon of visible light and that they operate close to the level of baseline thermal noise. They seem designed optimally to detect turbulent air flow produced by rapidly moving prey, such as flying insects, and their varying lengths and diameters provide tuning to diVerent stimulation frequencies. Slit sensilla are distributed in a wide range of patterns over the spider body, from single, isolated slits to complex arrangements of multiple slits, forming lyriform structures (Fig. 1). It is clear that slit sensilla respond to strain in the exoskeleton, produced by the animal’s movements or by vibrations conducted through the substrate. Measurements in models of spider leg cuticle indicate that the slits are optimally positioned to detect strain at the locations where it is maximized by normal loading and that slit orientations are matched to the directions of maximum natural stress. Most compound lyriform organs occur near the leg joints, while individual slits are often found at points of muscle attachment to the cuticle (Barth, 2001). The fine structure of an individual slit allows cuticular stress to apply a levered compression to the tips of the sensory dendrites. This arrangement has some similarities to the campaniform sensilla of insects, which seem to serve a similar stress‐detecting function but use singly innervated, circular structures. The varying lengths of the slits in a lyriform organ (typically 8 to 200 m long by 1 to 2 m wide) immediately suggest tuning to diVerent temporal frequencies, as in the eponymous lyre. There is some evidence that this occurs, but the varying lengths may also serve functions such as measuring the relative intensity of the strain by progressive recruitment of diVerent slits as strain increases (Barth, 2001).
V. MECHANOTRANSDUCTION IN SLIT SENSILLA Spider slit sensilla have provided important experimental preparations for research into mechanotransduction because of the following advantages. (1) Their mechanical structures, while complex, are approximately two‐ dimensional and relatively amenable to analysis and stimulation. (2) The exposed location of the sensory neurons inside the surface cuticle has allowed the development of preparations in which simultaneous mechanical stimulation and stable intracellular recording, including voltage‐clamp
1. Mechanotransduction in Spiders
7
recording can be conducted. (3) The sensory neurons are located within a hypodermal membrane that allows them to be removed from the animal intact. This has been particularly useful for studying their voltage‐activated conductances. (4) A complex eVerent innervation of the peripheral parts of sensory neurons promises to shed new light into understanding how mechanosensation is modulated. The remainder of this chapter will focus on major findings about mechanotransduction, sensory encoding, and eVerent modulation of these processes that have emerged from research on spider lyriform organs and trichobothria.
A. The Ionic Selectivity of Spider Mechanosensitive Channels Intracellular recording during mechanical stimulation has been achieved in two spider leg lyriform organs, VS‐3 on the patella (Juusola et al., 1994) and HS‐10 on the metatarsus (Gingl et al., 2006). In each case, all neurons innervating the slits were found to be mechanosensitive. Voltage‐clamp recording from the neuron cell bodies of VS‐3 revealed an inward, depolarizing receptor current with both adapting and long‐lasting components that saturated with slit indentations of about 3 m (Fig. 2). Note that the slit indentation used in these experiments does not represent a natural stimulus. Although the major functions of VS‐3 remain unclear, normal slit compression is presumably produced by cuticle strains. However, more natural stimulation of HS‐10 was achieved by moving the tarsus and this gave very similar results to the VS‐3 slit indentation. The receptor current in VS‐3 neurons could not be reversed, even with strong depolarization, and was completely eliminated when external sodium was replaced by choline (Fig. 3). Further tests with the common monovalent and divalent cations showed that, other than sodium, only lithium ions had detectable, but much lower, permeation (Ho¨ger et al., 1997). These experiments indicate that spider mechanosensitive channels are highly selective for sodium ions. Further support for this selectivity comes from measurements of the ionic composition of the solution in the lymph space that surrounds the dendrite tips (Fig. 1). Comparable insect mechanoreceptors have a high concentration of potassium ions in this region, as well as a potential that is positive compared to the normal extracellular space (Thurm and Ku¨ppers, 1980; Gru¨nert and Gnatzy, 1987), but in spiders this region not only lacks the high potassium and positive potential but also has a relatively high concentration of sodium ions (Rick et al., 1976).
8
French and Torkkeli 100 pA
Choline 100 mV
−100 mV
Control −100 pA
−200 pA
50 pA 50 ms
−300 pA FIGURE 3 Receptor current is carried by sodium ions in VS‐3 neurons. Graph shows typical peak receptor currents produced by step slit indentations of 3 m while the neuronal membrane was held at diVerent potentials. Note the failure to reverse, even at strong positive potentials. Replacement of the sodium ions in spider saline with the large choline cation completely eliminated the receptor current, but it returned when the normal saline solution was restored (control). On the basis of data from Ho¨ger et al. (1997).
B. The Location of VS‐3 Mechanosensitive Channels The bipolar structure of arthropod cuticular mechanoreceptor neurons (Fig. 2) has led to a long history of attempts to find the location of the mechanosensitive channels, as well as the location of the action potential‐initiating region. Although the obvious location for transduction would seem to be at the distal tips of the dendrites because of the close apposition to the initial mechanical stimulus and the specialized electrochemical gradient of the lymph space (Fig.1), there have also been theories that transduction occurs near the ciliary basal body and that action potentials might arise in the axosomatic region (reviewed by French, 1988). A direct test of the location of mechanotransduction was performed by applying small punctate stimuli to diVerent locations along the dendrites of VS‐3 neurons (Ho¨ger and Seyfarth, 2001). Only stimuli applied to the distal dendrites, close to the inner surface of the slits, produced electrical activity in the neurons, suggesting a distal location. The general direction of signal flow in a sensory receptor from distal to proximal implies that transduction should occur either at the site of action potential initiation or possibly distal to it. Gingl and French (2003) used several techniques to locate the site of action potential initiation in VS‐3 neurons, including the voltage jump method that measures collisions between
9
1. Mechanotransduction in Spiders
voltage waves started by the receptor potential and an artificially created potential step at the soma. These measurements all indicated that transduction and action potential initiation both start at the distal end of the dendrite. More recent work has directly observed action potentials flowing along the dendrite from the distal tips (Gingl et al., 2004). Although all these experiments support a distal location for the mechanosensitive channels, they cannot provide a more accurate position than somewhere within about 50 m from the end of the dendrite. The basal body occurs at the distal end of the dendritic enlargement in VS‐3 neurons (Fig. 1), which is close to the lymph space. More accurate localization will probably have to wait for better anatomical evidence such as antibodies to the mechanosensitive channels.
C. Mechanosensitive Channel Conductance, Density, and pH Sensitivity Single‐channel recordings of the mechanosensitive channels have not yet been achieved. Patch clamp recording from VS‐3 neurons is complicated by their location within a hypodermal membrane and extensive glial wrappings. The probable location of the channels near the tip of the sensory dendrite adds further diYculty. An alternative approach is to measure the variance, or noise, of the total receptor current to estimate the single‐channel conductance and number of channels (Traynelis and Jaramillo, 1998). This approach requires current variance measurements over a range of diVerent current amplitudes, which can be achieved by varying the stimulus used to open the channels being investigated. In VS‐3 neurons the receptor current adapts slowly after a step indentation of the slit, and this natural change in current was used to estimate the mechanosensitive channel properties. For a single group of identical ion channels, the total variance, s2, of the current flowing through a membrane is given by: s2 ¼ s20 þ IðV
EÞg
I 2=N
ð1Þ
where s02 is the background variance due to other sources, I is total membrane current, V is the voltage across the membrane, E is the equilibrium potential of the ions flowing through the channel, is the single‐channel conductance, and N is the number of channels in the membrane. Given the single‐channel conductance and number of channels, the open probability of the channels can be calculated from: Po ¼
I NðV
EÞg
ð2Þ
10
French and Torkkeli
Ho¨ger and French (1999a) showed that the mechanosensitive channels were almost completely open at the start of a step indentation, but then closed with several time constants over a period of several minutes (Fig. 4). Their single‐channel conductance estimate was about 7 pS and the number of channels per neuron was about 470. Neither of these parameters was sensitive to pH (Ho¨ger and French, 2002). However, acid conditions significantly raised the open probability of the channels, and hence the overall receptor current. From the estimated single‐channel conductance and number of channels, total mechanosensitive conductance was calculated to be about 3.5 nS in a single VS‐3 neuron. However, independent estimates of total charge flowing during a step indentation gave a significantly higher estimate of about 15 nS (Gingl and French, 2003). A possible cause of this diVerence lies in the cable properties of the sensory dendrite. The measured length constant of the sensory dendrites is about 200 m, which is comparable to the physical length of the dendrites (Gingl and French, 2003). Although the noise measurements were made at the neuronal resting potential to minimize the current requirements of the voltage clamp, it is possible that the current flowing through the mechanosensitive channels at the dendrite tip could depolarize the membrane beyond the control of the voltage clamp in the soma. This would reduce the estimated receptor current and its variance.
1.0
1.0
n = 9 n = 23 *
0.0
Popen
pH 8
pH 5
0.0 0
Time (s)
40
FIGURE 4 Noise analysis and pH sensitivity of VS‐3 receptor current. Step indentations of the slits lasting 40 s produced a slowly adapting receptor current. Noise analysis was used to estimate the number of mechanosensitive ion channels, single‐channel conductance, and channel open probability (Popen) during the step. Traces show Popen for a typical neuron at pH 8 (approximately normal conditions) and at pH 5. Inset shows mean values of Popen at 36 s after the step under normal and acid conditions. Asterisk indicates p < 0.05. On the basis of data from Ho¨ger and French (2002).
1. Mechanotransduction in Spiders
11
Therefore, the single‐channel conductance of the mechanosensitive channels could be 20 pS or more. This would be in better agreement with estimates from mammalian auditory hair cells based on single‐channel recordings, which are as high as 100 pS (Fettiplace et al., 1992).
D. Temperature Sensitivity of Mechanosensitive Channels Mechanotransduction has been found to be more thermally sensitive than would be predicted from simple ion channel conductance in a range of vertebrate and invertebrate sensory receptors (reviewed in Ho¨ger and French, 1999b). Most of these measurements were made on the action potential signals from sensory receptors so that the location of temperature sensitivity could not be clearly established. The VS‐3 organ provided the first direct measure of temperature sensitivity in the receptor current (Ho¨ger and French, 1999b). These data were well‐fitted by the Arrhenius rate equation to give a mean activation energy of 23 kcal/mol (97 kJ/mol or Q10 ¼ 3.2 at 20 C). This is the highest activation energy measured for mechanotransduction, although close to measurements in other systems (Ho¨ger and French, 1999b). It confirms the general finding that mechanotransduction involves a significant energy barrier, comparable to the energy required to break a covalent chemical bond. The reason for this relatively high activation energy is not clear but is probably associated with the mechanism that links mechanical stimulus to channel opening. It is much higher than the activation energy required for ionic movement through a water‐filled channel or for the production of action potentials by voltage‐activated ion channels.
E. Molecular Characterization of Spider Mechanosensitive Channels Two major groups of ion channel molecules have been associated with sensory mechanotransduction. Members of the transient receptor potential (TRP) family of channels have been implicated in a range of sensory functions of both vertebrates and invertebrates, including phototransduction, thermal transduction, mechanotransduction, pain, and osmosensation (Minke and Cook, 2002; Corey, 2003; Maroto et al., 2005; Montell, 2005; Dhaka et al., 2006; Kwan et al., 2006). TRP channels have been strongly linked to hearing and touch in Drosophila (Kim et al., 2003; Gong et al., 2004) and to touch in Caenorhabditis elegans (Goodman and Schwarz, 2003; Li et al., 2006). TRP1 channels have been found in vertebrate pain receptors (Kwan et al., 2006), as well as mouse, bullfrog, and zebrafish inner ear hair receptors (Corey, 2003), appearing at the same embryonic stage as sound
12
French and Torkkeli
sensitivity in mice (Lewin and Moshourab, 2004). However, a knockout mouse lacking TRP1 had an impaired response to painful stimuli but its hair cell transduction was not aVected (Kwan et al., 2006). None of the evidence yet gives clear proof that these channels are the primary source of the receptor current. The other channel family associated with mechanotransduction are the degenerin/acid‐sensitive/epithelial sodium channels (DEG/ASIC/ENaC), best known for the amiloride‐blockable epithelial sodium channels that conduct sodium flux through a wide range of epithelia (Bianchi and Driscoll, 2002). In C. elegans, two of the four proteins found only in mechanoreceptor cells are DEG molecules that have been proposed to form the core of the mechanotransduction channel, and the receptor current was carried by sodium ions (Goodman and Schwarz, 2003; Syntichaki and Tavernarakis, 2004). A DEG gene family was also associated with mechanosensitivity in Drosophila larvae (Adams et al., 1998). In rodents, several members of the DEG family have been found in dorsal root ganglia and in fine nerve endings surrounding tactile hairs (Price et al., 2000). Knockout animals for one channel, BNC1, showed reductions, but not elimination, of mechanosensation (Price et al., 2000), and none of these molecules have yet been identified in well known skin mechanoreceptors, such as Pacinian corpuscles or RuYni endings. Although the molecular evidence favors TRP channels in Drosophila mechanosensation (Kim et al., 2003), all the data from spider slit sensilla is more supportive of ASIC channels. The receptor current is highly selective for sodium and blocked by amiloride (Ho¨ger et al., 1997). Mechanosensitive channel open probability is strongly increased at low pH (Ho¨ger and French, 2002). These are all characteristic properties of ASIC channels. In contrast, TRP channels are quite strongly associated with calcium signaling, and at least some sensory TRP channels are calcium permeable (Montell, 2005), whereas spider mechanosensitive channels are probably not permeable to calcium (Ho¨ger et al., 2005). Two other commonly proposed features of sensory mechanically activated channels are heteromeric construction and connections to extracellular and intracellular structural proteins. Evidence from several preparations indicates that multiple proteins are required to form functioning eukaryotic mechanically activated channels, and this may explain the diYculty of demonstrating mechanosensitivity from proteins expressed in oocytes or other systems (Hamill and McBride, 1996; Emtage et al., 2004; Syntichaki and Tavernarakis, 2004). Mechanical connections to cytoskeletal and extracellular matrix structures have been proposed by several lines of evidence, including the amino acid sequences of proposed channel molecules (Emtage et al., 2004). It has also been argued that lipid membrane alone could not
13
1. Mechanotransduction in Spiders
provide enough force to open a protein channel (Sachs, 1997). Microtubules are often prominent in mechanoreceptor endings, and in some cases have been suggested to form a cytoskeletal anchor (Gillespie and Walker, 2001). Spider slit sensilla, like other arthropod cuticular mechanoreceptors, contain prominent arrangements of microtubules in the sensory dendrites that extend to the distal tips, but mechanotransduction in VS‐3 neurons and some insect cuticular mechanoreceptors persists after pharmacological destruction of microtubules (French, 1988; Ho¨ger and Seyfarth, 2001).
VI. DYNAMIC PROPERTIES OF MECHANOTRANSDUCTION AND ACTION POTENTIAL ENCODING Recordings of action potentials from spider tactile hairs and trichobothria show neurons that are normally silent, signaling brief touching or vibration (Barth, 2004). Slit sensilla neurons are also silent in their resting condition and respond preferentially to rapid changes. Each slit is innervated by two neurons that have diVerent dynamic properties. Type A neurons are very rapidly adapting, giving only one or two action potentials at the start of a step indentation, while Type B neurons give a longer burst of action potentials (Fig. 5). This pattern has been observed in both VS‐3 and HS‐10 lyriform organs (Seyfarth and French, 1994; Gingl et al., 2006) so it probably generalizes to most or all of the slit sensilla.
50 mV
Type A
100 ms
Type B
10 mV
FIGURE 5 Spider slit sensilla are innervated by pairs of functionally diVerent neurons. Intracellular recordings are shown from the two neuron types in a VS‐3 preparation receiving step indentations of 150‐ms duration. Upper traces show normal action potential responses from Types A (left) and B (right) neurons. Lower traces show receptor potentials produced by similar steps after action potentials were blocked by treatment with tetrodotoxin. On the basis of data from Juusola and French (1998).
14
French and Torkkeli
Recordings of the receptor current (Fig. 2) or the receptor potential (Fig. 5) do not show such strong adaptation or such a diVerence between the two neuron types (Juusola and French, 1998). Receptor potential in the Type A neurons does adapt more rapidly than in Type B neurons, but the diVerence is less dramatic than the firing behavior. This diVerence in action potential encoding can also be seen with direct electrical stimulation of the neurons, and can be explained by diVerences in the inactivation properties of the voltage‐activated sodium channels that cause the initial phase of the action potentials (Torkkeli and French, 2002). The time course of the receptor current and potential must be controlled by the combination of mechanical coupling components and mechanosensitive ion channels. However, little is known about the dynamic properties of either. Somatic measurements indicate that the receptor current decays with at least two time constants (Fig. 2), and voltage jump experiments indicated that there are larger, very transient components occurring in the distal dendrites (Gingl and French, 2003). It is possible that these diVerent time constants represent separate filtering by the mechanical components and the mechanosensitive ion channels. The existing evidence is compatible with the most parsimonious model of transduction, that is, that a single type of mechanosensitive channel is present in both Types A and B neurons.
VII. CALCIUM SIGNALING DURING TRANSDUCTION BY SPIDER MECHANORECEPTORS The membranes of VS‐3 neurons contain low‐voltage‐activated calcium‐ selective ion channels (Sekizawa et al., 2000). Measurements of intracellular calcium concentration during mechanical stimulation of the slits showed that calcium rises from a resting level of about 400 nM to a maximum level of about 2 M during rapid action potential firing (Ho¨ger et al., 2005). These experiments failed to show any change in calcium concentration without action potentials, even when there was a receptor potential of 10 mV amplitude or more, confirming that the mechanosensitive ion channels are not significantly permeable to calcium. They also failed to show any release of calcium from internal stores. The amount of calcium entering during action potential firing was compatible with the estimated conductance via voltage‐activated calcium channels. These data raise the question of what role the elevation of calcium plays during normal sensory transduction. There are no known calcium‐sensitive ion channels in VS‐3 neurons, and blockade of calcium entry does not reliably aVect action potential firing. Calcium rose by similar amounts throughout the VS‐3 neurons, but with diVerent time courses in diVerent regions
15
1. Mechanotransduction in Spiders Mid dendrite
Distal dendrite
Soma
Soma
500 nM 50 s FIGURE 6 Calcium concentration rises significantly in VS‐3 neurons when they are firing. Traces show calcium elevations in diVerent regions during stimulation at 10 action potentials per second. Resting calcium concentration was about 400 nM in all regions and the increases in diVerent regions were not significantly diVerent. However, the time course of elevation was significantly slower in the soma, as shown by the traces. On the basis of data from Ho¨ger et al. (2005).
(Fig. 6), suggesting that calcium channels are distributed throughout the cells. One possible role for calcium would be regulation of the mechanosensitive channels. Calcium ions play major roles in controlling the dynamic properties of auditory hair cells, and at least some of the time constants involved seem to depend on intracellular actions of calcium on mechanosensitive ion channels (Ricci et al., 2005).
VIII. SYNAPTIC MODULATION OF SPIDER MECHANORECEPTORS An interesting feature of arachnid mechanoreceptors is that even their most peripherally located parts receive extensive and complex eVerent innervation (Foelix, 1975; Fabian‐Fine et al., 2002), allowing an early modulation of the neuronal responses to mechanical stimuli. Several fine eVerent fibers in the legs of C. salei extend along the sensory nerves all the way to the tips of the sensory dendrites. They form many types of synaptic contacts with the sensory neurons, the glial cells that enwrap the sensory neurons, and they also synapse with other eVerents (Fabian‐Fine et al., 2002). The eVerent fibers have been shown to contain a variety of transmitters, including
‐aminobutyric acid (GABA), glutamate, acetylcholine (ACh), and octopamine (Fig. 7; Fabian‐Fine et al., 2002; Widmer et al., 2005), and the mechanosensory neurons respond to these transmitters (Panek et al., 2002; Panek
16
French and Torkkeli − Inhibitory glutamate receptor
+ Octopamine receptor
− Inhibitory GABA receptor GABAB receptor
− Inhibitory ACh receptor mACh receptor
ACh Glutamate GABA Octopamine
+
− −
?
−
?
−
AChE
− ?
Dendrite
− ?
−
− Inhibitory + Excitatory
?
?
? +
? ? ? ? − − − ? Axon
Soma
?
FIGURE 7 Schematic illustration of the arrangement of eVerent neurons and transmitter receptors on a Type A spider VS‐3 neuron based on immunocytochemical and electrophysiological evidence. The eVerent fibers contain GABA, glutamate, octopamine, and ACh. The sensory neurons have inhibitory ionotropic GABA and glutamate receptors and excitatory octopamine receptors. Type A neurons also have inhibitory ionotropic ACh receptors and they express acetylcholine esterase (AChE) activity. In addition, metabotropic GABAB and muscarinic ACh receptors are found in all VS‐3 neurons, but their physiological functions are unknown. Glutamate and mACh receptors are also present in the eVerent fibers. On the basis of data from Fabian‐Fine et al. (2002), Panek et al. (2002, 2003, 2005), Gingl et al. (2004), Panek and Torkkeli (2005), Widmer et al. (2005, 2006).
and Torkkeli, 2005; Widmer et al., 2005, 2006). In addition, antibodies against transmitter receptors labeled specific sites on the sensory neurons (Panek et al., 2003, 2005;Widmer et al., 2005, 2006; Fig. 7). GABA and glutamate both act on inhibitory ionotropic receptors that are Cl ‐gated ion channels. Although both transmitters blocked VS‐3 neurons’ responses to mechanical stimuli, GABA had a significantly stronger eVect than glutamate (Panek and Torkkeli, 2005). However, GABA only inhibited axonal action potentials while the glutamate eVect involved both dendritic and axonal action potentials and it also reduced the receptor current amplitude (Gingl et al., 2004; Panek and Torkkeli, 2005). Thus, glutamatergic eVerents may control the cellular response to mechanical stimuli at earlier stages than GABAergic eVerents. The VS‐3 neurons also have metabotropic GABAB receptors, concentrated on the most distal parts of the cell bodies and on the dendrites (Panek et al., 2003). Agonists of these receptors modulated voltage‐activated calcium and potassium currents, allowing a longer lasting modulatory eVect.
1. Mechanotransduction in Spiders
17
Application of octopamine, the invertebrate analogue of noradrenaline, enhanced trichobothria neuron sensitivity to mechanical stimuli (Widmer et al., 2005). Immunocytochemical evidence indicated that one eVerent fiber containing octopamine innervated each mechanosensory neuron in the spider leg and that octopamine receptors were concentrated at and close to the axon hillock (Widmer et al., 2005). These findings suggest that octopamine acts as a transmitter rather than a neurohormone on spider mechanoreceptors, controlling each sensory neuron individually. These recent findings only unravel a small part of the complex synaptic mechanisms that control the sensitivity and gain of spider mechanosensory neurons. For example, we still know very little about the cholinergic innervation that involves both muscarinic ACh receptors and ionotropic inhibitory receptors and is distinctly diVerent in the two diVerent types of VS‐3 neurons. IX. CONCLUSIONS Spider mechanoreceptors have yielded a great deal of information about their mechanosensitive ion channels and their mechanisms of activation and modulation. However, much remains to be discovered. The electrophysiological data from slit sensilla suggest that the channel molecules are related to ASIC channels and they are probably located near the tips of the sensory dendrites. The relatively low numbers of channel molecules per cell are one reason why molecular characterization has so far proved elusive as it has in other mechanoreceptor systems. However, the spider preparations should continue to provide useful models for identifying the molecular basis of mechanosensation and this knowledge can be expected to assist the broader investigation of this crucial sense in animals and humans. Acknowledgments We thank Ewald Gingl, Ulli Ho¨ger, Mikko Juusola, Izabela Panek, Shannon Meisner, Ernst‐ August Seyfarth, and Alexandre Widmer for all their contributions to work described here. Research in our laboratories has been funded by the Canadian Institutes of Health Research, the Natural Sciences and Engineering Council of Canada, NATO, the Canadian Foundation for Innovation, the Nova Scotia Research and Innovation Trust, and the Dalhousie Medical Research Foundation.
References Adams, C. M., Anderson, M. G., Motto, D. G., Price, M. P., Johnson, W. A., and Welsh, M. J. (1998). Ripped pocket and pickpocket, novel Drosophila DEG/ENaC subunits expressed in early development and in mechanosensory neurons. J. Cell Biol. 140, 143–152. Barth, F. G. (1985). Slit sensilla and the measurement of cuticular strains. In ‘‘Neurobiology of Arachnids’’ (F. G. Barth, ed.), pp. 162–188. Springer‐Verlag, Berlin.
18
French and Torkkeli
Barth, F. G. (2001). ‘‘A Spider’s World: Senses and Behavior.’’ Springer‐Verlag, Berlin. Barth, F. G. (2004). Spider mechanoreceptors. Curr. Opin. Neurobiol. 14, 415–422. Barth, F. G., and Libera, W. (1970). Ein Atlas der Spaltsinnesorgane von Cupiennius salei Keys. Chelicerata (Araneae). Z. Morph. Tiere. 68, 343–369. Bianchi, L., and Driscoll, M. (2002). Protons at the gate: DEG/ENaC ion channels help us feel and remember. Neuron 34, 337–340. Corey, D. P. (2003). New TRP channels in hearing and mechanosensation. Neuron 39, 585–588. Dhaka, A., Viswanath, V., and Patapoutian, A. (2006). TRP ion channels and temperature sensation. Annu. Rev. Neurosci. 29, 135–161. Emtage, L., Gu, G., Hartwieg, E., and Chalfie, M. (2004). Extracellular proteins organize the mechanosensory channel complex in C. elegans touch receptor neurons. Neuron 44, 795–807. Fabian‐Fine, R., Seyfarth, E.‐A., and Meinertzhagen, I. A. (2002). Peripheral synaptic contacts at mechanoreceptors in arachnids and crustaceans: Morphological and immunocytochemical characteristics. Microsc. Res. Tech. 58, 283–298. Fettiplace, R., Crawford, A. C., and Evans, M. G. (1992). The hair cell’s mechanoelectrical transducer channel. Ann. NY Acad. Sci. 656, 1–11. Foelix, R. F. (1975). Occurrence of synapses in peripheral sensory nerves of arachnids. Nature 254, 146–148. Foelix, R. F. (1996). ‘‘Biology of Spiders.’’ Oxford University Press, New York. French, A. S. (1988). Transduction mechanisms of mechanosensilla. Annu. Rev. Entomol. 33, 39–58. Gillespie, P. G., and Walker, R. G. (2001). Molecular basis of mechanosensory transduction. Nature 413, 194–202. Gingl, E., and French, A. S. (2003). Active signal conduction through the sensory dendrite of a spider mechanoreceptor neuron. J. Neurosci. 23, 6096–6101. Gingl, E., French, A. S., Panek, I., Meisner, S., and Torkkeli, P. H. (2004). Dendritic excitability and localization of GABA‐mediated inhibition in spider mechanoreceptor neurons. Eur. J. Neurosci. 20, 59–65. Gingl, E., Burger, A. M., and Barth, F. G. (2006). Intracellular recording from a spider vibration receptor. J. Comp. Physiol. A 192, 551–558. Gong, Z., Son, W., Chung, Y. D., Kim, J., Shin, D. W., McClung, C. A., Lee, Y., Lee, H. W., Chang, D. J., Kaang, B. K., Cho, H., Oh, U., et al. (2004). Two interdependent TRPV channel subunits, inactive and Nanchung, mediate hearing in Drosophila. J. Neurosci. 24, 9059–9066. Goodman, M. B., and Schwarz, E. M. (2003). Transducing touch in Caenorhabditis elegans. Annu. Rev. Physiol. 65, 429–452. Gru¨nert, U., and Gnatzy, W. (1987). Kþ and Caþþ in the receptor lymph of arthropod cuticular receptors. J. Comp. Physiol. A 161, 329–333. Hamill, O. P., and McBride, D. W. (1996). A supramolecular complex underlying touch sensitivity. Trends Neurosci. 19, 258–261. Ho¨ger, U., and French, A. S. (1999a). Estimated single‐channel conductance of mechanically‐ activated channels in a spider mechanoreceptor. Brain Res. 826, 230–235. Ho¨ger, U., and French, A. S. (1999b). Temperature sensitivity of transduction and action potential conduction in a spider mechanoreceptor. Pflu¨gers Arch. 438, 837–842. Ho¨ger, U., and French, A. S. (2002). Extracellular acid increases the open probability of transduction channels in spider mechanoreceptors. Eur. J. Neurosci. 16, 2311–2316. Ho¨ger, U., and Seyfarth, E.‐A. (2001). Structural correlates of mechanosensory transduction and adaptation in identified neurons of spider slit sensilla. J. Comp. Physiol. A 187, 727–736.
1. Mechanotransduction in Spiders
19
Ho¨ger, U., Torkkeli, P. H., Seyfarth, E.‐A., and French, A. S. (1997). Ionic selectivity of mechanically activated channels in spider mechanoreceptor neurons. J. Neurophysiol. 78, 2079–2085. Ho¨ger, U., Torkkeli, P. H., and French, A. S. (2005). Calcium concentration changes during sensory transduction in spider mechanoreceptor neurons. Eur. J. Neurosci. 22, 3171–3178. Juusola, M., and French, A. S. (1998). Adaptation properties of two types of sensory neurons in a spider mechanoreceptor organ. J. Neurophysiol. 80, 2781–2784. Juusola, M., Seyfarth, E.‐A., and French, A. S. (1994). Sodium‐dependent receptor current in a new mechanoreceptor preparation. J. Neurophysiol. 72, 3026–3028. Kim, J., Chung, Y. D., Park, D. Y., Choi, S., Shin, D. W., Soh, H., Lee, H. W., Son, W., Yim, J., Park, C. S., Kernan, M. J., and Kim, C. (2003). A TRPV family ion channel required for hearing in Drosophila. Nature 424, 28–29. Kwan, K. Y., Allchorne, A. J., Vollrath, M. A., Christensen, A. P., Zhang, D. S., Woolf, C. J., and Corey, D. P. (2006). TRPA1 contributes to cold, mechanical, and chemical nociception but is not essential for hair‐cell transduction. Neuron 50, 277–289. Lewin, G. R., and Moshourab, R. (2004). Mechanosensation and pain. J. Neurobiol. 61, 30–44. Li, W., Feng, Z., Sternberg, P. W., and Xu, X. Y. (2006). A C. elegans stretch receptor neuron revealed by a mechanosensitive TRP channel homologue. Nature 440, 684–687. Maroto, R., Raso, A., Wood, T. G., Kurosky, A., Martinac, B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nat. Cell Biol. 7, 179–185. Minke, B., and Cook, B. (2002). TRP channel proteins and signal transduction. Physiol. Rev. 82, 429–472. Montell, C. (2005). Drosophila TRP channels. Pflugers Arch. 451, 19–28. Panek, I., and Torkkeli, P. H. (2005). Inhibitory glutamate receptors in spider peripheral mechanosensory neurons. Eur. J. Neurosci. 22, 636–646. Panek, I., French, A. S., Seyfarth, E.‐A., Sekizawa, S.‐I., and Torkkeli, P. H. (2002). Peripheral GABAergic inhibition of spider mechanosensory aVerents. Eur. J. Neurosci. 16, 96–104. Panek, I., Meisner, S., and Torkkeli, P. H. (2003). The distribution and function of GABAB receptors in spider peripheral mechanosensilla. J. Neurophysiol. 90, 2571–2580. Panek, I., Meisner, S., and Torkkeli, P. H. (2005). Glutamate acts on inhibitory receptors on spider peripheral mechanoreceptors. Soc. Neurosci. Abstr. 31, 296.7. Patil, B., Prabhu, S., and Rajashekhar, K. P. (2006). Lyriform slit sense organs on the pedipalps and spinnerets of spiders. J. Biosci. 31, 75–84. Price, M. P., Lewin, G. R., McIlwrath, S. L., Cheng, C., Xie, J., Heppenstall, P. A., Stucky, C. L., Mannsfeldt, A. G., Brennan, T. J., Drummond, H. A., Qiao, J., Benson, C. J., et al. (2000). The mammalian sodium channel BNC1 is required for normal touch sensation. Nature 407, 1007–1011. Ricci, A. J., Kennedy, H. J., Crawford, A. C., and Fettiplace, R. (2005). The transduction channel filter in auditory hair cells. J. Neurosci. 25, 7831–7839. Rick, R., Barth, F. G., and Pawel, A. V. (1976). X‐ray microanalysis of receptor lymph in a cuticular arthropod sensillum. J. Comp. Physiol. 110, 89–95. Sachs, F. (1997). Mechanical transduction by ion channels: How forces reach the channel. In ‘‘Cytoskeletal Regulation of Membrane Function’’ (S. C. Froehner, ed.), pp. 209–218. Rockefeller University Press, New York. Sekizawa, S., French, A. S., and Torkkeli, P. H. (2000). Low‐voltage‐activated calcium current does not regulate the firing behavior in paired mechanosensory neurons with diVerent adaptation properties. J. Neurophysiol. 83, 746–753. Seyfarth, E.‐A. (1985). Spider proprioception: Receptors, reflexes, and control of locomotion. In ‘‘Neurobiology of Arachnids’’ (F. G. Barth, ed.), pp. 230–248. Springer‐Verlag, Berlin.
20
French and Torkkeli
Seyfarth, E.‐A., and French, A. S. (1994). Intracellular characterization of identified sensory cells in a new spider mechanoreceptor preparation. J. Neurophysiol. 71, 1422–1427. Syntichaki, P., and Tavernarakis, N. (2004). Genetic models of mechanotransduction: The nematode Caenorhabditis elegans. Physiol. Rev. 84, 1097–1153. Thurm, U., and Ku¨ppers, J. (1980). Epithelial physiology of insect sensilla. In ‘‘Insect Biology in the Future’’ (M. Locke and D. Smith, eds.), pp. 735–763. Academic Press, New York. Torkkeli, P. H., and French, A. S. (2002). Simulation of diVerent firing patterns in paired spider mechanoreceptor neurons: The role of Naþ channel inactivation. J. Neurophysiol. 87, 1363–1368. Traynelis, S. F., and Jaramillo, F. (1998). Getting the most out of noise in the central nervous system. Trends Neurosci. 21, 137–145. Widmer, A., Ho¨ger, U., Meisner, S., French, A. S., and Torkkeli, P. H. (2005). Spider peripheral mechanosensory neurons are directly innervated and modulated by octopaminergic eVerents. J. Neurosci. 25, 1588–1598. Widmer, A., Panek, I., Ho¨ger, U., Meisner, S., French, A. S., and Torkkeli, P. H. (2006). Acetylcholine receptors in spider peripheral mechanosensilla. J. Comp. Physiol. A 192, 85–95.
CHAPTER 2 Ion Channels for Mechanotransduction in the Crayfish Stretch Receptor Bo Rydqvist Department of Physiology and Pharmacology, Karolinska Institutet, SE‐171 77 Stockholm, Sweden
I. II. III. IV.
Overview Introduction Morphology of the SRO Functional Properties A. General Behavior B. Viscoelastic Properties of the Receptor Muscles C. MSCs in the Receptor Neurons D. Macroscopic Receptor Currents in the Stretch Receptor Neurons E. Pharmacology of the Crayfish MSCs F. Voltage‐Gated Ion Channels and the Generation of Impulse Response G. Adaptation: A Multifactor Property V. Summary and Discussion of Future Research Directions References
I. OVERVIEW Mechanosensitivity is found in almost every cell in all organisms from bacteria to vertebrates and covers a wide spectrum of function from osmosensing to mechanical sensing in the specialized receptors like the hair cells of the cochlea. The molecular substrate for such mechanosensitivity is thought to be mechanosensitive ion channels (MSCs). Since most development regarding the molecular aspects of the MSC has been made in nonsensory or sensory systems which have not been accessible to recordings from ion channels, it is important to focus on mechanosensitivity of sensory organs where their functional importance is undisputed. The stretch receptor organ Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59002-7
22
Bo Rydqvist
(SRO) of the crustaceans is a suitable preparation for such studies. Each organ contains two receptors: one slowly and one rapidly adapting receptor neurons. The primary mechanosensitivity is generated by two types of MSC of hitherto unknown molecular type located in the neuronal dendrites, which are inserted into a receptor muscle fiber. In addition to the MSCs, the neurons contain voltage‐gated Naþ channels which seem to be diVerently located in the slowly and rapidly adapting neurons. Finally, at least three types of voltage‐gated Kþ channels are present in the sensory neurons, the location of which is not known. The spatial distribution of ion channels and the kinetics of the channels, together with the viscoelastic properties of the receptor muscles, determine the overall transducer properties and impulse firing of the two receptor neurons including their typical adaptive characteristics.
II. INTRODUCTION The crustacean SRO has been a major preparation for the study of mechanotransduction both on the macroscopic and on the ion channel levels. The SRO is considered to be an organ analogous to the mammalian muscle spindle organ that is instrumental for proper skeletal muscle function. The receptor organ was first described in the lobster by the Polish‐British zoologist Alexandrowicz (1951, 1967). Later, Florey and Florey (1955) described the same type of organ in the crayfish (Astacus fluviatilis presently named Astacus astacus). Identical and similar muscle receptor organs can also be found in a number of other invertebrate phyla such as Mollusca, Chelicerata, and Uniramia (for a review see Rydqvist, 1992). The importance of this organ, and its accessibility relative to the human muscle spindle, and mechanotransduction in general, was soon acknowledged and triggered a number of electrophysiological studies in several laboratories (Wiersma et al., 1953; KuZer, 1954; Eyzaguirre and KuZer, 1955a,b; Edwards and Ottoson, 1958). The mechanosensory neurons of the SRO of the crayfish are of the nonciliated type and are diVerent from the ciliated type represented by the classical hair cells in the hearing organs. Most investigations regarding the molecular aspects of the MSC have been made in nonsensory systems, and it is thus important to focus on mechanosensitivity of sensory organs where the functional importance of these channels is undisputed. The SRO of the crustaceans is a suitable preparation for such studies. The SRO is experimentally accessible to mechanical manipulation and electrophysiological recordings using intracellular microelectrodes or patch clamp techniques for ion channel analysis, although the latter technique is not without problems since the sensory neuron is covered by supporting glial cells. It is,
2. Stretch Receptor Mechanotransduction
23
however, relatively easy to inject substances into the neuron, which makes the neuron accessible to measurements using fluorescent probes. In the present chapter, I have focused on the overall function of the SRO in the crayfish stretch receptor including results obtained with structural techniques, classical electrophysiology, and patch clamp techniques. The main emphasis will be on the mechanotransduction processes and the ion channels involved in the SRO of the species A. astacus, Pacifastacus leniusculus, Procambarus clarkii, and Orconectes limosus.
III. MORPHOLOGY OF THE SRO Since the crayfish stretch receptor is a genuine mechanosensory organ with several ion channels involved in the overall mechanotransduction, a brief description of the SRO seems relevant. The SRO has two sensory neurons, each connected to a receptor muscle, located in the extensor muscles of the abdomen (Florey and Florey, 1955; Purali, 2005). The sensory neuron is of the multipolar type (Fig. 1) with its dendrites inserted into the central (intercalated) part of the receptor muscle which consists of only one muscle cell (Tao‐Cheng et al., 1981). The receptor muscles insert on consecutive segments and the aVerent axons from the neurons join the dorsal segmental nerve to the ventral ganglion. The SROs also receive eVerent innervations: (1) one or two motor axons to the receptor muscle cell and (2) two or three accessory axons conveying inhibitory signals to the receptor muscle and the sensory neurons (Alexandrowicz, 1951, 1967; Elekes and Florey, 1987a,b). Functionally, the receptors are activated when stretched by flexion of the abdomen or contraction of the receptor muscles (KuZer, 1954) and the SROs are involved in the control of the extensor muscles. In the crayfish, both the slowly and the rapidly adapting receptor muscles consist of a single muscle fiber that is divided by invagination of the cell membrane into numerous cytoplasmic processes in the central region of the muscle, the intercalated tendon, which is mainly made up of collagen. Some of the myofibrils insert in the intercalated tendon but some pass this region. The slowly adapting muscle is in the order of 30–80 mm in the central region but considerably thinner in the distal ends. The rapidly adapting muscle has a more even diameter and is thicker 70–150 mm (Komuro, 1981). The sensory neurons are large (30–100 mm) multiterminal cells of mainly pyramidal or fusiform shape. They contain a nucleus (ca. 10 mm) with a clear nucleolus. The dendrites branch about four to five times and intermingle with the connective tissue and muscle strands in the intercalated tendon. The fine terminal branches are about 2‐mm long and about 0.l mm in diameter and are devoid of mitochondria (Tao‐Cheng et al., 1981). The axon is in the
24
Bo Rydqvist
FIGURE 1 (A) Abdomen and thorax of crayfish. (B) Slowly (top) and rapidly (bottom) adapting receptor with typical action potential firing pattern as a result of a ramp and hold extension of the receptor muscle. (C) Confocal microscopic image of a slowly adapting neuron injected with Fluo‐4 (Bruton and Rydqvist, unpublished data), a recurrent fiber is present at left. (D) Drawing of stretch receptor neuron with proposed channel distribution. (A, B) Adapted with kind permission from Springer Science and Business Media (Rydqvist, 1992).
order of 30 mm in diameter. The receptor neurons of the crayfish have several layers of sheet cells that surround them except for the dendritic tips. The fine structure of the inhibitory synapses has been investigated by several authors (Elekes and Florey, 1987a,b) using serial sectioning and immunohistochemical technique, which have revealed a complex array of GABAergic inhibitory synapses on the axon, neuron, and muscle fibers and also reciprocal synapses on the inhibitory axon.
IV. FUNCTIONAL PROPERTIES A. General Behavior The receptors are activated (stretched) by flexion of the abdomen or contraction of the receptor muscle. The receptors are involved in the motor control of the abdominal muscles and the physiological range is up to 40% of resting length (Alexandrowicz, 1951). The first measurements from the
25
2. Stretch Receptor Mechanotransduction
receptor neuron were done by Wiersma et al. (1953) and KuZer (1954) and subsequent studies by Eyzaguirre and KuZer (1955a,b) and Edwards and Ottoson (1958) on lobster and crayfish showed that stretching the receptor organs gave rise to a distinctive pattern of impulse discharge from the neurons. It was found that the firing properties of the two neurons were clearly diVerent, one neuron maintained firing as long as the stretch was applied (slowly adapting) whereas the other neuron generated a short high frequency discharge (rapidly adapting) at the onset of the stretch (Fig. 1B). The chain of events that leads from extension of the receptor muscle to action potential generation in the stretch receptor is represented by the steps outlined in Fig. 2. In a first step, the extension of the receptor muscle is
Action potential Na+
50 mV
Receptor potential TTX Receptor potential TTX, 4AP, TEA
Voltage-gated channels K+
cm gL
Passive membrane
SA-channels
Dendrites
Viscoelastic
Receptor muscle
Receptor current
100 nA
100 kPa
Muscle tension Extension 100 ms
FIGURE 2 Transduction processes in a stretch receptor neuron. Left: recorded responses of muscle tension, receptor current, receptor potential, and action potentials in response to a ramp and hold extension of the muscle. The receptor potential is seen both after block of Naþ channels with tetrodotoxin (TTX) and after additional block of Kþ channels with tetraethylammonium chloride (TEA) and 4‐aminopyridine (4‐AP). Right: functional blocks in transduction. Stretch‐activated channels; SA channels, MSC. Adapted from Swerup and Rydqvist, 1992 with permission from Elsevier.
26
Bo Rydqvist
converted to tension in the muscle, which leads to deformation of the dendritic membrane of the sensory neuron. This opens nonselective mechanosensitive (gated) ion channels (MSCs) permeable to Naþ, Kþ, and Ca2þ ions producing an inward generator current (Erxleben, 1989). The transformation from generator current to impulse response is a complex process determined by the passive membrane properties, that is capacitance (cm) and membrane resistance (rm or leak conductance gL), and the voltage‐gated ion conductances (gion) present in the neuron. At present only voltage‐gated Kþ and Naþ channels and a Ca2þ‐activated Kþ channel have been observed. Figure 2 shows the receptor potential after block by tetrodotoxin (TTX), 4‐aminopyridine (4‐AP), and tetraethylammonium chloride (TEA), and the receptor potential after block with TTX only. In addition, the geometry of the cell and the spatial distribution of the diVerent ion channels will contribute to the type of impulse response seen in the cell. The diVerence in responses reflects the relativity of the concept of receptor potential (see discussion in Swerup and Rydqvist, 1992).
B. Viscoelastic Properties of the Receptor Muscles In most mechanosensory cells, the accessory structures contribute to the overall behavior of the transducer function. In particular, it has been discussed to what extent the accessory structures contribute to the adaptation in sensory cells. This is obvious in, for example, the Pacinian corpuscle. The two neurons in the SRO have diVerent adaptive properties and this could arise solely from possible diVerences in passive mechanical properties in the two receptor muscles. Earlier studies (KuZer, 1954) observed that the contractile properties indeed diVered; the rapidly adapting muscle had properties resembling a fast twitch fiber, whereas the slowly adapting muscle behaved as a slow twitch fiber. The viscoelastic properties of the receptor muscles in the slowly and rapidly adapting receptors were investigated by extending the muscles while measuring the resulting force at one end of the muscle fiber (Rydqvist et al., 1991, 1994). It was found that the viscoelastic properties of the two muscles diVered considerably, the rapidly adapting receptor muscle having more dynamic characteristics (Fig. 3A). The muscles could be reasonably well described by a viscoelastic model consisting of a Voigt element (parallel spring and damping element, Fig. 3, inset) in series with a nonlinear spring (Rydqvist et al., 1991; Swerup and Rydqvist, 1996). The diVerence in viscoelastic properties probably relates to the morphological diVerences mentioned above. At least, part of the diVerence in adaptive properties between the slowly and rapidly adapting receptors is due to the diVerent viscoelastic properties of the muscle fibers (Rydqvist et al., 1994).
27
2. Stretch Receptor Mechanotransduction B
A Rapidly
50 kPa K1
C
Slowly
D
F
K2 B Voigt element
F
Nonlinear spring
25 kPa
30% 50 ms FIGURE 3 Typical tension responses due to imposed stretches in the (A) rapidly and (C) slowly adapting receptor muscle. (B and D) Calculated responses in the rapidly and slowly adapting receptor muscle, respectively, using the model seen in the inset and using diVerent spring and dashpot values. Adapted with permission from Blackwell (Rydqvist et al., 1994).
C. MSCs in the Receptor Neurons As is evident from other chapters in this volume, considerable advances have been made in the field of structure and function of MSCs. It is now over 20 years since the first MSC, or stretch‐activated (SA) ion channel, was reported and analyzed by Sachs and coworkers (Guharay and Sachs, 1984) in cultured chick embryonic muscle cells. This was soon followed by recordings from MSCs in bacteria which also gave rise to the first molecular structure of an MSC (Martinac et al., 1987; Sukharev et al., 1994; see also Kung, 2005). However, relatively few sensory systems have been studied in terms of MSCs despite the fact that few doubt that mechanotransduction in diVerent sensory cells is due to opening of MSCs, for example hair cells in the hearing organ, the crayfish stretch receptor, and touch receptors in the nematode Cenorhabditis elegans. Recordings from single MSCs using the patch clamp technique have been made only in a few pure mechanosensory neurons due to the fact that the MSC are situated in very fine cilia and thus not easily accessible to single‐channel recording. Instead, whole‐cell recordings and indirect methods like knockout techniques as in C. elegans
28
Bo Rydqvist
and molecular techniques such as in situ hybridization have been used to define the presence of MSC (Sukharev and Corey, 2004). The crayfish stretch receptor is one of the few undisputed mechanosensory organs where actual recordings and analysis of MSCs have been performed (Erxleben, 1989). On the other hand, the molecular structure of the channel is still not determined. The receptor neurons (slowly and rapidly adapting) are of the nonciliary type, which probably have implications for the type of MSC present in these neurons. The sensitivity of the crayfish MSCs is very high compared to other MSCs, for example the bacterial channels (Hamill and McBride, 1994). Erxleben (1989) reported the presence of two types of MSCs in the slowly adapting neuron of the crayfish O. limosus. These channels are believed to be present in high density in the extensive dendritic tree of the neuron (Fig. 1D) but are also present in the large dendrites and the soma. Since the dendrites are too small and buried in the intercalated zone of the receptor muscle, the only possible parts of the neuron accessible to patch clamp are the soma and the main dendrites. Erxleben (1989) found a marked increase in single‐channel activity when the membrane of the patch was deformed by applying suction to the pipette (Fig. 4). Two diVerent types of MSCs were reported on the soma and the primary dendrites of the neuron with similar conductance properties but diVerent voltage range of activation and diVerent sensitivity to membrane tension: (1) an inward‐rectifying SA (RSA) channel which responded only weakly to membrane tension and (2) an SA channel which was only weakly voltage dependent but was more sensitive to membrane tension. The RSA was inactive when no suction was applied to the pipette and showed a decreased open probability when the patch was depolarized (Fig. 3 in Erxleben, 1989). The RSA was also found mostly in the soma, whereas the SA channel was found predominantly in the large dendrites. Figure 4A and B show single‐channel recordings from RSA and SA ion channels. Whereas the RSA never reached saturation in the suction range used, the SA channel displayed a classical sigmoid relation between suction pressure and open probability (P0) with a saturating pressure of about 25 mmHg (Fig. 4C and D). These experimental curves could be described by either of the following Boltzmann equations: P0 ¼
1 1 þ k expð spÞ
ð1Þ
P0 ¼
1 1 þ k expð sp2 Þ
ð2Þ
29
2. Stretch Receptor Mechanotransduction B
A mmHg 0
mmHg 0
5
9
9
28
14
33
18
40
26
44 10 pA
D
C
100 ms
0.15 0.8 P0
0.10
0.4
0
0.05
0
10 20 Suction (mmHg)
0 30
0
10
30 40 20 Suction (mmHg)
50
FIGURE 4 The two types of mechanosensory channels observed in the soma of the crayfish (Orconectes limosus) slowly adapting neuron. (A) SA (stretch‐activated) type of mechnanosensitive ion channel. (B) RSA (rectifying SA) type mechanosensitive channel. (C) Stimulus– response relation of SA channel. (D) Stimulus–response relation of RSA channel. Reproduced from Erxleben (1989) by copyright permission of The Rockefeller University Press.
where k is a pressure insensitive term, s the sensitivity, p the applied negative pressure in the pipette, and P0 the open probability of the channel (Erxleben, 1989). The two equations illustrate the two types of gating mechanisms that have been proposed (Sokabe and Sachs, 1992; Kung, 2005). The open probability dependence of the squared suction pressure is derived from the model where the force in the lipid bilayer is changing/gating the MSC. In the linear model, the open probability is dependent on p, and is equivalent to the tethered type of gating where a spring is attached to the MSC controlling the gate. From available experimental data, it is obvious that it is not possible to distinguish between these two models for the RSA and SA in the crayfish
30
Bo Rydqvist
sensory neurons. However, the sensitivity of RSA is close to what was found for one of the MSCs in Escherichia coli, the MscL, a channel that is reported to have a gating mechanism that is best described by the lipid phase model (Martinac et al., 1987). This stems from the fact that mechanosensitivity was preserved even when the MscL was inserted in pure lipid artificial membranes where the force acting on the channels must come from the lipid phase. The squared model was also concluded by Guharay and Sachs (1984) to best fit the gating properties of the chicken muscle MSCs. Other animal cell MSCs are, however, thought to be gated through a tether (Kung, 2005). The issue of gating regarding the SA and RSA in the crayfish neuron is at the moment an open question, and it is not possible to deduce information about the structural nature of the RSA and SA in the crayfish neurons. The permeability properties of the crayfish stretch receptor MSC were studied by Brown et al. (1978) and Edwards et al. (1981) from macroscopic currents and by Erxleben (1989) from patch clamp recordings. Both types of studies show permeability for Naþ, Kþ, and Ca2þ ions. The single‐channel analysis gave a slope conductance for Kþ of 71 11 pS, Naþ of 50 7.4 pS, and Ca2þ of 22 3 pS. Assuming an average resting membrane potential of 70 mV, the reversal potential for the SA current was estimated by extrapolation to be about 0 mV. The permeability through the rectifying channel RSA was similar: 44 pS for Naþ and 22 pS for Ca2þ. This can be compared to the estimate based on macroscopic currents. Brown et al. (1978) found a value for the reversal potential of 13 6.5. The result was in general agreement with those reported by other groups (Obara, 1968; Klie and Wellho¨ner, 1973), even though these estimates were made by extrapolation. However, using Tris (Trizma) and arginine as substitute for Naþ, it was shown that even these ions could permeate the MSC in the receptor neuron and the PTris/PNaþ was 0.31 and the PArginine/PNaþ was 0.25 (Brown et al., 1978). This indicates that the pore of the crayfish MSC might be of considerable size. In Edwards et al. (1981), the size of the channel was discussed and it was concluded that the crayfish MSC must be of about the same size ˚. or slightly larger than the acetylcholine receptor channel, that is around 7 A The permeability of divalent cations was also investigated independently and based on the SA currents. The permeability for Ca2þ and Mg2þ compared to Naþ was estimated to be 0.3 and 0.4, respectively (Edwards et al., 1981), similar to what was found by Erxleben (1989) (see above). In the rapidly adapting receptor PCa/PNa ¼ 0.44 and PMg/PNa ¼ 0.60 (Rydqvist and Purali, 1993). These values were based on PNa/PK ¼ 1.6 and 1.5 for the slowly and rapidly adapting receptor neurons, respectively, which is slightly at odd with the values for Naþ and Kþ conductances obtained by Erxleben (1989). The influx of Ca2þ was suggested by Erxleben (1993) to be responsible for part of the adaptation of the receptor current by activating a Ca2þ‐ dependent Kþ channel in the neuron. In recordings from patches containing
2. Stretch Receptor Mechanotransduction
31
both SA and KCa, he was able to demonstrate that there was a concomitant increase in the open probability for the SA channel and the KCa channel. These results are in line with measurements of macroscopic stretch‐induced currents (Ottoson and Swerup, 1985a,b). However, it was also found that Ca2þ had direct eVects on the MSCs. When Ca2þ was reduced in the external solution from the normal concentration of 13.5 mM to 1.35 and 0.13 mM, the stretch‐induced generator current increased. This was interpreted as an eVect of Ca2þ, possibly on an internal site of the MSC, by which Ca2þ reduced the permeability to monovalent cations or decreased the open probability of the crayfish MSC (Brown et al., 1978). The eVect was observed even when Naþ was substituted with Tris or arginine.
D. Macroscopic Receptor Currents in the Stretch Receptor Neurons The findings at the single‐channel level can be compared to the macroscopic current response to stretch, that is the receptor current. If the sensory neuron of the stretch receptor is subjected to voltage clamp, it is possible to observe the receptor current generated by extending the receptor muscle without interference of voltage‐gated ion channels. This means that the receptor current reflects the activation of the MSCs of the receptor neuron. It should be observed, however, that a Ca2þ‐activated Kþ current (Erxleben, 1993) could be present, since the MSCs in both slowly and rapidly adapting receptor neurons are permeable to Ca2þ (Edwards et al., 1981; Rydqvist and Purali, 1993) as discussed in the previous section. However, the quantitative contribution at the macroscopic level is still unclear because using several blockers of Ca2þ‐dependent Kþ current did not indicate any eVect of either slowly or rapidly adapting neuron (Purali and Rydqvist, 1992). In addition, it cannot be altogether ruled out that that some non‐MSC could be involved in the generation of the receptor current. However, there is so far no evidence for such channels. As is seen in Fig. 2, there is an essential diVerence in shape between the time course of tension and current response to a ramp and hold extension (cf. Swerup and Rydqvist, 1992), which is also seen comparing Figs. 3 and 5. This suggests that there is no simple linear relation between tension and current but that the current must be related to the tension in the receptor muscle through the SA ion channels (MSCs). Consequently, the stimulus‐response relation for both the slowly and rapidly adapting receptors is not linear but typically sigmoid in character (Rydqvist and Swerup, 1991; Rydqvist and Purali, 1993), and the amplitude of the receptor current reaches a maximum probably determined by the number of MSCs being simultaneously open (Fig. 5C). Using a log–log relationship between the stimulus and the response, the sigmoid character
32
Bo Rydqvist
FIGURE 5 (A) Receptor currents in slowly and rapidly adapting neurons from the same organ due to ramp and hold extensions (6%, 12%, and 18%) of the receptor muscle. (B) The decay phase is quite diVerent which is more clearly seen in the log plot in B. (C) Typical stimulus response characteristics for a rapidly adapting neuron. TEA þ 4‐AP did not aVect the receptor current. With permission from Blackwell (Rydqvist and Purali, 1993).
of the relation was preserved indicating that a Stevens’ power law is not applicable over the entire stimulus range (Stevens, 1957). For the virtually linear part of the log–log curve for the peak current (10–20% extension), the n‐value as defined by Stevens was 3.0 in the slowly adapting neuron and 4.7 in the rapidly adapting neuron (Rydqvist and Swerup, 1991; Rydqvist and
2. Stretch Receptor Mechanotransduction
33
Purali, 1993). These are very steep relationships compared to many other mechanosensory systems, indicating that in this physiological range (10–20%) the muscle receptors have a considerable impact on the reflex arcs controlling the extensor muscles of the abdomen. This type of movement is characteristic for escape responses of the crayfish. However, for small extensions, less than 10%, the slope is close to 1 which indicates that for small length changes of the extensor muscle the SRO could have a more moderate impact on the motor system. The time course of the receptor current of the slowly and rapidly adapting stretch receptors diVer, the rapidly adapting receptor current having a typical dynamic character (Fig. 5A and B ). This can be compared to tension development in the receptor muscles, which also displays a more dynamic character in the rapidly adapting muscle fiber as compared to the slowly adapting muscle (Fig. 3A and C). To investigate if it was possible to calculate the receptor current and potential using available data for the SRO, Swerup and Rydqvist (1996) developed a model that took into account data from the slowly adapting stretch receptor. The model was based on viscoelastic properties of the muscle fiber, the biophysical properties of the MSCs, and the passive properties of the neuronal membrane (a lumped leak conductance and capacitance). The model could take into account a wide range of experimental data from the slowly adapting neuron provided that a time‐dependent shift of open probability of the MSC (MSC adaptation) was taken into account (Hamill and McBride, 1994; see also Section IV.G).
E. Pharmacology of the Crayfish MSCs Although a molecular characterization of the MSC in the crustacean stretch receptor has not been successful so far, despite several attempts, pharmacological characterization has given some clues to the molecular nature of the MSC (for a review on MSC pharmacology see Hamill and McBride, 1996). The MSCs of the stretch receptor neuron or the receptor current are not aVected by TTX, 4‐AP, or TEA (Ottoson and Swerup, 1985b, Erxleben, 1989; Rydqvist and Purali, 1993). The trivalent lanthanide gadolinium (Gd3þ) was found to block the stretch‐induced current in the stretch receptor neuron, although it also blocked both the voltage‐gated Naþ channel and the Kþ channel to some degree which indicates that it is not completely selective for the crayfish MSC. The crayfish receptor current was more sensitive to Gd3þ when Ca2þ was lowered, indicating some competitive interaction between these two ions (Swerup et al., 1991). Several local anesthetics were found to aVect the receptor current in the stretch receptor neuron (Fig. 6). Lidocaine at low concentrations facilitated
34
Bo Rydqvist
A
Control
B
100 nA
8 mm lidocaine
Relative receptor current (%)
150
Control 8 mm lidocaine 16 mm lidocaine
100
50
0 30%
0
10
20
30
40
Extension (%)
50 ms C
D Control
150 nA
4 mm Bupivacaine
30%
Relative receptor current (%)
100
Control 2 mm bupivacaine 4 mm bupivacaine
50
0 0
50 ms
10
20
30
40
Extension (%)
FIGURE 6 EVect of lidocaine and bupivacain on receptor currents. Stretch receptor neurons were stimulated by extensions of receptor muscle from 3% to 30% of resting muscle length. (A) Receptor currents in control solution (top) and in 8‐mM lidocaine (middle) to extensions (bottom). (B) Relative peak receptor current vs extension in control solution and in 8‐ and 16‐mM lidocaine. (C) Receptor currents in control solution (top) and in 4‐mM bupivacaine (middle) to extensions (bottom). (D) Relative peak receptor current vs extension in control solution and in 2‐ and 4‐mM bupivacaine. With permission from Blackwell (Lin and Rydqvist, 1999).
the receptor current, whereas tetracaine, bupivacaine, and an analogue to lidocaine (LL33) partially blocked the receptor current (Lin and Rydqvist, 1999a). There was a correlation between the eVect on receptor current and the oil:water distribution coeYcient which indicates that the local anesthetic
2. Stretch Receptor Mechanotransduction
35
blocking eVect is mediated through the lipid phase. In an earlier study, it was also shown that several nonionic detergents in the Triton series (i.e., Triton X‐100 and X‐45) could block the receptor current (Ottoson and Rydqvist, 1978). These results suggest that this particular MSC can be modulated through the lipid phase and that the gating mechanism is similar to what is found for bacterial MSCs (Martinac, 2004). Some transient receptor potential (TRP) ion channels have been suggested to be mechanosensitive (Martinac, 2004; Sukharev and Corey, 2004; Kung, 2005), for example the TRPV4, TRPA1, TRPN. This is an attractive possibility since the TRPN channel family (NompC) was shown to be responsible for mechanosensitivity in Drosophila—an invertebrate relative (Walker et al., 2004). We have used ruthenium red, known to block TRP channels, to investigate possible eVects on the stretch receptor neuron. However, no eVect on the receptor current in the stretch receptor neuron has been observed (Fernstro¨m and Rydqvist, unpublished observations). Ruthenium red is probably a crude tool in this respect, and considering the great number of possible TRP channels involved in mechanosensing these results must be interpreted with caution. A spider toxin (GsMTx4) from the spider Grammostola spatulata has been shown to block some MSCs in the heart (Bode et al., 2001), astrocytes, and kidney (Suchyna et al., 2000). In preliminary experiments, we have studied the eVects on the receptor current of the purified fraction GsMTx4 at concentrations up to 10 mM. Only minor eVects on the receptor current were found using the toxin (Fernstro¨m, Rydqvist, and Sachs, unpublished observations). In a similar type of experiment, the stretch receptor preparation was exposed to 1 mM amiloride, a substance known to block MSCs of the ENaC/DEG type (epithelial Naþ channel, degenerin channel protein), that are responsible for mechanotransduction in C. elegans and hair cells (Charfie and Sulton, 1981; Driscoll and Chalfie, 1991; Martinac, 2004; Sukharev and Corey, 2004). Our results clearly show that this substance has very small eVects on SA currents, since in three experiments no significant eVect could be demonstrated (Fernstro¨m and Rydqvist, unpublished observations). It is thus an open question as to what are the molecular constituents of the MSC in the crayfish stretch receptor neuron. It is observed in experiments using local anesthetics and detergents on the stretch receptor that substances that perturb the lipid phase have an increased tendency to aVect mechanotransduction. Further, the more hydrophobic the substance is the larger the blocking eVect. This indicates that the MSC in the sensory neuron is gated through the lipid phase and not through the cytoskeleton or extracellular matrix, that is a tethered model. This is supported by the fact that the local anesthetic shifted the stimulus–response curve indicating an eVect on gating
36
Bo Rydqvist
(Lin and Rydqvist, 1999a) and not simply a plugging of the ion channel pore. In addition, the relatively slow onset of the eVect in these experiments, similar to what was found by Martinac et al. (1990), could be explained by diVusion of the anesthetics into the lipid bilayer.
F. Voltage‐Gated Ion Channels and the Generation of Impulse Response 1. Na1 Channels Like many other neurons, the action potential or nervous impulse in the stretch receptor neuron is generated by Naþ and Kþ ion channels. No voltage‐gated Ca2þ channel has been implicated in this process. Generation of action potentials is fundamental in shaping the final mechanotransduction response in the sensory neuron and the properties of the ion channels involved in this process have important consequences for the particular type of response seen in the rapidly and slowly adapting neurons. It is thus not surprising that the shape of action potentials in the slowly and rapidly adapting neurons diVer. In the rapidly adapting receptor neuron the amplitude is around 55 mV, whereas in the slowly adapting neuron the amplitude is around 80 mV. It was also found that the duration of the action potential was longer in the slowly adapting receptor mainly due to a slower repolarization (Purali and Rydqvist, 1998). This is consistent with the diVerence in properties of the Naþ and Kþ channels. Naþ currents generate action potentials in both the slowly and rapidly adapting neurons. In the slowly adapting neuron, the Naþ current is larger and the inactivation (th) is slower and takes place at more negative potentials compared to that in the rapidly adapting neuron, consistent with the properties of the action potentials. It was also observed that pinching the axon of the rapidly adapting neuron close to the soma totally abolished the action potentials, contrary to what was found in the slowly adapting neuron. Further, in the rapidly adapting neuron, the action potentials recorded in the axon were larger and had slightly faster rise time than those recorded in the soma (Purali and Rydqvist, 1998). Taken together, these observations point toward a possible diVerence in Naþ channel distribution in the two types of neurons (Fig. 7B). Recordings of Naþ currents in the slowly adapting neuron using the two‐ electrode voltage clamp technique indicated the presence of two diVerent Naþ channel populations with diVerent kinetic properties (Purali and Rydqvist, 1998; Lin and Rydqvist, 1999b). Since this was not the case using macropatch clamp recordings in the soma of the slowly adapting neuron (Lin et al., 1999), the observations point toward a specific spatial distribution of at least two diVerent sets of Naþ channels. As a result of these
37
2. Stretch Receptor Mechanotransduction Rapidly
Slowly
A Er (mV) −55
L0 (mm) 115
−69
−10 −29
Er (mV) −60
400 nA
−58
−20
−55
L0 (mm) Slowly
−20 210
−23 215 −28 300
L0 (mm) 150
175 −63
B
−63 −3 −66
Em (mV)
2 ms
Na+ Na+ axon
250 −36 325
Em (mV)
Rapidly
2 ms þ
FIGURE 7 (A) Peak Na currents from four slowly and four rapidly adapting neurons in which the axon was cut as indicated in B. L0 indicates distance from center of nucleus to cut position. Er is resting membrane potential, which was also holding potential. Em is the potential value for peak Naþ current. (B) Suggested Naþ channel distribution in the slowly and rapidly adapting neurons. Two types of Naþ channels are proposed, one (Naþ) that can be present in both soma and axon and one (Naþ‐axon) present in axon only. Arrow indicates location of cutting the axon as stated in B. Reprinted from Lin and Rydqvist (1999) (A) with permission from Elsevier, and from the author (Lin, 2000) (B).
observations, we recorded Naþ currents from the soma of slowly and rapidly adapting neurons after cutting the axons at diVerent positions from about 350 to 100 mm from the soma. To our surprise, in the rapidly adapting neuron, the Naþ current was completely abolished when the axon was cut at about 150 mm from the soma, whereas in the slowly adapting neuron most of the Naþ current was preserved even if the axon was cut as close as 100 mm from the soma (Fig. 7). This indicates that in the slowly adapting neuron, one of the suggested Naþ channels is dominating in the soma but both channels may be present in the axon. In the rapidly adapting neuron, the results so far indicate a single Naþ channel population located at least 150–200 mm out in the axon (Lin and Rydqvist, 1999b). However, there appears to be a part of the axon even further out with a high density of Naþ channels, constituting a trigger area for the action potential as indicated from recordings of action potentials made concomitantly in the axon and the soma (Purali and Rydqvist, 1998). The diVerence in spatial distribution of the Naþ channels between the two neurons suggests that it might be important for the diVerence in adaptation. This will be discussed further in a later section.
38
Bo Rydqvist
2. K1 Channels There are presently over 100 genes coding for Kþ channel subunits, including the b‐units modulating Kþ channels. Many neurons and other cells contain a large number of diVerent Kþ channels. In the stretch receptor neuron of the crayfish, three types of voltage‐gated Kþ channels and one Ca2þ‐activated Kþ channel have been identified to date. In experiments using two intracellular electrode voltage clamp, it was shown that both the slowly and rapidly adapting receptor neurons contained at least two diVerent Kþ channels (Rydqvist and Zhou, 1989; Rydqvist and Purali, 1991; Purali and Rydqvist, 1992). The whole‐cell Kþ currents in the slowly (Brown et al., 1978; Rydqvist and Zhou, 1989) and the rapidly adapting neurons (Rydqvist and Purali, 1991) were characterized by a transient and an outwardly rectifying component. The activation time constant for the Kþ current in the rapidly adapting neuron was smaller and the activation took place at more negative potentials as compared to the slowly adapting receptor. The results were supported by macropatch recordings from the soma of the slowly adapting neuron which gave almost identical results (Lin et al., 1999). The inactivation as derived from whole‐cell currents had two time constants in both receptors, a fast component of about 0.5 ms and a slow component ranging from 2 to 8 s (Brown et al., 1978). Pharmacological dissection of the Kþ currents in the slowly and rapidly adapting neurons using 4‐AP and TEA suggested two diVerent populations of ionic channels: one channel having high aYnity to TEA and the other low aYnity to TEA. The results further indicated that the low TEA aYnity channel dominated in the slowly adapting neuron, whereas in the rapidly adapting neuron both channels were equally common (Purali and Rydqvist, 1992). Later experiments using patch clamp recordings from the slowly adapting neuronal soma have demonstrated the existence of three diVerent types of Kþ channels in this neuron (Fig. 8). First, an outward delayed rectifier has been analyzed in detail having a single‐channel conductance of 13 pS and a PK ¼ 6.5 10 14 cm3/s (mean values) with little inactivation (Figs. 8A and 9). First latency analysis suggested a two closed states preceding two open states. The channel displays properties similar to a Kþ channel of the Kv1.2 type (Lin and Rydqvist, 2001). Second, a Kþ channel with large conductance (53 pS) having properties suggesting a delayed outward rectifier with some inactivation as seen from cell‐attached recordings (Fig. 8B). The third Kþ channel is clearly a transient Kþ channel (Fig. 8C) with fast inactivation (estimated time constant in the order of 20–50 ms). This channel has a conductance of 23 pS. The 23 and 53 pS Kþ channels are diYcult to detect at resting membrane potential but could be activated at a depolarization of 10–20 mV. Other Kþ channels with single‐channel conductance of less than 10 pS have also been observed but further experiments are necessary to analyze these currents.
39
2. Stretch Receptor Mechanotransduction A
B
0
0
+20
+20
+40
+50
+60 +80
+80
+110 +110
5 pA 200 ms
i (pA) 3
+140
2 1
V m (mV)
0
−120 −80 −40 0 −1
−80 2 pA
40 80 50 ms C 0
+20 +50 +90 +100
2 pA 200 ms
+140 2 pA
−80 50 ms
FIGURE 8 Top row: cell‐attached patch clamp recordings from three diVerent Kþ channels in slowly adapting stretch receptor neurons. The patches were depolarized from resting level to the potential indicated; normal saline in the pipette. Bottom row: I–V curves for the Kþ channels. In A, the curve was based on single‐channel activity at potential steps; in B and C, voltage ramps were used. (A) Recordings from a 13‐pS Kþ channel. In this cell, the PK ¼ 5.4 10 14 cm3/s (I–V curve bottom A). (B) Top: single‐channel currents recorded in cell‐attached patch from a 53‐pS Kþ channel; representative currents at the potentials (marked on the left) to which the patch was depolarized from the resting state. The voltage step started and ended as indicated by the capacitative current. The bottom panel shows superimposed currents activated by voltage ramp from 80 to þ140 mV; conductance 53 pS. (C) Same as in B; this Kþ channel is typical transient with conductance of 23 pS. Reprinted from Lin and Rydqvist (1999) with permission from Elsevier and the author Lin (2000, Fig. 7, p. 25).
40
Bo Rydqvist
A
B 0 I (pA)
+20
4
+40
3 2
+60
1 +80 −80
−40
Vm (mV)
0 0
40
80
−1
+100
2 pS 20 ms
C
D
Number of events
1.0 O2 O1
10,000
C
Popen
2
1 0.5
0
5000
0 −1
0
1 2 3 Amplitude (pA)
4
0.0 −80 −40 0 40 Membrane potential (mV)
80
FIGURE 9 (A) Analysis of ensemble average currents from the same patch containing two Kþ channels. (A) Average currents initiated by depolarizations from 0 to þ100 mV from resting membrane potential. The number of sweeps was from 105 to 194. (B) Voltage dependence of the average currents plotted against membrane potential. (C) Amplitude histogram from the recording at 80‐mV depolarization (inset) and fitted to a third‐order Gaussian distribution. The peaks correspond to closed level (0), one channel open (1), and two channels open (2). (D) Open probability vs membrane potential using data from A and single‐channel data (C) Open probability, P0 ¼ I/(i N), where I is the average current, i is single‐channel current, and N the observed maximal number of channel in the patch (two in this patch). The smooth line is a fit to a Boltzmann equation. Reprinted from Lin and Rydqvist (2001) with permission from Elsevier.
It is thus clear that at least the slowly adapting neuron contains up to six diVerent Kþ channels and that there might be considerable diVerences compared to the rapidly adapting neuron. This would explain part of the diVerence in action potential properties between the two neurons and the firing properties of the two neurons.
2. Stretch Receptor Mechanotransduction
41
The spatial distribution and density of the Kþ channels have been diYcult to define but experiments using macropatch recordings from diVerent locations in the soma have given some clues (Lin et al., 1999). The experiments show that the Kþ current density is highest close to the axon hillock area as compared to current densities in the dendritic part of the soma. This suggests that Kþ channel distribution is diVerent over the neuronal surface and it can be anticipated that the slowly and rapidly adapting neurons diVer in their Kþ channel distribution. Finally, a Ca2þ‐activated Kþ channel was suggested by Ottoson and Swerup (1982, 1985a,b) who injected EGTA and TEA into the slowly adapting sensory neurons and found changes in receptor potential adaptation consistent with eVects on a Kþ channel. This was later confirmed by Erxleben (1993) who recorded simultaneously from SA (MSCs) and Kþ channels. He observed that when SA channels were stimulated by suction in the patch pipette, the Kþ channel in the same patch increased its activity. Since the Kþ channels in isolated patches were not activated by suction, he concluded that Ca2þ entering through the SA channel activated a nearby Kþ channel and thus was a Ca2þ‐activated channel. This is a strong support for the suggestion made earlier by Ottoson and Swerup (1985a,b) that this channel contributed to the early adaptation in the receptor potential.
G. Adaptation: A Multifactor Property The diVerence in adaptive properties between the two neurons in this receptor, as well as in other similar organs, has been a constant challenge. Figure 1B illustrates the distinct diVerence in adaptation of impulse discharge in responses to mechanical stimulation between the rapidly and slowly adapting receptors (cf. Rydqvist and Purali, 1993; Fig. 1). The rapidly adapting receptor gives a brief impulse discharge in response to a ramp and hold extension, whereas the slowly adapting receptor gives a sustained impulse discharge for the same stimulus. As already observed by Nakajima and Onodera (1969a,b), the same diVerence in impulse response is seen when the neurons are electrically stimulated (Rydqvist and Purali, 1993). Adaptation must therefore be a consequence of several processes in the receptor organs. As can be observed from Fig. 2, the viscoelastic properties of the receptor muscle must contribute to adaptation since the tension response to ramp and hold extensions of the muscle adapts (Rydqvist et al., 1991, 1994). Further, a distinct diVerence is seen between the tension response in the rapidly and slowly adapting receptor muscles (Fig. 3). The transient peak is more pronounced in the rapid muscle compared to that in the slow muscle.
42
Bo Rydqvist
This correlates well with the diVerence in receptor current seen in Fig. 5A. The receptor current of the rapidly adapting neuron has a dynamic phase that is more pronounced as compared to the slowly adapting neuron. The fast initial decay phase of the rapidly receptor current is also considerably faster as compared to that of the slowly adapting current (Fig. 5B). From studies of the viscoelastic properties and the properties of the MSCs, Swerup and Rydqvist (1996) developed a model of the primary transduction process which gave a reasonable fit to receptor potential responses of the slowly adapting stretch receptor. To achieve the fit, it was, however, necessary to include an MSC‐specific adaptation originally proposed by Hamill and McBride (1994) who observed this adaptation in oocyte MSCs. This type of MSC adaptation has not been experimentally observed in the crayfish stretch receptor neuron but from the model fit it is assumed to be present in the crayfish MSC. Due to lack of quantitative data on Ca2þ‐dependent Kþ currents (cf. Erxleben, 1993), a KCa current was not included but could be the additional factor which would make the fit even better. DiVerences in density of Ca2þ‐ dependent Kþ channels in the slowly and rapidly adapting neurons could be a factor determining the adaptation in receptor current. However, since electrical stimulation gave the same principal type of adaptation in impulse discharge in the slowly and rapidly adapting neurons as seen for extension of the receptor muscle, nonmechanical factors must also contribute to adaptation. Analysis of both Naþ and Kþ currents in the slowly and rapidly adaptive neurons as outlined above have revealed some diVerences in kinetic properties between the two neurons. Some of these kinetic changes are consistent with the adaptive properties seen in the two neurons. In an analysis of the Naþ currents in the slowly and rapidly adapting neurons, it was found that the inactivation parameter in the rapidly adapting neuron was moved in the negative direction (Purali and Rydqvist, 1998) that is compatible with inactivation of impulses occurring at less depolarized levels than in the slowly adapting neuron. In a simple model of the neurons using a Hodgkin‐ Huxley modeling of the voltage‐gated channels (Rydqvist and Swerup, 1991; Rydqvist et al., 2003), it was shown that a minor change of the activation (shift of þ8 mV in m parameter) and inactivation rate constants (shift of 8 mV of h parameter) could dramatically change the firing pattern in the neuron. An additional factor is the spatial distribution of the Naþ and Kþ channels in the two neuron types. This is suggested but not proved to be an important factor for the typical pattern of impulse discharge seen in these neurons. So far, it has only been possible to define tentatively the spatial distribution of the Naþ channels in the two neurons. As described above (Fig. 7), the rapidly adapting receptor neuron seems to have few Naþ channels in the soma and a high concentration of the Naþ channels in the axon, whereas the slowly neuron seems to have a similar density of Naþ channels
2. Stretch Receptor Mechanotransduction
43
in both axon and soma (Lin and Rydqvist, 1999b). It is possible that the distribution and density of Kþ channels are even more important but so far this has not been possible to determine. To ascertain the relation between spatial distribution of the diVerent channels and the impulse response characteristics, it is necessary to develop a compartmental model of the neuron. Presently, we are in the process of developing such a model to be able to better study how diVerent densities and distribution of channels influence impulse discharge in these receptor neurons. It will also be necessary to develop more eYcient methods (e.g., in situ hybridization) to determine experimentally the location of ion channels on the cell membrane.
V. SUMMARY AND DISCUSSION OF FUTURE RESEARCH DIRECTIONS The crayfish SRO is a very useful preparation for the study of mechanotransduction in all its aspects. It is a true mechanosensor, analogous to the vertebrate muscle spindle, with a clear‐cut function to monitor muscle length through activation of a sensory neuron equipped with MSCs that are proposed to generate the receptor current (stretch‐induced current). The receptor current activates voltage‐gated Naþ and Kþ ion channels present in the soma and axon, generating the final output of the organ: an impulse train that will reach the crayfish central nervous system. All channels thus contribute to the performance of the sensory organ. At this point, some caution should be expressed regarding the casual relation between MSC and the stretch‐generated current. Even though there is no indication, to the contrary it cannot be entirely excluded that some other channels could contribute to the stretch‐induced current. The MSCs have a small conductance as compared to the bacterial MSCs (MscL and MscS). The gating mechanism is not defined, but from pharmacological results could well be due to tension in the membrane lipids, that is, the quadratic model similar to what was found for the behavior of bacterial MscL that were inserted into pure lipid membranes (Kung, 2005). The stimulus– response relation is very steep, indicating that this MSC belong to the most sensitive MSCs discovered to date. This is also reflected by the stimulus– response relation of the macroscopic receptor current that is also very steep with a power function of between 3 and 5 in a Stevens’ power law concept. In addition, the receptor current amplitude is greater in the rapidly adapting neuron as compared to that in the slowly. It is obvious that the fact that the organ is specialized for mechanosensory detection is reflected in the properties and densities of the MSCs present in the neurons. This points toward a real challenge for the future, namely to define the molecular nature of the crayfish MSC represented in the slowly and rapidly adapting neurons.
44
Bo Rydqvist
The voltage‐gated Naþ and Kþ channels are important players in the generation of the action potentials and thus the final impulse response of the two neurons. It has been pointed out in this chapter that diVerences in adaptive properties between the two receptor neurons can be partly explained by the viscoelastic properties of the receptor muscles and by possible diVerences in MSC setup and by activation of Ca2þ‐activated Kþ currents. However, since the same adaptive properties seem to be present also with electrical stimulation (Rydqvist and Purali, 1993; Fig. 1), the voltage‐gated ion channels must also be involved in the adaptive characteristics of the two receptors. Several factors could contribute. First, the type of Naþ channels present in the two types of neurons. This is supported by the observation that two types of Naþ channels seem to be present in the neurons. The two Naþ channels, in all probability, seem to have diVerent kinetic properties. It is shown in model experiments (Rydqvist et al., 2003) that small diVerences in activation and inactivation of the Naþ channels can have profound eVects on the firing properties. Second, the distribution of Naþ channels is diVerent in the two neurons. This is supported by observations of Naþ currents in neurons with axons cut at diVerent positions (Lin and Rydqvist, 1999). The result indicates that in the rapidly adapting neuron, the Naþ channels are present in the axon only, whereas in the slowly adapting neuron Naþ channels are present in both axon and soma. This would aVect the initiation site of the action potential. In the future, this must be further investigated and determined using histological as well as patch clamp techniques. Third, the Kþ channels are not the same in the rapidly and slowly adapting neurons. Four types of Kþ channels have, so far, been tentatively defined in the slowly adapting neuron: three types of voltage‐gated Kþ channels and one Ca2þ‐activated Kþ channel. It is not known if the same type of Kþ channels are present in the rapidly adapting neuron or if the relative proportion of the channels is diVerent. Purali and Rydqvist (1992) demonstrated, using pharmacological dissection, that the type of voltage‐gated Kþ channels is not the same in the two neurons. Also, the action potentials in the two neurons are diVerent, the one in the rapidly adapting neuron having a much faster repolarization (Purali and Rydqvist, 1998). The explanation of these diVerences has to be further studied in the future. Fourth, the spatial distribution of Kþ channels in the two neurons is unknown. In particular, the relation between the Naþ channel densities, MSC densities, and Kþ channels densities are important. Kþ channels in the soma could act as a current sink for the current generated by the MSCs in the dendrites. This would aVect the generation of action potentials, particularly in relation to where the Naþ channels are located. To solve these
2. Stretch Receptor Mechanotransduction
45
problems we have to use a combination of morphological, electrophysiological, and model studies of this preparation. Investigations of this relatively simple invertebrate receptor are of fundamental importance for the general understanding of generation of signals in sensory receptors. The results will also give a better insight into the proprioceptive contribution to motor control and how it is possible to modify the function of such receptors, that is, the muscle spindle and tendon organs. In particular, it should be of interest to apply the results to clinically useful therapies (in diVerent pathological conditions where the reflexes involving the muscle receptors are aVected). Theoretically, this is possible through the muscle receptors since many motor conditions are generated through increased activity in reflex arcs dependent on mechanosensory organs. One such possibility should be to selectively block the MSCs. At present this has not been generally possible and until the molecular details of the invertebrate and vertebrate MSCs are known, this will be diYcult. In this context, it should also be noted that MSCs are probably not a homogenous entity. Several molecular constructions can probably be involved in this sensory modality. However, looking at the rate of development of molecular techniques, the determination of the MSCs in diVerent species including man should not be too far ahead. Acknowledgments This work was supported by grants from Karolinska Institutet. I thank Christer Swerup and Joseph Bruton for valuable discussions and criticisms.
References Alexandrowicz, J. S. (1951). Muscle receptor organs in the abdomen of Homarus vulgaris and Palinurus vulgaris. Q. J. Microsc. Sci. 92(Pt. 2), 163–200. Alexandrowicz, J. S. (1967). Receptor organs in thoracic and abdominal muscles of crustacea. Biol. Rev. 42, 288–326. Bode, F., Sachs, F., and Franz, M. R. (2001). Tarantula peptide inhibits atrial fibrillation. Nature 409, 35–36. Brown, H. M., Ottoson, D., and Rydqvist, B. (1978). Crayfish stretch receptor: An investigation with voltage‐clamp and ion‐sensitive electrodes. J. Physiol. 284, 155–179. Charfie, M., and Sulton, J. (1981). Developmental genetics of the mechanosensory neurons of Cenorhabditis elegans. Dev. Biol. 82, 358–370. Driscoll, M., and Chalfie, M. (1991). The mec‐4 gene is a member of a family of Cenorhabditis elegans genes that can mutate to induce neuronal degeneration. Nature 349, 588–593. Edwards, C., and Ottoson, D. (1958). The site of impulse initiation in a nerve cell of a crustacean stretch receptor. J. Physiol. 143, 138–148. Edwards, C., Ottoson, D., Rydqvist, B., and Swerup, C. (1981). The permeability of the transducer membrane of the crayfish stretch receptor to calcium and other divalent cations. Neuroscience 6, 1455–1460. Elekes, K., and Florey, E. (1987a). New types of synaptic connections in crayfish stretch receptor organs: An electron microscopic study. J. Neurocytol. 16, 613–626.
46
Bo Rydqvist
Elekes, K., and Florey, E. (1987b). Immunocytochemical evidence for the GABAergic innervation of the stretch receptor neurons in crayfish. Neuroscience 22, 1111–1122. Erxleben, C. (1989). Stretch activated current through single ion channels in the abdominal stretch receptor organ of the crayfish. J. Gen. Physiol. 94, 1071–1083. Erxleben, C. F. J. (1993). Calcium influx through stretch‐activated cation channels mediates adaptation by potassium current activation. Neuroreport 4, 616–618. Eyzaguirre, C., and KuZer, S. W. (1955a). Processes of excitation in the dendrites and in the soma of single isolated sensory nerve cells of the lobster and crayfish. J. Gen. Physiol. 39, 87–119. Eyzaguirre, C., and KuZer, S. W. (1955b). Further study of soma, dendrite, and axon excitation in single neurons. J. Gen. Physiol. 39, 121–153. Florey, E., and Florey, E. (1955). Microanatomy of the abdominal stretch receptors of the crayfish (Astacus Fluviatilis L). J. Gen. Physiol. 39, 69–85. Guharay, F., and Sachs, F. (1984). Stretch activated single ion‐channel currents in tissue‐cultured embryonic chick skeletal muscle. J. Physiol. 352, 685–701. Hamill, O. P., and McBride, D. W. (1992). The cloning of a mechano‐gated membrane ion channel. Trends Neurosci. 17, 439–443. Hamill, O. P., and McBride, D. W. (1994). Molecular mechanisms of mechanoreceptor adaptation. News Physiol. Sci. 9, 53–59. Hamill, O. P., and McBride, D. W. (1996). The pharmacology of mechano‐gated membrane ion channels. Pharmacol. Rev. 48, 231–252. Klie, W., and Wellho¨ner, H. H. (1973). Voltage clamp studies on the stretch response in the neuron of the slowly adapting crayfish stretch receptor. Pflugers Arch. 342, 93–104. Komuro, T. (1981). Fine structural study of the abdominal muscle receptor organs of the crayfish (Procambarus clarkii). Fast and slow receptor muscles. Tissue Cell 13, 79–92. KuZer, S. W. (1954). Mechanisms of activation and motor control of stretch receptors in lobster and crayfish. J. Neurophysiol. 17, 558–574. Kung, C. (2005). A possible unifying principle for mechanosensation. Nature 436, 647–654. Lin, J.‐H. (2000). Transducer properties of a mechanoreceptor. An electrophysiological and pharmacological study of the crayfish stretch receptor. Thesis. Karolinska Institutet, Stockholm. Lin, J.‐H., and Rydqvist, B. (1999a). The mechanotransduction of the crayfish stretch receptor neurone can be diVerentially activated or inactivated by local anaesthetics. Acta Physiol. Scand. 166, 65–74. Lin, J.‐H., and Rydqvist, B. (1999b). DiVerent spatial distributions of sodium channels in the slowly and rapidly adapting stretch receptor neuron of the crayfish. Brain Res. 830, 353–357. Lin, J.‐H., and Rydqvist, B. (2001). Characterization of a delayed rectifier potassium channel in the slowly adapting stretch receptor neuron of crayfish. Brain Res. 913, 1–9. Lin, J.‐H., Sand, P., and Rydqvist, B. (1999). Macrocurrents of voltage gated Naþ and Kþ channels from the crayfish stretch receptor neuronal soma. Neuroreport 10, 2503–2507. Martinac, B. (2004). Mechanosensitive ion channels: Molecules of mechanotransduction. J. Cell Sci. 117, 2449–2460. Martinac, B., Buechner, M., Delcour, A. H., Adler, J., and Kung, C. (1987). Pressure sensitive ion channels in Escherichia coli. Proc. Natl. Acad. Sci. USA 84, 2297–2301. Martinac, B., Adler, J., and Kung, C. (1990). Mechanosensitive ion channels of E. Coli activated by amphipaths. Nature 348, 261–263. Nakajima, S., and Onodera, K. (1969a). Membrane properties of the stretch receptor neurones of crayfish with particular reference to mechanisms of sensory adaptation. J. Physiol. 200, 161–185.
2. Stretch Receptor Mechanotransduction
47
Nakajima, S., and Onodera, K. (1969b). Adaptation of the generator potential in the crayfish stretch receptors under constant length and constant tension. J. Physiol. 200, 187–204. Obara, S. (1968). EVects of some organic cations on generator potential of crayfish stretch receptor. J. Gen. Physiol. 52, 363–386. Ottoson, D., and Rydqvist, B. (1978). The eVects of Triton‐detergents on the stretch receptor of the crayfish. Acta Physiol. Scand. 103, 9–18. Ottoson, D., and Swerup, C. (1982). Studies of the role of calcium in adaptation of the crustacean stretch receptor. EVects of intracellular injection of calcium, EGTA and TEA. Brain Res. 244, 337–341. Ottoson, D., and Swerup, C. (1985a). Ionic dependence of early adaptation in the crustacean stretch receptor. Brain Res. 336, 1–8. Ottoson, D., and Swerup, C. (1985b). EVects of intracellular TEA injections on early adaptation of crustacean stretch receptor. Brain Res. 336, 9–17. Purali, N. (2005). Structure and function relationship in the abdominal stretch receptor organs of the crayfish. J. Comp. Neurol. 488, 369–383. Purali, N., and Rydqvist, B. (1992). Block of potassium outward currents in the crayfish stretch receptor neurons by 4‐aminopyridine, tetraethylammonium chloride and some other chemical substances. Acta Physiol. Scand. 146, 67–77. Purali, N., and Rydqvist, B. (1998). Action potential and sodium current in the slowly and rapidly adapting stretch receptor neurons of the crayfish (Astacus astacus). J. Neurophysiol. 80, 2121–2132. Rydqvist, B. (1992). Muscle mechanoreceptors in invertebrates. In ‘‘Advances in Comparative and Environmental Physiology’’ (F. Ito, ed.), Vol. 10, pp. 233–260. Springer‐Verlag, Berlin Heidelberg. Rydqvist, B., and Purali, N. (1991). Potential dependent outward potassium currents in the rapidly adapting stretch receptor neuron. Acta Physiol. Scand. 142, 67–76. Rydqvist, B., and Purali, N. (1993). Transducer properties of the rapidly adapting stretch receptor of the crayfish (Pacifastacus leniusculus). J. Physiol. 469, 193–211. Rydqvist, B., and Swerup, C. (1991). Stimulus response properties of the slowly adapting stretch receptor of the crayfish. Acta Physiol. Scand. 143, 11–19. Rydqvist, B., and Zhou, J.‐Y. (1989). Potential dependent outward potassium currents in the slowly adapting crayfish stretch receptor neuron. Acta Physiol. Scand. 137, 409–419. Rydqvist, B., Swerup, C., and La¨nnergren, J. (1991). Viscoelastic properties of the receptor muscle of the slowly adapting stretch receptor organ of the crayfish. Acta Physiol. Scand. 143, 11–19. Rydqvist, B., Purali, N., and La¨nnergren, J. (1994). Viscoelastic properties of the receptor muscle of the rapidly adapting stretch receptor organ of the crayfish. Acta Physiol. Scand. 150, 151–159. Rydqvist, B., Swerup, C., and Sand, P. (2003). Voltage gated ion channels in transduction and adaptation in crayfish stretch receptor. In ‘‘Proc. 3rd Feps Congress’’ (P. Poujeol and O. Petersen, eds.), pp. 195–199. Monduzzi Editore S.p.A., Bologna, Italy. Sokabe, M., and Sachs, F. (1992). Towards molecular mechanism of activation in mechanosensitive ion channels. In ‘‘Advances in Comparative and Environmental Physiology’’ (F. Ito, ed.), Vol. 10, pp. 55–77. Springer‐Verlag, Berlin, Heidelberg. Stevens, S. S. (1957). On the psychophysical law. Psychol. Rev. 64, 153–181. Suchyna, T. M., Johnson, J. H., Hamer, K., Leykam, J. F., Gage, D. A., Clemo, H. F., Baumgarten, C. M., and Sachs, F. (2000). Identification of a peptide toxin from Grammostola spatulata spider venom that blocks cation‐selective stretch‐activated channels. J. Gen. Physiol. 115, 583–598.
48
Bo Rydqvist
Sukharev, S. I., and Corey, D. P. (2004). Mechanosensitive channels: Multiplicity of families and gating paradigms. Sci. STKE 219, 1–24, re4. Sukharev, S. I., Blount, P., Martinac, B., Blattner, F. R., and Kung, C. (1994). A large conductance mechanosensitive channel in E. coli encodes by mscL alone. Nature 368, 265–268. Swerup, C., and Rydqvist, B. (1992). The abdominal stretch receptor organ of the crayfish. Comp. Biochem. Physiol. 103A, 423–431. Swerup, C., and Rydqvist, B. (1996). A mathematical model of the crustacean stretch receptor organ. Biomechanics of the receptor muscle, mechanosensitive ion channels, and mechanotransducer properties. J. Neurophysiol. 76, 2211–2220. Swerup, C., Purali, N., and Rydqvist, B. (1991). Block of receptor response in the stretch receptor neurone of the crayfish by gadolinium. Acta Physiol. Scand. 143, 21–26. Tao‐Cheng, J.‐H., Hirosawa, K., and Nakajima, Y. (1981). Ultrastructure of the crayfish stretch receptor in relation to its function. J. Comp. Neurol. 200, 1–21. Walker, R. G., Willingham, A. T., and Zucker, C. S. (2004). A Drosophila mechanosensory transduction channel. Science 287, 2229–2234. Wiersma, C. A. G., Furshpan, E., and Florey, E. (1953). Physiological and pharmacological observations on muscle receptor organs of the crayfish, Cambarus Clarkii Girard. J. Exp. Biol. 30, 136–150.
CHAPTER 3 Mechanosensitive Ion Channels in Caenorhabditis elegans Dafne Bazopoulou and Nektarios Tavernarakis Institute of Molecular Biology and Biotechnology, Foundation for Research and Technology, Heraklion 71110, Crete, Greece
I. II. III. IV.
Overview Introduction C. elegans Mechanosensitive Behaviors C. elegans DEG/ENaCs A. MEC‐4 and MEC‐10 B. UNC‐8 and DEL‐1 C. UNC‐105 V. C. elegans TRP Ion Channels A. OSM‐9 and OCR‐2 B. TRP‐4 VI. Concluding Remarks References
I. OVERVIEW Caenorhabditis elegans depends critically on mechanosensory perception to negotiate its natural habitat, the soil. The worm displays a rich repertoire of mechanosensitive behaviors, which can be easily examined in the laboratory. This, coupled with the availability of sophisticated genetic and molecular biology tools, renders C. elegans a particularly attractive model organism to study the transduction of mechanical stimuli to biological responses. Systematic genetic analysis has facilitated the dissection of the molecular mechanisms that underlie mechanosensation in the nematode. Studies of various worm mechanosensitive behaviors have converged to
Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59003-9
50
Bazopoulou and Tavernarakis
identify highly specialized, plasma membrane ion channels that are required for the conversion of mechanical energy to cellular signals. Strikingly, similar mechanosensitive ion channels appear to function at the core of the mechanotransduction apparatus in higher organisms, including humans. Thus, the mechanisms responsible for the detection of mechanical stimuli are likely conserved across metazoans. The nematode oVers a powerful platform for elucidating the fundamental principles that govern the function of metazoan mechanotransducers. In this chapter, we survey the current understanding of mechanotransduction in C. elegans and focus on the role of mechanosensitive ion channels in specific mechanosensory behavioral responses. Further, we aspire to highlight potential unifying themes, common to mechanosensory transduction in diverse species.
II. INTRODUCTION C. elegans is a small soil‐dwelling nematode worm, with a simple body plan that is formed by just 959 somatic cells. C. elegans is primarily a hermaphroditic species but males, which can mate with hermaphodites are also found in natural populations at very low frequency. The transparent nature of both the egg and the cuticle of this nematode have facilitated exceptionally detailed developmental characterization of the animal. The complete sequence of cell divisions and the normal pattern of programmed cell deaths that occur as the fertilized egg develops into the 959‐celled adult are known (Sulston and Horvitz, 1977; Sulston et al., 1983). The anatomical characterization and understanding of neuronal connectivity in C. elegans are unparalleled in the metazoan world. Serial section electron microscopy has identified the pattern of synaptic connections made by each of the 302 neurons of the animal (including 5000 chemical synapses, 600 gap junctions, and 2000 neuromuscular junctions) so that the full ‘‘wiring diagram’’ of the animal is known (White et al., 1976, 1986). Although the overall number of neurons is small, 118 diVerent neuronal classes, including many neuronal types present in mammals, can be distinguished. Other animal model systems contain many more neurons of each class (there are about 10,000 more neurons in Drosophila with approximately the same repertoire of neuronal types). Thousands of mutations that disrupt development or various behaviors have been identified and positioned on a detailed genetic map (Brenner, 1974). Sequencing and high‐quality annotation of the complete genome organized in six chromosomes (five autosomes and the sex chromosome X) have been accomplished (The C. elegans Sequencing Consortium, 1998); http://www.wormbase.org). Primary cell culture methodologies are available
51
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
for the analysis of specific groups of cells and neurons ex vivo (Christensen et al., 2002). Electrophysiological study of nematode neurons and muscles has also become possible (Richmond and Jorgensen, 1999; O’Hagan et al., 2005). Overall, the broad range of genetic and molecular tools available in C. elegans allows in‐depth investigation of the cellular mechanisms underlying mechanotransduction.
III. C. ELEGANS MECHANOSENSITIVE BEHAVIORS Despite its anatomical simplicity, C. elegans displays an impressive repertoire of mechanosensitive behaviors (Table I). When touched gently on the posterior, an animal will move forward; when touched on the anterior body,
TABLE I Main C. elegans Mechanosensitive Behaviors Mechanosensitive behavior
Stimulus
Mechanosensory neurons
References
Gentle body touch response
Light touch on the body
ALM, AVM, PLM, PVM
Chalfie et al., 1985
Harsh touch response
Prodding with a stiV object on the body
PVD, PVC, others?
Way and Chalfie, 1989; Chalfie and Wolinsky, 1990
Head‐on collision response
Nose tip collision with an obstacle
ASH, FLP, OLQ
Kaplan and Horvitz, 1993; Colbert et al., 1997; Hart et al., 1999
Head withdrawal response
Light touch on nose side during foraging
OLQ, IL1
Kaplan and Horvitz, 1993; Hart et al., 1995
Proprioception
Muscle contractions and relaxations
Ventral nerve cord motorneurons, DVA
Wolinsky and Way, 1990; Francis and Waterston, 1991; Hresko et al., 1994; Tavernarakis et al., 1997; Li et al., 2006
Tap withdrawal reflex
Vibrations (taps) through the culture substrate
ALM, PVM, PLM, AVD
Wicks and Rankin, 1995
Basal slowing response
Mechanical input from the culture substrate texture (i.e., the presence of a bacterial lawn)
CEP, ADE, PDE
Sawin et al., 2000
52
Bazopoulou and Tavernarakis
it will move backward. This gentle body touch is sensed by the six touch receptor neurons (ALML/R, anterior lateral microtubule cell, left/right; AVM, anterior ventral microtubule cell; PLML/R, posterior lateral microtubule cell, left/right; PVM, posterior ventral microtubule cell; Chalfie, 1993, 1995; Chalfie and Sulston, 1981). The touch receptors are situated so that their processes run longitudinally along the body wall embedded in the hypodermis adjacent to the cuticle (Fig. 1). The position of the processes along the body axis correlates with the sensory field of the touch cell. Laser ablation of touch receptors, which have sensory receptor processes in the anterior half of the body, eliminates anterior touch sensitivity and laser ablation of the touch receptors, which have posterior dendritic processes, eliminates posterior touch sensitivity. In addition to mediating touch avoidance, the touch receptor neurons appear to control the spontaneous rate of locomotion since animals that lack functional touch cells are lethargic. The mechanical stimuli that drive spontaneous locomotion are unknown but could include encounters with objects in their environments or body stretch induced by locomotion itself. A
B
PVM AVM
ALMR
ALML
PLMR
PLML
FIGURE 1 The C. elegans touch receptor neurons. (A) Visualization of touch receptors. Worms are expressing the green fluorescent protein (GFP) under the control of the mec‐4 promoter, which is active only in the six touch receptor neurons. Arrows indicate touch receptor cell bodies. Some touch receptor axons are apparent. (B) Schematic diagram showing the position of the six touch receptor neurons in the body of the adult nematode. Note the two fields of touch sensitivity defined by the arrangement of these neurons along the body axis. The ALMs and AVM mediate the response to touch over the anterior field, whereas PLMs mediate the response to touch over the posterior field.
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
53
Animals with defective touch receptor neurons can still respond to a harsh stimulus (push with a platinum wire; Way and Chalfie, 1989; Chalfie and Wolinsky, 1990). This is indicative of the presence of a separate neuronal circuit, which is responsible for harsh touch sensitivity. Worms also respond to mechanical stimuli applied at the tip of their head by initiating a backward movement. This behavior known as nose touch response is mediated by nose touch neurons (Kaplan and Horvitz, 1993; Colbert et al., 1997). The nose of C. elegans is highly sensitive to mechanical stimuli. This region of the body is innervated by many sensory neurons which mediate mechanosensitivity. Responses to touch in the nose can be classified into two categories, the head‐on collision response and the foraging and head withdrawal response (Wicks and Rankin, 1995; Colbert et al., 1997; Bargmann and Kaplan, 1998; Hart et al., 1999). Additional mechanosensitive behaviors include proprioception (the regulation of coordinated locomotion), the tap withdrawal reflex, and the basal slowing response (Chiba and Rankin, 1990; Liu and Sternberg, 1995; Tavernarakis et al., 1997; Wicks and Rankin, 1997; Sawin et al., 2000). In the laboratory, C. elegans moves through a bacterial lawn on a Petri plate with a readily observed sinusoidal motion. Proprioception facilitates the coordinated movement of body parts by synchronization of muscle contractions that produce the characteristic sinusoidal locomotory pattern of the nematode. Interactions between excitatory and inhibitory motorneurons produce a pattern of alternating dorsal and ventral contractions (Francis and Waterston, 1991; Hresko et al., 1994). Distinct classes of motorneurons control dorsal and ventral body muscles. To generate and sustain the sinusoidal pattern of movement, the contraction of the dorsal and ventral body muscles must be out of phase. For example, to turn the body dorsally, the dorsal muscles contract while the opposing ventral muscles relax. The adult motor system involves five major types of ventral nerve cord motorneurons defined by axon morphologies and patterns of synaptic connectivity. The tap withdrawal reflex is a mechanosensitive behavior triggered by mechanical stimuli delivered as vibrations (taps) through the Petri dish and the agar medium on which the worms move. The response to taps consists of either accelerations or reversals (Wicks and Rankin, 1995). The basal slowing response occurs when moving worms encounter a bacterial lawn and is regulated by a circuit of dopaminergic mechanosensory neurons. Animals moving at high speed in the absence of food slow down when they enter a bacterial lawn. It is likely that mechanosensory input originating from textural diVerences in the substrate between areas with and without food drives this response. Indeed, the same response is observed if a lawn of sepharose beads is used instead of bacteria (Sawin et al., 2000).
54
Bazopoulou and Tavernarakis TABLE II Ion Channels Implicated in Mechanosensation, in C. elegans
Expression pattern
Associated mechanosensitive behavior
MEC‐4
þ
Epithelial Na channel (degenerin)
Touch receptor neurons
Gentle body touch response
Chalfie and Au, 1989; Driscoll and Chalfie, 1991; Hamill et al., 1992; O’Hagan et al., 2005
MEC‐10
Epithelial Naþ channel (degenerin)
Touch receptor neurons
Gentle body touch response
Huang and Chalfie, 1994; O’Hagan et al., 2005
UNC‐8
Epithelial Naþ channel (degenerin)
Motorneurons, interneurons, nose mechanosensory neurons
Proprioception, coordinated locomotion
Park and Horvitz, 1986b; ShreZer et al., 1995; Tavernarakis et al., 1997
DEL‐1
Epithelial Naþ channel (degenerin)
Motorneurons
Proprioception, coordinated locomotion
Tavernarakis et al., 1997
UNC‐105
Epithelial Naþ channel (degenerin)
Body wall muscles
Coordinated locomotion
Park and Horvitz, 1986a; Liu et al., 1996; Garcia‐Anoveros et al., 1998
OSM‐9
TRPV Ca2þ channel
Nose mechanosensory neurons, chemosensory neurons, osmosensory neurons
Nose touch response, nociception
Colbert et al., 1997; Tobin et al., 2002; Zhang et al., 2004b
OCR‐2
TRPV Ca2þ channel, OSM‐9/ capsaicin receptor related protein
Amphid sensory neurons, phasmid neurons
Nose touch response
Tobin et al., 2002
TRP‐4
TRPN Ca2þ channel
Dopaminergic mechanosensory neurons, interneurons
Proprioception, coordinated locomotion
Li et al., 2006
Ion channel
Sequence similarity
References
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
55
This behavior allows animals to spend more time in food‐rich areas and facilitates foraging. C. elegans displays several additional behaviors that are based on sensory mechanotransduction and have been characterized to a lesser extend. For example, mechanotransduction appears to play a regulatory role in processes such as matting, egg laying, feeding, defecation, and maintenance of the pseudocoelomic body cavity pressure (Thomas, 1990; Avery, 1993; Bargmann and Kaplan, 1998; Syntichaki and Tavernarakis, 2004). These behaviors add to the large repertoire of mechanosensitive phenomena, amenable to genetic and molecular dissection in the nematode. Extensive genetic studies have culminated in the identification and characterization of several genes, which encode components of specialized ion channels that mediate mechanosensitive behaviors in C. elegans. Similar channels with mechanosensitive properties have also been identified in diverse organisms including snails, flies, and vertebrates, and fall in two distinct classes: the degenerin (DEG)/epithelial Naþ channel (ENaC) family and the transient receptor potential (TRP) family of ion channels (Table II). Below, we review the role of these mechanosensitive ion channels in specific C. elegans mechanosensory behavioral responses and discuss the molecular mechanisms that govern the function of nematode mechanotransducers.
IV. C. ELEGANS DEG/ENaCs The DEG/ENaC family of ion channels is a large group of proteins sharing a high degree of sequence and overall structure similarity. Members of the DEG/ENaC family have been identified in organisms ranging from nematodes, snails, flies, and many vertebrates including humans and are expressed in tissues as diverse as kidney, epithelia, muscles, and neurons (reviewed by Kellenberger and Schild, 2002). Specific C. elegans ion channels are referred to as degenerins because unusual, gain‐of‐function mutations in several family members induce swelling or cell death (Chalfie and Wolinsky, 1990). C. elegans degenerins exhibit 25–30% sequence identity to subunits of the vertebrate amiloride sensitive, ENaCs, which are required for ion transport across epithelia (Hummler and Horisberger, 1999) and acid‐ sensing ion channels that may contribute to pain perception and mechanosensation (Waldmann and Lazdunski, 1998; Hummler and Horisberger, 1999; Kellenberger and Schild, 2002). Despite their functional diversity they share a few common properties such as Naþ selectivity and inhibition by amiloride, in addition to a highly
56
Bazopoulou and Tavernarakis
conserved overall structure. DEG/ENaC proteins range from about 550 to 950 amino acids in length and share several distinguishing blocks of sequence similarity. Subunit topology is invariable: all DEG/ENaC family members have two membrane‐spanning domains (MSDs) with cysteine‐rich domains (CRDs, the most conserved is designated CRD3) situated between these two transmembrane segments. DEG/ENaCs are situated in the membrane such that N‐ and C‐termini project into the intracellular cytoplasm while most of the protein, including the CRDs, is extracellular (Fig. 2). A MSD I
CRD I
CRD II
CRD III
MSD II
N-terminus
C-terminus
Thiol protease motif
Extracellular regulatory domain Neurotoxin-like domain (ERD) (NTD)
B
Membrane
Ala
Cytoplasm COOH NH2
FIGURE 2 Schematic representation of DEG/ENaC ion channel subunit structure and topology. (A) Functional/structural domains. Colored boxes indicate defined channel modules. These include the two membrane‐spanning domains (MSDs; dark‐blue shading) and the three cysteine‐rich domains (CRDs; red shading; the first CRD is absent in mammalian channels and is depicted by light red shading). The small light‐blue oval depicts the putative extracellular regulatory domain (ERD). The green box overlapping with CRDIII denotes the neurotoxin‐ related domain (NTD). The conserved intracellular region with similarity to thiol‐protease histidine active sites is shown in yellow. Shown in pink is the N‐terminal domain. (B) Transmembrane topology. Both termini are intracellular with the largest part of the protein situated outside the cell. The dot near MSD II represents the amino acid position (alanine 713 in MEC‐4) aVected in dominant, toxic degenerin mutants.
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
57
Highly conserved regions include the two MSDs (MSD I and II), a short amino acid stretch before the first MSD, the extracellular CRDs, an extracellular regulatory domain (ERD), and a neurotoxin‐related domain (NTD) before predicted transmembrane domain II (Tavernarakis and Driscoll, 2000; Tavernarakis et al., 2001). The high degree of conservation of cysteine residues in these extracellular domains suggests that the tertiary structure of this region is critical to the function of most channel subunits and may mediate interactions with extracellular structures. The strong sequence and structure conservation across species suggests that DEG/ENaC family members shared a common ancestor relatively early in evolution (Fig. 3). DEG/ENaC ion channels have been associated with mechanosensory responses in nematodes, flies, and mammals (Tavernarakis and Driscoll, 2001; Kellenberger and Schild, 2002; Syntichaki and Tavernarakis, 2004). At present, 30 genes encoding DEG/ENaC ion channels have been identified in the C. elegans genome. Genetic, molecular, and electrophysiological studies have implicated five nematode degenerins in mechanotransduction (DEL‐1, MEC‐4, MEC‐10, UNC‐8, and UNC‐105; Table II; reviewed by Syntichaki and Tavernarakis, 2004). Below, we discuss the role of degenerins in C. elegans mechanosensory behaviors.
A. MEC‐4 and MEC‐10 Genetic analysis revealed 18 genes, which, when mutated, disrupt specifically the gentle body touch sensation (Ernstrom and Chalfie, 2002). These genes are therefore thought to encode candidate mediators of touch sensitivity (these genes were named mec genes since when they are defective, animals are mechanosensory abnormal; Chalfie and Au, 1989). Almost all of the mec genes have now been molecularly identified and most of them encode proteins postulated to make up a touch‐transducing complex (Gu et al., 1996; Syntichaki and Tavernarakis, 2004). The core elements of this mechanosensory complex are the channel subunits MEC‐4 and MEC‐10, which can interact genetically and physically (Ernstrom and Chalfie, 2002; Goodman et al., 2002). Both these proteins are DEG/ENaC family members and interact to form the putative mechanotransducer in C. elegans touch receptor neurons, together with two other structural components, the stomatin‐like protein MEC‐2, and the paraoxonase‐like protein MEC‐6 (Chelur et al., 2002; Goodman et al., 2002). Loss‐of‐function mutations in mec‐4 or mec‐10 do not aVect the development and utlrastructure of the touch receptor neurons but render the animals touch insensitive (Chalfie and Au, 1989; Chalfie, 1995). The plasma membrane topology of these molecules has been elucidated by performing antibody and protease experiments (Lai et al., 1996). Evidence that MEC‐4
58
Bazopoulou and Tavernarakis F28A12.1FLR-1 C24G7.4 C24G7.1
C27C12.5 T28D9.7 F25D1.4 F26A3.6 F59F3.4
C24G7.2 F23B2.3 T28F2.7
C18B2.6
PPK
C11E4.4
RPK
T21C9.3
C46A5.2
D mNaCh
C11E4.3 aENaC FANaCh dENaC bENaC
ASIC-1
gENaC T28F4.2 b ASIC T28B8.5
BNaC2 BNaC1
UNC-105 ACCN3 ASIC3 Y69H2.13 F55G1.12 MEC-10 MEC-4 0.1
Y69H2.11
UNC-8 DEL-1 DEG-1
Y69H2.2
FIGURE 3 Phylogenetic relations among DEG/ENaC proteins. The nematode degenerins are shown with blue lines. The current degenerin content of the complete nematode genome is included. The seven genetically characterized (DEG‐1, DEL‐1, FLR‐1, MEC‐4, MEC‐10, UNC‐8, and UNC‐105) are shown in red. Representative DEG/ENaC proteins from a variety of organisms, ranging from snails to humans, are also included (mammalian: red lines; fly: green lines; snail: orange line). The scale bar denotes evolutionary distance equal to 0.1‐nucleotide substitutions per site.
and MEC‐10 coassemble into the same channel complex include that: (1) MEC‐4 and MEC‐10 subunits are coexpressed in the touch receptor neurons (Huang and Chalfie, 1994), (2) MEC‐4 and MEC‐10 proteins translated in vitro in the presence of microsomes can coimmunoprecipitate (Goodman et al., 2002), and (3) genetic interactions between mec‐4 and mec‐10 have been observed (Gu et al., 1996). MEC‐4 exhibits a punctuate distribution along the axon of the touch receptor neurons which may represent the subcellular localization of the mechanotransducing complexes (Fig. 4).
AVM
10 mm
FIGURE 4 Punctate localization of a putative mechanosensitive ion channel subunit. Image of an AVM touch receptor neuron expressing a GFP‐tagged MEC‐4 protein. Fluorescence is unevenly distributed along the process of the neuron in distinct puncta, which may represent the location of the mechanotransducing apparatus.
60
Bazopoulou and Tavernarakis
The MEC‐4/MEC‐10 mechanically gated ion channel is sensitive to the diuretic amiloride, which is a general inhibitor of mechanosensitive ion channels (Hamill et al., 1992; O’Hagan et al., 2005). It is proposed that at least two MEC‐4 and MEC‐10 subunits contribute to the channel formation (Huang and Chalfie, 1994). MEC‐4 is required for touch neuron activity induced by light touch stimuli in vivo, as shown by measurements of physiological neural responses using a fluorescent calcium indicator reporter fusion (cameleon; Suzuki et al., 2003). Absence of MEC‐4 does not alter the basic physiology of the touch neurons or their responses in harsh touch stimuli. Whole‐cell patch clamp recordings from C. elegans touch receptor neurons, in vivo, provided experimental verification that the MEC‐4/MEC‐10 channel is actually mechanically gated. These studies show that the MEC‐4/MEC‐10 channel is directly activated by external forces, which results in the generation of mechanosensory currents carried by Naþ and blocked by amiloride (O’Hagan et al., 2005). Gain‐of‐function (dominant, d) mutations in mec‐4 induce necrotic cell death of the six touch receptor neurons (Syntichaki and Tavernarakis, 2003). Most such mutations encode substitutions of an alanine, adjacent to the second transmembrane domain, near the channel pore. Substitution of the small side chain alanine by a large side chain amino acid causes toxicity. Steric interference conferred by a bulky amino acid side chain causes the channel to close less eVectively. Increased cation influx initiates neurodegeneration. That ion influx is critical for degeneration is supported by the fact that amino acid substitutions that disrupt the channel conducting pore can prevent neurodegeneration when present in cis to the A713 substitution. Other C. elegans degenerin family members (e.g., deg‐1 and mec‐10) can be altered by analogous amino acid substitutions to induce neurodegeneration (Syntichaki and Tavernarakis, 2002). The mutant MEC‐4(d) Naþ channel conducts Ca2þ both when heterologously expressed in Xenopus oocytes and in vivo. Thus, Ca2þ influx via the MEC‐4(d) channel directly contributes to the Ca2þ increase in the cytoplasm and signals the initiation of necrosis (Bianchi et al., 2004). Necrosis induced by MEC‐4(d) is similar in several respects to that associated with the excitotoxic cell death that occurs in higher organisms in response to injury, in stroke, and so on. Intragenic second‐site mutations in mec‐4(d) that encode amino acid substitutions near the pore domain disrupt the function of the hyperactive MEC‐4(d) channel. Such mutations appear to influence the traYcking of the channel and suppress necrosis induced by mec‐4(d) mutants in a temperature‐dependent manner (Royal et al., 2005). MEC‐4 and MEC‐10 together with MEC‐2 and MEC‐6 form the mechanosensitive channel complex that is thought to be linked to the extracellular mantle and to the cytoskeleton (Savage et al., 1989; Du et al., 1996).
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
61
These interactions are facilitated by other auxiliary molecules both extracellularly and intracellularly and may serve to convey mechanical forces to the channel. mec‐2 encodes a 481‐amino acid protein and is expressed in the touch receptor neurons and in a few additional neurons in the nerve ring region (Huang et al., 1995; Gu et al., 1996; Du and Chalfie, 2001). The MEC‐2 protein appears to be localized along the length of the touch receptor process as well as in the cell body (Huang et al., 1995), and shares sequence similarity with human stomatin, a protein that has been implicated in regulating red blood cell plasma membrane conductance (Stewart, 1997). The mammalian stomatin physically interacts with G‐protein‐coupled receptors and colocalizes with glycophosphoinositol (GPI)‐anchored proteins and lipid rafts (Snyers et al., 1999; Tavernarakis et al., 1999; Sedensky et al., 2001). MEC‐2 features a central region that encompasses an SPFH domain with a membrane‐associated hydrophobic part (AA 114–141) and a cytoplasmic hydrophilic part that together exhibit 65% identity to stomatin (Huang et al., 1995; Tavernarakis et al., 1999). The SPFH domain is the common denominator of stomatins, prohibitins, flotilins, and bacterial HflK/C proteins, all of which are membrane‐associated regulators (Tavernarakis et al., 1999). MEC‐2 activates the MEC‐4 channel in Xenopus oocytes and coimmunoprecipitates with the other members of the mechanosensitive complex (Goodman et al., 2002). It is also required for neural responses to gentle mechanical stimuli in vivo (Suzuki et al., 2003). MEC‐2 interacts in vitro and colocalizes with MEC‐4 through the SPFH domain. This interaction is necessary for channel activation (Zhang et al., 2004a). mec‐6 encodes a protein that is partially related to paraoxonases/acetylesterases and physically interacts with MEC‐4 and MEC‐10 (Chelur et al., 2002). Although animals bearing recessive mec‐6 mutations are touch insensitive, the touch receptor neurons exhibit an apparent wild‐type ultrastructure (Chalfie and Sulston, 1981). How MEC‐6 contributes to channel function is not yet known. In addition to MEC‐4, MEC‐10, MEC‐2, and MEC‐6, mechanotransduction in the touch receptor neurons also requires two groups of peripheral‐ associated proteins encoded by mec genes: the intracellular proteins MEC‐7 and MEC‐12 and the extracellular proteins MEC‐1, MEC‐5, and MEC‐9 (reviewed by Syntichaki and Tavernarakis, 2004). The mec‐7 and mec‐12 genes encode a b‐ and an a‐tubulin, respectively, expressed at high levels in the touch receptor neurons (Savage et al., 1989, 1994; Hamelin et al., 1992; Fukushige et al., 1999). These tubulins assemble to form 15‐protofilament microtubules specific to touch receptor neurons. mec‐7 and mec‐12 mutations, which cause a touch‐insensitive phenotype, disrupt tubulin subunit interactions, and protofilament assembly (Savage et al., 1989, 1994; Gu et al., 1996). The role of these microtubules in mechanosensation remains to be
62
Bazopoulou and Tavernarakis
determined. Perhaps these specialized structures are tethered to the mechanosensitive MEC‐4/MEC‐10 ion channel, providing an intracellular anchor required for channel gating. mec‐1 encodes an extracellular matrix (ECM) protein with multiple epidermal growth factor (EGF) and Kunitz domains (Emtage et al., 2004). In mec‐1 mutants, touch cells lack the mantle and other specializations of the cuticle and have displaced processes (Chalfie and Sulston, 1981; Chalfie, 1993; Savage et al., 1994; Gu et al., 1996). MEC‐1 colocalizes with MEC‐5 and the mechanosensory complex in the touch neurons (Emtage et al., 2004). The mec‐5 gene encodes a collagen that is secreted by cells of the hypodermis (Du et al., 1996). These two ECM components are required for the correct localization of the degenerin channel (Emtage et al., 2004). The mec‐9 gene encodes two transcripts which direct the synthesis of proteins secreted by the touch receptor neurons (Chalfie and Sulston, 1981; Du et al., 1996). MEC‐ 9L (encoded by one of the two mec‐9 transcripts) contains several domains related to the Kunitz type serine protease inhibitor domain, a Ca2þ‐binding EGF repeat, a non‐Ca2þ‐binding EGF repeat, and a glutamic acid‐rich domain (Du et al., 1996). How the extracellular MEC‐1, MEC‐5, and MEC‐9 proteins influence the activity of the MEC‐4/MEC‐10 ion channel is not known. It is proposed that these proteins are components of the ECM and collectively serve to anchor the channel to extracellular structures and convey external mechanical forces to the core mechanotransducer complex (Fig. 5).
B. UNC‐8 and DEL‐1 C. elegans shows a characteristic sinusoidal pattern of locomotion. Little is known about how the sinusoidal wave is propagated along the body axis. Adjacent muscle cells are electrically coupled via gap junctions, which could couple excitation of adjacent body muscles. Alternatively, ventral cord motorneurons could promote wave propagation since gap junctions connect adjacent motorneurons of a given class (White et al., 1976, 1986; Chalfie et al., 1985). A third possibility is that motorneurons could themselves act as stretch receptors so that contraction of body muscles could regulate adjacent motorneuron activities, thereby propagating the wave (Tavernarakis et al., 1997; Syntichaki and Tavernarakis, 2004). The adult neuronal circuit for locomotion comprises five major types of ventral nerve cord motorneurons (A motorneurons—12VA and 9 DA; B motorneurons—11VB and 7DB; D motorneurons—13 VD and 6 DD; AS motorneurons; and VC motorneurons; Francis and Waterston, 1991; Hresko et al., 1994). Mutations that aVect the neuronal circuit for locomotion disrupt the sinusoidal pattern of
63
3. Mechanosensitive Ion Channels in Caenorhabditis elegans Pressure Cuticle hypodermis Extracellular anchors Linker
Extracellular anchors Linker
Mantle
Membrane Linker
Degenerin channel
Degenerin channel
Linker Cytoplasm Na+
Cytoskeleton No stimulation (channel closed)
Cytoskeleton Stimulation (channel open)
FIGURE 5 A mechanotransducing complex in C. elegans touch receptor neurons. In the absence of mechanical stimulation, the channel is closed and therefore the sensory neuron is idle. Application of a mechanical force to the body of the animal results in distortion of a network of interacting molecules that opens the degenerin channel. Naþ influx depolarizes the neuron, initiating the perceptory integration of the stimulus.
movement and generate locomotory defects, uncoordination, and paralysis (Park and Horvitz, 1986b; Tavernarakis and Driscoll, 1997). Gain‐of‐function mutations in the unc‐8 gene (unc‐8(sd)) induce transient neuronal swelling and severe uncoordination (Park and Horvitz, 1986a; ShreZer et al., 1995; ShreZer and Wolinsky, 1997). unc‐8 encodes a degenerin, which shares high sequence similarity to other DEG/ENaC family members as well as the same overall structure and topology (two transmembrane domains, three Cysteine‐rich regions, and large extracellular region). It is expressed in several motorneuron classes and in some interneurons and nose touch sensory neurons (Tavernarakis et al., 1997). Interestingly, semidominant unc‐8 alleles alter an amino acid in the region hypothesized to be an extracellular channel‐closing domain, defined in studies of deg‐1 and mec‐4 degenerins (Garcia‐Anoveros et al., 1995; Tavernarakis et al., 1997). Another degenerin family member, del‐1 (for degenerin‐like) is coexpressed in a subset of neurons that express unc‐8 (the VA and VB motorneurons) and is likely to assemble into a channel complex with UNC‐8 in these cells (Tavernarakis et al., 1997). unc‐8 null mutants have a subtle locomotion defect (Tavernarakis et al., 1997). Wild‐type animals move through an E. coli lawn with a characteristic
64
Bazopoulou and Tavernarakis
sinusoidal pattern. unc‐8 null mutants inscribe a path in an E. coli lawn that is markedly reduced in both wavelength and amplitude as compared to wild type (Fig. 6). This phenotype indicates that the UNC‐8 degenerin channel functions to modulate the locomotory trajectory of the animal. How does the UNC‐8 motorneuron channel influence locomotion? One highly interesting morphological feature of some motorneurons (in particular, the VA and VB motorneurons that coexpress unc‐8 and del‐1) is that their processes include extended regions that do not participate in neuromuscular junctions or neuronal synapses. These ‘‘undiVerentiated’’ process regions have been hypothesized to be stretch‐sensitive (discussed in White et al., 1976). Given the morphological features of certain motorneurons and the sequence similarity of UNC‐8 and DEL‐1 to candidate mechanically gated channels, we have proposed that these subunits coassemble into a stretch‐sensitive channel that might be localized to the undiVerentiated regions of the motorneuron process (Tavernarakis et al., 1997; reviewed by Syntichaki and Tavernarakis, 2004). When activated by the localized body stretch that occurs during locomotion, this motorneuron channel potentiates signaling at the neuromuscular junction, which is situated at a distance from
A
Wild type
B
unc-8(lf) FIGURE 6 Proprioception in the nematode. (A) Wild‐type animals inscribe a sinusoidal track as they move on an agar plate evenly covered with an E. coli bacterial lawn. (B) The characteristic properties (amplitude and wavelength) of tracks inscribed by unc‐8(lf) mutants are drastically reduced.
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
65
the site of the stretch stimulus. In the absence of the stretch activation, the body wave and locomotion still occur, but with significantly reduced amplitude because the potentiating stretch signal is not transmitted. This model bears similarity to the chain reflex mechanism of movement pattern generation. However, it does not exclude a central oscillator that would be responsible for the rhythmic locomotion. Instead, we suggest that the output of such an oscillator is further enhanced and modulated by stretch‐sensitive motorneurons. Genetic data indicates that UNC‐8 interacts with UNC‐1, a protein which is similar to MEC‐2 and has an important role in determining volatile anesthetic sensitivity (Huang et al., 1995; Rajaram et al., 1998). UNC‐1 is a close homologue of the mammalian stomatin protein (Rajaram et al., 1998). UNC‐8 and UNC‐1 colocalize along with another stomatin‐like protein, UNC‐24, in lipid rafts isolated from C. elegans. unc‐1 mutations eliminate UNC‐8 from these structures (Sedensky et al., 2004). unc‐24 is expressed in a variety of motorneurons, interneurons, and sensory neurons, including the touch receptor neurons (Barnes et al., 1996; Zhang et al., 2002). Mutations in unc‐24 severely aVect forward locomotion. Similarly to UNC‐1, UNC‐24 also aVects anesthetic sensitivity and is required for the distribution of UNC‐1 in the lipid rafts (Sedensky et al., 2004). These findings suggest that, in motorneurons, UNC‐1 may play a role analogous to that of MEC‐2 in touch receptor neurons; tethering the UNC‐8/DEL‐1 ion channels to intracellular structures. One important corollary of the unc‐8 mutant studies is that the UNC‐8 channel does not appear to be essential for motorneuron function; if this was the case, animals lacking the unc‐8 gene would be severely paralyzed. This observation strengthens the argument that degenerin channels function directly in mechanotransduction, rather than merely serving to maintain the osmotic environment so that other channels can function. The model of UNC‐8 and DEL‐1 functions that is based on mutant phenotypes, cell morphologies, and molecular properties of degenerins remains to be tested by determining subcellular channel localization, subunit associations and, most importantly, channel‐gating properties.
C. UNC‐105 The unc‐105 gene encodes a member of the DEG/ENaC family of ion channels and is mainly expressed in body wall muscles of C. elegans, where it is believed to mediate stretch sensitivity (Park and Horvitz, 1986a; Liu et al., 1996). UNC‐105 contains 150 amino acids at the C‐terminus that are not represented in other degenerin proteins. Although loss‐of‐function
66
Bazopoulou and Tavernarakis
mutations in unc‐105 do not result in any readily observable phenotype, gain‐ of‐function mutations cause muscle hyper contraction and result in severe paralysis of the animal (Park and Horvitz, 1986a). These mutations disrupt extracellular residues situated near the predicted transmembrane domain, where degeneration‐causing mutations are found in MEC‐4, MEC‐10, and DEG‐1. Therefore, these mutations in unc‐105 may result in constitutive channel activation producing the hypercontraction phenotype (Liu et al., 1996). The muscle hyper contraction phenotype of dominant unc‐105 mutations can be suppressed by mutations near the C‐terminus of let‐2, a gene that encodes the a2 chain of type‐IV collagen found in the basement membrane between muscle cells and the hypodermis (Liu et al., 1996). The nature of the functional link, implied by the suppression eVect, between UNC‐105 and LET‐2 collagen is unknown. A possible interpretation is that LET‐2 normally carries gating tension to the UNC‐105 channel when the muscle is stretched, thus providing regulatory feedback for muscle contraction (Liu et al., 1996). Suppressor mutations in LET‐2 may relieve conformational alterations to the UNC‐105 channel induced by dominant mutations, allowing the channel to close. This putative connection between a collagen and a degenerin is reminiscent of a similar relationship between the MEC‐5 collagen and MEC‐4 in touch receptor neurons (Tavernarakis and Driscoll, 1997). Similarly, mechanosensory transduction in the auditory system requires the extracellular tip links that physically deliver mechanical energy to the mechanosensitive channels in the hair cell stereocillia of the inner ear (Section V.B; Pickles and Corey, 1992; Pickles, 1993). Expression of the wild‐type unc‐105 gene in two heterologous systems [Xenopus oocytes and human embryonic kidney (HEK) cells] resulted in no detectable currents, suggesting that the channel requires a mechanical stimulus for gating (Garcia‐Anoveros et al., 1998). By contrast, expression of two mutant forms of unc‐105, carrying gain‐of‐function mutations predicted to cause constitutive activation, resulted in constitutive currents in both heterologous systems (Garcia‐Anoveros et al., 1998). These currents occurred without additional exogenous proteins, indicating that UNC‐105 channels can assemble as homomultimers, at least in oocytes and HEK cells. Phylogenetic analysis suggests that UNC‐105 is one of the most ancient degenerins, and thus may have not developed dependencies on other subunits (Corey and Garcia‐Anoveros, 1996).
V. C. ELEGANS TRP ION CHANNELS TRP proteins are a family of cation‐permeable channels that are present in diverse species ranging from yeast, flies, and worms to humans (Fig. 7). These channels bear structural similarities to the Drosophila TRP protein
67
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
mTRPM7 mTRPM5 TRP-3 dTRPL mTRPM4 mTRPM6 mTRPM8 mTRPC2 mTRPM2 TRP-1 mTRPM3 mTRPC6
TRPMs
mTRPM1
TRPCs
mTRPC3
dTRPM
mTRPC7 dTRP
GON-2 GTL-2
TRP-2
GTL-1
dTRP? mTRPC1
CED-11
mTRPC4 mTRPML3
TRPMLs
mTRPC5
mTRPML1
dPAIN
mTRPML2
TRP-4
CUP-5
dNOMPC
TRPNs
mTRPP3
zTRPN1
mTRPP2
TRPA-1
mTRPP5
mTRPA1
PKD-2
TRPPs
dTRPA1
TRPAs OSM-9
dAMO LOV-1 mTRPV6 mTRPV5 mTRPV2 mTRPV1
OCR-3 OCR-1
mTRPV3 dNAN mTRPV4
OCR-2 OCR-4
0.1
TRPVs
FIGURE 7 Phylogenetic relations among TRP proteins. Nematode TRPs are indicated with blue lines together with mammalian TRP representatives (red lines), fly TRPs (green lines), and zebrafish TRPs (yellow line). The scale bar indicates relative evolutionary distance equal to 0.1 nucleotide substitution per site.
which is a light‐activated Ca2þ channel, expressed in photoreceptor cells (Montell and Rubin, 1989; Hardie and Minke, 1992; Montell, 2001). TRPs can form homo‐ or heteromultimeric channels composed of two or more TRP subunits and can associate with other macromolecular complexes to serve diverse cellular functions. Members of TRP family respond to several types of input such as mechanical and thermal stimuli, pH fluctuations, Ca2þ and Mg2þ ions, fatty acids, and chemicals that evoke thermal‐like responses (Kahn‐Kirby and Bargmann, 2006). Thus, TRP ion channels have been implicated in many physiological processes such as mechanosensation, thermosensation, osmosensation, phototransduction, responses to pheromones, ion absorption and homeostasis, lysosomal traYcking, and neurotransmitter release.
68
Bazopoulou and Tavernarakis
The TRP family of ion channels comprises seven subfamilies six of which include proteins that are conserved among worm, flies, and mammals (TRPC, classical/short TRP; TRPV, vallinoid TRP; TRPM, long/melastatin TRP; TRPM, mucolipin TRP; TRPP, polycystin TRP; TRPA; reviewed by Montell, 2005). The remaining subfamily, TRPN, contains members that are present in invertebrates and zebrafish (Walker et al., 2000; Sidi et al., 2003; Li et al., 2006), while a mammalian homologue has not been discovered yet. An additional distantly related subfamily, TRPY, named after the first member, the yeast vacuolar protein, Yvc1, includes proteins found only in fungi (Palmer et al., 2001; Bonilla and Cunningham, 2002; Denis and Cyert, 2002). All TRP members appear to form tetrameric assemblies and include six predicted transmembrane domains and a variable number of ankyrin motifs, which are suggested to mediate protein–protein interactions. Members of individual subfamilies may bear several other domains, such as coiled‐coil motifs, protein kinase domains, transmembrane segments, and TRP domains (reviewed by Montell, 2005). Sequence similarity searches of the C. elegans genome have identified 24 genes predicted to encode TRP proteins which are representatives of all seven TRP subfamilies. All the proteins contain the core regions of the TRP members, which include the six transmembrane domains, the gate domains, the pore loop, and the ankyrin repeats, distributed along the N‐terminus. The C‐terminus varies among diVerent subfamilies and may contain coiled‐coil motifs, lipid‐binding domains, or other domains. Both the N‐ and C‐termini are intracellular (reviewed by Kahn‐Kirby and Bargmann, 2006). Three C. elegans TRP ion channels have been implicated in mechanotransduction (Fig. 8). OSM‐9 and OCR‐2 are members of the TRPV subfamily, and TRP‐4 belongs to the TRPN group (Kahn‐Kirby and Bargmann, 2006; Li et al., 2006). Other members of the TRP ion channel family in C. elegans include GON‐2 and GTL‐1 which belong to the TRPM group and are localized in intestinal epithelial cells, where they control electrolyte homeostasis (Teramoto et al., 2005). GON‐2 is also required for proper gonadal development (Sun and Lambie, 1997; West et al., 2001; Church and Lambie, 2003). C. elegans TRP‐ 1, TRP‐2, and TRP‐3 are similar to TRPC ion channels. trp‐1 is expressed in motorneurons, sensory neurons, and interneurons, as well as in vulval and intestinal muscles (Colbert et al., 1997). TRP‐3 is required for sperm‐egg interactions during fertilization (Xu and Sternberg, 2003). LOV‐1 and PDK‐2 are the nematode homologues of mammalian PDK‐1 and PDK‐2 TRPP ion channels, respectively (Corey, 2003). Mutations in the mammalian PDK‐1 or PDK‐2 result in autosomal dominant polycystic kidney disease (ADPKD). PDK‐1 and PDK‐2 form a Ca2þ‐permeable ion
69
3. Mechanosensitive Ion Channels in Caenorhabditis elegans OSM-9 Membrane
Cytoplasm NH2
COOH OCR-2
Membrane
Cytoplasm NH2
COOH TRP-4
Membrane
Cytoplasm NH2
COOH
FIGURE 8 Structure and topology of mechanosensitive TRP ion channel in C. elegans. Each protein contains six transmembrane domains, with the last two contributing to channel pore formation. The N-terminus is cytoplasmic and bears a variable number of ankyrin repeats (yellow circles). The C-terminus is also cytoplasmic and may contain several functional domains (see text) such as coiled coil domains (red box).
channel which is mechanically activated by fluid flow in certain epithelial cells (Nauli et al., 2003). LOV‐1 and PDK‐2 act in nematode mating. C. elegans males deficient in either or both LOV‐1 and PDK‐2 are defective in attaching to hermaphrodites and locating the vulva (Barr and Sternberg, 1999; Barr et al., 2001). Both proteins are localized in the cilia of sensory neurons in the male tail and to the CEM head neurons, consistent with a chemo‐ or mechanosensory function for these channels (Qin et al., 2001).
70
Bazopoulou and Tavernarakis
The single TRPML ion channels in C. elegans, CUP‐5, appears to be localized in lysosomes of many cell types. Mutations of cup‐5 result in defective endocytosis and degradation of proteins, and in the formation of large vacuoles (Fares and Greenwald, 2001; Hersh et al., 2002). cup‐5 null mutants cause maternal eVect lethality, with an excess in lysosomes and high levels of apoptosis, which is rescued by the expression of mammalian TRPLM homologues (Treusch et al., 2004). Mammalian TRPML1, TRPML2, and TRPML3 also colocalize in the lysosomes and when mutated, they cause mucolipidosis IV, a disorder characterized by lysosomal dysfunction which leads to neurodegeneration (Qian and Noben‐Trauth, 2005; Venkatachalam et al., 2006).
A. OSM‐9 and OCR‐2 OSM‐9 is the C. elegans homologue of the mammalian TRPV4 ion channel. The OSM‐9 protein contains six predicted MSDs, three ankyrin motifs at the N‐terminus, and a hydrophilic C‐terminal domain. The osm‐9 gene is expressed in ciliated sensory neurons including QLQ, FLP, ADL, ADF, AWA, and ASH (Colbert et al., 1997). QLQ and ASH are polymodal nociceptive neurons which detect mechanical stimuli, osmotic pressure, and various odorants. These neurons have also been implicated in the response to light touch in the nose (Kaplan and Horvitz, 1993). The FLP neuron is a sensory neuron, also involved in nose touch responses. osm‐9 mutant animals fail to respond to nose touch stimuli while their response to gentle body touch, mediated by the six touch receptor neurons, is normal (Colbert et al., 1997). mec‐4 and mec‐10 are not required to sense nose touch and similarly osm‐9 is not required to sense body touch. These finding indicate that OSM‐9 functions as mechanosensory channel in ciliated nose sensory neurons. Furthermore, osm‐9 mutants are also defective in olfactory responses mediated by the AWA and AWC neurons, and in osmotic avoidance responses mediated by the ASH neuron. The OSM‐9 protein localizes to the sensory cilia of AWA and ASH, suggesting a direct role in sensory transduction (Colbert et al., 1997). Four additional osm‐9/capsaicin receptor‐related TRPV genes are coexpressed with osm‐9 in specific subsets of cells (ocr‐1, ocr‐2, ocr‐3, and ocr‐4). These TRPV genes encode proteins which are 20‐25% identical to OSM‐9 and, similarly to OSM‐9, contain six MSDs and three ankyrin repeats (Tobin et al., 2002). ocr‐1 is expressed in AWA and ADL chemosensory neurons, ocr‐2 is expressed in AWA, ADL, ASH, ADF, PHA, and PHB sensory neurons, ocr‐3 is expressed in the rectal gland cells and weakly in the glial socket cells, and finally ocr‐4 is expressed exclusively in the
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
71
mechanosensory QLQ neurons (Kahn‐Kirby and Bargmann, 2006). All of these neurons, as well as the rectal gland cells coexpress osm‐9. On the basis of the expression pattern of the ocr genes, OCR‐2 and OCR‐4 appear to be the strongest candidates for the formation of a TRPV mechanosensitive complex with OSM‐9. Consistent with this notion, the nociceptive functions of the ASH neurons, including nose touch sensation, are severely compromised in ocr‐2 mutants (Tobin et al., 2002). Both OSM‐9 and OCR‐2 are localized in the cilia of AWA and ASH cells and this localization is interdependent. In neurons that express osm‐9, the absence of OCR‐2 results in the translocation of OSM‐9 from cilia to the cell body. In addition, ectopic expression of OCR‐2 in the AWC drives OSM‐9 to the cilia. These findings suggest a physical interaction between OSM‐9 and OCR‐2 that is required for normal nose touch sensation (Tobin et al., 2002). Interestingly, in Drosophila the TRPV proteins NAN (Nanchung) and IAV (Inactive) interact to form a Ca2þ‐permeable channel which senses mechanical vibrations and is required for auditory transduction (Kim et al., 2003; Gong et al., 2004). This is also indicative of the conserved function of TRPV proteins to mediate mechanosensitive behaviors. OSM‐9 and OCR‐2 also regulate the social feeding behavior in C. elegans (de Bono et al., 2002). This behavior is characterized by a rapid movement toward the food source and the aggregation of animals during feeding (de Bono and Bargmann, 1998). Mutations in osm‐9 and ocr‐2 suppress this accumulation of animals in C. elegans strains, which are native social feeders. Several genetic studies suggest that the function of the putative OSM‐9/ OCR‐2 ion channel is regulated by G‐protein signaling and specific polyunsaturated fatty acids (PUFAs), which act upstream of OSM‐9/OCR‐2 to modulate nocipteptive responses in ASH neurons, including the mechanosensory nose touch avoidance behavior (Roayaie et al., 1998; Kahn‐Kirby et al., 2004). Rat TRPV4 expressed in the ASH neurons of nematode osm‐9 mutants rescues osmosensation and mechanosensation defects in these animals. However, this is not the case in ocr‐2 mutants (Liedtke et al., 2003). Another mammalian TRPV homologue, the TRPV1 capsaicin receptor, is also capable of restoring the impaired avoidance behaviors of osm‐9 and ocr‐2 mutants (Tobin et al., 2002). These results suggest that TRPV functions are at least partially conserved in metazoans.
B. TRP‐4 The C. elegans TRP‐4 is a member of the TRPN subfamily of ion channels (Li et al., 2006). This group also includes the zebrafish TRPN1 and the Drosophila NompC. TRPN1 is localized in the sensory hair cells of the inner
72
Bazopoulou and Tavernarakis
ear and is required for the response to vibrations and normal hearing (Sidi et al., 2003). NompC is a mechanosensory ion channel required for sensing bristle displacements (Walker et al., 2000). Similarly, TRP‐4 appears to be involved in mechanosensory signaling in C. elegans. The trp‐4 gene is expressed in three sets of dopaminergic neurons (CEP, ADE, and PDE; Walker et al., 2000, p. 104), and in two interneurons (DVA and DVC; Li et al., 2006). Dopaminergic neurons in C. elegans mediate the basal slowing response, which is a tactile mechanosensory behavior (Sawin et al., 2000). Essentially, wild‐type animals slow down when they encounter a bacterial lawn by sensing a mechanical attribute pertinent to the texture of the culture substrate and the bacterial lawn. This response is not specific to bacteria since animals respond similarly to sterile, artificial lawns made of sepharose beads (Sawin et al., 2000). Slowing originates from the decreased frequency of body bending and increases the amount of time animals spend in areas rich in food. trp‐4 mutant worms show fast and exaggerated body bending which is not modulated by the texture of the substrate. The frequency of body bending is regulated by dopaminergic neurons, while bending extend appears to be influenced by the DVA and DVC interneurons (Li et al., 2006). It is likely that TRP‐4 functions in these neurons as a sensor of body bending, which provides the feedback necessary to sustain sinusoidal locomotion. Indeed, measurements Ca2þ currents evoked by body bending suggest that the DVA interneuron is stretch sensitive and that the TRP‐4 ion channel mediates stretch sensitivity in this neuron to facilitate proprioception (Li et al., 2006).
VI. CONCLUDING REMARKS Genetic analyses have been highly successful in identifying genes needed for mechanosensitive behaviors (Chalfie, 1997; Eberl et al., 1997; Nicolson et al., 1998; Gillespie and Walker, 2001; Hamill and Martinac, 2001). However, there are several limitations associated with genetic approaches aiming to dissect mechanotransduction mechanisms. Genes that encode products needed for the activities of mechanotransducing complexes in multiple cell types or that perform multiple cellular functions might have evaded genetic detection because mutations in such genes would be expected to be severely uncoordinated or even lethal. Indeed, many mutations that aVect mechanosensation in Drosophila render animals severely uncoordinated and nearly inviable (Kernan et al., 1994; Eberl et al., 1997). Moreover, genes whose functions are redundantly encoded cannot be readily identified in genetic screens. Thus, additional cellular proteins essential for
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
73
the mechanotransducing complex in the well‐studied C. elegans body touch receptor neurons may still remain to be discovered. The detailed model for mechanotransduction in C. elegans touch receptor neurons accommodates genetic data and molecular properties of cloned genes. This model remains to be tested by determining subcellular channel localization, subunit associations and, most importantly, channel‐gating properties. The proposed direct interactions between proteins that build the mechanotransducing complex have begun to be addressed experimentally (Chelur et al., 2002; Goodman et al., 2002; O’Hagan et al., 2005). Despite the undeniably considerable progress that has been achieved during recent years in all fronts toward dissecting the process of sensory mechanotransduction at the molecular level, several thorny questions are still begging for answers. What is the gating mechanism of mechanosensitive ion channels? How is tension delivered to the mechanotransducing complex? What additional molecules play part in the biological response to mechanical stimuli? Are human sensory mechanotransducers similar in composition and function to nematode or Drosophila ones? It is important to emphasize that although specialized ion channels most likely comprise the core of every metazoan mechanotransducer, it is the other physically associated proteins that shape its properties. It is equally important to seek and identify these. Without them, our understanding of mechanical transduction will never be complete. Mechanical sensation at the molecular level in higher organisms is most likely a property of a complex structure involving many components and contacts and not of any single protein. Several tools could be employed toward this goal, such as yeast two hybrid screens and biochemical methods of copurification of channel complexes, together with anchoring proteins. Electrophysiological studies of sensory mechanotransduction in C. elegans became possible, allowing direct recordings from nematode touch receptor neurons (O’Hagan et al., 2005). In a complementary approach, noninvasive monitoring and measurement technologies have been developed that allow the functional characterization of degenerin or other ion channels, while they are kept embedded in their natural surroundings (Bouevitch et al., 1993; Khatchatouriants et al., 2000; Suzuki et al., 2003). Direct, nondestructive recordings from touch receptor neurons coupled with the powerful genetics of C. elegans will hopefully allow the complete dissection of a metazoan mechanotransducing complex. References Avery, L. (1993). The genetics of feeding in Caenorhabditis elegans. Genetics 133, 897–917. Bargmann, C. I., and Kaplan, J. M. (1998). Signal transduction in the Caenorhabditis elegans nervous system. Annu. Rev. Neurosci. 21, 279–308.
74
Bazopoulou and Tavernarakis
Barnes, T., Jin, Y., Horvitz, H., Ruvkun, G., and Hekimi, S. (1996). The Caenorhabditis elegans behavioral gene unc‐24 encodes a novel bipartite protein similar to both erythrocyte band 7.2 (stomatin) and nonspecific lipid transfer protein. J. Neurochem. 67, 46–57. Barr, M. M., and Sternberg, P. W. (1999). A polycystic kidney‐disease gene homologue required for male mating behaviour in C. elegans. Nature 401, 386–389. Barr, M. M., DeModena, J., Braun, D., Nguyen, C. Q., Hall, D. H., and Sternberg, P. W. (2001). The Caenorhabditis elegans autosomal dominant polycystic kidney disease gene homologs lov‐1 and pkd‐2 act in the same pathway. Curr. Biol. 11, 1341–1346. Bianchi, L., Gerstbrein, B., Frokjaer‐Jensen, C., Royal, D. C., Mukherjee, G., Royal, M. A., Xue, J., Schafer, W. R., and Driscoll, M. (2004). The neurotoxic MEC‐4(d) DEG/ENaC sodium channel conducts calcium: Implications for necrosis initiation. Nat. Neurosci. 7, 1337–1344. Bonilla, M., and Cunningham, K. W. (2002). Calcium release and influx in yeast: TRPC and VGCC rule another kingdom. Sci. STKE 2002, PE17. Bouevitch, O., Lewis, A., Pinevsky, I., Wuskell, J. P., and Loew, L. M. (1993). Probing membrane potential with nonlinear optics. Biophys. J. 65, 672–679. Brenner, S. (1974). The genetics of Caenorhabditis elegans. Genetics 77, 71–94. Chalfie, M. (1993). Touch receptor development and function in Caenorhabditis elegans. J. Neurobiol. 24, 1433–1441. Chalfie, M. (1995). The diVerentiation and function of the touch receptor neurons of Caenorhabditis elegans. Prog. Brain Res. 105, 179–182. Chalfie, M. (1997). A molecular model for mechanosensation in Caenorhabditis elegans. Biol. Bull. 192, 125. Chalfie, M., and Au, M. (1989). Genetic control of diVerentiation of the Caenorhabditis elegans touch receptor neurons. Science 243, 1027–1033. Chalfie, M., and Sulston, J. (1981). Developmental genetics of the mechanosensory neurons of Caenorhabditis elegans. Dev. Biol. 82, 358–370. Chalfie, M., and Wolinsky, E. (1990). The identification and suppression of inherited neurodegeneration in Caenorhabditis elegans. Nature 345, 410–416. Chalfie, M., Sulston, J. E., White, J. G., Southgate, E., Thomson, J. N., and Brenner, S. (1985). The neural circuit for touch sensitivity in Caenorhabditis elegans. J. Neurosci. 5, 956–964. Chelur, D., Ernstrom, G., Goodman, M., Yao, C., Chen, L., O’Hagan, R., and Chalfie, M. (2002). The mechanosensory protein MEC‐6 is a subunit of the C. elegans touch‐cell degenerin channel. Nature 420, 669–673. Chiba, C. M., and Rankin, C. H. (1990). A developmental analysis of spontaneous and reflexive reversals in the nematode Caenorhabditis elegans. J. Neurobiol. 21, 543–554. Christensen, M., Estevez, A., Yin, X., Fox, R., Morrison, R., McDonnell, M., Gleason, C., Miller, D. M., III, and Strange, K. (2002). A primary culture system for functional analysis of C. elegans neurons and muscle cells. Neuron 33, 503–514. Church, D. L., and Lambie, E. J. (2003). The promotion of gonadal cell divisions by the Caenorhabditis elegans TRPM cation channel GON‐2 is antagonized by GEM‐4 copine. Genetics 165, 563–574. Colbert, H. A., Smith, T. L., and Bargmann, C. I. (1997). OSM‐9, a novel protein with structural similarity to channels, is required for olfaction, mechanosensation, and olfactory adaptation in Caenorhabditis elegans. J. Neurosci. 17, 8259–8269. Corey, D. P. (2003). New TRP channels in hearing and mechanosensation. Neuron 39, 585–588. Corey, D. P., and Garcia‐Anoveros, J. (1996). Mechanosensation and the DEG/ENaC ion channels. Science 273, 323–324. de Bono, M., and Bargmann, C. I. (1998). Natural variation in a neuropeptide Y receptor homolog modifies social behavior and food response in C. elegans. Cell 94, 679–689.
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
75
de Bono, M., Tobin, D. M., Davis, M. W., Avery, L., and Bargmann, C. I. (2002). Social feeding in Caenorhabditis elegans is induced by neurons that detect aversive stimuli. Nature 419, 899–903. Denis, V., and Cyert, M. S. (2002). Internal Ca(2þ) release in yeast is triggered by hypertonic shock and mediated by a TRP channel homologue. J. Cell Biol. 156, 29–34. Driscoll, M., and Chalfie, M. (1991). The mec‐4 gene is a member of a family of Caenorhabditis elegans genes that can mutate to induce neuronal degeneration. Nature 349, 588–593. Du, H., and Chalfie, M. (2001). Genes regulating touch cell development in Caenorhabditis elegans. Genetics 158, 197–207. Du, H., Gu, G., William, C., and Chalfie, M. (1996). Extracellular proteins needed for C. elegans mechanosensation. Neuron 16, 183–194. Eberl, D. F., Duyk, G. M., and Perrimon, N. (1997). A genetic screen for mutations that disrupt an auditory response in Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 94, 14837–14842. Emtage, L., Gu, G., Hartwieg, E., and Chalfie, M. (2004). Extracellular proteins organize the mechanosensory channel complex in C. elegans touch receptor neurons. Neuron 44, 795–807. Ernstrom, G. G., and Chalfie, M. (2002). Genetics of sensory mechanotransduction. Annu. Rev. Genet. 36, 411–453. Fares, H., and Greenwald, I. (2001). Regulation of endocytosis by CUP‐5, the Caenorhabditis elegans mucolipin‐1 homolog. Nat. Genet. 28, 64–68. Francis, R., and Waterston, R. H. (1991). Muscle cell attachment in Caenorhabditis elegans. J. Cell Biol. 114, 465–479. Fukushige, T., Siddiqui, Z., Chou, M., Culotti, J., Gogonea, C., Siddiqui, S., and Hamelin, M. (1999). MEC‐12, an alpha‐tubulin required for touch sensitivity in C. elegans. J. Cell Sci. 112(Pt. 3), 395–403. Garcia‐Anoveros, J., Ma, C., and Chalfie, M. (1995). Regulation of Caenorhabditis elegans degenerin proteins by a putative extracellular domain. Curr. Biol. 5, 441–448. Garcia‐Anoveros, J., Garcia, J. A., Liu, J. D., and Corey, D. P. (1998). The nematode degenerin UNC‐105 forms ion channels that are activated by degeneration‐ or hypercontraction‐ causing mutations. Neuron 20, 1231–1241. Gillespie, P. G., and Walker, R. G. (2001). Molecular basis of mechanosensory transduction. Nature 413, 194–202. Gong, Z., Son, W., Chung, Y. D., Kim, J., Shin, D. W., McClung, C. A., Lee, Y., Lee, H. W., Chang, D. J., Kaang, B. K., Cho, H., Oh, U., et al. (2004). Two interdependent TRPV channel subunits, inactive and Nanchung, mediate hearing in Drosophila. J. Neurosci. 24, 9059–9066. Goodman, M., Ernstrom, G., Chelur, D., O’Hagan, R., Yao, C., and Chalfie, M. (2002). MEC‐2 regulates C. elegans DEG/ENaC channels needed for mechanosensation. Nature 415, 1039–1042. Gu, G., Caldwell, G., and Chalfie, M. (1996). Genetic interactions aVecting touch sensitivity in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 93, 6577–6582. Hamelin, M., Scott, I. M., Way, J. C., and Culotti, J. G. (1992). The mec‐7 beta‐tubulin gene of Caenorhabditis elegans is expressed primarily in the touch receptor Neurons. EMBO J. 11, 2885–2893. Hamill, O. P., and Martinac, B. (2001). Molecular basis of mechanotransduction in living cells. Physiol. Rev. 81, 685–740. Hamill, O. P., Lane, J. W., and McBride, D. W., Jr. (1992). Amiloride: A molecular probe for mechanosensitive channels. Trends Pharmacol. Sci. 13, 373–376. Hardie, R. C., and Minke, B. (1992). The trp gene is essential for a light‐activated Ca2þ channel in Drosophila photoreceptors. Neuron 8, 643–651.
76
Bazopoulou and Tavernarakis
Hart, A. C., Sims, S., and Kaplan, J. M. (1995). Synaptic code for sensory modalities revealed by C. elegans GLR‐1 glutamate receptor. Nature 378, 82–85. Hart, A. C., Kass, J., Shapiro, J. E., and Kaplan, J. M. (1999). Distinct signaling pathways mediate touch and osmosensory responses in a polymodal sensory Neuron. J. Neurosci. 19, 1952–1958. Hersh, B. M., Hartwieg, E., and Horvitz, H. R. (2002). The Caenorhabditis elegans mucolipin‐ like gene cup‐5 is essential for viability and regulates lysosomes in multiple cell types. Proc. Natl. Acad. Sci. USA 99, 4355–4360. Hresko, M. C., Williams, B. D., and Waterston, R. H. (1994). Assembly of body wall muscle and muscle cell attachment structures in Caenorhabditis elegans. J. Cell Biol. 124, 491–506. Huang, M., and Chalfie, M. (1994). Gene interactions aVecting mechanosensory transduction in Caenorhabditis elegans. Nature 367, 467–470. Huang, M., Gu, G., Ferguson, E., and Chalfie, M. (1995). A stomatin‐like protein necessary for mechanosensation in C. elegans. Nature 378, 292–295. Hummler, E., and Horisberger, J. D. (1999). Genetic disorders of membrane transport. V. The epithelial sodium channel and its implication in human diseases. Am. J. Physiol. 276, G567–G571. Kahn‐Kirby, A. H., and Bargmann, C. I. (2006). Trp channels in C. elegans. Annu. Rev. Physiol. 68, 719–736. Kahn‐Kirby, A. H., Dantzker, J. L., Apicella, A. J., Schafer, W. R., Browse, J., Bargmann, C. I., and Watts, J. L. (2004). Specific polyunsaturated fatty acids drive TRPV‐dependent sensory signaling in vivo. Cell 119, 889–900. Kaplan, J. M., and Horvitz, H. R. (1993). A dual mechanosensory and chemosensory neuron in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 90, 2227–2231. Kellenberger, S., and Schild, L. (2002). Epithelial sodium channel/degenerin family of ion channels: A variety of functions for a shared structure. Physiol. Rev. 82, 735–767. Kernan, M., Cowan, D., and Zuker, C. (1994). Genetic dissection of mechanosensory transduction: Mechanoreception‐defective mutations of Drosophila. Neuron 12, 1195–1206. Khatchatouriants, A., Lewis, A., Rothman, Z., Loew, L., and Treinin, M. (2000). GFP is a selective non‐linear optical sensor of electrophysiological processes in Caenorhabditis elegans. Biophys. J. 79, 2345–2352. Kim, J., Chung, Y. D., Park, D. Y., Choi, S., Shin, D. W., Soh, H., Lee, H. W., Son, W., Yim, J., Park, C. S., Kernan, M. J., Kim, C., et al. (2003). A TRPV family ion channel required for hearing in Drosophila. Nature 424, 81–84. Lai, C. C., Hong, K., Kinnell, M., Chalfie, M., and Driscoll, M. (1996). Sequence and transmembrane topology of MEC‐4, an ion channel subunit required for mechanotransduction in Caenorhabditis elegans. J. Cell Biol. 133, 1071–1081. Li, W., Feng, Z., Sternberg, P. W., and Xu, X. Z. (2006). A C. elegans stretch receptor Neuron revealed by a mechanosensitive TRP channel homologue. Nature 440, 684–687. Liedtke, W., Tobin, D. M., Bargmann, C. I., and Friedman, J. M. (2003). Mammalian TRPV4 (VR‐OAC) directs behavioral responses to osmotic and mechanical stimuli in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 100(Suppl. 2), 14531–14536. Liu, J., Schrank, B., and Waterston, R. H. (1996). Interaction between a putative mechanosensory membrane channel and a collagen. Science 273, 361–364. Liu, K. S., and Sternberg, P. W. (1995). Sensory regulation of male mating behavior in Caenorhabditis elegans. Neuron 14, 79–89. Montell, C. (2001). Physiology, phylogeny, and functions of the TRP superfamily of cation channels. Sci. STKE 2001, RE1. Montell, C. (2005). The TRP superfamily of cation channels. Sci. STKE 2005, re3.
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
77
Montell, C., and Rubin, G. M. (1989). Molecular characterization of the Drosophila trp locus: A putative integral membrane protein required for phototransduction. Neuron 2, 1313–1323. Nauli, S. M., Alenghat, F. J., Luo, Y., Williams, E., Vassilev, P., Li, X., Elia, A. E., Lu, W., Brown, E. M., Quinn, S. J., Ingber, D. E., Zhou, J., et al. (2003). Polycystins 1 and 2 mediate mechanosensation in the primary cilium of kidney cells. Nat. Genet. 33, 129–137. Nicolson, T., Rusch, A., Friedrich, R. W., Granato, M., Ruppersberg, J. P., and Nusslein‐ Volhard, C. (1998). Genetic analysis of vertebrate sensory hair cell mechanosensation: The zebrafish circler mutants. Neuron 20, 271–283. O’Hagan, R., Chalfie, M., and Goodman, M. B. (2005). The MEC‐4 DEG/ENaC channel of Caenorhabditis elegans touch receptor neurons transduces mechanical signals. Nat. Neurosci. 8, 43–50. Palmer, C. P., Zhou, X. L., Lin, J., Loukin, S. H., Kung, C., and Saimi, Y. (2001). A TRP homolog in Saccharomyces cerevisiae forms an intracellular Ca(2þ)‐permeable channel in the yeast vacuolar membrane. Proc. Natl. Acad. Sci. USA 98, 7801–7805. Park, E. C., and Horvitz, H. R. (1986a). C. elegans unc‐105 mutations aVect muscle and are suppressed by other mutations that aVect muscle. Genetics 113, 853–867. Park, E. C., and Horvitz, H. R. (1986b). Mutations with dominant eVects on the behavior and morphology of the nematode Caenorhabditis elegans. Genetics 113, 821–852. Pickles, J. O. (1993). A model for the mechanics of the stereociliar bundle on acousticolateral hair cells. Hear. Res. 68, 159–172. Pickles, J. O., and Corey, D. P. (1992). Mechanoelectrical transduction by hair cells. Trends Neurosci. 15, 254–259. Qian, F., and Noben‐Trauth, K. (2005). Cellular and molecular function of mucolipins (TRPML) and polycystin 2 (TRPP2). Pflugers Arch. 451, 277–285. Qin, H., Rosenbaum, J. L., and Barr, M. M. (2001). An autosomal recessive polycystic kidney disease gene homolog is involved in intraflagellar transport in C. elegans ciliated sensory neurons. Curr. Biol. 11, 457–461. Rajaram, S., Sedensky, M. M., and Morgan, P. G. (1998). Unc‐1: A stomatin homologue controls sensitivity to volatile anesthetics in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 95, 8761–8766. Richmond, J. E., and Jorgensen, E. M. (1999). One GABA and two acetylcholine receptors function at the C. elegans neuromuscular junction. Nat. Neurosci. 2, 791–797. Roayaie, K., Crump, J. G., Sagasti, A., and Bargmann, C. I. (1998). The G alpha protein ODR‐3 mediates olfactory and nociceptive function and controls cilium morphogenesis in C. elegans olfactory neurons. Neuron 20, 55–67. Royal, D. C., Bianchi, L., Royal, M. A., Lizzio, M., Jr., Mukherjee, G., Nunez, Y. O., and Driscoll, M. (2005). Temperature‐sensitive mutant of the Caenorhabditis elegans neurotoxic MEC‐4(d) DEG/ENaC channel identifies a site required for traYcking or surface maintenance. J. Biol. Chem. 280, 41976–41986. Savage, C., Hamelin, M., Culotti, J., Coulson, A., Albertson, D., and Chalfie, M. (1989). mec‐7 is a beta‐tubulin gene required for the production of 15‐protofilament microtubules in Caenorhabditis elegans. Genes Dev. 3, 870–881. Savage, C., Xue, Y., Mitani, S., Hall, D., Zakhary, R., and Chalfie, M. (1994). Mutations in the Caenorhabditis elegans beta‐tubulin gene mec‐7: EVects on microtubule assembly and stability and on tubulin autoregulation. J. Cell Sci. 107(Pt. 8), 2165–2175. Sawin, E. R., Ranganathan, R., and Horvitz, H. R. (2000). C. elegans locomotory rate is modulated by the environment through a dopaminergic pathway and by experience through a serotonergic pathway. Neuron 26, 619–631.
78
Bazopoulou and Tavernarakis
Sedensky, M. M., Siefker, J. M., and Morgan, P. G. (2001). Model organisms: New insights into ion channel and transporter function. Stomatin homologues interact in Caenorhabditis elegans. Am. J. Physiol. Cell Physiol. 280, C1340–C1348. Sedensky, M. M., Siefker, J. M., Koh, J. Y., Miller, D. M., III, and Morgan, P. G. (2004). A stomatin and a degenerin interact in lipid rafts of the nervous system of Caenorhabditis elegans. Am. J. Physiol. Cell Physiol. 287, C468–C474. ShreZer, W., and Wolinsky, E. (1997). Genes controlling ion permeability in both motorneurons and muscle. Behav. Genet. 27, 211–221. ShreZer, W., Magardino, T., Shekdar, K., and Wolinsky, E. (1995). The unc‐8 and sup‐40 genes regulate ion channel function in Caenorhabditis elegans motor neurons. Genetics 139, 1261–1272. Sidi, S., Friedrich, R. W., and Nicolson, T. (2003). NompC TRP channel required for vertebrate sensory hair cell mechanotransduction. Science 301, 96–99. Snyers, L., Umlauf, E., and Prohaska, R. (1999). Association of stomatin with lipid‐protein complexes in the plasma membrane and the endocytic compartment. Eur. J. Cell Biol. 78, 802–812. Stewart, G. W. (1997). Stomatin. Int. J. Biochem. Cell Biol. 29, 271–274. Sulston, J. E., and Horvitz, H. R. (1977). Post‐embryonic cell lineages of the nematode, Caenorhabditis elegans. Dev. Biol. 56, 110–156. Sulston, J. E., Schierenberg, E., White, J. G., and Thomson, J. N. (1983). The embryonic cell lineage of the nematode Caenorhabditis elegans. Dev. Biol. 100, 64–119. Sun, A. Y., and Lambie, E. J. (1997). gon‐2, a gene required for gonadogenesis in Caenorhabditis elegans. Genetics 147, 1077–1089. Suzuki, H., Kerr, R., Bianchi, L., Frokjaer‐Jensen, C., Slone, D., Xue, J., Gerstbrein, B., Driscoll, M., and Schafer, W. R. (2003). In vivo imaging of C. elegans mechanosensory neurons demonstrates a specific role for the MEC‐4 channel in the process of gentle touch sensation. Neuron 39, 1005–1017. Syntichaki, P., and Tavernarakis, N. (2002). Death by necrosis. Uncontrollable catastrophe, or is there order behind the chaos? EMBO Rep. 3, 604–609. Syntichaki, P., and Tavernarakis, N. (2003). The biochemistry of neuronal necrosis: Rogue biology? Nat. Rev. Neurosci. 4, 672–684. Syntichaki, P., and Tavernarakis, N. (2004). Genetic models of mechanotransduction: The nematode Caenorhabditis elegans. Physiol. Rev. 84, 1097–1153. Tavernarakis, N., and Driscoll, M. (1997). Molecular modeling of mechanotransduction in the nematode Caenorhabditis elegans. Annu. Rev. Physiol. 59, 659–689. Tavernarakis, N., and Driscoll, M. (2000). Caenorhabditis elegans degenerins and vertebrate ENaC ion channels contain an extracellular domain related to venom neurotoxins. J. Neurogenet. 13, 257–264. Tavernarakis, N., and Driscoll, M. (2001). Degenerins. At the core of the metazoan mechanotransducer? Ann. NY Acad. Sci. 940, 28–41. Tavernarakis, N., ShreZer, W., Wang, S., and Driscoll, M. (1997). unc‐8, a DEG/ENaC family member, encodes a subunit of a candidate mechanically gated channel that modulates C. elegans locomotion. Neuron 18, 107–119. Tavernarakis, N., Driscoll, M., and Kyrpides, N. C. (1999). The SPFH domain: Implicated in regulating targeted protein turnover in stomatins and other membrane‐associated proteins. Trends Biochem. Sci. 24, 425–427. Tavernarakis, N., Everett, J. K., Kyrpides, N. C., and Driscoll, M. (2001). Structural and functional features of the intracellular amino terminus of DEG/ENaC ion channels. Curr. Biol. 11, R205–R208.
3. Mechanosensitive Ion Channels in Caenorhabditis elegans
79
Teramoto, T., Lambie, E. J., and Iwasaki, K. (2005). DiVerential regulation of TRPM channels governs electrolyte homeostasis in the C. elegans intestine. Cell Metab. 1, 343–354. The C. elegans Sequencing Consortium (1998). Genome sequence of the nematode C. elegans: A platform for investigating biology. Science 282, 2012–2018. Thomas, J. H. (1990). Genetic analysis of defecation in Caenorhabditis elegans. Genetics 124, 855–872. Tobin, D., Madsen, D., Kahn‐Kirby, A., Peckol, E., Moulder, G., Barstead, R., Maricq, A., and Bargmann, C. (2002). Combinatorial expression of TRPV channel proteins defines their sensory functions and subcellular localization in C. elegans neurons. Neuron 35, 307–318. Treusch, S., Knuth, S., Slaugenhaupt, S. A., Goldin, E., Grant, B. D., and Fares, H. (2004). Caenorhabditis elegans functional orthologue of human protein h‐mucolipin‐1 is required for lysosome biogenesis. Proc. Natl. Acad. Sci. USA 101, 4483–4488. Venkatachalam, K., Hofmann, T., and Montell, C. (2006). Lysosomal localization of TRPML3 depends on TRPML2 and the mucolipidosis‐associated protein TRPML1. J. Biol. Chem. 281, 17517–17527. Waldmann, R., and Lazdunski, M. (1998). H(þ)‐gated cation channels: Neuronal acid sensors in the NaC/DEG family of ion channels. Curr. Opin. Neurobiol. 8, 418–424. Walker, R. G., Willingham, A. T., and Zuker, C. S. (2000). A Drosophila mechanosensory transduction channel. Science 287, 2229–2234. Way, J. C., and Chalfie, M. (1989). The mec‐3 gene of Caenorhabditis elegans requires its own product for maintained expression and is expressed in three neuronal cell types. Genes Dev. 3, 1823–1833. West, R. J., Sun, A. Y., Church, D. L., and Lambie, E. J. (2001). The C. elegans gon‐2 gene encodes a putative TRP cation channel protein required for mitotic cell cycle progression. Gene 266, 103–110. White, J. G., Southgate, E., Thomson, J. N., and Brenner, S. (1976). The structure of the ventral nerve cord of Caenorhabditis elegans. Philos. Trans. R. Soc. Lond. B Biol. Sci. 275, 327–348. White, J. G., Southgate, E., Thomson, J. N., and Brenner, S. (1986). The structure of the nervous system of Caenorhabditis elegans. Philos. Trans. R. Soc. Lond. B Biol. Sci. 314, 1–340. Wicks, S. R., and Rankin, C. H. (1995). Integration of mechanosensory stimuli in Caenorhabditis elegans. J. Neurosci. 15, 2434–2444. Wicks, S. R., and Rankin, C. H. (1997). EVects of tap withdrawal response habituation on other withdrawal behaviors: The localization of habituation in the nematode Caenorhabditis elegans. Behav. Neurosci. 111, 342–353. Wolinsky, E., and Way, J. (1990). The behavioral genetics of Caenorhabditis elegans. Behav. Genet. 20, 169–189. Xu, X. Z., and Sternberg, P. W. (2003). A C. elegans sperm TRP protein required for sperm‐egg interactions during fertilization. Cell 114, 285–297. Zhang, S., Arnadottir, J., Keller, C., Caldwell, G. A., Yao, C. A., and Chalfie, M. (2004a). MEC‐2 is recruited to the putative mechanosensory complex in C. elegans touch receptor neurons through its stomatin‐like domain. Curr. Biol. 14, 1888–1896. Zhang, S., Sokolchik, I., Blanco, G., and Sze, J. Y. (2004b). Caenorhabditis elegans TRPV ion channel regulates 5HT biosynthesis in chemosensory neurons. Development 131, 1629–1638. Zhang, Y., Ma, C., Delohery, T., Nasipak, B., Foat, B. C., Bounoutas, A., Bussemaker, H. J., Kim, S. K., and Chalfie, M. (2002). Identification of genes expressed in C. elegans touch receptor neurons. Nature 418, 331–335.
This page intentionally left blank
CHAPTER 4 Properties and Mechanism of the Mechanosensitive Ion Channel Inhibitor GsMTx4, a Therapeutic Peptide Derived from Tarantula Venom Philip A. Gottlieb, Thomas M. Suchyna, and Frederick Sachs The Department of Physiology and Biophysics, Center for Single Molecule Biophysics, SUNY at BuValo, BuValo, New York 14214
I. Overview II. Introduction III. Properties and Specificity of GsMTx4 A. Biochemical and Structural B. Biophysical and Mechanistic C. Specificity IV. Cellular Sites for GsMTx4 A. TRPC1 Channel B. TRPC6 Channel V. Potential Therapeutic Uses for GsMTx4 A. Cardiac Myocytes and Atrial Fibrillation B. Muscular Dystrophy C. Astrocytes and Gliosis D. Neurite Growth Extension VI. Conclusions References
I. OVERVIEW Mechanosensitive ion channels (MSCs) are found in all types of cells ranging from Escherichia coli to morning glories to humans. They seem to fall into two families: those in specialized receptors such as the hair cells of the cochlea and Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59004-0
82
Gottlieb et al.
those in cells not clearly diVerentiated for sensory duty. The physiological function of the channels in nonspecialized cells has not been demonstrated, although their activity has been demonstrated innumerable times in vitro. The only specific reagent to block MSCs is GsMTx4, a 4‐kDa peptide isolated from tarantula venom. Despite being isolated from venom, it is nontoxic to mice. GsMTx4 is specific for an MSC subtype, the nonselective cation channels that may be members of the TRP family. GsMTx4 acts as a gating modifier, increasing the energy of the open state relative to the closed state. Surprisingly, the mirror image D enantiomer of GsMTx4 is equally active, so mode of action is not via the traditional lock and key model. GsMTx4 probably acts in the boundary lipid of the channel by changing local curvature and mechanically stressing the channel toward the closed state. Despite the lack of definitive physiological data on the function of the cationic MSCs, GsMTx4 may prove useful as a drug or lead compound that can aVect physiological processes. These processes would be those driven by mechanical stress such as blood vessel autoregulation, stress‐induced contraction of smooth muscle, and Ca2þ loading in muscular dystrophy.
II. INTRODUCTION Mechanical sensitivity plays an essential role in cells and higher organisms. Specialized exteroceptors transduce external stimuli such as sound, vibration, touch, and local gravity. Interoceptors regulate for the voluntary musculature and the filling of the hollow organs, as in regulation of blood pressure. MSCs may serve as sensors for local control of blood flow, regulation of cell volume, deposition of bone, and so on (Sachs and Morris, 1998; Hamill and Martinac, 2001). The channels may also drive some of the hormonally coupled mechanical systems, such as renin‐angiotensin and atrial natriuretic peptide that regulate fluid volume. They may also serve some of the autocrine and paracrine transducers that generate second messengers such as endothelin (ET) (Ostrow et al., 2000; Ostrow and Sachs, 2005). Mechanical transduction is ubiquitous and is present in cells of all phyla. In higher plants, mechanical transducers guide root, stem, and leaf growth in response to gravity. MSCs serve as sensory transducers in bacteria and other microorganisms where they may be the sensors for volume regulation (Martinac, 2001; Sachs, 2002). The fact that E. coli has as many as five different MSCs argues for their functional importance (Sachs, 2002). Mechanical transduction is presumed to have developed early in evolution, probably as a necessity for controlling cell volume when conducting metabolism in a membrane‐limited compartment.
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
83
The pervasive nature of MSCs indicates that we will find genetic and environmental factors that create human pathologies related to MSC malfunction. For example, studies on dystrophic muscle cells show that the dystrophin mutations lead to weakening of the membrane, thereby activating a Ca2þ influx through MSCs (Patel et al., 2001; Yeung et al., 2003). This influx can be blocked by gadolinium (Yeung et al., 2003) and the peptide GsMTx4 (Yeung et al., 2005). Although mechanical sensitivity of ion channels appears across phyla (Martinac and Kloda, 2003), there appears to be no homology associated with the primary structure. For example, in E. coli, the two dominant mechanosensitive channels MscL and MscS (generically noted MSCs) diVer fundamentally in sequence and structure. MscL is a pentamer (Chang et al., 1998) and MscS is a heptamer (Bass et al., 2003), and the primary sequences have little homology. The only well‐characterized MSC cloned from eukaryotes is the Kþ selective 2P channels such as TREK‐1 (Patel et al., 2001), and these channels have no sequence homology to the bacterial channels. Thus, mechanosensitivity, while universal, does not obey the delightful homologies of many of the voltage‐ and ligand‐gated channels, an example of convergent evolution. Moreover, from a mechanistic viewpoint, bacterial MSCs are almost certainly diVerent from eukaryotic channels given the diVerence in cytoskeletal structure that influences the mechanics. What we learn from bacteria does not necessarily apply to eukaryotic MSCs. Within the phenotypic MSC families, however, there appears to be a useful discriminator— channels that are stimulated by stress in the cytoskeleton and extracellular matrix (Corey, 2003a,b), as in the cochlea, and those that are stimulated by stress in the bilayer, as in bacterial MSCs. The intrinsic mechanosensitivity of channels depends on dimensional changes between the closed and open states (Sachs and Morris, 1998; Sukharev et al., 1999; Hamill and Martinac, 2001). One detailed kinetic study of MscL shows that these prototype channels require at least eight rate constants to characterize the gating reaction, but only a single rate constant is significantly sensitive to tension (Sukharev et al., 1999). While most MSCs appear to be stretch‐activated channels (SACs), stretch‐inactivated channel (SIC) activity has also been described (Vandorpe et al., 1994), although this may be an artifactual response from SACs subjected to stress at rest (Honore et al., 2006). Only recently have cationic MSCs from nonspecialized tissues, TRPC1, been cloned or reconstituted (Maroto et al., 2005). Mechanosensitivity is not the domain of a particular class of ion channels. Any channel that changes dimensions between closed and open states may be mechanosensitive, in the same way that most ion channels are voltage sensitive. Ligand‐gated and voltage‐sensitive channels have been shown to be mechanically sensitive (Gu et al., 2001; Calabrese et al., 2002; Laitko and Morris, 2004;
84
Gottlieb et al.
Morris, 2004; see Chapter 11). The generality of mechanosensitivity poses an intriguing problem in evolution: how to design structures with the necessary flexibility to support large conformational changes (Jiang et al., 2003a,b) while avoiding unnecessary mechanical activation. MSCs are phenotypically described as channels whose kinetics are substantially altered by mechanical input. The key parameter that makes channels mechanosensitive is that they have large dimensional changes between the closed and open conformations (Howard and Hudspeth, 1988; Sachs et al., 1998; Sukharev et al., 1999; Hudspeth et al., 2000; Hamill et al., 2001; see chapter by Markin and Sachs in this series, vol. 58, pp. 87–119). MSCs are embedded in a heterogeneous, non‐Newtonian mechanical structure consisting of the extracellular matrix, the bilayer and its embedded proteins, and the cytoskeleton (Garcia‐Anoveros and Corey, 1996; Gillespie and Walker, 2001). The stress that activates MSCs may come from the lipid bilayer (Akinlaja and Sachs, 1998), but that tension depends on the cytoskeleton, the preparation geometry, and the boundary conditions (Suchyna and Sachs, 2004). Despite this complexity, it appears that MSCs from nonspecialized tissues are activated by tension in the lipid bilayer (Sukharev et al., 1994; Suchyna et al., 2004). The tension depends on the cortical structure, since the applied stresses are borne by cytoskeletal elements in parallel and in series with MSCs (Wan et al., 1995; Mills and Morris, 1998). This is also true not only for patch clamp experiments but also for global stimuli such as hypotonic or shear stress. To define an absolute sensitivity of a channel requires working in lipid bilayers where the stress is reasonably well defined (Sukharev et al., 1999; Suchyna et al., 2004). The physiological function of MSCs in nonspecialized tissues has not been demonstrated. One common ground (Sachs, 2002), however, may be volume regulation (Christensen, 1987), although preliminary data using GsMTx4 suggests that the volume sensor is not a cationic MSC (Hua, Gottlieb, and Sachs, in preparation). In general, to test the physiological role of a channel requires that one activate or inactivate the target by nonphysiological stimuli. Pharmacologic agents are one approach and genetic knockouts the other (Corey, 2003b). There is only one specific pharmacological agent for MSCs to date: GsMTx4 and its mutants (Suchyna et al., 2000). The search has been hampered, in part, by technical diYculties in defining the stimulus (Hamill and McBride, 1995; Besch et al., 2002). While stimulators for electrically gated and ligand‐gated channels have long been available (ALA Scientific Instruments Inc., Westbury, NY), until recently none were available for mechanically gated channels. However, even with controlled pressure stimuli for patch clamp experiments, defining the stimulus that actually reaches the channel requires knowledge of preparation geometry and constitutive mechanical properties of the cell cortex (Sachs and Morris, 1998; see chapter by
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
85
Markin and Sachs in this series, vol. 58, pp. 87–119), factors that are generally unknown. The focus of this chapter is the peptide GsMTx4 and how it aVects MSCs. The initial part will describe our eVort to characterize its chemical and structural properties. We then detail the biophysical properties as well as issue of specificity. These results have shown the peptide to work on MSCs in an unconventional manner. Finally, we survey some potential therapeutic uses that may emerge for this peptide and similar compounds which remain undiscovered.
III. PROPERTIES AND SPECIFICITY OF GsMTx4 A. Biochemical and Structural A number of years ago we branched into natural products biochemistry with a blind search of invertebrate venoms to find anything that aVects MSCs. Screening required an outside‐out patch preparation that retained mechanical sensitivity, but most cell types lose mechanical activity after excision. Consequently, we developed an assay using primary rat astrocytes that were reasonably stable (Suchyna et al., 2000). At the same time, we developed a pressure clamp to control cell stimulation (Besch et al., 2002). These eVorts led to the identification and isolated the peptide GsMTx4 from Grammostola rosea, the only peptide or other drug known to specifically aVect cationic MSCs. The properties of this peptide were analyzed in detail and are summarized below. Interestingly, no scorpion venoms and only one other spider venom had an eVect on MSCs. The correct sequence for the peptide GsMTx4 was deduced by isolating the GsMTx4 gene. A cDNA copy was made from RNA extract derived from the glands of G. spatulata and was sequenced (Ostrow et al., 2003). The protein exists in a pre‐proform (Fig. 1) where the first 21 amino acids (light gray) are a predicted signal sequence and are removed during protein translocation (bimas.dcrt.nih.gov/molbio/signal). The last two amino acids are glycine‐lysine (dark gray), a known site for amidation (Gomez et al., 1984). The arginine adjacent to the active peptide molecule (prosequence in black) is presumably the cleavage site (indicated by arrow) to release active GsMTx4 peptide (gray). Next, we chemically synthesized GsMTx4 peptide with a phenylalanine amide at the C‐terminal and determined the conditions for folding (Ostrow et al., 2003). Reduced peptide (10 4 M) is dissolved in 0.1 M Tris pH 7.8 containing glutathione (oxidized:reduced; 1:10 mM). The reaction is carried out at RT and completed within 24 h. Folding is easily achieved from the
86
Gottlieb et al.
ATG AAG ACA TCT GTG GTG T TC GTC AT T GCA GGC T TA GCT CTG CT T TCA GT T GTC M K T S V V F V I A G L A L L S V V TGT TAT GCT TCA GAA CTG AAG GAG CAA AGT TCC GTC AAT GAA GTG CT T TCT ACA C Y A S E L K E Q S S V N E V L S T AT T T T T CAT T T T GAA CAA CCT GAG GAA AGA GGC TGT T TG GAA T T T TGG TGG AAA I F H F E Q P E E R G C L E F W W K TGC AAC CCT AAC GAC GAC AAA TGC TGT CGT CCA AAA T TG AAA TGC AGT AAACTG= C N P N D D K C C R P K L K C S K L A TAA T TC AAG T TG TGT AAC T T T TCA T TC GGC AAG TT AA A F K L C N F S F G K Stop
FIGURE 1 cDNA of the gene encoding GsMTx4 with the open reading frame. The full length protein is processed. The first 21 amino acids are removed as a signal sequence. The protein is cleaved at an arginine (arrow) and the last two amino acids are removed during amidation (Ostrow et al., 2003).
misfolded peptide as well. The synthetic peptide is indistinguishable from wild‐type peptide in all physical‐chemical properties. MALDI‐MS reveals a mass of 4093.9 [MþHþ] while GsMTx4 from venom had a mass of 4094.0 [MþHþ]. Reverse phase liquid chromatography shows identical retention times, and co‐injection of the two compounds produced a single peak. Circular dichroism for both peptides is similar, having minima at around 192 and 202 nm. A comparison of the NMR spectra of the wild‐type peptide with that of the synthetic peptide demonstrated that the structures were in good agreement (Ostrow et al., 2003). Finally, the peptide produced the same physiological response on SACs when compared to the spider peptide. The primary sequence of GsMTx4 has six cysteines. The spacing of these residues is identical to a family of peptides called inhibitory cysteine knot (ICK). These peptides adopt stable three‐dimensional structures by forming three disulfide bonds (Pallaghy et al., 1994). The structure was confirmed by NMR spectroscopy (Oswald et al., 2002) and revealed that the peptide is amphipathic (Fig. 2). GsMTx4 is shown with its hydrophobic face in green and the charged residues in red (negative) and blue (positive) and illustrates the hydrophobic surface (at bottom) surrounded by mainly positive charges. This architectural design suggests that the peptide binds to membranes using its hydrophobic face to penetrate the lipid bilayer. While the net charge of GsMTx4 is þ5, the charge itself is not the essential component since polylysine has no eVect on MSCs. The distribution of charge close to the membrane, however, is probably essential for activity. GsMTx4 is homologous to other peptides derived from spider venoms. Figure 3 compares a number of peptides that have been recently isolated for
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
87
FIGURE 2 Solution structure of GsMTx4 determined by NMR spectroscopy. Disulfide bonds are shown in yellow, the hydrophobic residues are shown in green, the acidic residues in red, and the basic residues in blue. GsMTx4 has a predicted net charge of þ5 at neutral pH (Suchyna et al., 2004).
GsMTx4 GCLEFW W KCNPNDDK CCRPKLK CSKLFKL CNFSF- NH2 SgTx1
TCRYLFGG CKTTAD
CCKH LA CRSAGKY CAW DGTF
HnTx1
ECRYLFGG CKTTSD
CCKH LG CKFRDKY CAW DFTFS
VsTX1
ECGKFMWK CKNSND
CCKD LV CSSRW KW CVLASPF
FIGURE 3 Sequence comparison of four peptides derived from spider venoms. All peptides belong to the ICK structural family and they are all gating modifiers. Hydrophobic residues are indicated in green. Charged residues are shown with red for acidic residues and blue for basic. Cysteine residues are in yellow and boxed.
various spiders. While their targets are diVerent, as they inhibit various voltage‐gated channels, they nonetheless have features that are common. All the peptides listed belong to the ICK structural family. All of these peptides, including GsMTx4, are thought to be gating modifiers, and all of them have aromatic groups in hydrophobic regions at the C‐ and N‐termini, which in the three‐dimensional structure form a hydrophobic face, that enables them to interact with membranes.
88
Gottlieb et al.
The peptide’s ability to bind to membranes was suggested as a means for understanding how GsMTx4 inhibits MSCs (Suchyna et al., 2004). Simultaneously, the MacKinnon group demonstrated that the peptide VsTx1, a gating modifier for voltage‐dependent Kþ channels, also partitions into membranes to eVect inhibition of these channels (Lee and MacKinnon, 2004). Similar observations were made by the Swartz group for the SgTx1 peptide (Lee et al., 2004; Wang et al., 2004). The ability to partition into membranes appears to be an essential feature for these gating modifiers. The peptide’s ability to interact with membranes was measured using model systems of large unilamellar vesicles (LUVs). On the basis of previous work (White et al., 1998), the Ladhokin group developed a sensitive method for determining partitioning of peptides into LUVs using diVerential iodine quenching (Posokhov et al., 2006). They showed that the peptide partitions well into zwitterionic vesicles such as PC or PE (G ¼ 6 kcal/mol) and to anionic vesicles (75%PS, G ¼ 8 kcal/mol). Considering the charge of the peptide, the modest increase in binding in the presence of PS vs PE argues for a low eVective charge and a lack of additivity between hydrophobic and charged interactions (Posokhov, Gottlieb, and Ladhokin, submitted for publication). To answer the question of the peptide’s orientation, the accessibility of tryptophanyl groups was examined in the absence and presence of vesicles. In solution, both Csþ and I quenched the fluorescence showing that the tryptophan residues were not shielded and that electrostatic interactions were negligible. This was similar to other soluble peptides containing tryptophan residues. However, in the presence of vesicles to the peptide’s tryptophans were protected from quenching, consistent with the hydrophobic face being buried in the lipids. The depth of the peptide penetration into the lipid bilayer was also measured using brominated lipids. Bromine, that quenches tryptophan fluorescence, was anchored in the acyl chains at diVerent distances from the headgroup of the phospholipid, and the location of maximal quenching provided a measure ˚ from the center of of the depth of penetration‐GsMTx4 penetrates about 9 A the lipid bilayer (Posokhov, Gottlieb, and Ladhokin, submitted for publication).
B. Biophysical and Mechanistic The phenomenological dissociation rate constant was determined by averaging the current from several patches during washout, and curve fitting the recovery data (Fig. 4) (Suchyna et al., 2000). To measure the association rate constant, the channels were activated with a three‐second pressure step and GsMTx4 was rapidly applied after one second into the stimulus (<10‐ms
89
Average current (pA)
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels Average patch current GsMTx4 Applied
4 3 2
Average current
1
Pressure
Control
0 0
20
40
60
Time (s)
Pressure GsMTx4
GsMTx4
Average current (pA)
Recovery rate
t = 590 + 10 ms
4 3
Superimposed
Difference current
1 pA 2s
t = 4.7 + 1.7 s
2 1 0 40
48
56
64
Time (s) FIGURE 4 Left panel: dissociation rates of GsMTx4 from adult astrocyte SAC currents using steps of 500 ms at 2 s intervals and averaged (n ¼ 7). The current recovered after 5 s of washout. The data were fitted with a single exponential to obtain the dissociation rate. Right panel: association rates of GsMTx4 to astrocyte MSCs. Average SAC currents with GsMTx4 applied are subtracted from the control currents to generate a diVerence current. The diVerence current was fit with a single exponential to determine the association rate (Suchyna et al., 2000).
switching time). The exponential decay of this current was taken to be the association rate (Fig. 4). We calculated the equilibrium constant from KD ¼ kd/ka. The kinetic method of determining KD is important since it is hard to prevent rundown or run‐up of the patch over time, particularly with the need to produce saturation of the starting stimulus. The association rate was 3.3 105 M 1 s 1, the dissociation rate 0.2 s 1, providing KD ¼ 0.5 mM. This is a lower aYnity than many of the well‐known channel‐inhibiting peptides, but the association rate is similar to that for other peptides (Lewis and Garcia, 2003). We then tried to determine the mechanism of action. There are two basic types of channel inhibition: kinetic (gating) and pore blocking (Suchyna et al., 2004). Pore blockers will produce a reduction in current but retain the same time course. Gating modifiers will change the time course. Figure 5 shows the eVect of GsMTx4 on SACs. GsMTx4 shifts the activation curve to the
90
A
Gottlieb et al. mmHg 160-
500 ms
B
90-
5-mM GsMTx4
1 pA Control
Average current (pA)
2.0 Control GsMTx4
1.5 1.0 0.5 0.0
100
120 140 mmHg
160
FIGURE 5 GsMTx4 is a gating modifier since the block can overcome inhibition by increasing suction. (A) Ensemble average current of an outside‐out patch from rat astrocytes as a function of suction in the presence and absence of GsMTx4. (B) The peak average currents plotted as a function of suction showing a rightward shift for activation indicative of a gating modifier (Suchyna et al., 2004).
right—the hallmark of a gating modifier, since the inhibition is overcome by an increased stimulus. Note that seal breakage limits the permissible pressure used in this assay so that the channels could not be driven to saturation. We looked for possible interactions between GsMTx4 and channels. A clue to GsMTx4 mode of action was revealed when we measure the activity of a peptide with the same amino acid sequence as GsMTx4 but having only D amino acids (enGsMTx4). We chemically synthesized the linear peptide and were able to fold it using a standard folding protocol (Ostrow et al., 2003; Suchyna et al., 2004). The resulting peptide was identical in all its chemical properties to the wild‐type peptide as measured by reverse phase chromatography and mass spectrometry. The only distinction was evident from circular dichroism measurements where the D amino acid peptide’s spectrum was a mirror image of the wild‐type peptide. We tested this peptide and found it to be nearly identical to the wild‐type peptide activity. Figure 6 shows that the application of 5 mM enGsMTx4 was suYcient to achieve complete inhibition and this eVect was reversible by washout of the peptide. These unexpected results argue for a mechanism that does not rely on direct stereochemical interactions with the protein channel, since the enGsMTx4 is incapable of correct three‐dimensional peptide‐protein interactions that would be postulated for the wild‐type peptide. Given that the GsMTx4 peptide does not follow the standard lock and key model for interactions, it is worth asking how the peptide elicits an
91
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
500 ms/130 mmHg Control
GsMTx4
Washout
3 pA
en GsMTx4
Washout
FIGURE 6 The eVect of enGsMTx4 on MSCs from rat astrocytes. Outside‐out patches were first treated with GsMTx4, then washed out and the enantiomeric peptide was then added and shown to have the same eVect. The inhibition by enGsMTx4 was reversible (Suchyna et al., 2004).
inhibitory response for MSCs. There are two basic models, global and local. The first is that the peptide alters the general properties of the lipid bilayer in which the channel is embedded. This would be analogous to the eVect of Gd3þ on the lipid bilayer (Ermakov et al., 1998, 2001), which in turn alters channels activity in the membrane (Yeung et al., 2003). If this were the case, we might expect GsMTx4 to change membrane capacitance. Indeed, when Gd3þ was perfused onto rat astrocyte membrane, there was a substantial decrease in capacitance (Suchyna and Sachs, unpublished data). At a saturating dose concentration of GsMTx4, there was no eVect on capacitance. These observations were supported by work on red blood cells, where their very soft bending rigidity makes them sensitive to the presence of amphipaths (Sheetz and Singer, 1974; Sheetz et al., 1976). Iwasa (personal communication) showed that GsMTx4 at saturating doses did not alter the cells shape. This suggests that despite the partitioning of GsMTx4 into the membrane (above), the alteration in local shape or global number density is too small to significantly alter the general membrane structure. A second possibility is that the peptide resides at the interface between the protein and lipid membrane. To ascertain whether the peptide was in proximity of the channel, we took advantage of the peptide’s charge (þ5). When the peptide is next to the channel, within a Debye length of the conducting pore, cations passing through the channel will experience the peptide’s electric field and there should be a reduction in current as the local concentration of cations
92
Gottlieb et al.
is reduced by Poisson‐Boltzmann factors. Since the peptide acts on the extracellular side of the membrane, it will aVect inward current more than outward current. The results of these experiments are shown in Fig. 7 (Suchyna et al., 2004). The outward current is shown at the left and inward at the right with a ‘‘typical’’ single‐channel trace at the top and the average just below. The bottom graph is the amplitude histogram of the single currents that made up the average trace with the control data in black and the GsMTx4 data in red. The outward unitary current is unaVected by GsMTx4. However, the inward current is reduced by about 10% as predicted. Note, this experiment is only possible because GsMTx4 is a gating modifier not a pore blocker. These results established that the peptide is close to the pore, probably in the boundary lipid region. The question still remains as to how the peptide aVects mechanosensitivity of these ion channels. Recent work of many groups have begun to dissect the forces used for channel gating (Sukharev et al., 1997; Blount et al., 1999; Markin and Sukharev, 2000; Perozo et al., 2002; Sachs, 2002; Sukharev and Anishkin, 2004; Wiggins and Phillips, 2004). Clearly, the protein channel has to undergo a conformational change as result of tension in the bilayer since some MSCs have been reconstituted into artificial lipid
Outward current
Inward current
70 mmHg/500 ms
70 mmHg/500 ms
Single channel recording 2 pA Ensemble average 0.8
1.0 Control GsMTx4
0.6 0.4 0.2
0.4 0.2 0.0
0.0 −1
Control GsMTx4
0.6
Counts
Counts
0.8
0
1
pA
2
3
4
−5
−4
−3
−2
−1
0
1
pA
FIGURE 7 GsMTx4 aVects the inward but not outward currents of SACs from adult rat astrocytes. The pressure pulse is indicated at the top. Outward current is to the left and inward current is to the right. Single‐channel recordings are shown underneath the pressure pulse. Ensemble averages in the presence (red) and absence (black) of GsMTx4 are indicated. The single‐channel amplitude histograms are shown below. These data were collected from outside‐out patches from rat astrocytes (Suchyna et al., 2004).
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
93
bilayers (Betanzos et al., 2002). The simplest gating model is that the closed and open channels diVer in size so that in a tension field T, the two conformations diVer in energy by TA, where A is the diVerence of in‐plane area between close and open conformations. This change in dimensions need not be large; for the 2P channel TREK‐1, it has been estimated to correspond to a ˚ (Honore et al., 2006). We think GsMTx4 alters the change in radius of 1 A curvature of the boundary region of the channel. In a thermodynamic sense, GsMTx4 alters the line tension surrounding the channel or applies a torque that prestresses the channel in a manner that favors the closed state (see chapter by Markin and Sachs). The eVect of GsMTx4 on Gramicidin (gA) in lipid bilayers underscores the role of physical eVects on drug activity. The gA is a tubular peptide having the thickness of a lipid monolayer (1.5 nm). Two monomers, one from each leaflet of the bilayer, associate end to end forming a channel. The kinetics of channel creation and dissociation are aVected by the thickness of the bilayer and the bending stiVness of the lipids (Andersen et al., 1999) and these properties can be altered by amphipathic agents or by far field tension (Hwang et al., 2003). GsMTx4 aVects gA gating (Fig. 8). Figure 8A shows that the addition of GsMTx4 to either AgA(15) or gA (13), two Gramicidin analogues, independently increased the opening rate, indicating that GsMTx4 increases dimer association as though the membrane is thinner. Similar to our observations for eukaryotic channels, GsMTx4 also decreased the unitary conductance for both gA analogues (Fig. 8B). The fact that the amplitude histograms display a monomodal distribution suggests a uniform association of GsMTx4 with the channels, but the concentration dependence of the unitary current further suggests that the average population of GsMTx4 around the channels is time dependent, but averaged at the observation time scale, that is there is rapid exchange. The lifetime of the dimers (channels) increased 25‐fold at 400‐nM GsMTx4 (Fig. 8C). The qualitative eVect of GsMTx4 was independent of the length of the gA as well as the chirality of the sequence (Fig. 8D). These results can be explained by the eVect of GsMTx4 on local lipid curvature at the gA‐lipid interface, reducing the hydrophobic mismatch by making the membrane look thinner.
C. Specificity GsMTx4 is remarkably specific given its apparent achiral site(s) of action, although it may be useful to remember that phospholipids are chiral. GsMTx4 does not aVect voltage‐gated channels in rat astrocytes (Suchyna et al., 2000) or rabbit heart (Baumgarten, 2004; Sachs, 2004a). As shown in
94
Gottlieb et al. A
B Control AgA(15) gA−(13)
2 pA 10 s
Events per Bin
200 nM GsMTx4
AgA(15) gA−(13)
2 pA 10 s
gA−(13)
1000 800 600 400 200 0
AgA(15)
Control
1000 800 600 400 200 0
200 nM
1600
400 nM
1200 800 400
C
0 0
gA−(13)
1.0
N(t )/N(0)
0.8
1 Current (pA)
2
D
0.6
Control 200 nM 400 nM
0.4
1000
gA−(13)
AgA−(15)
AgA(15)
gT(15)
0.2 0.0 50
100
150
200
250
300 t/ms
0 1.0
AgA(15)
100
N(t )/N(0)
0.8 0.6
Control 200 nM 400 nM
0.4 0.2
10 0 200
0.0 0
500 1000 1500 2000 2500 3000
0 200 0 200 [GsMTx4] (nM)
0 200
Time (ms)
FIGURE 8 The eVect of GsMTx4 on gramicidin gating. (A) Raw channel data showing GsMTx4 increased open channel activity. (B) GsMTx4 decreased unitary conductance. (C) Open channel lifetimes increased in the presence of GsMTx4, that is, a decrease in the closing rate. (D) GsMTx4 increases of the lifetime of gA channels that vary in structure, including enantiomers (Suchyna et al., 2004).
Fig. 9, it has no measurable eVect on the action potential of resting isolated atrial myocytes from the rabbit (Bode et al., 2001) so that all of the channels and transporters responsible for the action potential are unaVected. A residual question in the literature has been the background activity of MSCs in cells. This could not be tested with a patch since adhesion of the patch to the glass always produces significant tension (Akinlaja and Sachs, 1998). However, adding GsMTx4 to resting cells has little eVect on cell electrophysiology,
95
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
Membrane voltage (mV)
60 40 Control 0.4-µM GsMTx4 4-µM GsMTx4
20 0 −20 −40 −60 −80 −100 0
50
100 Time (ms)
150
200
FIGURE 9 GsMtx4 does not aVect the action potential of resting rabbit atrial cells at 8 the KD. Perforated patch (amphotericin), n ¼ 5, 37 C (Sachs, 2004a,b).
although it does have eVects on spontaneous contractions of mdx muscle cells (Suchyna and Sachs, 2007). It has no significant eVect on the beating rabbit or sheep heart (Kalifa, Jalife, Gottlieb, and Sachs, unpublished data), unless the heart is distended (Bode et al., 2001). Preliminary experiments have shown that intravenous injection of GsMTx4 into mice (at 4 the Kd of MSCs in patch clamp) produces no behavioral change except for a slightly reduced water consumption over 24 h, perhaps by acting on the thirst center of the hypothalamus (Oliet and Bourque, 1996) or in the kidney. Even among MSCs, GsMTx4 is specific as it does not aVect auditory transduction (Marcotti et al., 2001) that may originate in a TRPA1 channel (Lin et al., 2005), or the activity of TREK channels (E. Honore, personal communication). The basis of this specificity is unclear, but probably arises from the channel itself; the only channels that respond to GsMTx4 are those for which local stress in the boundary layer couples to the energy of the closed or open states. IV. CELLULAR SITES FOR GsMTx4 A. TRPC1 Channel One of the major issues still outstanding is the identity of the channels aVected by GsMTx4. Although we are able to study various aspects of mechanosensitive channels, the exact protein stimulated by membrane
96
Gottlieb et al.
tension is being studied by a number of laboratories. The work by Hamill, Martinac, and collaborators suggests the TRPC1 channel is mechanically activated (Maroto et al., 2005), making it an excellent candidate for inhibition by GsMTx4. The transient response potential (TRP) ion channels were discovered in Drosophila when a mutation in the trp gene altered the organisms response to light (Minke, 1977; Montell et al., 1985). Human homologues were soon identified and 30 genes have been isolated, representing 6 families of ion channels (Clapham, 2003; Beech et al., 2004). One of these families is called the TRPC (Classical) and is a class of channels found predominantly in smooth muscle. TRPC1 presumably can form a homomeric, nonselective, cation channel, although it may also form heteromers with diVerent TRP subunits, making TRPC1 a critical ion channel in physiological function (Clapham, 2003; Beech et al., 2004). Maroto et al. (2005) introduced the human TRPC clone into oocytes, isolated the membrane proteins, fractionated them, and reconstituted an active fraction into liposomes. The mechanosensitivity they observed is associated with a protein whose molecular weight is consistent with that of TRPC1 and is inhibited by antibodies to TRPC1 and antisense RNA. The introduction of the TRPC1 gene into the Chinese hamster ovary (CHO) cell line seemed to support the ability of TRPC1 to be stretch sensitive. A natural question is whether GsMTx4 inhibits TRPC1. Following Maroto et al. (2005), we transfected CHO channels with TRPC1 DNA and were able to observe stretch sensitivity. Outside‐out patches from transfected cells were inhibited by 5 mM GsMTx4 acting as gating modifier (Gottlieb, Suchyna, Bowman, and Sachs, unpublished results; Hamill also found that GsMTx4 inhibits TRPC1, personal communication). However, more recent work on CHO and other cell lines in our laboratory has revealed that in all cell types there is a significant endogenous population of mechanosensitive channels exhibiting activity similar to the purported TRPC1. These channels may not be immediately visible in the recordings of native cells, but they are revealed by treatment with cytochalasin. As shown by Patel (Lauritzen et al., 2005), transfection with the gene for GFP is not a good control for the eVects of transfection‐the cell cares about what it is expressing. Expression of active or inactive 2P ion channels causes extensive modification of the cytoskeleton, whereas transfection by GFP has no eVect. To distinguish whether the channel under observation incorporates the subunit used for transfection requires the candidate be modified in a way that makes it diVerent from the endogenous channels. As a start we made dual mutants of the pore domain of TRPC1 that has a diVerent ion selectivity from the endogenous channel, and when expressed in COS7, these channels can be blocked by GsMTx4 (Gottlieb et al., in preparation).
97
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels 100 µM OAG 20
20 pA/pF
pA/pF
Control
10
+80 mV
Control
10
5 µM GsMTx4
GsMTx4
+80 mV 0
−100
0 −10
50 mV
−80 mV −80 mV
100 0
25 Time (s)
50
FIGURE 10 TRPC6 channel is blocked by GsMTx4 in whole‐cell recording using voltage ramps. TRPC6 was overexpressed in CHO cells and activated by membrane permeable DAG derivative, OAG. Left panel shows the I/V curve taken near the peak of the OAG response and shows the inhibition by GsMTx4. The outward rectifying current is typical for TRPC6. The right panel shows the time course of OAG stimulation at 80 mV. For illustrating GsMTx4 eVects, the cells had been incubated for 30 min in 5‐mM GsMTx4 (red symbols) prior to OAG. Courtesy M. Spassova. (See Color Insert.)
B. TRPC6 Channel TRPC6 channels are abundant in cells of tissue exposed to hydrostatic pressure changes such as vascular smooth muscle and glomerular podocytes where they may play a role in modulating myogenic tone. Recent data from Don Gill’s laboratory (Spassova et al., 2006) shows that TRPC6, a stretch sensitive channel found in smooth muscle, is blocked by GsMTx4. TRPC6 is a nonselective cation channel that is activated by receptor‐induced phospholipase C (PLC) activation probably via a direct eVect of diacylglycerol (DAG). GsMTx4 inhibited both stretch‐ and DAG‐activated currents that argue for a common mechanism of activation (Fig. 10) such as stress within the boundary lipids.
V. POTENTIAL THERAPEUTIC USES FOR GsMTx4 A. Cardiac Myocytes and Atrial Fibrillation The connection between mechanical stress and excitability in the heart (known as mechanoelectric feedback or MEF) has been appreciated for over a century (Kohl and Sachs, 2002; Kohl et al., 2003). The initiation of arrhythmias is a classic eVect of stretch on the heart (Baumgarten, 2004;
98
Gottlieb et al.
Sachs, 2004b). MEF is not a trivial modulation as the ventricles of an intact heart can be reliably stimulated to repeatedly contract in phase with inflation of an intraventricular balloon (Franz et al., 1992). Numerous studies have shown that stretch can aVect both the cardiac action potential and whole‐cell currents (Sachs, 2004a), and MSC activity (Bett and Sachs, 1997; Hu and Sachs, 1997). Recently, a mechanosensitive maxiK channel was cloned from chick heart (Tang et al., 2003). Voltage‐sensitive ion channels in the heart have also been shown to be mechanosensitive (see Chapter 11). Given the presence of MSCs, pharmacological agents that aVect mechanosensitivity can produce therapeutic eVects. It has been known for a long time that atrial fibrillation is potentiated by mechanical stress. If that is driven by cationic MSCs, the stretch‐induced eVects may be inhibited by GsMTx4. In LangendorV‐perfused rabbit hearts, GsMTx4 at 170 nM eVectively blocked atrial fibrillation potentiated by atrial dilation (Fig. 11; Bode et al., 2001). GsMTx4 provided complete protection against fibrillation up to a diastolic pressure of 15 cmH2O, and shortened the duration of fibrillation at all pressures. In the unstressed rabbit heart, ionotropy and the action potential were unaVected by GsMTx4. At nearly 10 the Kd for MSCs, the action potential of resting atrial cells was unaVected suggesting that there is little MSC activity at rest (Sachs, 2004a).
AF inducibility (%)
100 80 60 40 20 0 0
5
10 15 20 Intraatrial pressure (cmH2O)
25
30
FIGURE 11 Block of dilation potentiated atrial fibrillation (AF) by 170‐nM GsMTx4 in the LangendorV‐perfused rabbit heart. AF was induced by a burst of rapid electrical stimulation, and we have plotted how often that resulted in AF lasting longer than 60 s. Inflation increased the probability of AF (open symbols), and GsMTx4 shifted the dose–response curve to higher pressures (filled symbols; Bode et al., 2001).
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
99
The work on human heart tissue has also shown minimal eVect of GsMTx4 on ionotropy (Kockskamper et al., submitted for publication). Marban’s group described an arrhythmia with after depolarizations that occurred on termination of high‐speed pacing, and these were not blocked by any inhibitors of known channels (Nuss et al., 1999). Baumgarten and Clemo reproduced this result with isolated rabbit ventricular and atrial (Sachs, 2004a), and surprisingly found that GsMTx4 blocked the after depolarizations. Perhaps high‐speed pacing overloads the SR causing spontaneous Ca2þ release, and in turn this causes local contractions. The nonuniform stresses then activate the cationic MSCs. This example of GsMTx4 diagnosis may be a prototype of how these compounds can be used for physiological diagnosis of MSC function, as well as serving as lead compounds for drug development.
B. Muscular Dystrophy An intriguing connection between MSCs and physiological function exists for the muscular dystrophies that involve mutations in the dystroglycan complex (DGC), but not MSCs. The DGC is a multiprotein group that connects the internal actin cytoskeleton to the extracellular matrix through attachments to laminin (reviewed in Blake and Martin‐Rendon, 2002), and presumably provide structural support to the bilayer. Mutations in this complex lead to diVerent forms of muscular dystrophy that vary in severity, susceptible muscle groups, and the age of phenotypic onset (Khurana and Davies, 2003). Duchenne muscular dystrophy (DMD) is the most common of these disorders, caused by mutations to the large cytoskeletal protein dystrophin. Most dystrophies associated with mutations to proteins in the DGC are characterized by elevated internal Naþ and Ca2þ levels, and excessive protein degradation (Ruegg et al., 2002). Thus, a major strategy for acute therapy development can be to improve Ca2þ homeostasis (Khurana and Davies, 2003). Franco‐Obregon and Lansman (2002) have studied the eVect of pressure stimuli on patches from mouse mdx myoblasts (a mouse DMD model). They demonstrated that the SACs observed in wild‐type cells become hyperactive in mdx cells with prolonged stimulation, as though a reinforcing structure was being disrupted in the mdx cells. They proposed a decoupling mechanism whereby disruption of viscoelastic elements associated with the membrane led to a loss of membrane tension regulation and increased channel open probability. Suchyna and Sachs (Suchyna et al., 2004) and others (Hamill et al., 1992; Niu and Sachs, 2003) have shown that MSCs not only activate with stretch but inactivate with time (Honore et al., 2006). In rat astrocytes, channel activation closely follows the change in membrane tension, but when inactivation is disrupted, the channels stay open during continued stress since
100
Gottlieb et al.
inactivation, which is only accessible through the open states, depends on cytoskeletal integrity (Suchyna et al., 2004). The lack of membrane reinforcement by dystrophin can cause hyperactivity of SACs in dystrophic muscle. Yeung et al. (2003) linked the pathology of mdx muscle fibers with the influx of cations through SACs. They showed that both Gd3þ and streptomycin, nonspecific blockers of SACs, reduce elevated Naþ levels in stretched mdx fibers, while having little eVect on normal muscle fibers. They postulated that the cation selective SACs in mdx fibers are responsible for the increased influx of Naþ and Ca2þ (Naþ influx can lead to additional Ca2þ influx through the Naþ/Ca2þ exchanger; Bosteels et al., 1999; Arnon et al., 2000). Further work by Allen’s group evaluated the eVect of GsMTx4 on mdx muscle fibers and Ca2þ leakage currents that are presumed to be responsible for muscular atrophy (Yeung et al., 2005). Single muscle fibers isolated from mdx mice were isolated and subjected to a series of stretched (eccentric) tetanic contractions while measuring intracellular calcium with flou‐3 and confocal microscopy. In the absence of GsMTx4, there was a slow rise in resting intracellular Ca2þ levels after tetanic stimulation, and both Ca2þ influx and the force generated during tetanus were reduced. GsMTx4 reduced this eVect by reducing Ca2þ influx through hyperactive SACs, and this in turn inhibited the increased force.
C. Astrocytes and Gliosis Astrocytes are the most abundant cells in the CNS. They secrete substances that act on themselves, on other types of glia, vascular endothelial cells, and neurons (Araque et al., 2001). They play an important role in the induction and maintenance of the blood‐brain barrier (Zonta et al., 2003) and communicate intracellular Ca2þ signals bidirectionally with neurons (Carmignoto, 2000; Zonta and Carmignoto, 2002; Burgo et al., 2003). SACs exist in fetal, neonatal, and adult astrocytes (Ding et al., 1988; Bowman et al., 1992; Sontheimer, 1994; Ostrow et al., 2001) suggesting they may have a role in a variety of brain pathology including glial tumors (Ostrow and Sachs, 2005). Glial stretch not only activates SACs but also releases endothelin‐1 (ET‐1). ET is a peptide and potent autocrine mitogen produced by reactive and neoplastic astrocytes. The relationship between mechanical stretch and ET production in other cell types, combined with the observation that ET‐1‐positive reactive astrocytes appear in the mechanically deformed periphery of CNS pathology, led us to hypothesize that mechanical stress may regulate ET secretion in astrocytes through SACs. Ostrow and Sachs demonstrated that mechanical stimulation of adult rat astrocyte cultures causes a dramatic increase in ET‐1 production and secretion
101
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
into the culture medium (Ostrow et al., 2000). These experiments represented the first demonstration that the astrocytic ET system can be directly stimulated by cell deformation (Fig. 12). Since virtually all brain pathology is associated with some degree of mechanical deformation of the surrounding parenchyma, and ET induces the proliferation of astrocytes (and glioma cells), the possibility arises that mechanical induction of the ET system represents a general pathway for activating/augmenting astroglial proliferation. ET‐1 exerts its mitogenic eVects in most cell types, including astrocytes and glioma cells, through Ca2þ homeostasis, possibly explaining how ET potentiates its own production and secretion (Supattapone et al., 1989; Marsault et al., 1990; Supattapone et al., 1990; Marin et al., 1991; Holzwarth et al., 1992; Jacques et al., 2000). Nonmechanical stimuli including growth factors and cytokines can also enhance ET expression in astrocytes (Masaki et al., 1991; Tasaka et al., 1991; Ehrenreich et al., 1993; Brunner, 1995; Goto et al., 1996). Regardless of the stimulus, alterations in cell Ca2þ seem to be the common link in regulating the ET system in all cell types (Yanagisawa et al., 1989; Masaki et al., 1991;
80 (6)
ET-1 (pg/well) in culture media
70
(4) Stretched
60
(3)
50 40 Unstretched
(3) 30 20
(3)
(2)
(6)
(3)
27:00
48:00
(2)
10 00:00
09:00
18:00
Duration of stretch (hh:mm) FIGURE 12 EVects of cyclic stretch (0.1 Hz, 0–20%) on ET‐1 secretion. Both control and stretched cell cultures were incubated in Starvation Medium (0.1% serum) for 24 h prior to stretching. The media was aspirated and replaced with fresh Starvation Medium 5 min before stretching commenced. The green points represent data from a separate 2‐day stretching protocol on cells of a similar age and passage number. (Numbers in parentheses are the number of culture wells per data point. Each well was run in duplicate in the ET‐1 ELISA. Error bars ¼ standard error; Ostrow et al., 2000).
102
Gottlieb et al.
Kitazumi and Tasaka, 1992; Corder et al., 1993; Brunner et al., 1994; Morita et al., 1994; Brunner, 1995). The Ca2þ eVects seem to ultimately depend on activation of the mitogen‐activated protein kinase (MAPK) pathway and the induction of PLC and protein kinase C (PKC) (Goto et al., 1996). An inhibitor of SACs could provide a novel therapeutic approach to gliomas by controlling a mitogen whose overproduction is associated with pathological states. Supportive of this view, Vaz et al. (1998) demonstrated that the degree of edema following head trauma is reduced by the nonspecific SAC inhibitors Gd3þ, amiloride, and gentamicin, implying that SACs are activated by swelling. Having already demonstrated that GsMTx4 blocks SACs in astrocytes, we showed it also inhibits the stretch‐induced ET‐1 production by astrocytes. It is striking that GsMTx4 does not alter basal ET secretion in the absence of stretch, so that its action is specific. As expected, stretch‐induced ET secretion was also decreased by lowering extracellular Ca2þ (Ostrow, 2003). Gliomas could grow in a positive feedback cycle where cell division increases local stress, causing ET‐1 secretion and further division. An inhibitor of SACs might block or at least slow glioma development. Particularly appealing is the use of the D form of GsMTx4, since it is nonhydrolyzable, significantly hydrophobic, and might be able to be administered orally. It is not yet known whether GsMTx4 will cross an intact blood–brain barrier, but it is known that the barrier can be temporarily relaxed (Borlongan and Emerich, 2003).
D. Neurite Growth Extension GsMTx4 was shown to stimulate neurite growth extension (Jacques‐Fricke et al., 2006). The importance of intracellular Ca2þ signals in the regulation of neurite outgrowth is well known, although the mode of entry is not well defined. To test whether calcium influx through MSCs is important for neurite growth, Gomez and colleagues used a series of agents, both specific and nonspecific, including GsMTx4, in an assay that directly measures neurite extension. Of all the reagents used, GsMTx4 was the most potent for stimulating neurite growth in Xenopus neurons. The role of GsMTx4 in this physiological response was shown to involve a reduction of Ca2þ influx through channels activated by cell swelling. The eVective inhibition of GsMTx4 is not via global Ca2þ levels, meaning the mean cytoplasmic Ca2þ is unaVected by the presence of the peptide. Rather, GsMTx4 inhibits SACs in microdomains, a fact deduced by the sensitivity of the outgrowth to diVerent Ca2þ buVers (Neher, 1998). The inactivation of these channels by GsMTx4 prevents activation of downstream processes that inhibit neurite outgrowth.
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
103
VI. CONCLUSIONS GsMTx4 is the first compound known to specifically inhibit cationic MSCs in eukaryotes. The mechanism of action is a gating modifier, but it does not fit the familiar hand–glove model of peptide channel interactions since the D form is also active. The mode of action is probably to create stress within the boundary lipid region of the channel. GsMTx4 has begun to serve for the first time as a tool to study the physiological role of MSCs in situ. Since all physiology is subject to pathology, GsMTx4 may serve as a therapeutic tool or a lead compound to treat diseases resulting from defects in the mechanotransducing pathways. References Akinlaja, J., and Sachs, F. (1998). The breakdown of cell membranes by electrical and mechanical stress. Biophys. J. 75, 247–254. Andersen, O. S., Nielsen, C., Maer, A. M., Lundbæk, J. A., Goulian, M., and Koeppe, R. E. (1999). Ion channels as tools to monitor lipid bilayer‐membrane protein interactions: Gramicidin channels as molecular force transducers. Methods Enzymol. 294, 208–224. Araque, A., Carmignoto, G., and Haydon, P. G. (2001). Dynamic signaling between astrocytes and neurons. Annu. Rev. Physiol. 63, 795–813. Arnon, A., Hamlyn, J. M., and Blaustein, M. P. (2000). Naþ entry via store‐operated channels modulates Ca2þ signaling in arterial myocytes. Am. J. Physiol. Cell Physiol. 278, C163–C173. Bass, R. B., Locher, K. P., Borths, E., Poon, Y., Strop, P., Lee, A., and Rees, D. C. (2003). The structures of BtuCD and MscS and their implications for transporter and channel function. FEBS Lett. 555, 111–115. Baumgarten, C. M. (2004). Cell‐volume activated ion channels in cardiac cells. In ‘‘Cardiac Mechano‐Electric Feedback and Arrhythmias: From Pipette to Patient’’ (P. Kohl, M. R. Franz, and F. Sachs, eds.). Elsevier, Philadelphia. Beech, D. J., Muraki, K., and Flemming, R. (2004). Non‐selective cationic channels of smooth muscle and the mammalian homologues of Drosophila TRP. J. Phys. (Lond.), 559, 685–706. Besch, S. R., Suchyna, T. M., and Sachs, F. (2002). High‐speed pressure clamp. Pflugers Arch. 445, 161–166. Betanzos, M., Chiang, C. S., Guy, H. R., and Sukharev, S. I. (2002). A large iris‐like expansion of a mechanosensitive channel protein induced by membrane tension. Nat. Struct. Biol. 9, 704–710. Bett, G. C. L., and Sachs, F. (1997). Cardiac mechanosensitivity and stretch activated ion channels. Trends Cardiovasc. Med. 7, 4–9. Blake, D. J., and Martin‐Rendon, E. (2002). Intermediate filaments and the function of the dystrophin‐protein complex. Trends Cardiovasc. Med. 12, 224–228. Blount, P., Sukharev, S. I., Moe, P. C., Martinac, B., and Kung, C. (1999). Mechanosensitive channels of bacteria. Methods Enzymol. 294, 458–482. Bode, F., Sachs, F., and Franz, M. R. (2001). Tarantula peptide inhibits atrial fibrillation. Nature 409, 35–36. Borlongan, C. V., and Emerich, D. F. (2003). Facilitation of drug entry into the CNS via transient permeation of blood brain barrier: Laboratory and preliminary clinical evidence from bradykinin receptor agonist, Cereport. Brain Res. Bull. 60, 297–306.
104
Gottlieb et al.
Bosteels, S., Matejovic, P., Flameng, W., and Mubagwa, K. (1999). Sodium influx via a non‐ selective pathway activated by the removal of extracellular divalent cations: Possible role in the calcium paradox. Cardiovasc. Res. 43, 417–425. Bowman, C. L., Ding, J. P., Sachs, F., and Sokabe, M. (1992). Mechanotransducing ion channels in astrocytes. Brain Res. 584, 272–286. Brunner, F. (1995). Dependence of endothelin‐1 secretion on Ca2þ. Biochem. Pharmacol. 49, 1785–1791. Brunner, F., Stessel, H., Simecek, S., Graier, W., and Kukovetz, W. R. (1994). EVect of intracellular Ca2þ concentration on endothelin‐1 secretion. FEBS Lett. 350, 33–36. Burgo, A., Carmignoto, G., Pizzo, P., Pozzan, T., and Fasolato, C. (2003). Paradoxical Ca2þ rises induced by low external Ca2þ in rat hippocampal neurones. J. Physiol. 549, 537–552. Calabrese, B., Tabarean, I. V., Juranka, P., and Morris, C. E. (2002). Mechanosensitivity of N‐type calcium channel currents. Biophys. J. 83, 2560–2574. Carmignoto, G. (2000). Astrocyte‐neurone crosstalk: Variants of the same language? Trends Pharmacol. Sci. 21, 373–375. Chang, G., Spencer, R. H., Lee, A. T., Barclay, M. T., and Rees, D. C. (1998). Structure of the MscL homolog from Mycobacterium tuberculosis: A gated mechanosensitive ion channel. Science 282, 2220–2226. Christensen, O. (1987). Mediation of cell volume regulation by Ca2þ influx through stretch‐ activated channels. Nature 330, 66–68. Clapham, D. E. (2003). TRP channels as cellular sensors. Nature 426, 517–524. Corder, R., Khan, N., Anggard, E. E., and Vane, J. R. (1993). Calcium ionophores inhibit the release of endothelin‐1 from endothelial cells. J. Cardiovasc. Pharmacol. 22(Suppl. 8), S42–S45. Corey, D. (2003a). Sensory transduction in the ear. J. Cell Sci. 116, 1–3. Corey, D. P. (2003b). New TRP channels in hearing and mechanosensation. Neuron 39, 585–588. Ding, J. P., Yang, X. C., Bowman, C. L., and Sachs, F. (1988). A stretch‐activated ion channel in rat astrocytes in primary cell culture. Soc. Neurosci. Abstr. 14(Pt. 2), 1056. Ehrenreich, H., Costa, T., Clouse, K. A., Pluta, R. M., Ogino, Y., Coligan, J. E., and Burd, P. R. (1993). Thrombin is a regulator of astrocytic endothelin‐1. Brain Res. 600, 201–207. Ermakov, Y., Averbakh, A. Z., Arbuzova, A. B., and Sukharev, S. I. (1998). Lipid and cell membranes in the presence of gadolinium and other ions with high aYnity to lipids. 2. A dipole component of the boundary potential on membranes with diVerent surface charge. Membr. Cell Biol. 12, 411–426. Ermakov, Y. A., Averbakh, A. Z., Yusipovich, A. I., and Sukharev, S. I. (2001). Dipole potentials indicate restructuring of the membrane interface induced by gadolinium and beryllium ions. Biophys. J. 80, 1851–1862. Franco‐Obregon, A., and Lansman, J. B. (2002). Changes in mechanosensitive channel gating following mechanical stimulation in skeletal muscle myotubes from the mdx mouse. J. Physiol. 539, 391–407. Franz, M. R., Cima, R., Wang, D., Profitt, D., and Kurz, R. (1992). Electrophysiological eVects of myocardial stretch and mechanical determinants of stretch‐activated arrhythmias [published erratum appears in Cir. 86: 1663 (1992)]. Circulation 86, 968–978. Garcia‐Anoveros, J., and Corey, D. P. (1996). Mechanosensation: Touch at the molecular level. Curr. Biol. 6, 541–543. Gillespie, P. G., and Walker, R. G. (2001). Molecular basis of mechanosensory transduction. Nature 413, 194–202. Gomez, S., Dibello, C., Hung, L. T., Genet, R., Morgat, J. L., Fromageot, P., and Cohen, P. (1984). C‐terminal amidation of neuropeptides—gly‐lys‐arg extension an eYcient precursor of C‐terminal amide. FEBS Lett. 167, 160–164.
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
105
Goto, K., Hama, H., and Kasuya, Y. (1996). Molecular pharmacology and pathophysiological significance of endothelin. Jpn. J. Pharmacol. 72, 261–290. Gu, C. X., Juranka, P. F., and Morris, C. E. (2001). Stretch‐activation and stretch‐inactivation of Shaker‐IR, a voltage‐gated Kþ channel. Biophys. J. 80, 2678–2693. Hamill, O. P., and Martinac, B. (2001). Molecular basis of mechanotransduction in living cells. Physiol. Rev. 81, 685–740. Hamill, O. P., and McBride, D. W., Jr. (1995). Pressure/patch‐clamp methods. In ‘‘Neuromethods’’ (A. Boulton, G. Baker, and W. Walz, eds.), pp. 75–87. Humana Press Inc., Totowa, NJ. Hamill, O. P., Lane, J. W., and McBride, D. W., Jr. (1992). Amiloride: A molecular probe for mechanosensitive channels. Trends Pharmacol. Sci. 13, 373–376. Holzwarth, J. A., Glaum, S. R., and Miller, R. J. (1992). Activation of endothelin receptors by sarafotoxin regulates Ca2þ homeostasis in cerebellar astrocytes. GLIA 5, 239–250. Honore, E., Suchyna, T. M., Patel, A. A., and Sachs, F. (2006). Desensitization of cloned 2P domain K channels. Proc. Natl. Acad. Sci. USA 103, 6859–6864. Howard, J., and Hudspeth, A. J. (1988). Compliance of the hair bundle associated with gating of mechanoelectrical transduction channels in the Bullfrog’s saccular hair cell. Neuron 1, 189–199. Hu, H., and Sachs, F. (1997). Stretch‐activated ion channels in the heart. J. Mol. Cell Cardiol. 29, 1511–1523. Hudspeth, A. J., Choe, Y., Mehta, A. D., and Martin, P. (2000). Putting ion channels to work: Mechanoelectrical transduction, adaptation, and amplification by hair cells. Proc. Natl. Acad. Sci. USA 97, 11765–11772. Hwang, T. C., Koeppe, R. E., and Andersen, O. S. (2003). Genistein can modulate channel function by a phosphorylation‐independent mechanism: Importance of hydrophobic mismatch and bilayer mechanics. Biochemistry 42, 13646–13658. Jacques, D., Sader, S., Choufani, S., Orleans‐Juste, P., and Charest, D. (2000). Endothelin‐1 regulates cytosolic and nuclear Ca2þ in human endocardial endothelium. J. Cardiovasc. Pharmacol. 36, S397–S400. Jacques‐Fricke, B. T., Seow, Y., Gottlieb, P. A., Sachs, F., and Gomez, T. M. (2006). Ca2þ influx through a mechanosensitive channel and release from intracellular stores have opposite eVects on neurite outgrowth., J. Neurosci. 26. Jiang, Y., Lee, A., Chen, J., Ruta, V., Cadene, M., Chait, B. T., and MacKinnon, R. (2003a). X‐ray structure of a voltage‐dependent Kþ channel. Nature 423, 33–41. Jiang, Y., Ruta, V., Chen, J., Lee, A., and MacKinnon, R. (2003b). The principle of gating charge movement in a voltage‐dependent Kþ channel. Nature 423, 42–48. Khurana, T. S., and Davies, K. E. (2003). Pharmacological strategies for muscular dystrophy. Nat. Rev. Drug Discov. 2, 379–390. Kitazumi, K., and Tasaka, K. (1992). Thrombin‐stimulated phosphorylation of myosin light chain and its possible involvement in endothelin‐1 secretion from porcine aortic endothelial‐cells. Biochem. Pharmacol. 43, 1701–1709. Kohl, P., and Sachs, F. (2002). Mechanoelectric feedback in cardiac cells. Proc. R. Soc. Lond. Math. Phys. Eng. Sci. 359, 1173–1185. Kohl, P., Cooper, P. J., and Holloway, H. (2003). EVects of acute ventricular volume manipulation on in situ cardiomyocyte cell membrane configuration. Prog. Biophys. Mol. Biol. 82, 221–227. Laitko, U., and Morris, C. E. (2004). Membrane tension accelerates rate‐limiting voltage‐ dependent activation and slow inactivation steps in a shaker channel. J. Gen. Physiol. 123, 135–154.
106
Gottlieb et al.
Lauritzen, I., Chemin, J., Honore, E., Jodar, M., Guy, N., Lazdunski, M., and Patel, A. J. (2005). Cross‐talk between the mechano‐gated K‐2P channel TREK‐1 and the actin cytoskeleton. EMBO Rep. 6, 642–648. Lee, C. W., Kim, S., Roh, S. H., Endoh, H., Kodera, Y., Maeda, T., Kohno, T., Wang, J. M., Swartz, K. J., and Kim, J. I. (2004). Solution structure and functional characterization of SGTx1, a modifier of Kv2.1 channel gating. Biochemistry 43, 890–897. Lee, S. Y., and MacKinnon, R. (2004). A membrane‐access mechanism of ion channel inhibition by voltage sensor toxins from spider venom. Nature 430, 232–235. Lewis, R. J., and Garcia, M. L. (2003). Therapeutic potential of venom peptides. Nat. Rev. Drug Discov. 2, 790–802. Lin, S. Y., Holt, J. R., Vollrath, M. A., Garcia‐Anoveros, J., Geleoc, G., Kwan, K., HoVman, M. P., Zhang, D. S., and Corey, D. P. (2005). TRPA1 is a candidate for the mechanosensitive transduction channel of vertebrate hair cells. Biophys. J. 88, 287A–288A. Marcotti, W., Sachs, F., Ashmore, J. F., and Kros, C. J. (2001). EVect of a peptide tarantula toxin on mechano‐transduction in neonatal mouse cochlear hair cells. Int. J. Audiol. 41, 231. Marin, P., Delumeau, J. C., Durieu‐Trautmann, O., Le Nguyen, D., Premont, J. S., and Couraud, P. O. (1991). Are several G proteins involved in the diVerent eVects of endothelin‐1 in mouse striatal astrocytes? J. Neurochem. 56, 1270–1275. Markin, V., and Sukharev, S. (2000). A physical model of the bacterial large‐conductance mechanosensitive channel, MscL. Biophys. J. 78, 473A. Maroto, R., Raso, A., Wood, T. G., Kurosky, A., Martinac, B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nat. Cell Biol. 7, 179–189. Marsault, R., Vigne, P., Breittmayer, J. P., and Frelin, C. (1990). Astrocytes are target cells for endothelins and sarafotoxin. J. Neurochem. 54(6), 2142–2144. Martinac, B. (2001). Mechanosensitive channels in prokaryotes. Cell Physiol. Biochem. 11, 61–76. Martinac, B., and Kloda, A. (2003). Evolutionary origins of mechanosensitive ion channels. Prog. Biophys. Mol. Biol. 82, 11–24. Masaki, T., Kimura, S., Yanagisawa, M., and Goto, K. (1991). Molecular and cellular mechanism of endothelin regulation. Implications for vascular function. Circulation 84, 1457–1468 [Review]. Mills, L. R., and Morris, C. E. (1998). Neuronal plasma membrane dynamics evoked by osmomechanical perturbations. J. Membr. Biol. 166, 223–235. Minke, B. (1977). Drosophila mutant with a transducer defect. Biophys. Struct. Mech. 3, 59–64. Montell, C., Jones, K., Hafen, E., and Rubin, G. (1985). Rescue of the drosophila phototransduction mutation Trp by germline transformation. Science 230, 1040–1043. Morita, T., Kurihara, H., Maemura, K., Yoshizumi, M., Nagai, R., and Yazaki, Y. (1994). Role of Ca2þ and protein kinase C in shear stress‐induced actin depolymerization and endothelin 1 gene expression. Circ. Res. 75, 630–636. Morris, C. E. (2004). Stretch eVects on voltage‐gated channels in cardiac cells. In ‘‘Cardiac Mechano‐Electric Feedback and Arrhythmias: From Pipette to Patient’’ (P. Kohl, F. Sachs, and M. R. Franz, eds.). Elsevier, Philadelphia. Neher, E. (1998). Usefulness and limitations of linear approximations to the understanding of Caþþ signals. Cell Calcium 24, 345–357. Niu, W. Z., and Sachs, F. (2003). Dynamic properties of stretch‐activated Kþ channels in adult rat atrial myocytes. Prog. Biophys. Mol. Biol. 82, 121–135. Nuss, H. B., Kaab, S., Kass, D. A., Tomaselli, G. F., and Marban, E. (1999). Cellular basis of ventricular arrhythmias and abnormal automaticity in heart failure. Am. J. Physiol. Heart Circ. Physiol. 277, H80–H91.
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
107
Oliet, S. H. R., and Bourque, C. W. (1996). Gadolinium uncouples mechanical detection and osmoreceptor potential in supraoptic neurons. Neuron 16, 175–181. Ostrow, K. L., Mammoser, A., Suchyna, T., Sachs, F., Oswald, R. E., Kubo, S., Chino, N., and Gottlieb, P. A. (2003). cDNA sequence and in vitro folding of GsMTx4, a specific peptide inhibitor of mechanosensitive channels. Toxicon 42, 263–274. Ostrow, L., and Sachs, F. (2005). Mechanosensation and endothelin in astrocytes‐potential roles in reactive gliosis. Brain Res. Rev. 48, 488–508. Ostrow, L., Suchnya, T. M., and Sachs, F. (2001). Mechanical induction of endothelin‐1 is inhibited by gsmtx‐4, the first specific blocker of stretch‐activated ion channels. ET‐7: The Seventh International conference on Endothelin. Ostrow, L., Langan, T., and Sachs, F. (2000). Stretch‐induced endothelin‐1 production by astrocytes. J. Cardiovasc. Pharmacol. 36, S274–S277. Ostrow, L. W. (2003). Mechanosensation and endothelin in astrocytes. Thesis/Dissertation. Department of Physiology and Biophysics, School of Medicine and Biomedical Sciences, State University of New York at BuValo. Oswald, R. E., Suchyna, T. M., McFeeters, R., Gottlieb, P.,A., and Sachs, F. (2002). Solution structure of peptide toxins that block mechanosensitive ion channels. J. Biol. Chem. 277, 34443–34450. Pallaghy, P. K., Nielsen, K. J., Craik, D. J., and Norton, R. S. (1994). A common structural motif incorporating a cystine knot and a triple‐stranded beta‐sheet in toxic and inhibitory polypeptides. Protein Sci. 3, 1833–1839. Patel, A. J., Lazdunski, M., and Honore, E. (2001). Lipid and mechano‐gated 2P domain K(þ) channels. Curr. Opin. Cell Biol. 13, 422–428. Perozo, E., Kloda, A., Cortes, D. M., and Martinac, B. (2002). Physical principles underlying the transduction of bilayer deformation forces during mechanosensitive channel gating. Nat. Struct. Biol. 9, 696–703. Posokhov, Y. G., Gottlieb, P., Morales, M. J., Sachs, F., and Ladokhin, A. S. (2006). Quantitative characterization of membrane interactions of the cysteine knot (CK) family of ion‐channel blockers. Biophys. J. Pos145. Ruegg, U. T., Nicolas‐Metral, V., Challet, C., Bernard‐Helary, K., Dorchies, O. M., Wagner, S., and Buetler, T. M. (2002). Pharmacological control of cellular calcium handling in dystrophic skeletal muscle. Neuromuscul. Disord. 12(Suppl. 1), S155–S161. Sachs, F. (2002). Retaining your identity under stress. Nat. Struct. Biol. 9, 636–637. Sachs, F. (2004a). Heart mechanoelectric transduction. In ‘‘Cardiac Electrophysiology; From Cell to Bedside’’ (J. Jalife and D. Zipes, eds.), pp. 96–102. Saunders (Elsevier), Philadelphia. Sachs, F. (2004b). Stretch activated channels in the heart. In ‘‘Cardiac Mechano‐Electric Feedback and Arrhythmias: From Pipette to Patient’’ (P. Kohl, M. R. Franz, and F. Sachs, eds.). Saunders (Elsevier), Philadelphia. Sachs, F., and Morris, C. E. (1998). Mechanosensitive ion channels in nonspecialized cells. Rev. Physiol. Biochem. Pharmacol. 132, 1–77. Sheetz, M. P., and Singer, S. J. (1974). Biological membranes as bilayer couples. A molecular mechanism of drug‐erythrocyte interactions. Proc. Natl. Acad. Sci. USA 71, 4457–4461. Sheetz, M. P., Painter, R. G., and Singer, S. J. (1976). Biological membranes as bilayer couples. III. Compensatory shape changes induced in membranes. J. Cell Biol. 70, 193–203. Sontheimer, H. (1994). Voltage‐dependent ion channels in glial cells. GLIA 11, 156–172 [Review]. Spassova, M. A., Hewavitharana, T., Xu, W., Soboloff, J., and Gill, D. L. (2006). A common mechanism underlies stretch activation and receptor activation of TRPC6 channels. Proc. Natl. Acad. Sci. USA 103, 16586–16591.
108
Gottlieb et al.
Suchyna, T., and Sachs, F. (2007). Mechanical and electrical properties of membranes from dystrophic and normal mouse muscle. J. Physiol. (Lond.). Suchyna, T. M., and Sachs, F. (2004). Dynamic regulation of mechanosensitive channels; capacitance used to monitor patch tension in real time. Phys. Biol. 1, 1–18. Suchyna, T. M., Johnson, J. H., Hamer, K., Leykam, J. F., Gage, D. A., Clemo, H. F., Baumgarten, C. M., and Sachs, F. (2000). Identification of a peptide toxin from Grammostola spatulata spider venom that blocks cation‐selective stretch‐activated channels. J. Gen. Physiol. 115, 583–598. Suchyna, T. M., Tape, S. E., Koeppe, R. E., Andersen, O. S., Sachs, F., and Gottlieb, P. A. (2004). Bilayer‐dependent inhibition of mechanosensitive channels by neuroactive peptide enantiomers. Nature 430, 235–240. Sukharev, S. I., and Anishkin, A. (2004). Mechanosensitive channels: What can we learn from ‘simple’ model systems? Trends Neurosci. 27, 345–351. Sukharev, S. I., Blount, P., Martinac, B., Blattner, F. R., and Kung, C. (1994). A large conductance mechanosensitive channel in, E. coli encoded by mscL alone. Nature 368, 265–268. Sukharev, S. I., Blount, P., Martinac, B., and Kung, C. (1997). Mechanosensitive channels of Escherichia coli: The MscL gene, protein and activities. Annu. Rev. Physiol. 59, 633–657. Sukharev, S. I., Sigurdson, W., Kung, C., and Sachs, F. (1999). Energetic and spatial parameters for gating of the bacterial large conductance mechanosensitive channel, MscL. J. Gen. Physiol. 113, 525–539. Supattapone, S., Simpson, A. W., and Ashley, C. C. (1989). Free calcium rise and mitogenesis in glial cells caused by endothelin. Biochem. Biophys. Res. Commun. 165, 1115–1122. Supattapone, S., Simpson, A. W., and Ashley, C. C. (1990). Endothelin causes a free calcium rise in rat astrocytes. Adv. Second Messenger Phosphoprotein Res. 24, 429–433. Tang, Q. Y., Qi, Z., Naruse, K., and Sokabe, M. (2003). Characterization of a functionally expressed stretch‐activated BKca channel cloned from chick ventricular myocytes. J. Membr. Biol. 196, 185–200. Tasaka, K., Mio, M., Akagi, M., Fujisawa, K., and Aoki, I. (1991). Role of the cytoskeleton in Ca2þ release from the intracellular Ca store of rat peritoneal mast cells. Agents Actions 33, 44–47. Vandorpe, D. H., Small, D. L., Dabrowski, A. R., and Morris, C. E. (1994). FMRFamide and membrane stretch as activators of the Aplysia S‐channel. Biophys. J. 66, 46–58. Vaz, R., Sarmento, A., Borges, N., Cruz, C., and Azevedo, I. (1998). EVect of mechanogated membrane ion channel blockers on experimental traumatic brain oedema. Acta Neurochir. 140, 371–375. Wan, X., Harris, J. A., and Morris, C. E. (1995). Responses of neurons to extreme osmo‐ mechanical stress. J. Membr. Biol. 145, 21–31. Wang, J. M., Roh, S. H., Kim, S., Lee, C. W., Kim, J. I., and Swartz, K. J. (2004). Molecular surface of tarantula toxins interacting with voltage sensors in kv channels. J. Gen. Physiol. 123, 455–467. White, S. H., Wimley, W. C., Ladokhin, A. S., and Hristova, K. (1998). Protein folding in membranes: Determining energetics of peptide‐bilayer interactions. Methods Enzymol. 295, 62–87. Wiggins, P., and Phillips, R. (2004). Analytic models for mechanotransduction: Gating a mechanosensitive channel. Proc. Natl. Acad. Sci. USA 101, 4071–4076. Yanagisawa, M., Inoue, A., Takuwa, Y., Mitsui, Y, Kobayashi, M, and Masaki, T. (1989). The human preproendothelin‐1 gene: Possible regulation by endothelial phosphoinositide turnover signaling. J. Cardiovasc. Pharmacol. 13(Suppl. 5), S13–S17.
4. GsMTx4 Inhibition of Mechanosensitive Ion Channels
109
Yeung, E. W., Head, S. I., and Allen, D. G. (2003). Gadolinium reduces short‐term stretch‐induced muscle damage in isolated mdx mouse muscle fibres. J. Physiol. 552, 449–458. Yeung, E. W., Whitehead, N. P., Suchyna, T. M., Gottlieb, P. A., Sachs, F., and Allen, D. G. (2005). EVects of stretch‐activated channel blockers on [Ca2þ]i and muscle damage in the mdx mouse. J. Physiol. Lond. 562, 367–380. Zonta, M., and Carmignoto, G. (2002). Calcium oscillations encoding neuron‐to‐astrocyte communication. J. Physiol. Paris 96, 193–198. Zonta, M., Angulo, M. C., Gobbo, S., Rosengarten, B., Hossmann, K. A., Pozzan, T., and Carmignoto, G. (2003). Neuron‐to‐astrocyte signaling is central to the dynamic control of brain microcirculation. Nat. Neurosci. 6, 43–50.
This page intentionally left blank
CHAPTER 5 Mechanosensitive Channels in Neurite Outgrowth Mario Pellegrino* and Monica Pellegrini{ *Dipartimento di Fisiologia Umana ‘‘G. Moruzzi,’’ Universita` di Pisa, Pisa, Italy { Scuola Normale Superiore, Pisa, Italy
I. II. III. IV. V. VI. VII.
Overview Introduction Encoding of Guidance Cues in Axon Pathfinding Requirement of TRP Channels in Calcium‐Dependent Axon Pathfinding Physical Guidance Cues and Role of Mechanosensitive Ion Channels Ion Channels as Molecular Integrators Concluding Remarks References
I. OVERVIEW The past few years have seen the convergence of two areas of investigation: the first is the study of the molecular basis for Ca2þ‐dependent axon pathfinding and the second is the molecular and functional characterization of mechanosensitive Ca2þ‐permeant cation channels (MscCa). The convergence of these two fields has reached a pivotal point when some ion channels belonging to the transient receptor potential (TRP) superfamily of proteins, denoted as TRPCs, were reported to play essential roles in the growth cone guidance and, independently, some of these channels were found to form MscCa of vertebrate cells. Various lines of evidence taken together make likely the idea that MscCa can substantially contribute to the spatial and temporal shaping of Ca2þ responses in growing neurites. These findings will be described and the possible contribution of MscCa to the neurite outgrowth will be discussed.
Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59005-2
112
Pellegrino and Pellegrini
II. INTRODUCTION During development, growth cones at the tips of extending neurites guide axons to appropriate target regions, migrating through complex environments. Regeneration and development share basic mechanisms (Letourneau et al., 1991). Although the majority of studies concerning growth cone motility have been performed in vitro, the main findings have been validated on neurons developing in vivo (Gomez and Spitzer, 1999). The guidance of nerve fibers to their final destination can be considered as a series of short‐range projections under the influence of local cues. A variety of simultaneous environmental stimuli is likely to confront the growth cone that must therefore integrate inputs and choose an appropriate final response, consisting in a reorganization of cytoskeleton and adhesion complexes (Lin et al., 1994; Gomez and Zheng, 2006; Wen and Zheng, 2006). Both in vitro (Shaw and Bray, 1977) and in vivo (Harris et al., 1987) experiments demonstrate that individual growth cones are largely independent of cell body in their responses to environmental cues; accordingly, the essential components for extension and guidance are locally regulated (Ming et al., 2002). In the next section, a basic summary of current knowledge regarding the mechanisms of calcium‐dependent axon pathfindings will be provided. Since many excellent reviews are available in the field (Letourneau et al., 1991; Henley and Poo, 2004; Bolsover, 2005; Chilton, 2006; Gomez and Zheng, 2006; Wen and Zheng, 2006), attention will be focused on a few emerging concepts, which are relevant to answer the question whether MscCa may have a role in the control of growth cone dynamics.
III. ENCODING OF GUIDANCE CUES IN AXON PATHFINDING A large variety of guidance cues have been found to be encoded into a limited number of intracellular signals. Among others, intracellular free Ca2þ concentration ð½Ca2þ i Þ and cyclic nucleotides have been extensively studied (Song and Poo, 1999; Ming et al., 2001; Xiang et al., 2002; Nishiyama et al., 2003; Gomez and Zheng, 2006; Wen and Zheng, 2006). DiVerent cues, such as neurotransmitters, growth factors, components of the extracellular matrix (ECM), and cell adhesion molecules (CAMs), through distinct receptor complexes, contribute to change the ½Ca2þ i . Elevations of ½Ca2þ i are recognized as a cellular signaling event aVecting a variety of cellular processes from cytoskeletal remodeling to transcriptional regulation (Berridge et al., 2003). In many neuronal types, there is an optimal range of ½Ca2þ i that supports maximal neurite outgrowth, while large ½Ca2þ i elevations induce growth cones to slow down or retract, and reductions in ½Ca2þ i promote neurite
5. Mechanosensitive Channels in Neurite Outgrowth
113
growth (Bixby and Spitzer, 1984; Kater et al., 1988; Gomez and Spitzer, 2000; Henley and Poo, 2004). Since the discovery that diVerent classes of growth cones, both in vitro and in vivo, share the property of generating periodic elevations of ½Ca2þ i as they migrate (Gomez et al., 1995), growth‐associated calcium transients have been extensively investigated. These appear as a natural signaling mechanism regulating the axon extension because the rate of axon outgrowth has been found inversely proportional to the frequency of calcium transients (Goldberg and Grabham, 1999; Gomez and Spitzer, 1999). Both Ca2þ‐induced Ca2þ release and influx through plasma membrane channels are contributors to the transients. Besides transmitter‐ and voltage‐gated channels, nontraditional calcium channels had been suspected to sustain calcium transients in growth cones. Calcium transients were only partly aVected by organic blockers of voltage‐ dependent Ca2þ channels (VDCCs), while they were inhibited by nonspecific inorganic cations (Gu et al., 1994; Gomez et al., 1995, 2001; Williams and Cohan, 1995). Analysis at higher resolution shows that adhesion generates in neurons short‐lived and highly localized calcium rises (Dunican and Doherty, 2000), and studies elegantly demonstrate that these transients in the growth cone are suYcient to modulate its motility (Zheng, 2000). Furthermore, frequency of filopodial local calcium transients has been found to depend on substrate type and concentration (Gomez et al., 2001). Thus, the spatiotemporal shaping of the ½Ca2þ i responses represents a key determinant of signal specificity. One expects that such a shaping of the calcium signal requires mechanisms which are capable of producing large but local changes of ½Ca2þ i along with a strategic assembly of proteins in signal complexes. A successful paradigm to study in vitro calcium‐dependent axon pathfinding is the analysis of growth cone turning responses to guidance cues. Local Ca2þ signals regulate growth cone steering toward or away from a source of specific neurotransmitters, neurotrophic factors, or components of ECM and cell surface. Thus, well‐defined responses have been established for diVerent classes of neurons. For example, among others, brain‐derived neurotrophic factor (BDNF) and netrin‐1 are chemoattractants for several classes of developing neurons, such as Xenopus spinal neurons and rat cerebellar granule cells (Li et al., 2005; Wang and Poo, 2005), whereas, myelin‐associated glycoprotein (MAG) is a chemorepellent for Xenopus spinal neurons (Song et al., 1998). This experimental approach was also exploited to highlight the unexpected role of intracellular cAMP ([cAMP]i) in determining the direction of growth cone turning to a particular guidance cue. Low levels of [cAMP]i in growth cones lead to the conversion of netrin‐induced attraction into repulsion (Song et al., 1997). Interestingly, both laminin‐1 (Hopker et al., 1999) and N‐cadherin or L1 (Ooashi et al., 2005), via cAMP, regulate the steering of growth cones.
114
Pellegrino and Pellegrini
Therefore, both ECM molecules and CAMs modify the growth cone response to guidance cues. Other cyclic nucleotides, such as cGMP, can play such a role and also the reverse shift, that is repulsive factors converted to attraction, can occur (Song et al., 1998). Thus, a general implication here is that a given guidance cue may act to either repel or attract growth cones, depending on the cellular context. The discovery of spontaneous changes of [cAMP]i in neurons (Bacskai et al., 1993; Hempel et al., 1996) and, particularly, the demonstrated interdependence of ½Ca2þ i and [cAMP]i oscillations (Gorbunova and Spitzer, 2002) open the possibility that cell uses specific dynamics of intracellular second messengers to encode complex environmental stimuli, converging onto a limited number of intracellular signals (Zaccolo et al., 2002). Many potential sites of cross talk between the ½Ca2þ i and the [cAMP]i signaling pathways have been found. Among the main ones, Ca2þ activates some isoforms of adenylyl cyclase and phosphodiesterases, while cAMP, in turn, modulates Ca2þ channels and pumps (Bruce et al., 2003). The mechanisms of growth cone guidance exhibit adaptation. High concentration of a given cue reversibly desensitizes the turning response, not only to the same guidance cue but also to those sharing common cytosolic transduction mechanisms (Ming et al., 2002). Such adaptive behavior enables growth cones to modulate the gain of their guidance signal transduction.
IV. REQUIREMENT OF TRP CHANNELS IN CALCIUM‐DEPENDENT AXON PATHFINDING VDCCs have been found only partly responsible for the netrin‐1‐mediated Ca2þ influx into Xenopus growth cones. Furthermore, the mechanisms of their activation by netrin remained elusive (Hong et al., 2000; Nishiyama et al., 2003). Investigations have provided conclusive evidence that some cationic channels, belonging to the TRP superfamily (for reviews see Clapham et al., 2001; Minke and Cook, 2002; Montell et al., 2002; Clapham, 2003; Moran et al., 2004; Lin and Corey, 2005; Pederson et al., 2005; Owsianik et al., 2006; Ramsey et al., 2006), included in the canonical family (TRPC), are involved in chemotropic axon guidance. Although all TRPCs are tetramers of six transmembrane polypeptide subunits, are nonselective cation channels, and share invariant sequences in both C‐terminal and part of N‐terminal tails, their selectivity ratio varies significantly. Members of diVerent subfamilies of TRPCs are reported to be required for growth cone guidance in various neuronal types. TRPC5 have been found to regulate length and morphology of hippocampal neurites, since neurite growth is inhibited or enhanced by overexpression of TRPC5 or a dominant‐negative (DN) pore mutant, respectively (Greka et al., 2003). Furthermore, in rat
5. Mechanosensitive Channels in Neurite Outgrowth
115
cerebellar granule cells, the block of influx through endogenous TRPC3 by siRNA or by overexpression of DN‐TRPC3 or DN‐TRPC6 destroys growth cone attraction toward BDNF (Li et al., 2005). Parallel studies demonstrated that TRPC1 is essential for netrin‐induced axon turning in Xenopus spinal neurons, because this response is abolished by pharmacological inhibition with SKF‐96365 or La3þ, knockdown of protein by morpholino injection or expression of DN‐TRPC1 (Wang and Poo, 2005). In summary, these results demonstrate that netrin or BDNF stimulates its receptor, DCC or TrkB, respectively, which in turn promotes through activation of a diVerent phospholipase C the production of inositol‐1,4,5‐triphosphate (IP3), a messenger causing the release of Ca2þ from internal stores. While in amphibian spinal neurons, the consequent activation of TRPCs produce membrane depolarization and additional Ca2þ influx through VDCCs, this step does not seem to be involved in the response to BDNF of mammalian neurons in which only TRPCs are activated (Li et al., 2005; Wang and Poo, 2005). However, interesting points emerging from these remarkable findings are: (1) TRPCs are activated by guidance cues such as netrin or BDNF; (2) their local applications produce localized and asymmetrical increases of ½Ca2þ i ; (3) both Ca2þ influx and Ca2þ release are necessary for the turning response; (4) TRPCs, with or without VDCC coactivation, have a clear‐cut role in the amplification of the Ca2þ response. As far as the activation mechanism of TRP is concerned, these channels have often been associated with the store‐operated Ca2þ entry (SOCE; Parekh and Putney, 2005), also reported as capacitative Ca2þ entry because it allows stores to be replenished, but the process responsible for TRP activation is still a matter of intense debate (Clapham, 2003; Parekh and Putney, 2005). The physiological role of TRPC1 channels has been confirmed both in vitro and in vivo, demonstrating that they are essential for the proper formation of commissural axon tracts in Xenopus spinal cord (Shim et al., 2005). DiVerent mechanisms can be considered to account for the ability of cyclic nucleotides to shift the growth cone turning direction in response to a given guidance cue. Cyclic nucleotides might modulate proteins involved in voltage‐ dependent Ca2þ influx, in ATPase activities, as well as in Ca2þ buVering or exchange, but the possibility that TRPs themselves might be potential targets of modulation should be taken into account. Four main findings concerning these channels are relevant in this context. The first is that Xenopus TRPC1, which can be activated by netrin and BDNF, was also found to form MscCa (Maroto et al., 2005), appearing as a molecular integrator of chemical and physical stimuli. Interestingly, other members of TRPs have been found mechanosensitive (MS); TRPA1 has been proposed as a candidate for the transduction channel of vertebrate hair cells
116
Pellegrino and Pellegrini
(Corey et al., 2004, but see Kwan et al., 2006). The second outcome is that TRPC1 and TRPC3 coimmunoprecipitate and colocalize with caveolins, and all members of TRPCs have a conserved motif, adjacent to the first transmembrane domain in the cytosolic N‐terminus, which is similar to the caveolin‐1 binding region. TRPCs have also binding domains for calmodulin, as well as for PLC and scaVolding proteins (Kiselyov et al., 2005), indicating that they can be localized within Ca2þ signaling microdomains (Ambudkar, 2006). The third interesting finding is that regulation of TRPC5 in response to growth factors involves rapid (within few minutes) and reversible insertion of vesicles containing constitutively active channels into the plasma membrane. In principle, this phenomenon enables the membrane to accomplish a tight spatiotemporal control of Ca2þ influx (Bezzerides et al., 2004). The last relevant outcome is the ability of TRPCs to exist as heterotetramers, both when heterologously expressed and in vivo, resulting a wide range of channel subtypes (Hofmann et al., 2002; Strubing et al., 2003).
V. PHYSICAL GUIDANCE CUES AND ROLE OF MECHANOSENSITIVE ION CHANNELS The means by which physical interactions at the cell–substrate interface can guide cell movement has been investigated for many years. Cells show diVerent morphologies and motility rates on substrates of given chemical properties but diVerent rigidities (Pelham and Wang, 1997; Lo, 2006). Moreover, it has been shown that substrate topography alone contains neurite guidance information (Rajnicek et al., 1997). The complexity of the cell responses to mechanical stimuli can be appreciated in some excellent reviews (Hamill and Martinac, 2001; Ingber, 2006), which present mechanotransduction as a cascade of multiscale events in which forces are unevenly distributed in order to force specific target molecules, while protecting most other cellular components. Prestress of membrane domains appears as a key factor that enables rapid responses to mechanical deformation. Environmental mechanical stimuli can directly aVect intracellular targets since communication by means of physical signals seems to be as important as that carried out by chemical messengers, in mechanisms of cellular interaction with the substrate (Wang et al., 1993). On the other hand, mechanical stimuli from the substrate can be transduced via MS proteins (receptors, channels, and enzymes). The adhesive interactions, mediated by cell surface receptors that bind to ligands in the ECM and on other cells, trigger intracellular signaling events that regulate diVerent cellular functions, including cell growth and diVerentiation (Clark and Brugge, 1995; Condic and Letourneau, 1997). There is increasing evidence that signaling via adhesion takes two forms:
5. Mechanosensitive Channels in Neurite Outgrowth
117
inside‐out signaling (regulation of expression, conformation, and aYnity of the receptors) and outside‐in signaling (triggering of intracellular responses by ligand occupation of surface adhesion receptors; Hynes, 1992). Dynamics of integrin‐mediated adhesions depends on local tension. Thus, application of force to focal adhesions strengthen the cytoskeletal anchorage, increasing the activation of integrin signaling (Geiger and Bershadsky, 2002). In keeping with this, a substrate‐dependent calcium dynamics has been reported (Gomez et al., 2001). MS channels appear to be involved in both cell–substrate and cell–cell interactions. On the one hand, MscCa of epithelial keratocytes regulate cell movement, mediating detachment of the cell margin (Lee et al., 1999). On the other hand, mechanical forces applied to adherens junctions of human gingival fibroblasts activate MscCa and increase actin polymerization (Ko et al., 2001). To better understand the mechanisms by which cells transduce changes in membrane tension into diVerent biochemical responses which regulate growth, we should explain how diVerent signals are integrated inside the cell. In particular, growth cone membranes undergo large changes in tension during their dynamics (Lamoureux et al., 1989) and a link between pulling mechanical tension and neurite elongation has been demonstrated in cultured hippocampal neurons (Lamoureux et al., 2002). The conceivable involvement in neurite elongation of MS channels, as tension‐dependent modulators of membrane voltage, has already been put forward (Sigurdson and Morris, 1989). However, the failure to evoke macroscopic Kþ currents, in response to various mechanical stresses applied to the growth cones, raised doubts about the physiological relevance of single‐ channel MS currents (Morris and Horn, 1991). Cytoskeleton may provide a mechanoprotective shock absorber and this can be one of the factors which accounts for the failure to elicit MS currents from cells expressing MS channels (Wan et al., 1999; Ko and McCulloch, 2000). More recently, the hypothesis of a role of MS channels in neurite growth has been newly supported. For example, gentamicin, which blocks single‐channel currents of MscCa in leech neurons, was found to increase their axon outgrowth in culture, as shown in Fig. 1 (Calabrese et al., 1999). In the last few years, the hypothesis was strongly supported by the convergence of two fields of investigation: the study of the molecular basis for calcium‐dependent axon pathfinding and that concerning the molecular identification of MscCa. On the one hand, as reported above, TRPC cation channels have been found to play essential roles in the control of neurite length and growth cone morphology (Greka et al., 2003) as well as in the growth cones guidance (Li et al., 2005; Shim et al., 2005; Wang and Poo, 2005). On the other hand, TRPC was reported to form MscCa of vertebrate cells. In frog oocytes, the protein responsible for MscCa was identified as TRPC1, and liposome
118
Pellegrino and Pellegrini No gentamicin Gentamicin
100
1500
75
1000
50
500
25
0
Density of arborization (%)
Total length of arborization (µm)
2000
0 0
10
20 30 Time (h)
40
FIGURE 1 EVects of 200‐mM gentamicin on the total length and density of arborization of leech AP neurons, at diVerent times after plating. Each point of the plot is expressed as mean values of at least 11 cells SEM. The statistical significance of the diVerences between the means (p < 0.025 with the Mann‐Whitney test) is denoted by an asterisk. Reproduced with permission from Calabrese et al. (1999).
reconstitution showed that this channel can be gated by tension developed purely in the lipid bilayer (Maroto et al., 2005). TRPCs are activated in response to G‐protein–coupled receptor activation and/or depletion of intracellular Ca2þ stores (Minke and Cook, 2002; Montell et al., 2002). Three activation mechanisms have been hypothesized. First, a direct link between IP3R and TRP channel; second, the action of a diVusible second messenger released from the Ca2þ‐depleted endoplasmic reticulum; third, a fusion of vesicles containing constitutively active TRPs, induced by Ca2þ release. Which mechanism is used by which channel subtype is a matter of intense debate (Putney and McKay, 1999; Beech, 2005). Whatever the mechanism involved in the TRPC1 response to G‐protein–coupled receptor activation, Maroto et al. (2005) demonstrated that membrane stretch alone can activate TRPC1. Thus, the multiplicity of activation makes these channels suitable to integrate chemical and physical stimuli, meeting the requirements for a context‐dependent sensor. Other findings are consistent with the idea that also other MS ion channels are involved in neurite growth. It has been demonstrated that NGF‐ TrkA regulates the ENaC expression in PC12 cells and that blocking protein transcription blunts neurite formation (Drummond et al., 2006).
5. Mechanosensitive Channels in Neurite Outgrowth
119
VI. ION CHANNELS AS MOLECULAR INTEGRATORS The previous section pointed out the ability of TRPC1 to integrate diVerent stimuli. The multimodal activation is emerging as a notable common feature of MS, TRP channels. In the nervous system of the leech Hirudo medicinalis, we have identified large conductance cation channels, activated by negative pressure applied to the membrane in inside‐out configuration or by hypotonic swelling in the cell attached (Pellegrino et al., 1990). These channels are expressed by mechanosensory cells as well as by neurons not involved in sensory mechanotransduction. Both selectivity and ½Ca2þ i ‐imaging studies have shown that these channels admit cations and exhibit a substantial Ca2þ permeability (Calabrese et al., 1999; Barsanti et al., 2006b). Their pharmacological features are similar to those of typical MscCa of vertebrate cells, in particular, gentamicin produces a complete voltage‐dependent block (Calabrese et al., 1999). Both cell bodies and growth cones of leech neurons growing in culture express MscCa. Interestingly, two activity modes diVering in kinetics and single‐channel subconductances were identified. The first, denoted as spikelike (SL) mode, was mainly displayed in membrane patches excised from freshly desheathed quiescent cell bodies, while the second, called multiconductance (MC) mode, was commonly found in cultured cell bodies and mainly in growth cones (Pellegrini et al., 2001). As previously reported, we found that addition to the culture medium of gentamicin, a nonspecific blocker of MscCa, which does not aVect voltage‐dependent Ca2þ currents in the leech neurons, increased the neurite extension in culture (Calabrese et al., 1999). MS channels of leech neurons have been further characterized as polymodal cation channels: both SL and MC increase their open probability with depolarization (Menconi et al., 2001) and with intracellular acidosis, while only SL was activated by intracellular calcium in the range 1–10 mM (Barsanti et al., 2006b). MC activity is quickly and robustly increased by intracellular ATP in excised membrane patches, whereas intracellular cAMP is capable to slowly overcome the ATP activation to reach a complete inhibition, as illustrated in Fig. 2 (Barsanti et al., 2006a). Thus, these channels exhibit typical biophysical and pharmacological features of TRPs, with the clear‐cut ability to integrate. In this context, the major new finding is the powerful antagonistic modulation by intracellular ATP and cAMP. The diVerent time dependence of the two regulations might enable these channels to participate in the interplay between intracellular Ca2þ and cAMP. Other ion channels which are MS and display multimodal properties might be involved in cell motility. For example, recombinant N‐type VDCCs expressed in HEK cells have been reported to be MS (Calabrese et al., 2002). Among the Kþ channels expressed in neuronal tissues, it has been also hypothesized that some members (TREK‐1 and TRAAK) of the 2P/4TM
120
Pellegrino and Pellegrini
A
ATP
30
ATP b
25 Im (pA)
B
ATP
20 15 a
10
10 pA 5 0
* 0
200
C 80
400 Time (s)
600
2s
800
ATP 1 mM cAMP 1 mM
Im (pA)
60 40 20 0 0
200
400 600 Time (s)
800
FIGURE 2 The cytoplasmic side of a membrane patch containing four MscCa of leech AP neurons was perfused twice with 1‐mM MgATP. The membrane potential was held at þ80 mV. Columns in the plot (A) represent the mean patch current calculated from 1‐s‐long consecutive data segments. Trace (B) illustrates the transition between a and b to estimate the activation delay. The asterisk (*) marks the opening of the electrovalve in the solution changer. (C) Addition of cAMP during sustained activation by ATP produced a slow, complete, and reversible inhibition of eight MscCa. The membrane potential was held at þ80 mV. The plot shows the mean patch current measured at consecutive intervals of 1 s. Modified from Barsanti et al. (2006a).
structural class of mammalian Kþ channels might be involved in cell growth (Maingret et al., 1999). These channels are MS and exhibit polymodal activation (Patel and Honore´, 2001).
VII. CONCLUDING REMARKS It is clear from this chapter that TRPCs are involved in neurite growth, with distinct mechanisms in diVerent species, playing the general role of Ca2þ signal amplifier. Since the Ca2þ release from intracellular stores is involved
5. Mechanosensitive Channels in Neurite Outgrowth
121
in signal amplification, the question arises whether cells need an additional amplifier in the plasma membrane. It is tempting to speculate that TRPCs because of their capabilities of integration and translocation are suitable for the special function of developing powerful Ca2þ responses in restricted membrane domains of temporary structures such as lamellae and filopodia. Much of the work discussed in this chapter indicates that MS ion channels should be included in the list of molecules that control the ½Ca2þ i homeostasis in the growing neurites. Excessive activation of MscCa is potentially cytotoxic; therefore, it is conceivable that they are controlled by various mechanoprotective mechanisms. These can consist in the cytoskeletal action, both as absorber structure and as regulator of prestress, as well as in metabolic modulation of the channels themselves. In addition, the multimodal activation of these channels can produce local eVects, enabling them to work in a context‐dependent mode, in membrane microdomains where diVerent stimuli can be integrated. Although the field has recently made impressive progress, major questions remain to be answered. The next stages will be to clarify how MscCa contribute to the integration of diVerent signals, such as Ca2þ and cyclic nucleotides, and how distinct patterns of oscillation of these messengers can be decoded by downstream eVector molecules, to determine the growth cone behavior. It will be also important to investigate the relationships between the fusion of vesicles containing TRPCs and integrin dynamics, in order to explore the possible participation of MS ion channels in the reorganization of focal adhesions during cell movement.
NOTE ADDED IN PROOF B. T. Jacques‐Fricke, Y. Seow, P. A. Gottlieb, F. Sachs, and T. Gomez (J. Neurosci. 26, 5656–5664, 2006) showed that inhibition of Ca2þ influx through stretch‐activated channels, with gentamicin or GsMTx4, a peptide isolated from G. spatulata spider venom, enhances the rate of neurite extension of Xenopus neurons. These results are in keeping with those obtained with leech neurons. These authors also found that SKF‐96365, which blocks TRPCs, slows neurite outgrowth in Xenopus. This suggests that diVerent TRP channels can antagonistically modulate neurite growth. Acknowledgments We wish to thank the colleagues of our laboratory who contributed to the experiments on the leech neurons, as Ph.D. students, Dr. Barbara Calabrese, Dr. Maria Carla Menconi, and Dr. Cristina Barsanti, and as a student, Dr. Debora Ricci.
122
Pellegrino and Pellegrini
References Ambudkar, I. S. (2006). Ca2þ signaling microdomains: Platforms for the assembly and regulation of TRPC channels. Trends Pharmacol. Sci. 27, 25–32. Bacskai, B. J., Hochner, B., Mahaut‐Smith, M., Adams, S. R., Kaang, B. K., Kandel, E. R., and Tsien, R. Y. (1993). Spatially resolved dynamics of cAMP and protein kinase A subunits in Aplysia sensory neurons. Science 260, 222–226. Barsanti, C., Pellegrini, M., and Pellegrino, M. (2006a). Regulation of the mechanosensitive cation channels by ATP and cAMP in leech neurons. Biochim. Biophys. Acta 1758, 666–672. Barsanti, C., Pellegrini, M., Ricci, D., and Pellegrino, M. (2006b). EVects of intracellular pH and Ca2þ on the activity of stretch‐sensitive cation channels in leech neurons. Pflugers Arch. 452, 435–443. Beech, D. J. (2005). Emerging functions of 10 types of TRP cationic channel in vascular smooth muscle. Clin. Exp. Pharmacol. Physiol. 32, 597–603. Berridge, M. J., Bootman, M. D., and Roderick, M. L. (2003). Calcium signalling: Dynamics, homeostasis and remodelling. Nat. Rev. Mol. Cell. Biol. 4, 517–529. Bezzerides, V. J., Ramsey, I. S., Kotecha, S., Greka, A., and Clapham, D. E. (2004). Rapid vesicular translocation and insertion of TRP channels. Nat. Cell. Biol. 6, 709–720. Bixby, J. L., and Spitzer, N. C. (1984). Early diVerentiation of vertebrate spinal neurons in absence of voltage‐dependent Ca2þ and Naþ influx. Dev. Biol. 106, 89–96. Bolsover, S. R. (2005). Calcium signalling in growth cone migration. Cell Calcium 37, 395–402. Bruce, J. I. E., Straub, S. V., and Yule, D. I. (2003). Crosstalk between cAMP and Ca2þ signaling in non‐excitable cells. Cell Calcium 34, 431–444. Calabrese, B., Manzi, S., Pellegrini, M., and Pellegrino, M. (1999). Stretch‐activated cation channels of leech neurons: Characterization and role in neurite outgrowth. Eur. J. Neurosci. 11, 2275–2284. Calabrese, B., Tabarean, I. V., Juranka, P., and Morris, C. E. (2002). Mechanosensitivity of N‐type calcium channel currents. Biophys. J. 83, 2560–2574. Chilton, J. K. (2006). Molecular mechanisms of axon guidance. Dev. Biol. 292, 13–24. Clapham, D. E. (2003). TRP channels as cellular sensors. Nature 426, 517–524. Clapham, D. E., Runnels, L. W., and Strubing, C. (2001). The TRP ion channel family. Nat. Rev. Neurosci. 2, 387–396. Clark, E. A., and Brugge, J. S. (1995). Integrins and signal transduction pathways: The road taken. Science 268, 233–239. Condic, M. L., and Letourneau, P. C. (1997). Ligand‐induced changes in integrin expression regulate neuronal adhesion and neurite outgrowth. Nature 389, 852–856. Corey, D. P., Garcia‐Anoveros, J., Holt, J. R., Kwan, K. Y., Lin, S. Y., Vollrath, M. A., Amalfitano, A., Cheung, E. L., Derfler, B. H., Duggan, A., Geleoc, G. S., Gray, P. A., et al. (2004). TRPA1 is a candidate for the mechanosensitive transduction channel of vertebrate hair cells. Nature 432, 723–730. Drummond, H. A., Furtado, M. M., Myers, S., Grifoni, S., Parker, K. A., Hoover, A., and Stec, D. E. (2006). ENaC proteins are required for NGF‐induced neurite growth. Am. J. Physiol. Cell Physiol. 290, C404–C410. Dunican, D. J., and Doherty, P. (2000). The generation of localized calcium rises mediated by cell adhesion molecules and their role in neuronal growth cone motility. Mol. Cell. Biol. Res. Commun. 3, 255–263. Geiger, B., and Bershadsky, A. (2002). Exploring the neighborhood: Adhesion‐coupled cell mechanosensors. Cell 110, 139–142. Goldberg, D. J., and Grabham, P. W. (1999). Braking news: Calcium in the growth cone. Neuron 22, 423–425.
5. Mechanosensitive Channels in Neurite Outgrowth
123
Gomez, T. M., and Spitzer, N. C. (1999). In vivo regulation of axon extension and pathfinding by growth‐cone calcium transients. Nature 397, 350–355. Gomez, T. M., and Spitzer, N. C. (2000). Regulation of growth cone behavior by calcium: New dynamics to earlier perspectives. J. Neurobiol. 44, 174–183. Gomez, T. M., and Zheng, J. Q. (2006). The molecular basis of calcium‐dependent axon pathfinding. Nat. Rev. 7, 115–125. Gomez, T. M., Snow, D. M., and Letourneau, P. C. (1995). Characterization of spontaneous calcium transients in nerve growth cones and their eVect on growth cone migration. Neuron 14, 1233–1246. Gomez, T. M., Robles, E., Poo, M., and Spitzer, N. C. (2001). Filopodial calcium transients promote substrate‐dependent growth cone turning. Science 291, 1983–1987. Gorbunova, Y. V., and Spitzer, N. C. (2002). Dynamic interactions of cyclic AMP transients and spontaneous Ca2þ spikes. Nature 418, 93–96. Greka, A., Navarro, B., Oancea, E., Duggan, A., and Clapham, D. E. (2003). TRPC5 is a regulator of hippocampal neurite length and growth cone morphology. Nat. Neurosci. 6, 837–845. Gu, X., Olson, E. C., and Spitzer, N. C. (1994). Spontaneous neuronal calcium spikes and waves during early diVerentiation. J. Neurosci. 14, 6325–6335. Hamill, O. P., and Martinac, B. (2001). Molecular basis of mechanotransduction in living cells. Physiol. Rev. 81, 685–740. Harris, W. A., Holt, C. E., and BonhoeVer, F. (1987). Retinal axons with and without their somata, growing to and arborizing in the tectum of Xenopus embryos: A time‐lapse video study of single fibres in vivo. Development 101, 123–133. Hempel, C. M., Vincent, P., Adams, S. R., Tsien, R. Y., and Selverston, A. I. (1996). Spatio‐temporal dynamics of cyclic AMP signals in an intact neural circuit. Nature 384, 166–169. Henley, J., and Poo, M. M. (2004). Guiding neuronal growth cones using Ca2þ signals. Trends Cell Biol. 14, 320–330. Hofmann, T., Schaefer, M., Schultz, G., and Gudermann, T. (2002). Subunit composition of mammalian transient receptor potential channels in living cells. Proc. Natl. Acad. Sci. USA 99, 7461–7466. Hong, K., Nishiyama, M., Henley, J., Tessier‐Lavigne, M., and Poo, M. M. (2000). Calcium signalling in the guidance of nerve growth by netrin‐1. Nature 403, 93–98. Hopker, V. H., Shewan, D., Tessier‐Lavigne, M., Poo, M., and Holt, C. (1999). Growth‐cone attraction to netrin‐1 is converted to repulsion by laminin‐1. Nature 401, 69–73. Hynes, R. O. (1992). Integrins: Versatility, modulation, and signaling in cell adhesion. Cell 69, 11–25. Ingber, D. E. (2006). Cellular mechanotransduction: Putting all the pieces together again. FASEB J. 20, 811–827. Kater, S. B., Mattson, M. P., Cohan, C., and Connor, J. (1988). Calcium regulation of the neuronal growth cone. Trends Neurosci. 11, 315–321. Kiselyov, K., Kim, J. Y., Zeng, W., and Muallem, S. (2005). Protein‐protein interaction and functionTRPC channels. Pflugers Arch. 451, 116–124. Ko, K. S., and McCulloch, C. A. (2000). Partners in protection: Interdependence of cytoskeleton and plasma membrane in adaptations to applied forces. J. Membr. Biol. 174, 85–95. Ko, K. S., Arora, P. D., and McCulloch, A. G. (2001). Cadherins mediate intercellular mechanical signaling in fibroblasts by activation of stretch‐sensitive calcium‐permeable channels. J. Biol. Chem. 276, 35967–35977.
124
Pellegrino and Pellegrini
Kwan, K. Y., Allchorne, A. J., Vollrath, M. A., Christensen, A. P., Zhang, D. S., Woolf, C. J., and Corey, D. P. (2006). TRPA1 contributes to cold, mechanical and chemical nociception but is not essential for hair cell transdcution. Neuron 50, 277–289. Lamoureux, P., Buxbaum, R. E., and Heidemann, S. R. (1989). Direct evidence that growth cones pull. Nature 340, 159–162. Lamoureux, P., Ruthel, G., Buxbaum, R. E., and Heidemann, S. R. (2002). Mechanical tension can specify axonal fate in hippocampal neurons. J. Cell Biol. 159, 499–508. Lee, J., Ishihara, A., Oxford, G., Johnson, B., and Jacobson, K. (1999). Regulation of cell movement is mediated by stretch‐activated calcium channels. Nature 400, 382–386. Letourneau, P. C., Kater, S. B., and Macagno, E. R. (1991). ‘‘The Nerve Growth Cone.’’ Raven Press, New York. Li, Y., Jia, Y. C., Cui, K., Li, N., Zheng, Z. Y., Wang, Y. Z., and Yuan, X. R. (2005). Essential role of TRPC channels in the guidance of nerve growth cones by brain‐derived neurotrophic factor. Nature 434, 894–898. Lin, C. H., Thompson, C. A., and Forscher, P. (1994). Cytoskeletal reorganization underlying growth cone motility. Curr. Opin. Neurobiol. 4, 640–647. Lin, S. Y., and Corey, D. P. (2005). TRP channels in mechanosensation. Curr. Opin. Neurobiol. 15, 350–357. Lo, S. H. (2006). Focal adhesions: What’s new inside. Dev. Biol. 294, 280–291. Maingret, F., Fosset, M., Lesage, F., and Lazdunski, M. (1999). TRAAK is a mammalian neuronal mechano‐gated Kþ channel. J. Biol. Chem. 274, 1381–1387. Maroto, R., Raso, A., Wood, T. G., Kurosky, A., Martinac, B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nat. Cell Biol. 7, 179–185. Menconi, M. C., Pellegrini, M., and Pellegrino, M. (2001). Voltage‐induced activation of mechanosensitive cation channels of leech neurons. J. Membr. Biol. 180, 65–72. Ming, G., Henley, J., Tessier Lavigne, M., Song, H., and Poo, M. (2001). Electrical activity modulates growth cone guidance by diVusible factors. Neuron 29, 441–452. Ming, G., Wong, S., Henley, J., Yuan, X., Song, H., Spitzer, N. C., and Poo, M. (2002). Adaptation in the chemotactic guidance of nerve growth cones. Nature 417, 411–418. Minke, B., and Cook, B. (2002). TRP channel proteins and signal transduction. Physiol. Rev. 82, 429–472. Montell, C., Birnbaumer, L., and Flockerzi, V. (2002). The TRP channels, a remarkably functional family. Cell 108, 595–598. Moran, M. M., Xu, H., and Clapham, D. E. (2004). TRP ion channels in the nervous system. Curr. Opin. Neurobiol. 14, 362–369. Morris, C. E., and Horn, R. (1991). Failure to elicit neuronal macroscopic mechanosensitive currents anticipated by single‐channel studies. Science 251, 1246–1249. Nishiyama, M., Hoshino, A., Tsai, L., Henley, J. R., Goshima, Y., Tessier Lavigne, M., Poo, M. M., and Hong, K. (2003). Cyclic AMP/GMP‐dependent modulation of Ca2þ channels sets the polarity of nerve growth‐cone turning. Nature 423, 990–995. Ooashi, N., Futatsugi, A., Yoshihara, F., Mikoshiba, K., and Kamiguchi, H. (2005). Cell adhesion molecules regulate Ca2þ‐mediated steering of growth cones via cyclic AMP and ryanodine receptor type 3. J. Cell Biol. 170, 1159–1167. Owsianik, K., Talavera, K., Voets, T., and Nilius, B. (2006). Permeation and selectivity of TRP channels. Annu. Rev. Physiol. 68, 685–717. Parekh, A. B., and Putney, J. W., Jr. (2005). Store‐operated calcium channels. Physiol. Rev. 85, 757–810. Patel, A., and Honore´, E. (2001). Properties and modulation of mammalian 2P domain Kþ channels. Trends Neurosci. 24, 339–346.
5. Mechanosensitive Channels in Neurite Outgrowth
125
Pederson, S. F., Owsianik, K. G., and Nilius, B. (2005). TRP channels: An overview. Cell Calcium 38, 233–252. Pelham, R. J., Jr., and Wang, Y. (1997). Cell locomotion and focal adhesions are regulated by substrate flexibility. Proc. Natl. Acad. Sci. USA 94, 13661–13665. Pellegrini, M., Menconi, M. C., and Pellegrino, M. (2001). Stretch‐activated cation channels of leech neurons exhibit two activity modes. Eur. J. Neurosci. 13, 503–511. Pellegrino, M., Pellegrini, M., Simoni, A., and Gargini, C. (1990). Stretch‐activated cation channels with large unitary conductance in leech central neurons. Brain Res. 525, 322–326. Putney, J. W., Jr., and McKay, R. R. (1999). Capacitative calcium entry channels. Bioessays 21, 38–46. Rajnicek, A., Britland, S., and McCaig, C. (1997). Contact guidance of CNS neurites on grooved quartz: Influence of groove dimensions, neuronal age and cell type. J. Cell Sci. 110, 2905–2913. Ramsey, I. S., Delling, M., and Clapham, D. E. (2006). An introduction to trp channels. Annu. Rev. Physiol. 68, 619–647. Shaw, G., and Bray, D. (1977). Movement and extension of isolated growth cones. Exp. Cell Res. 104, 55–62. Shim, S., Goh, E. L., Ge, S., Sailor, K., Yuan, J. P., Roderick, H. L., Bootman, M. D., Worley, P. F., Song, H., and Ming, G. (2005). XTRPC1‐dependent chemotropic guidance of neuronal growth cones. Nat. Neurosci. 8, 730–735. Sigurdson, W. J., and Morris, C. E. (1989). Stretch‐activated ion channels in growth cones of snail neurons. J. Neurosci. 9, 2801–2808. Song, H. J., and Poo, M. M. (1999). Signal trasduction underlying growth cone guidance by diVusible factors. Curr. Opin. Neurobiol. 9, 355–363. Song, H. J., Ming, G. L., and Poo, M. M. (1997). cAMP‐induced switching in turning direction of nerve growth cones. Nature 388, 275–279. Song, H., Ming, G., He, Z., Lehmann, M., McKerracher, L., Tessier‐Lavigne, M., and Poo, M. (1998). Conversion of neuronal growth cone responses from repulsion to attraction by cyclic nucleotides. Science 281, 1515–1518. Strubing, C., Krapivinsky, G., Krapivinsky, L., and Clapham, D. E. (2003). Formation of novel TRPC channels by complex subunit interactions in embryonic brain. J. Biol. Chem. 278, 39014–39019. Wan, X., Juranka, P., and Morris, C. E. (1999). Activation of mechanosensitive currents in traumatized membrane. Am. J. Physiol. 276, C318–C327. Wang, N., Butler, J. P., and Ingber, D. E. (1993). Mechanotransduction across the cell surface and through the cytoskeleton. Science 260, 1124–1127. Wang, X. P., and Poo, M. M. (2005). Requirement of TRPC channels in netrin‐1‐induced chemotropic turning of nerve growth cones. Nature 434, 898–904. Wen, Z., and Zheng, J. Q. (2006). Directional guidance of nerve growth cone. Curr. Opin. Neurobiol. 16, 52–58. Williams, D. K., and Cohan, C. S. (1995). Calcium transients in growth cones and axons of cultured Helisoma neurons in response to conditioning factors. J. Neurobiol. 27, 60–75. Xiang, Y., Li, Y., Zhang, Z., Cui, K., Wang, S., Yuan, X., Wu, C., Poo, M., and Duan, D. (2002). Nerve growth cone guidance mediated by G protein‐coupled receptors. Nat. Neurosci. 5, 843–848. Zaccolo, M., Magalhaes, P., and Pozzan, T. (2002). Compartmentalisation of cAMP and Ca2þ signals. Curr. Opin. Cell Biol. 14, 160–166. Zheng, J. Q. (2000). Turning of nerve growth cones induced by localized increases in intracellular calcium ions. Nature 403, 89–93.
This page intentionally left blank
CHAPTER 6 ENaC Proteins in Vascular Smooth Muscle Mechanotransduction Heather A. Drummond Department of Physiology, The Center for Excellence in Cardiovascular–Renal Research, University of Mississippi Medical Center, Jackson, Mississippi 39216
I. Overview II. Introduction III. DEG/ENaC/ASIC Proteins are Members of a Diverse Protein Family Involved in Mechanotransduction A. ENaC Proteins B. Genetic Link to Mechanotransduction C. Mechanotransduction in C. elegans D. ENaC and Mechanotransduction IV. Involvement of ENaC Proteins in Vascular Smooth Muscle Mechanotransduction A. ENaC Proteins in Pressure‐Mediated Myogenic Constriction V. Summary and Future Directions References
I. OVERVIEW Mechanotransduction influences many aspects of biological function. In the cardiovascular system, mechanotransduction has a significant impact on vascular function. Multiple signal transduction pathways participate in the transmission of biomechanical forces into cellular signals, including integrins, adhesion molecules, cytoskeleton, and the activation of membrane‐bound transporters and ion channels. In this chapter, the potential role of a novel family of ion channels with evolutionarily conserved involvement in mechanotransduction, the degenerin/epithelial Naþ channel (DEG/ENaC) family, is
Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59006-4
128
Heather A. Drummond
discussed. Members of the family have been identified in a diverse range of species and tissue types. Most members of the DEG/ENaC family form cation selective ion channels, some of which are believed to form mechanically gated ion channels. While most research has focused on degenerins in sensory neuron mechanotransduction, emerging evidence suggests ENaC proteins may also participate in vascular smooth muscle mechanotransduction. This chapter addresses the potential role and physiological importance of ENaC proteins as mechanosensors in vascular smooth muscle.
II. INTRODUCTION Mechanotransduction is the conversion of a mechanical force (pressure, strain, shear stress) into a biological response (channel activation, gene expression, contraction, and so on). It is a fundamental biological process, occurring in numerous cell types (epithelial, neuronal, and muscle; Sachs, 1988; Morris, 1990; Lingueglia et al., 1994; Ingber, 1997; Tavernarakis and Driscoll, 1997; Hamill and Martinac, 2001; Syntichaki and Tavernarakis, 2004). While mechanosensation influences many biological functions, including, bone growth, neuritogenesis, touch sensation, and hearing, it has significant impact on the cardiovascular system. For example, hemodynamic forces sculpt the developing heart and blood vessels (Riha et al., 2005). Mechanoreceptors in the heart, aortic arch, and carotid sinuses instantaneously regulate arterial blood pressure (Paintal, 1973; Sheperd and Mancia, 1986). Mechanical forces contribute to the regulation of vascular tone, local blood flow, and vascular remodeling (Davis and Hill, 1999; Ingber, 2006). Given the vast influence of mechanical forces on biological functions, it is not surprising that animal cells have evolved a variety of diVerent signaling mechanisms including the transduction of mechanical forces into changes in gene expression and channel activity via the integrins, adhesion molecules, and the cytoskeleton, as well as activation of membrane‐associated ion channels. Members of the transient receptor potential (TRP) and DEG/ ENaC ion channel families have received attention for their potential involvement in mechanotransductory responses. Excellent reviews on the (1) involvement of the integrins and the cytoskeleton in mechanotransduction, (2) functioning of TRP channels as sensors, and (3) role of ENaC and acid‐sensitive ion channel (ASIC) proteins in baroreception can be found elsewhere (Davis et al., 2001; Alenghat and Ingber, 2002; Lin and Corey, 2005; Martinez‐Lemus et al., 2005; see also Chapter 21). The current chapter will focus on the evidence supporting a role for ENaC proteins in vascular smooth muscle cell (VSMC) mechanotransduction.
6. ENaC and Cardiovascular Mechanotransduction
129
III. DEG/ENaC/ASIC PROTEINS ARE MEMBERS OF A DIVERSE PROTEIN FAMILY INVOLVED IN MECHANOTRANSDUCTION A. ENaC Proteins ENaC proteins are members of a large protein family, termed DEG/ ENaC. Members of this family have been identified in the nematode, Caenorhabditis elegans; the fly, Drosophila melanogaster; the snail, Helix aspersa; and mammals. Members share a common structure that consists of short intracellular N‐ and C‐termini and two membrane‐spanning domains separated by a large extracellular domain. Most members of the DEG/ENaC family form homo‐ and/or heteromultimeric cation channels that are selective for Naþ, but some also conduct Ca2þ and other cations (Garcia‐Anoveros and Corey, 1997; Benos and Stanton, 1999; Mano and Driscoll, 1999; Alvarez de la Rosa et al., 2000; Kellenberger and Schild, 2002; Syntichaki and Tavernarakis, 2004). ENaC proteins are commonly found in epithelial cells of the kidney, lung, and colon, where they play a rate‐limiting role in Naþ and water transport (Benos and Stanton, 1999; Mano and Driscoll, 1999; Kellenberger and Schild, 2002). In epithelial tissue, a‐, ‐, and g‐subunits form a heteromultimeric channel inhibitable by the diuretic amiloride. The stoichiometry of the ENaC channel is presumed to be a2 1g1; however, an alternate stoichiometry of a3 3g3 has also been proposed (Cheng et al., 1998; Firsov et al., 1998; Kosari et al., 1998; Snyder et al., 1998; Dijkink et al., 2002). The channel is Naþ selective, nonvoltage‐gated and constitutively active. ENaC channels usually have long open and closed time(s); however, populations of channels with short opening times (50 ms) and long closed states have also been reported (Duchatelle et al., 1992; Palmer and Frindt, 1996; Caldwell et al., 2004). Gain‐of‐function mutations are associated with severe hypertension due to excess salt and water retention (Liddle’s syndrome), while loss‐of‐ function mutations are associated with salt wasting and hypotension (pseudohypoaldosteronism type I, PHA; Lifton, 1995; Luft, 1998, 2001; Pradervand et al., 1999; Oh and Warnock, 2000; Kellenberger and Schild, 2002; Hummler and Vallon, 2005). aENaC, ENaC, and gENaC are presumed to be the main players, but at least two other novel ENaC subunits have been identified with species‐specific expression. Delta (d)ENaC is expressed in the pancreas, testes, ovaries, and brain in human tissues (Waldmann et al., 1995) and can substitute for aENaC and associate with ENaC and gENaC to form a channel. Similarly, in Xenopus, epsilon (e)ENaC can replace aENaC and form a channel with ENaC and gENaC (Babini et al., 2003). It remains to be determined, if these
130
Heather A. Drummond
subunits are also expressed in rodents since the rat and mouse homologues of those ENaC subunits have not been fully cloned.
B. Genetic Link to Mechanotransduction ENaC proteins are members of a larger family of proteins with a strong genetic link to mechanotransduction. The importance of DEG/ENaC proteins in mechanotransduction originated from work in the nematode. Specific mutations in the C. elegans degenerins, which are expressed in neurons, hypodermis, and muscle, produce animals with an abnormal response to light touch or uncoordinated locomotion (Chalfie and Wolinsky, 1990; Driscoll and Chalfie, 1991; Gu et al., 1996; Tavernarakis et al., 1997; Mano and Driscoll, 1999; Syntichaki and Tavernarakis, 2004). The similarity between the C. elegans proteins and mammalian family members lead investigators to hypothesize that mammalian DEG/ENaC proteins may also be involved in mechanotransduction. Subsequently, research has provided genetic evidence that another group of DEG/ENaC proteins found in neural tissue and sensory epithelia, termed the ASIC, are required for the normal mechanotransduction in specific populations of sensory neurons (Ugawa et al., 1998; Mano and Driscoll, 1999; Price et al., 2000; Price et al., 2001; Waldmann, 2001). Studies in ASIC null mice suggest some of the ASIC proteins are required for normal mechanotransduction in arterial baroreceptor neurons, touch receptors, and visceral mechanoreceptor (Price et al., 2000, 2001; see Chapter 21). Thus, genetic evidence demonstrates C. elegans degenerins and mammalian ASIC proteins are required for normal mechanosensory responses. Expression of ENaC proteins was originally considered to be limited to epithelial tissue. However, numerous studies show that ENaC expression can also be found at several important mammalian sites of mechanotransduction. In neural tissue, ENaC transcripts and proteins are found in certain hypothalamic nuclei, nodose, dorsal root, and trigeminal sensory ganglia (Drummond et al., 1998; Fricke et al., 2000; Chapleau et al., 2001; Amin et al., 2005; Yamamoto and Taniguchi, 2006). Further, ENaC proteins are expressed at the site of mechanotransduction in a wide variety of peripheral mechanoreceptor nerve endings. ENaC expression has also been detected in cells that are typically exposed to mechanical forces such as the placental trophoblasts, uroepithelia, osteoblasts, keratinocytes, and VSMCs (Kizer et al., 1997; Drummond et al., 1998; Kopp et al., 1998; McCarter et al., 1999; Garcia‐Anoveros et al., 2001; Mauro et al., 2002; Drummond et al., 2004; Jernigan and Drummond, 2006a). Genetic evidence demonstrating a link between ENaC proteins and mechanotransduction is not available as ENaC null mice die shortly following birth (Hummler and Rossier, 1996; Koyama et al., 1999; Bonny and Hummler, 2000; Hummler and Beermann, 2000; Snitsarev et al., 2002).
6. ENaC and Cardiovascular Mechanotransduction
131
In Drosophila, degenerins are expressed in diverse tissues. Pickpocket is expressed in multidendritic neurons, a subset of neurons similar to peripheral sensory neurons that play a role in touch sensation and proprioception (Adams et al., 1998). Disruption of this degenerin alters rhythmic locomotion in Drosophila larvae (Ainsley et al., 2003). Additional degenerin proteins contribute to salt taste and are expressed in trachea and contribute to salt and water transport, much like ENaC proteins in airway epithelia (Liu et al., 2003a,b). This suggests that Drosophila degenerin proteins may function as mechanosensors in neurons and Naþ transporters in epithelia. C. Mechanotransduction in C. elegans C. elegans geneticists have identified degenerin protein expression in diverse range of cell types (sensory neurons, motorneurons, interneurons, hypodermis, muscle) and involvement in diverse functions including neurodegeneration, proprioception, and control of locomotion and touch sensation. The molecular basis of the latter response, touch sensation, has been extensively characterized in the nematode (Sulston et al., 1975; Chalfie and Sulston, 1981; Chalfie et al., 1986; Chalfie and Au, 1989; Huang and Chalfie, 1994; Gu et al., 1996; Du and Chalfie, 2001). Most of our understanding of how mammalian DEG/ENaC proteins may participate in mechanotransduction is based on touch sensation in the nematode. A model of the mechanotransducing complex responsible for touch responses in the nematode is discussed in the following section. For an in‐depth review of degenerin channel structure, molecular attributes, and role in mechanotransduction, the reader is referred to a comprehensive review (Syntichaki and Tavernarakis, 2004). 1. Model of C. elegans Mechanotransducer A tethered model of the mechanotransducer in C. elegans has been proposed (Driscoll and Tavernarakis, 1997; Tavernarakis and Driscoll, 1997, 2001; Gillespie and Walker, 2001; Ernstrom and Chalfie, 2002; Syntichaki and Tavernarakis, 2004). In this model, the channel is fixed in the membrane and tethered intracellularly to the cytoskeleton and extracellularly to the extracellular matrix. Mechanical force is transmitted to the channel through the extracellular matrix and cytoskeleton to gate the channel. On the basis of extensive studies, a model of a C. elegans mechanosensor was developed. A cartoon of this model is shown in Fig. 1. The mechanosensor is composed of core and accessory components. a. The Mechanosensor Core. The channel pore and critical associated proteins form the core of the mechanosensor (Driscoll and Tavernarakis, 1997; Tavernarakis and Driscoll, 1997, 2001; Ernstrom and Chalfie, 2002; Syntichaki and Tavernarakis, 2004). The channel pore is formed by MEC‐4
132
Heather A. Drummond Extracellular matrix MEC-5 (collagen)
MEC-9
MEC-4/MEC-10
Stomatin-like (MEC-2)
Microtubules (MEC-7/MEC-12) FIGURE 1 Model of the mechanotransduction complex in C. elegans. The proposed model of a mechanotransducer in C. elegans consists of a channel pore formed by degenerins MEC‐4 and MEC‐10. The stomatin‐like protein MEC‐2 links the channel pore is linked to the microtubules (MEC‐7/MEC‐12). The extracellular MEC‐9 protein links the channel pore to collagen (MEC‐5) in the extracellular matrix.
and MEC‐10, members of the degenerin family and closely related to ENaC proteins (i.e., they have a similar structure of intracellular N‐ and C‐termini, two membrane‐spanning domains, and a large extracellular domain). Evidence suggests MEC‐4 and MEC‐10 can form a pore that conducts Naþ (Goodman et al., 2002). In addition to the pore‐forming subunits, MEC‐2 and MEC‐6 are two intracellular proteins that contribute to the core of the mechanosensor. MEC‐2 is a stomatin‐like protein that is proposed to link the pore of the mechanotransducer to the membrane and cytoskeleton. Stomatin, also called Band 7, is a protein expressed in a diverse range of cell types in mammals. Along with other stomatin proteins, MEC‐2 may stabilize the channel in the membrane. MEC‐6 encodes a protein with similarity to paraoxonases and physically interacts with MEC‐4 and MEC‐10. MEC‐6 is proposed to participate in channel assembly or stabilization. Although the precise role of MEC‐6 is unknown, it is proposed to play a role in channel assembly and/or localization. While MEC‐2 and MEC‐6 do not form the pore
6. ENaC and Cardiovascular Mechanotransduction
133
of the mechanosensor, they are critical to modulating gating properties (O’Hagan et al., 2005). b. Associated Proteins. Additional intracellular and extracellular proteins associate with the mechanosensor core to tether the channel with the cytoskeleton and extracellular matrix (Driscoll and Tavernarakis, 1997; Tavernarakis and Driscoll, 1997, 2001; Ernstrom and Chalfie, 2002; Syntichaki and Tavernarakis, 2004). Intracellular associated proteins include MEC‐7 and MEC‐12, which encode for ‐ and a‐tubulin proteins, respectively. The tubulin proteins form microtubules and serve as an anchor/link for the mechanosensor channel to the cytoskeleton. MEC‐2 may link the microtubules (MEC‐7/MEC‐12) to the pore (MEC‐4/MEC‐10). At least three diVerent extracellular proteins (MEC‐1, MEC‐5, and MEC‐9) are associated with the mechanotransducer. MEC‐1 contains protein interaction domains (EGF‐like domains and Kunitz repeats). A mammalian homologue of this protein has not been identified. MEC‐5 is a collagen protein. A direct interaction between MEC‐5 and degenerin proteins has not been proven; however, a genetic interaction has been shown. MEC‐9 is a component of the extracellular matrix. It contains Kunitz type serine protease inhibitor domains. While the role of MEC‐9 is undetermined, it is required for touch responses. 2. C. elegans UNC‐105: A Muscle Mechanotransducer While degenerin expression and function have been mostly studied in neuronal tissue, it is clear that degenerin expression is not limited to neuronal tissue. At least one degenerin protein is predominantly expressed in muscle tissue. UNC‐105 is thought to be part of a mechanosensitive ion channel important in the control of locomotion (Liu et al., 1996; ShreZer and Wolinsky, 1997; Garcia‐Anoveros et al., 1998). UNC‐105 interacts with an extracellular collagen (LET‐2), presumably to gate the channel (Liu et al., 1996). UNC‐105 is another example where a pore‐forming degenerin (MEC‐4/ MEC‐10 or UNC‐105) can interact with an extracellular collagen (MEC‐5 or LET‐2) presumably to gate the mechanosensitive channel in response to mechanical stimuli. Although the role of degenerin proteins as mechanotransducers in muscle has received less attention, it is clear that degenerins are expressed in, and required for, stretch sensitivity in muscle tissue.
D. ENaC and Mechanotransduction In addition to being related to proteins involved in mechanosensory responses, there is direct evidence that ENaC channels can be gated by mechanical factors such as pressure and shear stress. Further, a growing body of
134
Heather A. Drummond
functional evidence demonstrates ENaC proteins play an important role in mechanotransduction. 1. Mechanically Gated ENaC Activity a. Stretch Activation of ENaC. Early investigations into the direct mechanosensitivity of ENaC in heterologous expression systems have yielded equivocal results. When expressed in a lipid bilayer, aENaC and a gENaC channels are activated by the application of negative hydrostatic pressure (Awayda et al., 1995; Ismailov et al., 1996a,b, 1997a,b). Similarly, expression of aENaC in a fibroblast cell line conferred the presence of stretch‐activated cation channels (Kizer et al., 1997) using an electrophysiological approach. However, contrasting results were found in the Xenopus oocyte expression system in response to osmotic‐induced swelling and shrinking. Hypoosmotic‐ induced swelling either had no eVect on or inhibited a gENaC current (Awayda and Subramanyam, 1998; Ji et al., 1998) and hyperosmotic‐ induced shrinking inhibited a gENaC current. Thus, diVerences in mechanosensitivity of ENaC in heterologous systems may be due to the expression system itself (bilayer or fibroblast vs oocyte) or may reflect the ability of pressure‐ vs osmotic‐induced stretch to generate membrane tension and activate ENaC channels. Mechanical gating of ENaC channels has also been demonstrated in endogenously expressing tissue. Palmer and Frindt (1996) found that the application of negative pressure to isolated channels in cortical‐collecting duct cells increased the open probability of native ENaC channels in 27% of patches. Subsequent work from Ma et al. (2002) suggests this low probability is likely due to inhibition by purinergic receptors. B lymphocytes also express a mechanically gated, amiloride‐sensitive current (Achard et al., 1996; Ma et al., 2004). A study suggests B lymphocytes express aENaC and ENaC, but not gENaC; however, only aENaC appears to contribute to mechanically gated ENaC activity in these cells (Ma et al., 2004). b. Shear Stress Activation of ENaC. While studies evaluating activation of ENaC currents by pressure‐ vs hypoosmotic‐induced stretch have provided equivocal results, experiments evaluating activation of ENaC to a diVerent mechanical stimulus, shear stress, have provided more consistent findings (Satlin et al., 2001; Carattino et al., 2004, 2005; Morimoto et al., 2006). In isolated rabbit cortical‐collecting ducts, Naþ reabsorption is dependent on tubular flow rate; increases in flow rate increase Naþ reabsorption (Satlin et al., 2001). In subsequent studies, Carattino et al. (2004) demonstrated shear stress activation of Naþ current in oocytes expressing
6. ENaC and Cardiovascular Mechanotransduction
135
a gENaC, a finding that provides direct evidence of the mechanosensitivity of ENaC. Further, residues within the pore participate in the transduction of shear stress (Carattino et al., 2005). Shear stress likely activates a gENaC by increasing open probability because ENaC channels with a high intrinsic open probability were not activated by shear stress (Palmer and Frindt, 1996; Carattino et al., 2004, 2005). The contrasting eVects of shear stress and osmotic‐induced swelling on ENaC activation suggest mechanical gating of a gENaC may be stimulus specific (i.e., shear stress vs osmotic stretch). The reasons underlying modality‐specific activation of a gENaC in oocytes are unknown, but could reflect the importance of the appropriate complement of intracellular and extracellular proteins necessary to gate the channel in response to stretch or strain. Alternatively, modality‐ specific activation of ENaC channels may be conferred by the specific ENaC subunits forming the channel. It is unknown if shear stress sensitivity is due to the presence of aENaC. Whether a channel formed by gENaC can be activated by shear stress, or another mechanical stimulus such as strain, has never been addressed. Constitutive activity is a common feature reported for ENaC channels. However, electrically silent channels may also be expressed at the membrane (Caldwell et al., 2004; Morimoto et al., 2006). Despite the decreased constitutive activity of this pool of channels, it is likely that these channels can be gated mechanically. Emerging evidence suggests proteolytic cleavage of ENaC subunits contributes to baseline activity of ENaC channels (Lewis and Alles, 1986; Palmer and Frindt, 1986; Vallet et al., 1997; Jovov et al., 2001; Caldwell et al., 2004; Hughey et al., 2004a; Olivieri et al., 2005; Carattino et al., 2006). Satlin et al. (2001) have proposed that uncleaved channels represent a pool of ENaC channels that can be put into action (Hughey et al., 2004b). Interestingly, even though uncleaved, or protease resistant, channels exhibit reduced activity under basal conditions, they can still be gated by mechanical stimulation with shear stress (Morimoto et al., 2006). This is a significant finding because it suggests that an ENaC channel that has little or no constitutive activity can be mechanically gated. 2. ENaC Proteins Are Expressed in Mechanosensitive Tissue and Activity Is Required for Mechanosensory Responses If ENaC proteins are to be considered as mechanosensors, then at least two criteria must be met. First, ENaC proteins must be expressed at the site of mechanotransduction. Second, inhibition or disruption of ENaC activity should inhibit the mechanosensitive response. Since ENaC null mice die shortly after birth (Hummler and Rossier, 1996; McDonald et al., 1999; Bonny and Hummler, 2000; Hummler and Beermann, 2000; Snitsarev et al., 2002),
136
Heather A. Drummond
genetic evidence for ENaC involvement in mechanotransduction is lacking. As an alternative, selective ENaC inhibitors, such as amiloride and benzamil, are useful tools to determine ENaC involvement. Although most degenerin channels are sensitive to amiloride, a gENaC is the most sensitive, blocked by as little as 100 nM (Kellenberger and Schild, 2002). Other mammalian degenerins such as the ASIC channels require up to 100 M (Kellenberger and Schild, 2002). Amiloride has been used as a probe for mechanosensitive channel in Xenopus oocytes and hair cells (Hamill et al., 1992; Rusch et al., 1994). However, the doses used to inhibit these channels were significantly greater than the Ki for a gENaC channels. The use of pharmacological inhibitors is limited as they cannot discern among the importance of the individual ENaC subunits in a given mechanosensory response. Thus, the development of animal models with tissue selective knockdown of ENaC proteins will be necessary to determine the importance of specific ENaC subunits in mechanotransduction. ENaC proteins are expressed in populations of somatic and visceral sensory neurons in the dorsal root, trigeminal, and nodose ganglia. In mammals, ENaC proteins have been identified in the aVerent nerve terminals innervating subsets of mechanoreceptors such as arterial baroreceptors, whiskers, larynx, tooth pulp, and touch receptors of hairless skin (Pacinian corpuscles, Merkel cells, Meissner corpuscles; Drummond et al., 2000; Fricke et al., 2000; Ichikawa et al., 2005; Yamamoto and Taniguchi, 2006). The specific ENaC proteins that are expressed in sensory neurons and their nerve endings vary, but include a gENaC and gENaC. While a few studies have demonstrated that activation of mechanically gated currents or ion transients in sensory neurons are inhibited by amiloride, or its analogue benzamil, genetic evidence for a role of any ENaC protein in peripheral sensory neuron mechanotransduction is lacking (McCarter et al., 1999; Carr et al., 2001; Drummond et al., 2001; Snitsarev et al., 2002). To date, expression of ENaC proteins has been reported in a variety of cells other than epithelial or neuronal and includes placental trophoblasts, chondrocytes, osteoblasts, endothelial cells, epidermal cells, and VSMCs (Kizer et al., 1997; Brouard et al., 1999; Trujillo et al., 1999; Golestaneh et al., 2001; Mauro et al., 2002; Driver et al., 2003; Page et al., 2003; Shakibaei and Mobasheri, 2003; Drummond et al., 2004; Jernigan and Drummond, 2006a). Mechanical factors influence responses in many of these cell types, in particular cardiovascular cells. Endothelial cells and VSMCs are continually exposed to mechanical forces such as shear stress, pressure, and strain and, as will be discussed in the following sections, recent investigations show ENaC proteins are expressed in VSMCs and are required for responses dependent on mechanical signaling.
6. ENaC and Cardiovascular Mechanotransduction
137
IV. INVOLVEMENT OF ENaC PROTEINS IN VASCULAR SMOOTH MUSCLE MECHANOTRANSDUCTION The VSMC is an excellent model to study ENaC proteins as mechanosensors for three reasons. First, VSMCs are known to express mechanosensitive ion channels of a largely unknown molecular identity. Second, quantitative assays for VSMC responses that are dependent on mechanical signaling are available. Pressure‐mediated vasoconstriction (also referred to as myogenic constriction) is one widely used assay. Third, the mechano‐dependent response, myogenic constriction, has physiological and pathophysiological significance. The evidence supporting a role for ENaC proteins in VSMCs and the physiological significance are discussed below. A. ENaC Proteins in Pressure‐Mediated Myogenic Constriction 1. What is Myogenic Constriction? Myogenic constriction is an inherent characteristic of resistance vessels that is characterized by a decrease in luminal diameter in response to an increase in transmural pressure. The response is important in establishing basal vascular tone and autoregulation of blood flow. While myogenic constriction occurs in many vascular beds, it is an important regulatory mechanism for blood flow autoregulation in cerebral, mesenteric, and renal beds (Davis and Hill, 1999). Our understanding of the response is mechanical stimulation, produced by pressure‐induced vessel wall strain, initiates a signaling pathway that leads to membrane depolarization and subsequent calcium influx via voltage‐gated calcium channels, which in turns causes VSMC contraction (Harder, 1984; Meininger and Davis, 1992; Knot and Nelson, 1995; Davis and Hill, 1999; Hill et al., 2006). The mechanism(s) transmitting vascular smooth muscle stretch into a cellular response is unknown, but several possible mechanisms have been postulated including extracellular matrix– integrin interactions, membrane‐bound enzyme and second messenger systems, ion transporters and exchangers, or direct activation of mechano‐sensitive ion channels on the vascular smooth muscle plasmalemmal membrane (Davis and Hill, 1999). Of these potential mediators, mechanosensitive ion channels have received the most attention. Recordings from dissociated VSM cells suggest mechanosensitive ion channels tend to be cation selective, with Ca2þ and Naþ as the principal conductors (Kirber et al., 1988; Davis et al., 1992; Wellner and Isenberg, 1993; Ohya et al., 1998; Ernstrom and Chalfie, 2002). Stretch‐ mediated Naþ and/or Ca2þ entry depolarizes the membrane and leads to activation of voltage‐gated cation channels. Activation of voltage‐gated cation
138
Heather A. Drummond
channels produces a larger Ca2þ influx, which, in turn, stimulates the release of Ca2þ from intracellular stores. This large increase in cytosolic Ca2þ drives VSM cell contraction (Brayden and Nelson, 1992; Nelson et al., 1997; Wu and Davis, 2001). TRP and ENaC channels are candidates for these channels. 2. Quantifying Myogenic Constriction The myogenic response can be quantified using in isolated arteries (Fig. 2). In this approach, artery segments are dissected from surrounding tissue and mounted on glass pipettes (Fig. 2A). Vessels are filled and bathed with a Ca2þ containing physiologic salt solution. To determine pressure‐induced constrictor responses, pressure in the vessel is raised in 25 mmHg increments and equilibrated for 5 min (Fig. 2B). At the end of each pressure step equilibration, an image of the vessel is collected for determination of internal diameter under ‘‘active’’ conditions. As shown in Fig. 2C, despite increasing pressure, arteries maintain or decrease diameter. This sequence is repeated in the absence of Ca2þ in the bathing solution. In the absence of external Ca2þ, arteries are unable to constrict and passively dilate (Fig. 2C). Myogenic tone at each pressure step is calculated as the myogenic tone (%) ¼ (passive diameter active diameter/ passive diameter) 100 and plotted, and a pressure–diameter curve constructed (Fig. 2D). 3. Importance of ENaC Proteins in Pressure‐Mediated Vasoconstriction a. ENaC Inhibition Abolishes Pressure‐Mediated Vasoconstriction. The importance of ENaC proteins in pressure‐induced vasoconstriction has been examined in two diVerent preparations, rat middle cerebral and mouse renal interlobar arteries. In both arteries, pharmacological inhibition of ENaC with amiloride or benzamil inhibits pressure‐mediated constriction (Oyabe et al., 2000; Drummond et al., 2004; Jernigan and Drummond, 2006a). In rat middle cerebral and mouse renal interlobar arteries, myogenic constriction is inhibited with submicromolar and low micromolar doses of benzamil (30 nM to 1 M) and amiloride (1–5 M; Drummond et al., 2004; Jernigan and Drummond, 2006a). Approximately 40% of myogenic tone is blocked with 1‐M amiloride (Fig. 2C and D, representative of the response). At submicromolar and low micromolar doses, amiloride and benzamil are selective ENaC inhibitors. Although higher doses of amiloride and benzamil can inhibit other ion transporters and channels such as the Naþ/Ca2þ exchanger, Naþ/Hþ exchanger, Naþ and Ca2þ channels, and other degenerin channels, the doses used to evaluate the role of ENaC proteins in pressure‐mediated constriction are selective for ENaC (100 nM to 5 M; Kleyman and Cragoe, 1988; Kellenberger and Schild, 2002). To address the concern that ENaC inhibition
139
6. ENaC and Cardiovascular Mechanotransduction A
Isolated renal interlobar artery inner diameter ~90 mm
Pressure (mmHg)
B
150 125 100 75 50 25 0 0
5
10
15 20 Time (min)
C
25
D
(Ca2+ plus) “active”
Pressure
“Active” tone
Myogenic tone (%)
Diameter
(Ca2+ free) “passive” ENaC blockade
30
Tone after ENaC blockade Pressure
FIGURE 2 Assessment of myogenic constriction. (A) Artery segments, 90–100 m in diameter, dissected from surrounding tissue are mounted on two glass pipettes. (B) Arteries are exposed to stepwise (25 mmHg, 5 min) increase in luminal pressure. Artery diameter is determined at the end of each 5 min equilibration. (C) Cartoon of the myogenic response. The myogenic response is represented as the change in vessel diameter in response to changes in pressure. An artery with an active myogenic response maintains or decreases diameter when pressure increases (solid line). The passive response of an artery is determined by repeating the pressure steps using a Ca2þ‐free bathing solution (dashed line). Following ENaC blockade with amiloride (5 M) or benzamil (1 M) in a Ca2þ‐containing solution, the myogenic response is abolished. (D) Myogenic tone (passive diameter–active diameter/passive diameter) develops in an untreated artery. Following ENaC blockade, artery segments develop very little myogenic tone.
140
Heather A. Drummond
could be blocking the ability of the artery to constrict, vasoconstrictor responses to the a agonist phenylephrine were examined and found to be unchanged (Jernigan and Drummond, 2006a). These findings suggest ENaC inhibition (1–5 M amiloride; 100 nM, 1 M benzamil) specifically blocks pressure‐ mediated, but not agonist‐induced vasoconstriction. Similar to epithelial cells and heterologous expression systems, ENaC channels formed in VSMCs likely conduct Naþ ions. In mouse renal interlobar arteries, step increases in pressure activate Naþ transients that are abolished following ENaC inhibition with amiloride or benzamil (Jernigan and Drummond, 2006a). Although pressure also activates Ca2þ transients, the importance of Ca2þ moving through the pore vs release of intracellular Ca2þ was not determined. b. ENaC Subunit Expression in VSMCs. In VSMCs freshly dissociated from mouse kidney and rat brain arteries, ENaC and gENaC, but not aENaC, expression is detected by RT‐PCR and immunolabeling (Drummond et al., 2004). Immunolocalization studies reveal ENaC and gENaC are found together and are concentrated at or near the membrane and frequently found localized in puncta. Since a similar punctate‐staining pattern of MEC‐4 along touch neuron processes has been suggested to reflect the presence of mechano‐ transducing ion channel complexes near the membrane, by analogy, localization of ENaC and gENaC in punctae along sarcolemma may reflect the distribution of mechanotransducing ion channel complexes along the smooth muscle membrane (Syntichaki and Tavernarakis, 2004). c. Electrophysiological Evidence. Direct evidence of the ENaC channel in VSMCs is not available. However, Van Renterghem and Lazdunski (1991) reported an epithelial‐like Naþ current in VSMCs. Similar to a gENaC, the channel reported in VSMCs is nonvoltage‐gated and has a 10‐pS conductance and high Naþ:Kþ selectivity. Unlike a gENaC, the channel is insensitive to amiloride (100 M). While the amiloride characteristics of this channel are not consistent with the reported amiloride sensitivity of a gENaC and gENaC channels in heterologous expression systems, this finding supports the potential presence of an ENaC‐like Naþ channel in VSMCs. d. Can bENaC and gENaC Form a Channel in the Absence of aENaC? Evidence from Rossier’s laboratory suggests aENaC is not required for ENaC and gENaC to form a channel; however, aENaC is required to form a fully functional channel (Bonny et al., 1999). Bonny et al. (1999) demonstrated that oocytes expressing ENaC and gENaC generate amiloride sensitive currents in the absence of aENaC, when provided a longer incubation period
6. ENaC and Cardiovascular Mechanotransduction
141
(6 days). Channels formed by ENaC and gENaC have a greater selectivity for Naþ and significantly less current. Additionally, gENaC channels have a tenfold higher Ki for amiloride (2 M in gENaC vs 0.2M in a gENaC). Thus, channels formed by gENaC are not the same as channels formed by a gENaC. The finding by Jernigan and Drummond that 40% of myogenic constrictor responses are blocked with 1‐M amiloride is consistent with the amiloride Ki for gENaC channels. Compared to a gENaC channels, traYcking of gENaC channels to the surface membrane in Xenopus oocytes is delayed and results in protein localization in the intracellular compartment (Bonny et al., 1999). The delayed traYcking of gENaC channels may be the basis for lack of current generated by gENaC in heterologous expression systems. In freshly dissociated VSMCs, traYcking of ENaC and gENaC does not appear to be impaired as they are expressed at or near the cell surface (Fig. 3; Drummond et al., 2004; Jernigan and Drummond, 2006a). These findings suggest, ENaC and gENaC are traYcked to or near the surface in the absence of aENaC. The mechanism(s) mediating membrane localization of ENaC and gENaC in the absence of aENaC is unknown; however, there are a few possible explanations. First, VSMCs may express another protein that associates with and stabilizes ENaC and gENaC, perhaps a protein similar to the C. elegans degenerin, MEC‐6 or MEC‐2. Second, another pore‐forming subunit may interact with ENaC and gENaC, such as dENaC, or an ASIC protein. Third, aENaC may be expressed in VSMCs, but we are unable to detect it and the small amount expressed is suYcient to stabilize the channel. Lastly, the presence of proteins within the dense extracellular matrix of blood vessels may help stabilize gENaC channels that reach the membrane. Regardless of the mechanism, in the absence of detectable levels of aENaC, ENaC and gENaC appear to traYc to the cell surface of VSMC in vivo and shortly following enzymatic dissociation. It is likely that ENaC proteins expressed in VSMCs from renal and cerebral circulations are not proteolytically cleaved. On the basis of Hughey et al. (2004b), demonstrating proteolytic cleavage of one subunit requires coexpression of all three subunits (aENaC, ENaC, and gENaC), it is likely that the absence aENaC in VSMCs prevents proteolytic cleavage of ENaC and gENaC. If gENaC channels were not cleaved, they would be expected to be electrically silent (Hughey et al., 2004b). Since electrically silent a gENaC channels can be mechanically stimulated, it is probable to speculate that electrically silent cell surface gENaC channels may also be activated by mechanical stimulation. However, this has not been directly evaluated. On the basis of these findings, we speculate that ENaC and gENaC subunits are the predominant subunits forming ENaC channels in VSMC. Although the channels to traYc to the cell surface, they are probably electrically silent,
142 A
Heather A. Drummond b ENaC
80 mm
B
a -actin
g ENaC
Merged
8 mm
C
Green/b ENaC
Red/g ENaC
Blue/a -actin
Merged
8 mm
FIGURE 3 ENaC localization in vascular smooth muscle. (A) ENaC immunolocalization in an isolated rat middle cerebral artery segment. The horizontal banding pattern is characteristic of vascular smooth muscle. (B) High magnification image of gENaC localization with a‐actin in another rat cerebral artery segment. At least two individual cells can be identified in the image. Note the highly punctate gENaC staining, suggestive of localization of mechanosensory complexes in the VSMCs. (C) ENaC localization in an individual VSMC enzymatically dissociated from a rat middle cerebral artery. Image is a single optical section taken through the middle of an isolated VSMC. Note that ENaC and gENaC are colocalized near the membrane.
since the channels are not likely to be proteolytically processed, but likely to be activated by mechanical stimuli. e. Gene Silencing of bENaC and gENaC Inhibits Pressure‐Mediated Vasoconstriction in Mouse Renal Interlobar Arteries. Although amiloride and benzamil are great tools to screen for DEG/ENaC channel involvement, determining which specific DEG/ENaC proteins (i.e., ENaC or gENaC) are
6. ENaC and Cardiovascular Mechanotransduction
143
involved requires selective gene silencing. Using two independent approaches to silence expression of specific ENaC genes, expression of dominant‐negative ENaC isoforms and small interfering RNA (siRNA), Jernigan and Drummond (2006b) demonstrate suppression of ENaC or gENaC inhibits myogenic constriction. These findings suggest that both ENaC and gENaC subunits are required for mechanotransduction. f. How Might ENaCs Transduce Vessel Strain In Vivo? The answer to this question is not known; however, they probably fit a model similar to C. elegans, where ENaC and gENaC form the pore of the mechanotransducing complex. The channel pore is likely anchored to the cytoskeleton and extracellular matrix in a similar manner. A potential intracellular protein associated the mechanosensor is stomatin. Stomatin is related to MEC‐2, an essential component of the C. elegans mechanosensor. Stomatin colocalizes with ENaC and gENaC in sensory neurons of the trigeminal ganglia. Additionally, Price et al. (2004) have shown that stomatin interacts with and helps gate another degenerin family member. However, it is unknown if stomatin, or a related stomatin protein, regulates activity of ENaC channels or is required for mechanotransduction in VSMCs. ENaC channels are known to interact with cytoskeletal proteins including spectrin, anykrin, and actin (Rotin et al., 1994; Jovov et al., 1999; Zuckerman et al., 1999; Berdiev et al., 2001; Copeland et al., 2001; Mazzochi et al., 2006). In isolated expression systems, cytoskeletal proteins can regulate gating properties of ENaC channels (Achard et al., 1996; Jovov et al., 1999). Other investigators have suggested mechanotransduction in VSMCs requires a link between extracellular matrix proteins, integrins, and ion channels (Hill et al., 2006); however, the identities of extracellular proteins that bind to ENaC are unknown. We speculate activation of ENaC proteins, by pressure or strain, leads to an influx of Naþ and perhaps Ca2þ (Fig. 4). The cation influx depolarizes the membrane and activates voltage‐gated Ca2þ channels to initiate the Ca2þ‐signaling cascade and lead to vasoconstriction. In addition to degenerins, evolving evidence suggests members of the TRP channel family may also form mechanosensitive channels in VSMC. TRPC6 and TRPM4 have both been implicated as mechanosensors and mediators of myogenic constriction in VSMCs (Welsh et al., 2002; Earley et al., 2004). In independent investigations using similar preparations, inhibition of ENaC function or TRP channel function produces a near total loss of myogenic function. If ENaC and TRP channels function independently of each other, then why does suppression of one mechanosensitive channel class abolish myogenic control? One might expect suppressing one channel would allow the other to compensate, at least partially. This leaves an alternative explanation that the function of ENaC and TRP channels are somehow linked.
144
Heather A. Drummond Mechanical stimuli (strain)
ENaC activation
Na+ influx
Membrane depolarization
Voltage-gated Ca2+ channel activation
Vasoconstriction FIGURE 4 Proposed role of ENaC in myogenic constriction. Activation of ENaC channels by pressure‐induced strain allows Naþ influx and leads to membrane depolarization. Depolarization‐induced activation of voltage‐gated Ca2þ channels and Ca2þ influx stimulates release of intracellular Ca2þ stores, which leads to vasoconstriction.
Whether the channels are located in the same domain, linked directly, that is interacting via protein–protein interactions, or by a common signaling pathway, or some other way, have never been addressed. g. What Is the Potential Physiological Importance of ENaC in Myogenic Constriction? Myogenic constriction contributes to the regulation of blood flow by allowing resistance arteries to adjust tone to luminal pressure; vessels constrict to increases and dilate to decreases in luminal pressure. In the kidney, myogenic constriction plays a critical role in protecting against hypertension‐induced renal injury (Bidani et al., 1987; Hayashi et al., 1992; Van Dokkum et al., 1999; Wang et al., 2000; Loutzenhiser et al., 2004) by preventing increases in blood pressure from being transmitted to the glomerulus, a primary determinant of renal injury (Bidani et al., 1987; GriYn et al., 2000; Bidani and GriYn, 2002; GriYn and Bidani, 2004). These findings suggest myogenic constriction may help protect the kidney from pressure‐ induced renal injury. It is important to note that the role of ENaC proteins in the myogenic constrictor was evaluated only in the larger middle cerebral and renal interlobar arteries. The role of ENaC proteins in myogenic constriction in the small resistance arterioles, the primary site of local blood flow regulation, was not determined. Thus, the role of ENaC proteins in
6. ENaC and Cardiovascular Mechanotransduction
145
myogenic constriction in small resistance arterioles is unknown and remains an important area of future investigation. A better understanding of the physiological importance of ENaC proteins in vascular function will likely be accomplished through the evaluation of vascular function in tissue‐ specific knockout animal models as well as Liddle’s and PHA type I patient populations.
V. SUMMARY AND FUTURE DIRECTIONS In the kidney, the role of ENaC in the regulation of blood pressure through Naþ and water homeostasis is well established. Numerous investigators suggested an additional role for ENaC proteins in mechanotransduction because they are members of a protein family required for mechanotransduction in C. elegans. Most investigations addressing the mechanosensitivity of ENaC demonstrate the ENaC channels can be gated by mechanical factors. In addition to their potential role in neural regulation of cardiovascular function, recent evidence also suggests ENaC may contribute to cardiovascular homeostasis by functioning as mechanosensors in VSMCs. However, whether ENaC proteins truly form mechanically gated channels in VSMC, how vascular ENaC channels contribute to the blood flow regulation in vivo, and if altered ENaC channel function contributes to cardiovascular disease remain important unanswered questions. References Achard, J. M., Bubien, J. K., Benos, D. J., and Warnock, D. G. (1996). Stretch modulates amiloride sensitivity and cation selectivity of sodium channels in human B lymphocytes. Am. J. Physiol. 270, C224–C234. Adams, C. M., Anderson, M. G., Motto, D. G., Price, M. P., Johnson, W. A., and Welsh, M. J. (1998). Ripped pocket and pickpocket, novel Drosophila DEG/ENaC subunits expressed in early development and in mechanosensory neurons. J. Cell Biol. 140, 143–152. Ainsley, J. A., Pettus, J. M., Bosenko, D., Gerstein, C. E., Zinkevich, N., Anderson, M. G., Adams, C. M., Welsh, M. J., and Johnson, W. A. (2003). Enhanced locomotion caused by loss of the Drosophila DEG/ENaC protein Pickpocket1. Curr. Biol. 13, 1557–1563. Alenghat, F. J., and Ingber, D. E. (2002). Mechanotransduction: All signals point to cytoskeleton, matrix, and integrins. Sci. STKE 119, PE6. Alvarez de la Rosa, D., Canessa, C. M., Fyfe, G. K., and Zhang, P. (2000). Structure and regulation of amiloride‐sensitive sodium channels. Annu. Rev. Physiol. 62, 573–594. Amin, M. S., Wang, H. W., Reza, E., Whitman, S. C., Tuana, B. S., and Leenen, F. H. (2005). Distribution of epithelial sodium channels and mineralocorticoid receptors in cardiovascular regulatory centers in rat brain. Am. J. Physiol. Regul. Integr. Comp. Physiol. 289, R1787–R1797. Awayda, M. S., and Subramanyam, M. (1998). Regulation of the epithelial Naþ channel by membrane tension. J. Gen. Physiol. 112, 97–111.
146
Heather A. Drummond
Awayda, M. S., Ismailov, I. I., Berdiev, B. K., and Benos, D. J. (1995). A cloned renal epithelial Naþ channel protein displays stretch activation in planar lipid bilayers. Am. J. Physiol. 268, C1450–C1459. Babini, E., Geisler, H. S., Siba, M., and Grunder, S. (2003). A new subunit of the epithelial Naþ channel identifies regions involved in Naþ self‐inhibition. J. Biol. Chem. 278, 28418–28426. Benos, D. J., and Stanton, B. A. (1999). Functional domains within the degenerin/epithelial sodium channel (Deg/ENaC) superfamily of ion channels. J. Physiol. 520(Pt. 3), 631–644. Berdiev, B. K., Latorre, R., Benos, D. J., and Ismailov, I. I. (2001). Actin modifies Ca2þ block of epithelial Naþ channels in planar lipid bilayers. Biophys. J. 80, 2176–2186. Bidani, A. K., and GriYn, K. A. (2002). Long‐term renal consequences of hypertension for normal and diseased kidneys. Curr. Opin. Nephrol. Hypertens. 11, 73–80. Bidani, A. K., Schwartz, M. M., and Lewis, E. J. (1987). Renal autoregulation and vulnerability to hypertensive injury in remnant kidney. Am. J. Physiol. 252, F1003–F1010. Bonny, O., and Hummler, E. (2000). Dysfunction of epithelial sodium transport: From human to mouse. Kidney Int. 57, 1313–1318. Bonny, O., Chraibi, A., LoYng, J., Jaeger, N. F., Grunder, S., Horisberger, J. D., and Rossier, B. C. (1999). Functional expression of a pseudohypoaldosteronism type I mutated epithelial Naþ channel lacking the pore‐forming region of its alpha subunit. J. Clin. Invest. 104, 967–974. Brayden, J. E., and Nelson, M. T. (1992). Regulation of arterial tone by activation of calcium‐ dependent potassium channels. Science 256, 532–535. Brouard, M., Casado, M., Djelidi, S., Barrandon, Y., and Farman, N. (1999). Epithelial sodium channel in human epidermal keratinocytes: Expression of its subunits and relation to sodium transport and diVerentiation. J. Cell Sci. 112, 3343–3352. Caldwell, R. A., Boucher, R. C., and Stutts, M. J. (2004). Serine protease activation of near‐ silent epithelial Naþ channels. Am. J. Physiol. Cell Physiol. 286, C190–C194. Carattino, M. D., Sheng, S., and Kleyman, T. R. (2004). Epithelial Naþ channels are activated by laminar shear stress. J. Biol. Chem. 279, 4120–4126. Carattino, M. D., Sheng, S., and Kleyman, T. R. (2005). Mutations in the pore region modify epithelial sodium channel gating by shear stress. J. Biol. Chem. 280, 4393–4401. Carattino, M. D., Sheng, S., Bruns, J. B., Pilewski, J. M., Hughey, R. P., and Kleyman, T. R. (2006). The epithelial Naþchannel is inhibited by a peptide derived from proteolytic processing of its alpha subunit. J. Biol. Chem. 281(27), 18901–18907. Carr, M. J., Gover, T. D., Weinreich, D., and Undem, B. J. (2001). Inhibition of mechanical activation of guinea‐pig airway aVerent neurons by amiloride analogues. Br. J. Pharmacol. 133, 1255–1262. Chalfie, M., and Sulston, J. (1981). Developmental genetics of the mechanosensory neurons of Caenorhabditis elegans. Dev. Biol. 82, 358–370. Chalfie, M., and Au, M. (1989). Genetic control of diVerentiation of the Caenorhabditis elegans touch receptor neurons. Science 243, 1027–1033. Chalfie, M., and Wolinsky, E. (1990). The identification and suppression of inherited neurodegeneration in Caenorhabditis elegans. Nature 345, 410–416. Chalfie, M., Dean, E., Reilly, E., Buck, K., and Thomson, J. N. (1986). Mutations aVecting microtubule structure in Caenorhabditis elegans. J. Cell Sci. Suppl. 5, 257–271. Chapleau, M. W., Li, Z., Meyrelles, S. S., Ma, X., and Abboud, F. M. (2001). Mechanisms determining sensitivity of baroreceptor aVerents in health and disease. Ann. NY Acad. Sci. 940, 1–19.
6. ENaC and Cardiovascular Mechanotransduction
147
Cheng, C., Prince, L. S., Snyder, P. M., and Welsh, M. J. (1998). Assembly of the epithelial Naþ channel evaluated using sucrose gradient sedimentation analysis. J. Biol. Chem. 273, 22693–22700. Copeland, S. J., Berdiev, B. K., Ji, H. L., Lockhart, J., Parker, S., Fuller, C. M., and Benos, D. J. (2001). Regions in the carboxy terminus of alpha‐bENaC involved in gating and functional eVects of actin. Am. J. Physiol. Cell Physiol. 281, C231–C240. Davis, M. J., and Hill, M. A. (1999). Signaling mechanisms underlying the vascular myogenic response. Physiol. Rev. 79, 387–423. Davis, M. J., Donovitz, J. A., and Hood, J. D. (1992). Stretch‐activated single‐channel and whole cell currents in vascular smooth muscle cells. Am. J. Physiol. 262, C1083–C1088. Davis, M. J., Wu, X., Nurkiewicz, T. R., Kawasaki, J., Davis, G. E., Hill, M. A., and Meininger, G. A. (2001). Integrins and mechanotransduction of the vascular myogenic response. Am. J. Physiol. Heart Circ. Physiol. 280, H1427–H1433. Dijkink, L., Hartog, A., van Os, C. H., and Bindels, R. J. (2002). The epithelial sodium channel (ENaC) is intracellularly located as a tetramer. Pflugers Arch. 444, 549–555. Driscoll, M., and Chalfie, M. (1991). The MEC‐4 gene is a member of a family of Caenorhabditis elegans genes that can mutate to induce neuronal degeneration. Nature 349, 588–593. Driscoll, M., and Tavernarakis, N. (1997). Molecules that mediate touch transduction in the nematode Caenorhabditis elegans. Gravit. Space Biol. Bull. 10, 33–42. Driver, P. M., Rauz, S., Walker, E. A., Hewison, M., Kilby, M. D., and Stewart, P. M. (2003). Characterization of human trophoblast as a mineralocorticoid target tissue. Mol. Hum. Reprod. 9, 793–798. Drummond, H. A., Price, M. P., Welsh, M. J., and Abboud, F. M. (1998). A molecular component of the arterial baroreceptor mechanotransducer. Neuron 21, 1435–1441. Drummond, H. A., Abboud, F. M., and Welsh, M. J. (2000). Localization of beta and gamma subunits of ENaC in sensory nerve endings in the rat foot pad. Brain Res. 884, 1–12. Drummond, H. A., Welsh, M. J., and Abboud, F. M. (2001). ENaC subunits are molecular components of the arterial baroreceptor complex. Ann. NY Acad. Sci. 940, 42–47. Drummond, H. A., Gebremedhin, D., and Harder, D. R. (2004). Degenerin/epithelial Naþ channel proteins. Components of a vascular mechanosensor. Hypertension 44, 643. Du, H., and Chalfie, M. (2001). Genes regulating touch cell development in Caenorhabditis elegans. Genetics 158, 197–207. Duchatelle, P., Ohara, A., Ling, B. N., Kemendy, A. E., Kokko, K. E., Matsumoto, P. S., and Eaton, D. C. (1992). Regulation of renal epithelial sodium channels. Mol. Cell. Biochem. 114, 27–34. Earley, S., Waldron, B. J., and Brayden, J. E. (2004). Critical role for transient receptor potential channel TRPM4 in myogenic constriction of cerebral arteries. Circ. Res. 95, 922–929. Ernstrom, G. G., and Chalfie, M. (2002). Genetics of sensory mechanotransduction. Annu. Rev. Genet. 36, 411–453. Firsov, D., Gautschi, I., Merillat, A. M., Rossier, B. C., and Schild, L. (1998). The heterotetrameric architecture of the epithelial sodium channel (ENaC). EMBO J. 17, 344–352. Fricke, B., Lints, R., Stewart, G., Drummond, H., Dodt, G., Driscoll, M., and von During, M. (2000). Epithelial Naþ channels and stomatin are expressed in rat trigeminal mechanosensory neurons. Cell Tissue Res. 299, 327–334. Garcia‐Anoveros, J., and Corey, D. P. (1997). The molecules of mechanosensation. Annu. Rev. Neurosci. 20, 567–594.
148
Heather A. Drummond
Garcia‐Anoveros, J., Garcia, J. A., Liu, J. D., and Corey, D. P. (1998). The nematode degenerin UNC‐105 forms ion channels that are activated by degeneration‐ or hypercontraction‐ causing mutations. Neuron 20, 1231–1241. Garcia‐Anoveros, J., Samad, T. A., Zuvela‐Jelaska, L., Woolf, C. J., and Corey, D. P. (2001). Transport and localization of the DEG/ENaC ion channel BNaC1alpha to peripheral mechanosensory terminals of dorsal root ganglia neurons. J. Neurosci. 21, 2678–2686. Gillespie, P. G., and Walker, R. G. (2001). Molecular basis of mechanosensory transduction. Nature 413, 194–202. Golestaneh, N., Klein, C., Valamanesh, F., Suarez, G., Agarwal, M. K., and Mirshahi, M. (2001). Mineralocorticoid receptor‐mediated signaling regulates the ion gated sodium channel in vascular endothelial cells and requires an intact cytoskeleton. Biochem. Biophys. Res. Commun. 280, 1300–1306. Goodman, M. B., Ernstrom, G. G., Chelur, D. S., O’Hagan, R., Yao, C. A., and Chalfie, M. (2002). MEC‐2 regulates C. elegans DEG/ENaC channels needed for mechanosensation. Nature 415, 1039–1042. GriYn, K. A., and Bidani, A. K. (2004). Hypertensive renal damage: Insights from animal models and clinical relevance. Curr. Hypertens. Rep. 6, 145–153. GriYn, K. A., Picken, M. M., Churchill, M., Churchill, P., and Bidani, A. K. (2000). Functional and structural correlates of glomerulosclerosis after renal mass reduction in the rat. J. Am. Soc. Nephrol. 11, 497–506. Gu, G., Caldwell, G. A., and Chalfie, M. (1996). Genetic interactions aVecting touch sensitivity in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 93, 6577–6582. Hamill, O. P., and Martinac, B. (2001). Molecular basis of mechanotransduction in living cells. Physiol. Rev. 81, 685–740. Hamill, O. P., Lane, J. W., and McBride, D. W., Jr. (1992). Amiloride: A molecular probe for mechanosensitive channels. Trends Pharmacol. Sci. 13, 373–376. Harder, D. R. (1984). Pressure‐dependent membrane depolarization in cat middle cerebral artery. Circ. Res. 55, 197–202. Hayashi, K., Epstein, M., and Loutzenhiser, R. (1992). Enhanced myogenic responsiveness of renal interlobular arteries in spontaneously hypertensive rats. Hypertension 19, 153–160. Hill, M. A., Davis, M. J., Meininger, G. A., Potocnik, S. J., and Murphy, T. V. (2006). Arteriolar myogenic signalling mechanisms: Implications for local vascular function. Clin. Hemorheol. Microcirc. 34, 67–79. Huang, M., and Chalfie, M. (1994). Gene interactions aVecting mechanosensory transduction in Caenorhabditis elegans. Nature 367, 467–470. Hughey, R. P., Bruns, J. B., Kinlough, C. L., Harkleroad, K. L., Tong, Q., Carattino, M. D., Johnson, J. P., Stockand, J. D., and Kleyman, T. R. (2004a). Epithelial sodium channels are activated by furin‐dependent proteolysis. J. Biol. Chem. 279, 18111–18114. Hughey, R. P., Bruns, J. B., Kinlough, C. L., and Kleyman, T. R. (2004b). Distinct pools of epithelial sodium channels are expressed at the plasma membrane. J. Biol. Chem. 279, 48491–48494. Hummler, E., and Rossier, B. C. (1996). Physiological and pathophysiological role of the epithelial sodium channel in the control of blood pressure. Kidney Blood Press. Res. 19, 160–165. Hummler, E., and Beermann, F. (2000). Scnn1 sodium channel gene family in genetically engineered mice. J. Am. Soc. Nephrol. 11(Suppl. 16), S129–S134. Hummler, E., and Vallon, V. (2005). Lessons from mouse mutants of epithelial sodium channel and its regulatory proteins. J. Am. Soc. Nephrol. 16, 3160–3166.
6. ENaC and Cardiovascular Mechanotransduction
149
Ichikawa, H., Fukuda, T., Terayama, R., Yamaai, T., Kuboki, T., and Sugimoto, T. (2005). Immunohistochemical localization of gamma and beta subunits of epithelial Naþ channel in the rat molar tooth pulp. Brain Res. 1065, 138–141. Ingber, D. E. (1997). Tensegrity: The architectural basis of cellular mechanotransduction. Annu. Rev. Physiol. 59, 575–599. Ingber, D. E. (2006). Cellular mechanotransduction: Putting all the pieces together again. FASEB J. 20, 811–827. Ismailov, I. I., Berdiev, B. K., Bradford, A. L., Awayda, M. S., Fuller, C. M., and Benos, D. J. (1996a). Associated proteins and renal epithelial Naþ channel function. J. Membr. Biol. 149, 123–132. Ismailov, I. I., Awayda, M. S., Berdiev, B. K., Bubien, J. K., Lucas, J. E., Fuller, C. M., and Benos, D. J. (1996b). Triple‐barrel organization of ENaC, a cloned epithelial Naþ channel. J. Biol. Chem. 271, 807–816. Ismailov, I. I., Berdiev, B. K., Shlyonsky, V. G., and Benos, D. J. (1997a). Mechanosensitivity of an epithelial Naþ channel in planar lipid bilayers: Release from Ca2þ block. Biophys. J. 72, 1182–1192. Ismailov, I. I., Shlyonsky, V. G., and Benos, D. J. (1997b). Streaming potential measurements in alphabetagamma‐rat epithelial Naþ channel in planar lipid bilayers. Proc. Natl. Acad. Sci. USA 94, 7651–7654. Jernigan, N. L., and Drummond, H. A. (2006a). Myogenic vasoconstriction in mouse renal interlobar arteries: Role of endogenous beta and gamma ENac. Am. J. Physiol. Renal Physiol. 296(6), F1184–1191. Jernigan, N. L., and Drummond, H. A. (2006b). Suppression of endogenous beta and gamma ENaC abolishes myogenic vasoconstriction in mouse interlobar arteries. FASEB J. 20, 2006 Experimental Biology meeting abstracts [on CD‐ROM], Abstract #472.9. Ji, H.‐L., Fuller, C. M., and Benos, D. J. (1998). Osmotic pressure regulates a g‐rENaC expressed in Xenopus oocytes. Am. J. Physiol. 275, C1182–C1190. Jovov, B., Tousson, A., Ji, H. L., Keeton, D., Shlyonsky, V., Ripoll, P. J., Fuller, C. M., and Benos, D. J. (1999). Regulation of epithelial Na(þ) channels by actin in planar lipid bilayers and in the Xenopus oocyte expression system. J. Biol. Chem. 274, 37845–37854. Jovov, B., Berdiev, B. K., Fuller, C. M., Ji, H. L., and Benos, D. J. (2001). The serine‐protease trypsin cleaves C‐termini of beta and gamma subunits of ENaC. J. Biol. Chem. 5, 5. Kellenberger, S., and Schild, L. (2002). Epithelial sodium channel/degenerin family of ion channels: A variety of functions for a shared structure. Physiol. Rev. 82, 735–767. Kirber, M. T., Walsh, J. V., Jr., and Singer, J. J. (1988). Stretch‐activated ion channels in smooth muscle: A mechanism for the initiation of stretch‐induced contraction. Pflugers Arch. 412, 339–345. Kizer, N., Guo, X.‐L., and Hruska, K. (1997). Reconstitution of stretch‐activated cation channels by expression of the a‐subunit of the epithelial sodium channel cloned from osteoblasts. Proc. Natl. Acad. Sci. USA 94, 1013–1018. Kleyman, T. R., and Cragoe, E. J., Jr. (1988). Amiloride and its analogs as tools in the study of ion transport. J. Membr. Biol. 105, 1–21. Knot, H. J., and Nelson, M. T. (1995). Regulation of membrane potential and diameter by voltage‐dependent Kþ channels in rabbit myogenic cerebral arteries. Am. J. Physiol. 269, H348–H355. Kopp, U. C., Matsushita, K., Sigmund, R. D., Smith, L. A., Watanabe, S., and Stokes, J. B. (1998). Amiloride‐sensitive Naþ channels in pelvic uroepithelium involved in renal sensory receptor activation. Am. J. Physiol. 275, R1780–R1792. Kosari, F., Sheng, S., Li, J., Mak, D. O., Foskett, J. K., and Kleyman, T. R. (1998). Subunit stoichiometry of the epithelial sodium channel. J. Biol. Chem. 273, 13469–13474.
150
Heather A. Drummond
Koyama, K., Sasaki, I., Naito, H., Funayama, Y., Fukushima, K., Unno, M., Matsuno, S., Hayashi, H., and Suzuki, Y. (1999). Induction of epithelial Naþ channel in rat ileum after proctocolectomy. Am. J. Physiol. 276, G975–G984. Lewis, S. A., and Alles, W. P. (1986). Urinary kallikrein: A physiological regulator of epithelial Naþ absorption. Proc. Natl. Acad. Sci. USA 83, 5345–5348. Lifton, R. P. (1995). Genetic determinants of human hypertension. Proc. Natl. Acad. Sci. USA 92, 8545–8551. Lin, S. Y., and Corey, D. P. (2005). TRP channels in mechanosensation. Curr. Opin. Neurobiol. 15, 350–357. Lingueglia, E., Renard, S., Waldmann, R., Voilley, N., Champigny, G., Plass, H., Lazdunski, M., and Barbry, P. (1994). DiVerent homologous subunits of the amiloride‐sensitive Naþ channel are diVerently regulated by aldosterone. J. Biol. Chem. 269, 13736–13739. Liu, J., Schrank, B., and Waterston, R. H. (1996). Interaction between a putative mechanosensory membrane channel and a collagen. Science 273, 361–364. Liu, L., Johnson, W. A., and Welsh, M. J. (2003a). Drosophila DEG/ENaC pickpocket genes are expressed in the tracheal system, where they may be involved in liquid clearance. Proc. Natl. Acad. Sci. USA 100, 2128–2133. Liu, L., Leonard, A. S., Motto, D. G., Feller, M. A., Price, M. P., Johnson, W. A., and Welsh, M. J. (2003b). Contribution of Drosophila DEG/ENaC genes to salt taste. Neuron 39, 133–146. Loutzenhiser, R., Bidani, A. K., and Wang, X. (2004). Systolic pressure and the myogenic response of the renal aVerent arteriole. Acta Physiol. Scand. 181, 407–413. Luft, F. C. (1998). Molecular genetics of human hypertension. J. Hypertens. 16, 1871–1878. Luft, F. C. (2001). Molecular genetics of salt‐sensitivity and hypertension. Drug Metab. Dispos. 29, 500–504. Ma, H. P., Al‐Khalili, O., Ramosevac, S., Saxena, S., Liang, Y. Y., Warnock, D. G., and Eaton, D. C. (2004). Steroids and exogenous gamma‐ENaC subunit modulate cation channels formed by alpha‐ENaC in human B lymphocytes. J. Biol. Chem. 279, 33206–33212. Ma, H. P., Li, L., Zhou, Z. H., Eaton, D. C., and Warnock, D. G. (2002). ATP masks stretch activation of epithelial sodium channels in A6 distal nephron cells. Am. J. Physiol. Renal Physiol. 282, F501–F505. Mano, I., and Driscoll, M. (1999). DEG/ENaC channels: A touchy superfamily that watches its salt. Bioessays 21, 568–578. Martinez‐Lemus, L. A., Sun, Z., Trache, A., Trzciakowski, J. P., and Meininger, G. A. (2005). Integrins and regulation of the microcirculation: From arterioles to molecular studies using atomic force microscopy. Microcirculation 12, 99–112. Mauro, T., Guitard, M., Behne, M., Oda, Y., Crumrine, D., Komuves, L., Rassner, U., Elias, P. M., and Hummler, E. (2002). The ENaC channel is required for normal epidermal diVerentiation. J. Invest. Dermatol. 118, 589–594. Mazzochi, C., Bubien, J. K., Smith, P. R., and Benos, D. J. (2006). The carboxyl terminus of the alpha‐subunit of the amiloride‐sensitive epithelial sodium channel binds to F‐actin. J. Biol. Chem. 281, 6528–6538. McCarter, G. C., Reichling, D. B., and Levine, J. D. (1999). Mechanical transduction by rat dorsal root ganglion neurons in vitro. Neurosci. Lett. 273, 179–182. McDonald, F. J., Yang, B., Hrstka, R. F., Drummond, H. A., Tarr, D. E., McCray, P. B., Jr., Stokes, J. B., Welsh, M. J., and Williamson, R. A. (1999). Disruption of the beta subunit of the epithelial Naþ channel in mice: Hyperkalemia and neonatal death associated with a pseudohypoaldosteronism phenotype. Proc. Natl. Acad. Sci. USA 96, 1727–1731.
6. ENaC and Cardiovascular Mechanotransduction
151
Meininger, G. A., and Davis, M. J. (1992). Cellular mechanisms involved in the vascular myogenic response. Am. J. Physiol. 263, H647–H659. Morimoto, T., Liu, W., Woda, C., Carattino, M., Wei, Y., Hughey, R., Apodaca, G., Satlin, L. M., and Kleyman, T. R. (2006). Mechanism underlying flow‐stimulation of sodium absorption in the mammalian collecting duct. Am. J. Physiol. Renal Physiol. 291, F663–F669. Morris, C. E. (1990). Mechanosensitive ion channels. J. Membr. Biol. 113, 93–107. Nelson, M. T., Conway, M. A., Knot, H. J., and Brayden, J. E. (1997). Chloride channel blockers inhibit myogenic tone in rat cerebral arteries. J. Physiol. 502(Pt. 2), 259–264. O’Hagan, R., Chalfie, M., and Goodman, M. B. (2005). The MEC‐4 DEG/ENaC channel of Caenorhabditis elegans touch receptor neurons transduces mechanical signals. Nat. Neurosci. 8, 43–50. Oh, Y. S., and Warnock, D. G. (2000). Disorders of the epithelial Na(þ) channel in Liddle’s syndrome and autosomal recessive pseudohypoaldosteronism type 1. Exp. Nephrol. 8, 320–325. Ohya, Y., Adachi, N., Nakamura, Y., Setoguchi, M., Abe, I., and Fujishima, M. (1998). Stretch‐activated channels in arterial smooth muscle of genetic hypertensive rats. Hypertension 31, 254–258. Olivieri, O., Castagna, A., Guarini, P., Chiecchi, L., Sabaini, G., Pizzolo, F., Corrocher, R., and Righetti, P. G. (2005). Urinary prostasin: A candidate marker of epithelial sodium channel activation in humans. Hypertension 46, 683–688. Oyabe, A., Masumoto, N., Ueta, K., and Nakayama, K. (2000). Amiloride‐sensitive pressure‐ induced myogenic contraction in rat cerebral artery. Fundam. Clin. Pharmacol. 14, 369–377. Page, K. R., Ashworth, C. J., McArdle, H. J., Finch, A. M., and Nwagwu, M. O. (2003). Sodium transport across the chorioallantoic membrane of porcine placenta involves the epithelial sodium channel (ENaC). J. Physiol. 547, 849–857. Paintal, A. S. (1973). Vagal sensory receptors and their reflex eVects. Physiol. Rev. 53, 159–227. Palmer, L. G., and Frindt, G. (1986). Epithelial sodium channels: Characterization by using the patch‐clamp technique. Fed. Proc. 45, 2708–2712. Palmer, L. G., and Frindt, G. (1996). Gating of Na channels in the rat cortical collecting tubule: EVects of voltage and membrane stretch. J. Gen. Physiol. 107, 35–45. Pradervand, S., Barker, P. M., Wang, Q., Ernst, S. A., Beermann, F., Grubb, B. R., Burnier, M., Schmidt, A., Bindels, R. J., Gatzy, J. T., Rossier, B. C., and Hummler, E. (1999). Salt restriction induces pseudohypoaldosteronism type 1 in mice expressing low levels of the beta‐ subunit of the amiloride‐sensitive epithelial sodium channel. Proc. Natl. Acad. Sci. USA 96, 1732–1737. Price, M. P., Lewin, G. R., Mcllwrath, S. L., Cheng, C., Xie, J., Heppenstall, P. A., Stucky, C. L., Mannsfeldt, A. G., Brennan, T. J., Drummond, H. A., Qiao, J., Benson, C. J., et al. (2000). The mammalian sodium channel BNC1 is required for normal touch sensation. Nature 407, 1007–1011. Price, M. P., McIlwrath, S. L., Xie, J., Cheng, C., Qiao, J., Tarr, D. E., Sluka, K. A., Brennan, T. J., Lewin, G. R., and Welsh, M. J. (2001). The DRASIC cation channel contributes to the detection of cutaneous touch and acid stimuli in mice. Neuron 32, 1071–1083. Price, M. P., Thompson, R. J., Eshcol, J. O., Wemmie, J. A., and Benson, C. J. (2004). Stomatin modulates gating of acid‐sensing ion channels. J. Biol. Chem. 279, 53886–53891. Riha, G. M., Lin, P. H., Lumsden, A. B., Yao, Q., and Chen, C. (2005). Roles of hemodynamic forces in vascular cell diVerentiation. Ann. Biomed. Eng. 33, 772–779. Rotin, D., Bar‐Sagi, D., O’Brodovich, H., Merilainen, J., Lehto, V. P., Canessa, C. M., Rossier, B. C., and Downey, G. P. (1994). An SH3 binding region in the epithelial Naþ channel (alpha rENaC) mediates its localization at the apical membrane. EMBO J. 13, 4440–4450.
152
Heather A. Drummond
Rusch, A., Kros, C. J., and Richardson, G. P. (1994). Block by amiloride and its derivatives of mechano‐electrical transduction in outer hair cells of mouse cochlear cultures. J. Physiol. 474, 75–86. Sachs, F. (1988). Mechanical transduction in biological systems. Crit. Rev. Biomed. Eng. 16, 141–169. Satlin, L. M., Sheng, S., Woda, C. B., and Kleyman, T. R. (2001). Epithelial Na(þ) channels are regulated by flow. Am. J. Physiol. Renal Physiol. 280, F1010–F1018. Shakibaei, M., and Mobasheri, A. (2003). Beta1‐integrins co‐localize with Na, K‐ATPase, epithelial sodium channels (ENaC) and voltage activated calcium channels (VACC) in mechanoreceptor complexes of mouse limb‐bud chondrocytes. Histol. Histopathol. 18, 343–351. Sheperd, J.T, and Mancia, G. (1986). Reflex control of the human cardiovascular system. Rev. Physiol. Biochem. Pharmacol. 105, 1–99. ShreZer, W., and Wolinsky, E. (1997). Genes controlling ion permeability in both motorneurons and muscle. Behav. Genet. 27, 211–221. Snitsarev, V., Whiteis, C. A., Abboud, F. M., and Chapleau, M. W. (2002). Mechanosensory transduction of vagal and baroreceptor aVerents revealed by study of isolated nodose neurons in culture. Auton. Neurosci. 98, 59–63. Snyder, P. M., Cheng, C., Prince, L. S., Rogers, J. C., and Welsh, M. J. (1998). Electrophysiological and biochemical evidence that DEG/ENaC cation channels are composed of nine subunits. J. Biol. Chem. 273, 681–684. Sulston, J., Dew, M., and Brenner, S. (1975). Dopaminergic neurons in the nematode Caenorhabditis elegans. J. Comp. Neurol. 163, 215–226. Syntichaki, P., and Tavernarakis, N. (2004). Genetic models of mechanotransduction: The nematode Caenorhabditis elegans. Physiol. Rev. 84, 1097–1153. Tavernarakis, N., and Driscoll, M. (1997). Molecular modeling of mechanotransduction in the nematode Caenorhabditis elegans. Annu. Rev. Physiol. 59, 659–689. Tavernarakis, N., and Driscoll, M. (2001). Degenerins. At the core of the metazoan mechanotransducer? Ann. NY Acad. Sci. 940, 28–41. Tavernarakis, N., ShreZer, W., Wang, S., and Driscoll, M. (1997). unc‐8, a DEG/ENaC family member, encodes a subunit of a candidate mechanically gated channel that modulates C. elegans locomotion. Neuron 18, 107–119. Trujillo, E., Alvarez de la Rosa, D., Mobasheri, A., Gonzalez, T., Canessa, C. M., and Martin‐ Vasallo, P. (1999). Sodium transport systems in human chondrocytes. II. Expression of ENaC, Naþ/Kþ/2Cl cotransporter and Naþ/Hþ exchangers in healthy and arthritic chondrocytes. Histol. Histopathol. 14, 1023–1031. Ugawa, S., Minami, Y., Guo, W., Saishin, Y., Takatsuji, K., Yamamoto, T., Tohyama, M., and Shimada, S. (1998). Receptor that leaves a sour taste in the mouth. Nature 395, 555–556. Vallet, V., Chraibi, A., Gaeggeler, H. P., Horisberger, J. D., and Rossier, B. C. (1997). An epithelial serine protease activates the amiloride‐sensitive sodium channel. Nature 389, 607–610. Van Dokkum, R. P., Alonso‐Galicia, M., Provoost, A. P., Jacob, H. J., and Roman, R. J. (1999). Impaired autoregulation of renal blood flow in the fawn‐hooded rat. Am. J. Physiol. 276, R189–R196. Van Renterghem, C., and Lazdunski, M. (1991). A new non‐voltage‐dependent, epithelial‐like Naþ channel in vascular smooth muscle cells. Pflugers Arch. 419, 401–408. Waldmann, R. (2001). Proton‐gated cation channels—neuronal acid sensors in the central and peripheral nervous system. Adv. Exp. Med. Biol. 502, 293–304. Waldmann, R., Champigny, G., Bassilana, F., Voilley, N., and Lazdunski, M. (1995). Molecular cloning and functional expression of a novel amiloride‐sensitive Naþ channel. J. Biol. Chem. 270, 27411–27414.
6. ENaC and Cardiovascular Mechanotransduction
153
Wang, X., Ajikobi, D. O., Salevsky, F. C., and Cupples, W. A. (2000). Impaired myogenic autoregulation in kidneys of Brown Norway rats. Am. J. Physiol. Renal Physiol. 278, F962–F969. Wellner, M. C., and Isenberg, G. (1993). Stretch‐activated nonselective cation channels in urinary bladder myocytes: Importance for pacemaker potentials and myogenic response. EXS 66, 93–99. Welsh, D. G., Morielli, A. D., Nelson, M. T., and Brayden, J. E. (2002). Transient receptor potential channels regulate myogenic tone of resistance arteries. Circ. Res. 90, 248–250. Wu, X., and Davis, M. J. (2001). Characterization of stretch‐activated cation current in coronary smooth muscle cells. Am. J. Physiol. Heart Circ. Physiol. 280, H1751–H1761. Yamamoto, Y., and Taniguchi, K. (2006). Expression of ENaC subunits in sensory nerve endings in the rat larynx. Neurosci. Lett. 402, 227–232. Zuckerman, J. B., Chen, X., Jacobs, J. D., Hu, B., Kleyman, T. R., and Smith, P. R. (1999). Association of the epithelial sodium channel with Apx and alpha‐spectrin in A6 renal epithelial cells. J. Biol. Chem. 274, 23286–23295.
This page intentionally left blank
CHAPTER 7 Regulation of the Mechano‐Gated K2P Channel TREK‐1 by Membrane Phospholipids Jean Chemin,* Amanda Jane Patel,{ Patrick Delmas,{ Frederick Sachs,} Michel Lazdunski,{ and Eric Honore{ *Institut de Genomique Fonctionnelle, UPR 2580 CNRS, F‐34094 Montpellier cedex 05, France { Institut de Pharmacologie Mole´culaire et cellulaire, UMR 6097 CNRS, 06560 Valbonne, France { Laboratoire de Neurophysiologie Cellulaire, Faculte de Medecine, UMR 6150 CNRS, 13916 Marseille Cedex 20, France } The Department of Physiology and Biophysics, Center for Single Molecule Biophysics, SUNY at Buffalo, BuValo, New York 14214
I. II. III. IV.
Overview Introduction TREK‐1 Stimulation by Membrane Phospholipids TREK‐1 Inhibition by Membrane Phospholipids References
I. OVERVIEW TREK‐1 (KCNK2 or K2P2.1) is a polymodal Kþ channel that is activated by membrane stretch, intracellular acidosis, heat, and cellular lipids such as arachidonic acid (AA). Phospholipids, including PIP2, exert a dual dose‐ dependent eVect on TREK‐1. Low concentrations transform the mechano‐ gated Kþ channel TREK‐1 into a leak Kþ channel. The phospholipid‐sensing domain is a positively charged cluster in the proximal C‐terminal domain. This region also encompasses the proton sensor E306 that is required for activation of TREK‐1 by cytosolic acidosis. Protonation of E306 increases Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59007-6
156
Chemin et al.
channel‐phospholipid interaction leading to TREK‐1 opening without direct mechanical stimulation. At higher concentrations, intracellular phospholipids inhibit channel activation by stretch, intracellular acidosis, and AA. Binding endogenous negative inner leaflet phospholipids with polylysine reduces the inhibition and reveals channel stimulation by exogenous intracellular phospholipids. Both stimulatory and inhibitory eVects are observed with phosphatidylinositol (PI), phosphatidylethanolamine (PE), phosphatidylserine (PS), and phosphatidic acid (PA), but not diacylglycerol (DG), suggesting that the phosphate at position 3 is required, although the net charge is not critical. In conclusion, membrane phospholipids, including PIP2, are major regulators of TREK‐1 channel activity.
II. INTRODUCTION Mammalian K2P channel subunits have 4 transmembrane segments and 2P domains in tandem, and the family has at least 15 members (Patel and Honore´, 2001). The P domains are part of the Kþ conduction pathway. Functional K2P channels are dimers of subunits and heteromultimerization occurs (Lesage et al., 1996a,b, 1997; Czirjak and Enyedi, 2001; Lauritzen et al., 2003). Although the subunits appear to have the same structural motif, they only share moderate sequence homology outside the P regions (Patel and Honore´, 2001). Human TREK‐1 mRNA is highly expressed in the brain, the spinal cord, the stomach, and, to a lesser extent, in the small intestine (Fink et al., 1996; Medhurst et al., 2001). In the central nervous system, TREK‐1 shows the greatest expression in the caudate nucleus and the putamen (Medhurst et al., 2001). Besides hippocampal glutamatergic neurons, TREK‐1 is localized in GABAergic interneurons (Hervieu et al., 2001) and sensory neurons of dorsal root ganglia (Maingret et al., 2000a; Medhurst et al., 2001). TREK‐1 is present at both synaptic and nonsynaptic sites (Maingret et al., 2000a). TREK channels are activated by mechanical stress (Patel et al., 1998). The single‐channel conductance is about 100 pS (positive potentials, 155 mM Kþ), and the outward rectification of TREK‐1 in symmetric saline is a combination of a mild intrinsic voltage‐dependency and external Mg2þ block at negative potentials (Bockenhauer et al., 2001; Maingret et al., 2002). The typical resting activity suggests that TREK‐1 will influence both resting excitability and the action potential duration (Patel and Honore´, 2001). At the whole‐cell level, TREK‐1 is modulated by cellular volume, with hyperosmolarity closing the channel (Patel et al., 1998). Mechanical force may be transmitted directly to the channel via the lipid bilayer, while the cytoskeleton tonically represses channel activity (Maingret et al., 1999a;
7. Phospholipids and K2P Channels
157
Patel et al., 2001). Heat reversibly opens TREK‐1 with about a sevenfold increase in current amplitude for a temperature jump of 10 C (Maingret et al., 2000a). Acidic pHi shifts the pressure–activation relationship toward lower pressures, leading to channel opening without applied stimuli (Maingret et al., 1999b; Lesage et al., 2000; Honore´ et al., 2002). Intracellular acidosis converts a TREK mechano‐gated channel into a constitutively active channel. Deletional and chimeric analyses demonstrate that the C‐terminal domain, but not the N‐terminal domain, is critical for TREK‐1 activation by stretch, temperature, and intracellular acidosis (Maingret et al., 1999b, 2000a; Honore´ et al., 2002). E306 in the proximal C‐terminal domain is the key intracellular proton sensor that regulates TREK‐1 channel activity (Honore´ et al., 2002). TREK channels are also sensitive to amphipaths. For instance, TREK‐1 is reversibly activated by polyunsaturated fatty acids (PUFAs) including AA, but not by saturated fatty acids (Fink et al., 1998; Patel et al., 1998; Kim et al., 2001a,b). Activation of TREK and TRAAK channels by PUFA in excised patches indicates that the eVect does not require the intact cell (Patel et al., 2001), but is due to a direct interaction with the channel in its local membrane environment (Patel et al., 2001). Extracellular lysophospholipids (LP), including lysophosphatidylcholine (LPC), open TREK channels (Lesage et al., 2000; Maingret et al., 2000b). Deletional analysis again indicates that the C‐terminal domain, but not the N‐terminal domain and the extracellular loop M1P1 of TREK‐1, is critical for both AA and LPC activation (Patel et al., 1998; Maingret et al., 2000b). TREK‐1 is also opened by intracellular lysophosphatidic acid (LPA; Chemin et al., 2005a). The pharmacological properties of TREK‐1 have been reviewed elsewhere (Lesage, 2003; Franks and Honore, 2004). When coexpressed with the 5HT4 receptor, serotonin inhibits TREK‐1 (Fink et al., 1996; Patel et al., 1998). This eVect is mimicked by a membrane permeant cAMP derivative and the eVect is mediated by protein kinase A‐mediated phosphorylation of S333 in the C‐terminal domain (Patel et al., 1998). In contrast, sodium nitroprusside and 8‐Br‐cGMP increase TREK‐1 currents (Koh et al., 2001) probably through the PKG consensus sequence at S351 (Koh et al., 2001). Although a minor component of the plasma membrane, PIP2 is increasingly recognized as a key physiological regulator of ion channel activity (Hilgemann et al., 2001). PIP2 controls the nucleotide sensitivity of KATP channels and promotes channel opening (Baukrowitz et al., 1998; Shyng and Nichols, 1998). It directly activates inward rectifier IRK potassium channels and is essential for Gbg‐protein activation of GIRK channels (Huang et al., 1998; Mirshahi et al., 2003; Delmas and Brown, 2005); PIP2 also regulates both activation and inactivation of voltage‐gated Kþ channels (Loussouarn, 2003
158
Chemin et al.
#1031; Oliver et al., 2004). Genetic alterations of Kþ channel–PIP2 interactions can lead to channelopathies such as Andersen’s and Bartter’s syndromes (Lopes et al., 2002). Receptor‐mediated PIP2 hydrolysis plays a key role in the regulation of several ion channel types (Kobrinsky et al., 2000; Chuang et al., 2001; Runnels et al., 2002; Suh and Hille, 2002; Zhang et al., 2003), including the TASK K2P channels (Czirjak et al., 2001; Chemin et al., 2003). In the present chapter, we review the evidence that membrane phospholipids, including PIP2, are major regulators of TREK‐1 channel activity. Regulation of mechano‐gated K2P channels is complex with acute stimulation at low concentration and inhibition at high concentration.
III. TREK‐1 STIMULATION BY MEMBRANE PHOSPHOLIPIDS Both the cloned and endogenous neuronal TREK‐1 channels are highly dependent for their activity on membrane phospholipids (Chemin et al., 2005b; Lopes et al., 2005). The cationic molecules polylysine (pL) and spermine, which have a high aYnity for phospholipids, inhibit TREK‐1 channel activity at rest, but the activity is restored and even stimulated by PIP2 (Chemin et al., 2005b; Lopes et al., 2005). The inhibitory eVect of polycationic molecules is thought to be due to the fact that they interact with the negative charges of essential membrane phospholipids, thus removing them from their electrostatic interaction with specific positively charged segments in the channel protein (Huang et al., 1998; Lopes et al., 2002). In the presence of intracellular polylysine or other polyamines, TREK‐1 remains in an inactive state (Chemin et al., 2005b; Lopes et al., 2005; Fig. 1A). We have identified a cluster of five positive charges in the proximal C‐terminal domain of TREK‐1 that is central to the eVect of phospholipids (Chemin et al., 2005b; Fig. 1A). When the positive charges are deleted, TREK‐1 becomes more resistant to the activation by PIP2 (Chemin et al., 2005b). Using a complementation yeast system designed to assay for membrane– protein interactions (Aronheim, 2001), we have shown that the C‐terminal domain of TREK‐1 is in close proximity to the plasma membrane (Chemin et al., 2005b). Moreover, the positive charges in this cluster, crucial for PIP2 stimulation of the channel, are also associated with the bilayer as confirmed by fluorescence experiments (Chemin et al., 2005b). The model presented in Fig. 1 summarizes our results and our mechanistic interpretation concerning the eVects of PIP2. When the positively charged segment cannot interact with the inner leaflet phospholipids, as for instance in the presence of polyamines, TREK‐1 is inactive and stretch or internal acidification cannot promote channel activity (Fig. 1A). When PIP2 neutralizes the positively charged
159
7. Phospholipids and K2P Channels TMS4
A
TMS4 Closed
E306 +
++
++
PIP2
B
++
++
E306 +
Polyamines
+++ + + Acidosis
C
PIP2 pHi
+++ + +
Depolarization
Stretch Gated
E306A
Leak + 0++++
+ + + +0 +
FIGURE 1 Model of TREK‐1 gating. (A) Positive charges in the C‐terminal domain of TREK‐1 are critically required for PIP2 stimulation. In the 5 þ A mutant, R297, K301, K302, K304, and R311 are substituted by an alanine. In the E306A mutant, glutamate at position 306 is substituted by an alanine. In the presence of polylysine (pL) or endogenous polyamines, TREK‐1 is in the closed state and not activable (closed state). (B) Phospholipids, including PIP2, electrostatically interact with the positively charged cluster (5þ) in the cytosolic proximal C‐terminal domain of TREK‐1. Insertion of PIP2 in the inner leaflet of the bilayer controls coupling of the C‐terminal domain of TREK‐1 with the plasma membrane. When partially coupled, TREK‐1 is in the closed state but activable by membrane stretch, depolarization, and cytosolic acidosis (gated state). (C) This membrane interaction is favored when the negative charge of the proton sensor E306 is masked by either protonation at acidic pHi or by substitution with an alanine (E306A). The E306A mutant is locked open and behaves as a leak Kþ channel. Similarly, in the presence of exogenous phospholipids, cytosolic acidosis irreversibly locks TREK‐1 open (leak state).
C‐terminal segment, it inserts in the membrane near the channel making it activable by stretch, depolarization, or cytosolic acidification (Fig. 1B). This model proposes that E306 plays an important regulatory role in the interaction between membrane phospholipids and the C‐terminal domain of
160
Chemin et al.
TREK‐1 (Fig. 1C). Protonation of E306 makes the region more positive, and probably increases the electrostatic interaction with the negative charges of inner leaflet phospholipids. The interaction is then suYciently strong that dissociation rate is slower than the fluid exchange time when the cytosolic pH is brought back to normal. The TREK‐1 channel becomes constitutively open (Chemin et al., 2005b). Removal of the negative charge of E306 by substituting an alanine mimics the eVect of acidic pHi (Fig. 1C). PI that lacks the phosphates on the inositol group stimulates TREK‐1 channel activity. Furthermore, PS and PE (Fig. 6A) that do not belong to the phosphoinositide family, but are also inner leaflet phospholipids, stimulate TREK‐1 (Chemin et al., 2005b). Similar to PIP2, these phospholipids lock TREK‐1 in the open conformation during cytosolic acidosis (Chemin et al., 2005b). The presence of a large polar head is not an absolute requirement since PA (Fig. 6A) also stimulates TREK‐1 after a previous polylysine treatment (Chemin et al., 2005b). However, the negative phosphate group at position 3 of the glycerol is critical as DG (Fig. 6A) does not stimulate channel activity. Thus, although PIP2 is generally considered as the key phospholipid for regulation of many ion channels, other inner leaflet phospholipids such as PI, PE, or PS can fulfill the same role (Chemin et al., 2005b; Fig. 6). TREK‐1 conduction has an intrinsic voltage‐dependency (Bockenhauer et al., 2001; Maylie and Adelman, 2001; Maingret et al., 2002; Chemin et al., 2005b; Lopes et al., 2005). When TREK‐1 is locked open by PIP2, the outward rectification disappears. This may reflect modification of the local electric field by the negative charge on PIP2. Glutamate is a major excitatory neurotransmitter in the central nervous system via ionotropic receptors. Glutamate also activates metabotropic receptors (mGluRs) that modulate neuronal excitability and synaptic transmission, resulting in persistent depolarization and increased cell firing. mGluR1 and mGluR5 are located primarily in postsynaptic areas where they can tonically modify cellular excitability. Group I mGluRs are coupled to Gq that in turn stimulates phospholipase C (PLC), increasing phosphoinositide hydrolysis and the generation of inositol trisphosphate (InsP3) and DG. TREK‐1 channel activity, recorded in the presence of AA, is reversibly inhibited by DG and this mechanism may contribute to the down‐ modulation of TREK‐1 by stimulation of mGluR1 and mGluR5 (Chemin et al., 2003). Furthermore, wortmannin (that inhibits PI 4‐kinase, depleting PIP2) inhibits TREK‐1 and slows recovery of TREK‐1 from Gq‐coupled receptor‐induced inhibition (Lopes et al., 2005). PIP2 hydrolysis may thus also contribute to Gq‐coupled receptor‐mediated TREK‐1 inhibition (Lopes et al., 2005).
161
7. Phospholipids and K2P Channels
IV. TREK‐1 INHIBITION BY MEMBRANE PHOSPHOLIPIDS Intracellular PIP2 stimulates TREK‐1 in 46% of patches (Chemin et al., 2005b; Fig. 2A and C), while it inhibits the other 54% (Fig. 2B and C). This dual behavior is observed with excised patches from cells of the same culture dish and is independent of the level of channel expression. A transient stimulation sometimes occurs before complete inhibition (Fig. 2B). The stimulation
C
I (nA)
10
25 s −100
V (mV)
Control
Inhibition
30
PIP2 Cells (%)
1 nA
A
20 Stimulation 10
100 0 0
PIP2
B
D
1
2 3 4 I PIP2/Icontrol
5
6
Control
25 s
−100
Hippocampus
I (nA) 200 pA
1 PIP2
V (mV)
100
COS cells
FIGURE 2 Dual regulation of TREK‐1 by PIP2. (A) Inside‐out patch excised from a transiently transfected COS cell expressing TREK‐1. The holding potential was 80 mV and the patch was stimulated every 5 s by a voltage ramp of 600 ms in duration from 100 to 100 mV. The current amplitude was measured at 90 mV. 5 mM PIP2 added intracellularly as indicated by a horizontal green line stimulates TREK‐1 channel activity. The horizontal dashed line indicates the zero current. The inset illustrates the stimulatory eVect of PIP2 on the I–V curve of TREK‐1 recorded in a physiological Kþ gradient. (B) Same in a patch where TREK‐1 is inhibited by PIP2. The application of PIP2 (5 mM) is indicated by a red line. (C) Relative distribution of the diVerential eVect of PIP2 (5 mM) on TREK‐1 channel activity (bins represent 33% of variation, n ¼ 81). Two populations of patches were identified showing inhibition and stimulation. Both populations were fitted with Gaussian curves. The wheel representation illustrates the percentage of patches showing stimulation (green) and inhibition (red) (n ¼ 81). In stimulated and inhibited patches IPIP2/Icontrol was 3.4 0.7 (n ¼ 37) and 0.3 0.03 (n ¼ 44), respectively. (D) EVect of PIP2 on TREK‐1 expressed in infected cultured hippocampal neurons (n ¼ 8) and transfected COS cells. The inhibitory and stimulatory eVects of PIP2 are illustrated in red and green, respectively. In stimulated and inhibited patches from hippocampal neurons IPIP2/Icontrol was 3.3 0.6 (n ¼ 4) and 0.08 0.05 (n ¼ 4), respectively.
162
Chemin et al.
or inhibition persists after washout (Fig. 2A and B), suggesting that the channel‐associated PIP2 has a very slow washout rate. The bimodal eVect is not specific to COS7 cells, since we found similar results in hippocampal neurons that express TREK‐1 (Fig. 2D). Inhibition by PIP2 decreases both basal activity and mechanically stimulated activity (Fig. 3A–C). TREK‐1 activation by intracellular acidosis (Fig. 4A and B) or AA (Fig. 4C and D) is similarly inhibited by PIP2. However, it is only when TREK‐1 was first activated by AA that the inhibitory eVect of PIP2 was completely reversible (n ¼ 6; Fig. 4C). Competition of the negative AA with PIP2 may increase the dissociation rate. Polylysine converts the inhibitory eVect of PIP2 to stimulation (Fig. 5A–D). Remarkably, the dose–eVect curve of PIP2 is biphasic, with net stimulation at lower concentration and inhibition at higher concentrations (Fig. 5D). A
B
Control 6
I (nA)
2 nA
200 ms 4 PIP2 2 PIP2
Wash 0 0
−20 −40 P (mmHg)
−60 mmHg
I (nA)
C
−60
3 PIP2
2 1 0
0
−45 0 P (mmHg)
−45
FIGURE 3 PIP2 inhibits TREK‐1 activation by membrane stretch. (A) EVect of PIP2 (5 mM) in an inside‐out patch excised from a transfected COS cell at a holding potential of 0 mV. A membrane stretch of 60 mmHg was applied as indicated by a horizontal line. (B) Pressure–eVect curve of TREK‐1 in control (P0.5: 33.2 mmHg, k: 9.5), in the presence of intracellular PIP2 (5 mM; P0.5: 40.8 mmHg, k: 10.6), and after washout (P0.5: 39.7 mmHg, k:11.4). P0.5 is the midpoint of the pressure‐effect curve and k is the slope factor. Same patch as A. (C) EVect of PIP2 (5 mM) on TREK‐1 current amplitude measured at rest and during a stimulation of 45 mmHg (n ¼ 10).
163
7. Phospholipids and K2P Channels A
C PIP2
pHi 5.5
pHi 5.5
AA PIP2
5 pHi 5.5
pHi 5.5 PIP2
0 −100
V (mV)
100
2 nA 5
AA AA PIP2
I (nA)
I (nA)
pHi 5.5 PIP2
D 7.5
I (nA)
10
pHi 5.5
AA PIP2
0 −100
AA
I (nA)
B
50 s
2 nA
100 s
V (mV)
100
FIGURE 4 PIP2 inhibits TREK‐1 activation by acidic intracellular pH and the polyunsaturated fatty acid AA. (A) EVect of PIP2 (5 mM) on the acidic (pHi 5.5) activation of TREK‐1. Currents were measured at 90 mV during voltage ramps. (B) TREK‐1 current at pHi 5.5 in the absence and in the presence of PIP2 (5 mM). The holding potential was 80 mV and the patch was stimulated every 5 s by a voltage ramp of 600 ms in duration from 100 to 100 mV. The current amplitude was measured at 90 mV (n ¼ 18). (C) EVect of PIP2 (5 mM) in the presence of 10‐mM AA applied intracellularly. (D) TREK‐1 currents stimulated by AA in the presence and in the absence of PIP2 (5 mM). The holding potential was 80 mV and the patch was stimulated every 5 s by a voltage ramp of 600 ms in duration from 100 to 100 mV. The current amplitude was measured at 90 mV (n ¼ 6).
Either there are heterogeneous sites of interaction or there is drug accumulation at a multioccupancy ‘‘site.’’ Inhibition is not specific for PIP2 with PI being the most potent inhibitor (Fig. 6A–D). When the first application of phospholipids (shown in red) does not produce a significant eVect on current amplitude (see for instance PE and PS; Fig. 6B–D), a brief treatment with polylysine leads to dramatic stimulation during a second application (Fig. 6B–D, shown in green). If one remembers the earlier finding that PIP2 stimulates at low concentration and inhibits at high concentration, then an immediate conclusion is that polylysine is probably reducing the number of PIP2 molecules interacting with the channel, particularly at sites responsible for the inhibitory eVect. As pointed out above, the polar head of the phospholipid is not critical for channel inhibition as PA also inhibits TREK‐1 (Fig. 6D). However, the phosphate at position 3 of the glycerol moiety apparently is required (as observed for
164
Chemin et al.
100 s pHi 5.5
pHi 5.5
B
PIP2 2nd
I (nA)
6
PIP2
I/Icontrol
Percentage of cells
5 nA
Polylysine
PIP2 2nd
5
C
PIP2 2nd
PIP2 1st
30 20
0
PIP2 1st Control polylysine
10 0
0
D
2
4 I/Icontrol
15 Ipip2/Icontrol
A
6
8
PIP2 2nd
10 PIP2 1st
5
1st
0 −100
V (mV)
100
0
0.1 1 PIP2 (µ M)
10
FIGURE 5 Polylysine lowers PIP2 inhibition. (A) EVect of PIP2 (5 mM) before (1st) (shown in red) and after (2nd) polylysine pretreatment (30 mg/ml; shown in green). The patch was stimulated by intracellular acidosis to pHi 5.5 before and after the first application of PIP2. The current amplitude was measured at 90 mV. (B) I‐V curves corresponding to the experiment illustrated in A. (C) Distribution of the diVerential eVect of PIP2 (5 mM) on TREK‐1. Before polylysine pretreatment, two populations of patches were identified showing inhibition and stimulation (gray circles). Both populations were fitted with Gaussian curves. The green circles indicate the distribution of the patches after polylysine pretreatment (30 mg/ml for 1 min). The inset shows the eVect of polylysine pretreatment on the magnitude of the eVect of PIP2 (5 mM) on inside‐out patches excised from TREK‐1 transfected COS cells (n ¼ 68). Patches with both stimulatory and inhibitory PIP2 responses were averaged. (D) Dose–eVect curve of PIP2 on TREK‐1 expressed in COS cells before (red) and after (green) pretreatment with polylysine (30 mg/ml for 1 min) on the same inside‐out patch excised from a TREK‐1 transfected COS cell. In these experiments, inside‐out patches were held at 80 mV and stimulated by voltage ramps of 600 ms in duration from 100 to 100 mV every 5 s.
TREK‐1 stimulation), since DG fails to aVect the basal TREK‐1 current (Fig. 6D shown in red). When all five basic residues required for the stimulatory eVect of PIP2 (Fig. 1) are substituted by alanines, inhibition by PIP2 is not aVected (Fig. 7A–D). Similarly, deletion of the N‐terminal domain fails to aVect either the inhibitory or stimulatory eVect of PIP2 (Fig. 7A and B). The C‐terminal domain can be progressively truncated (leaving intact the polycationic cluster), removing PIP2 inhibition but maintaining PIP2 stimulation (Fig. 7A and B and 7C inset). The deletion of the same distal C‐terminal domain (deletion of the last
165
7. Phospholipids and K2P Channels A + N
P−
B
P−
−OOC +N
P−
P−
PS 1st
PS 2nd
pHi 5.5
P−
P−
PE
PS
PIP2
C 12
PA
Polylysine
5 nA
50 s DG
PS 2nd
PA
D
15 2nd Application PS
PS 1st
−100
V(mV)
I/Icontrol
I (nA)
10 PE 5
PIP2 PI DG
100 0 1st Application
FIGURE 6 Specificity of the eVect of membrane phospholipids on TREK‐1. (A) Diagram of the diVerent phospholipid molecules. (B) EVect of PS (5 mM) on TREK‐1 channel activity. PS was applied before and after polylysine (30 mg/ml) treatment. The patch was stimulated by pHi 5.5 before PS application. In this experiment, the inside‐out patch was held at 80 mV and stimulated by voltage ramps of 600 ms in duration from 100 to 100 mV every 5 s. The current amplitude was measured at 90 mV. (C) I–V curves recorded with voltage ramps showing the eVect of PS (5 mM) before (shown in red) and after (shown in green) polylysine treatment (30 mg/ml). Same patch as B. (D) EVect of phospholipids (5 mM) on TREK‐1. PIP2 (n ¼ 81), PI (n ¼ 14), PE (n ¼ 12), PS (n ¼ 9), PA (n ¼ 18), DG (n ¼ 11) before (red bars) and PIP2 (n ¼ 68), PI (n ¼ 11), PE (n ¼ 5), PS (n ¼ 14), PA (n ¼ 6), DG (n ¼ 5) after polylysine treatment (30 mg/ml for 1 min; green bars). Patches with both stimulatory and inhibitory PIP2 responses were averaged.
100 amino acids) shifts the pressure–eVect curve of TREK‐1 toward more negative pressures (Maingret et al., 1999b). In other words, truncation of TREK‐1 gradually makes channels more resistant to stretch and leads to a decrease in inhibition by PIP2. A possible mechanism to explain PIP2 inhibition is membrane insertion. Amphipaths can alter membrane curvature by inserting asymmetrically into one monolayer or another (Sheetz and Singer, 1974). This partitioning has two eVects: compression of the lipids in that monolayer (since the opposing monolayer has not expanded) and the lateral expansion pressure causing curvature toward the other monolayer. If either of these eVects favors the
166
Chemin et al.
** 10
A
C
IPIP2/Icontrol
NS
TMS4 5 + A +++++
NS
*
89
∆C 100 0.5 nA
*** **
59 76
PIP2 50 s
** 0
pHi 5.5
WT∆N 59 76 89 100 5 + A
pHi 5.5
∆C 100
5+A 7
∆N 17 10 81
D 1.5
I (nA)
Patches inhibited (%)
B
pHi 5.5
pHi 5.5 PIP2
9
0 WT
8
8
59 76 89 100
−100
V (mV)
100
∆C FIGURE 7 The inhibitory eVect of PIP2 is abolished by truncation of the C‐terminal domain. (A) EVect of PIP2 (5 mM) on TREK‐1 WT and N‐ and C‐terminal mutants before (red) and after (green) polylysine treatment. Patches with both stimulatory and inhibitory PIP2 responses were averaged. WT (n ¼ 81), N (n ¼ 17), C59 (n ¼ 10), C76 (n ¼ 9), C89 (n ¼ 8), C100 (n ¼ 8), 5 þ A (n ¼ 7) before and WT (n ¼ 68), N (n ¼ 8), C59 (n ¼ 3), C76 (n ¼ 9), C89 (n ¼ 4), C100 (n ¼ 4), 5 þ A (n ¼ 18) after polylysine treatment (30 mg/ml for 1 min). Statistical diVerences between PIP2 eVects before and after polylysine treatment are indicated by stars (*p < 0.05, **p < 0.01, ***p < 0.001). (B) Percentage of patches inhibited by PIP2 (5 mM) when applied in the absence of polylysine treatment (first application). The number of experiments is indicated. (C) EVect of PIP2 (5 mM) on the 5 þ A mutant. The patch was stimulated by pHi 5.5 before and after PIP2 application. In this experiment, the inside‐out patch was held at 80 mV and stimulated by voltage ramps of 600 ms in duration from 100 to 100 mV every 5 s. The current amplitude was measured at 90 mV. The inset shows an illustration of the mutants in the C‐terminal domain. In the 5 þ A mutant, residues R297, K101, K102, K104, and R111 were substituted with alanines. This cluster of positive charges is required for the stimulatory eVect of PIP2 (shown in green), while the region between Ala352 and Thr322 (between C59 and C89) in the C‐terminal domain is required for inhibition (shown in red). (D) I–V curves showing the eVect of intracellular pH 5.5 before and after PIP2 (5 mM) stimulation. Same patch as C.
7. Phospholipids and K2P Channels
167
open state, amphipaths will activate the channels. Anionic amphipaths such as AA preferentially insert in the external leaflet of the bilayer because of the electrostatic repulsion with the inner leaflet anionic PS (Sheetz and Singer, 1974; Patel et al., 1998). AA and other anionic amphipaths including trinitrophenol are potent openers of TREK‐1 (Patel et al., 1998, 2001). Conversely, binding of amphipaths to the inner monolayer can explain inhibition by the cationic amphipaths including chlorpromazine and tetracaine (Patel et al., 1998). Intracellular PIP2 inhibits TREK with patches precurved by positive or negative hydrostatic pressures (Honore´, unpublished data). Thus, if amphipath‐induced curvature is important to TREK activity, it must be local to the channels. Why PIP2 stimulates some patches and inhibits others remains unclear. There are hundreds of active channels in a patch, so the variability cannot be a result of sampling. If there are heterogeneous domains in the membrane with diVerent lipid compositions or diVerent TREK‐1 assemblies including with the cytoskeleton, the domains must be of patch dimensions. Possibly, the heterogeneity is the result of diVerences in global stress, resulting from diVerent balances in cytoskeletal stress and adhesion energy of the seal that cause whole patches to vary in their properties. Activation and inhibition by PIP2 occur at diVerent PIP2 concentrations, requiring that two or more PIP2 molecules/ channels are involved. This complex behavior can arise from the existence of two diVerent types of binding sites: a specific one, the proximal cationic cluster, and less specific ‘‘sites’’ such as the boundary lipid domain (Suchyna et al., 2004) that can absorb multiple PIP2 molecules. Alternatively, it may be that binding of PIP2 to one of the monolayer may lead to saturation and then favor PIP2 flipping to the other side of the bilayer with opposite curvature eVects. Is the dual modulation of TREK‐1 activity observed by direct application of PIP2 related to TREK‐1 regulation by neurotransmitters? PIP2 is dynamically regulated by kinases and phospholipases, and Gq‐coupled receptors stimulate PLC, causing hydrolysis of PIP2 and generating second messengers DG and InsP3. PIP2 depletion has been proposed as the mechanism of Gq‐related inhibitory eVects on TREK‐1 (Lopes et al., 2005). Since PIP2 can cause inhibition or stimulation of TREK‐1 activity, in situ hydrolysis under the influence of hormones or neurotransmitters could create very complex and diverse responses in neurons. DiVerent types of regulation by a given neurotransmitter type could be spatially distributed among the diVerent types of neuronal compartments such as the soma, dendrites, growth cones, or nerve terminals. This has not yet been systematically investigated. At present, activation of Gq‐coupled receptors, such as mGluR1 and mGluR5 receptors, has only been seen to cause consistent inhibition of TREK‐1 (Chemin et al., 2003). It should also be kept in mind that other pathways have also been
168
Chemin et al.
proposed to be involved in the regulation of K2P channels by membrane receptors to hormones and neurotransmitters (Chemin et al., 2003; Murbartian et al., 2005; Chen et al., 2006). In conclusion, gating of TREK‐1 is tightly linked to the regulation by membrane phospholipids. Similar types of regulation exist in other families of ion channels, particularly TRPV, which like the TREK channels may serve as sensors of stretch, osmotic pressure, pH, and temperature (Voets and Nilius, 2003). Acknowledgments We are grateful to the ANR 2005 Cardiovasculaire‐obe´site´‐diabe`te, to the Association for Information and Research on Genetic Kidney Disease France, and to the Foundation del Duca for support.
References Aronheim, A. (2001). Ras signaling pathway for analysis of protein‐protein interactions. Methods Enzymol. 332, 260–270. Baukrowitz, T., Schulte, U., Oliver, D., Herlitze, S., Krauter, T., Tucker, S. J., Ruppersberg, J. P., and Fakler, B. (1998). PIP2 and PIP as determinants for ATP inhibition of KATP channels. Science 282, 1141–1144. Bockenhauer, D., Zilberberg, N., and Goldstein, S. A. (2001). KCNK2: Reversible conversion of a hippocampal potassium leak into a voltage‐dependent channel. Nat. Neurosci. 4, 486–491. Chemin, J., Girard, C., Duprat, F., Lesage, F., Romey, G., and Lazdunski, M. (2003). Mechanisms underlying excitatory eVects of group I metabotropic glutamate receptors via inhibition of 2P domain Kþ channels. EMBO J. 22, 5403–5411. Chemin, J., Patel, A., Duprat, F., Zanzouri, M., Lazdunski, M., and Honore, E. (2005a). Lysophosphatidic acid‐operated Kþ channels. J. Biol. Chem. 280, 4415–4421. Chemin, J., Patel, A. J., Duprat, F., Lauritzen, I., Lazdunski, M., and Honore, E. (2005b). A phospholipid sensor controls mechanogating of the Kþ channel TREK‐1. EMBO J. 24, 44–53. Chen, X., Talley, E. M., Patel, N., Gomis, A., McIntire, W. E., Dong, B., Viana, F., Garrison, J. C., and Bayliss, D. A. (2006). Inhibition of a background potassium channel by Gq protein {alpha}‐subunits. Proc. Natl. Acad. Sci. USA 103, 3422–3427. Chuang, H. H., Prescott, E. D., Kong, H., Shields, S., Jordt, S. E., Basbaum, A. I., Chao, M. V., and Julius, D. (2001). Bradykinin and nerve growth factor release the capsaicin receptor from PtdIns(4,5)P2‐mediated inhibition. Nature 411, 957–962. Czirjak, G., and Enyedi, P. (2002). Formation of functional heterodimers between the TASK‐1 and TASK‐3 two pore domain potassium channel subunits. J. Biol. Chem. 277, 5426–5432. Czirjak, G., Petheo, G. L., Spat, A., and Enyedi, P. (2001). Inhibition of TASK‐1 potassium channel by phospholipase C. Am. J. Physiol. Cell Physiol. 281, C700–C708. Delmas, P., and Brown, D. A. (2005). Pathways modulating neural KCNQ/M (Kv7) potassium channels. Nat. Rev. Neurosci. 6, 850–862. Fink, M., Duprat, F., Lesage, F., Reyes, R., Romey, G., Heurteaux, C., and Lazdunski, M. (1996). Cloning, functional expression and brain localization of a novel unconventional outward rectifier Kþ channel. EMBO J. 15, 6854–6862. Fink, M., Lesage, F., Duprat, F., Heurteaux, C., Reyes, R., Fosset, M., and Lazdunski, M. (1998). A neuronal two P domain Kþ channel activated by arachidonic acid and polyunsaturated fatty acid. EMBO J. 17, 3297–3308.
7. Phospholipids and K2P Channels
169
Franks, N. P., and Honore, E. (2004). The TREK K2P channels and their role in general anaesthesia and neuroprotection. Trends Pharmacol. Sci. 25, 601–608. Hervieu, G. J., Cluderay, J. E., Gray, C. W., Green, P. J., Ranson, J. L., Randall, A. D., and Meadows, H. J. (2001). Distribution and expression of TREK‐1, a two‐pore‐domain potassium channel, in the adult rat CNS. Neuroscience 103, 899–919. Hilgemann, D. W., Feng, S., and Nasuhoglu, C. (2001). The complex and intriguing lives of PIP2 with ion channels and transporters. Sci. STKE 2001, RE19. Honore´, E., Maingret, F., Lazdunski, M., and Patel, A. J. (2002). An intracellular proton sensor commands lipid‐ and mechano‐gating of the Kþ channel TREK‐1. EMBO J. 21, 2968–2976. Huang, C. L., Feng, S., and Hilgemann, D. W. (1998). Direct activation of inward rectifier potassium channels by PIP2 and its stabilization by Gbetagamma. Nature 391, 803–806. Kim, Y., Bang, H., Gnatenco, C., and Kim, D. (2001a). Synergistic interaction and the role of C‐terminus in the activation of TRAAK Kþ channels by pressure, free fatty acids and alkali. Pflugers Arch. 442, 64–72. Kim, Y., Gnatenco, C., Bang, H., and Kim, D. (2001b). Localization of TREK‐2 Kþ channel domains that regulate channel kinetics and sensitivity to pressure, fatty acids and pHi. Pflugers Arch. 2001, 952–960. Kobrinsky, E., Mirshahi, T., Zhang, H., Jin, T., and Logothetis, D. E. (2000). Receptor‐ mediated hydrolysis of plasma membrane messenger PIP2 leads to Kþ‐current desensitization. Nat. Cell Biol. 2, 507–514. Koh, S. D., Monaghan, K. M., Sergeant, G. P., Ro, S., Walker, R. L., Sanders, K. M., and Horowitz, B. (2001). TREK‐1 regulation by nitric oxide and cGMP‐dependent protein kinase. J. Biol. Chem. 47, 44338–44346. Lauritzen, I., Zanzouri, M., Honore´, E., Duprat, F., Ehrengruber, M. U., Lazdunski, M., and Patel, A. J. (2003). Kþ‐dependent cerebellar granule neuron apoptosis: Role of TASK leak Kþ channels. J. Biol. Chem. 278, 32068–32076. Lesage, F. (2003). Pharmacology of neuronal background potassium channels. Neuropharmacology 44, 1–7. Lesage, F., Guillemare, E., Fink, M., Duprat, F., Lazdunski, M., Romey, G., and Barhanin, J. (1996a). TWIK‐1, a ubiquitous human weakly inward rectifying Kþ channel with a novel structure. EMBO J. 15, 1004–1011. Lesage, F., Reyes, R., Fink, M., Duprat, F., Guillemare, E., and Lazdunski, M. (1996b). Dimerization of TWIK‐1 Kþ channel subunits via a disulfide bridge. EMBO J. 15, 6400–6407. Lesage, F., Lauritzen, I., Duprat, F., Reyes, R., Fink, M., Heurteaux, C., and Lazdunski, M. (1997). The structure, function and distribution of the mouse TWIK‐1 Kþ channel. FEBS Lett. 402, 28–32. Lesage, F., Terrenoire, C., Romey, G., and Lazdunski, M. (2000). Human TREK‐2, a 2P domain mechano‐sensitive Kþ channel with multiple regulations by polyunsaturated fatty acids, lysophospholipids, and Gs, Gi, and Gq protein‐coupled receptors. J. Biol. Chem. 275, 28398–28405. Lopes, C. M., Zhang, H., Rohacs, T., Jin, T., Yang, J., and Logothetis, D. E. (2002). Alterations in conserved Kir channel‐PIP2 interactions underlie channelopathies. Neuron 34, 933–944. Lopes, C. M., Rohacs, T., Czirjak, G., Balla, T., Enyedi, P., and Logothetis, D. E. (2005). PiP2‐ hydrolysis underlies agonist‐induced inhibition and regulates voltage‐gating of 2‐P domain Kþ channels. J. Physiol. 564, 117–129. Loussouarn, G. (2003). Phosphatidylinositol‐4,5‐bisphosphate, PIP2, controls KCNQ1/ KCNE1 voltage‐gated potassium channels: A functional homology between voltage‐gated and inward rectifier Kþ channels. EMBO J. 22, 5412–5421.
170
Chemin et al.
Maingret, F., Fosset, M., Lesage, F., Lazdunski, M., and Honore´, E. (1999a). TRAAK is a mammalian neuronal mechano‐gated Kþ channel. J. Biol. Chem. 274, 1381–1387. Maingret, F., Patel, A. J., Lesage, F., Lazdunski, M., and Honore´, E. (1999b). Mechano‐ or acid stimulation, two interactive modes of activation of the TREK‐1 potassium channel. J. Biol. Chem. 274, 26691–26696. Maingret, F., Lauritzen, I., Patel, A., Heurteaux, C., Reyes, R., Lesage, F., Lazdunski, M., and Honore´, E. (2000a). TREK‐1 is a heat‐activated background Kþ channel. EMBO J. 19, 2483–2491. Maingret, F., Patel, A. J., Lesage, F., Lazdunski, M., and Honore´, E. (2000b). Lysophospholipids open the two P domain mechano‐gated Kþ channels TREK‐1 and TRAAK. J. Biol. Chem. 275, 10128–10133. Maingret, F., Honore´, E., Lazdunski, M., and Patel, A. J. (2002). Molecular basis of the voltage‐dependent gating of TREK‐1, a mechano‐sensitive Kþ channel. Biochem. Biophys. Res. Commun. 292, 339–346. Maylie, J., and Adelman, J. P. (2001). Beam me up, Scottie! TREK channels swing both ways. Nat. Neurosci. 4, 457–458. Medhurst, A. D., Rennie, G., Chapman, C. G., Meadows, H., Duckworth, M. D., Kelsell, R. E., Gloger, I. I., and Pangalos, M. N. (2001). Distribution analysis of human two pore domain potassium channels in tissues of the central nervous system and periphery. Mol. Brain Res. 86, 101–114. Mirshahi, T., Jin, T., and Logothetis, D. E. (2003). G beta gamma and KACh: Old story, new insights. Sci. STKE 2003, PE32. Murbartian, J., Lei, Q., Sando, J. J., and Bayliss, D. A. (2005). Sequential phosphorylation mediates receptor‐ and kinase‐induced inhibition of TREK‐1 background potassium channels. J. Biol. Chem. 280, 30175–30184. Oliver, D., Lien, C. C., Soom, M., Baukrowitz, T., Jonas, P., and Fakler, B. (2004). Functional conversion between A‐type and delayed rectifier Kþ channels by membrane lipids. Science 304, 265–270. Patel, A. J., and Honore´, E. (2001). Properties and modulation of mammalian 2P domain Kþ channels. Trends Neurosci. 24, 339–346. Patel, A. J., Honore´, E., Maingret, F., Lesage, F., Fink, M., Duprat, F., and Lazdunski, M. (1998). A mammalian two pore domain mechano‐gated S‐like Kþ channel. EMBO J. 17, 4283–4290. Patel, A. J., Lazdunski, M., and Honore´, E. (2001). Lipid and mechano‐gated 2P domain Kþ channels. Curr. Opin. Cell Biol. 13, 422–428. Runnels, L. W., Yue, L., and Clapham, D. E. (2002). The TRPM7 channel is inactivated by PIP2 hydrolysis. Nat. Cell Biol. 4, 329–336. Sheetz, M. P., and Singer, S. J. (1974). Biological membranes as bilayer couples. A molecular mechanism of drug‐erythocyte interactions. Proc. Natl. Acad. Sci. USA 71, 4457–4461. Shyng, S. L., and Nichols, C. G. (1998). Membrane phospholipid control of nucleotide sensitivity of KATP channels. Science 282, 1138–1141. Suchyna, T. M., Tape, S. E., Koeppe, R. E., II, Andersen, O. S., Sachs, F., and Gottlieb, P. A. (2004). Bilayer‐dependent inhibition of mechanosensitive channels by neuroactive peptide enantiomers. Nature 430, 235–240. Suh, B. C., and Hille, B. (2002). Recovery from muscarinic modulation of M current channels requires phosphatidylinositol 4,5‐bisphosphate synthesis. Neuron 35, 507–520. Voets, T., and Nilius, B. (2003). TRPs make sense. J. Membr. Biol. 192, 1–8. Zhang, H., Craciun, L. C., Mirshahi, T., Rohacs, T., Lopes, C. M., Jin, T., and Logothetis, D. E. (2003). PIP2 activates KCNQ channels, and its hydrolysis underlies receptor‐ mediated inhibition of M currents. Neuron 37, 963–975.
CHAPTER 8 MechanoTRPs and TRPA1 Andrew J. Castiglioni and Jaime Garcı´a‐An˜overos Departments of Anesthesiology, Physiology, and Neurology, Northwestern University Institute for Neuroscience, Feinberg School of Medicine, Northwestern University, Chicago, Illinois 60611
I. II. III. IV.
Overview MechanoTRP Channels Characteristics of TRPA1 Gene and Protein TRPA1 Expression in Mechanosensory Organs A. Somatosensory Neurons B. Inner Ear V. Function of TRPA1 A. Nociception B. Auditory and Vestibular C. Channel Similarities Between Heterologously Expressed TRPA1 and Endogenous Mechanotransducers VI. Proposed Biological Roles for TRPA1 References
I. OVERVIEW Genetic and molecular searches in animals identify two families of ion channels used by specialized mechanosensory cells (Duggan et al., 2000). These are Deg/ENaC channels (reviewed by Corey and Garcı´a‐An˜overos, 1996; Garcı´a‐An˜overos and Corey, 1997; Ernstrom and Chalfie, 2002; and elsewhere in this volume) and TRP channels (reviewed in this and other chapters of this volume). Some of these channels open in response to mechanical forces and/ or mediate cellular responses to mechanical stimulation (Table I). TRPA1 is expressed in nociceptive neurons of peripheral ganglia and in the sensory Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59008-8
172 TABLE I TRP Channels Implicated in Mechanical Sensitivity Channel
Mechanical sensitivity
Evidence
References
TRPY TRPY1
Osmotic swelling; pipette pressure
Yeast mutant phenotype
Zhou et al., 2003
TRPV1
Response to bladder filling (bladder stretch)
Mouse mutant phenotype
Birder et al., 2002
TRPV2
Osmotic swelling of cell; pipette pressure on patch
Antisense oligo eVect; functional expression
Muraki et al., 2003
TRPV4
Osmotic swelling of cell; touch
Functional expression; mouse mutant phenotype
Liedtke et al., 2000; Strotmann et al., 2000; Mizuno et al., 2003; Suzuki et al., 2003; Vriens et al., 2004
OSM‐9
Osmotic shock; touch
Worm mutant phenotype
Colbert et al., 1997
NAN
Auditory
Fly mutant phenotype
Kim et al., 2003
IAV
Auditory
Fly mutant phenotype
Gong et al., 2004
Fluid flow in epithelia (cilia deflection)
Mouse mutant phenotype
Pennekamp et al., 2002; McGrath et al., 2003; Nauli et al., 2003
TRPV
TRPP PKD1 and PKD2 (TRPP2)
TRPC Stretch; membrane tension; pipette pressure
Functional expression; siRNA eVect
Maroto et al., 2005
dmTRPN1
Touch and proprioception (bristle deflection)
Fly mutant phenotype; altered transient receptor currents on deflection
Walker et al., 2000
drTRPN1
Auditory and vestibular (hair cells)
Morpholino treatment alters extracellular receptor potentials in zebrafish hair cells
Sidi et al., 2003
ceTRPN1 (TRP‐4)
Proprioception (body bending)
Worm mutant phenotype
Li et al., 2006
TRPC1 TRPN
TRPA Painless
Mechanothermal (heated prodding)
Fly larvae mutant phenotype
Tracey et al., 2003
drTRPA1
Lateral line and vestibular (hair cells)
Morpholino treatment alters extracellular receptor potentials in zebrafish hair cells
Corey et al., 2004
mmTRPA1
Mechanical nociception
Mouse mutant phenotype; functional expression; expression pattern
Corey et al., 2004; Nagata et al., 2005; Kwan et al., 2006
Hair cells
Mouse mutant is deaf
Di Palma et al., 2002
TRPML TRPML3
173
174
Castiglioni and Garcı´a‐An˜overos
epithelia of the inner ear. In nociceptors, TRPA1 forms chemosensitive channels that mediate the response to exogenous pain‐producing chemicals as well as to the endogenous proalgesic bradykinin (BK). More indirect evidence suggests that TRPA1 might also form mechanosensory channels.
II. MECHANOTRP CHANNELS Some of the TRP channels that mediate mechanical responses are not necessarily mechanically gated. For example, TRPV4 mutant mice have reduced sensitivity to noxious tactile stimulation, and heterologously expressed TRPV4 opens in response to hypotonic solution (which induces cell swelling and thus stretches membranes). However, this response appears to be mediated by lipid second messengers gating TRPV4 (Liedtke et al., 2000; Strotmann et al., 2000; Mizuno et al., 2003; Suzuki et al., 2003; Vriens et al., 2004). In addition, the capsaicin receptor TRPV1 is expressed in bladder epithelium, which is less sensitive to stretch in TRPV1 mutant mice (Birder et al., 2002). However, heterologously expressed TRPV1 and TRPV4 have not been shown to be activated by membrane stretch. Rather, they have been shown to be activated by nonmechanical stimuli like hot temperature and extracellular acidity. Other TRP channels have been shown to be activated directly by membrane stretch elicited in membrane patches by pipette suction. TRPY1 of yeast has not been heterologously expressed, but TRPY mutants lack a stretch‐activated and osmosensitive channel in their vacuoles (Zhou et al., 2003). The vertebrate TRPC1 generates stretch‐sensitive channels when heterologously expressed (Maroto et al., 2005). In addition, gene overexpression and inhibition studies suggest it to be the long‐sought stretch‐activated channel of frog oocytes. Although mechanosensitive cation channels like the one expressed in the oocyte have been implicated in cell‐volume regulation, cell locomotion, muscle dystrophy, and cardiac arrhythmias, there is so far no data supporting a sensory function for the mechanosensitivity of TRPC1 channels. Finally, a large number of TRP channels, primarily in flies and worms, are required for the specialized sensory neurons that mediate auditory, proprioceptive, tactile, and nociceptive mechanotransduction. One of these genes encodes TRPN1 (originally called NompC), which is expressed in the bristle organs of flies, in the lateral line hair cells of zebrafish, as well as in ciliated dopaminergic mechanosensory neurons (ADEs and CEPs) and in stretch‐activated neurons (DVA) of worms (Walker et al., 2000; Sidi et al., 2003; Li et al., 2006). When the TRPN1 gene is mutated (in flies and worms), the mechanoreceptor currents of these sensory cells are eliminated or reduced. Furthermore, a point mutation in the fly TRPN1 alters the speed of mechanoreceptor current adaptation,
8. MechanoTRPs and TRPA1
175
strongly suggesting that TRPN1 is the mechanically gated channel in these cells. In zebrafish, TRPN1 mRNA was detected in hair cells of the inner ear and its inhibition with morpholinos produced deaf and unbalanced fish whose hair cells failed to generate mechanoreceptor potentials. A very similar eVect was obtained by morpholinos against a structurally analogous protein, TRPA1 (Corey et al., 2004). It is likely that TRPN1, perhaps in association with other proteins like TRPA1, contributes to the mechanosensitive channel in zebrafish hair cells. However, mammals have hair cells but their genomes do not appear to have an orthologue of TRPN1. Other proteins must constitute mechanosensory channels in mammals. A search for TRP channel expression in specialized mechanosensory organs revealed several candidate genes expressed in dorsal root ganglia, trigeminal ganglia, or inner ear. One of them is TRPA1 (also known as ANKTM1 and P120), which is absent from (or very weakly expressed in) most organs yet is expressed in both somatosensory neurons and sensory epithelia of inner ear. Although mice with mutations in TRPA1 have no auditory or vestibular defects reported, they do have nociceptive phenotypes, some of which may be attributed to defective mechanotransduction.
III. CHARACTERISTICS OF TRPA1 GENE AND PROTEIN TRPA1 genes in mammals are large, occupy around 50 kb of chromosomal DNA, and are encoded by at least 27 exons. In humans, the TRPA1 gene is located on chromosome 8q13 (contig NT_008183.18; June, 2006). In mice, the TRPA1 gene is located on chromosome 1A3. TRPA1 homologues have been found in many species including human, rat, mouse, zebrafish, puVerfish, sea squirt, fly, mosquito, and nematode, which suggests an evolutionary conservation of TRPA1 function. TRPA1 protein is predicted to have canonical TRP structure in its pore‐forming region with six transmembrane domains (S1–S6), a putative pore loop between S5 and S6, together with cytoplasmic N‐ and C‐termini. The most distinguishing feature of TRPA1 is a long N‐terminal region containing up to 18 predicted ankyrin repeat motifs (the exact number depending on the consensus definition used by the computer prediction). Ankyrin repeats help to cluster and organize ion channels and receptors at specialized regions of cells (Bennett and Chen, 2001; Lee et al., 2006); however, ankyrin repeat domains may do more than clustering and subcellular organization. Force measurements demonstrate that ankyrin repeats are elastic and can constitute molecular springs (Lee et al., 2006), supporting the hypothesis that ankyrin repeats of TRPN1 and other TRP channels form mechanical gating springs (Howard and Bechstedt, 2004; Sotomayor et al., 2005). Eighteen is the greatest number of ankyrin repeats known in a mammalian TRP channel and helps to explain TRPA1’s predicted
176
Castiglioni and Garcı´a‐An˜overos
molecular weight of 127.4 kD. Other TRPs possess between 0 and 8 ankyrin repeats, with the notable exception of TRPN1 (NompC) which has 29 predicted ankyrin repeats and is found in flies (Walker et al., 2000), nematodes (Walker et al., 2000; Li et al., 2006), and fish (Sidi et al., 2003). Intriguingly, mammalian genomes do not appear to encode a homologue of TRPN1, begging the question of which mammalian protein may play the mechanosensory roles that TRPN1 has in nematodes, flies, and fish. TRPA1 has been proposed as a candidate because of its structural similarity to TRPN1 and its expression pattern in mechanosensory organs (Corey et al., 2004; Nagata et al., 2005).
IV. TRPA1 EXPRESSION IN MECHANOSENSORY ORGANS Initial Northern blot analysis showed TRPA1 mRNA present at very low levels in human fibroblasts but this expression was lost on oncogenic transformation (Jaquemar et al., 1999). Thus far, the only tissues expressing detectable TRPA1 mRNA and protein are those specialized in sensation: somatosensory neurons in the peripheral nervous system and mechanosensory epithelia of the inner ear. Northern blot and in situ hybridization did not detect TRPA1 mRNA in most organs of late embryonic or adult mice including brain, heart, liver, kidney, skeletal muscle, lung, spleen, testis, whisker pad skin, and superior cervical ganglia (Jaquemar et al., 1999; Nagata et al., 2005). Lack of detection suggests lack of TRPA1 expression; however, these findings do not rule out the possibility that TRPA1 mRNA is present below detection levels. For example, TRPA1 could be expressed in restricted regions of the brain. In support of this localized or low‐level expression, five cDNA clones of mouse TRPA1 were derived from mesencephalon (see the UniGene cluster Mm.186329).
A. Somatosensory Neurons In dorsal root, trigeminal, and nodose ganglia of mouse and rat, in situ hybridization detected TRPA1 mRNA exclusively in small diameter neurons, most of which are nociceptors. Neurons expressing TRPA1 constitute a sizeable percentage (between 20% and 56%) of all neurons in a ganglia (Jaquemar et al., 1999; Kobayashi et al., 2005; Nagata et al., 2005). Precise estimates on the proportion of TRPA1‐expressing neurons per ganglion vary from study to study, although this might largely be attributed to diVerences in species, strain, age, type of ganglia, sensitivity, and perhaps the nociceptive history of the ganglion. Prior nociception may aVect the proportion of
8. MechanoTRPs and TRPA1
177
neurons that express TRPA1 (Obata et al., 2005). TRPA1‐expressing peripheral neurons do not coexpress neurofilament 200 or the growth factor receptors TrkC and TrkB, but do express the NGF receptor TrkA, a marker of nociceptive neurons (Bautista et al., 2005; Nagata et al., 2005). Western blots confirm TRPA1 protein expression in peripheral ganglia and their nerves (i.e., sciatic nerve; Nagata et al., 2005). Using antibodies raised against either the mouse N‐terminus (Nagata et al., 2005) or the rat C‐terminus (Bautista et al., 2005), most TRPA1‐expressing neurons are reported to also express other markers of nociceptive neurons, including the capsaicin receptor (TRPV1), calcitonin gene‐related peptide (CGRP), substance P (SP), and peripherin (Bautista et al., 2005; Kobayashi et al., 2005; Nagata et al., 2005). These immunoreactivities also demonstrate the presence of TRPA1 protein in the peripheral nerve endings of target organs, such as the bladder epithelium and the cornea, where sensation initially occurs. All these experiments confirm that small diameter C‐fiber nociceptors, not large diameter A‐fiber innocuous mechanoreceptor neurons, express TRPA1.
B. Inner Ear Analyses of mRNA by RT‐PCR and in situ hybridization as well as analyses of protein by Western blot and immunohistochemistry also detected TRPA1 mRNA in vestibular and auditory sensory epithelia (Corey et al., 2004; Nagata et al., 2005). In neonatal organ of Corti, in situ hybridization revealed clear expression in supporting cells which is confirmed by immunohistochemistry with antibodies to both the N‐terminus (Nagata et al., 2005; our unpublished results) and C‐terminus (Corey et al., 2004) of mouse TRPA1. Weak immunoreactivities were also detected in the sensory hair cells apically at the cuticular plates and in the mechanosensory stereocilia. Important functional evidence that TRPA1 is present in hair cells was provided with a report that TRPA1‐agonists allyl isothiocyanate (AITC) and icilin (AG 3‐5) activate currents in hair cells (Stepanyan et al., 2006). Although we are not sure how or if these agonists specifically activate TRPA1, at present we know no other ion channel activated by both AITC and AG 3‐5.
V. FUNCTION OF TRPA1 A. Nociception Mice with a deletion that eliminates the pore domain of TRPA1 have impaired nociception. TRPA1 mutant mice show reduced or eliminated nocifensive responses to pain‐producing TRPA1‐agonists like mustard oil,
178
Castiglioni and Garcı´a‐An˜overos A Withdrawal response (%)
60
40
von Frey +/+ +/− −/−
20
0
0.01 0.02 0.04 0.07 0.16 0.40 0.60 1.0 1.4 2.0 Force (g)
B
BK sensitization
Withdrawal threshold (g)
1.0 0.8 0.6 0.4 0.2 0
+/+
+/− Pre
−/−
+/+
+/− −/− 2-hr-post
C Nerve injury
Withdrawal threshold (g)
1.0 0.8
*
0.6 0.4 0.2 0 /
+/+ +/− −/− Baseline þ/
Day 7
Day 14
FIGURE 1 TRPA1 and TRPA1 mice have reduced responses to mechanical stimuli. Loss of TRPA1 diVerentially protects from development of hypersensitivity. (A) Plantar withdrawal response measured with calibrated von Frey hairs for 10 presentations at each force. Fourteen animals (seven males and seven females) of each genotype were tested; there was no significant diVerence between sexes, so results were pooled. Mean SEM. (B) Plantar
8. MechanoTRPs and TRPA1
179
allicin, and acrolein (and their trigeminal neurons, in culture, do not uptake Ca2þ in response to these agonists), whereas these mutant animals do respond nocifensively (and their trigeminal neurons uptake Ca2þ) to capsaicin, a pungent that is not a TRPA1 agonist (Bautista et al., 2006; Kwan et al., 2006). Clearly TRPA1 mediates the response of nociceptors to certain pungent chemicals. These compounds, however, are not endogenous and are not normally in contact with most TRPA1‐expressing nociceptive neurons. Typically, we encounter these chemicals when we eat them and they mainly contact the nerves innervating the epithelia lining the buccal cavity. Although it is possible that there are similar substances endogenously produced by tissue damage or inflammation, these have not been reported. It is also possible that these compounds are agonists that bear no structural relationship with the nociceptive signals that under normal physiological conditions activate TRPA1, like BK. TRPA1 is also a mediator of BK proalgesic eVects (Bandell et al., 2004; Bautista et al., 2006). Not only do TRPA1 channels open if coexpressed with the BK receptor and exposed to BK, but TRPA1 mutant mice do not develop hyperalgesia when exposed to BK. As the BK receptor activates phospholipase C, and this enzyme also mediates the eVects of other proalgesics like ATP, monoamines, and neurotrophins, TRPA1 channels may also mediate their eVects. Surprisingly, induction of hyperalgesia by BK requires both TRPA1 and TRPV1 (Bautista et al., 2006). Therefore, TRPV1 has two diVerent roles: (1) to mediate BK‐induced hyperalgesia in concert with TRPA1 and (2) to transduce noxious thermal and acidic stimuli. By analogy, TRPA1 might also play a role as a primary sensory transducer in addition to mediating the action of proalgesics. Indeed, one of the reports on TRPA1 knockout mice reveals that they have a reduced response to punctate mechanical stimulation. Knockout mice and heterozygous mice display significantly lower withdrawal responses and elevated withdrawal thresholds as compared to wild‐type mice (Kwan et al., 2006; Fig. 1). This impairment of behavioral response to mechanonociceptive stimuli is consistent with a role of TRPA1 in mechanical transduction, although further tests will be necessary to determine this. For example, nerve recordings showing a reduction in mechanically induced firing rates from nociceptors of TRPA1 mutant mice would further the argument that TRPA1 withdrawal threshold measured before and 2 h after injection of BK into footpad. Bradykinin increased the sensitivity of wild‐type mice to von Frey stimulation, reducing thresholds by fivefold, but in knockout mice, thresholds were not significantly reduced. Thirteen animals of each genotype were tested. (C) Plantar withdrawal thresholds measured before and at 7 and 14 days after spared nerve injury. Knockout mice showed significantly higher thresholds than wild type (*, p ¼ 0.03) before nerve injury. Figure and legend adapted from Kwan et al. 2006. Copyright 2006 by Elsevier.
180
Castiglioni and Garcı´a‐An˜overos
is acting in these cells as a mechanotransducer. This is especially important given the potential for TRPA1 expression in the CNS previously mentioned. Still, to demonstrate that TRPA1 channels are mechanically gated we would need (1) recreation of mechanical gating of TRPA1 in heterologous cells and (2) demonstration of impaired mechanoreceptive currents in TRPA1 knockout neurons. Neither experiment has been reported yet, largely because of the prevailing thought that, although some mechanically gated channels can be opened by direct pressure applied through the recording pipette, those channels used by specialized mechanosensory cells are part of macromolecular complexes where the channel physically interacts with structural components that transmit the gating force (Garcı´a‐An˜overos and Corey, 1997). In this context, receptor potentials would have to be recorded from the extremely fine free nerve endings of nociceptors embedded in other tissues, something never performed. Heterologous expression would necessitate other proteins, which are presently unknown.
B. Auditory and Vestibular To date, there are no reports of mechanical activation of heterologously expressed TRPA1, although there are reasons to suspect that mechanical activation might occur in the appropriate cellular context. Mice with a deletion of the pore domain of TRPA1 have no overt vestibular or auditory defect, as determined by behavioral observation, auditory brainstem response thresholds, and distortion product otoacoustic emissions (Bautista et al., 2006; Kwan et al., 2006). In addition, the hair cell transduction currents in wild‐type mice and mice with a deletion of the TRPA1 pore are indistinguishable (Kwan et al., 2006). Importantly, both strains of knockout mice delete a small part of the TRPA1 gene, so it will be critical to determine if they are functional nulls. If not, these mice could make a truncated TRPA1 protein bearing the N‐terminal ankyrin repeats and the initial transmembrane domains. Such a truncated protein would not form a channel by itself, but it could function as part of a multimeric complex. Data obtained prior to knockout studies indicated that acute inhibition of TRPA1 mRNA using viral‐mediated siRNA in mouse utricular hair cells and morpholinos in zebrafish hair cells reduced but did not eliminate the mechanotransduction currents of these cells (Corey et al., 2004). These apparently contradictory results could be reconciled if TRPA1 participates in, but its pore domain is not essential for, hair cell mechanotransduction complexes. An alternative explanation is that TRPA1 plays no role in hair cells but rather functions in support cells, where its expression is most prominent (our unpublished data). Presently, the role of TRPA1 in the inner ear is unclear.
8. MechanoTRPs and TRPA1
181
C. Channel Similarities Between Heterologously Expressed TRPA1 and Endogenous Mechanotransducers Without a clear mechanical stimulus paradigm for heterologously expressed TRPA1, these channels have primarily been studied by activation with agonists like AITC (Nagata et al., 2005). This permitted a comparison of TRPA1 with other mechanosensory channels, primarily those of hair cells (unfortunately, for reasons stated above, there is little information regarding the mechanotransducer channels in nociceptors). Although data obtained from two knockout analyses of TRPA1 suggests that TRPA1 does not play a crucial role in hair cell mechanotransduction, there is pharmacological evidence that the pore formed by TRPA1 is similar to the pore of mechanotransduction channels, including the one in hair cells. While numerous pain‐producing chemicals act as TRPA1 agonists (Bandell et al., 2004; Jordt et al., 2004; Bautista et al., 2005, 2006; Nagata et al., 2005), four antagonists have also been studied: gentamycin, ruthenium red, gadolinium (Gd3þ), and amiloride (Nagata et al., 2005; Table II). Although each of these antagonists blocks other types of channels, block by all four is characteristic of mechanosensory channels from various cell types, including hair cells (Hamill and McBride, 1996). All four antagonists show similar Hill coeYcients of block between heterologously expressed TRPA1 and the hair cell mechanotransducer. This suggests a shared mechanism of action on TRPA1 channels and the hair cell mechanotransducer. Further, IC50 values for gentamycin and ruthenium red (both simple pore blockers) are also indistinguishable between heterologously expressed TRPA1 and the hair cell mechanotransducer. However, the IC50 values for amiloride and Gd3þ diVer by factors of 10 and 100, respectively. On the basis of these data, Gd3þ is a more potent blocker of heterologously expressed TRPA1 than of any other channel, suggesting a potential analgesic use (Nagata et al., 2005). Other studies of heterologously expressed TRPA1 revealed that these channels are permeable to both monovalent and divalent cations, although the ionic selectivity has not been calculated in detail. The estimated conductance of a single TRPA1 channel in the negative voltage range was 100 pS, but traces with estimated values as low as 38 pS were observed. In fact, single‐channel current amplitude varied widely, even within one sweep of single‐channel openings, suggesting that TRPA1 channels have a range of conductance levels. For example, addition of external Ca2þ decreased the amplitude of single TRPA1 channel inward current to 54% of the level in the absence of Ca2þ (Nagata et al., 2005). Curiously, the conductance of the hair cell transducer is similar in magnitude (although it also varies, as it ranges tonotopically along the cochlea) and it is reduced by Ca2þ to the same extent (Table II).
TABLE II Properties of Heterologous TRPA1 Channels Compared to Mechanotransduction Channels of Hair Cells
Channel property
Heterologous TRPA1 (Nagata et al., 2005)
Hair cell transducers
References for hair cell transducers
Gentamycin block (IC50)
6.7 0.7 mM
7.6 mM
Kroese et al., 1989
Gentamycin block (Hill coeYcient)
1.2
1.2
Kroese et al., 1989
Ruthenium red block (IC50)
3.4 0.1 mM
3.6 0.3 mM
Farris et al., 2004
Ruthenium red block (Hill coeYcient)
1.1 0.03
1.4 0.2
Farris et al., 2004
Gd3þ block (IC50)
0.1 0.02 mM
10.1 mM
Kimitsuki et al., 1996
Gd3þ block (Hill coeYcient)
1.2 0.25
1.1
Kimitsuki et al., 1996
Amiloride block (IC50)
511 12 mM
50 mM
Jorgensen and Ohmori, 1988; Rusch et al., 1994
Amiloride block (Hill coeYcient)
2.4 0.1
2.2 0.1
Ricci, 2002
Ca ‐induced potentiation
Yes
Yes
Ricci et al., 2003
Ca2þ‐induced closure (adaptation)
Yes (t ¼ 3.5 s)
Yes (t ¼ few millisecond)
Howard and Hudspeth, 1988; Ricci et al., 1998; Kennedy et al., 2003
Reopening by depolarization after Ca2þ‐induced closure
Yes
Yes
Ricci et al., 2000
Reduction of unitary conductance by Ca2þ
53.7 2.0%
54%
Ricci et al., 2003
2þ
Heterologous TRPA1 channels were expressed in HEK293 cells and opened by AITC. Hair cell transducers were opened by mechanical force.
8. MechanoTRPs and TRPA1
183
Indeed, Ca2þ exerts similar pronounced eVects on the gating of both TRPA1 and the mechanotransducer of hair cells (Table II). Whole‐cell recordings of AITC‐activated currents in HEK293 cells transiently expressing TRPA1, voltage‐clamped at 80 mV, and in the presence of physiological external Ca2þ reveal a multiphasic inward current. This current initially turns on slowly, but potentiates further before inactivating. In the absence of external Ca2þ, the slowly developing inward currents do not potentiate or inactivate as they do in the presence of Ca2þ. However, if Ca2þ is suddenly added to the bath solution when the channels are already opened by AITC, the currents potentiate and then inactivate (Fig. 2). The Ca2þ‐induced phenomena are also voltage dependent. At more depolarized potentials ( 20 mV in Fig. 2A), inactivation of channels (activated in the continuous presence of 2 mM external Ca2þ) does not occur as it does at 80 mV. At þ80 mV, the potentiation and inactivation induced by sudden exposure to external Ca2þ at 80 mV are absent or much reduced. In addition, channels inactivated at 80 mV will reopen if held at þ80 mV. It appears that Ca2þ must enter the cell to exert its eVects on TRPA1 channels. However, these eVects occurred whether the cytoplasm of the cells was loaded with Ca2þ (up to 3 mM) or with Ca2þ chelators (EGTA or BAPTA). Therefore, it is unlikely that Ca2þ is exerting these eVects through the cytoplasm. Single channel recordings from outside‐out patches of cells heterologously expressing TRPA1 further reveal the eVects of Ca2þ on channel conductance and open probability (Po). In the presence of AITC, but no external Ca2þ, heterologous TRPA1 channels adopt a high conductance (5.7 0.4 pA) and low Po state (II of Fig. 2B). On external Ca2þ addition, the channel conductance reduces (3.1 0.4 pA) but Po increases (potentiation; III in Fig. 2B) for several seconds before the Po drops essentially to zero (inactivation; IV in Fig. 2B). It seems that entering Ca2þ binds to the channel and causes a brief potentiation (rise in Po) followed by closure (adaptation or inactivation; Fig. 2B and C). Interestingly, similar eVects of Ca2þ have been reported for the mechanotransduction channel of hair cells, albeit at a rate 1000 faster (Table II). This quantitative diVerence may be accounted for by a diVerence in aYnity of the Ca2þ‐binding site, which could result from small molecular diVerences in the channel proteins. Nonetheless, the same mechanistic model, presented on Fig. 2C, accounts for the behavior of both TRPA1 and hair cell mechanotransduction channels. Although TRPA1 is not essential for hair cell transduction, the above‐mentioned similarities suggest that (1) the mechanotransducer of hair cells is similar to TRPA1 and (2) TRPA1 may form a mechanosensory channel in other cells where it is expressed, namely supporting cells of inner ear and somatosensory nociceptors.
184
Castiglioni and Garcı´a‐An˜overos A
10-µM AITC, 2-mM [Ca2+]0
200 pA 50 s −20 mV
−80 mV
B 10-µM AITC, No [Ca2+]0 I
+ 2-mM [Ca2+]0
IV
II 500 pA
20 s
III Control C O1 O2
I 10-µM AITC, No [Ca2+]0
C
II
O1 O2
+2-m [Ca2+]0
C O1 O2
III +2-m [Ca2+]0
C O1 O2
C Ca
IV 10 pA 200 ms K, Na I
II
Ca2+-binding site
III
IV
*(+80 mV)
8. MechanoTRPs and TRPA1
185
VI. PROPOSED BIOLOGICAL ROLES FOR TRPA1 The highly restricted expression of TRPA1 in nociceptive neurons of peripheral ganglia and in mechanosensory epithelia of the inner ear strongly suggests sensory roles. The activation of heterologously expressed TRPA1 by various pain‐producing chemicals (Bandell et al., 2004; Jordt et al., 2004; Bautista et al., 2005; Nagata et al., 2005; Bautista et al., 2006), together with the impaired nociception of mice bearing engineered deletions in TRPA1, demonstrate a role in nociception. The voltage sensitivity of TRPA1’s Ca2þ‐dependent inactivation renders it well suited for a role in pain (Nagata et al., 2005). In a sensory terminal insuYciently depolarized by low threshold (i.e., innocuous) stimulation, any TRPA1 channels that open would quickly inactivate. But with stimulation above threshold (i.e., noxious), enough TRPA1 channels would open and depolarize the terminal. Under conditions of sustained depolarization, TRPA1 channels inactivate incompletely and slowly compared to conditions of weak or no depolarization. Thus, TRPA1 may have the inherent abilities to (1) distinguish between innocuous (even if persistent) stimuli and noxious stimuli and (2) respond continuously to prolonged noxious stimulation by failing to desensitize. These two abilities are characteristic of nociceptive sensation. In nociceptors, TRPA1 responds to pain‐producing chemicals, mediating both their acute noxious eVects as well as the sensitization that these chemicals generate. These compounds are either synthetic or plant derived and they are not known to be produced endogenously. TRPA1 is also responsive to the proalgesic BK, which activates TRPA1 via its G‐protein–coupled receptor. FIGURE 2 Biophysical properties of heterologously expressed TRPA1 channels. (A) When opened by AITC in the presence of external Ca2þ, TRPA1 channels inactivate completely at strongly hyperpolarized ( 80 mV) holding potentials but not at more depolarized ( 20 mV) holding potentials. (B) Recapitulation of TRPA1 slow activation, Ca2þ‐induced potentiation, and subsequent inactivation at the single‐channel level recorded from outside‐out patches (at a holding potential of 80 mV). In the absence of Ca2þ, AITC activates a large conductance, flickery channel (II, activation). On exposure to extracellular Ca2þ, this channel transitions to a low conductance but high Po state (III, potentiation) followed by a low Po state (IV, inactivation). The upper trace is a whole‐cell record, obtained just before patch formation, indicating the general times that correspond to the single‐channel traces. (C) A model for how Ca2þ and voltage aVects TRPA1 channels, which can also describe similar behaviors of hair cell transducers. (I) At rest, the channel opens spontaneously. (II) AITC‐activated channels, with no external Ca2þ, have a high unitary conductance, but a low Po. (III) AITC‐activated channels in the presence of external Ca2þ have a low unitary conductance, but a high Po. (IV) Prolonged exposure to external Ca2þ causes channels to enter very low Po state. (*) This inactivation can be relieved by strong depolarization (þ80 mV) by an unknown mechanism. Figure and legend adapted from Nagata et al., 2005. Copyright 2005 by Society for Neuroscience.
186
Castiglioni and Garcı´a‐An˜overos
Clearly, a biological role for TRPA1 is to mediate the proalgesic eVects of BK, and perhaps of other proalgesics, in conjunction with TRPV1. The question remains whether TRPA1 has, like TRPV1, a role as a sensory transducer of other pain stimuli, like extreme temperature or physical deformation. Taken together, (1) the similar pharmacology and pore properties between heterologously expressed TRPA1 and hair cell mechanotransducers, (2) the structural or phylogenetic similarity of TRPA1 with mechanosensory channels from invertebrates (TRPN1 and Painless), (3) the wide distribution of TRPA1 mRNA among nociceptors (many of which are mechanosensory), and (4) the elevated withdrawal thresholds to punctate mechanical stimulation of one of the TRPA1 knockout mouse lines suggest a role for TRPA1 in mechanonociception. However, it should be noted that the elevated withdrawal thresholds were not detected in both strains of TRPA1 knockout mice. Further experimentation will be necessary. What exactly TRPA1 does in the inner ear is a very intriguing question. Although not stated on current publications, the expression of TRPA1 is weak in hair cells and more prominent in the adjacent supporting cells. It is possible that TRPA1 has a role in supporting cells and that its low‐level presence in hair cells is an accident of their location and developmental history (both hair and support cells come from the same precursors). Alternatively, TRPA1 may play a nontransduction role in hair cells that is not essential to hearing. A role in mechanotransduction is not completely eliminated by the current data, although one would have to assume that TRPA1 is acting redundantly. Hair cell transduction channels appear functionally heterogeneous, with tens to hundreds of conductance levels arranged tonotopically (Ricci et al., 2003). We think that molecular diversity may underlie this functional diversity. Heterogeneity could be accomplished by multimerization of several channel subunits, perhaps including TRPA1 and/or other TRP proteins that we and others find in hair cells. An interesting candidate in this regard is TRPML3, which is also present in hair cell stereocilia. Mutations in TRPML3 produce stereocilia malformations and deafness in mice (Di Palma et al., 2002). Unfortunately, these mutations act in a dominant fashion, so they do not clarify the function of TRPML3 in hair cells. It will be interesting to determine if conditionally induced loss of function mutations in TRPML3 specifically impair hair cell mechanotransduction. References Bandell, M., Story, G. M., Hwang, S. W., Viswanath, V., Eid, S. R., Petrus, M. J., Earley, T. J., and Patapoutian, A. (2004). Noxious cold ion channel TRPA1 is activated by pungent compounds and bradykinin. Neuron 41, 849–857. Bautista, D. M., Movahed, P., Hinman, A., Axelsson, H. E., Sterner, O., Hogestatt, E. D., Julius, D., Jordt, S. E., and Zygmunt, P. M. (2005). Pungent products from garlic activate the sensory ion channel TRPA1. Proc. Natl. Acad. Sci. USA 102, 12248–12252.
8. MechanoTRPs and TRPA1
187
Bautista, D. M., Jordt, S. E., Nikai, T., Tsuruda, P. R., Read, A. J., Poblete, J., Yamoah, E. N., Basbaum, A. I., and Julius, D. (2006). TRPA1 mediates the inflammatory actions of environmental irritants and proalgesic agents. Cell 124, 1269–1282. Bennett, V., and Chen, L. (2001). Ankyrins and cellular targeting of diverse membrane proteins to physiological sites. Curr. Opin. Cell Biol. 13, 61–67. Birder, L. A., Nakamura, Y., Kiss, S., Nealen, M. L., Barrick, S., Kanai, A. J., Wang, E., Ruiz, G., De Groat, W. C., Apodaca, G., Watkins, S., and Caterina, M. J. (2002). Altered urinary bladder function in mice lacking the vanilloid receptor TRPV1. Nat. Neurosci. 5, 856–860. Colbert, H. A., Smith, T. L., and Bargmann, C. I. (1997). OSM‐9, a novel protein with structural similarity to channels, is required for olfaction, mechanosensation, and olfactory adaptation in Caenorhabditis elegans. J. Neurosci. 17, 8259–8269. Corey, D. P., and Garcı´a‐An˜overos, J. (1996). Mechanosensation and the DEG/ENaC ion channels. Science 273, 323–324. Corey, D. P., Garcı´a‐An˜overos, J., Holt, J. R., Kwan, K. Y., Lin, S. Y., Vollrath, M. A., Amalfitano, A., Cheung, E. L., Derfler, B. H., Duggan, A., Geleoc, G. S., Gray, P. A., et al. (2004). TRPA1 is a candidate for the mechanosensitive transduction channel of vertebrate hair cells. Nature 432, 723–730. Di Palma, F., Belyantseva, I. A., Kim, H. J., Vogt, T. F., Kachar, B., and Noben‐Trauth, K. (2002). Mutations in Mcoln3 associated with deafness and pigmentation defects in varitint‐ waddler (Va) mice. Proc. Natl. Acad. Sci. USA 99, 14994–14999. Duggan, A., Garcı´a‐An˜overos, J., and Corey, D. P. (2000). Insect mechanoreception: What a long, strange TRP it’s been. Curr. Biol. 10, R384–R387. Ernstrom, G. G., and Chalfie, M. (2002). Genetics of sensory mechanotransduction. Annu. Rev. Genet. 36, 411–453. Farris, H. E., LeBlanc, C. L., Goswami, J., and Ricci, A. J. (2004). Probing the pore of the auditory hair cell mechanotransducer channel in turtle. J. Physiol. 558, 769–792. Garcı´a‐An˜overos, J., and Corey, D. P. (1997). The molecules of mechanosensation. Annu. Rev. Neurosci. 20, 567–594. Gong, Z., Son, W., Chung, Y. D., Kim, J., Shin, D. W., McClung, C. A., Lee, Y., Lee, H. W., Chang, D. J., Kaang, B. K., Cho, H., Oh, U., et al. (2004). Two interdependent TRPV channel subunits, inactive and Nanchung, mediate hearing in Drosophila. J. Neurosci. 24, 9059–9066. Hamill, O. P., and McBride, D. W., Jr. (1996). The pharmacology of mechanogated membrane ion channels. Pharmacol. Rev. 48, 231–252. Howard, J., and Bechstedt, S. (2004). Hypothesis: A helix of ankyrin repeats of the NOMPC‐ TRP ion channel is the gating spring of mechanoreceptors. Curr. Biol. 14, R224–R226. Howard, J., and Hudspeth, A. J. (1988). Compliance of the hair bundle associated with gating of mechanoelectrical transduction channels in the bullfrog’s saccular hair cell. Neuron 1, 189–199. Jaquemar, D., Schenker, T., and Trueb, B. (1999). An ankyrin‐like protein with transmembrane domains is specifically lost after oncogenic transformation of human fibroblasts. J. Biol. Chem. 274, 7325–7333. Jordt, S. E., Bautista, D. M., Chuang, H. H., McKemy, D. D., Zygmunt, P. M., Hogestatt, E. D., Meng, I. D., and Julius, D. (2004). Mustard oils and cannabinoids excite sensory nerve fibres through the TRP channel ANKTM1. Nature 427, 260–265. Jorgensen, F., and Ohmori, H. (1988). Amiloride blocks the mechano‐electrical transduction channel of hair cells of the chick. J. Physiol. 403, 577–588. Kennedy, H. J., Evans, M. G., Crawford, A. C., and Fettiplace, R. (2003). Fast adaptation of mechanoelectrical transducer channels in mammalian cochlear hair cells. Nat. Neurosci. 6, 832–836.
188
Castiglioni and Garcı´a‐An˜overos
Kim, J., Chung, Y. D., Park, D. Y., Choi, S., Shin, D. W., Soh, H., Lee, H. W., Son, W., Yim, J., Park, C. S., Kernan, M. J., and Kim, C. (2003). A TRPV family ion channel required for hearing in Drosophila. Nature 424, 81–84. Kimitsuki, T., Nakagawa, T., Hisashi, K., Komune, S., and Komiyama, S. (1996). Gadolinium blocks mechano‐electric transducer current in chick cochlear hair cells. Hear. Res. 101, 75–80. Kobayashi, K., Fukuoka, T., Obata, K., Yamanaka, H., Dai, Y., Tokunaga, A., and Noguchi, K. (2005). Distinct expression of TRPM8, TRPA1, and TRPV1 mRNAs in rat primary aVerent neurons with adelta/c‐fibers and colocalization with trk receptors. J. Comp. Neurol. 493, 596–606. Kroese, A. B., Das, A., and Hudspeth, A. J. (1989). Blockage of the transduction channels of hair cells in the bullfrog’s sacculus by aminoglycoside antibiotics. Hear. Res. 37, 203–217. Kwan, K. Y., Allchorne, A. J., Vollrath, M. A., Christensen, A. P., Zhang, D. S., Woolf, C. J., and Corey, D. P. (2006). TRPA1 contributes to cold, mechanical, and chemical nociception but is not essential for hair‐cell transduction. Neuron 50, 277–289. Lee, G., Abdi, K., Jiang, Y., Michaely, P., Bennett, V., and Marszalek, P. E. (2006). Nanospring behaviour of ankyrin repeats. Nature 440, 246–249. Li, W., Feng, Z., Sternberg, P. W., and Xu, X. Z. (2006). A C. elegans stretch receptor neuron revealed by a mechanosensitive TRP channel homologue. Nature 440, 684–687. Liedtke, W., Choe, Y., Marti‐Renom, M. A., Bell, A. M., Denis, C. S., Sali, A., Hudspeth, A. J., Friedman, J. M., and Heller, S. (2000). Vanilloid receptor‐related osmotically activated channel (VR‐OAC), a candidate vertebrate osmoreceptor. Cell 103, 525–535. Maroto, R., Raso, A., Wood, T. G., Kurosky, A., Martinac, B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nat. Cell Biol. 7, 179–185. McGrath, J., Somlo, S., Makova, S., Tian, X., and Brueckner, M. (2003). Two populations of node monocilia initiate left‐right asymmetry in the mouse. Cell 114, 61–73. Mizuno, A., Matsumoto, N., Imai, M., and Suzuki, M. (2003). Impaired osmotic sensation in mice lacking TRPV4. Am. J. Physiol. Cell Physiol. 285, C96–C101. Muraki, K., Iwata, Y., Katanosaka, Y., Ito, T., Ohya, S., Shigekawa, M., and Imaizumi, Y. (2003). TRPV2 is a component of osmotically sensitive cation channels in murine aortic myocytes. Circ. Res. 93, 829–838. Nagata, K., Duggan, A., Kumar, G., and Garcı´a‐An˜overos, J. (2005). Nociceptor and hair cell transducer properties of TRPA1, a channel for pain and hearing. J. Neurosci. 25, 4052–4061. Nauli, S. M., Alenghat, F. J., Luo, Y., Williams, E., Vassilev, P., Li, X., Elia, A. E., Lu, W., Brown, E. M., Quinn, S. J., Ingber, D. E., Zhou, J., et al. (2003). Polycystins 1 and 2 mediate mechanosensation in the primary cilium of kidney cells. Nat. Genet. 33, 129–137. Obata, K., Katsura, H., Mizushima, T., Yamanaka, H., Kobayashi, K., Dai, Y., Fukuoka, T., Tokunaga, A., Tominaga, M., and Noguchi, K. (2005). TRPA1 induced in sensory neurons contributes to cold hyperalgesia after inflammation and nerve injury. J. Clin. Invest. 115, 2393–2401. Pennekamp, P., Karcher, C., Fischer, A., Schweickert, A., Skryabin, B., Horst, J., Blum, M., and Dworniczak, B. (2002). The ion channel polycystin‐2 is required for left‐right axis determination in mice. Curr. Biol. 12, 938–943. Ricci, A. (2002). DiVerences in mechano‐transducer channel kinetics underlie tonotopic distribution of fast adaptation in auditory hair cells. J. Neurophysiol. 87, 1738–1748. Ricci, A. J., Wu, Y. C., and Fettiplace, R. (1998). The endogenous calcium buVer and the time course of transducer adaptation in auditory hair cells. J. Neurosci. 18, 8261–8277.
8. MechanoTRPs and TRPA1
189
Ricci, A. J., Crawford, A. C., and Fettiplace, R. (2000). Active hair bundle motion linked to fast transducer adaptation in auditory hair cells. J. Neurosci. 20, 7131–7142. Ricci, A. J., Crawford, A. C., and Fettiplace, R. (2003). Tonotopic variation in the conductance of the hair cell mechanotransducer channel. Neuron 40, 983–990. Rusch, A., Kros, C. J., and Richardson, G. P. (1994). Block by amiloride and its derivatives of mechano‐electrical transduction in outer hair cells of mouse cochlear cultures. J. Physiol. 474, 75–86. Sidi, S., Friedrich, R. W., and Nicolson, T. (2003). NompC TRP channel required for vertebrate sensory hair cell mechanotransduction. Science 301, 96–99. Sotomayor, M., Corey, D. P., and Schulten, K. (2005). In search of the hair‐cell gating spring elastic properties of ankyrin and cadherin repeats. Structure 13, 669–682. Stepanyan, R. B. E., Friedman, T. B., and Frolenkov, G. L. (2006). TRPA1, a hair cell channel with unknown function? In ‘‘29th ARO Midwinter Meeting,’’ pp. 211–212. Baltimore, MD. Strotmann, R., Harteneck, C., Nunnenmacher, K., Schultz, G., and Plant, T. D. (2000). OTRPC4, a nonselective cation channel that confers sensitivity to extracellular osmolarity. Nat. Cell Biol. 2, 695–702. Suzuki, M., Mizuno, A., Kodaira, K., and Imai, M. (2003). Impaired pressure sensation in mice lacking TRPV4. J. Biol. Chem. 278, 22664–22668. Tracey, W. D., Jr., Wilson, R. I., Laurent, G., and Benzer, S. (2003). Painless, a Drosophila gene essential for nociception. Cell 113, 261–273. Vriens, J., Watanabe, H., Janssens, A., Droogmans, G., Voets, T., and Nilius, B. (2004). Cell swelling, heat, and chemical agonists use distinct pathways for the activation of the cation channel TRPV4. Proc. Natl. Acad. Sci. USA 101, 396–401. Walker, R. G., Willingham, A. T., and Zuker, C. S. (2000). A Drosophila mechanosensory transduction channel. Science 287, 2229–2234. Zhou, X. L., Batiza, A. F., Loukin, S. H., Palmer, C. P., Kung, C., and Saimi, Y. (2003). The transient receptor potential channel on the yeast vacuole is mechanosensitive. Proc. Natl. Acad. Sci. USA 100, 7105–7110.
This page intentionally left blank
CHAPTER 9 TRPCs as MS Channels Owen P. Hamill and Rosario Maroto Department of Neuroscience and Cell Biology, University of Texas Medical Branch, Galveston, Texas 77555
I. II. III. IV. V. VI. VII.
Overview Introduction Practical Aspects of Recording MS Channels Distinguishing Direct vs Indirect MS Channels Extrinsic Regulation of Stretch Sensitivity Strategies to Identify MS Channel Proteins General Properties of TRPCs A. TRPC Expression B. TRPC Activation and Function C. TRPC–TRPC Interactions D. TRPC Interactions with ScaVolding Proteins E. Single TRPC Channel Conductance F. TRPC Pharmacology VIII. Evidence for TRPC Mechanosensitivity A. TRPC1 B. TRPC2 C. TRPC3 D. TRPC4 E. TRPC6 IX. Conclusions References
I. OVERVIEW This chapter reviews recent evidence indicating canonical or classical transient receptor potential (TRPC) channels are directly or indirectly mechanosensitive (MS) and can therefore be designated as mechano‐operated channels (MOCs). The MS functions of TRPCs may be mechanistically related to Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59009-X
192
Hamill and Maroto
their better known functions as store‐operated and receptor‐operated channels (SOCs and ROCs). In particular, mechanical forces may be conveyed to TRPC channels through the ‘‘conformational coupling’’ mechanism that transmits information regarding the status of internal Ca2þ stores. Furthermore, all TRPCs are regulated by receptors coupled to phospholipases that are themselves MS and can regulate channels via lipidic second messengers. Accordingly, there may be several nonexclusive mechanisms by which mechanical forces may regulate TRPC channels, including direct sensitivity to bilayer mechanics, physical coupling to internal membranes, and/or cytoskeletal proteins, and sensitivity to lipidic second messengers generated by MS enzymes. Various strategies that can be used to separate out diVerent MS‐gating mechanisms and their possible role in specific TRPCs are discussed.
II. INTRODUCTION MS ion channels transduce mechanical force into ion flux. To exhibit direct mechanosensitivity, a channel protein must be sensitive to some membrane parameter that changes with mechanical deformation. In many cases, the mechanotransduction step involves a shift in the equilibrium between closed and open channel conformations caused by changes in bilayer mechanics (e.g., lipid packing, bilayer thickness, curvature, and/or lateral pressure profile) or by direct ‘‘tugging’’ on the protein by cytoskeletal and/or extracellular tethers (Hamill and Martinac, 2001; Kung, 2005; Markin and Sachs, 2007; Matthews et al., 2007; Powl and Lee, 2007). However, some channels may be indirectly MS in that they derive their mechanosensitivity from being functionally linked to MS enzymes that regulate the channel via second messenger or phosphorylation. Apart from mechanosensation, MS channels have been implicated in several basic cellular functions, including the regulation of cell volume, cell shape, motility, growth, and cell death. Because abnormalities in MS channels may also contribute to major human diseases, including muscular dystrophy, kidney disease, cardiac arrhythmias, hypertension, and tumor cell invasion (‘‘mechanochannelopathies’’), there is great interest in identifying the molecules that form MS channels and discovering agents that can selectively block their activity and/or expression (Chapter 4, Gottlieb et al.; Chapter 10, Cantiello et al.; Chapter 15, Drew et al.; Chapter 16, Lansman; Chapter 17, Maroto and Hamill; Chapter 19, Chapleau et al.). In eukaryotic cells, three membrane protein families, epithelial Naþ channel (ENaC), two pore domain Kþ (TREK), and TRP families have been implicated in forming MS Naþ (MscNa), Kþ (MscK), and cation/Ca2þ (MscCa) channels, respectively (Chapter 3, Bazopoulou and Tavernarakis;
9. MS TRPCs
193
Chapter 6, Drummond; Chapter 7, Chemin et al.; Chapter 8, Castiglioni and Garcı´a‐An˜overos; Chapter 10, Cantiello et al.). Here, we focus on the TRPCs, which have been implicated in forming the ubiquitous stretch‐ activated MscCa (Maroto et al., 2005).
III. PRACTICAL ASPECTS OF RECORDING MS CHANNELS The most direct method to determine if an ion channel is MS is to apply a hydrostatic or osmotic pressure gradient across the membrane patch while monitoring single‐channel currents (Hamill et al., 1981; Hamill, 1983; Guharay and Sachs, 1984; Hamill, 2006). This method led directly to the discovery of MscK and MscCa in frog red blood cells and cultured chick myotubes, respectively (Hamill, 1983; Guharay and Sachs, 1984). Subsequently, MscK and MscCa were shown to be widely expressed in sensory and nonsensory animal cells and proposed to function in various physiological processes including regulatory volume decrease (RVD) in response to osmotic swelling (Sachs, 1988; Morris, 1990; Sackin, 1995; Sachs and Morris, 1998; Hamill and Martinac, 2001; Patel and Honore, 2001). In several cases of RVD, it was possible to demonstrate that the same channel (e.g., MscK, MscCa, MscL, and MscS) was activated by cell swelling and membrane stretch (Christensen, 1987; Sackin, 1989; Cemerikic and Sackin, 1993; Levina et al., 1999; Vanoye and Reuss, 1999). However, in other cases, most notably the vanilloid transient receptor potential 4 (TRPV4), the channel was sensitive to cell volume changes without displaying stretch sensitivity (Strotmann et al., 2000). This discrepancy may arise because TRPV4 is not directly MS but instead derives its volume sensitivity from being coupled to one or more MS enzymes (Watanabe et al., 2003; Xu et al., 2003; Vriens et al., 2004; Cohen, 2005a). In particular, one group has proposed that TRPV4 is coupled to an osmotic‐sensitive Src protein tyrosine kinase that regulates channel activation by tyrosine phosphorylation (Xu et al., 2003; Cohen, 2005b). Another group (Watanabe et al., 2003; Vriens et al., 2004) has proposed that TRPV4 is coupled to the volume‐sensitive phospholipase A2 (PLA2; Basavappa et al., 1988; Lehtonen and Kinnunen, 1995) that releases arachidonic acid (AA) from membrane phospholipids, which is then metabolized, via the action of cytochrome P450, into 50 ,60 ‐epoxyeicosatrienoic acid (50 ,60 ‐EET). In support of the latter scheme, it was shown that blocking either PLA2 or cytochrome P450 inhibits TRPV4 activation, whereas direct application of 50 ,60 ‐EET activates TRPV4 in a membrane‐delimited manner (Watanabe et al., 2003; Vriens et al., 2004). The group that carried out the PLA2 study was unable to reproduce the Src results (Cohen, 2005b), indicating the mechanism(s) that activates TRPV4 may vary with cell type and/or experimental conditions.
194
Hamill and Maroto A
Turn-on mmHg −20 −40 −60
P
1 2270 µs
2
−20
936 µs 3
−40
390 µs 20 pA −60 2 ms B
Turn-off
mmHg +30 0 −30
P 1 23 ms
−30 2
0 15 ms
3
+30 5 ms 100 pA 100 ms
10 ms
FIGURE 1 Fast turn‐on and turn‐oV of MS channel currents measured in response to suction steps applied with a pressure clamp. A shows in the top trace (labeled P) three superimposed suction steps of 20, 40, and 60 mmHg applied to a cell‐attached patch on a Xenopus oocyte. The lower three traces show the change in latency and the rate of turn‐on of the currents in response to the increasing suction steps. The numbers in microseconds alongside each trace reflect the time from 20% to 80% of the peak current. B shows recordings designed to show the pressure dependence of the current turn‐oV. In the left‐hand panel, the
9. MS TRPCs
195
In any case, the results indicate that while TRPV4 may function as a mechano‐eVector, it is not directly MS (O’Neil and Heller, 2005). There are added complications with other channels because they can be activated both by membrane stretch and by lipidic second messengers including AA and lysophospholipids (Martinac et al., 1990; Kim, 1992; Hamill and McBride, 1996; Casado and Ascher, 1998; Patel et al., 2001; Chapter 7, Chemin et al.). In this case, the issue becomes how to distinguish between direct and indirect mechanisms of mechanosensitivity. IV. DISTINGUISHING DIRECT VS INDIRECT MS CHANNELS Channels that are directly MS should only be limited by the conformational transitions of the channel protein, and may therefore be activated and deactivated with relatively brief delays (i.e., in the submillisecond or millisecond range). In comparison, channels dependent on enzymatic reactions and/or diVusion of second messenger may be expected to show much longer delays in opening and closing (e.g., 1 s). Figure 1 illustrates the activation and deactivation of the oocyte MS channel in response to increasing pressure steps. The transition time for the pressure step is limited by the speed of the pressure clamp (McBride and Hamill, 1992; 1995, 1999; Besch et al., 2002). However, once the threshold pressure for activation is reached, the MS current turns on in few hundred microseconds (Fig. 1A). With increasing step size, both the latency and rise time of the MS current decreases consistent with the pressure reaching threshold faster. Similarly, the current turn‐ oV indicates the channels close faster with larger pressure steps (Fig. 1B). The slower time for turn‐oV compared with turn‐on presumably reflects the relatively slower rate of MS channel closure under these conditions (Fig. 1B). Similar brief delays and fast channel opening have been reported for activation and deactivation of the expressed a TWIK (tandem of P domains in a weak inward rectifier Kþ channel)‐related arachidonic acid stimulated Kþ channel (TRAAK) (Honore´ et al., 2006). In contrast, an MscK expressed in snail neurons, which like TRAAK is a two‐pore domain Kþ channel (Vandorpe and Morris, 1992), shows activation delays of up to several seconds (Small and Morris, 1994). However, because the delays can superimposed suction pulse waveforms are shown in the upper trace and the corresponding current responses are shown in the lower traces. The initial activating suction was 40 mmHg for all three pulses. To turn oV the currents the suction was stepped back to three increasing positive pressures. The right‐hand panel shows on an expanded timescale the regions of the turn‐ oV that were highlighted in the left panel with the numbers in microseconds representing the turn‐oV times (20–80%), and indicate that as the turn‐off step size increases the channels turn oV faster (Reproduced from McBride and Hamill, 1993).
196
Hamill and Maroto
be abolished by mechanical or chemical disruption of the cytoskeleton (CSK), they presumably arise from CSK constraint of the bilayer that prevents rapid transmission of tension to the channel. So far, studies measuring possible delays in pressure activation of TRPs that are suspected of being indirectly MS have not been performed. In the case of TRPV4, which has been functionally linked to PLA2 (Vriens et al., 2004), it will be interesting to determine whether its apparent lack of stretch sensitivity when measured in the patch was overlooked because of long delays and slow channel activation in response to applied pressure. A further strategy for discriminating between direct and indirect MS channel mechanisms is to use specific inhibitors to test for involvement of MS enzymes (e.g., p‐bromophenacyl bromide for PLA2, 4‐amino‐5‐(4‐chlorophenyl)‐7‐(t‐ butyl)pyrazolo[3,4,d] pyrimidine (PP2) for Src tyrosine kinase, and U73122 for phospholipase C (PLC). In particular, the stretch sensitivity of the MS channel in arterial smooth muscle has been reported to be abolished by the PLC inhibitor U73122 (Park et al., 2003). Furthermore, Ca2þ influx in dystrophic muscle that is mediated by a TRPC‐dependent SOC and/or MOC (Vandebrouck et al., 2002; Ducret et al., 2006) can be abolished by inhibitors of PLA2 (Lindahl et al., 1995; Boittin et al., 2006; Section VIII.A.3). The most unequivocal method for distinguishing direct from indirect mechanosensitivity is to examine whether the detergent‐solubilized channel protein retains stretch sensitivity when reconstituted in pure liposomes. So far, this test has been applied to several MS channels in prokaryotes and MscCa expressed in the frog oocyte (Sukharev et al., 1993; Kloda and Martinac, 2001a,b; Sukharev, 2002; Maroto et al., 2005). This approach also oVers the potential of definitive evidence on whether lipidic second messengers [e.g., diacylglycerol (DAG), AA, lysophospholipids and 50 ,60 ‐EET] activate the channel by binding directly to the channel protein and/or its surrounding lipid without intermediate steps. Furthermore, the same method may also be applied to determine whether multiprotein component MS‐signaling complexes can be functionally reconstituted from their specific elements (e.g., TRPV4, PLA2, and so on). Although stretch sensitivity measured in the patch can be used to demonstrate a channel protein is MS at the biophysical level, it cannot prove the channel functions as a physiological mechanotransducer (Hamill, 2006). Indeed, many structurally diverse voltage‐ and receptor‐gated channels [e.g., Shaker, L‐type Ca2þ channels, N‐methyl‐D‐aspartate receptor (NMDAR), S‐type Kþ channels], as well as the simple model peptide channels alamethicin and gramicidin A, display stretch sensitivity in patch recordings (Opsahl and Webb, 1994; Paoletti and Ascher, 1994; Martinac and Hamill, 2002; Chapter 11, Morris and Juranka). In order to demonstrate functionality, one also needs to show that blocking the channel (pharmacologically and/or genetically) inhibits a mechanically induced cellular/physiological process.
9. MS TRPCs
197
V. EXTRINSIC REGULATION OF STRETCH SENSITIVITY Stretch sensitivity is unlikely to be accounted for by a single structural domain analogous to the S‐4 voltage sensor‐domain shared by voltage‐gated Naþ, Kþ, and Ca2þ channels (Hille, 2001); even the relatively simple peptide channels, gramicidin and alamethicin, which have dramatically diVerent structures and gating mechanisms, exhibit stretch sensitivity (Hamill and Martinac, 2001; Martinac and Hamill, 2002). Furthermore, stretch sensitivity is not a fixed channel property but rather can undergo significant changes with changing extrinsic conditions. For example, mechanical and/or chemical disruption of the CSK can either enhance or abolish the stretch sensitivity of specific channels (Guharay and Sachs, 1984; Hamill and McBride, 1992, 1997; Small and Morris, 1994; Patel and Honore, 2001; Hamill, 2006); changes in bilayer thickness (Martinac and Hamill, 2002), membrane voltage (Gu et al., 2001; Chapter 11, Morris and Juranka), or dystrophin expression (Franco‐Obregon and Lansman, 2002; Chapter 16, Lansman) can switch specific MS channels from being stretch‐activated to stretch‐inactivated; specific lipids (Patel and Honore, 2001; Chemin et al., 2005), nucleotides (Barsanti et al., 2006a and references therein), and increased internal acidosis (Honore´ et al., 2002; Barsanti et al., 2006b) can convert MS channels into constitutively open ‘‘leak’’ channels. The basis for these changes is often because changes in the bilayer, CSK, and/or ECM alter how mechanical forces are conveyed to the channel protein. The practical consequence may be that the specific conditions associated with reconstitution and/or heterologous expression may alter the stretch sensitivity of the reconstituted/expressed channel.
VI. STRATEGIES TO IDENTIFY MS CHANNEL PROTEINS Once a channel has been functionally identified as stretch sensitive, there are several strategies that can be used to identify the membrane protein. The first strategy of ‘‘expression cloning’’ involves generating a cDNA library from cells expressing the channel, and then screening the library, typically in Xenopus oocytes or a mammalian cell line. This strategy has been used to clone several voltage‐ and receptor‐gated channels, including the first vanilloid receptor TRP channel TRPV1 (Caterina et al., 1997). However, its application to MS channels has proven problematic because the expression vehicles express their own endogenous MS channels. The second strategy of ‘‘functional protein reconstitution’’ involves detergent solubilizing and reconstituting membrane proteins into liposomes and then screening for stretch sensitivity using patch clamp recording. This strategy has been used to successfully identify/clone a number of MS channel proteins from bacteria and archaea
198
Hamill and Maroto
(Sukharev et al., 1993, 1994; Sukharev, 2002; Martinac, 2007). It was also used to implicate a TRPC in forming MscCa (Maroto et al., 2005).
VII. GENERAL PROPERTIES OF TRPCs This section provides an overview of the TRPC subfamily (for reviews see Minke and Cook, 2002; Vazquez et al., 2004a; Montell, 2005; Nilius and Voets, 2005; Parekh and Putney, 2005; Owsianik et al., 2006). The first TRP was discovered in a Drosophila mutant that showed a transient rather than a sustained receptor potential in response to light (Cosens and Manning, 1969; Minke et al., 1975; Montell and Rubin, 1989). On the basis of these kinetics, the protein was designated TRP. Subsequently seven mammalian TRP homologues were discovered that together with TRP now make up the TRPC1–7. Other TRP subfamilies include TRPV (vanilloid), TRPA (ankyrin), TRPP (polycystin), TRPM (melastatin), TRPML (mucolipid), TRPN (NompC), and TRPY (yeast), and these together with TRPCs form the TRP superfamily. In addition to MS TRPCs, specific members of the other subfamilies have also been implicated in mechanotransduction so that the MS mechanisms discussed below may be general (Walker et al., 2000; Palmer et al., 2001; Zhou et al., 2003; Nauli and Zhou, 2004; O’Neil and Heller, 2005; Voets et al., 2005, Saimi et al., 2007). The proposed transmembrane topology of TRPCs is reminiscent of voltage‐gated channels—sharing six transmembrane‐spanning helices (TM1–6), cytoplasmic N‐ and C‐termini, and a pore region between TM5 and TM6—but lacking the positively charged residues in the TM4 domain that forms the voltage sensor. The seven mammalian TRPC channels also share an invariant sequence in the C‐terminal tail called a TRP box (E‐W‐K‐F‐A‐R), as well as 3–4 N‐terminal ankyrin repeats. Although the ankyrin repeats may act as gating springs for MS channels (Howard and Bechstedt, 2004; Saimi et al., 2007; see also Chapter 8, Castiglioni and Garcı´a‐An˜overos; Chapter 10, Cantiello et al.), their exact role and that of the TRP box remains to be verified (Vazquez et al., 2004a; Owsianik et al., 2006). The TRPCs share very little sequence identity in the region that is C‐terminal of the TRP box, except for the common feature of CaM‐ and IP3R‐binding domains that have been implicated in Ca2þ feedback inhibition and activation by store depletion, respectively (Kiselyov et al., 1998; Vaca and Sampieri, 2002). On the basis of sequence homology, the TRPCs have been divided into two major subgroups TRPC3/6/7 (70–80% homology) and TRPC1/4/5 (65% homology). TRPC2 is in a special class because multiple stop codons within its open reading frame make it a pseudogene in humans. However, it does form a functional channel in rodents (Section VIII.B).
9. MS TRPCs
199
A. TRPC Expression TRPCs are widely expressed in mammalian tissues, with some human cells expressing all six and others expressing only one or two (Riccio et al., 2002; Antoniotti et al., 2006; Goel et al., 2006; Hill et al., 2006). The latter cells may prove useful for dissecting out specific TRPC functions, but it is necessary that selective expression be verified at both transcriptional and protein levels, since low turnover proteins may require little mRNA, and high mRNA levels need not translate into high membrane protein levels (Andersen and Seilhamer, 1997). Another caveat is that TRPC expression patterns may vary during development and with culture conditions (e.g., presence or absence of growth factors). For example, TRPC1 expression is upregulated by (1) serum deprivation where it contributes to increased proliferation of pulmonary arterial smooth muscle cells (Golovina et al., 2001), (2) tumor necrosis factor a where it enhances endothelial cell death (Paria et al., 2003), and (3) vascular injury in vivo which contributes to human neoitimal hyperplasia (Kumar et al., 2006); TRPC6 expression in pulmonary arterial smooth muscle cells is enhanced by platelet‐derived growth factor and by idiopathic pulmonary arterial hypertension (Yu et al., 2003, 2004).
B. TRPC Activation and Function Studies of TRPC activation and function are complicated by their polymodal activation and splice variants that display diVerent activation mechanisms (Ramsey et al., 2006). However, all TRPCs are regulated by PLC‐coupled receptors (i.e., G‐protein–coupled receptors or tyrosine kinase receptors). PLC hydrolyzes a component of the bilayer, PIP2, into two distinct messengers—the soluble InsP3 that activates the IP3R in the ER to release Ca2þ from internal stores and the lipophilic DAG that may regulate TRPs indirectly via protein kinase C (PKC) or by interacting directly with the TRPCs in a membrane‐ delimited manner (Delmas et al., 2002; Clapham, 2003; Ramsey et al., 2006). Furthermore, Bolotina and colleagues have shown that a diffusible second messenger produced by depletion of Ca2þ stores activates a Ca2þ independent phospholipase (iPLA2) that generates lysophospholipids, which are themselves capable of activating SOCs when exogenously applied to inside‐out patches (Smani et al., 2003; Bolotina and Csutora, 2005). Therefore, although all TRPCs could be classified as ROCs (but see Janssen and Kwan, 2007), they are more often subdivided into either SOCs, based on their sensitivity to Ca2þ store depletion that may or may not depend on PLC–IP3R signaling, or ROCs
200
Hamill and Maroto
that are activated by DAG or its byproducts, but are insensitive to Ca2þ store depletion (Hofmann et al., 1999). To be classified as a SOC, the channel should be gated by a variety of procedures that only share the common feature of reducing Ca2þ stores (Parekh and Putney, 2005). Unfortunately, there have been conflicting reports for all seven TRPCs on whether they function as SOCs, ROCs, or both. Here, we focus on a further complication that the same mechanisms that make a channel a SOC or a ROC may also contribute to it being MS.
C. TRPC–TRPC Interactions If all 7 TRPC subunits are expressed in a given cell and 4 subunits are required to form a channel (i.e., homotetrameric and heterotetrameric), then there could be as many as 100 diVerent TRPC channels types. However, the number would be smaller if only certain TRPC–TRPC combination can occur. Two diVerent models for TRPC interactions have been proposed: a homotypic model in which only subunits within each subfamily can interact to form channels with TRPC1/4/5 forming SOCs and TRPC3/6/7 forming ROCs (Hofmann et al., 2002; Sinkins et al., 2004), and a heterotypic model that also allows interactions between subfamily members, in this case with TRPC1, TRPC3, and TRPC7 proposed to form SOCs (i.e., without TRPC4 and TRPC6) and TRPC3, TRPC4, TRPC6, and TRPC7 proposed to form ROCs (without TRPC1; Zagranichnaya et al., 2005). In the heterotypic model, TRPC1’s role is limited to SOCs and TRPC4’s and TRPC6’s roles are limited to ROCs, while TRPC3 and TRPC7 can participate as both SOCs and ROCs (Zagranichnaya et al., 2005). Interestingly, both models were generated from studies of the human embryonic kidney cell line, HEK‐293, with the homotypic model based on gain‐of‐function (i.e., from TRPC overexpression) and the heterotypic model based on loss‐of‐function (i.e., from TRPC suppression). However, one complication with the former approach is that the level of TRPC expression can determine channel function. In particular, it has been shown that low TRPC3 expression result in SOCs, while high expression result in ROCs (Vazquez et al., 2003). This eVect presumably occurs because high expression promotes homomeric TRPC3 channels, whereas low levels allow for heteromers that include endogenous subunits (Brereton et al., 2001; Vazquez et al., 2003). DiVerences in channel function may also arise depending on whether the cell is permanently or transiently transfected, presumably because stable transfection allows time adaptive changes in endogenous protein expression (Lievremont et al., 2004).
9. MS TRPCs
201
D. TRPC Interactions with Scaffolding Proteins TRPCs also interact with a variety of regulatory and scaVolding proteins that may add further diversity and segregation of the channels (Ambudkar, 2006). In particular, it has been shown that several TRPCs assemble into multiprotein and lipid‐signaling complexes that result in physical and functional interactions between the plasma membrane and CSK and ER resident proteins. These interactions may also allow for mechanical forces to be conveyed via a tethered mechanism to gate the channel (Howard et al., 1988; Hamill and Martinac, 2001; Matthews et al., 2007; Chapter 3, Bazopoulou and Tavernarakis; Chapter 10, Cantiello et al.). Alternatively, the interactions may also serve to constrain the development or transmission of bilayer tension to the TRPC and thereby ‘‘protect’’ it from being mechanically activated (Small and Morris, 1994; Hamill and McBride, 1997). For all TRPCs, the C‐terminal coiled‐coil domains and N‐terminal ankyrin repeats have the potential to mediate protein–CSK interactions. All TRP family members also encode a conserved proline‐rich sequence LP(P/X)PFN in their C‐termini that is similar to the consensus‐binding site for Homer, a scaVold protein that has been shown to facilitate TRPC1 interaction with IP3R—disruption of which has been proposed to promote SOC activity (Yuan et al., 2003). For example, TRPC1 mutants lacking Homer protein‐ binding sites show diminished interaction between TRPC1 and IP3R and the TRPC1 channels are constitutively active. Similarly, coexpression of a dominant‐negative form of Homer increases basal TRPC1 channel activity (Yuan et al., 2003). Another protein I‐mfa, which inhibits helix‐loop‐helix transcription factors, also binds to TRPC1 and blocks SOC function (Ma et al., 2003). TRPC1 also expresses a dystrophin domain in its C‐terminus (Wes et al., 1995) that may allow interaction with dystrophin, and possibly explain why the absence of dystrophin in Duchenne muscular dystrophic muscle results in TRPC1 channels being abnormally gated open (Section VIII.A.3). TRPC1 also shows a putative caveolin‐1‐binding domain that may promote its functional recruitment into lipid rafts and increase SOC activity (Lockwich et al., 2000; Brazier et al., 2003; Ambudkar, 2006). TRPC1 has been shown to interact with stromal interaction molecule (STIM), the putative ER Ca2þ sensor that can apparently regulate TRPC1 SOC function (Huang et al., 2006). Junctate is another IP3R‐associated protein, and it interacts with TRPC2, TRPC3, and TRPC5 (but not TRPC1) to regulate their SOC/ROC function (Treves et al., 2004; Stamboulian et al., 2005). In pulmonary endothelial cells, TRPC4 is localized to cell–cell adhesions in cholesterol‐rich caveolae and has been shown to interact with the spectrin CSK via the protein 4.1 (Torihashi et al., 2002; CioY et al., 2005).
202
Hamill and Maroto
Furthermore, either deletion of the putative 4.1 protein‐binding site on the TRPC4 C‐terminus of TRPC4 or addition of peptides that competitively bind to that site are able to reduce SOC activity. Another site for TRPC4–CSK interaction involves the PSD‐95/disk‐large protein/zona occludens 1 (PDZ)‐ binding domain located at the TRPC4 distal C‐terminus that binds to the Naþ/Hþ exchange regulatory factor (NHERF)‐scaVolding protein (Tang et al., 2000; Mery et al., 2002). TRPC6 interacts with the stomatin‐like protein podocin that may modulate its MOC function in the renal slit diaphragm (Reiser et al., 2005). Interestingly, another stomatin homologue, MEC‐2, links the putative MS channel to the microtubular CSK in Caenorhabditis elegans neurons (Chapter 3, Bazopoulou and Tavernarakis). In summary, TRPCs undergo dynamic interactions with various scaVolding proteins that may act to inhibit or promote a particular mode of channel activation. Any one of these interactions may be important in modulating the mechanosensitivity of TRPC by focusing mechanical force on the channel or constraining the channel and/or bilayer from responding to mechanical stretch. It may be that the right combination of TRPC proteins and accessory proteins are needed to produce channels that are not constitutively active but are responsive to store depletion and/or mechanical stimulation.
E. Single TRPC Channel Conductance Single‐channel conductance provides a good identifying fingerprint of specific channels. However, compared with whole‐cell current recording studies, there have been relatively few studies of the single‐channel currents that are either enhanced or deleted by TRPC overexpression or suppression, respectively. Furthermore, there is no simple way to determine if a channel reflects a homomeric rather than a heteromeric TRPC. Another practical issue for comparisons has been the lack of standardized recording conditions. Nevertheless, a survey of the TRPC single‐channel literature indicates the following order for conductance values TRPC3 (65 pS) > TRPC5 (50 pS) > TRPC4 (32 pS) TRPC6 (31 pS) > TRPC1 (20 pS) for estimates made from cell‐attached recordings with 100‐ to 150‐mM Naþ/Csþ, 1‐ to 4‐mM Ca2þ/ Mg2þ at 40 to 100 mV (Hurst et al., 1998; Kiselyov et al., 1998; Hofmann et al., 1999; Yamada et al., 2000; Liu et al., 2003; Stru¨bing et al., 2003; Bugaj et al., 2005; Maroto et al., 2005; Inoue et al., 2006). The only available estimates for TRPC2 (42 pS) and TRPC7 (60 pS) were made with no divalents (Perraud et al., 2001; Zufall et al., 2005). These numbers may serve as a baseline for the future conductance measurements of the purified/ reconstituted TRPCs.
9. MS TRPCs
203
F. TRPC Pharmacology The pharmacological tools to study TRPCs are limited with the following agents reported to block, stimulate, or have no eVect on diVerent TRPCs (Ramsey et al., 2006): SKF‐96365 blocks TRPC3 and TRPC6 currents (at 5 mM), and is considered a ROC more than a SOC blocker; 2APB (2‐aminoethoxydiphenyl borate) blocks TRPC1 (80 mM), TRPC5 (20 mM), and TRPC6 (10 mM) but not TRPC3 (75 mM), and is considered more a SOC than a ROC blocker; Gd3þ and La3þ block TRPC1 and TRPC6, but potentiate TRPC4 and TRPC5 (in micromolars; Jung et al., 2003); flufenamate blocks TRPC3, TRPC5, and TRPC7 (100 mM), but potentiates TRPC6; and tarantula venom peptide, GsmTX4, (Gottlieb et al., 2004) blocks TRPC1 in mammalian cells but not in Xenopus oocytes (Hamill, 2006; Chapter 4, Gottlieb et al.). Other agents of interest that need to be systematically tested on both SOC and ROC activity include gentamicin, ruthenium red, GsmTX4, and amiloride (Lane et al., 1991, 1992; Ru¨sch et al., 1994; Flemming et al., 2003; Suchyna et al., 1998, 2004; Jacques‐Fricke et al., 2006).
VIII. EVIDENCE FOR TRPC MECHANOSENSITIVITY Below, we consider the MS role of specific TRPCs. At this time, the main evidence exists for TRPC1 (SOC), TRPC6 (a DAG‐activated ROC), and to a lesser extent TRPC4 (an AA‐activated ROC). However, as discussed in Section IX, a basic issue is whether the mechanisms that confer SOC and ROC activity on TRPC channels also contributes to there mechanonsensitivity. In this case, all TRPs may end up expressing some degree of mechanosensitivity.
A. TRPC1 TRPC1 was the first identified vertebrate TRP homologue (Wes et al., 1995; Zhu et al., 1995) and initial heterologous expression of human TRPC1 (hTRPC1) in Chinese hamster ovary (CHO) and sf9 cells enhanced SOC currents (Zitt et al., 1996). However, a subsequent study indicated hTRPC1 expression in sf9 cells induced a constitutively active nonselective cation channel that was not sensitive to store depletion (Sinkins et al., 1998). This early discrepancy raises the possibility that store sensitivity (and perhaps stretch sensitivity) may depend on a variety of conditions (e.g., expression levels, presence of endogenous TRPCs, and state of phosphorylation). For example,
204
Hamill and Maroto
TRPC1 has multiple serine/threonine phosphorylation sites in the putative pore‐forming region and the N‐ and C‐termini, and at least one report indicates that PKCa‐dependent phosphorylation of TRPC1 can enhance Ca2þ entry induced by store depletion (Ahmmed et al., 2004). Despite this early discrepancy, many studies now point to TRPC1 forming a SOC (Liu et al., 2000, 2003; Xu and Beech, 2001; Kunichika et al., 2004; for reviews see Beech et al., 2003; Beech, 2005) and in cases where TRPC1 expression has not resulted in enhanced SOC (Sinkins et al., 1998; Lintschinger et al., 2000; Stru¨bing et al., 2001), it has been argued that TRPC1 was not traYcked to the membrane (Hofmann et al., 2002). This does not seem to be the case when hTRPC1 is expressed in the oocyte (Brereton et al., 2000; see Figs. 2 and 3). In any case, any direct TRPC1 involvement in forming the highly Ca2þ‐selective SOC or Ca2þ release‐activated current (ICRAC) seems to be reduced by the finding that a novel protein family (i.e., CRAM1 or Orai1) forms ICRAC channels (Peinelt et al., 2006; but see Mori et al., 2002; Huang et al., 2006). 1. Maitotoxin Activates TRPC1 and MscCa In 1999, xTRPC1 was cloned from Xenopus oocytes and shown to be 90% identical in sequence to the hTRPC1 (Bobanovic et al., 1999). An anti‐TRPC1 antibody (T1E3) targeted to an extracellular loop of the predicted protein was generated and shown to recognize an 80‐kDa protein. Immunofluorescent staining indicated an irregular ‘‘punctuate’’ expression pattern of xTRPC1 that was uniformly evident over the animal and vegetal hemispheres. Patch clamp studies also indicate that MscCa is uniformly expressed over both hemispheres (Zhang and Hamill, 2000a). This is in contrast to the polarized expression of the ER and the phosphatidylinositol second messenger system, which are more abundantly expressed in the animal hemisphere (Callamaras et al., 1998; Jaconi et al., 1999). These results indicate that neither TRPC1 nor MscCa are tightly coupled to ER internal Ca2þ stores and IP3 signaling. Originally, it was speculated that punctuate TRPC1 expression reflected discrete channel clusters but it could also indicate the channels are localized in microvilli, which make up >50% of the membrane surface (Zhang et al., 2000). In another study, testing the idea that xTRPC1 formed a SOC, Brereton et al. (2000) found that antisense oligonuceotides targeting diVerent regions of xTRPC1 sequence did not inhibit IP3‐ or thapsigargin‐stimulated Ca2þ inflow (cf., Tomita et al., 1998). Furthermore, overexpreesion of hTRPC1 did not enhance the basal or IP3‐stimulated Ca2þ inflow (Brereton et al., 2000). However, they did see enhancement of a lysophosphatidic acid (LPA)‐stimulated Ca2þ influx. Interestingly, LPA also enhances mechanically induced Ca2þ influx in a variety of other cells (Ohata et al., 2001). On the basis of the apparent lack of TRPC1‐related SOC activity, Brereton et al. (2000) speculated that TRPC1 might form the endogenous cation channel activated
9. MS TRPCs
205
by the marine toxin, maitotoxin (MTX). To test this idea, they compared the properties of MTX‐activated conductance in normal and in TRPC1‐transfected rat liver cells (Brereton et al., 2001), and found that the endogenous MTX‐activated conductance displayed properties diVerent from the enhanced MTX‐activated conductance expressed in the hTRPC1‐transfected cells. In particular, the endogenous conductance showed a higher selectivity for Naþ over Ca2þ and a higher sensitivity to Gd3þ block (K50% block ¼ 1 mM vs 3 mM). These diVerences were taken to indicate that other endogenous TRPC subunits may normally combine with TRPC1 to form the endogenous MTX‐activated conductance, whereas hTRPC1 alone forms the enhanced MTX‐activated conductance (Brereton et al., 2001). Unlike with oocytes, it was found that heterologous expression of hTRPC1 in rat liver cells did increase thapsigarin‐induced Ca2þ inflow. Evidence from several studies indicates that the oocyte MTX‐activated conductance may be mediated by MscCa (Bielfeld‐Ackermann et al., 1998; Weber et al., 2000; Diakov et al., 2001). In particular, both display the same cation selectivity, are blocked by 1‐mM amiloride and 10‐mM Gd3þ, are insensitive to flufenamic and niflumic acid, and have a conductance of 25 pS (measured in symmetrical 140‐mM Kþ and 2‐mM external Ca2þ). Because MTX is a highly amphipathic molecule (Escobar et al., 1998), it may activate MscCa by changing bilayer‐membrane interactions, as has been proposed for other amphipaths that can activate MS channels in the absence of membrane stretch (Martinac et al., 1990; Kim, 1992; Hamill and McBride, 1996; Casado and Ascher, 1998; Perozo et al., 2002). 2. TRPC1 and Cell Swelling To directly test whether TRPC1 might be MS, Chen and Barritt (2003) selectively suppressed TRPC1 expression in rat liver cells and measured their response to osmotic cell swelling. Liver cells express MscCa (Bear, 1990) and previous studies had shown that osmotic swelling of epithelial cells activates an MscCa‐dependent Ca2þ influx that stimulates Ca2þ‐activated Kþ eZux accompanied by Cl /H2O eZux and RVD (Christensen, 1987). However, in the TRPC1‐suppressed liver cells, hypotonic stress caused a greater swelling and faster RVD than observed in control liver cells (Chen and Barritt, 2003). This opposite response may occur because TRPC1 suppression results in a compensatory overexpression of other TRPCs (or redundant RVD mechanisms) that enhance cell swelling and RVD. It should also be recognized that cell swelling does not always activate MscCa. For example, although hypotonic solution activates a robust Ca2þ‐independent Cl conductance in Xenopus oocytes that should contribute to RVD, it fails to activate the endogenous MscCa (Ackerman et al., 1994; Zhang and Hamill, 2000a,b).
206
Hamill and Maroto
3. Abnormal TRPC1/MscCa Activity in Duchenne Muscular Dystrophy Both TRPC1 and MscCa are expressed in skeletal muscle and both have been implicated in the muscular degeneration that occurs in Duchenne muscular dystrophy (DMD). In particular, muscle fibers from the mdx mouse (i.e., an animal model of DMD) show an increased vulnerability to stretch‐induced membrane wounding (Yeung and Allen, 2004; Allen et al., 2005) that has been linked to elevated [Ca2þ]i levels caused by increased Ca2þ leak channel activity (Fong et al., 1990) and/or abnormal MscCa activity (Franco and Lansman, 1990; Chapter 16, Lansman). On the basis of the observation that the channel activity was increased by thapsigargin‐induced store depletion, it was proposed that the channel may also be a SOC belonging to the TRPC family (Vandebrouck et al., 2002, see also Hopf et al., 1996). To test this idea, mdx and normal muscle were transfected with antisense oligonucleotides designed against the most conserved TRPC regions. The transfected‐muscles showed a significant reduction in expression of TRPC1 and TRPC4 but not TRPC6 (all three TRPCs are expressed in normal and mdx muscle) and a decrease in the Ca2þ leak channel activity. Previous studies indicate that MscCa behaves more like a Ca2þ leak channel in mdx patches (Franco‐Obregon and Lansman, 2002) and in some oocyte patches (Reifarth et al., 1999). It has also been reported that SOC and MscCa in mdx muscle display the same single‐ channel conductance and sensitivity to block by Gd3þ, SKF96365, 2APB, and GsMTx4 (Ducret et al., 2006). These studies implicate TRPC1 as being a subunit of both the SOC and MscCa, which given the presence of a dystrophin domain on the C‐terminus of TRPC1 (Wes et al., 1995) could explain the shift in gating mode in mdx muscle. 4. TRPC1 and Polycystic Kidney Disease TRPC1 interacts with the putative MS channel TRPP2 when both are heterologously expressed in HEK‐293 (Tsiokas et al., 1999), and there is evidence that TRPC1 and TRPP2 may form functional heteromers (Delmas, 2004). TRPP2 is a distant member of the TRP family (polycystin subfamily) and has been shown to form a Ca2þ‐permeable cation channel that is mutated in the autosomal dominant polycystic kidney disease (ADPKD; Nauli et al., 2003; Nauli and Zhou, 2004; Giamarchi et al., 2006; Chapter 10, Cantiello et al.). TRPP2 was originally designated as polycystin kidney disease 2 (PKD2) and shown to combine with PKD1, a membrane protein with a large extracellular N‐terminal domain proposed to act as an extracellular sensing antenna for mechanical stimuli. Both TRPP2 and PKD1 are localized in the primary cilium of renal epithelial cells, which is essential for detecting laminar fluid flow (Praetorius and Spring, 2005). However, TRPV4, which is expressed in renal epithelial cells, may also associate with TRPP2
9. MS TRPCs
207
(Giamarchi et al., 2006). It remains to be determined if TRPC1 combines with TRPP2 in renal epithelial cells and whether knock out of TRPC1 and/or TRPV4 blocks fluid flow detection. 5. TRPC1 Is Expressed in Specialized Mechanosensory Nerve Endings If TRPC1 is a mechanosensory channel, it should be expressed in specialized mechanosensory nerve endings. Glazebrook et al. (2005) used immunocytochemical techniques to examine the distribution of TRPC1 and TRPC3–7 in the soma, axons, and sensory terminals of arterial mechanoreceptors, and found that TRPC1, TRPC3, TRPC4, and TRPC5 were expressed in the peripheral axons and the mechanosensory terminals. However, only TRPC1 and TRPC3 extended into the low‐threshold mechanosensory complex endings, with TRPC4 and TRPC5 mainly limited to the major branches of the nerve. Although these results are consistent with TRPC1 (and possibly TRPC3) involvement in baroreception, it was concluded that because TRPC1 was not present in all fine terminals that it more likely modulated than directly mediated mechanotransduction. However, it is not clear that all fine endings are capable of transduction. Furthermore, other putative MS proteins (i.e., b and g ENaC subunits) are expressed in baroreceptor nerve terminals (Drummond et al., 1998) in which case diVerent classes of MS channels (i.e., ENaC and TRPC) may mediate mechanotransduction in diVerent mechanosensory nerves. 6. TRPC1 Involvement in Wound Closure and Cell Migration The first study to implicate TRPC1 in cell migration was by Moore et al. (1998). They proposed that shape changes induced in endothelial cells by activation of TRPC1 were necessary step for angiogenesis. In another study, it was demonstrated that TRPC1 overexpression promoted, while TRPC1 supression inhibited intestinal cell migration measured by wound closure assay (Rao et al. (2006). On the basis of the proposal that MscCa regulates fish keratocyte cell migration (Lee et al., 1999) and the identification of TRPC1 as a MscCa subunit (Maroto et al., 2005), the role of TRPC1 was tested on migration of the highly invasive/metastatic prostate tumor cell line PC‐3. TRPC1 activity was shown to be essential for PC‐3 cell migration and Gd3þ, GsMTx4, anti‐TRPC1 antibody, and siRNA‐targeting TRPC1 were shown to block PC‐3 migration by inhibiting Ca2þ dynamics required of cell migration (Maroto et al., 2007, submitted for publication). 7. Reconstitution of TRPC1 as an MS Channel To identify the protein forming the oocyte MscCa, oocyte membrane proteins were detergent solubilized, fractionated by FPLC, reconstituted
208
Hamill and Maroto
in liposomes, and assayed for MscCa activity using patch recording (Maroto et al., 2005). A specific protein fraction that ran with a conductivity of 16 mS cm 1 was shown to reconstitute the highest MscCa activity and silver‐stained gels indicated it displayed the highest abundance of 80‐kDa protein. On the basis of previous studies that identified xTRPC1 and hTRPC1 as forming an 80‐kDa protein when expressed oocytes (Bobanovic et al., 1999; Brereton et al., 2000), immunological methods were used to demonstrate that TRPC1 was present in the MscCa active fraction. Furthermore, heterologous expression of the hTRPC1 was shown to greatly increase the MscCa activity expressed in the transfected oocyte, whereas TRPC1‐ antisense reduced the endogenous MscCa activity (Maroto et al., 2005). Figure 2 compares MscCa activity in cell‐attached patches on a control oocyte (Fig. 2A) and an oocyte that had been injected with hTRPC1 (Fig. 2B). Despite the almost tenfold increase in current density in the TRPC1‐injected oocyte, channel activation and deactivation kinetics in the two patches were similar. However, in some patches, even on the same oocyte, the kinetics of the TRPC1‐dependent channels show delayed activation and deactivation kinetics. An example of the slow kinetics is illustrated for a patch that was formed on an oocyte that had been injected with TRPC1 with enhanced green fluorescence protein (eGFP) attached to the C‐terminus. Figure 3 shows confocal fluorescence images of the oocyte at low magnification and at high magnification indicating eGFP‐TRPC1 concentrated in the surface membrane (Fig. 3). Figure 4 compares the patch response on a control oocyte and the slow kinetics response of a patch formed on the oocyte displayed in Fig. 3. The basis for the heterogeneity in kinetics of TRPC1 channels may reflect local diVerences in the underlying CSK and/or bilayer or even the MscCa subunit composition that occurs with TRPC1 overexpression. Maroto et al. (2005) also demonstrated that hTRPC1 expression in CHO cells results in increased MscCa activity, consistent with an approximately fivefold greater increase in channel density. The presence of endogenous MscCa activity is consistent with previous reports that indicate CHO cells express TRPC1 along with TRPC2–6 (Vaca and Sampieir, 2002). Although the above results provide compelling evidence that TRPC1 is a structural component on the MscCa, the current increase in TRPC1‐transfected oocytes and CHO cells is relatively low compared with that achieved by overexpression of other channel types. This may be because endogenous TRPC1 needs to combine with endogenous TRPCs or other ancillary proteins. On the other hand, the ability to reconstitute MscCa activity following 5000‐fold protein to lipid dilution would seem to argue against the requirement of at least ancillary proteins that are not firmly attached to the channel complex.
209
9. MS TRPCs A
Control oocyte
B
hTRPC1 expressing oocyte
100 mmHg
4s 400 ms
175 pA
FIGURE 2 MS current activity measured in a control and an hTRPC1‐expressing oocyte. (A) Stepwise increase in suction (top trace) applied to a cell‐attached patch formed on a control oocyte (i.e., that was injected with 50 nl of water 4 days earlier) induced a current of 12 pA. (B) Similar to A except that the patch was formed on an hTRPC1‐expressing oocyte (i.e., injected 4 days earlier with 50 nl of TRPC1 transcripts). In this case, the peak current produced was 175 pA. Examination of the residual channels immediately after the steps indicates the same single‐channel currents of 2 pA. Both recordings were made at a patch potential of 50 mV. (From Maroto et al., 2005).
210
Hamill and Maroto
FIGURE 3 Fluorescence images of an oocyte that had been injected 3 days earlier with mRNA construct encoding enhanced green fluorescence protein attached to the C‐terminus of Trpc1. Upper panel shows whole oocyte. Lower panel shows confocal images focussed on the oocyte edge.
B. TRPC2 At present there is no evidence, direct or indirect, to indicate TRPC2 forms an MS channel. The current view is that it may function as a ROC or a SOC depending on cell type (Vannier et al., 1999; Gailly and Colson‐Van Schoor, 2001; Chu et al., 2004; Zufall et al., 2005). For example, TRPC2 has been implicated in pheromone detection in the rodent vomeronasal organ (VNO; Liman et al., 1999) because TRPC2 / mice lack gender discrimination (Zufall et al., 2005). Because a DAG‐activated channel in VNO neurons is downregulated in TRPC2 / mice and TRPC2 is localized in the sensory microvilli that lack Ca2þ stores, it seems that TRPC2 functions as a ROC rather than a SOC at least in VNO neurons (Spehr et al., 2002; Zufall et al., 2005). However, in erythroblasts, and possibly sperm, TRPC2 has been reported to be activated by store depletion. In both cell types, the long splice
211
9. MS TRPCs A
100 mmHg Control
B hTRPC1
50 pA
5s FIGURE 4 Oocyte patches formed on TRPC1‐expressing oocytes can show slow turn‐on and delayed turn‐oV in addition to the greatly enhanced currents. (A) Cell‐attached patch on a control oocyte showing step responses similar to Fig. 2A. (B) Cell‐attached patch on the oocyte shown in Fig. 3 that had been injected with hTRPC1 transcripts. In this case, the pressure‐ stimulated currents were not only much larger than the wild‐type responses but also failed to saturate and exhibited a pronounced delay in both its turn‐on and turn‐oV with the pressure steps. Fast responses similar to Fig. 2B were also seen on this oocyte.
variants of TRPC2 were detected (Yildrin et al., 2003), whereas VNO neurons express the short splice variant (Hofmann et al., 2000; Chu et al., 2002). In sperm, TRPC2 may participate in the acrosome reaction‐based inhibition by a TRPC2 antibody in vitro (Jungnickel et al., 2001). However, TRPC2 / mice display normal fertility therefore casting doubt on this role (Stamboulian et al., 2005). In hematopoiesis, erthyropoietin modulates Ca2þ influx via TRPC2 in possible combination with TRPC6 (Chu et al., 2002, 2004).
212
Hamill and Maroto
C. TRPC3 As with TRPC2, there is no evidence yet for TRPC3 mechanosensitivity. However, TRPC3 does colocalize with TRPC1 in specialized mechanosensory nerve endings, indicating that both may combine to form an MS channel (Glazebrook et al., 2005). The growing consensus is that TRPC3 can contribute to both SOC and ROC channels depending on expression levels (Zitt et al., 1997; Hofmann et al., 1999; Hurst et al., 1998; Kamouchi et al., 1999; Trebak et al., 2002; Putney et al., 2004; Vazquez et al., 2005; Groschner and Rosker, 2005; Liu et al., 2005; Zagranichnaya et al., 2005; Kawasaki et al., 2006). Suppression of TRPC3 in cerebral arterial smooth muscle while suppressing pyridine receptor‐induced depolarization does not alter the pressure‐ increased depolarization and contraction, which appears to be dependent on TRPC6 (Reading et al., 2005). On the other hand, TRPC3 activation appears to depend upon Src kinase that may be MS (Vazquez et al., 2004b) and like TRPC6 is directly activated by OAG (Hofmann et al., 1999). D. TRPC4 There is disagreement on whether TRPC4 functions as a SOC and/or ROC (Philipp et al., 1996; Tomita et al., 1998; McKay et al., 2000; Schaefer et al., 2000; Plant and Schaefer, 2005). TRPC4 has been suggested to form a ROC activated by AA (Wu et al., 2002; Zagranichnaya et al., 2005). In particular, using siRNA and antisense strategies to reduce endogenous TRPC4 expression, TRPC4 was shown to be required for the OAG‐induced and receptor‐operated Ca2þ entry as well as the AA‐induced Ca2þ oscillations but not for SOC function. This AA activation may have implications for the mechanosensitivity of TRPC4 since AA has been shown to activate a variety of MS channels in the absence of applied stretch where it appears to act by directly altering mechanical properties of the bilayer surrounding the channel (Kim, 1992; Hamill and McBride, 1996; Casado and Ascher, 1998; Patel et al., 2001). Studies of TRPC4 / mice indicate TRPC4 is an essential determinant of endothelial vascular tone and endothelial permeability as well as neurotransmitter release from central neurons (reviewed by Freichel et al., 2004). E. TRPC6 The general consensus is that TRPC6 forms a ROC that is activated by DAG in a membrane‐delimited fashion and is insensitive to activation by IP3 and store depletion (Boulay et al., 1997; Hofmann et al., 1999; Estacion
9. MS TRPCs
213
et al., 2004; Zagranichnaya et al., 2005; Zhang et al., 2006). Although TRPC6 is a member of the TRPC3/6/7 subfamily, it shows distinct functional and structural properties. Functionally, while TRPC6 only forms a ROC, TRPC3 and TRPC7 appear capable of participating in forming both ROCs and SOCs (Zagranichnaya et al., 2005); structurally, whereas TRPC6 carries two extracellular glycosylation sites, TRPC3 carries only one (Dietrich et al., 2003). Furthermore, exogenously expressed TRPC6 shows low basal activity compared with TRPC3 and elimination of the extra glycosylation site that is missing in TRPC3 transforms TRPC6 into a constitutively active TRPC‐ 3‐like channel. Conversely, engineering of an additional glycosylation site in TRPC3 markedly reduces TRPC3 basal activity. 1. TRPC6 as a Regulator of Myogenic Tone TRPC6 is proposed to mediate the depolarization and constriction of small arteries and arterioles in response to adrenergic stimulation (Inoue et al., 2001; Jung et al., 2002; Inoue et al., 2006), and elevation of intravascular pressure consistent with TRPC6 forming a MOC as well as a ROC (Welsh et al., 2000, 2002). The cationic current activated by pressure in vascular smooth muscle is suppressed by antisense‐DNA to TRPC6 (Welsh et al., 2000). Furthermore, because the cation entry was stimulated by OAG and inhibited by PLC inhibitor (Park et al., 2003), it was proposed that TRPC6 forms an MS channel that is activated indirectly by pressure according to the pathway: " intravasular pressure !" PLC !" ½DAG !" TRPC !" ½Ca2þ !" myogenic tone In this scheme, it is PLC rather than TRPC that is MS. This would imply that since all TRPCs are coupled to PLC‐dependent receptors, they may all display mechanosensitivity. However, while there are reports that PLC can be mechanically stimulated independent of external Ca2þ (Mitchell et al., 1997; Rosales et al., 1997; Moore et al., 2002), there are also studies that indicate the mechanosensitivity of PLC derives from stimulation by Ca2þ influx via MscCa (Matsumoto et al., 1995; Ryan et al., 2000; Ruwhof et al., 2001). In this case, it becomes important to demonstrate that TRPC6 can be mechanically activated in the absence of external Ca2þ (e.g., using Ba2þ). There is other evidence to indicate TRPC6 may be coupled to other MS enzymes. For example, TRPC6 is similar to TRPV4 in that it is activated by 20‐hydroxyeicosatetraenoic acid (20‐HETE), which is the dominant AA metabolite produced by cytochrome P‐450 w‐hydroxylase enzymes (Basora et al., 2003). TRPC6 may also be activated by Src family protein tyrosine kinase (PTK)‐mediated tyrosoine phosphorylation (Hisatsune et al., 2004). Indeed, PP2, a specific inhibitor of Src
214
Hamill and Maroto
PTKs, abolishes TRPC6 (and TRPC3) activation and strongly inhibits OAG‐ induced Ca2þ entry (Soboloff et al., 2005). OAG may operate solely through TRPC6 homomers, the cation of vasopressin (VP) may also include the OAG‐ insensitive TRPC heteromers (e.g., TRPC1 and TRPC6). A further complication is that DAG‐dependent activation of PKC appears to stimulate the myogenic channels based on their block by the PKC inhibitor chelerythrine (Slish et al., 2002), whereas PKC activation seems to inhibit TRPC6 channels, which would seem more consistent with direct activation by DAG/OAG (SoboloV et al., 2005). Despite the above evidence implicating TRPC6 as the ‘‘myogenic’’ channel, TRPC6‐deficient mice show enhanced rather than reduced myotonic tone and increased rather than reduced responsiveness to constrictor agonist in small arteries. These eVects result in both a higher elevated mean arterial blood pressure and a shift in the onset of the myogenic tone toward lower intravascular pressures, again opposite to what would be expected if TRPC6 was critical for myoconstriction (Dietrich et al., 2005). Furthermore, isolated smooth muscle from TRPC6 / mice show increased basal cation entry and more depolarized resting potentials, but both eVects are blocked if the muscles are also transfected with siRNA‐targeting TRPC3. On the basis of this last observation, it was suggested that constitutively active TRPC3 channels are upregulated in TRPC6 / mice. However, the TRPC3 subunits are unable to functionally replace the lost TRPC6 function that involves suppression of high basal TRPC3 activity (i.e., the TRPC3/TRPC6 heteromer is a more tightly regulated ROC and/or MOC). In summary, although evidence indicates TRPC6 may be a pressure or stretch‐sensitive channel and contribute to MOC, the TRPC6 knockout mouse indicates a phenotype that cannot be explained if TRPC6 alone forms the vasoconstrictor channel. It may also be relevant that another study could find no evidence that Gd3þ‐sensitive MscCa contributes to myogenic tone in isolated arterioles from rat skeletal muscle (Bakker et al., 1999). 2. TRPC6 as a Regulator of the Kidney Slit Diaphragm Autosomal dominant focal segmental glomerulosclerosis (FSGS) is a kidney disease that leads to progressive renal kidney failure characterized by leakage of plasma proteins like albumin into the urine (proteinuria). Mutations in TRPC6 were associated with familial FSGS and implicated in aberrant Ca2þ signaling that leads to podocyte injury (Reiser et al., 2005; Winn et al., 2005). Furthermore, two of the mutants were demonstrated to be gain‐of‐function mutations that produce larger ROCs than the wild‐type TRPC6 expressed in HEK‐293 cells. Ultrafiltration of plasma by the renal glomeruli is mediated mainly by the podocyte, which is an epithelial cell that
9. MS TRPCs
215
lies external to the glomerular basement membrane (GBM) and lines the outer endothelium of the capillary tuft located inside the Bowman’s capsule. The podocyte covers the GBM and forms interdigitating foot processes that are connected by slit diaphragms—ultra‐thin membrane structures that form at the center of the slit a zipper‐like structure with pores smaller than albumin (Kriz, 2005; Tryggvason and Wartiovarara, 2005). The podocyte‐ specific proteins, nephrin and podocin, are localized in the slit diaphragm and the extracellular domains of nephrin molecules of neighboring foot processes interact to form the zipper structure. Podocin, a member of the stomatin family, is a scaVolding protein that accumulates in lipid rafts and interacts with the cytoplasmic domain of nephrin (Durvasula and Shankland, 2006). Both nephrin and podocin have been shown to be mutated in diVerent familial forms of FSGS. Furthermore, TRPC6 interacts with both nephrin and podocin and a nephrin deficiency in mice leads to overexpression and mislocalization of TRPC6 in podocyte as well as disruption of the slit diaphragm (Reiser et al., 2005). Mechanical forces play an important role in ultrafiltration both in terms of the high transmural distending forces arising from the capillary perfusion pressure as well as the intrinsic forces generated by the contractile actin network in the foot process that control, in a Ca2þ‐dependent manner, the width of the filtration slits. As a consequence, TRPC6 may act as the central signaling component mediating pressure‐ induced constriction at the slit. In summary, two quite diverse physiological functions, myogenic tone and renal ultrafiltration, implicate TRPC6 as an MS channel. However, whether TRPC6 acts as a direct mechanosensor as in the case of TRPC1 or is indirectly MS like TRPV4 remains to be determined.
IX. CONCLUSIONS At least three basic mechanisms referred to as ‘‘bilayer,’’ ‘‘conformational coupling,’’ and ‘‘enzymatic’’ may confer mechanosensitivity on TRPCs. The bilayer mechanism should operate if the TRPC, in shifting between closed and open states, undergoes a change in its membrane‐occupied area, thickness, and/or cross‐sectional shape. Any one of these changes would confer mechanosensitivity on the channel. A bilayer mechanism may also underlie the ability of lipidic second messengers (e.g., DAG/OAG, AA, lysophospholipid and 50 ,60 ‐EET) to directly activate TRPCs by inserting in the bilayer to alter its local bilayer packing, curvature, and/or the lateral pressure profile. The only unequivocal way to demonstrate that a bilayer mechanism operates is to show that stretch sensitivity is retained when the purified channel protein is reconstituted in liposomes. At this stage, one can go onto measure
216
Hamill and Maroto
channel activity as a function of changing bilayer thickness (i.e., by using phospholipids with diVerent acyl length chains) and local curvature/pressure profile (i.e., by using lysophospholipids with diVerent shapes; Perozo et al., 2002; Martinac, 2007; Markin and Sachs, 2007; Powl and Lee, 2007). The second mechanism involves conformational coupling (CC) that has been evoked to account for TRPC sensitivity to depletion of internal Ca2þ stores. CC was originally used to explain excitation–contraction (E–C) coupling, involving the physical coupling between L‐type Ca2þ channel (i.e., dihydropyridine receptors, DHPR) in the plasma membrane and ryanodine receptors (RyR1) that release Ca2þ from the sarcoplasmic reticulum (SR; Protasi, 2002). Subsequently, a retrograde form of CC was discovered between the same two proteins that regulate the organization of the DHPR into tetrads and the magnitude of the Ca2þ current carried by DHPR (Wang et al., 2001; Paolini et al., 2004; Yin et al., 2005). Another form of CC was demonstrated associated with physiological stimuli that do not deplete Ca2þ stores yet activate Ca2þ entry through channels referred to as excitation‐ coupled Ca2þ entry channels to distinguish them from SOC (Cherednichenko et al., 2004). Interestingly, RyR1 is functionally coupled to both TRPC1‐ dependent SOC and TRPC3‐dependent SR Ca2þ release (Sampieri et al., 2005; Lee et al., 2006). A key issue for all forms of CC is whether the direct physical link that conveys mechanical conformational energy from one protein to another can also act as a pathway to either focus applied mechanical forces on the channel or alternatively constrain the channel from responding to mechanical forces generated within the bilayer. Another possibility is that reorganization or clustering of the resident ER protein (i.e., STIM) that senses Ca2þ stores may alter channel mechanosensitivity by increasing the strength of CC coupling (Kwan et al., 2003). Some insights into these possibilities can be provided by the process of ‘‘membrane blebbing,’’ which involves decoupling of the plasma membrane from the underlying CSK and has been shown to either increase or decrease the mechanosensitivity of MS channels depending on the channel (Hamill and McBride, 1997; Hamill, 2006). Since membrane blebbing would also be expected to disrupt any dynamic interactions between TRPC and scaVolding proteins, it should alter TRPC function. In one case it has been reported that Ca2þ store depletion carried out after but not before formation of a tight seal is eVective in blocking the activation of SOC channels in the frog oocyte patches (Yao et al., 1999). Presumably, this occurs because the sealing process physically decouples the channels from ER proteins that sense internal Ca2þ stores. Tight seal formation using strong suction can also reduce MscCa mechanosensitivity and gating kinetics possibly by a related mechanism (Hamill and McBride, 1992). On the other hand, it has been reported that
9. MS TRPCs
217
ICRAC is retained following cell ‘‘ballooning’’ (i.e., a form of reversible membrane blebbing), indicating that the coupling between the channel and the Ca2þ sensor STIM may be relatively resistant to decoupling (Bakowski et al., 2001). In any case, in order to directly demonstrate a role for CC in mechanosensitivity, one needs to show that stretch sensitivity can be altered in mutants in which TRPC–ancillary protein interactions are disrupted (Section VII.D). The third mechanism of mechanosensitivity relates to functional coupling between TRPCs and MS enzymes. Apart from the PLA2 and Src that are MS and have been implicated in conferring mechanosensitivity on TRPV4 (Vriens et al., 2004; Cohen, 2005a,b), there is growing evidence that PLC is also MS with reports indicating that mechanosensitivity is either dependent on external Ca2þ and Ca2þ influx (Matsumoto et al., 1995; Ryan et al., 2000; Ruwhof et al., 2001; Alexander et al., 2004) or Ca2þ independent (Mitchell et al., 1997; Rosales et al., 1997; Moore et al., 2002). In either case, these studies indicate that mechanical forces transduced by MscCa and/or by MS enzymes may modulate the gating of all TRP channels. It remains to be determined what are the physiological and/or pathological eVects of this MS modulation? The methods discussed in this chapter, including the applications of pressure steps to measure the kinetics of MS enzyme–channel coupling and the use of membrane protein liposome reconstitution for identifying specific protein–lipid interactions, should play an increasing role in understanding the importance of the diVerent MS mechanisms underlying TRPC functions.
Note Added in Proof Spassova, M. A., Hewavitharana, T., Xu, W., Soboloff, J., and Gill, D. L. (Proc. Natl. Acad. Sci. USA 103, 16586–16591) have reported that overexpression of hTRPC6 in mammalian cells results in increased OAG‐ and swelling‐activated whole cell currents and increased stretch‐activated channel activity in inside‐out patches. The TRPC6 activity was blocked by GsmTX4 but was insensitive to block by the PLC inhibitor U73122 (c.f., Park et al., 2003). Furthermore, they found that the long delays associated with stretch activation of TRPC6 channels could be reduced by treatment of cells with cytochalasin D. These results are consistent with TRPC6 being directly MS and a common bilayer mechanism underlying OAG‐ and stretch‐activation of TRPC6.
218
Hamill and Maroto
Acknowledgments We thank the Department of Defense, Prostate Cancer Research Program and the National Cancer Institute for their funding support.
References Ackerman, M. J., Wickman, K. D., and Clapham, D. E. (1994). Hypotonicity activates a native chloride current in Xenopus oocyte. J. Gen. Physiol. 103, 153–179. Ahmmed, G. U., Mehta, D., Vogel, S., Holinstat, M., Paria, B. C., Tiruppathi, C., and Malik, A. B. (2004). Protein kinase Ca phosphorylates the TRPC1 channels and regulates store‐operated Ca2þ entry in endothelial cells. J. Biol. Chem. 279, 20941–20949. Alexander, L. D., Alagarsamy, S., and Douglas, J. G. (2004). Cyclic stretch‐induced cPLA2 mediates ERK ½ signaling in rabbit proximal tubule cells. Kidney Int. 65, 551–563. Allen, D. G., Whitehead, N. P., and Yeung, E. W. (2005). Mechanisms of stretch‐induced muscle damage in normal and dystrophic muscle: Role of ionic changes. J. Physiol. (Lond.) 567(Pt. 3), 723–735. Ambudkar, I. S. (2006). Ca2þ signaling microdomains: Platforms for the assembly and regulation of the assembly and regulation of TRPC channels. Trends Pharmacol. Sci. 27, 25–32. Andersen, L., and Seilhamer, J. A. (1997). Comparison of selected mRNA and protein abundance in human liver. Electrophoresesis 18, 533–537. Antoniotti, S., Pla, A. F., Barrel, S., Scalabrino, L., and Vovisolo, D. (2006). Interaction between TRPC subunits in endothelial cells. J. Recept. Signal. Transduc. 26, 225–240. Bakker, E. N. T. P., Krkhof, C. J. M., and Sipkema, P. (1999). Signal transduction in spontaneous myogenic tone insolated arterioles from rat skeletal muscle. Cardiovasc. Res. 41, 229–236. Bakowski, D., Glitsch, M. D., and Parekh, A. B. (2001). An examination of the secretion‐like coupling model for the activation of the Ca2þ release‐activated Ca2þ current Icrac in RBL‐1 cells. J. Physiol. 532(Pt. 1), 55–71. Barsanti, C., Pellegrini, M., and Pellegrino, M. (2006a). Regulation of the mechanosensitive cation channels by ATP and cAMP in leech neurons. Biochem. Biophys. Acta 1758, 666–672. Barsanti, C., Pellegrini, M., Ricci, D., and Pellegrino, M. (2006b). EVects of intracellular pH and Ca2þ on the activity of stretch‐sensitive cation channels in leech neurons. Pflu¨gers Arch. 452, 435–443. Basavappa, S., Pesersen, S. F., Jorgense, N. K., Ellory, J. C., and HoVmann, E. K. (1988). Swelling‐induced arachidonic acid release via a 85 kda cPLA2 in human neuroblastoma cells. J. Neurophysiol. 79, 1441–1449. Basora, N., Boulay, G., Biloddeau, L., Rousseau, E., and Marcel, D. P. (2003). 20‐Hydroxyeicosatetraenocic acid (20‐HETE) activates mouse TRPC6 channels expressed in HEK293 cells. J. Biol. Chem. 278, 31709–31716. Bear, C. E. (1990). A nonselective cation channel in rat liver cells is activated by membrane stretch. Am. J. Physiol. 258, C421–C428. Beech, D. J. (2005). TRPC1: Store‐operated channel and more. Pflu¨gers Arch. 451, 53–60. Beech, D. J., Xu, S. Z., and Flemming, M. R. (2003). TRPC1 store operated cationic channel subunit. Cell Calcium 33, 433–440. Besch, S. R., Suchyna, T., and Sachs, F. (2002). High speed pressure clamp. Pflu¨gers Arch. 445, 161–166.
9. MS TRPCs
219
Bielfeld‐Ackermann, A., Range, C., and Korbmacher, C. (1998). Maitotoxin (MTX) activates a nonselective cation channel in Xenopus laevis oocytes. Pflu¨gers Arch. 436, 329–337. Bobanovic, L. K., Laine, M., Petersen, C. C. H., Bennett, D. L., Berridge, M. J., Lipp, P., Ripley, S. J., and Bootman, M. D. (1999). Molecular cloning and immunolocalization of a novel vertebrate trp homologue from Xenopus. Biochem. J. 340, 593–599. Boittin, F. X., Pettermann, O., Hirn, C., Mittaud, P., Dorchies, O. M., Roulet, E., and Ruegg, U. T. (2006). Ca2þ‐independent phospholipase A2 enhances store‐operated Ca2þ entry in dystrophic skeletal muscle fibres. J. Cell Sci. 119, 3733–3742. Bolotina, V. M., and Csutora, P. (2005). CIF and other mysteries of the store‐operated Ca2þ‐ entry pathway. Trends Neurosci. 30, 378–387. Boulay, G., Zhu, X., Peyton, M., Jiang, M., Hurst, R., Stefani, E., and Birnbaumer, L. (1997). Cloning and expression of a novel mammalian homolog of Drosophila transient receptor potential (Trp) involved in calcium entry secondary to activation of receptors coupled by the Gq class of G protein. J. Biol. Chem. 272, 29672–29680. Brazier, S. W., Singh, B. B., Liu, X., Swaim, W., and Ambudkar, I. S. (2003). Caveolin‐1 contributes to assembly of store‐operated Ca2þ influx channels by regulating plasma membrane localization of TRPC1. J. Biol. Chem. 29, 27208–27215. Brereton, H. M., Harland, M. L., Auld, A. M., and Barritt, G. J. (2000). Evidence that the TRP‐1 protein is unlikely to account for store‐operated Ca2þ inflow in Xenopus laevis oocytes. Mol. Cell. Biochem. 214, 63–74. Brereton, H. M., Chen, J., Rychkov, G., Harland, M. L., and Barritt, G. J. (2001). Maitotoxin activates an endogenous non‐selective cation channel and is an eVective initiator of the activation of the heterologously expressed hTRPC‐1 (transient receptor potential) non‐selective cation channel in H4‐IIE liver cells. Biochem. Biophys. Acta 1540, 107–126. Bugaj, V., Alexeenko, V., Zubov, A., Glushankova, L., Nikalaev, A., Wang, Z., Kaznaceyeva, I., and Mozhayeva, G. N. (2005). Functional properties of endogenous receptor and store operated calcium influx channels in HEK293 cells. J. Biol. Chem. 280, 16790–16797. Callamaras, N., Sun, X. P., Ivorra, I., and Parker, I. (1998). Hemispheric asymmetry of macroscopic and elementary calcium signals mediated by InsP3 in Xenopus oocytes. J. Physiol. 511, 395–405. Casado, M., and Ascher, P. (1998). Opposite modulation of NMDA receptors by lysophospholipids and arachidonic acid: Common features with mechanosensitivity. J. Physiol. (Lond.) 513, 317–330. Caterina, M. J., Schumacher, M. A., Tominaga, M., Rosen, T. A., Levine, J. D., and Julius, D. (1997). The capsacian receptor: A heat activated ion channel in the pain pathway. Nature 389, 816–824. Cemerikic, D., and Sackin, H. (1993). Substrate activation of mechanosensitive, whole cell currents in renal proximal tubule. Am. J. Physiol. 264, F697–F714. Chen, J., and Barritt, G. J. (2003). Evidence that TRPC1 (transient receptor potential canonical 1) forms a Ca2þ‐permeable channels linked to the regulation of cell volume in liver cells obtaining using small interfering RNA targeted against TRPC1. Biochem. J. 373, 327–336. Chemin, J., Patel, A. J., Duprat, F., Lauritzen, I., Lazdunski, M., and Honore, E. (2005). A phospholipid sensor controls mechanogating of the Kþ channel TREK‐1. EMBO J. 24, 44–53. Cherednichenko, G., Hurne, A. M., Fessenden, J. D., Lee, E. H., Allen, P. D., Beam, K. G., and Pessah, I. N. (2004). Conformational activation of Ca2þ entry by depolarization of skeletal myotubes. Proc. Natl. Acad. Sci. USA 101, 15793–15798.
220
Hamill and Maroto
Christensen, O. (1987). Mediation of cell volume regulation by Ca2þ influx through stretch activated cation channels. Nature 330, 66–68. Chu, X., Cheung, J. Y., Barber, D. L., Birnbaumer, L., Rothblum, L. I., Conrad, K., Abrason, V., Chan, Y., Stahl, R., Carey, D. J., and Miller, B. A. (2002). Erythropoietin modulates calcium influx through TRPC2. J. Biol. Chem. 277, 34375–34382. Chu, X., Tong, Q., Cheung, J. Y., Wozney, J., Conrad, K., Maznack, V., Zhang, W., Stahl, R., Barber, D. L., and Miller, B. A. (2004). Interaction of TRPC2 and TRPC6 in erythropoietin modulation of Calcium influx. J. Biol. Chem. 279, 10514–10522. CioY, D. L., Wu, S., Alexeyev, M., Goodman, S. R., Zhu, M. X., and Stevens, T. (2005). Activation of the endothelial store‐operated ISOC Ca2þ channel requires interaction of protein 4.1 with TRPC4. Circ. Res. 97, 1164–1172. Clapham, D. E. (2003). TRP channels as cellular sensors. Nature 426, 517–524. Cohen, D. M. (2005a). SRC family kinases in cell volume regulation. Am. J. Physiol. 288, C483–C493. Cohen, D. M. (2005b). TRPV4 and the mammalian kidney. Pflu¨gers Arch. 451, 168–175. Cosens, D. J., and Manning, A. (1969). Abnormal electroretinogram from a Drosophila mutant. Nature 224, 285–287. Delmas, P. (2004). Assembly and gating of TRPC channels in signaling microdomains. Novartis Found. Symp. 258, 75–97. Delmas, P., Wanaverbecq, N., Abogadie, F. C., Mistry, M., and Brown, D. A. (2002). Signaling microdomains define the specificity of receptor‐mediated INsP3 pathways in neurons. Neuron 14, 209–220. Diakov, A., Koch, J. P., Ducoudret, O., Mueler‐Berger, S., and Fro¨mter, E. (2001). The disulfoic stilbene DIDS and the marine toxin maitotoxin activated the same two types of endogenous cation conductance in the cell membrane of Xenopus laevis oocytes. Pflu¨gers Arch. 442, 700–708. Dietrich, A. M., Schnitzler, M. M., Emmel, J., Kallwa, H., Hofmann, T., and Gundermann, T. (2003). N‐linked protein glycosylation is a major determinant for basal TRPC3 and TRPC6 channel activity. J. Biol. Chem. 278, 47842–47852. Dietrich, A., Schnitzker, M. M., Gollasch, M., Gross, V., Storch, U., Dubrovska, G., Obst, M., Yildirim, E., Salanova, B., Kalwa, H., Essin, K., Pinkenburg, O., et al. (2005). Increased vascular smooth muscle contractility in TRPC6 / mice. Mol. Cell. Biol. 25, 6980–6989. Drummond, H. A., Price, M. P., Welsh, M. J., and Abboud, F. M. (1998). A molecular component of the arterial baroreceptor mecahotransducer. Neuron 21, 1435–1441. Ducret, T., Vanebrouck, C., Cao, M. L., Lebacq, J., and Gailly, P. (2006). Functional role of store‐operated and stretch‐activated channels in murine adult skeltal muscle fibers. J. Physiol. 575(Pt. 3), 913–924. Durvasula, R. V., and Shankland, S. J. (2006). Podocyte injury and targeting therapy: An update. Curr. Opin. Nephrol. Hypertens. 15, 1–7. Escobar, L. I., Salvador, C., Martinez, M., and Vaca, L. (1998). Maitotoxin, a cationic channel activator. Neurobiology 6, 59–74. Estacion, M., Li, S., Sinkins, W. G., Gosling, M., Bahra, P., Poll, C., Westwick, J., and Schilling, W. P. (2004). Activation of human TRPC6 channels by receptor stimulation, J. Biol. Chem. 279, 22047–22056. Flemming, R., Xu, S. Z., and Beech, D. J. (2003). Pharmacological profile of store‐operated channels in cerebral arteriolar smooth muscle cells. Br. J. Pharmacol. 139, 955–965. Fong, P., Turner, P. R., Denetclaw, W. F., and Steinhardt, R. A. (1990). Increased activity of calcium leak channels in myotubes of Ducheene human and mdx mouse origin. Science 250, 673–676.
9. MS TRPCs
221
Franco, A., and Lansman, J. B. (1990). Calcium entry through stretch‐inactivated channels in mdx myotubes. Nature 344, 670–673. Franco‐Obregon, A., and Lansman, J. B. (2002). Changes in mechanosensitive channel gating following mechanical stimulation in skeletal muscle myotubes from the mdx mouse. J. Physiol. (Lond.) 539(Pt. 2), 391–407. Freichel, M., Vennekens, R., Olausson, J., HoVmann, M., Mu¨ller, C., Stolz, S., Scheunemann, J., Weissgerber, P., and Flockerzi, V. (2004). Functional role of TRPC proteins in vivo: Lessons from TRPC‐deficient mouse models. Biochem. Biophys. Res. Commun. 322, 1352–1358. Gailly, P., and Colson‐Van Schoor, M. (2001). Involvement of TRP2 protein in store‐operated influx of calcium in fibroblasts. Cell Calcium 30, 157–165. Giamarchi, A., Padilla, F., Coste, B., Raoux, M., Crest, M., Honore, E., and Delmas, P. (2006). The versatile nature of the calcium‐permeable cation channel TRPP2. EMBO Rep. 7, 787–793. Glazebrook, P. A., Schilling, W. P., and Kunze, D. L. (2005). TRPC channels as signal transducers. Pflu¨gers Arch. 451, 125–130. Goel, M., Sinkins, W. G., Zuo, C. D., Estacion, M., and Schilling, W. P. (2006). Indentification and localization of TRPC channels in the rat kidney. Am. J. Physiol. 290, F1241–F1252. Golovina, V. A., Platoshyn, O., Bailey, C. L., Wang, J., Limsuwan, A., Sweeney, M., Rubin, L. J., and Yuan, J. X. (2001). Upregulated TRP and enhanced capacitative Ca2þ entry in human pulmonary artery myocytes during proliferation. Am. J. Physiol. 280, H746–H755. Groschner, K., and Rosker, C. (2005). TRPC3: A versatile transducer molecule that serves integration and diversification of cellular signals. Naunyn Schmiedebergs Arch. Pharmacol. 371, 251–256. Gottlieb, P. A., Suchyna, T. M., Ostrow, L. W., and Sachs, F. (2004). Mechanosensitive ion channels as drug targets. Curr. Drug Targets 3, 287–295. Gu, C. X., Juranka, P. F., and Morris, C. E. (2001). Stretch‐activation and stretch‐inactivation of Shaker‐IR, a voltage‐gated Kþ channel. Biophys. J. 80, 2678–2693. Guharay, F., and Sachs, F. (1984). Stretch‐activated single ion channel currents in tissue cultured embryonic chick skeletal muscle. J. Physiol. (Lond.) 352, 685–701. Hamill, O. P. (1983). Potassium and chloride channels in red blood cells. In ‘‘Single Channel Recording’’ (B. Sakmann and E. Neher, eds.), pp. 451–471. Plenum Press, New York. Hamill, O. P. (2006). Twenty odd years of stretch activated channels. Pflu¨gers Arch. 453, 333–351. Hamill, O. P., and Martinac, B. (2001). Molecular basis of mechanotransduction in living cells. Physiol. Rev. 81, 685–740. Hamill, O. P., and McBride, D. W., Jr. (1992). Rapid adaptation of the mechanosensitive channel in Xenopus oocytes. Proc. Natl. Acad. Sci. USA 89, 7462–7466. Hamill, O. P., and McBride, D. W., Jr. (1996). The pharmacology of mechanogated membrane ion channels. Physiol. Rev. 48, 231–252. Hamill, O. P., and McBride, D. W., Jr. (1997). Induced membrane hypo‐/hyper‐mechanosensitivity: A limitation of patch clamp recording. Annu. Rev. Physiol. 59, 621–631. Hamill, O. P., Marty, A., Neher, E., Sakmann, B., and Sigworth, F. (1981). Improved patch clamp techniques for high current resolution from cells and cell‐free membrane patches. Pflu¨gers Arch. 391, 85–100. Hill, A. J., Hinton, J. M., Cheng, H., Gao, Z., Bates, D. O., Hancox, J. C., Langton, P. D., and James, A. F. (2006). A TRPC‐like non‐selective cation current activated by a‐adrenoceptors in rat mesenteric artery smooth muscle cells. Cell Calcium 40, 29–40. Hille, B. (2001). ‘‘Ion Channels of Excitable Membranes’’ (3rd edn.), pp. 1–814. Sinauer, MA.
222
Hamill and Maroto
Hisatsune, C., Kuroda, Y., Nakamura, K., Inoue, T., Nakamura, T., Michikawa, T., Mizuntani, A., and Mikoshiba, K. (2004). Regulation of TRPC6 channel activity by tyrosine phosphorylation. J. Biol. Chem. 279, 18887–18894. Hofmann, T., Obukhov, A. G., Schaefer, M., Harteneck, C., Gudermann, T., and Schultz, G. (1999). Direct activation of human TRPC6 and TRPC3 channels by diacylglycerol. Nature 397, 259–263. Hofmann, T., SchaeVer, M., Schultz, G., and Gudermann, T. (2000). Cloning, expression and subcellular localization of two novel splice variants of mouse transient receptor potential 2. Biochem. J. 351, 115–122. Hofmann, T., SchaeVer, M., Schultz, G., and Gudermann, T. (2002). Subunit composition of mammalian transient receptor potential channels in living cells. Proc. Natl. Acad. Sci. USA 99, 7461–7466. Honore´, E., Maingret, F., Lazdunski, M., and Patel, A. J. (2002). An intracellular proton sensor commands lipid and mechano‐gating of the Kþ channel TREK‐1. EMBO J. 21, 2968–2976. Honore´, E., Patel, A. J., Chemin, J., Suchyna, T., and Sachs, F. (2006). Desensitization of mechano‐gated K2p channels. Proc. Natl. Acad. Sci. USA 103, 6859–6864. Hopf, F. W., Reddy, P., Hong, J., and Steinhardt, R. A. (1996). A capacitive calcium current in cultured skeletal muscle cells is medicated by the Ca2þ‐specific leak channel and inhibited by dihyropyridine compounds. J. Biol. Chem. 271, 22358–22367. Howard, J., and Bechstedt, S. (2004). Hypothesis: A helix of ankyrin repeats of the NOMPC‐ TRP ion channel is the gating spring of mechanoreceptors. Curr. Biol. 14, 224–226. Howard, J., Roberts, W. M., and Hudspeth, A. J. (1988). Mechanoelectrical transduction be hair cells. Annu. Rev. Biophys. Biophys. Chem. 17, 99–124. Huang, G. N., Zeng, W., Kim, J. Y., Yuan, J. P., Han, L., Muallem, S., and Worley, P. F. (2006). STIM1 carboxl‐terminus activates native SOC, ICRAC and TRPC1 channels. Nat. Cell Biol. 8, 1003–1010. Hurst, R. S., Zhu, X., Boulay, G., Birnbaumer, L., and Stefani, E. (1998). Ionic currents underlying HTRP3 mediated agonist‐dependent Ca2þ influx in stably transfected HEK293 cells. FEBS Lett. 422, 333–338. Inoue, R., Okada, T., Onoue, H., Hara, Y., Shimizu, S., Naitoh, S., Ito, Y., and Mori, Y. (2001). The transient receptor potential protein homologue TRP6 is the essential component of vascular a1‐adrenceptor‐activated Ca2þ‐permeable cation channel. Circ. Res. 88, 325–332. Inoue, R., Jensen, L. J., Shi, J., Morita, H., Nishida, M., Honda, A., and Ito, Y. (2006). Transient receptor potential channels in cardiovascular function and disease. Circ. Res. 99, 119–131. Jaconi, M., Pyle, J., Bortolon, R., Ou, J., and Clapham, D. (1999). Calcium release and influx colocalize to the endoplasmic recticulum. Curr. Biol. 7, 599–602. Jacques‐Fricke, B. T., Seow, Y., Gottlieb, P. C., Sachs, F., and Gomez, T. M. (2006). Ca2þ influx through mechanosensitive channels inhibits neurite outgrowth in opposition to other influx pathways and release of intracellular stores. J. Neurosci. 26, 5656–5664. Janssen, L. J., and Kwan, C. Y. (2007). ROCs and SOCs: What’s in a name? Cell Calcium. (published online). Jung, S., Strotmann, R., Schultz, G., and Plant, T. D. (2002). TRPC6 is a candidate channel involved in receptor‐stimulated cation currents in A7r5 smooth muscle cells. Am. J. Physiol. 282, C347–C359. Jung, S., Mu¨hle, A., Shaefer, M., Strotmann, R., Schultz, G., and Plant, T. D. (2003). Lanthanides potentiate TRPC5 currents by an action at the extracellular sites close to the pore mouth. J. Biol. Chem. 278, 3562–3571.
9. MS TRPCs
223
Jungnickel, M. K., Marrero, H., Birnbaumer, L., Lemos, J. R., and Florman, H. M. (2001). Trp2 regulates entry of Ca2þ into mouse sperm triggered by ZP3. Nat. Cell Biol. 3, 499–502. Kamouchi, M., Philipp, S., Flockerzi, V., Wissenbach, U., Mamin, A., Raemaekers, L., Eggermont, J., Droogmans, G., and Nilius, B. (1999). Properties of heterologously expressed hTRP3 channels in bovine pulmonary artery endothelial cells. J. Physiol. (Lond.) 518(Pt. 2), 345–359. Kawasaki, B. T., Liao, Y., and Birnbaumer, L. (2006). Role of Src in C3 transient receptor potential channel function and evidence for a heterogeneous makeup of receptor‐ and store‐operated Ca2þ entry channels. Proc. Natl. Acad. Sci. USA 103, 335–340. Kim, D. (1992). A mechanosensitive Kþ channel in heart‐cells: Activation by arachidonic acid. J. Gen. Physiol. 100, 1021–1040. Kiselyov, K., Xu, X., Mozayeva, G., Kuo, T., Pessah, I., Mignery, G., Zhu, X., Birnbaumer, L., and Muallen, S. (1998). Functional interaction between InsP3 receptors and store‐operated Htrp3 channels. Nature 396, 478–482. Kloda, A., and Martinac, B. (2001a). Structural and functional diVerences between two homolgous mechanosensitive channels of Methanocccus jannaschii. EMBO J. 20, 1888–1896. Kloda, A., and Martinac, B. (2001b). Mechanosensitive channel of Thermoplasma, the cell wall‐less Archaea: Cloning and molecular characterization. Cell. Biochem. Biophys. 34, 321–347. Kriz, W. (2005). TRPC6: A new podocyte gene involved in focal segmental glomerulosclerosis. Trends Mol. Med. 11, 527–530. Kumar, B., Dreja, K., Shah, S. S., Cheong, A., Xu, S. Z., Sukumar, P., Naylor, J., Forte, A., Cipollaro, M., McHugh, D., Kingston, P. A., Heagerty, A. M., et al. (2006). Upregulated TRPC1 channel in vascular injury in vivo and its role in human neoitimal hyperplasia. Circ. Res. 98, 557–563. Kung, C. (2005). A possible unifying principle for mechanosensation. Nature 436, 647–654. Kunichika, N., Yu, Y., Remillard, C. V., Platoshyn, O., Zhang, S., and Yuan, J. X. L. (2004). Overexpression of TRPC1 enhances pulmonary vasoconstriction induced by capacitative Ca2þ entry. Am. J. Physiol. 287, L962–L969. Kwan, H. Y., Leung, P. C., Huang, Y., and Yao, X. (2003). Depletion of intracellular Ca2þ stores sensitizes the flow‐induced Ca2þ influx in rat endothelial cells. Circ. Res. 92, 286–292. Lane, J. W., McBride, D. W., Jr., and Hamill, O. P. (1991). Amiloride block of the mechanosensitive cation channel in Xenopus oocytes. J. Physiol. 441, 347–366. Lane, J. W., McBride, D. W., Jr., and Hamill, O. P. (1992). Structure‐activity relations of amiloride and some of its analogues in blocking the mechanosensitive channel in Xenopus oocytes. Br. J. Pharmacol. 106(2), 283–286. Lee, E. H., Cherednichenko, G., Pessah, I. N., and Allen, P. D. (2006). Functional coupling between TRPC3 and RyR1 regulates the expressions of key triadic proteins. J. Biol. Chem. 281, 10042–10048. Lee, J., Ishihara, A., Oxford, G., Johnson, B., and Jacobson, K. (1999). Regulation of cell movement is mediated by stretch‐activated calcium channels. Nature 400, 382–386. Lehtonen, J. Y., and Kinnunen, P. K. (1995). Phospholipase A2 as a mechansensor. Biophys. J. 68, 1888–1894. Levina, N., To¨temeyer, S., Stokes, N. R., Louis, P., Jones, M. A., and Booth, I. R. (1999). Protection of Escherichia coli cells against extreme turgor pressure by activation of MscS and MscL mechanosensitive channels: Identification of genes for MscS activity. EMBO J. 18, 1730–1737.
224
Hamill and Maroto
Lievremont, J. P., Bird, G. S., and Putney, J. W., Jr. (2004). Canonical transient receptor potential TRPC7 can function as both a receptor‐and store‐operated channel in HEK‐293 cells. Am. J. Physiol. 287, C1709–C1716. Liman, E., Corey, D. P., and Dulac, C. (1999). TRP2: A candidate transduction channel for mammalian pheromone sensory signaling. Proc. Natl. Acad. Sci. USA 96, 5791–5796. Lindahl, M., Backman, E., Henriksson, K. G., Gorospe, J. R., and HoVman, E. P. (1995). Phospholipase A2 activity in dystrophinopathies. Neuromuscul. Disord. 5, 193–199. Lintschinger, B., Balzer‐Geldsetzer, M., Baskaran, T., Graier, W. F., Romanin, C., Zhu, M. X., and Groschner, K. (2000). Coassembly of Trp1 and Trp3 proteins generates diacylglycerol‐ and Ca2þ‐sensitive cation channels. J. Biol. Chem. 275, 27799–27805. Liu, X., Wang, W., Singh, B. B., Lockwich, T., Jadlowiec, J., O’Connell, B., Wellner, R., Zhu, M. X., and Ambudkar, I. S. (2000). TRP1, a candidate protein for the store‐operated Ca2þ influx mechanism in salivary gland cells. J. Biol. Chem. 275, 3403–3411. Liu, X., Singh, B. B., and Ambudkar, I. S. (2003). TRPC1 is required for functional store‐ operated channels. J. Biol. Chem. 278, 11337–11343. Liu, X., Bandyopadhyay, B. C., Singh, B. B., Groschner, K., and Ambudkar, I. S. (2005). Molecular analysis of a store‐operated and 2‐acetyl‐sn‐glycerol‐sensitive non‐selective cation channel. J. Biol. Chem. 280, 21600–21606. Lockwich, T. P., Liu, X., Singh, B. B., Jadlowiec, J., Weiland, S., and Ambudkar, I. S. (2000). Assembly of Trp1 in a signaling complex associated with caveolin‐scaVolding lipid raft domains. J. Biol. Chem. 275, 11934–11942. Ma, R., Rundle, D., Jacks, J., Koch, M., Downs, T., and Tsiokas, L. (2003). Inhibitor of myogenic family, a novel suppressor of store‐operated currents through an interaction with TRPC1. J. Biol. Chem. 278, 52763–52772. Markin, V. S., and Sachs, F. (2007). Thermodynamics of mechanosensitivity. Curr. Tops. Membr. 58, 87–119. Maroto, R., Raso, A., Wood, T. G., Kurosky, A., Martinac, B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nat. Cell. Biol. 7, 1443–1446. Maroto, R., Kurosky, A., and Hamill, O. P. (2007). The role of MscCa in prostate tumor cell migration. (submitted for publication). Martinac, B. (2007). 3.5 Billion years of mechanosensory transduction: Structure and function of mechanosensitive channels in prokaryotes. Curr. Tops. Membr. 58, 25–57. Martinac, B., and Hamill, O. P. (2002). Gramicidin A channels switch between stretch‐ activation and stretch‐inactivation depending upon bilayer thickness. Proc. Natl. Acad. Sci. USA 99, 4308–4312. Martinac, B., Adler, J., and Kung, C. (1990). Mechanosensitive channels of E. coli activated by amphipaths. Nature 348, 261–263. Matthews, B. D., Thodeki, C. K., and Ingber, D. E. (2007). Activation of mechanosensitive ion channels by forces transmitted through integrins and the cytoskeleton. Curr. Tops. Membr. 58, 59–85. Matsumoto, H., Baron, C. B., and Coburn, R. F. (1995). Smooth muscle stretch‐activated phospholipase C activity. Am. J. Physiol. 268, C458–C465. McBride, D. W., Jr., and Hamill, O. P. (1992). Pressure‐clamp: A method for rapid step perturbation of mechanosensitive channels. Pflu¨gers Arch. 421, 606–612. McBride, D. W., Jr., and Hammill, O. P. (1993). Pressure‐clamp techniques for measurement of the relaxation kinetics of mechanosensitive channels. Trends Neurosci. 16, 341–345. McBride, D. W., Jr., and Hamill, O. P. (1995). A fast pressure clamp technique for studying mechano‐gated channels. In ‘‘Single Channel Recording’’ (B. Sakmann and E. Neher, eds.), (2nd edn.), pp. 329–340. Plenum Press, New York.
9. MS TRPCs
225
McBride, D. W., Jr., and Hamill, O. P. (1999). A simplified fast pressure‐clamp technique for studying mechanically‐gated channels. Methods Enzymol. 294, 482–489. McKay, R. R., Szymeczek‐Seay, C. L., Lievremont, J. P., Bird, G. S., Zitt, C., Ju¨ngling, E., Lu¨ckhoV, A., and Putney, J. W., Jr. (2000). Cloning and expression of the human transient receptor potential 4 (TRP4) gene: Localization and functional expression of human TRP4 and TRP3. Biochem. J. 351, 735–746. Mery, L., Strauss, B., Dufour, J. F., Krause, K. H., and Hoth, M. (2002). The PDZ‐interacting domain of TRPC4 controls its localization and surface expression in HEK293 cells. J. Cell. Sci. 15, 3497–3508. Minke, B., and Cook, B. (2002). TRP channel proteins and signal transduction. Physiol. Rev. 82, 429–472. Minke, B., Wu, C., and Pak, W. L. (1975). Induction of photoreceptor voltage noise in the dark in Drosophila mutant. Nature 258, 84–87. Mitchell, C. H., Zhang, J. J., Wang, L., and Jacob, T. J. C. (1997). Volume‐sensitive chloride current in pigmented ciliary epithelial cells: Role of phospholipases. Am. J. Physiol. 272, C212–C222. Montell, C. (2005). The TRP superfamily of cation channels. Sci. STKE re3, 1–24. Montell, C., and Rubin, G. M. (1989). Molecular characterization of the Drosphila trp locus: A putative integral membrane protein required for phototransduction. Neuron 2, 1313–1323. Moore, A. L., Roe, M. W., Melnick, R. F., and Lidofsky, S. D. (2002). Calcium mobilization evoked by hepatocellular swelling is linked to activation of phospholase Cg. J. Biol. Chem. 277, 34030–34035. Moore, T. M., Brough, G. H., Babal, P., Kelly, J. J., Li, M., and Stevens, T. (1998). Store‐ operated calcium entry promotes shape changes in pulmonary endothelial cells expressing TRP1. Am. J. Physiol. Lung Cell Mol. Physiol. 275, L574–L582. Mori, Y., Wakamori, M., Miyakawa, T., Hermosura, M., Hara, Y., Nishida, M., Hirose, K., Mizushima, A., Kurosaki, M., Mori, E., Gotoh, K., Okada, T., et al. (2002). Transient receptor potential 1 regulates capacitative Ca2þ entry and Ca2þ release from endoplasmic reticulum in B lymphocytes. J. Exp. Med. 195, 673–681. Morris, C. E. (1990). Mechanosensitive ion channels. J. Membr. Biol. 113, 93–107. Nauli, S. M., and Zhou, J. (2004). Polycystins and mechanosensation in renal and nodal cilia. Bioessays 26, 844–856. Nauli, S. M., Alenghat, F. J., Luo, Y., Williams, E., Vassilev, P., Elia, A., Lu, W., Brown, E. M., Quinn, S. J., Ingber, D. E., and Zhou, J. (2003). Polycystins 1 and 2 mediate mechanosensation in primary cilium of kidney cells. Nat. Genet. 33, 129–137. Nilius, B., and Voets, T. (2005). TRP channels: A TR(I)P through a world of multifunctional cation channels. Pflu¨gers Arch. 451, 1–10. O’Neil, R. G., and Heller, S. (2005). The mechanosensitive nature of TRPV channels. Pflu¨gers Arch. 451, 193–203. Ohata, H., Tanaka, K., Maeyama, N., Ikeuchi, T., Kamada, A., Yamamoto, M., and Momose, K. (2001). Physiological and pharmacological role of lysophosphatidic acid as modulator in mechanotransduction. Jpn. J. Physiol. 87, 171–176. Opsahl, L. R., and Webb, W. W. (1994). Transduction of membrane tension by the ion channel alamethicin. Biophys. J. 66, 71–74. Owsianik, G., D’Hoedt, D., Voets, T., and Nilius, B. (2006). Structure‐function relationship of the TRP channel superfamily. Rev. Physiol. Biochem. Pharmacol. 156, 61–90. Palmer, C. P., Zhou, X. L., Lin, J., Loukin, S. H., Kung, C., and Saimi, Y. (2001). A TRP homolog in Saccharomyces cerevisiae forms an intracellular Ca permeable channel in the yeast vaculola membrane. Proc. Natl. Acad. Sci. USA 98, 7801–7805.
226
Hamill and Maroto
Paoletti, P., and Ascher, P. (1994). Mechanosensitivity of NMDA receptors in cultured mouse central neurons. Neuron 13, 645–655. Paolini, C., Fessenden, J. D., Pessah, I. N., and Franzini‐Armstrong, C. (2004). Evidence for conformational coupling between two calcium channels. Proc. Natl. Acad. Sci. USA 101, 12748–12752. Parekh, A. B., and Putney, J. W., Jr. (2005). Store‐operated calcium channels. Physiol. Rev. 85, 757–810. Paria, P. C., Malik, A. B., Kwiatek, A. M., Rahman, A., May, M. J., Ghosh, S., and Tiruppathi, C. (2003). Tumor necrosis factor‐a induces nuclear factor‐kB‐dependent TRPC1 expression in endothelial cells. J. Biol. Chem. 278, 37195–37203. Park, K. S., Kim, Y., Lee, Y. H., Earm, Y. E., and Ho, W. K. (2003). Mechanosensitive cation channels in arterial smooth muscle cells are activated by diacylglycerol and inhibited by phospholipase C inhibitor. Circ. Res. 93, 557–564. Patel, A. J., and Honore, E. (2001). Properties and modulation of mammalian 2P domain Kþ channels. Trends Neurosci. 24, 339–346. Patel, A. J., Lazdunski, M., and Honore, E. (2001). Lipid and mechano‐gated 2P domain Kþ channels. Curr. Opin. Cell Biol. 13, 422–428. Peinelt, C., Vig, M., Koomoa, D. L., Beck, A., Nadler, M. J. S., Koblan‐Huberson, M., Lis, A., Fleig, A., Penner, R., and Kinet, J. P. (2006). Amplification of CRA current by STIM1 and CRACM1 (Orai1). Nat. Cell Biol. 8, 771–773. Perozo, E., Kloda, A., Cortes, D. M., and Martinac, B. (2002). Physical principles underlying the transduction of bilayer deformation forces during mechanosensitive channel gating. Nat. Struct. Biol. 9, 696–703. Perraud, A. L., Fleig, A., Dunn, C. A., Bagley, L. A., Launay, P., Schmitz, C., Stokes, A. J., Zhu, Q., Bessman, M. J., Penner, R., Kinet, J. P., and Scharenberg, A. W. (2001). ADP‐ ribose gating of the calcium‐pereable LTRPC2 channel revealed by Nudix motif homology. Nature 411, 595–599. Philipp, S., Cavalie´, A., Freichel, M., Wissenbach, U., Zimmer, S., Torst, C., Marquart, A., Murakami, M., and Flockerzi, V. (1996). A mammalian capacitative calcium entry channel homologous to Drosophila TRP and TRPL. EMBO J. 15, 6166–6171. Plant, T. D., and Schaefer, M. (2005). Receptor‐operated cation channels formed by TRPC4 and TRPC5. Naunyn Schmiedebergs Arch. Pharmacol. 371, 266–276. Powl, A. M., and Lee, A. G. (2007). Lipid effects on mechanosensitive channels. Curr. Tops. Membr. 58, 151–178. Praetorius, H. A., and Spring, K. R. (2005). A physiological view of the primary cilium. Annu. Rev. Physiol. 67, 515–529. Protasi, F. (2002). Structural interactions between RYRs and DHPs in calcium release units of cardiac and skeletal muscle cells. Front. Biosci. 7, 650–658. Putney, J. W., Jr., Trebak, M., Vazquez, G., Wedel, B., and Bird, G. S. (2004). Signaling mechanisms for TRPC3 channels. Novartis Found. Symp. 258, 123–139. Ramsey, I. S., Delling, M., and Clapham, D. E. (2006). An introduction to TRP channels. Annu. Rev. Physiol. 68, 619–647. Rao, J. N., Platoshyn, O., Golovina, V. A., Liu, L., Zou, T., Marasa, B. S., Turner, D. J., Yuan, J. X. J., and Wang, J. Y. (2006). TRPC1 functions as a store‐operated Ca2þ channel in intestinal epithelial cells and regulates mucosal restitution after wounding. Am. J. Physiol. 290, G782–G792. Reading, S. A., Earley, S., Waldron, B. J., Welsh, D. J., and Brayden, J. E. (2005). TRPC3 mediates pyridine receptor‐induced depolarization of cerebral arteries. Am. J. Physiol. 288, H2055–H2061.
9. MS TRPCs
227
Reifarth, F. W., Clauss, W., and Weber, W. M. (1999). Stretch‐independent activation of the mechanosensitive cation channel in oocytes of Xenopus laevis. Biochim. Biophys. Acta 1417, 63–76. Reiser, J., Polu, K. R., Mo¨ller, C. C., Kemlan, P., Altinas, M. M., Wei, C., Faul, C., Herbert, S., Villegas, I., Avila‐Casado, C., McGee, M., Sugmoto, H., et al. (2005). TRPC6 is a glomerular slit diaphragm‐associated channel required for normal renal function. Nat. Genet. 37, 739–744. Riccio, A., Medhurst, A. D., Mattei, C., Kelsell, R. E., Calver, A. R., Randall, A. D., Benham, C. D., and Pangalos, M. N. (2002). mRNA distribution analysis of human TRPC family in CNS and peripheral tissues. Brain Res. Mol. Brain Res. 109, 95–104. Rosales, O. R., Isales, C. M., Barrett, P. Q., Brophy, C., and Sumpio, B. E. (1997). Exposures of endothelial cells to cyclic strain induces elevations of cytosolic Ca2þ concentration through mobilization of intracellular and extracellular pools. Biochem. J. 326, 385–392. Ru¨sch, A, Kros, C. J., and Richardson, G. P. (1994). Block by amiloride and its derivatives of mechano‐electrical transduction in outer hair cells of mouse cochlear cultures. J. Physiol. 474, 75–86. Ruwhof, C., Van Wamel, J. E. T., Noordzij, L. A. W., Aydin, S., harper, J. C. R., and Van Der Laarse, A. (2001). Mechanical stress stimulates phospholipase C activity and intracellular calcium ion levels in neonatal rat cardiomyocytes. Cell Calcium 29, 73–83. Ryan, M. J., Gross, K. W., and Hajduczok, G. (2000). Calcium‐dependent activation of phospholipase C by mechanical distension in renin‐expressing As4. 1 cells. Am. J. Physiol. 279, E823–E829. Sachs, F. (1988). Mechanical transduction in biological systems. CRC Crit. Rev. Biomed. Eng. 16, 141–169. Sachs, F., and Morris, C. E. (1998). Mechanosensitive ion channels in nonspecialized cells. Rev. Physiol. Biochem. Pharmacol. 132, 1–77. Sackin, H. (1989). A stretch‐activated Kþ channel sensitive to cell volume. Proc. Natl. Acad. Sci. USA 86, 1731–1735. Sackin, H. (1995). Mechanosensitive channels. Annu. Rev. Physiol. 57, 333–353. Saimi, Y., Zhou, X., Loukin, S. H., Haynes, W. J., and Kung, C. (2007). Microbial TRP channels and their mechanosensitivity. Curr. Tops. Membr. 58, 311–327. Sampieri, A., Diaz‐Munoz, M., Antaramian, A., and Vaca, L. (2005). The foot structure form the type 1 ryanodine receptor is required for functional coupling to store‐operated channels. J. Biol. Chem. 280, 24804–24815. Schaefer, M., Plant, T. D., Obukhov, A. G., Hofmann, T., Gudermann, T., and Shultz, G. (2000). Receptor‐mediated regulation of the nonselctive cation channels TRPC4 and TRPC5. J. Biol. Chem. 275, 17517–17526. Sinkins, W. G., Estacion, M., and Schilling, W. P. (1998). Functional expression of TRPC1: A human homologe of the Drosophila TRP channel. Biochem. J. 331, 331–339. Sinkins, W. G., Goel, M., Estacion, M., and Schilling, W. P. (2004). Association of immunophilins with mammalian TRPC channels. J. Biol. Chem. 279, 34521–34529. Slish, D. F., Welsh, D. G., and Brayden, J. E. (2002). Diacylglycerol and protein kinase C activate cation channels in myogenic tone. Am. J. Physiol. 283, H2196–H2201. Small, D. L., and Morris, C. E. (1994). Delayed activation of single mechanosensitive channels in Lymnaea neurons. Am. J. Physiol. 267, C598–C606. Smani, T., Zakharov, S. I., Leno, E., Csutoras, P., Trepakova, E. S., and Bolotina, V. M. (2003). Ca2þ‐independent phospholpase A2 is a novel determinant of store‐operated Ca2þ entry. J. Biol. Chem. 278, 11909–11915.
228
Hamill and Maroto
SoboloV, J., Spassova, M., Xu, W., He, L. P., Cuesta, N., and Gill, D. L. (2005). Role of endogenous TRPC6 channels in Ca2þ signal generation in A7r5 Smooth muscle cells. J. Biol. Chem. 280, 39786–39794. Spehr, M., Hatt, H., and Wetzel, C. H. (2002). Arachidonic acid plays a role in rat vomeronasal signal transduction. J. Neurosci. 22, 8429–8437. Stamboulian, S., Moutin, M. J., Treves, S., Pochon, N., Grunwald, D., Zorzato, F., Waard, M. D., Ronjat, M., and Arnoult, C. (2005). Junctate, an inositol 1,4,5‐triphosphate receptor associated protein is present in sperm and binds TRPC2 and TRPC5 but not TRPC1 channels. Dev. Biol. 286, 326–337. Strotmann, R., Harteneck, C., Nunnemacher, K., Schultz, G., and Plant, T. D. (2000). OTRPC4, a nonselective cation channe that confers sensitivity to extracellular osmolarity. Nat. Cell Biol. 2, 695–702. Stru¨bing, C., Krapivinsky, G., Krapivinsky, L., and Clapham, D. E. (2001). TRPC1 and TRPC5 from a novel cation channel in mammalian brain. Neuron 29, 645–655. Stru¨bing, C., Krapivinsky, G., Krapivinsky, L., and Clapham, D. E. (2003). Formation of novel TRPC channels by complex subunit interactions in embryonic brain. J. Biol. Chem. 278, 39014–39019. Suchyna, T. M., Johnson, J. H., Hamer, K., Leykam, J. F., Hage, D. A., Clemo, H. F., Baumgarten, C. M., and Sachs, F. (1998). Identification of a peptide toxin from Grammostola spatula spider venom that blocks cation selective stretch‐activated channels. J. Gen. Physiol. 115, 583–598. Suchyna, T. M., Tape, S. E., Koeppe, R. E., III, Anderson, O. S., Sachs, F., and Gottlieb, P. A. (2004). Bilayer‐dependent inhibition of mechanosensitive channels by neuroactive peptide enatiomers. Nature 430, 235–240. Sukharev, S. (2002). Purification of the small mechanosensitive channel in Escherichia coli (MScS): The subunit structure, conduction and gating characteristics. Biophys. J. 83, 290–298. Sukharev, S. I., Martinac, B., Arshavsky, V. Y., and Kung, C. (1993). Two types of mechanosensitive channels in the E. coli cell envelope: Solubilization and functional reconstitution. Biophys. J. 65, 177–183. Sukharev, S. I., Blount, P., Martinac, B., Blattner, F. R., and Kung, C. (1994). A large‐ conductance mechanosensitive channel in E. coli encoded by MscL alone. Nature 368, 265–268. Tang, Y., Tang, J., Chen, Z., Torst, C., Flockerzi, V., Li, M., Ramesh, V., and Zhu, M. X. (2000). Association of mammalian trp4 and phospholipase C isozymes with a PDZ domain‐containing protein, NHERF. J. Biol. Chem. 275, 37559–37564. Tomita, Y., Kaneko, S., Funayama, M., Kondo, H., Satoh, M., and Akaike, A. (1998). Intracellular Ca2þ store operated influx of Ca2þ through TRP‐R a rat homolog of TRP, expressed in Xenopus oocyte. Neurosci. Lett. 248, 195–198. Torihashi, S., Fujimoto, T., Trost, C., and Nakayama, S. (2002). Calcium oscillation linked to pacemaking of intestinal cells of Cajal: Requirement of calcium influx and localization of TRPC4 in caveolae. J. Biol. Chem. 277, 19191–19197. Trebak, M., Bird, G. S., Mckay, R. R., and Putney, J. W. (2002). Comparison of human TRPC3 channels in receptor‐activated and store‐operated modes. DiVerential sensitivity to channel blockers suggests fundamental diVerences in channel composition. J. Biol. Chem. 277, 21617–21623. Treves, S., Franzini‐Armstrong, C., Moccagatta, L., Arnoult, C., Grasso, C., Schrum, A., Ronjat, M., and Zorzato, F. (2004). Junctate is a key element in calcium entry induced by activation of InsP3 receptors and/or calcium store depletion. J. Cell Biol. 166, 537–548.
9. MS TRPCs
229
Tryggvason, K., and Wartiovaara, J. (2005). How does the kidney filter plasma? Physiology (Bethesda) 20, 96–101. Tsiokas, L., Arnould, T., Zhu, C., Kim, E., Walz, G., and Sukhatme, V. P. (1999). Specific association of the gene product of pkD2 with the TRPC1 channel. Proc. Natl. Acad. Sci. USA 96, 3934–3939. Vaca, L., and Sampieri, A. (2002). Calmodulin modulates the delay period between the release of calcium from internal stores and activation of calcium influx via endogenous TRP1 channels. J. Biol. Chem. 277, 42178–42187. Vandebrouck, C., Martin, D., Colson‐Van Schoor, M, Debaix, H., and Gailly, P. (2002). Involvement of TRPC in the abnormal calcium influx observed in dystrophic (mdx) mouse skeletal muscle fibers. J. Cell Biol. 158, 1089–1096. Vandorpe, D. H., and Morris, C. E. (1992). Stretch activation of the Aplysia S‐channel. J. Membr. Biol. 127, 205–214. Vannier, B., Peyton, M., Boulay, G., Brown, D., Qin, N., Jiang, M., Zhu, X., and Birnbaumer, L. (1999). Mouse trp2, the homologue of the human trpc2 pseudogene encodes mTrp2, a store depletion‐activated capacitive Ca2þ entry channel. Proc. Natl. Acad. Sci. USA 96, 2060–2064. Vanoye, C. G., and Reuss, L. (1999). Stretch‐activated single Kþ channels account for whole‐ cell currents elicited by swelling. Proc. Natl. Acad. Sci. USA 96, 6511–6516. Vazquez, G., Lievermont, P. P., Bird, G. S., and Putney, J. W., Jr. (2001). Human Trp3 forms both inositol trisphosphate receptor‐dependent and receptor‐independent store‐operated cation channels in DT40 avian B lymphocytes. Proc. Natl. Acad. Sci. USA 98, 11777–11782. Vazquez, G., Wedel, B. J., Trebak, M., Bird, G. S., and Putney, J. W., Jr. (2003). Expression levels of the canonical transient receptor potential 3 (TRPC3) channels determine its mechanism of activation. J. Biol. Chem. 278, 21649–21654. Vazquez, G., Wedel, B. J., Aziz, O., Trebak, M., and Putney, J. W., Jr. (2004a). The mammalian TRPC cation channels. Biochim. Biophys. Acta 1742, 21–36. Vazquez, G., Wedel, B. J., Kawasaki, B. T., Bird, G. S., and Putney, J. W. (2004b). Obligatory role of src kinase in the signaling mechanism for TRPC3 cation channels. J. Biol. Chem. 279, 40521–40528. Voets, T., Talavera, K., Owsiannik, G., and Nilius, B. (2005). Sensing with TRP channels. Nat. Chem. Biol. 1, 85–92. Vriens, J., Watanabe, H., Janssens, A., Droogmans, G., Voets, T., and Nilius, B. (2004). Cell swelling, heat, and chemical agonists use distinct pathways for the activation of the cation channel TRPV4. Proc. Natl. Acad. Sci. USA 101, 396–401. Walker, R. G., Willingham, A. T., and Zucker, C. S. (2000). A Drosophilia mechanosensory transduction channel. Science 287, 2229–2234. Wang, S. Q., Song, L. S., Lakatta, E. G., and Cheng, H. (2001). Ca2þ signaling between single L‐type Ca2þ channels and ryanodine receptors in heart cells. Nature 410, 592–596. Watanabe, H., Vriens, J., Prenen, J., Droogmans, G., Voets, T., and Nilius, B. (2003). Anandamide and arachidonic acid use epoxyeicosatrienoic acids to activate TRPV4 channels. Nature 424, 434–438. Weber, W. M., Popp, C., Clauss, W., and van Driessche, W. (2000). Maitotoxin induces insertion of diVerent ion channels into the Xenopus oocyte plasma membrane via Ca2þ‐ stimulated exocytosis. Pflu¨gers Arch. 439, 363–369. Welsh, D. G., Nelson, M. T., Eckman, D. M., and Brayden, J. E. (2000). Swelling activated cation channels mediate depolarization of rat cerebrovascular smooth muscle by hypotonicity and intravascular pressure. J. Physiol. 527(Pt. 1), 139–148.
230
Hamill and Maroto
Welsh, D. G., Morielli, A. D., Nelson, M. T., and Brayden, J. E. (2002). Transient receptor potential channels regulate myogenic tone of resistance arteries. Circ. Res. 90, 248–250. Wes, P. D., Chevesich, J., Jeromin, A., Rosenberg, C., Stetten, G., and Montell, C. (1995). TRPC1, a human homolog of a Drosophila store operated channel. Proc. Natl. Acad. Sci. USA 92, 9652–9656. Winn, M. P., Conlon, P. J., Lynn, K. L., Farrington, M. K., Creazzo, T., Hawkins, A. F., Daskalakis, N., Kwan, S. Y., Ebersviller, S., Burchette, J. L., Pericak‐Vance, M. A., Howell, D. N., et al. (2005). A mutation in the TRPC6 cation channel causes familial focal segmental Glomerulosclerosis. Science 308, 1801–1804. Wu, X., Babnigg, G., Zagranichnaya, T., and Villereal, M. L. (2002). The role of endogenous human TRP4 in regulating carbachol‐induced Calcium oscillations in HEK‐293 cells. J. Biol. Chem. 277, 13597–13608. Xu, H., Zhao, H., Tian, W., Yoshida, K., Roullet, J. P., and Cohen, D. M. (2003). Regulation of a transient receptor potential (TRP) channel by tyrosine phosphorylation. J. Biol. Chem. 278, 11520–11527. Xu, S. Z., and Beech, D. J. (2001). TRPC1 is a membrane‐spanning subunit of store‐operated Ca2þ channels in native vascular smooth muscle cells. Circ. Res. 88, 84–87. Yamada, H., Wakamori, M., Hara, Y., Takahashi, Y., Konishi, K., Imoto, K., and Mori, Y. (2000). Spontaneous single channel activity of neuronal TRP5 channel recombinantly expressed in HEK293 cells. Neurosci. Lett. 285, 111–114. Yao, Y., Ferrer‐Montiel, A. V., Montal, M., and Tsien, R. Y. (1999). Activation of store‐ operated Ca2þ current in Xenopus oocytes requires SNAP‐25 but not a diVusible messenger. Cell 98, 475–485. Yeung, E. W., and Allen, D. G. (2004). Stretch‐activated channels in stretch‐induced muscle damage: Role in muscular dystrophy. Clin. Exp. Pharmacol. Physiol. 31, 551–556. Yildrin, E., Dietrich, A., and Birnbaumer, L. (2003). The mouse C‐type transient receptor potential 2 (TRPC2) channel: Alternative splicing and calmodulin binding to its N terminus. Proc. Natl. Acad. Sci. USA 100, 2220–2225. Yin, C. C., Blayney, L. M., and Lai, F. A. (2005). Physical coupling between ryanodine receptor‐calcium release channels. J. Mol. Biol. 349, 538–546. Yu, Y., Fantozzi, I., Remillard, C. V., Landsberg, J. W., Kunichika, N., Platoshyn, O., Tigno, D. D., Thistlethwaite, P. A., Rubin, L. J., and Yuan, J. X‐L. (2004). Enhanced expression of transient receptor potential channels in idiopathic pulmonary arterial hypertension. Proc. Natl. Acad. Sci. USA 101, 13861–13866. Yu, Y., Sweeney, M., Zhang, S., Platoshyn, O., Landsberg, J., Rothman, A., and Yuan, J. X. J. (2003). PDGF stimulates pulmonary vascular smooth muscle cells proliferation by upregulating TRPC6 expression. Am. J. Physiol. 284, C316–C330. Yuan, J. P., Kislyoy, K., Shin, D. M., Chen, J., Shcheynikov, N., Kang, S. H., DehoV, M. H., Schwarz, M. K., Seeberg, P. H., Muallem, S., and Worley, P. F. (2003). Homer binds TRPC family channels and is required for gating of TRPC1 by IP3 receptors. Cell 114, 777–789. Zagranichnaya, T. K., Wu, X., and Villereal, M. L. (2005). Endogenous TRPC1, TRPC3 and TRPC7 proteins combined to form native store‐operated channels in HEK‐293 cells. J. Biol. Chem. 280, 29559–29569. Zhang, Y., and Hamill, O. P. (2000a). Calcium‐, voltage‐ and osmotic stress‐sensitive currents in Xenopus oocytes and their relationship to single mechanically gated channels. J. Physiol. 523(Pt. 1), 83–99. Zhang, Y., and Hamill, O. P. (2000b). On the discrepancy between membrane patch and whole cell mechanosensitivity in Xenopus oocytes. J. Physiol. 523(Pt. 1), 101–115.
9. MS TRPCs
231
Zhang, Y., Gao, F., Popov, V., Wan, J., and Hamill, O. P. (2000). Mechanically‐gated channel activity in cytoskeleton deficient blebs and vesicles from Xenopus oocytes. J. Physiol. 523 (Pt. 1), 117–129. Zhang, Y., Guo, F., Kim, J. Y., and SaVen, D. (2006). Muscarinic acetylcholine receptors activated TRPC6 channels in PC12D cells via Ca2þ store‐independent mechanisms. J. Biochem. 139, 459–470. Zhou, X. L., Batiza, A. F., Loukin, S. H., Palmer, C. P., Kung, C., and Saimi, Y. (2003). The transient receptor potential channels on the yeast vacuole is mechanosensitive. Proc. Natl. Acad. Sci. USA 100, 7105–7110. Zhu, X., Chu, P. B., Peyton, M., and Birnbaumer., L. (1995). Molecular cloning of a widely expressed human homologue for the Drosophilia trp gene. FEBS Lett. 373, 193–198. Zitt, C., Zobei, A., Obukhov, A. G., Harteneck, C., Kalkbrenner, F., Lu¨ckhoV, A., and Schultz, G. (1996). Cloning and functional expression of a human Ca2þ‐permeable cation channel activated by calcium store depletion. Neuron 16, 1189–1196. Zitt, C., Obukhov, A. G., Stru¨bing, C., Zobel, A., Kalkbrenner, F., Lu¨ckhoV, A., and Schultz, G. (1997). Expression of TRPC3 in Chinese hamster ovary cells results in calcium‐activated cation currents not related to store depletion. J. Cell Biol. 138, 1333–1341. Zufall, F., Ukhanov, K., Lucas, P., and Leinders‐Zufall, T. (2005). Neurobiology of TRPC2: From gene to behavior. Pflu¨gers Arch. 451, 61–71.
This page intentionally left blank
CHAPTER 10 The Cytoskeletal Connection to Ion Channels as a Potential Mechanosensory Mechanism: Lessons from Polycystin‐2 (TRPP2) Horacio F. Cantiello,*,{,{ Nicola´s Montalbetti,{ Qiang Li,} and Xing‐Zhen Chen} *Renal Unit, Massachusetts General Hospital East, Charlestown, Massachusetts 02129 { Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115 { Laboratorio de Canales Io´nicos, Departamento de Fisicoquı´mica y Quı´mica Analı´tica, Facultad de Farmacia y Bioquı´mica, Buenos Aires 1113, Argentina } Department of Physiology, University of Alberta, Edmonton T6G 2H7, Canada
I. Overview II. Introduction A. The Channel–Cytoskeleton Connection B. Actin Filaments and Their Disruption: EVect of Cytochalasins C. The Superfamily of TRP Channels D. TRP Channels and Mechanosensation E. Cytoskeletal Connections in TRP Channels III. Role of Actin Cytoskeletal Dynamics in PC2‐Mediated Channel Function A. Role of PC2 in Health and Disease B. Presence of Actin and Associated Proteins and EVect of CD on Channel Activity in hST C. EVect of Gelsolin and Actin on PC2 Channel Activity in hST IV. Identification of Actin‐Binding Protein Interactions with Polycystin‐2 A. Interaction Between PC2 and a‐Actinin Revealed by Yeast Two‐Hybrid System B. In Vitro and In Vivo Binding of PC2 with a‐Actinins C. Functional Modulation of PC2 by a‐Actinin V. EVect of Hydroosmotic Pressure on PC2 Channel Function: Role of the Cytoskeleton in Osmosensory Function A. EVect of Hydrostatic and Osmotic Pressure on PC2 Channel Regulation
Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59010-6
234
Cantiello et al.
VI. The Channel–Cytoskeleton Interface: Structural–Functional Correlates A. Mechanosensitivity and the Lipid Bilayer B. Cytoskeletal Interactions with PC2 C. In Search of the Molecular Link D. Elastic Properties of Actin Networks E. Sensory Role of the Actin Cytoskeleton in PC2 Channel Function VII. Perspective and Future Directions References
I. OVERVIEW Mechanosensitivity of ion channels, or the ability to transfer mechanical forces into a gating mechanism of channel regulation, has been split into two main working (not mutually exclusive) hypotheses. One is that elastic and/or structural changes in membrane properties act as a transducing mechanism of channel regulation. The other hypothesis involves tertiary elements, such as the cytoskeleton which, itself by dynamic interaction(s) with the ion channel, may convey conformational changes including those ascribed to mechanical forces. This hypothesis is supported by numerous instances of regulatory changes in channel behavior by alterations in cytoskeletal structures/interactions. However, only recently, the molecular nature of these interactions has slowly emerged. Recently, a surge of evidence has emerged to indicate that transient receptor potential (TRP) channels are key elements in the transduction of a variety of environmental signals. Herein, we summarize recent work, which in brief, define the molecular linkage and regulatory elements of polycystin‐2 (PC2), a TRP‐type (TRPP2) nonselective cation channel whose mutations cause autosomal dominant polycystic kidney disease (ADPKD). We provide evidence for the involvement of cytoskeletal structures in the regulation of PC2 and assess how these connections are the transducing mechanism of environmental signals to its channel function. We conclude and propose that the actin network, which attaches to the PC2 channel, is a novel osmosensitive device, where the three‐dimensional structure of the actin gel apposed to the channel, elicits its regulation. Thus, ‘‘environmental forces’’ such as hydroosmotic pressure control PC2 channel activity, by conveying a sensory mechanism to the channel, and through the integrity of the cytoskeleton. Taken together, our findings strongly support the hypothesis that the channel‐cytoskeletal interface is a functional unit, with general and important implications in mechanosensitivity.
10. The Cytoskeletal Connection to Ion Channels
235
II. INTRODUCTION All cells are continually exposed to external physical forces to which they adapt. Therefore, a large part of the cell’s physiology is placed in detecting and responding to environmental stresses, including (osmotic, hydrostatic) pressure, shear, flow, stretch, and compression. Conversion of mechanical sensation into intracellular signals that modifies the cellular response is called ‘‘mechanotransduction.’’ Long before, molecular physiological techniques identified mechanosensitive molecular devices in most cells, it was accepted that mechanosensitivity was only observed in specialized sensory organs. Thus, mechanosensitivity had only been identified in crayfish stretch receptors (Brown et al., 1978), Pacinian frog muscle spindle (Katz, 1950) and specialized skin cells as the Paciniam corpuscle (Mendelson and Loewenstein, 1964), and the stretch receptor in crickets (Coillot and Boistel, 1969). Specialized organs responsible for hearing and touch are also expected to mediate their function through sensory receptors. However, many other organs such as the brain, spinal cord, bladder, and joints are subjected to mechanical forces, which require mechanotransduction. Vascular and alveolar distension, and the mechanical response of several organs to pressure, such as the urinary bladder, the intestines, the placenta, and the kidney also require mechanosensitive responses. Changes in cerebrospinal fluid production in the brain, for example, may raise intracranial pressure, causing anatomical deformities as hydrocephalus. In the eye, all tissues are continuously subjected to variations in intraocular pressure whose dysfunction may cause glaucoma (Vittitow and Borras, 2002; Kalapesi et al., 2005), corneal edema, iris ischemia, and changes in the trabecular meshwork lens opacity (Johnstone and Grant, 1973; Borras, 2003; Kalapesi et al., 2005). Other examples of exposure to excessive mechanical force such as pressure or compression include peripheral nerve entrapments, obstructive nephropathy, hypertensive hypertrophic cardiomyopathy, hypertensive glomerulosclerosis, and compression of vertebral fractures and disc herniation. The blood vessel endothelium and tubular epithelia of the mammalian nephron are also subjected to shear, stretch, and tension, and skin and mucosal epithelia are subjected to continuous stretch. Striated skeletal muscle is also under tension, whose compression aVects sarcomere contraction. Therefore, mechanically active environments impose various forms of pressure from both inside their plasma membrane (i.e., osmotic, cytoskeletal) and externally, including forces such as stretch, and when intracellular pressure may rise, for instance in hypoosmotic shock or due to elevated extracellular pressure such as in increased flow and shear.
236
Cantiello et al.
Cells respond to environmental stress by changing their morphologic appearance (Ives et al., 1986; Dartsch and Hammerle, 1986; Dahlin et al., 1987) activating signaling pathways and the production of a number of second messengers (Skinner et al., 1992; Matsuo and Matsuo, 1996; Matsuo et al., 1996; Okada et al., 1998; Yokoyama et al., 1999; Mikuni‐Takagaki, 1999; Mallouk and Allard, 2000; Endlich et al., 2002; Ji et al., 2002) and/or by aVecting gene and protein expression (Terakawa et al., 2002; Voisin and Bourque, 2002; Borras, 2003; Kelly et al., 2006; Zhou et al., 2006). It should be expected, therefore, that several not mutually exclusive adaptive mechanisms are present in most cells to help cope with environmental stress. However, such mechanisms at the cellular level have just began to be defined. Mechanosensing molecular devices in cells described to date include mechanically gated ion channels, plasma membrane‐bound enzymes such as phospholipases A2 and C, cytoskeletal structures, and receptor complexes with cell–cell and cell– matrix adhesion properties, such as cadherins, selectins, and integrins (Ingber, 1997; Ko and McCulloch, 2000; Stamenovic and Ingber, 2002). Among these, the cytoskeleton provides underlying support to the plasma membrane and forms part of the linkage to the extracellular matrix (Ingber, 1997). Integrins are transmembrane receptors, which link to the extracellular matrix components of the cytoskeleton such as actin and intermediate filaments. Integrin receptors are accepted as sensory transducers of mechanical stress via the extracellular matrix, through the plasma membrane, into the cytoskeleton to elicit the activation of intracellular signaling pathways (Ingber, 1997). Inhibition of integrins aVect mechanotransduction (Yoshida et al., 1996; Mobasheri et al., 2002). The interdependence of the plasma membrane and cytoskeleton in adaptation to applied forces has also been recognized. Membrane‐bound enzymes and proteins such as phospholipase A2, phospholipase C, and tyrosine kinases have been implicated in mechanosensory function. Thus, membrane stretch, including by osmotic swelling, induces release of prostaglandins and cAMP, aVecting the hydrolytic production of phospholipids (Kreisberg et al., 1982; Skinner et al., 1992; Yokoyama et al., 1999; Ko and McCulloch, 2000). In vitro studies on stretched mesangial cells, cardiac myocytes, and fetal lungs, as well as flow and shear on human umbilical vein endothelial cells show activation of phospholipase C, generating diacylglycerol (DAG) and the ensuing molecules in the phosphatidylinositol pathway. In addition to this, the cytoskeleton itself is capable of modifying the cellular environment. Physical stress has been shown to induce changes in actin polymerization and thus mechanotransduction by providing additional sites for actin–myosin interaction, thereby enhancing force generation in response to increased intravascular pressure (Zhelev and Hochmuth, 1995; Cipolla et al., 2002). Another class of molecular mechanosensory devices in cells implicates mechanosensitive channels, which are phenomenologically defined as channel
10. The Cytoskeletal Connection to Ion Channels
237
structures whose gating and activation is identified in response to forces acting at the plasma membrane. The widespread discovery of mechanosensitive channels arose with the advent of the patch clamping techniques, which allowed the identification of channel phenotypes in situ (Hamill et al., 1981). The technical problems came about with this milestone arose and its paradigmatic shift in membrane physiology. Reportedly, the first observation of mechanosensitive ion channel activity is attributed to Guharay and Sachs (1984), who noted that channel activity increased with suction while trying to form patch clamp seals in cultured chick skeletal muscle. The authors found that the channel’s open probability increased by the applied vacuum pressure (Guharay and Sachs, 1984). Their results provided the first direct evidence for cells to respond to applied pressure with changes in ionic conductance. The conclusion of this original report was that the cytoskeleton controls mechanically gated ion channels by tonically repressing their activity (Guharay and Sachs, 1984). Thus, patch clamping studies in particular those of cell‐attached and excised patches had an intrinsic caveat in that the very acquisition of the patch had associated with it a consequent rearrangement of intrinsic cytoskeletal structures. This problem, which may also plays a relevant role in the very definition of mechanosensitive channel adaptation was clearly manifested in phenomenological diVerences between the altered channel activity of cell‐attached patches, which appear quite diVerent from data obtained in whole‐cell studies (Zhang and Hamill, 2000). Thus, an intrinsic uncertainty lies in the fact that every patch has an unknown albeit relevant attached cytoskeleton. Serious attempts at avoiding this ‘‘contaminating’’ factor have provided strong evidence for mechanosensitive channels whose function is intrinsically associated with structural changes to the membrane itself (Hamill and Martinac, 2001, see below). To what extent channels sensitive to membrane stretch are intrinsically (i.e., under physiological conditions) regulated by cytoskeletal components instead is to date, a largely open question. Nonetheless, a number of mechanosensitive channels have been identified across various cell types ranging from prokaryotes, such as bacteria and archea, to eukaryotic cells in mammalian organs, including the central and peripheral nervous system, myocytes, blood vessel endothelium, renal epithelia, hair cells, and fibroblasts (reviewed in Hamill and Martinac, 2001). Stretch‐regulated channels have been particularly described in a number of excitable cells, including snail neurons, mammalian astrocytes, atrial myocytes, dystrophic muscle from mdx mice and toad gastric smooth muscle, and possibly many other cells (reviewed in Morris, 1990; Hamill and Martinac, 2001). To date, only a few channels have been molecularly identified on the basis of specific responses to stress forces (Sukharev and Corey, 2004). Mechanosensitive channels, whose gating has been associated to activation by changes in membrane
238
Cantiello et al.
tension, appear to share functional similarities (Hamill and Martinac, 2001), including weakly selective cation perm‐selectivity (Morris, 1990), and permeability to divalent cations, allowing significant Ca2þ influx during stretch (Sackin, 1994). In prokaryotes, a stretch‐activated mechanosensitive channel was first evidenced by experiments on Escherichia coli, where the MscL channel was identified, which open at pressures just under those which would disrupt membranes (Gu et al., 1998; Martinac, 2004). Imposed vacuum pressure is often one of the identifying forces in determining the presence of mechanosensitive channel activity. Nonetheless, other forces such as osmotic pressure are equally relevant in defining the activating mechanism of mechanosensitive channels. In cultured trabecular meshwork cells, for example, high conductance Ca2þ‐activated Kþ channels are activated in response to either membrane stretch or hypotonic shock (Gasull et al., 2003). Several instances of claimed mechanosensitivity have arisen from studies where ‘‘membrane stress’’ has been achieved by osmotic shock. Due to the universality of anisoosmotic cell responses, and in particular cell volume regulation, which implicates cytoskeletal structures, the true nature of mechanosensitivity requires further exploration. Interestingly, both mechanosensitive channel activation and inactivation have been reported (Morris and Sigurdson, 1989), which brings to the issue of the techniques with which mechanosensory channel function is described and further explored. To date, at least three families of channels have been identified as functionally linked to mechanosensory function, by this being understood, that the phenotypes underlying the channels later identified, were of a mechanical nature. The two‐pore domain potassium channels TREK and TRAAK are a group of four‐ transmembrane domain channels preferentially found in the CNS (Patel et al., 1998; Patel and Honore, 2001). These channels, whose intrinsic properties are regulated by stretch of the plasma membrane, are likely implicated in the response to various ‘‘environmental’’ forces including mechanical or osmotic stress, intracellular pH, or temperature. As expected from mechanosensitive channels, and indeed associated with its activation by membrane stretch, TREK‐1 is also modulated by osmotic cell swelling (Patel et al., 1998). Studies also showed that apart from stretch or membrane tension, TRAAK and TREKs appear to be activated by arachidonic acid metabolites and other ligands (Fink et al., 1998; Maingret et al., 2000), which interestingly enough is a common mechanism of activating TRP channels as well (Minke, 2001). The DEG/ENaC channel family has been implicated in mechanosensitive channel activity, despite the fact that to date, no clear evidence is available as to whether channels of this family indeed respond to stretch activation (Sukharev and Corey, 2004). The first member of this family was originally identified after the long search for the first member, a highly‐Naþ selective,
10. The Cytoskeletal Connection to Ion Channels
239
and amiloride‐sensitive cation channel of tubular epithelia (ENaC) of the distal nephron (Canessa et al., 1993). Interestingly, although not previously sought after as a mechanosensitive channel, ENaC was found to share homology with genes associated with neural degeneration, the nematode degenerins (DEGs; Canessa et al., 1993). Subsequently, ENaC activity has been found to be mechanosensitive, at least in response to osmotic changes, when heterologously expressed in Xenopus oocytes. Further confirmation of the intrinsic mechanosensitivity by ENaC has been postulated in reconstituted channels in a lipid bilayer (Awayda et al., 1995). The encompassed conclusions have not been without controversy (Rossier, 1998), and the true mechanosensitive nature of ENaC remains as open question. However, ENaC function has been previously observed to require cytoskeletal components, namely actin, and Benos’ group provided strong evidence for ENaC to bind actin directly (Mazzochi et al., 2006). This, in combination with the fact that the epithelial Naþ channel complex also contains at least another pore, Apx, and cytoskeletal proteins such as spectrin suggests that in vivo conditions may trigger activating responses, which are much more complicated than originally expected. Other members of the DEG/ENaC family, the DEGs are nematode gene products of several genes (Hamill and McBride, 1993; Goodman and Schwarz, 2003), which include those which encode cytoskeletal proteins (i.e., tubulins, MEC‐7, ‐12), structural channel proteins (MEC‐4, MEC‐10, DEG‐1, and UNC‐105) and matrix proteins (MEC‐1, MEC‐5, and MEC‐9). This superfamily also includes, the acid‐sensitive channels of vertebrate neurons (ASICs; also known as BNCs and BNaCs), and Drosophila PPKs (Sukharev and Corey, 2004). Channels of this superfamily are likely involved in touch and other mechanosensations (Welsh et al., 2002; Goodman and Schwarz, 2003), although direct proof is still lacking. By far the most appealing superfamily of sensory channels discovered in recent years is that of the TRP channels (Section II.C), with clear connections to a number of sensory responses, including mechanosensitivity.
A. The Channel–Cytoskeleton Connection Due to the techniques used to assess mechanosensitivity in ion channels, it is inherently clear that except for few exceptions (Perozo et al., 2002; Maroto et al., 2005) many mechanosensitive channels likely work in concert with cytoskeletal structures. Actin‐based networks are implicated in such diverse cellular functions as phagocytosis, regulation of cellular shape, locomotion, and hormone action (Painter and McIntosh, 1979; Stendahl et al., 1980; Hall, 1984; Stossel, 1984; Smith, 1988; Stossel, 1993). Thus, it is not surprising that a
240
Cantiello et al.
wide number of channel species have been linked to the cytoskeleton. Disregarding epiphenomenological aspects of cell conductance, such as traYcking and vesicle fusion of channel‐containing membranes, most channels, whose function has been looked in detail, associate directly and/or indirectly, to cytoskeletal proteins. Most studies have employed drugs that selectively stabilize or destabilize either actin filaments or microtubules, with resulting eVects on specific ion channel activity, either as changes in whole‐cell conductance or single‐channel activity by the patch clamping techniques. However, direct proof, namely the anchoring and binding of specific cytoskeletal proteins to individual channel proteins is only beginning to emerge. A body of earlier evidence demonstrated that various cytoskeletal components, including actin and actin‐associated proteins, anchor, colocalize, and regulate both the spatial stability as well as the function of ion transport proteins. Ankyrin and spectrin, colocalize with the band 3 anion exchanger (Drenckhahn et al., 1985), the a‐subunit of the Naþ, Kþ‐ATPase in epithelial cells (Morrow, 1989), rat brain voltage‐sensitive (Edelstein et al., 1988), and epithelial Naþ channels (Smith et al., 1991). Direct cytoskeletal connections have been found to a number of identified channel structures including ligand‐gated channels such as the NMDA (Lei et al., 2001; Yuen et al., 2005), AMPA (Kim and Lisman, 2001), and acetylcholine receptors (Bloch et al., 1997; Mitsui et al., 2000; Shoop et al., 2000), voltage‐gated Naþ (Srinivasan et al., 1988, 1992; Undrovinas et al., 1996), Kþ (Mazzanti et al., 1996; Jing et al., 1997; Nakahira et al., 1999) and Ca2þ (Johnson and Byerly, 1994; Lader et al., 1999; Johnson et al., 2005) channels, and Cl channels as the GABA(A) (Wang et al., 1999; Luccardini et al., 2004) and glycine (van Zundert et al., 2002, 2004) receptors. Channels in nonexcitable cells are also linked to the cytoskeleton, including amiloride‐sensitive cation channels such as ENaC (Berdiev et al., 1996; Ismailov et al., 1997; Mazzochi et al., 2006) and Apx (Prat et al., 1996; Zuckerman et al., 1999), and Cl channels such as CFTR (Prat et al., 1994, 1995; Ismailov et al., 1997), and CLC channels (Ahmed et al., 2000; Dhani et al., 2003). Direct interactions include the binding of key cytoskeletal components such as actin itself to channels like ENaC (Mazzochi et al., 2006) and CFTR (Chasan et al., 2002), or tubulin, as it has been observed for TRPV1 (Goswami et al., 2004) and GABA and glycine receptors (Kirsch et al., 1991; Coyle et al., 2002). Spectrin, for example, has been observed to bind to voltage‐gated Naþ channels in the brain (Srinivasan et al., 1988) and may be a structural component of the epithelial channel complex containing both ENaC and Apx in cells from the distal nephron (Zuckerman et al., 1999). Linker proteins as a‐actinin have been found to bind directly to the C‐terminal end of the glutamate receptor (Wyszynski et al., 1997) and to L‐type Ca2þ channels (Sadeghi et al., 2002). Similar interactions also occur between the a‐actinin and voltage‐gated Kþ channels (Maruoka et al., 2000), and TRPP2 (PC2; Li et al., 2005).
10. The Cytoskeletal Connection to Ion Channels
241
The actin cytoskeleton may modify mechanosensitive channels by attaching and modifying directly the plasma membrane, and/or structural dynamics of channel function, directly (reviewed in Morris, 1990; Hamill and McBride, 1993; Sackin, 1994). Actually, the original description of mechanosensitive channels strongly implicates the actin cytoskeleton as a modulator of mechanosensitivity (Guharay and Sachs, 1984). Following the application of pressure or suction to a membrane patch, there is a lag response noted in channel activation and deactivation, suggesting that an elastic component must exist to ‘‘transduce’’ membrane tension toward the channels. This lag response is likely due to the viscoelastic actin network relaxing with time, transferring the membrane tension to the mechanosensitive channels (Guharay and Sachs, 1984). Thus, cytoskeletal dynamics may be an important component of the delay of channel activation, likely acting as a stabilizing or restraining force on stretch‐regulated channel function (Guharay and Sachs, 1984; Small and Morris, 1994; Laitko et al., 2006).
B. Actin Filaments and Their Disruption: Effect of Cytochalasins It is important therefore to consider how the actin cytoskeleton is assembled, and how it can be disrupted. The formation of actin filaments (F‐actin) from actin monomers (G‐actin) is generally viewed as a condensation polymerization in which G‐actin monomers, condense, in a rate‐limiting step, to form nuclei which then rapidly elongate to form F‐actin in equilibrium with G‐actin at its critical monomer concentration (Bray, 1992). Normally, for every actin monomer added to the polymer, one G‐actin‐bound ATP is hydrolyzed to F‐actin‐bound ADP. This hydrolysis of ATP allows actin to polymerize in a ‘‘head‐to‐tail orientation,’’ where association and dissociation at either end occur at diVerent rate constants (Wegner, 1976; Cartier et al., 1984; Korn et al., 1987). The degree of polymerization is therefore a steady state condition, which requires the constant supply of actin monomers and nucleotides. Interestingly, ATP hydrolysis by F‐actin modifies elastic properties to the filament, which in turn, may render actin networks of distinct mechanical properties (Janmey et al., 1990b). Most of the work related to the actin cytoskeletal control of ion channels, whether mechanosensitive or not, has relied on natural toxins, which aVect actin polymerization. It is imperative; therefore, that a clearer understanding of their modes of action, and eVect(s) in cells are better understood. Cytochalasins are fungal metabolites, which were originally described by their inhibiting eVect on a wide variety of cellular movements (Bray, 1992). It is interesting that cytochalasin B (CB), one of the most popular cytochalasins in changing cell motility, was originally claimed to have no eVect on
242
Cantiello et al.
several cell functions (Wessells et al., 1971), including a lack of eVect on actin filament disruption (Wessells et al., 1971; Forer et al., 1972). Conversely, cytochalasins have been observed to cause eVects, which may not be directly associated with cytoskeletal derangement but are nonetheless important when claiming a change in a particular cellular function. CB has been found, for example, to bind and inhibit glucose transport (Fay et al., 1990), aVect ascorbic acid uptake in fibroblasts (Fay et al., 1990), inhibit protein synthesis by releasing mRNA (Ornelles et al., 1986), and induce mitochondrial contraction (Lin et al., 1973). Soon after the discovery of the cytochalasin‐ induced inhibition of cellular motility, it was observed that these compounds actually aVect the rate of actin polymerization (Brenner and Korn, 1979; MacLean‐Fletcher and Pollard, 1980). Early experiments demonstrated that high concentrations (100 mM) of CB lower by 30–50% the viscosity of actin filaments (Spudich and Lin, 1972). Hartwig and Stossel (1976) also showed that CB inhibits F‐actin gelation by a high molecular weight actin cross‐ linking protein, later to be identified as filamin A. Thus, CB inhibition of actin gelation has been used as a criterion to disrupt the cell’s cytoskeleton. Nevertheless, the final outcome of this interaction lies on the concentration and incubation times to which cells are exposed to cytochalasins. Substoichiometric concentrations of CB strongly inhibit network formation by actin filaments themselves (Brenner and Korn, 1979), and suggest that this is a direct eVect on actin, rather than an eVect on actin cross‐linking proteins (Brenner and Korn, 1979). Similar data were obtained independently using diVerent cross‐linking proteins (Hartwig and Stossel, 1979). To date it is accepted that cytochalasins have distinct and multiple eVects on actin polymerization (Bray, 1992). CB reduces the rate of actin polymerization, in a process, which involves inhibition of monomer addition to the barbed, or fast growing, end of actin filaments, which is favored for elongation (Brenner and Korn, 1979; Brown and Spudich, 1979; Lin and Lin, 1979; Lin et al., 1980). Similar eVects have been observed for cytochalasin D (CD; Brown and Spudich, 1979; Lin et al., 1980). Addition of polymers to the barbed end of F‐actin is fundamentally diVerent from the addition of G‐actin. This is based on the fact that this reaction involves the binding of one actin in the monomer conformation to two actin molecules in the polymer conformation. Thus, a key element in determining the eVect of a given cytochalasin is its putative interaction with G‐actin. While binding of CB to G‐actin has never been demonstrated, Goddette and Frieden, found that CD actually binds to monomeric actin with 1: l stoichiometry and a dissociation constant of 18 mM (Goddette and Frieden, 1986a,b). This stoichiometry and aYnity can be changed in the presence of low Mg2þ (Goddette and Frieden, 1986a,b). Thus, it has been concluded that CD induces dimer formation (Goddette and Frieden, 1986a). These studies are highly relevant, as the presence of CD
10. The Cytoskeletal Connection to Ion Channels
243
actually induces actin polymerization by an enhancement of the initial rate of nucleation. The large decrease in final extent of F‐actin is also attributed to the formation of such dimers (Goddette and Frieden, 1986b). Therefore, binding of CB is restricted to the polymeric form of actin, and restricted even further to sites exposed at the barbed end of the filaments. In contrast, CD treatment may actually induce a pool of short actin oligomers. The outcome of the interaction between cytochalasins and actin networks depends on the toxin species, the concentration, and the time of incubation. Direct eVects on actin occur at low CB concentrations, comparable to those which inhibit cell movements, and at least the relative potencies of CB and CD on cell movements (Atlas and Lin, 1978) and on actin are similar (Brenner and Korn, 1979; Brown and Spudich, 1979; Lin et al., 1980; MacLean‐Fletcher and Pollard, 1980). However, CE inhibits cell movements (Atlas and Lin, 1978) at much lower concentrations than it inhibits actin polymerization (Lin et al., 1980), binds to actin (Lin et al., 1980), or inhibits network formation (MacLean‐ Fletcher and Pollard, 1980). Species‐specific eVects are also relevant. CB, for example, failed to inhibit actin polymerization in the sperm acrosomal reaction (Sanger and Sanger, 1975). It can be concluded therefore, that cytochalasins modify the cell’s architecture, and subsequently cell motility and other cell functions by a number of interrelated eVects, including the alteration of the steady state interaction between G‐ and F‐actin, and by their respective interactions with supramolecular actin networks (Bray, 1992). Cytochalasins reduce the viscosity of actin gels, by both decreasing the average filament length through a change in the steady state between net polymerizing and depolymerizing ends, and by inhibiting the reannealing of spontaneous breaks in F‐actin. It is important to know, however, that the capping eVect of cytochalasins is not shared by other actin depolymerizing toxins, such as latrunculin A, which only elicits a tight 1:1 binding interaction with G‐actin (Bray, 1992). The steady state pool of actin organization, namely the actin network in the presence of cytochalasins might also shift the interaction with proteins that either block actin polymerization and, conversely, by proteins that block the net depolymerizing ends of actin filaments. This steady state of F‐actin pool in vivo would be therefore strongly influenced, by the disruptive eVect of cytochalasin itself, by localized ionic gradients created by channel function, and by proteins such as profilin which specifically interact with monomeric actin (Markey et al., 1978; Reichstein and Korn, 1979). On the basis of the dynamic steady state of the actin cytoskeleton, and in particular cortical cytoskeleton, it is necessary that each instance, in which cytochalasins are used to assess mechanosensitive channel function, is carefully evaluated. Important parameters to consider include in the endogenous state of the actin cytoskeleton, the concentration of the drugs used, and most importantly, the time of incubation with the drug.
244
Cantiello et al.
Agents that disrupt cytoskeletal organization, such as the fungal toxins cytochalasin or colchicine, which disrupt actin filaments and microtubules, respectively, have been shown to regulate stretch‐activated ion channel activity (Guharay and Sachs, 1984; Patel et al., 1998; Wan et al., 1999). In Lymnaea neurons, Johnson and Byerly (1993) originally determined that agents that modify cytoskeletal organization also alter Ca2þ channel activity. The cytoskeletal disrupters, colchicine and CB were both found to speed Ca2þ channel decline in ATP, whereas the cytoskeletal stabilizers, taxol and phalloidin, were found to prolong Ca2þ channel activity without ATP. In addition, cytoskeletal stabilizers reduced Ca2þ‐dependent channel inactivation (Johnson and Byerly, 1994). Thus, it was concluded that both channel metabolic dependence and Ca2þ‐dependent inactivation might be controlled by cytoskeletal interactions. Indeed, the state of cortical cytoskeleton organization is important in the control of voltage‐gated Ca2þ channels. In cultured neonatal mouse cardiac myocytes (NMCM), for example, we determined that CD disruption of the actin cytoskeleton blunts L‐type Ca2þ currents (Lader et al., 1999). This phenomenon, which is largely prevented by addition of the actin stabilizer phalloidin, could be mimicked in NMCM genetically deficient in the actin‐severing protein gelsolin (Lader et al., 1999). Whole‐cell and single‐channel recordings were obtained in retinal bipolar neurons of the tiger salamander (Maguire et al., 1998). In that study, we showed that acute (20–30 min) disruption of endogenous actin filaments with CD instead activated voltage‐gated Kþ currents in these cells, which was largely prevented by intracellular perfusion with phalloidin. Interestingly, direct addition of actin to excised, inside‐out patches activated and/or increased single Kþ channels. This is an important control experiment, as it strongly supports a direct cytoskeletal interaction, rather than a membrane‐induced change in channel function. The above evidence is thus indicative of a more general and quite appealing mechanism by which cytoskeletal structures control feedback mechanisms in voltage‐gated cation channels. Both activation and inhibition can be elicited by dynamic changes in cytoskeletal conformations. Insofar as mechanosensitive channel function is concerned, in Lymnaea neurons, for example, treatment with CB, CD, or N‐ethylmaleimide enhances mechanosensitive channel activity (Small and Morris, 1994; Wan et al., 1999). This was viewed as evidence of an eVect by cytoskeletal structures on channel function. Nonetheless, it is important to consider that only when the channel phenotypes are identified at the molecular level, we will know the precise nature of the interaction. In Cos‐7 cells transfected with either the mechanosensitive TRAAK or TREK‐1 CD reduces delay time in activation and enhances peak amplitude of Kþ channel activity (Patel et al., 1998; Maingret et al., 1999; Patel et al., 2001).
10. The Cytoskeletal Connection to Ion Channels
245
Colchicine has also been shown to enhance TRAAK channel activity (Patel et al., 1998). Thus, cytoskeletal connections may also have important consequences to the kinetic properties of mechanosensitive channels. In epithelial cell physiology, cytoskeletal integrity was early on implicated in transepithelial ion transport (Pratley and McQuillen, 1973) and the response to vasopressin (Hardy and DiBona, 1982). However, the role of actin filament organization on apical epithelial Naþ channel activity was unknown at the time. The first direct demonstration that cytoskeletal dynamics provide a functional interface with Naþ channels, came from work in our laboratory, where we tested whether the eVect of CD on Naþ channel activation was indeed mediated by a cytoskeletal connection and not a change in the membrane elasticity. In this study, we observed that addition of exogenous actin plus ATP, and actin–gelsolin complexes of known size to excised inside‐out patches (Cantiello et al., 1991), mimicked the eVect of CD. We first found that Naþ channels immunocolocalize with F‐actin from the cortical cytoskeleton suggesting that actin is always present in proximity to apical epithelial Naþ channels in A6 cells. Addition of the actin filament disrupter, CD (5 mg/ml) to cell–attached patches, induced Naþ channel activity within 5 min of addition. It is important to note that contrary to previous reports where other cytochalasins and exposure times were applied (Guharay and Sachs, 1984), our data strongly suggested a direct cytoskeletal‐gating mechanism. CD also increased Naþ channel activity in excised patches, further suggesting the remaining presence of cytoskeletal structures interacting with the channel. The first indication that indeed it was the cytoskeletal connection, and not changes in the plasma membranes arose from the fact that addition of short actin filaments (>5 mM) to excised patches also induced Naþ channel activity. Further, the eVect of actin on Naþ channel activity was reversed by addition of the G‐actin‐binding protein DNase I, and completely prevented by treatment of the excised patches with this protein (Cantiello et al., 1991). Addition of the actin cross‐linking protein, filamin A, reversibly inhibited both spontaneous and actin‐induced Naþ channels. Conversely, addition of short actin filaments in the form of actin‐gelsolin complexes in molar ratios <8:1 was also eVective in activating Naþ channels. This cytoskeletal interaction was found essential for the regulation by vasopressin and the PKA‐induced activation of epithelial Naþ channels ascribed to both Apx (Prat et al., 1993a,b), and ENaC (Berdiev et al., 1996). The Naþ channel complex of renal cells copurifies with ankyrin, fodrin, and actin itself (Smith et al., 1991). Thus, it is likely that actin either binds directly to the channel, or that actin may first interact with actin‐binding proteins, which in turn regulate‐by binding or other indirect interaction‐ion channel function. In favor of the former possibility,
246
Cantiello et al.
studies by Benos’ group provided strong evidence with regard to the role of actin on the heterotrimeric abg‐rENaC (Ismailov et al., 1997). Functional reconstitution of rENaC in lipid bilayers, where the components were restricted to this epithelial Naþ channel, indicated that actin itself modified rENaC channel kinetics. Further, only short actin filaments, but neither G‐ nor F‐actin, were able to regulate rENaC channel activity. Evidence for a direct interaction between ENaC and actin has only surfaced recently, confirming this hypothesis (Mazzochi et al., 2006). Interestingly, ENaC has been claimed to respond to membrane stretch both in Xenopus oocytes (Awayda and Subramanyan, 1998), and in reconstituted lipid bilayers (Awayda et al., 1995). The fact that this channel complex interacts with spectrin (Zuckerman et al., 1999) and binds directly to actin (Mazzochi et al., 2006), forwards the likely possibility that a cytoskeletal connection is at the center of epithelial Naþ channel putative mechanosensitivity, despite early controversy in this regard (Rossier, 1998). Although impossible to generalize presently, other ion channels not associated with either mechanosensitivity or excitable membranes have followed the path of the epithelial Naþ channels. Independent studies demonstrated that membrane‐resident CFTR, a cAMP‐ activated channel of nonexcitable cells, is functionally regulated by actin filament organization (Prat et al., 1994, 1995), in a process, which is independent of PKA stimulation (Prat et al., 1995). Interestingly, actin was found to bind and functionally and directly interact with CFTR (Chasan et al., 2002), providing further support to the idea that channel function can indeed be controlled by cytoskeletal dynamics. It is important to indicate, however, that this regulation has to be explored in detail, and no assumptions can be made, unless the state of the cytoskeleton in the expression system is clearly defined. Failure of eliciting amiloride‐sensitive Naþ currents in Xenopus oocytes, Apx was postulated to be a regulatory protein instead of a channel protein (Staub et al., 1992). However, expression of Apx in filamin A actin‐ binding protein (ABP‐280)‐deficient, and thus cytoskeletally deranged human melanoma cells clearly showed that Apx indeed induces a cation‐ selective conductance (Prat et al., 1996; Cantiello, 1999). These human melanoma cells also proved eYcient in further confirming the important role of actin filament organization in the control of CFTR function (Prat et al., 1999), and most relevantly, the response to cell volume regulation (Cantiello et al., 1993). In this study, we observed that ABP‐280‐deficient human melanoma cells, failed to respond to hypoosmotic shock with Kþ channel activation and the regulatory volume decrease (RVD) observed in most cells. Genetic rescue by expression of ABP‐280, however, restored RVD, indicating that the basal Kþ permeability of control cells is tonically inhibited, but rapidly activates by osmotic stress, in the presence of organized actin networks (Cantiello and Prat, 1996).
10. The Cytoskeletal Connection to Ion Channels
247
C. The Superfamily of TRP Channels The TRP superfamily comprises a large group of related cation channels that display a wide diversity of activating modes and cation selectivity (Montell, 2001; Vennekens et al., 2002; Voets and Nilius, 2003). All members of this superfamily share the same structural features: six putative transmembrane (6TM) domains, S1–S6, with a characteristic pore region between transmembrane segments S5 and S6, which is typical of voltage‐gated channels (Phillips et al., 1992; Birnbaumer et al., 1996). The sequence of this region is highly conserved across the superfamily. TRP proteins also display moderate homology with the extended family of voltage and cyclic nucleotide‐gated channels in particular the transmembrane domains of the pore region (Phillips et al., 1992). However, the voltage sensor present in the S4 domain of voltage‐gated cation channels is missing in TRP channels (reviewed in Montell, 2001; Minke and Cook, 2002). Furthermore, a 25‐amino acid ‘‘TRP domain’’ with unknown function is also present in the C‐terminus of most, but not all TRP channels. There are large variations in the amino and carboxyl cytoplasmic tails present in diVerent TRP channels, which may partially reflect their functional diversity. For example, the N‐termini of TRPC and TRPV channels contain several ankyrin repeats, whereas the TRPC and TRPM C‐termini contain proline‐rich motifs. Interestingly, some TRP channels also contain distinct putative motifs for binding of calmodulin (CaM), dystrophin, and the PDZ‐scaVolding protein ENAD in their C‐terminal tails (reviewed in Minke and Cook, 2002). These motifs may be relevant in regulation of channel function by such mechanisms as Ca2þ‐induced channel inactivation and cytoskeletal control of channel function. Evidence is mounting to suggest that TRP channels oligomerize with homolog or heterologous partners enabling channel complexes with distinct functional features (Birnbaumer et al., 1996; Xu et al., 1997; Tsiokas et al., 1999; Lintschinger et al., 2000). An emerging consensus is that most TRP channels play important roles in various sensory physiologies (Clapham, 2002, 2003; Voets and Nilius, 2003). TRP channels participate in numerous sensory transduction responses, including hearing, vision, and thermosensation, as well as recently identified responses to a variety of physical stimuli including heat, cold, osmolarity, stretch, shear flow, and pressure (Montell, 2001; Montell et al., 2002a; Voets and Nilius, 2003). From the evolutionary point of view, the TRP channel superfamily is highly conserved throughout animal phylogeny. There are at least 28 mammalian members identified thus far, which have been classified into multiple subfamilies (Montell, 2001; Montell et al., 2002a; Voets and Nilius, 2003). In terms of their function, TRP proteins are nonselective cation channels with diverse cation perm‐selectivity properties, including a high Ca2þ
248
Cantiello et al.
selectivity in some, but not all TRP channels (Minke and Cook, 2002). The first two TRP channels were identified in Drosophila as light sensitive channels with either very high (PCa:PNa>100:1) or moderate (PCa:PNa, 4:1) Ca2þ permeability. The low Ca2þ selective channel is encoded at least in part by the TRP‐like (TRPL) gene, which shares 40% homology with the canonical TRP (Phillips et al., 1992; Niemeyer et al., 1996). Subsequently, a number of new members have also been isolated from various eukaryotic species, which are associated with diVerent sensory stimuli, including cold and heat, osmotic challenges, and other receptor stimulatory responses (Montell, 2001; Voets and Nilius, 2003). The TRP family was originally classified (Harteneck et al., 2000) into short (TRPC), osm‐9‐related (TRPV, vanilloid), and long (melanostatin‐related, TRPM) channels, based on the protein length, and thus potential regulation of their cytoplasmic tails (Harteneck et al., 2000). The TRPC (canonical) subfamily is the closest to the Drosophila TRP. The osm‐9‐like gene encodes a TRP protein associated with osmotic responses in Caenorhabditis elegans (Strotmann et al., 2000). TRP have now been extended to other 6TM transmembrane proteins (TRPP, TRPML) with weaker homology but potentially similar topological features and regulatory roles in cell function. Thus, a new encompassing and comprehensive nomenclature for the TRP superfamily has recently been adopted, where there are group‐1 comprising five subfamilies (TRPC, TRPV, TRPM, TRPN, and TRPA) and more distant group‐2 comprising two subfamilies (TRPP and TRPML). The recently discovered epithelial Ca2þ channels CaT1 and ECaC, have now been renamed TRPV6, and TRPV5, respectively (Peng et al., 1999; Yue et al., 2001; Hoenderop et al., 2002). TRPV5–6 may represent a major contributing factor to the apical Ca2þ absorption step in transporting epithelia. The Ca2þ permeable nonselective cation channels PC2 and PCL (Mochizuki et al., 1996; Chen et al., 1999) have been incorporated to the TRPP subfamily of TRP proteins (Montell et al., 2002b). The PC2 topologically similar protein, mucolipin‐1 (TRPML subfamily), which is genetically linked to mucolipidosis type IV is also a cation channel (Raychowdhury et al., 2004). Mucolipin homologs have also been implicated in sensory functions (Di Palma et al., 2002). The widespread distribution of TRP channels among excitable and nonexcitable cells and the fact that most TRPs permeate Ca2þ has forwarded the working hypothesis that TRP channel function underlies the ubiquitous ‘‘capacitative’’ Ca2þ response (Birnbaumer et al., 1996; Golovina et al., 2001; Putney et al., 2001). This was originally supported by the finding that signals as PLC activation, lead to the opening of the Drosophila TRP. Hydrolysis of phosphatidylinositol‐4,5‐bisphosphate (PIP2) by PLC generates second messengers such as inositol 1,4,5 trisphosphate (IP3) and DAG. This leads to cascades involved in the production of polyunsaturated fatty acids and elicit Ca2þ store‐activation responses (Xu et al., 1997;
10. The Cytoskeletal Connection to Ion Channels
249
Chorna‐Ornan et al., 2001). A number of TRP channels are regulated by various branches of this metabolic pathway (Minke, 2001; Minke and Cook, 2002). However, TRP channels appear to be activated by a wider variety of stimuli, chiefly among which is mechanical stimulation (see below). TRP channels are also regulated by a number of specific intracellular ligands mostly associated with their cytoplasmic domains. These include putative PKC (TRPC, TRPV), and other kinases PKA (TRPP1‐TRP2, TRPV1–2, 5–6‐ CaT1, ECaC1), and PI3K (TRPC2–3, 5–7, TRPV1, 4, 6), Ins3P receptor and CaM (all TRPC), and PDZ domain‐containing proteins (TRP, TRPC4, 5), including INAD and NINAC myosin III (reviewed in Montell, 2001; Li et al., 2002). Other putative ligands for intracellular regulation include cytoskeletal proteins, such as dystrophin‐like motifs (TRPC1), and interactions with troponin‐I, and tropomyosin‐1 (PC2 or TRPP2). Other expected ligands for intracellular regulation may involve directly or indirectly trimeric and small G‐proteins, ATP, InsP3, and DAG, and arachidonic acid ligands such as arachidonic acid (AA) itself. AA byproducts of potential DAG kinases and lipases reactions include anandamides (TRPV1), and HPETE (TRPV), which are linked to regulation of the vanilloid receptors (reviewed in Benham et al., 2002). Interplay among activation mechanisms is apparent by the fact that cell swelling, a common activating factor of various TRP channels also activates phospholipase A2 (PLA2). This activity increases arachidonic acid production (Basavappa et al., 1998), and downstream metabolites, such as 50 ,60 ‐ epoxyeicosatrienoic acid, which has been shown to activate TRPV4 (Watanabe et al., 2002). Further, inhibition of either PLA2 or cytochrome P450 strongly inhibited the hypotonicity‐induced TRPV4 channel currents (Vriens et al., 2004).
D. TRP Channels and Mechanosensation Much attention has been placed on potential roles of TRP channels in mechanosensory responses, particularly in view of the fact that they are targets of various environmental cues such as sound, light, pressure, and osmotic imbalance (Clapham, 2002, 2003; Corey, 2003; Voets and Nilius, 2003). TRP channel members may thus be considered novel mechanosensitive channels (Birnbaumer et al., 1996; Clapham, 2003; Corey, 2003). Two models have been proposed for the signal transduction mechanism, and gating of mechanosensitive channels. Some mechanosensitive channels (Hamill and Martinac, 2001) appear to be gated by direct changes in membrane tension, which is generated in the lipid bilayer on osmotic imbalance. A study identified the mechanosensitive channel of Xenopus oocytes, as TRPC1, which may not require cytoskeletal components, but dynamic
250
Cantiello et al.
changes in membrane structure to elicit activation (Maroto et al., 2005). Nonetheless, the phenomenon known as adaptation of mechanosensitive channels may implicate cytoskeletal elements (Hamill and McBride, 1992; Small and Morris, 1994; Laitko et al., 2006). An alternative model may rely more on the tethering by cytoskeletal proteins which play a key role in transducing mechanical deformation (Corey, 2003). The first mechanosensory TRP channel to be described in flies was ‘‘no mechanoreceptor potential C’’ NompC (TRPN1), the defining member of the TRPN subfamily (Walker et al., 2000). Mechanoelectrical responses in bristle sensory neurons occur rapidly on deflection of the bristle hair shaft and result from the opening of the NompC. Another Drosophila TRP known as ‘‘painless’’ (Tracey et al., 2003) is also involved in mechanosensation. The painless mutant is required for mechanical nociception (Tracey et al., 2003). ‘‘Painless’’ contains eight tandem ankyrin repeats in the N‐terminal domain. A distinguishing feature of NompC is the presence of 29 ankyrin repeats between the N‐terminus and the first transmembrane segment (Walker et al., 2000). The role of this large tandem array of ankyrin motifs is not known, but it is proposed to form the gating spring that leads to opening of the channel pore (Howard and Bechstedt, 2004). Other sensory responses implicate the C. elegans osmotin‐like protein‐9 (OSM‐9) channel. OSM‐9 was originally thought to be a mechanosensitive channel because osm‐9 mutants are defective in osmotic avoidance and in sensitivity to nose touch. OSM‐9 is now considered a member of the TRP vanilloid (TRPV)‐related subfamily of TRP channels, which contains three ankyrin repeat domains at its N‐terminal intracellular domain. One particular subfamily of TRP members (vanilloid, TRPV) is emerging as quintessentially sensory channels, implicated in mechanosensation (Montell et al., 2002a; Clapham, 2002, 2003; Voets and Nilius, 2003). Vertebrate TRPV channels are sensitive to various forms of physical and chemical stimuli (O’Neil and Heller, 2005), whose response eVects an increased Ca2þ permeability. TRPV1–4 channels are moderately Ca2þ‐selective, while TRPV5 and TRPV6 are highly selective Ca2þ channels (den Dekker et al., 2003). Much of this information was originally obtained from studies in lower organisms. The Drosophila genome, for example, harbors two genes for TRPV‐like channels. One gene (iav) encodes inactive (IAV), a protein that is related to OSM‐9 and the second gene (nan) encodes Nanchung (NAN). Both genes encode channels implicated in hearing in Drosophila (Kim et al., 2003; Gong et al., 2004), which when expressed functionally in vitro are activated by hypoosmolarity (Kim et al., 2003; Gong et al., 2004). This feature is very similar to the osmosensitivity observed in cells expressing the mammalian TRPV4 (Liedtke et al., 2000; Strotmann et al., 2000; Wissenbach et al., 2000). Other TRPV channels including TRPV2 share
10. The Cytoskeletal Connection to Ion Channels
251
some gating characteristics of TRPV4, as TRPV2‐expressing cells, respond to both hypotonic cell swelling and application of membrane stretch, with increased Ca2þ influx (Muraki et al., 2003). TRPV4 osmosensitivity was originally surmised by its activation by hypotonic cell swelling (Liedtke et al., 2000; Strotmann et al., 2000). The molecular mechanism(s) of the TRP‐mediated osmosensing response, is far from been clearly understood, however, exposure to hypotonic media on both sides of the membranes failed to activate TRPV4, suggesting that cell swelling induced by hypotonicity in the media, aVected the activation signal. This was hypothesized to reflect a response to membrane stretch instead. However, direct application of pressure to the cell failed to activate the channel at room temperature (Strotmann et al., 2000). This could be reversed at higher temperatures (Guler et al., 2002; Watanabe et al., 2002). Thus, other physical parameters sensitize the channel, including hypotonic media and application of shear stress (Gao et al., 2003; O’Neil and Heller, 2005). TRP channels from the TRPN and TRPA subfamilies are additional candidates for mechanosensation, in particular for transduction of sound. Interestingly, TRPA1 was first cloned from mammalian fibroblasts (Jaquemar et al., 1999) and originally thought to be instead a thermo‐sensitive and ligand‐gated channel (Story et al., 2003; Jordt et al., 2004). Evidence indicates that TRPA1 is also a mechanosensitive channel (Corey et al., 2004). Knockdown of TRPA1 expression in zebrafish hair cells markedly impaired the transduction channel electrical activity of the otocyst to sound vibrations (Corey, 2003; Corey et al., 2004). Similarly, knockdown of TRPA1 in cultured mouse utricle hair cells inhibited electrical activity associated with hair‐cell transduction (Corey, 2003; Corey et al., 2004).
E. Cytoskeletal Connections in TRP Channels The widespread display of sensory functions in which TRP superfamily members are involved underlies the relevant role of associated proteins, which may help sense, or otherwise transduce physical forces into their regulation. Montell et al. have extensively reported on the connections made by TRPs, to the anchoring cytoskeleton (Wes et al., 1999; Xu et al., 2001; reviewed in Montell, 2005). In brief, TRPC binds directly to a scaVold protein, ‘‘inactivation no afterpotential‐D’’ (INAD), which consists of five protein interaction modules referred to as postsynaptic density/discs‐large/ zonula occludens (PDZ) domains (Huber et al., 1996; Shieh and Zhu, 1996). In addition, this ‘‘core complex’’ forms a large macromolecular assembly known as the signalplex (Xu et al., 1998), containing other signaling proteins such as the ‘‘neither inactivation nor afterpotential‐C’’ (NINAC) myosin III (Wes et al., 1999). In addition, the mechanosensitive properties of TRP
252
Cantiello et al.
channels have been associated with the cytoplasmic tails of the protein. The N‐terminus of most TRP channels (but notably not TRPP subfamily members) contains several ankyrin repeat domains. Ankyrin repeat motifs are present in tandem copies, which are considered to associate with the cytoskeleton and thus mediate protein–protein interactions of a tethered mechanism required for mechanical gating. To support this contention, studies indicate that ankryn motifs contain elastic properties (Sotomayor et al., 2005), whose reversible spring behavior may be relevant in mechanosensation (Lee et al., 2006). TRPV4, for example, has three ankyrin repeat domains (Liedtke et al., 2000), a fact that may help explain its implication in vertebrate mechanosensation, in that, it can sense hypoosmotic stress (Alessandri‐Haber et al., 2003). However, a direct correlation between the lengths of this potential tethering TRPV channels, most known for their mechanosensitive properties, contains three to five ankyrin repeat motifs. Cloning of the nompC gene revealed a TRP channel protein (NompC, TRPN) with 29 N‐terminal ankyrin repeats. TRPN is only distantly related to other TRP families (Corey, 2003). Corey et al. (2004) showed that TRPA1 (also called ANKTM1), which contains 17 ankyrin motifs, constitutes, or is a component of, the mechanosensitive transduction channel of vertebrate hair cells. Despite the fact that the activation mechanism elicited by mechanical force is unknown, the spring‐like structure of the ankyrin repeats is consistent with a ‘‘tethered channel’’ model (Howard and Bechstedt, 2004; Sotomayor et al., 2005; Lee et al., 2006). Given that TRP channels are most likely tetramers (Hoenderop et al., 2003; other stoichiometric interactions have been postulated as well; Flockerzi et al., 2005), the ankyrin motifs in each monomer may also help the channel subunits to assemble. Arniges et al. (2006) provided evidence that the ankyrin motifs are implicated role in the multimeric assembly of TRPV4 channels. Specific cytoskeletal connections in TRP channels may actually be a distinguishing feature among diVerent TRP subfamilies, thus providing wider variety of regulatory connections. Goel et al. (2005) identified several proteins that interact with the TRPC5 and TRPC6 channels, which are chiefly localized to specialized postsynaptic dendritic spines in the rat brain, where they may play a critical role in synaptic responses to neurotransmitters. Twenty‐eight proteins were identified in the TRPC5 immunoprecipitate from rat cerebral cortex, including the prominent actin, and other cytoskeletal proteins including spectrin, nonmuscle myosin, a‐actinin, and tubulin. The interaction between TRPC5 and TRPC6 with a‐actinin, actin, and drebrin, was confirmed by immunoprecipitation followed by Western blot analysis. Remarkably, the a‐3 subunit of the Naþ, Kþ‐ATPase, the main component of the Naþ pump was also identified as an interacting partner with both TRPC5‐ and TRPC6‐channel proteins (Goel et al., 2005).
10. The Cytoskeletal Connection to Ion Channels
253
The Naþ pump is known to bind ankyrin, which in turn binds to spectrin, linking the transporter to the underlying cytoskeleton. Thus, the Naþ pump may immunoprecipitate with TRPC5 via an interaction with spectrin or may directly bind to the ankyrin‐like repeats found in the channel proteins. A role of other TRP channel members in mechanosensory function has been dismissed based on the fact that no discrete consensus sequences—namely ankyrin repeats are found in TRPP2 (PC2), for example (Delmas, 2005). However, clear cytoskeletal connections have been established between PC2 and the various components of the cytoskeleton. Originally, Gallagher et al. (2000) determined that Hax‐1, a cytoskeletal protein that interacts with the F‐actin‐binding protein cortactin, interacts with PC2. Chen and collaborators have extensively explored PC2 interacting proteins and found that both cytoskeletal proteins troponin‐I (Li et al., 2003b) and tropomyosin‐1 (Li et al., 2003a) directly bind to PC2, further strengthening a link between the cytoskeleton and the PC2 channel. We have expanded on these findings, to demonstrate that the cytoskeletal connections of PC2 are a key component of a novel sensory mechanism based on dynamic changes in the actin cytoskeleton attached to the channel.
III. ROLE OF ACTIN CYTOSKELETAL DYNAMICS IN PC2‐MEDIATED CHANNEL FUNCTION A. Role of PC2 in Health and Disease ADPKD describes a group of genetic disorders with almost identical clinical features, collectively aVecting 1:1000 of the world’s population. ADPKD is largely (95%) caused by mutations in the PKD1 and PKD2 genes. Clinical manifestation of mutations in either gene are largely similar, both in human and animal models, suggesting the current working hypothesis that the encoded transmembrane proteins, PC1 (polycystin‐1, TRPP1) and PC2 (polycystin‐2, TRPP2), both recent additions to the superfamily of TRP channels, form a functional complex associated with cell‐signaling events. Thus, a partnership between PC1 and PC2 converges into a common signaling cascade which is now thought to involve Ca2þ transport. The molecular steps linking this signaling pathway to renal cell function have only recently become apparent. Studies implicate the PC1‐regulated and PC2‐mediated Ca2þ entry, as a sensory mechanism for renal epithelial cell function (Nauli et al., 2003). Further understanding of the molecular steps in this signaling pathway stems from the fact that PC2 is a TRP channel (Montell et al., 2002b). Studies (Gonza´lez‐Perrett et al., 2001; Vassilev et al., 2001; Koulen et al., 2002) confirmed the original hypothesis
254
Cantiello et al.
that the PKD2 gene product is a cation‐selective channel (Mochizuki et al., 1996). Nevertheless, the actual mechanism(s) associated with the onset and development of cyst formation in ADPKD remains largely unknown. It is speculated that cyst formation in ADPKD implicates dysfunctional ion transport and/or abnormal cell growth in target epithelia, likely elicited by dysregulation of, PC2‐modified Ca2þ signals. Our understanding of the role(s) PC2 plays in renal cystogenesis derives from two issues still being studied and assessed. First, the cellular localization and putative interaction of PC2 with expected partners such as PC1, remains a matter of current interest (Koulen et al., 2002; Luo et al., 2003). PC1 and PC2 cell expression is developmentally disconnected, such that either protein may be independent of each other, and/or associated with other likely partners (Delmas, 2004). Work by Wilson and collaborators indicates that PC1 is part of cell adhesion complexes, implying a role in cell‐cell interactions, and cell–matrix adhesion. These complexes entail supermacromolecular structures containing focal adhesion proteins, including talin, vinculin, p130Cas, FAK, paxillin, pp60c‐src, and a‐actinin in human renal epithelial cells, when cell–matrix interactions prevail (Geng et al., 2000). Conversely, PC2, acting as a TRP channel is capable of interactivity with PC1, as expected (Nauli et al., 2003), but also with other TRP channels (Tsiokas et al., 1999), such that its functional and regulatory properties may diVer depending on location, abundance, and complexing with specific partners (Delmas, 2004). Second, diVerent functional profiles as to how PC2 acts as a channel have been depicted from work in diVerent cell systems (Gonza´lez‐Perrett et al., 2001; Vassilev et al., 2001; Koulen et al., 2002). Nonetheless, the current hypothesis is that PC2 function is implicated in Ca2þ signaling, as evidence suggests the ability of PC2 to mediate Ca2þ influx into renal epithelial cells (Luo et al., 2003; Nauli et al., 2003). Interestingly, other forms of renal cystic disease do not implicate directly PKD genes, but instead proteins associated with the axonemal machinery and ciliary structures (reviewed in Calvet, 2002, 2003; Cantiello, 2003). Thus, it is possible, that diVerent pathophysiological events likely converge to common points of PC2 regulation. For example, the C‐terminus of PC1 structurally interacts (Tsiokas et al., 1997), and regulates PC2 channel function (Xu et al., 2003). This provides a recognizable molecular aspect of the PC1–PC2 complex. A normal PC1– PC2 complex is required for proper sensory function of renal epithelial cells (Nauli et al., 2003), whose primary cilia respond to bending by volume flow and shear stress, with Ca2þ entry and cell activation (Praetorius and Spring, 2001; Praetorius et al., 2003). PC2 also interacts with a number of other proteins, including TRP channels (Tsiokas et al., 1999; Delmas, 2004), and cytoskeletal proteins (Gallagher et al., 2000; Li et al., 2003a,b), providing testable hypotheses as to putative mechanisms of PC2 regulation. In the
10. The Cytoskeletal Connection to Ion Channels
255
following sections, we assessed the various aspects of PC2 regulation by the actin cytoskeleton. We determined the role of cytoskeletal dynamics in PC2 regulation, the molecular links that anchor this channel protein to actin networks, and how the three‐dimensional structure of such a gel, likely acts as a tethering structure of the channel, in brief, eliciting a novel signal transducer, the PC2 channel–cytoskeleton interface.
B. Presence of Actin and Associated Proteins and Effect of CD on Channel Activity in hST To assess a regulatory role of the actin cytoskeleton on PC2‐mediated channel function, enriched human syncytiotrophoblast (hST), apical vesicles were reconstituted in a lipid bilayer system. Experiments were conducted in the presence of a Kþ chemical gradient, with 150 mM in the cis compartment and 15‐mM KCl in the trans compartment, respectively as originally reported (Gonza´lez‐Perrett et al., 2001). Experiments, where no (or little) spontaneous activity was originally observed at the beginning of the experiment, were chosen (Fig. 1A). Addition of CD (5 mg/ml; n ¼ 17) to the cis compartment initiated Kþ‐permeable ion currents (Fig. 1A–D). CD‐activated membrane currents increased eightfold, from 0.023 0.019 pA, to 0.217 0.154 pA (n ¼ 8, p < 0.01) at 7.42 0.28 min, n ¼ 5, after exposure to the drug. Currents were highly cation selective (Fig. 1C), and further characterized as those previously observed as mediated by PC2 (Gonza´lez‐Perrett et al., 2001), with a single‐channel conductance of 135 pS (n ¼ 3, Fig. 1C), and inhibition by La3þ and 50 mM amiloride (data not shown). The hST apical membranes that were incubated for 1–3 days at 4 C in the presence of CD (5 mg/ml), to completely collapse the actin networks, only displayed flickery, sporadic channel openings, and very little activity (Fig. 1D). Inmunofluorescence analysis of the hST vesicles (Fig. 1E), indicated that PC2 colocalized with F‐actin, which was disrupted by addition of CD. Western blot analysis (Fig. 1F) determined the presence of PC2, actin and the actin‐binding proteins, a‐actinin, and gelsolin in the hST vesicles. This was confirmed by the colabeling of TRITC‐phalloidin and antiactin antibodies, to label F‐actin, and the entire actin pool, respectively (Montalbetti et al., 2005b). Although both monomeric and polymeric actin, were observed in the intravesicular compartment, most F‐actin displayed stronger labeling in proximity to the membrane. Incubation of hST apical membranes with CD (10 mg/ml, Fig. 1E, bottom) for 1 h at 4 C aVected the presence of F‐actin in the vicinity of the membrane. An extended incubation with CD (>24 h), further collapsed cytoskeleton where most actin was ‘‘detached’’ from the plasma membrane. These findings provided the first indication that a
256
Cantiello et al.
A
B
+CD
30 mV
Control
After CD 10 pA
Npo
25 s
1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0
5 pA 400 ms
+CD
−200
−400
0 200 Time (s)
400
600
D
C 25
Control
30 mV
Self-inhibition
40 mV
After CD
40 mV
I, pA
20
1.2
5 pA
*
1.0
15
100 ms
F
Npo
0.8 10
0.4
5
150 100 75
0.6
*
0.2 Vh, mV −⎯100−80−60 −40 −20
20 40 60 80 100
−5
E
Actin label
*
0.0
PC2 label
Control 4 ⬚C 15 min >24 h >24 h Cytochalasin D
Merged
Polycystin-2
150 100 75
a -Actinin
150 100 75 Gelsolin
Control
75 50
Actin
37 1 2 3 Sample A
1 2 3 Sample B
+CD
FIGURE 1 EVect of cytochalasin D on PC2 channel activity in hST. (A) Top: spontaneous ion channel activity increased after addition of cytochalasin D (CD, 5 mg/ml) to the cis chamber. The hST apical vesicles were reconstituted in the presence of a KCl chemical gradient, with 150 mM KCl in cis, and 15‐mM trans compartments, respectively. Bottom: average channel activity before and after CD addition, is expressed as the number of channels, times the open channel probability (n ¼ 19). (B) Expanded tracings of single‐channel currents. CD activation preserves the single‐channel conductance (135 pS, n ¼ 3; n ¼ 19). (C) Highly cation‐selective single‐channel conductance of the CD‐activated Kþ permeable channels was 135 pS conductance. Experimental data (filled circles) are indicated as mean SEM (n ¼ 3). The solid line is the fitting of data with the GHK equation. Dashed line indicates spontaneous PC2 single‐channel conductance
10. The Cytoskeletal Connection to Ion Channels
257
PC2–actin cytoskeleton interface is a functional unit whereby changes in the actin structures control channel function and thus ion transport. C. Effect of Gelsolin and Actin on PC2 Channel Activity in hST To further test the nature of the changes in endogenous actin filament organization, which mediate the regulation of PC2, the eVect of cytoskeletal reorganization on PC2 was further explored by addition of the Ca2þ‐ dependent actin‐severing protein, gelsolin (Montalbetti et al., 2005b). To support the physiological relevance of this maneuver, we first confirmed the presence of gelsolin in the hST vesicle preparation by Western blot analysis (Fig. 1F). This was the first direct demonstration of gelsolin in cortical cytoskeleton of hST (Montalbetti et al., 2005b). Addition of gelsolin (40 nM) to control membranes, in which the Ca2þ concentration in the cis compartment was kept at 10 mM, increased Kþ channel activity in 15 out of 16 experiments (Fig. 2A). The Ca2þ concentration used for this study was similar to that used for assessing hST spontaneous Kþ channel activity (Gonza´lez‐Perrett et al., 2001). The current–voltage relationship remained as observed for the control channel (Fig. 2B). Membrane currents increased 85‐fold, from 0.022 0.016 pA, n ¼ 14 to 1.89 1.60 pA, n ¼ 7, for control vs gelsolin‐treated membranes, respectively (p < 0.01). Addition of gelsolin, either in the absence of Ca2þ or the presence of Ca2þ (10 mM) plus EDTA (10 mM), was without eVect on channel activity (0.024 0.018 pA, n ¼ 10 vs 0.051 0.027 pA, n ¼ 6, p < 0.2, Fig. 2C, bottom). However, this lack of gelsolin eVect in the absence of cis Ca2þ reversed (in five out of six experiments) after further addition of Ca2þ (10 mM) to the trans chamber (Fig. 2D). Under these conditions, membrane currents increased to 1.22 0.69 pA, n ¼ 5 (p < 0.01) in the presence of cis gelsolin (40 nM). The activation observed after addition of Ca2þ to the trans chamber was mediated by Ca2þ (Gonza´lez‐Perrett et al., 2001). (D) Average data for mean currents before, after 10 min addition of CD, and after chronic treatment with the drug (n¼8, p<0.05 between control and acute CD treatment). (E) Top: colocalization of actin filaments and PC2 in control hST vesicles. The hST apical membranes were colabeled with TRITC‐phalloidin and an anti‐PC2 antibody to determine the presence of actin filaments and the channel, respectively. Colocalization of F‐actin and PC2 was confirmed in both the plasma membrane and the intravesicular compartment. Bottom: treatment of hST membrane vesicles with CD (10 mg/ml) for 24 h modified the actin cytoskeleton, with extensive detachment from the membrane. (F) Western blot analysis of hST apical membrane vesicles confirmed the presence of PC2 and cytoskeletal components, including actin, and the PC2‐ associated proteins a‐actinin, and gelsolin. Treatment of hST membrane vesicles with CD (10 mg/ml) for either 1 (labeled 2) or 24 h (labeled 3) had little eVect on the amount of actin and actin‐associated proteins present in the vesicle preparation. Data reproduced from the Journal of Physiology (Montalbetti et al., 2005b), with permission.
258
Cantiello et al.
A
B Control
I, pA
25
20 mV 20
Self inhibition, 5 min
40 mV
After gelsolin
40 mV
15
10
5
C
5 pA
3.0 2.5
20
−80 −60 −40 −20
40
60 80 Vh, mV
200 ms
*
−5
D
Npo
2.0 1.5
After gelsolin (cis) + Ca2+ 90 mM (trans)
1.0
40 mV
0.5 0.0
* Control
+Gelsolin
−60 mV
2.0 *
1.5
1 pA
Npo
500 ms
1.0
F 20
I, pA
0.5 0.0
16
* Control
+Ca2+
+Gelsolin
E
12 K+ 150/15 mM
1.0 8 0.8 4
0.6 * 0.4 Npo
*
*
*
* 0.2
K+ 150/15 mM Ca2+ 0/90 mM
*
*
*
−80 −60 −40 −20
* *
20
40
60 80 Vh, mV
−4
0.0 −0.2 0
40
80
120
160
Time (s)
FIGURE 2 EVect of gelsolin on PC2 channel activity in hST. (A) The actin‐severing protein gelsolin activated PC2 channel activity in hST membranes in the presence of 10 mM Ca2þ (n ¼ 7). (B) Current–voltage relationship of cation‐selective channel activated by addition of gelsolin is identical to that activated by CD. (C) Top: average data are the mean SEM of 13 and
10. The Cytoskeletal Connection to Ion Channels
259
transport through the Kþ permeable channels (Fig. 2D). The regulatory eVect of actin network remodeling on PC2 channel function was confirmed by inwardly (Ca2þ‐carrying) currents in the presence of external Ca2þ (90 mM). This is reflected in an increased open probability of the channel (Fig. 2E), while the single‐channel conductance was much lower (Fig. 2F) as originally reported (Gonza´lez‐Perrett et al., 2001). The data are consistent with a feedback mechanism where the transported Ca2þ from the trans chamber activates ‘‘cytoplasmic’’ gelsolin (cis compartment) thus remodeling the endogenous actin network. This was confirmed by a diVerent strategy, namely, the direct addition of exogenous actin (1 mg/ml) to spontaneously active, control hST apical membranes. Interestingly, addition of monomeric actin inhibited Kþ channel activity in 15 out of 17 experiments (Fig. 3). The mean membrane current decreased by 74.2%, from 0.048 0.011 pA, n ¼ 5, to 0.012 0.007 pA, n ¼ 5 (p < 0.03). This was unexpected, as most channel function activated by short CD treatment is usually mimicked by addition of polymerizing concentrations of monomeric actin (Cantiello et al., 1991; Prat et al., 1994, 1995). However, this inhibitory eVect of actin was not observed in the presence of CD (< 15 min), in which channel activity remained high (data not shown). In chronically CD‐treated membranes (>24 h), in contrast, PC2 channel activity was reactivated by addition of actin (Fig. 3C). In three out of three experiments, membrane currents increased from 0.0005 0.0001 pA to 0.026 0.014 pA, n ¼ 3, p < 0.05, for the absence and presence of actin, respectively. These data are most consistent with a scenario in which competition occurs between exogenous (monomeric) actin and the endogenous pool of actin filaments, and likely channel‐associated proteins. The CD and gelsolin data are in agreement with a role of actin network remodeling in the regulation of PC2 channel activity.
6 experiments, for control (n ¼ 13) and gelsolin‐treated (n ¼ 6) membranes, respectively. Bottom: in a Ca2þ‐free solution, however, gelsolin (30 mM) was unable to induce channel activation. Further addition of Ca2þ (10 mM) to the cis chamber restored stimulation. Average data are the mean SEM for the gelsolin eVect in the absence (n ¼ 6) and presence of Ca2þ (n ¼ 5) added to the cis chamber (30 mM). Control data were the average of 10 experiments. (D) Addition of Ca2þ to the trans chamber, reactivated channel activity in the presence, but not the absence of cis gelsolin. The data indicate that Ca2þ transport feeds back into gelsolin activation, and thus reinitiation of hST cation channel activity (n ¼ 5). (E) The presence of gelsolin (100 nM) in the cis compartment decreases the inhibitory eVect of Ca2þ transport through cation‐permeable channels in hST. Data are the mean SEM (n ¼ 5), obtained as Npo of current activity at positive potentials. Asterisks indicate statistical diVerence at least p < 0.05. (F) Ca2þ transport through the channel reduces the single‐channel conductance and shifts the reversal potential by 60 mV ( 54 to þ10 mV). Data reproduced from the Journal of Physiology (Montalbetti et al., 2005b), with permission.
260
Cantiello et al. B
A
+G-actin 30 mV
Control
1.0
Npo
0.8
1 min after G-actin
0.6 0.4 0.2 0.0
5 min after G-actin
−60 −40 −20
C
0
20
80 100
60
0.6
5 pA 500 ms
0.5
60 mV
0.4 Npo
Control, CD >24 h
40
Time (s)
D
*
0.3 0.2
After G-actin
0.1 0.0 5 pA
* * CD >24 h +G-actin
Control
500 ms
F 40 mV
−5
1000 N/Div
After F-actin
15
After amiloride 100 mM
5 pA 500 ms
−5
*
0.5 0.4 0.3
0 5 10 Amplitude (pA)
15
1000 N/Div
−5
0.6 0 5 10 Amplitude (pA)
Npo
Control
1000 N/Div
E
0.2 0.1 0.0
0 5 10 Amplitude (pA)
* Control +F-actin
+Amil
15
FIGURE 3 EVect of actin on cation channel activity in hST. (A) Representative single‐ channel tracings of hST apical membranes in asymmetrical KCl. Addition of actin (1 mg/ml) to the cis chamber inhibited spontaneously active ion channel currents (n ¼ 17). (B) Average channel activity expressed as Npo, where N is the total number of active channels, and po is the channel’s open probability indicate that monomeric actin inhibited channel activity within 1 min (arrow; n ¼ 5). (C) Addition of actin to chronically CD‐treated membranes (>24 h), in contrast, stimulated otherwise largely quiescent membranes. (D) Summarized data are indicated as means SEM for control conditions (n ¼ 7), and after addition of actin to acutely (15 min, center, n ¼ 7) and chronically (24 h, right, n ¼ 3) CD‐treated membranes. While actin addition was inhibitory to control membranes, the same actin concentration was stimulatory in
10. The Cytoskeletal Connection to Ion Channels
261
IV. IDENTIFICATION OF ACTIN‐BINDING PROTEIN INTERACTIONS WITH POLYCYSTIN‐2 A. Interaction Between PC2 and a‐Actinin Revealed by Yeast Two‐Hybrid System To begin a search for the molecular determinants of the cytoskeleton–PC2 functional interaction, we used the yeast two‐hybrid system to screen proteins which interact with the intracellular N‐terminus of PC2 (PC2N, amino acids M1‐K215; Li et al., 2005). A bait construct, pGBKT7‐PC2N, was used to screen a human heart cDNA library. One plasmid isolated from the library represented the C‐terminal two‐thirds of human muscle‐type a‐actinins, which forms a complex in vivo in cells and tissues (Fig. 4A). The identified a‐actinin‐2 cDNA is a 3‐kb fragment starting at nucleotide C1187 (aa L341), which encodes most of the central domains (including three spectrin‐like repeats) and the C‐terminus of a‐actinin. Given that mammalian a‐actinin has four isoforms with high sequence similarity, we further explored whether nonmuscle a‐actinin‐1 can also bind PC2. Indeed, a‐actinin‐1 associated with PC2N albeit not as strongly as the a‐actinin‐2–PC2N interaction (Fig. 4A). We also performed a similar yeast two‐hybrid assay to determine whether the C‐terminus of PC2 (PC2C, aa 682–968) also interacts with a‐actinins (Fig. 4A). Interestingly, while the entire cytoplasmic PC2C showed no interaction with a‐actinins 1 or 2, shorter segments within PC2C displayed strong interaction with both a‐actinins (Fig. 4A). This finding suggested the possibility that a domain within the first part (aa 682–820) of PC2C may inhibit the PC2C–a‐actinin interaction. We narrowed down this interaction to a smaller segment of 58 amino acids (PC2CC, aa 821–878), which is responsible for association with a‐actinins. Interestingly, the segment PC2CC, which binds a‐actinin, overlaps with the domain that 35 interacts with tropomyosin‐1 (Li et al., 2003a). On the other hand, while the spectrin‐like domain II of a‐actinin‐2 seemed to be required for association with the PC2 C‐terminal segments PC2CA (aa 821–968) and PC2CC, we found that the domain IV alone was responsible for mediating association with PC2N. Our findings were most consistent with the possibility that the PC2–a‐actinin interaction actually entails at least two discrete domains in the channel protein.
chronically CD‐treated membranes (p < 0.05 in both cases). (E) Addition of prepolymerized actin (2 h in 150‐mM KCl plus 1‐mM MgCl2) to the cis chamber was without eVect on spontaneous ion channel activity (middle tracing). Channel activity was however, readily inhibited by amiloride (100 mM, bottom tracing; n ¼ 5). All‐point histograms to the right of each tracing indicate current amplitude. (F) Average data for conditions in (E). Data reproduced from the Journal of Physiology (Montalbetti et al., 2005b), with permission.
262
A
N
TMs
C
Buffer
GST
PC2CD
PC2CC
A
PC2CB
Binding partners ACTN2 ACTN1 (341−894) (334−892)
PC2CA
968
680 EF-hand ER-retention
PC2C
PC2 224
1
PC2N
B
A
50 215
1
PC2N PC2CA PC2CB PC2CC
821 821
968 927
821 878
++
+
++
+
++
+
26
++
+
B
−
−
114
37
872 927
PC2CD
NM ACTN
81
B
ACTN2 1
252
N (actin-binding) SR I
371
486
SR II
Binding partners 607
720 EF-hands 894
SR III SR IV
C
PC2N (1−215)
C
PC2CA PC2CC (821−968) (821−878)
114 81
341
894 423
894 596 596
++
++
++
+
−
−
+++
−
−
+
−
−
894 728
M ACTN
C
D B
A
C
PC2
100
75 BM752
+
−
+
−
+
E
D
− F
NM ACTN
100
75 IAII
+
−
HEK293
+
−
MDCK
+
−
Rat kidney
263
FIGURE 4 Evidence for the interaction between PC2 and a‐actinin. Schematics of segments of PC2 and a‐actinin (ACTN2) used to reveal interaction by yeast two‐hybrid system. Solid bars indicate the interacting candidate (bait) constructs in the initial library screening. (A) Human PC2 segments with marked starting and ending amino acid residues and their association with human a‐actinin‐1 (ACTN1) and a‐actinin‐2 (ACTN2) indicated by ‘‘þþþ,’’ ‘‘þþ,’’ ‘‘þ,’’ and ‘‘ ’’ for development of blue color within 1, 3, and 24 h, and no development of blue color within 24 h, respectively. (B) ACTN2 segments and their association with PC2N, PC2CA, and PC2CC. (B) Interaction between PC2 segments and a‐actinins as shown by the GST pull‐down assay. (A) E. coli extracts expressing GST‐tagged PC2 polypeptides PC2N, PC2C, PC2CA, PC2CB, PC2CC, PC2CD, or GST alone were visualized by the GST antibody. (B and C) Fusion proteins were incubated with purified nonmuscle a‐actinin (NM‐ACTN) from chicken gizzard (B) and muscle a‐actinin protein (M‐ACTN) from rabbit skeletal muscle (C). Glutathione–agarose beads were used to precipitate GST epitope‐binding proteins. The resultant protein samples were immunoblotted with a‐actinin antibodies BM75.2 (nonmuscle) or EA53 (muscle). Molecular mass markers (in kDa) are shown. (C) Interaction between endogenous PC2 and a‐actinin in cultured cells, and the rat kidney. Total cell lysate from HEK293 (A) and MDCK cells (B) and total protein from rat kidney (C) were precipitated with either nonmuscle a‐actinin antibody (BM75.2) or nonimmune mouse IgG. Precipitates were detected with anti‐PC2 antibody (1A11). In reciprocal co‐IP experiments, cell lysates from HEK293 (D) and MDCK cells (E) and total protein from rat kidney (F) were precipitated with 1A11 or nonimmune mouse IgG. The precipitates were detected with BM75.2. (D) Cellular colocalization of PC2 (green, 1A11 antibody) and nonmuscle a‐actinin (red, BM75.2 antibody) in subconfluent MDCK and IMCD cells. Triangles and arrows indicate the plasma membrane and cell–cell junction localization, respectively. Data reproduced from Human Molecular Genetics (Li et al., 2005), with permission.
264
Cantiello et al.
B. In Vitro and In Vivo Binding of PC2 with a‐Actinins To confirm the interaction between PC2 and a‐actinins, we used a glutathione S‐transferase (GST) fusion protein aYnity‐binding method (Fig. 4B). Polypeptides, including PC2N, PC2C, PC2CA, PC2CB (aa 821–927), PC2CC, and PC2CD (aa 872–927) were fused in frame with a GST epitope and expressed in bacteria (BL21; Li et al., 2005). The PC2 fusion peptides were incubated with purified a‐actinin. Using monoclonal a‐actinin antibodies, we observed that GST‐PC2N, ‐PC2C, ‐PC2CA, ‐PC2CB, and ‐PC2CC, but not GST‐PC2CD or controls (GST alone and buVer without fusion protein lysates), coprecipitated with both nonmuscle and muscle a‐actinin (Fig. 4B). These results confirmed that both the PC2 amino terminus and the small C‐terminal segment 821–878 of PC2 interact with a‐actinin. GST‐ PC2C showed a clear binding with a‐actinins. However, less amount of GST‐PC2CA still pulled down slightly more a‐actinin, compared with GST‐PC2C. This finding is in agreement with the observation from our yeast two‐hybrid assay that the PC2C segment (aa 682–820) exhibits an inhibitory eVect on the PC2C–a‐actinin interaction. To determine whether PC2 also interacts with a‐actinin in vivo, we coimmunoprecipitated (co‐IP, Fig. 4C) both proteins from MDCK cells, and rat kidney (other cell lines and tissues were also tested; Li et al., 2005). Using antibodies against PC2 (1A11), muscle a‐actinin (EA53), and nonmuscle a‐actinin (BM75.2), we detected the associated proteins by immunoblotting. PC2 was detected in the immunoprecipitate from rat kidney tissue using muscle a‐actinin antibody EA53. Reciprocal co‐IP using a PC2‐specific antibody (1A11) also precipitated muscle a‐actinin. These data confirmed that PC2 interacts with muscle a‐actinin (Fig. 4C) in vivo. Under the same conditions PC2 was immunoprecipitated from MDCK cells using BM75.2 (Fig. 4C). A reciprocal signal corresponding to nonmuscle a‐actinin from these cell or tissue lysates was observed in the immunoprecipitate of the same cell/tissue lysates using 1A11. Taken together, these results demonstrate that endogenous PC2 and a‐actinins form a complex in vivo in cells and tissues. The colocalization of endogenous PC2 and a‐actinin was further explored by immunofluorescence in epithelial MDCK cells (Fig. 4D). We found that nonmuscle a‐actinin exhibited clear cell surface/periphery localization, notably at cell‐ cell contacts. Faint staining was also observed along stress fibers and in the perinuclear region. Consistent with recent reports (ScheVers et al., 2002, 2004; Luo et al., 2003), we also found that subconfluent MDCK cells express endogenous PC2 in both the cytoplasm and the plasma membrane and at the cell–cell contacts (Fig. 4D). The fluorescence patterns of a‐actinin and PC2 substantially overlapped both in the plasma membrane and at cell–cell junctions, indicating that the two proteins colocalize in these cells.
10. The Cytoskeletal Connection to Ion Channels
265
C. Functional Modulation of PC2 by a‐Actinin a‐Actinin has been shown to regulate a number of ion channels, including Ca2þ and Kþ channels and NMDA receptors (Krupp et al., 1999; Maruoka et al., 2000; Sadeghi et al., 2002). We thus examined whether a‐actinin also modulates PC2 channel function in hST apical membrane vesicles (Fig. 5). Substantial colocalization of PC2 with a‐actinin was observed in hST apical membrane vesicles and partially altered by the actin filament‐disrupting agent cytochalasin D (10 mg/ml; Fig. 5C). This is in agreement with the overlapping distribution of PC2 and actin filaments in hST where changes in actin filament organization modifies PC2 channel activity (Montalbetti et al., 2005b). After hST vesicles were reconstituted in the lipid bilayer system, addition of a‐actinin to the cis chamber substantially increased PC2 cation channel activity (Fig. 5A and B). To further assess whether a‐actinin regulates PC2 channels by direct interaction with the channel protein, we replicated the experiment with purified PC2 (Li et al., 2005). We found that addition of nonmuscle a‐actinin to the cis chamber (cytoplasm) elicited a substantial increase in PC2 single‐channel activity (Fig. 6A). In average, the mean single‐channel currents increased 15‐fold by a‐actinin. The currents averaged 1.7 0.5 pA, n ¼ 7 for controls, and 27.0 5.4 pA, n ¼ 6, in the presence of a‐actinin (p < 0.01). However, a-actinin did not significantly alter the current-voltage relationship of the main (largest) conductance of PC2 (Fig. 6A, right), confirming that the stimulatory eVect of a‐actinin on PC2 is indeed mediated by the control of its open probability. Interestingly, addition of either monomeric, or F‐actin, was without eVect on the isolated channel protein (Fig. 6B), strengthening the contention that anchoring proteins are a requirement for the channel to connect to the actin cytoskeleton. This also poses the interesting possibility that other actin‐binding proteins already known to interact with PC2, may exert novel regulatory eVects, based on competition with a‐actinin, and/or direct binding to PC2 (Li et al., 2003a,b; Section IV.A–B).
V. EFFECT OF HYDROOSMOTIC PRESSURE ON PC2 CHANNEL FUNCTION: ROLE OF THE CYTOSKELETON IN OSMOSENSORY FUNCTION A. Effect of Hydrostatic and Osmotic Pressure on PC2 Channel Regulation Studies have raised the interesting possibility of a possible role of PC2 in mechanotransduction of environmental signals in renal epithelial cells. PC2 has been localized to the primary cilium of renal cells (Pazour et al., 2002;
266
Cantiello et al. +a -Actinin
A
10 pA 20 s
Normalized lmean
0.006 +a -Actinin
1.0 0.8 0.6 0.4 0.2 0.0
Npo
0.004 0.002
−100
−50
0
50 Time (s)
100
150
0.000 Control
a -actinin
6000 N
B Control
4000 2000 0
5 pA
4 8 I (pA)
12
0
4 8 I (pA)
12
500 ms 6000 N
+a -Actinin
0
4000 2000 0
C
−CD
+CD
FIGURE 5 Colocalization of, and regulation by, a‐actinin with PC2 in hST vesicles. (A) Upper left: representative tracings at 40 mV before and after addition of a‐actinin (as indicated). Lower left: average data are indicated as normalized mean currents (n ¼ 6).
10. The Cytoskeletal Connection to Ion Channels
267
Yoder et al., 2002), which is thought to be a sensory organelle (Praetorius and Spring, 2001; Pazour and Witman, 2003). We further observed that PC2 is a functional channel in primary cilia (Raychowdhury et al., 2005). The nature of a possible mechanotransduction in PC2 is, however, still unknown. To test a potential role of PC2 in mechanotransduction, we reconstituted hST apical membranes (Montalbetti et al., 2005a). For these studies, we relied on changes in two ‘‘environmental’’ forces, which may convey physical changes to the preparation, namely, changes in osmotic and/or hydrostatic pressure. The experimental setup we used to test whether hydroosmotic pressure might control PC2 function is shown in Fig. 7A, inset. Basal PC2 channel activity was observed in the presence of an osmotic gradient (), initially imposed in a KCl channel gradient (150‐ and 15‐mM KCl, cis and trans, respectively, Fig. 7A) as originally reported (Gonza´lez‐Perrett et al., 2001). We tested both, changes in osmotic pressure elicited by either compensation of the by addition of salt to, and/or imposition of a H by addition of volume to the trans compartment (or reduction in cis volume or both). Channel activity was first observed in the absence of (150‐mM KCl in cis, and 150‐mM NaCl in trans compartments, respectively, Fig. 7A). A change in hydrostatic pressure (H), namely a decrease in volume in the cis chamber, accompanied by an augment in the trans solution, increased PC2 channel activity after rundown. This finding would indicate that changes in the physical parameters of the reconstituted membrane convey a regulatory mechanism to PC2 channel function. To further assess this phenomenon, the osmotic contribution of this activating eVect was also assessed. We first determined PC2 channel function in the presence of a KCl chemical gradient (Fig. 7B). Under these conditions, an ‘‘outward’’ (in‐to‐out) osmotic gradient was imposed. This basal was then eliminated by addition of NaCl (150 mM, Fig. 7B) to the trans compartment. Channel activity immediately (<5 s) changed after addition of salt, as expected by the elimination of . This is shown as a decrease in both single‐channel conductance (Fig. 7B, center, bottom), and a change in the channel’s open probability (po, n ¼ 5, Fig. 7B, bottom right). Interestingly,
Right: average channel activity is shown as the number of channels times the channel’s open probability (Npo) before and after application of a‐actinin (n ¼ 6). (B) Representative single‐ channel tracings (left panels) at 40 mV of reconstituted hST apical membranes in asymmetrical KCl solution. The corresponding all‐point histograms are shown on the right. (C) Distribution of endogenous a‐actinin and PC2 was also determined in hST apical membrane vesicles. The two proteins substantially colocalized with each other (PC2, green; a‐actinin, red). Treatment with cytochalasin D (10 mg/ml for 24 h) showed sporadic detachment of this interaction (merge, yellow). Horizontal bars ¼ 20 mm. Data reproduced from Human Molecular Genetics (Li et al., 2005), with permission.
268
Cantiello et al.
A
Control
15
2 pA
6000
0.5 s
I (pA)
N
4000 2000
10
I (pA)
0 0
2
4
6
8 10 12 14
+a -Actinin
N
5
4 pA
12,000 10,000 8000 6000 4000 2000 0
−80 −60 −40 −20
0.5 s
0
20
40
60
80
V h (mV)
I (pA) 0
2
4
6
8
10 12 14
B
0.6 60 mV 1000 N/Div
Control
Npo
0.4
−5
0
5 10 15 Amplitude (pA)
20
0
20
0.2
1000 N/Div
After G-actin
10 pA
−5 500 ms
0.0 5 10 15 Amplitude (pA)
Control
+ G-actin
FIGURE 6 EVect of a‐actinin and actin on the isolated PC2 channel. (A) Top left: representative single‐channel tracings and corresponding all‐point histograms (below each tracing) of in vitro translated PC2, reconstituted in a lipid bilayer system. The data indicate the increase in single‐channel conductance after addition of nonmuscle a‐actinin to the cis chamber. Data were obtained at 40 mV. Right: current–voltage relationship is shown of the a‐actinin‐activated channel. Dashed and solid lines correspond to the control condition (Gonza´lez‐Perrett et al., 2001) and in the presence of a‐actinin (n ¼ 5), respectively. (B) EVect of actin on the purified PC2 channel. Left: monomeric actin (1 mg/ml) was added to the cis compartment, with no eVect on PC2 channel. All‐point histograms of current amplitudes are shown on the right. Right: average data (n ¼ 5) for control and þactin conditions. Data were not statistically diVerent, p < 0.4. Data reproduced from Human Molecular Genetics (Li et al., 2005), with permission.
269
10. The Cytoskeletal Connection to Ion Channels A
∆H
R1 ∆H
Cis Trans 150−mM KCI = 150−mM NaCI
10 mV
B
10 pA
NaCI
10 s
10 pA 10 s
14
0.8
Open probability
Current (pA)
12 10 8 6
0.6
0.4
0.2
0.0
4 0
100
200 Time (s)
300
400
0
100
200
300
400
Time (s)
FIGURE 7 EVect of hydroosmotic pressure on hST PC2 channel activity. (A) PC2 channel activity was observed under conditions resembling a cell membrane in situ, namely the presence of outward KCl, and inward NaCl chemical gradients (isoosmotic condition). After spontaneous channel rundown, the membrane was exposed to a H by addition of saline solution to the trans compartment, which increased the membrane currents, with an onset time varying among the various preparations (n ¼ 7). Inset: experimental setup to assess the role of the physical membrane parameters on PC2 channel function. Basal channel activity was originally observed
270
Cantiello et al.
the decay kinetics of po increase (Fig. 7B, bottom right) followed a time constant several seconds long, which is inconsistent with the establishment of . A similar phenomenon was observed by addition of KCl (data not shown), suggesting an osmotic response (Montalbetti et al., 2005a). To further assess the nature of this activation process, osmotically uncompensated ( > 0) membranes were subjected to a hydrostatic gradient (H), in the presence of a KCl chemical gradient (150 and 15 mM, cis and trans, respectively; Fig. 8A). Cation channel activity was first allowed to spontaneously inactivate. We then imposed a H, which induced channel reactivation, which was delayed in average by 9 s (Fig. 8A, right). The delay activation is again inconsistent with a ‘‘compliance’’ eVect on membrane elasticity induced by H, which should be immediate. This stimulatory effect was thought to implicate instead changes in the cytoskeletal structures present in the vicinity of the channels. To test the hypothesis that a PC2– cytoskeletal interaction is actually required for the hydroosmotic activation of the channel, we first reconstituted the purified channel protein instead (Fig. 8B). PC2, exposed to H (in the presence of a KCl chemical gradient), was unable to respond with an increased channel activity. This is consistent with the finding that addition of actin alone was unable to activate the isolated PC2 channel, further suggesting the requirement of actin‐associated proteins such as a‐actinin (Section IV, above). To confirm this hypothesis, hST vesicles were first treated with the actin cytoskeleton disrupter cytochalasin D (10 mg/ml) for 24 h to collapse the intravesicular cytoskeleton prior to reconstitution. CD‐treated apical hST vesicles displayed low channel activity (Fig. 8C). Establishment of H was without a stimulatory eVect, such that channel activity disappeared even in the presence of a hydroosmotic gradient. Thus, the stimulatory eVect
in osmotically challenged () membranes, which were then subjected to a compensatory change in osmotic pressure ( ¼ 0), by addition of either KCl or NaCl, to the trans compartment. Conversely, a hydrostatic gradient (H) was imposed under and ¼ 0 conditions, by means of either a decrease in cis volume, and/or in addition of volume to the trans compartment. Thus, changes in were either used instead, or in addition to H, to assess hydroosmotic changes in PC2 channel activity. (B) EVect of osmotic compensation on hST PC2 cation channel activity. A reduction of the imposed osmotic gradient (150‐mM vs 15‐mM KCl in cis and trans compartments, respectively) by addition of NaCl (150 mM) to the trans compartment, increased PC2 channel activity. Top: both the single‐channel conductance and the open probability of the channel were modified by the osmotic compensation (addition of trans NaCl, arrow). Expanded tracings below indicate each region of the top tracing expanded under a horizontal bar. Bottom: kinetic changes in single‐channel currents and open probability associated with the compensatory osmotic decrease. While the single‐channel currents remained lower than their respective controls after elimination of the chemical gradient, (left) the open probability rapidly increased, followed by a slower decrease to control levels (right, n ¼ 3). Data reproduced from Pflu¨gers Archiv (Montalbetti et al., 2005a), with permission.
271
10. The Cytoskeletal Connection to Ion Channels A
∆H
0 mV
1.5
10
20 mV
1.0
20 pA 15 s
After ∆H
Before ∆H
Mean current (pA)
0.5 0.0 0
1 2 Time (s)
3
5
0 0
3
6 9 Time (s)
12
B −40 mV 40 mV
40 mV
∆H
3 pA 10 s 4.0
*
C Npo
3.0
2.0
+CD, 24 h
∆H 60 mV
1.0
0.0
* Control
∆H
∆H + CD
FIGURE 8 EVect of hydrostatic pressure on PC2 cation channel activity in hST. (A) Left: PC2 channel activity was observed under basal ( > 0) conditions in the presence of an outward KCl chemical gradient. Asterisks indicate spontaneous channel activity. After spontaneous channel rundown, the membrane was exposed to H by addition of saline solution to the trans chamber, which increased the membrane currents, with a lag time. The upward arrow indicates breaking of the membrane (n ¼ 3). Right: average currents from three diVerent experiments are shown. The kinetics of channel activation was fitted with a single exponential in two, and a straight line in one experiment. Inset shows activation kinetics after addition of NaCl to the trans compartment, for comparison. (B) The eVect of hydroosmotic pressure was also determined in the reconstituted purified PC2 channel. Channel activity was first observed in the presence of a KCl chemical ( > 0) gradient. Channel activity was voltage‐inactivated as
272
Cantiello et al.
of H observed under control conditions requires an organized actin cytoskeleton to respond. The combined studies indicate that the PC2‐cytoskeleton interface is a critical part of the transduction mechanism for channel regulation.
VI. THE CHANNEL–CYTOSKELETON INTERFACE: STRUCTURAL–FUNCTIONAL CORRELATES A. Mechanosensitivity and the Lipid Bilayer Mechanosensitivity of ion channels has been associated with two main likely transducers of mechanical force. These are either changes in the properties of the membrane lipid bilayer, or the more complicated, and thus less described in detail, interactions with channel‐associated proteins, such as the cytoskeletal components (Sackin, 1994). It is thus not surprising that a clearer molecular picture has only been developed for the specific components of the lipid bilayer, forwarding the well‐established description of channel regulation by changes in the physical properties of the membrane (Hamill and McBride, 1993; Hamill and Martinac, 2001; Markin and Sachs, 2004). Several types of deformation have been postulated for mechanosensitive channels to undergo transitions from closed to open as transduced directly by restructuring of membrane lipids (Markin and Sachs, 2004). A channel can change its in‐plane area, such that an increase in the in‐plane area induces stretch activation, while the opposite elicits stretch inactivation (Morris and Sigurdson, 1989). Another type of mechanosensitivity occurs if the channel changes its shape, transducing tension to the bilayer, which combined, can result in a torque, manifested as a tendency of the membrane to bend, and change its shape (Volkov et al., 1998). The channel’s open probability would then be sensitive to this torque (Petrov and Usherwood, 1994; Markin and Sachs, 2004). Yet another type of deformation may entail changes in length of the channel complex without a change in shape, such that as a result, a hydrophobic mismatch between the channel and the surrounding lipid bilayer occurs. A stretched bilayer would decrease its
reported (Gonza´lez‐Perrett et al., 2002), by switching the holding potential to 40 mV. Further exposure of the membrane to H did not comparatively further increased PC2 channel activity (n ¼ 3). (C) Left: PC2 channel activity from CD‐treated (>24 h) hST membranes was observed in the presence of a KCl chemical ( > 0) gradient. Imposition of H had no eVect on the membrane currents. Right: mean data SEM for the control and conditions, and for the CD‐ treated membranes (n ¼ 7). Data reproduced from Pflu¨gers Archiv. (Montalbetti et al., 2005a), with permission.
10. The Cytoskeletal Connection to Ion Channels
273
thickness, changing the lipid–channel interaction, which in turn would transfer this mismatch to the energetic equilibrium between open and closed states of the channel under tension. In the context of TRP channel function, however, which form mechanosensitive transducers, and have potential links to the cytoskeleton, the possibility exists for more than one form of mechanotransduction to take place. Maroto et al. (2005) showed that TRPC1 is a component of the amphibian mechanosensitive channel. In those studies it was observed that changes consistent with membrane stretch, but not cytoskeletal rearrangement might be responsible for this response. Future experiments may be required to assess the functional role of the N‐terminal ankyrin domains in this response. A similar question applies to TRPA1 also implicated in mechanosensory function (Corey et al., 2004), which contains an even longer ankyrin structure, consistent with a tethering mechanism of gating and regulation. B. Cytoskeletal Interactions with PC2 Our studies demonstrate that changes in the actin cytoskeleton play an important role in conveying regulatory properties to PC2 (TRPP2). These studies clearly define specific molecular interactions between the cytoskeletal components and the ion channel, such that PC2 channel function is largely mediated by the state of the actin cytoskeleton. Most of our studies were conducted in hST apical membranes, where endogenous PC2 is abundantly expressed and functional (Gonza´lez‐Perrett et al., 2001). We originally observed that early disruption of the attached cytoskeleton by CD activates PC2 channel function in the hST. We expanded this evidence to place it in the context of endogenous regulation by physiological components of the actin cytoskeleton. The CD eVect on PC2 was mimicked by the actin‐ severing protein gelsolin, in the presence, but not the absence of Ca2þ. Thus, the stimulatory eVect of the Ca2þ–gelsolin complex but not gelsolin alone indicates that cleavage of endogenous actin filaments by Ca2þ–activated gelsolin is the triggering mechanism, in agreement with the acute eVect of CD. Further, we also found that the activation of gelsolin can also be elicited by trans Ca2þ, suggesting that Ca2þ transport through the channel is itself a feedback mechanism, involving channel‐associated proteins. Interestingly, we found that addition of monomeric actin to control membranes had a largely inhibitory eVect under control conditions. This was surprising in view that actin rapidly polymerizes under our experimental conditions. However, we observed in contrast a stimulatory eVect after complete collapse of the endogenous cytoskeleton induced by a chronic CD incubation period (>24 h). This suggested to us a PC2 regulatory mechanism by which competition of
274
Cantiello et al.
actin monomers with either the channels themselves or tightly associated proteins, allow binding and likely changes in the actin conformations not observed for the monomer itself. The most likely scenario is that actin monomers compete with PC2‐associated prepolymerized actin filaments. Regulation of cation channels in hST by gelsolin and Ca2þ further suggests an involvement of organized actin (F‐actin) in the vicinity of the channel responsible for the currents reported in the present study.
C. In Search of the Molecular Link The above experiments provided a clear picture that a dynamic remodeling of cortical cytoskeleton in proximity to the membrane aVects and indeed modulate the channel properties of PC2. Insofar as this is concerned it was also clear that despite this phenomenon had been evidenced in numerous channels with diVerent outcomes, actin binding to the channel protein was not in itself the molecular mechanism of regulation. Interestingly, PC2 does not display ankyrin repeats or any obvious consensus domains in the protein that would allow tight interaction between the channel complex and the cytoskeleton. However, studies have suggested a potential mechanosensory for the PC1–PC2 complexes, which are expressed in primary cilia of cultured renal epithelial cells (Pazour et al., 2002; Yoder et al., 2002), might function in transducing environmental information (Nauli et al., 2003). Interestingly, proteins necessary for the assembly or function of primary cilia including cystin, polaris, inversin, and kinesin II also cause polycystic kidney diseases (Ong and Wheatley, 2003). Primary cilia are microtubular organelles, which seem to exclude structural actin. Recent studies do indicate, however, that PC2 also interacts with elements of the actin cytoskeleton. Hax‐1, a cytoskeletal protein that interacts with the F‐actin‐binding protein cortactin, was observed to interact with PC2 (Gallagher et al., 2000). Moreover, Chen and collaborators found that two cytoskeletal proteins, troponin‐I (Li et al., 2003b) and tropomyosin‐1 (Li et al., 2003a) directly bind to PC2, further strengthening a link between cytoskeletal dynamics and the PC2 channel. a‐Actinin, is a widely distributed actin‐bundling protein, which is prominently located in cell–cell and cell–matrix adhesion complexes, which associate with integrin receptors at cell–matrix focal contacts and the cell‐cell adhesion belt (Otey et al., 1993; Nieset et al., 1997). Alpha‐actinin is present in a number of renal cell types, including epithelial and blood vessel cells. Evidence for the role of a‐actinin in renal disease includes the finding that experimental nephritic syndrome can be induced by upregulation of a‐actinin (Kaplan and Pollak, 2001), and that mutations in a‐actinin‐4 cause familial focal segmental glomerulonephritis (Kaplan et al., 2000). a‐Actinin has also
10. The Cytoskeletal Connection to Ion Channels
275
been reported to exhibit tumor suppresser activity. Nonmuscle a‐actinin overexpressing mouse NIH 3T3 fibroblasts, for example, display a significant reduction in cell motility, while cells with reduced expression of a‐actinin have an increased cell motility (Gluck and Ben Ze’ev, 1994). A number of reports have established that a‐actinin also regulates the activity of a number of channels. The muscle actinin isoform, a‐actinin‐2, for example, directly binds and modulates channel gating and current density of the voltage‐gated Kþ channel Kv1.5 (Maruoka et al., 2000). Alpha‐ actinin‐2 was also found to bind the NR1 and NR2B subunits of the NMDA‐type glutamate receptor and regulate its channel function (Shieh and Zhu, 1996; Rycroft and Gibb, 2004). It was revealed that muscle type a‐actinin regulates L‐type Ca2þ channel function (Sadeghi et al., 2002). Our present data demonstrate that a‐actinin binds directly to, and connects PC2 to the actin cytoskeleton (Li et al., 2005). These data demonstrate that PC2 physically and functionally interacts with a‐actinins. This interaction was documented by a variety of methods, including the yeast two‐hybrid system, that helped identify it, and in vitro biochemical assays, immunofluorescence, and coimmunoprecipitation in cultured renal cells and tissues, which confirmed its widespread distribution and physiological relevance of this interaction. Further, a functional interaction exists between a‐actinin and PC2, whereby its channel activity can be substantially increased in the presence of this actin‐bundling protein but the absence of any other cytoskeletal elements. It is thus likely that we have determined one of the key elements that anchor PC2 to the actin cytoskeleton. It is, however, interesting to note that a‐actinin is also an amphipathic protein (Meyer and Aebi, 1989), which has been reported to bind specific phospholipids (Greenwood et al., 2000; Fraley et al., 2003; Corgan et al., 2004). This association decreases a‐actinin bundling activity through competitive block of the interaction between its actin‐binding domain (ABD) and the actin filament. In the presence of lipids the binding aYnity between a‐actinin and F‐actin changes (Meyer and Aebi, 1989), such that cytoskeletal proteins interaction to PC2 may also be regulated by lipid bilayer components, which likely aVect this interaction.
D. Elastic Properties of Actin Networks F‐actin networks are a major constituent of the cellular cytoskeleton, which determines, to a large degree, the mechanical properties of cells (Janmey et al., 1990a; Xu et al., 2000). However, the rheological properties of in vitro F‐actin networks are quite diVerent from those of cells, often by several orders of magnitude (Gardel et al., 2004a,b). This stresses the
276
Cantiello et al. A P04
Fimbrin/plastin
a-Actinin
ABD ABD
ABD
Actin-binding domain
ABD ABD P04
ABD
Spectrin/fodrin
ABP-120
ABD
ABD P04
ABD
P04
Filamin/ABP-280 ABD
20 nm B Human T-plastin Chicken fimbrin Chicken a-actinin Human fimbrin Human dytrophin Filamin/ABP280 C
(208) (210) (107) (477) (95) (121)
VVNIGAEDLRAGKPHLVLGLLWQIIKI VVNIGSQDLQEGKPHLVLGLLWQIIKV LVSIGAEEIVDGNVKMTLGMIWTIILR LVGIAGHDLNEGNPTLTLALIWQLMRR LVNIGSTDIVDGNHKLTLGLIWNIILH LVSIDSKAIVDGNLKLILGLIWTIILH
Filamin
15 pA 10 s
(234) (236) (133) (503) (121) (147)
10. The Cytoskeletal Connection to Ion Channels
277
importance of actin‐binding proteins and their regulatory mechanisms in the mechanical stability of cells. Actin assembly in the presence of actin cross‐ linking proteins such as filamin, for example, acquires mechanical properties resembling those of a gel (Brotschi et al., 1978; Stossel, 1982), the rigidity of which is proportional to the concentration of actin‐binding protein added (Hartwig and Stossel, 1981; Stossel, 1982). Thus, cross‐linked F‐actin is more resistant to deformation and water flow in response to stress than an equivalent concentration of pure actin (Ito et al., 1987). During the last 30 years at least 30 proteins have been identified (Kreis and Vale, 1999), which either cross‐link or bundle F‐actin in various cell domains. Several proteins that cross‐link actin filaments also promote bundling of actin filaments (Shizuta et al., 1976; Bretscher and Weber, 1980; Bretscher, 1981; Blanchard et al., 1989). It is expected that this abundance in likely similar protein function generate a redundancy and complementation in cellular responses. Fascin, a‐actinin, and the filamins, may all have overlapping roles, and complementary functions (Tseng et al., 2001), enabling wider variety to the mechanical responses of cellular actin networks. The perpendicular branching of actin by filamins, for example, increases the isotropy of F‐actin by preventing bundle formation (Hartwig and Stossel, 1981). This suggests that, in vivo, F‐actin may be at any given time, cross‐linked by a variable number of actin‐binding proteins, which are essential in determining then specific elastic properties of the actin cytoskeleton. It is interesting to note, however, that little is known about the consequences of the role of the dynamic changes in cytoskeletal structures as they interact with, and potentially regulate ion channels. Even when a number of actin‐binding proteins may share similar ABD, their expected role in cytoskeletal organization, and thus cell function may be quite distinct and varied (Weeds, 1982; Matsudaira, 1991; Weeds and Maciver, 1993; Otto, 1994). The consensus sequence of the ABD of a number of ABPs is shared for both actin bundling‐, and actin cross‐linking‐proteins (Matsudaira, 1991). This generic domain, contains a highly conserved actin‐binding tandem repeat (Matsudaira, 1991) (Fig. 9A–B). Thus, the ultimate role of ABPs in cell FIGURE 9 Features of actin‐binding proteins (ABPs), and eVect of filamin on PC2 from hST. (A) The consensus sequence of the ABD of a number of ABPs is shared for both actin bundling‐ and actin cross‐linking proteins. This generic domain is present as a tandem repeat, of which the most conserved region in repeat A is shown in (B) (Matsudaira, 1991). The ultimate role of ABPs in channel regulation, thus largely depend in the ABP’s topology and ability to interact with more than one actin filament. Thus, a filamin (ABP‐280) homodimer, would convey a three‐dimensional structure to the cross‐linked actin network, most consistent with a gel, which is largely diVerent from the expected bundling role of a‐actinin (Matsudaira, 1991). (C) Consistent with the importance of the three‐dimensional structure of the actin gel in PC2 channel regulation, filamin had a strong inhibitory eVect, which is exactly opposite to the eVect observed for a‐actinin (Fig. 6). Interestingly, a similar finding has been observed for epithelial Naþ channels (Cantiello, 1995).
278
Cantiello et al. A
B
Ca2+
Ca2+
K+
K+
Out Actin monomers (G-actin)
Out
Long actin filaments (F-actin)
In (−)
a-Actinin
Cytochalasin D gelsolin (+)
Actin In monomers (G-actin)
K+
Out a-Actinin Long actin filaments Actin In (F-actin) monomers (G-actin) (+)
2+ Short actin Ca filaments
Ca2+
Ca2+ Ca2+
Ca2+ Out
K+
a-Actinin
In (+) Cytochalasin D chronic treatment >24 h
(+)
Short actin filaments Ca2+
D ∆H
Long actin filaments (F-actin)
(+)
Ca2+
K+
(−)
a-Actinin Ca2+
Cytochalasin D gelsolin
Ca2+ Out
Gelsolin
In
Gelsolin
Ca2+
C
Actin monomers (G-actin)
K+
Out
Out
In
In
FIGURE 10 Cytoskeletal regulation and osmosensory control of PC2. The encompassed evidence, mostly from endogenous hST‐PC2 and the purified channel protein, allows a comprehensive understanding of the role the dynamic changes in the actin cytoskeleton convey on PC2 channel function. (A) Under basal conditions, a‐actinin (and/or other actin‐binding proteins) link the PC2 channel, to an endogenous cross‐linked network of actin filaments. (B) Endogenous F‐actin depolymerization by actin‐severing proteins, such as gelsolin (likewise by toxins such as CD) activates PC2 channel function, which in turn is modulated the Ca2þ transport through the channel, which elicits a feedback mechanism mediated by the cytoskeleton. (C) This is in agreement with the inhibitory eVect of monomeric actin on the cytoskeletally associated channel and the lack of a direct functional eVect of actin on the isolated channel. In contrast, chronically CD treated membranes (low channel activity) to completely collapse the cytoskeleton can be functionally restored to almost control levels of channel activity by addition of G‐actin, which likely replenishes the exhausted pool of endogenous actin. Conversely, the molecular anchor between actin networks and PC2 is clearly demonstrated to be a‐actinins, which in turn directly modulate, and may be competed by other actin‐associated proteins with similar binding domains. (D) The stimulatory eVect of and H in cytoskeletally competent
10. The Cytoskeletal Connection to Ion Channels
279
function, largely depends in the ABP’s topology and ability to interact with more than one actin filament. As indicated above, a filamin (ABP–280) homodimer, would convey a three‐dimensional structure to the cross‐linked actin network, most consistent with a gel, which is largely diVerent from the expected bundling role of a‐actinin (Matsudaira, 1991). The fact that the three‐ dimensional structure of the filamin‐cross‐linked actin gel would be largely diVerent from that observed with a‐actinin, should also be reflected in distinct functional interactions with PC2. We found that addition of filamin (ABP– 280) has a strong inhibitory eVect on PC2 channel regulation, exactly opposite to the eVect observed for a‐actinin (Fig. 9). Interestingly, a similar finding has been observed for epithelial Naþ channels (Cantiello, 1995). Thus, our data are consistent with the hypothesis that the three‐dimensional structure of the actin gel is a key element in controlling channel regulation (Fig. 10). A membrane‐cytoskeleton functional interface may be critical for the cation‐ dependent signaling pathway(s) normally associated with various cell functions, including cell cycle, vesicle traYcking, and ion transport. As an example, Ca2þ signals which play essential role in cell function, may be under the control of a feedback mechanisms linking Ca2þ transport to the remodeling of the cytoskeleton, and in particular channel function, as described in our studies (Montalbetti et al., 2005b, data herein). A universal cell response known as cell volume regulation is elicited by an osmotic stress. This response, often associated with the activation of ion channels and other solute transporters often entails a combined response associated with both, changed in the structure of the cytoskeleton, and changes in the geometry and properties of the lipid bilayer underlying the plasma membrane. These two widely distributed responses are both linked to the mechanotransduction of physical forces to ensure the phenomenologically defined mechanosensitivity of ion channels which respond to such forces. Placed in the context of the PC2–cytoskeletal interface (Fig. 10), it can be expected that the channel, tethered by the a‐actinin (or another anchoring protein) to the cytoskeleton, will respond and thus elicit nonlinear responses mediated by the actin networks. Accepting the strong linkage and regulatory function of cytoskeletal components on PC2 function, we speculated and herein
membranes is consistent with a dynamic contribution of the PC2‐adjacent actin cytoskeleton, which remains under ‘‘swelling equilibrium.’’ This actin network serves as an interface, which provides a sensory mechanism whereby changes in hydroosmotic pressure modulate this actin– PC2 interaction, likely by the anchoring eVect of a‐actinin, which also control channel function. F‐actin disorganization (osmotic shock, CD treatment) uncouples this physical interaction, rendering channels, first active, and an unresponsive to environmental changes.
280
Cantiello et al.
determined that the PC2‐connected actin network is a key regulatory component of the channel. E. Sensory Role of the Actin Cytoskeleton in PC2 Channel Function Studies have turned attention to the potential sensory transducer properties of PC2 in renal epithelial and other target cells (Nauli et al., 2003; Cantiello et al., 2004; Nauli and Zhou, 2004; Pazour, 2004). A functional PC1–PC2 channel complex has been implicated in the mechanosensory transduction associated with cilia bending and Ca2þ influx in renal epithelial cells (Nauli et al., 2003). It is noteworthy, that ciliary structures are largely microtubular, rather than actin filamental organelles (Pazour and Witman, 2003). Consistent with our contention of the relevance of cytoskeletal structures in PC2 channel regulation, preliminary evidence from our laboratories would suggest that indeed microtubular organization also regulate PC2 in hST apical membranes (Montalbetti et al., 2006). However, little is known as to how the PC1–PC2 channel complex, or more particularly PC2 senses internal and/or external environmental responses, particularly in the plasma membrane. Physical forces such as hydrostatic (H) and osmotic () pressure may play an important role in signal transduction elicited by such ‘‘environmental forces,’’ which control ion transport, and thus hydroelectrolytic homeostasis. The spontaneous cation channel activity observed in hST vesicles is normally elicited in the presence of an osmotic gradient () imposed by a KCl chemical gradient to the plasma membrane (Gonza´lez‐Perrett et al., 2001, 2002). Thus, changes in the properties of the membrane may control PC2 channel activity in hST. Further, the elastic pressure arising from osmotically induced membrane deformation may be compensated by hydrostatic pressure diVerences imposed to the membrane. Both physical changes, osmotically induced changes in cytoskeletal structures in the vicinity of the channels and elastic changes in membrane compliance, can act as regulators of PC2 channel function and thus ion transport. To determine the potential regulatory role of hydroosmotic forces in PC2 function we determined that a compensatory hydrostatic gradient in osmotically challenged hST membranes increases PC2 channel activity. We found (Montalbetti et al., 2005a) that either a decrease in cis volume, and/or increase in volume to the trans side of the reconstitution chamber, was suYcient to stimulate PC2 channel activity in hST membranes. Interestingly, this phenomenon was no longer observed after pretreatment of the hST membranes with the actin filament‐disrupting agent CD. These data strongly suggested to us that a sensory mechanism of endogenous PC2 in human placenta entails a structural‐functional interaction between the channel and adjacent cytoskeletal proteins. In this
10. The Cytoskeletal Connection to Ion Channels
281
hypothetical model of sensory transduction, a ‘‘swelling equilibrium’’ in the actin gel (Hodson and Earlman, 1993) associated with the PC2 channel is established under basal (osmotic) conditions, such that any displacement from such equilibrium entails changes in the cortical actin network of the apical hST membrane (Fig. 10). Similarly, a physical challenge such as a change in H, which modifies the elastic compliance of the membrane, also regulates PC2 function. This phenomenon is mimicked by compensatory osmotic changes, such as > 0 to ¼ 0, whose equivalent would be an isoosmotic cell volume increase associated with cell volume regulation (MacKnight, 1988; Strange, 1994), a phenomenon also requiring cytoskeletal structures (Ziyadeh et al., 1992; Cantiello and Prat, 1996). The requirement of the cytoskeleton in the transmission of force to PC2 was confirmed by our finding that H did not regulate the isolated PC2 channel (Fig. 8). This is further supported by the lack of eVect in the CD chronically treated hST vesicles, the requirement of a‐actinin to link the channel to the cytoskeleton, and the fact that addition of actin alone is also without eVect on the isolated protein. A regulatory pathway of PC2 channel function by the cytoskeleton may be considered a novel sensory mechanism linking physical forces to the swelling equilibrium of the adjacent actin networks, rather than the membrane or channel itself.
VII. PERSPECTIVE AND FUTURE DIRECTIONS The present studies and conclusions tried to provide a comprehensive understanding of the steps involved in the regulation by the cytoskeleton of a prototypical TRP channel, namely polycystin‐2 (TRPP2). Our studies provide a comprehensive analysis of the molecular elements, which underlie the PC2–actin cytoskeleton interface. Although membrane compliance and elastic properties may aVect channel function, as it is postulated for most mechanosensitive ion channels, our studies suggest that PC2 channel regulation is instead controlled by physical changes imposed to the cortical actin cytoskeletal structures that link the ion channel to the plasma membrane. The evidence strongly suggest a mechanical interaction, where environmental forces target a true gating mechanism by linking PC2 to the actin cytoskeleton via a‐actinin. Both, changes in cytoskeletal dynamics and direct binding of actin‐associated proteins convey forces that translate into conformational changes in the channel protein. This has to be viewed in a broader scope, as most ion channels studied to date, are indeed either linked or can be regulated by the various components of the cellular cytoskeleton. We are tempted to postulate, that based on the prevalence of this interaction, most ion channels may display, under specific conditions, some component
282
Cantiello et al.
of mechanosensitivity. This is further strengthened by the prevalence of putative cytoskeletal‐binding domains in the various TRP channels, whose multimeric structure, makes them ideal structures to be modulated and controlled by cytoskeletal interacting proteins. Among these, a‐actinins, are rather interesting because not only are known actin‐bundling proteins, but also amphipathic proteins that directly interact with lipid components of the plasma membrane. In this regard, two points are noteworthy. First, even in conditions under which the cytoskeleton is apparently disrupted, actin‐ associated proteins may elicit a response, which may be misconstrued as cytoskeleton independent. Genuine eVorts in addressing whether ion channels remain ‘‘mechanosensitive’’ can be unwittingly biased toward the remaining presence of actin‐binding proteins. In the context of heterologous expression of TRP channel isotypes, for example, the possibility exists, for heterocomplexes to convey unexpected regulatory properties not observed in the native protein. Thus, a careful analysis of the isolated channel protein (not without experimental problems itself) may help provide a more complete picture of this pervading and likely general mechanism of channel regulation. References Ahmed, N., Ramjeesingh, M., Wong, S., Varga, A., Garami, E., and Bear, C. E. (2000). Chloride channel activity of ClC‐2 is modified by the actin cytoskeleton. Biochem. J. 352, 789–794. Alessandri‐Haber, N., Yeh, J. J., Boyd, A. E., Parada, C. A., Chen, X., Reichling, D. B., and Levine, J. D. (2003). Hypotonicity induces TRPV4‐mediated nociception in rat. Neuron 39, 497–511. Arniges, M., Fernandez‐Fernandez, J. M., Albrecht, N., Schaefer, M., and Valverde, M. A. (2006). Human TRPV4 channel splice variants revealed a key role of ankyrin domains in multimerization and traYcking. J. Biol. Chem. 281, 1580–1586. Atlas, S., and Lin, S. (1978). Dihydrocytochalasin‐B‐biological eVects and binding to 3T3 cells. J. Cell Biol. 73, 360–370. Awayda, M. S., and Subramanyan, M. (1998). Regulation of the epithelial Naþ channel by membrane tension. J. Gen. Physiol. 112, 97–111. Awayda, M. S., Ismailov, I., Berdiev, B., and Benos, D. (1995). A cloned renal epithelial Naþ‐ channel protein displays stretch activation in planar lipid bilayers. Am. J. Physiol. 268, C1450–C1459. Basavappa, S., Pedersen, S. F., Jorgensen, N. K., Ellory, J. C., and HoVmann, E. K. (1998). Swelling‐induced arachidonic acid release via the 85‐kDa cPLA2 in human neuroblastoma cells. J. Neurophysiol. 79, 1441–1449. Benham, C. D., Davis, J. B., and Randall, A. D. (2002). Vanilloid and TRP channels: A family of lipid‐gated cation channels. Neuropharmacology 42, 873–888. Berdiev, B., Prat, A., Cantiello, H., Ausiello, D., Fuller, C., Jovov, B., Benos, D., and Ismailov, I. (1996). Regulation of epithelial sodium channels by short actin filaments. J. Biol. Chem. 271, 17704–17710. Birnbaumer, L., Zhu, X., Jiang, M., Boulay, G., Peyton, M., Vannier, B., Brown, D., Platano, D., Sadeghi, H., Stefani, E., and Birnbaumer, M. (1996). On the molecular basis and regulation of
10. The Cytoskeletal Connection to Ion Channels
283
cellular capacitative calcium entry: Roles for Trp proteins. Proc. Natl. Acad. Sci. USA 93, 15195–15202. Blanchard, A., Ohanian, V., and Critchley, D. (1989). The structure and function of a‐actinin. J. Muscle Res. Cell. Motil. 10, 280–289. Bloch, R. J., Bezakova, G., Ursitti, J. A., Zhou, D., and Pumplin, D. W. (1997). A membrane skeleton that clusters nicotinic acetylcholine receptors in muscle. Soc. Gen. Physiol. Ser. 52, 177–195. Borras, T. (2003). Gene expression in the trabecular meshwork and the influence of intraocular pressure. Prog. Retin. Eye Res. 22, 435–463. Bray, D. (1992). Actin filaments: Structure and assembly. In ‘‘Cell Movements’’ (D. Bray, ed.), pp. 75–92. Garland Publishing Co., New York. Brenner, S. L., and Korn, E. D. (1979). Substoichiometric concentrations of cytochalasin D inhibit actin polymerization. J. Biol. Chem. 254, 9982–9985. Bretscher, A. (1981). Fimbrin is a cytoskeletal protein that crosslinks F‐actin in vitro. Proc. Natl. Acad. Sci. USA 78, 6849–6853. Bretscher, A., and Weber, K. (1980). Villin is a major protein of the microvillus cytoskeleton which binds both G and F actin in a calcium‐dependent manner. Cell 20, 839–847. Brotschi, E. A., Hartwig, J. H., and Stossel, T. P. (1978). The gelation of actin by actin‐binding protein. J. Biol. Chem. 253, 8988–8993. Brown, H. M., Ottoson, D., and Rudqvist, B. (1978). Crayfish stretch receptor: And investigation with voltage‐clamp and ion‐selective electrodes. J. Physiol. 284, 155–179. Brown, S. S., and Spudich, J. A. (1979). Cytochalasin inhibits the rate of elongation of actin filament fragments. J. Cell Biol. 83, 657–662. Calvet, J. C. (2002). Cilia in PKD‐Letting it all hang out. J. Am. Soc. Nephrol. 13, 2614–2616. Calvet, J. P. (2003). Ciliary signaling goes down. Nat. Genet. 33, 113–114. Canessa, C., Horisberger, J.‐D., and Rossier, B. (1993). Epithelial sodium channel related to proteins involved in neurodegeneration. Nature 361, 467–470. Cantiello, H. F. (1995). Role of the actin cytoskeleton on epithelial Naþ channel regulation. Kidney Int. 48, 970–984. Cantiello, H. F. (1999). Regulatory aspects of Apx, a novel Naþ channel with connections to the cytoskeleton. In ‘‘Current Topics in Membranes’’ (D. J. Benos, ed.), pp. 177–194. Academic Press, New York. Cantiello, H. F. (2003). A tale of two tails: Ciliary mechanotransduction in ADPKD. Trends Mol. Med. 9, 234–236. Cantiello, H. F., and Prat, A. G. (1996). Role of actin filament organization in ion channel activity and cell volume regulation. In ‘‘Membrane Protein‐Cytoskeleton Interactions’’ (W. J. Nelson, ed.), pp. 373–396. Academic Press, San Diego. Cantiello, H. F., Stow, J., Prat, A. G., and Ausiello, D. A. (1991). Actin filaments control epithelial Naþ channel activity. Am. J. Physiol. 261, C882–C888. Cantiello, H. F., Prat, A. G., Bonventre, J. V., Cunningham, C. C., Hartwig, J., and Ausiello, D. A. (1993). Actin‐binding protein contributes to cell volume regulatory ion channel activation in melanoma cells. J. Biol. Chem. 268, 4596–4599. Cantiello, H. F., Montalbetti, N., Timpanaro, G. A., and Gonza´lez‐Perrett, S. (2004). Polycystin‐2 as a signal transducer. In ‘‘Cell Volume and Signaling’’ (P. K. Lauf and N. C. Adragna, eds.), pp. 235–244. Springer Sci. & Business Media, Inc., New York. Cartier, M. F., Pantaloni, D., and Korn, E. D. (1984). Evidence for an ATP cap at the end of actin filaments and its regulation of the F‐actin steady state. J. Biol. Chem. 259, 9983–9986. Chasan, B., Geisse, N. A., Pedatella, K., Wooster, D. G., Teintze, M., Carattino, M. D., Goldmann, W. H., and Cantiello, H. F. (2002). Evidence for direct interaction between actin and the cystic fibrosis transmembrane conductance regulator. Eur. Biophys. J. 30, 617–624.
284
Cantiello et al.
Chen, X. Z., Vassilev, P. M., Basora, N., Peng, J. B., Nomura, H., Segal, Y., Brown, E. M., Reeders, S. T., Hediger, M. A., and Zhou, J. (1999). Polycystin‐L is a calcium‐regulated cation channel permeable to calcium ions. Nature 401, 383–386. Chorna‐Ornan, I., Joel‐Almagor, T., Ben‐Ami, H. C., Frechter, S., Gillo, B., Selinger, Z., Gill, D. L., and Minke, B. (2001). A common mechanism underlies vertebrate calcium signaling and Drosophila phototransduction. J. Neurosci. 21, 2622–2629. Cipolla, M. J., Gokina, N. I., and Osol, G. (2002). Pressure‐induced actin polymerization in vascular smooth muscle as a mechanism underlying myogenic behavior. FASEB J. 16, 72–76. Clapham, D. E. (2002). Hot and cold TRP ion channels. Science 295, 2228–2229. Clapham, D. E. (2003). TRP channels as cellular sensors. Nature 426, 517–524. Coillot, J. P., and Boistel, J. (1969). Electrical activity along the stretch receptors of the metathoracic leg of the cricket, Schistocerca gregaria. J. Insect Physiol. 15, 1449–1470. Corey, D. P. (2003). New TRP channels in hearing and mechanosensation. Neuron 39, 585–588. Corey, D. P., Garcia‐Anoveros, J., Holt, J. R., Kwan, K. Y., Lin, S.‐Y., Vollrath, M. A., Amalfitano, A., Cheung, E. L.‐M., Derfler, B. H., Duggan, A., Geleoc, G. S. G., Gray, P. A., et al. (2004). TRPA1 is a candidate for the mechanosensitive transduction channel of vertebrate hair cells. Nature 432, 723–730. Corgan, A. M., Singleton, C., Santoso, C. B., and Greenwood, J. A. (2004). Phosphoinositides diVerentially regulate alpha‐actinin flexibility and function. Biochem. J. 378, 1067–1072. Coyle, J. E., Qamar, S., Rajashankar, K. R., and Nikolov, D. B. (2002). Structure of GABARAP in two conformations: Implications for GABA(A) receptor localization and tubulin binding. Neuron 33, 63–74. Dahlin, L. B., Nordborg, C., and Lundborg, G. (1987). Morphologic changes in nerve cell bodies induced by experimental graded nerve compression. Exp. Neurol. 95, 611–621. Dartsch, P. C., and Hammerle, H. (1986). Orientation response of arterial smooth muscle cells to mechanical stimulation. Eur. J. Cell Biol. 41, 339–346. Delmas, P. (2004). Assembly and gating of TRPC channels in signalling microdomains. Novartis Found. Symp. 258, 75–97. Delmas, P. (2005). Polycystins: Polymodal receptor/ion channel cellular sensors. Pflu¨gers Arch. 451, 264–276. den Dekker, E., Hoenderop, J. G. J., Nilius, B., and Bindels, R. J. M. (2003). The epithelial calcium channels, TRPV5 & TRPV6: From identification towards regulation. Cell Calcium 33, 497–507. Dhani, S. U., Mohammad‐Pannah, R., Ahmed, N., Ackerley, C., Ramjeesing, M., and Bear, C. E. (2003). Evidence for a functional interaction between the ClC‐2 chloride channel and the retrograde motor dynein complex. J. Biol. Chem. 278, 16262–16270. Di Palma, F., Belyantseva, I., Kim, H. J., Vogt, T. F., Kachar, B., and Noben‐Trauth, K. (2002). Mutations in Mcoln3 associated with deafness and pigmentation defects in varitint‐ waddler (Va) mice. Proc. Natl. Acad. Sci. USA 99, 14994–14999. Drenckhahn, D., Schluter, K., Allen, D. P., and Bennett, V. (1985). Colocalization of Band 3 with ankyrin and spectrin at the basal membrane of intercalated cells in the rat kidney. Science 230, 1287–1290. Edelstein, N. G., Catteral, W. A., and Moon, R. T. (1988). Identification of a 33 kD cytoskeletal protein with high aYnity for the sodium channel. Biochemistry 27, 1818–1822. Endlich, N., Sunohara, M., Nietfeld, W., Wolski, E. W., Schiwek, D., Kranzlin, B., Gretz, N., Kriz, W., EickhoV, H., and Endlich, K. (2002). Analysis of diVerential gene expression in stretched podocytes: Osteopontin enhances adaptation of podocytes to mechanical stress. FASEB J. 16, 1850–1852.
10. The Cytoskeletal Connection to Ion Channels
285
Fay, M. J., Bush, M. J., and Verlangieri, A. J. (1990). EVects of cytochalasin B on the uptake of ascorbic acid and glucose by 3T3 fibroblasts: Mechanism of impaired ascorbate transport in diabetes. Life Sci. 46, 619–624. Fink, M., Lesage, F., Duprat, F., Heurteaux, C., Reyes, R., Fosset, M., and Lazdunski, M. (1998). A neuronal two P domain Kþ channel stimulated by arachidonic acid and polyunsaturated fatty acids. EMBO J. 17, 3297–3308. Flockerzi, V., Jung, C., Aberle, T., Meissner, M., Freichel, M., Phillip, S. E., Nastainczyk, W., Maurer, P., and Zimmermann, R. (2005). Specific detection and semi‐quantitative analysis of TRPC4 protein expression by antibodies. Pflu¨gers Arch. 451, 81–86. Forer, A., Emmersen, J., and Behnke, O. (1972). Cytochalasin B: Does it aVect actin‐like filaments? Science 175, 774–776. Fraley, T. S., Tran, T. C., Corgan, A. M., Nash, C. A., Hao, J., Critchley, D. R., and Greenwood, J. A. (2003). Phosphoinositide binding inhibits alpha‐actinin bundling activity. J. Biol. Chem. 278, 24039–24045. Gallagher, A. R., Cedzich, A., Gretz, N., Somlo, S., and Witzgall, R. (2000). The polycystic kidney disease protein PKD2 interacts with Hax‐1, a protein associated with the actin cytoskeleton. Proc. Natl. Acad. Sci. USA 97, 4017–4022. Gao, X., Wu, L., and O’Neil, R. G. (2003). Temperature‐modulated diversity of TRPV4 channel gating: Activation by physical stresses and phorbol ester derivatives through protein kinase C‐ dependent and ‐independent pathways. J. Biol. Chem. 278, 27129–27137. Gardel, M. L., Shin, J. H., MacKintosh, F. C., Mahadevan, L., Matsudaira, P., and Weitz, D. A. (2004a). Elastic behavior of cross‐linked and bundled actin networks. Science 304, 1301–1305. Gardel, M. L., Shin, J. H., MacKintosh, F. C., Mahadevan, L., Matsudaira, P. A., and Weitz, D. A. (2004b). Scaling of F‐actin network rheology to probe single filament elasticity and dynamics. Phys. Rev. Lett. 93, 188102. Gasull, X., Ferrer, E., Llobet, A., Castellano, A., Nicolas, J. N., Pales, J., and Gual, A. (2003). Cell membrane stretch modulates the high‐conductance Ca2þ‐activated Kþ channel in bovine trabecular meshwork cells. Invest. Ophthalmol. Vis. Sci. 44, 706–714. Geng, L., B. C. R., Li, H. P., and Wilson, P. D. (2000). Modification of the composition of polycystin‐1 multiprotein complexes by calcium and tyrosine phosphorylation. Biochim. Biophys. Acta 1535, 21–35. Gluck, U., and Ben Ze’ev, A. (1994). Modulation of alpha‐actinin levels aVects cell motility and confers tumorigenicity on 3T3 cells. J. Cell Sci. 107, 1773–1782. Goddette, D. W., and Frieden, C. (1986a). Actin polymerization: The mechanism of action of cytochalasin D. J. Biol. Chem. 261, 15974–15980. Goddette, D. W., and Frieden, C. (1986b). The kinetics of cytochalasin D binding to monomeric actin. J. Biol. Chem. 261, 15970–15973. Goel, M., Sinkins, W., Keightley, A., Kinter, M., and Schilling, W. P. (2005). Proteomic analysis of TRPC5‐ and TRPC6‐binding partners reveals interaction with the plasmalemmal Naþ/Kþ‐ATPase. Pflu¨gers Arch. 451, 87–98. Golovina, V. A., Platoshyn, O., Bailey, C. L., Wang, J., Limsuwan, A., Sweeney, M., Rubin, L. J., and Yuan, J. X. (2001). Upregulated TRP and enhanced capacitative Ca2þ entry in human pulmonary artery myocytes during proliferation. Am. J. Physiol. 280, H746–H755. Gong, Z., Son, W., Chung, Y. D., Kim, J., Shin, D. W., McClung, C. A., Lee, Y., Lee, H. W., Chang, D. J., Kaang, B. K., Cho, H., Oh, H., et al. (2004). Two interdependent TRPV channel subunits, inactive and Nanchung, mediate hearing in Drosophila. J. Neurosci. 24, 9059–9066.
286
Cantiello et al.
Gonza´lez‐Perrett, S., Kim, K., Ibarra, C., Damiano, A. E., Zotta, E., Batelli, M., Harris, P. C., Reisin, I. L., Arnaout, M. A., and Cantiello, H. F. (2001). Polycystin‐2, the protein mutated in autosomal dominant polycystic kidney disease (ADPKD), is a Ca2þ‐permeable nonselective cation channel. Proc. Natl. Acad. Sci. USA 98, 1182–1187. Gonza´lez‐Perrett, S., Batelli, M., Kim, K., Essafi, M., Timpanaro, G., Montalbetti, N., Reisin, I. L., Arnaout, M. A., and Cantiello, H. F. (2002). Voltage dependence and pH regulation of human polycystin‐2 mediated cation channel activity. J. Biol. Chem. 277, 24959–24966. Goodman, M. B., and Schwarz, E. M. (2003). Transducing touch in Caenorhabditis elegans. Annu. Rev. Physiol. 65, 429–452. Goswami, C., Dreger, M., Jahnel, R., Bogen, O., Gillen, C., and Hucho, F. (2004). Identification and characterization of a Ca2þ‐sensitive interaction of the vanilloid receptor TRPV1 with tubulin. J. Neurosci. 91, 1092–1103. Greenwood, J. A., Theibert, A. B., Prestwich, G. D., and Murphy‐Ullrich, J. E. (2000). Restructuring of focal adhesion plaques by PI 3‐kinase. Regulation by PtdIns (3,4,5)‐p(3) binding to alpha‐actinin. J. Cell Biol. 150, 627–642. Gu, L., Liu, W., and Martinac, B. (1998). Electromechanical coupling model of gating the large mechanosensitive ion channel (MscL) of Escherichia coli by mechanical force. Biophys. J. 74, 2889–2902. Guharay, F., and Sachs, F. (1984). Stretch‐activated single ion channel currents in tissue‐ cultured embryonic chick skeletal muscle. J. Physiol. 352, 685–701. Guler, A. D., Lee, H., Iida, T., Shimizu, I., Tominaga, M., and Caterina, M. (2002). Heat‐ evoked activation of the ion channel, TRPV4. J. Neurosci. 2002, 6408–6414. Hall, P. F. (1984). The role of the cytoskeleton in hormone action. Can. J. Biochem. Cell Biol. 62, 653–665. Hamill, O. P., and McBride, D. W. (1992). Rapid adaptation of the mechanosensitive channel in Xenopus oocytes. Proc. Natl. Acad. Sci. USA 89, 7462–7466. Hamill, O. P., and McBride, D. (1993). Molecular clues to mechanosensitivity. Biophys. J. 66, 17–18. Hamill, O. P., and Martinac, B. (2001). Molecular basis of mechanotransduction in living cells. Physiol. Rev. 81, 685–740. Hamill, O. P., Marty, A., Neher, E., Sakmann, B., and Sigworth, F. J. (1981). Improved patch‐ clamp techniques for high‐resolution current recording from cells and cell‐free membrane patches. Pflu¨gers Arch. 391, 85–100. Hardy, M., and DiBona, D. (1982). Microfilaments and the hydrosmotic action of vasopressin in toad urinary bladder. Am. J. Physiol. 243, C200–C204. Harteneck, C., Plant, T. D., and Schultz, G. (2000). From worm to man: Three subfamilies of TRP channels. Trends Neurosci. 23, 159–166. Hartwig, J. H., and Stossel, T. P. (1976). Interactions of actin, myosin and an actin‐binding protein of rabbit pulmonary macrophages. 3. EVects of cytochalasin B. J. Cell Biol. 77, 295–302. Hartwig, J. H., and Stossel, T. P. (1979). Cytochalasin B and the structure of actin gels. J. Mol. Biol. 734, 539–553. Hartwig, J. H., and Stossel, T. S. (1981). The structure of actin‐binding protein molecules in solution and interacting with actin filaments. J. Mol. Biol. 145, 563–581. Hodson, S., and Earlman, R. (1993). The incorporation of gel pressure into the irreversible thermodynamic equation of fluid flow in order to explain biological tissue swelling. J. Theor. Biol. 163, 173–180. Hoenderop, J. G., Nilius, B., and Bindels, R. J. (2002). ECaC: The gatekeeper of transepithelial Ca2þ transport. Biochim. Biophys. Acta 1600, 6–11.
10. The Cytoskeletal Connection to Ion Channels
287
Hoenderop, J. G., Voets, T., Hoefs, S., Weidema, F., Prenen, J., Nilius, B., and Bindels, R. J. (2003). Homo‐ and heterotetrameric architecture of the epithelial Ca2þ channels TRPV5 and TRPV6. EMBO J. 22, 776–785. Howard, J., and Bechstedt, S. (2004). Hypothesis: A helix of ankyrin repeats of the NOMPC‐TRP ion channel is the gating spring of mechanoreceptors. Curr. Biol. 14, R224–R226. Huber, A., Sander, P., Gobert, A., Bahner, M., Hermann, R., and Paulsen, R. (1996). The transient receptor potential protein (Trp), a putative store‐operated Ca2þ channel essential for phosphoinositide‐mediated photoreception, forms a signaling complex with NorpA, InaC and InaD. EMBO J. 15, 7036–7045. Ingber, D. E. (1997). Tensegrity: The architectural basis of cellular mechanotransduction. Annu. Rev. Physiol. 59, 575–599. Ismailov, I. I., Berdiev, B. K., Shlyonsky, V. G., Fuller, C. M., Prat, A. G., Jovov, B., Cantiello, H. F., Ausiello, D. A., and Benos, D. J. (1997). Role of actin in regulation of epithelial sodium channels by CFTR. Am. J. Physiol. 272, C1077–C1086. Ito, T., Zaner, K. S., and Stossel, T. P. (1987). Nonideality of volume flows and phase transitions of F‐actin solutions in response to osmotic stress. Biophys. J. 51, 745–753. Ives, C. L., Eskin, S. G., and McIntire, L. V. (1986). Mechanical eVects on endothelial cell morphology: In vitro assessment. In Vitro Cell. Dev. Biol. 22, 500–507. Janmey, P. A., Hvidt, S., Lamb, J., and Stossel, T. (1990a). Resemblance of actin binding proteins/actin gels to crosslinked networks. Nature 347, 95–99. Janmey, P. A., Hvidt, S., Oster, G. F., Lamb, J., Stossel, T. P., and Hartwig, J. H. (1990b). EVect of ATP on actin filament stiVness. Nature 347, 96–99. Jaquemar, D., Schenker, T., and Truch, B. (1999). An ankyrin‐like protein with transmembrane domains is specifically lost after oncogenic transformation of human fibroblasts. J. Biol. Chem. 274, 7325–7333. Ji, G., Barsotti, R. J., Feldman, M. E., and KotlikoV, M. I. (2002). Stretch‐induced calcium release in smooth muscle. J. Gen. Physiol. 119, 533–544. Jing, J., Peretz, T., Singer‐Lahat, D., Chikvashvili, D., Thornhill, W. B., and Lotan, I. (1997). Inactivation of a voltage‐dependent Kþ channel by beta subunit. Modulation by a phosphorylation‐dependent interaction between the distal C terminus of alpha subunit and cytoskeleton. J. Biol. Chem. 272, 14021–14024. Johnson, B. D., and Byerly, L. (1993). A cytoskeletal mechanism for Ca2þ channel metabolic dependence and inactivation by intracellular Ca2þ. Neuron 10, 797–804. Johnson, B. D., and Byerly, L. (1994). Ca2þ channel Ca2þ‐dependent inactivation in a mammalian central neuron involves the cytoskeleton. Pflu¨gers Arch. 429, 14–21. Johnson, B. D., Scheuer, T., and Catterall, W. A. (2005). Convergent regulation of skeletal muscle Ca2þ channels by dystrophin, the actin cytoskeleton, and cAMP‐dependent protein kinase. Proc. Natl. Acad. Sci. USA 102, 4191–4196. Johnstone, M. A., and Grant, W. G. (1973). Pressure‐dependent changes in structures of the aqueous outflow system of human and monkey eyes. Am. J. Ophthalmol. 75, 365–383. Jordt, S. E., Bautista, D. M., Chuang, H. H., McKemy, D. D., Zygmunt, P. M., Hogestatt, E. D., Meng, I. D., and Julius, D. (2004). Mustard oils and cannabinoids excite sensory nerve fibres through the TRP channel ANKTM1. Nature 427, 260–265. Kalapesi, F. B., Tan, J. C. H., and Coroneo, M. T. (2005). Stretch‐activated channels: A mini‐ review. Are stretch‐activated channels an ocular barometer? Clin. Exp. Ophthalmol. 33, 210–217. Kaplan, J. M., and Pollak, M. R. (2001). Familial focal segmental glomerulosclerosis. Curr. Opin. Nephrol. Hypertens. 10, 183–187.
288
Cantiello et al.
Kaplan, J. M., Kim, S. H., North, K. N., Rennke, H., Correia, L. A., Tong, H. Q., Mathis, B. J., Rodriguez‐Perez, J. C., Allen, P. G., Beggs, A. H., and Pollak, M. R. (2000). Mutations in ACTN4, encoding alpha‐actinin‐4, cause familial focal segmental glomerulosclerosis. Nat. Genet. 24, 251–256. Katz, B. (1950). Depolarization of sensory terminals and the initiation of impulses in the muscle spindle. J. Physiol. 111, 261–282. Kelly, D., Mackenzie, L., Hunter, P., Smaill, B., and Saint, D. (2006). Gene expression of stretch‐activated channels and mechanoelectric feedback in the heart. Clin. Exp. Pharmacol. Physiol. 33, 642–648. Kim, C. H., and Lisman, J. E. (2001). A labile component of AMPA receptor‐mediated synaptic transmission is dependent on microtubule motors, actin, and N‐ethylmaleimide‐ sensitive factor. J. Neurosci. 21, 4188–4194. Kim, J., Chung, Y. D., Park, D. Y., Choi, S., Shin, D. W., Soh, H., Lee, H. W., Son, W., Yim, J., Park, C. S., Kernan, M. J., and Kim, C. (2003). A TRPV family ion channel required for hearing in Drosophila. Nature 424, 81–84. Kirsch, J., Langosch, D., Prior, P., Littauer, U. Z., Schmitt, B., and Betz, H. (1991). The 93‐ kDa glycine receptor‐associated protein binds to tubulin. J. Biol. Chem. 266, 22242–22245. Ko, K. S., and McCulloch, C. A. (2000). Partners in protection: Interdependence of cytoskeleton and plasma membrane in adaptations to applied forces. J. Membr. Biol. 174, 85–95. Korn, E. D., Carlier, M. F., and Pantaloni, D. (1987). Actin polymerization and ATP hydrolysis. Science 238, 638–644. Koulen, P., Cai, Y., Geng, L., Maeda, Y., Nishimura, S., Witzgall, R., Ehrlich, B. E., and Somlo, S. (2002). Polycystin‐2 is an intracellular calcium release channel. Nat. Cell Biol. 4, 191–197. Kreis, T., and Vale, R. (1999). ‘‘Guidebook to the Cytoskeletal and Motor Proteins.’’ Oxford University Press, Oxford, UK. Kreisberg, J. I., Karnovsky, M. J., and Levine, L. (1982). Prostaglandin production by homogeneous cultures of rat glomerular epithelial and mesangial cells. Kidney Int. 22, 355–359. Krupp, J. J., Vissel, B., Thomas, C. G., Heinemann, S. F., and Westbrook, G. L. (1999). Interactions of calmodulin and a‐actinin with the NR1 subunit modulate Ca2þ‐dependent inactivation of NMDA receptors. J. Neurosci. 19, 1165–1178. Lader, A. S., Kwiatkowski, D. J., and Cantiello, H. F. (1999). Role of gelsolin in the actin filament regulation of cardiac L‐type calcium channels. Am. J. Physiol. 277, C1277–C1283. Laitko, U., Juranka, P. F., and Morris, C. E. (2006). Membrane stretch slows the concerted step prior to opening of the Kv channel. J. Gen. Physiol. 127, 687–701. Lee, G., Abdi, K., Jiang, Y., Michaely, P., Bennett, V., and Marszalek, P. (2006). Nanospring behaviour of ankyrin repeats. Nature 440, 246–249. Lei, S., Czerwinska, E., Czerwinski, W., Walsh, M. P., and MacDonald, J. F. (2001). Regulation of NMDA receptor activity by F‐actin and myosin light chain kinase. J. Neurosci. 21, 8464–8472. Li, Q., Dai, Y., Guo, L., Liu, Y., Hao, C., Wu, G., Basora, N., Michalak, M., and Chen, X. Z. (2003a). Polycystin‐2 associates with tropomyosin‐1, an actin microfilament component. J. Mol. Biol. 325, 949–962. Li, Q., Shen, P. Y., Wu, G., and Chen, X. Z. (2003b). Polycystin‐2 interacts with troponin I, an angiogenesis inhibitor. Biochemistry 42, 450–457. Li, Q., Montalbetti, N., Shen, P. Y., Dai, X.‐Q., Cheeseman, C. I., Karpinski, E., Wu, G., Cantiello, H. F., and Chen, X.‐Z. (2005). Alpha‐actinin associates with polycystin‐2 and regulates its channel activity. Hum. Mol. Genet. 14, 1587–1603.
10. The Cytoskeletal Connection to Ion Channels
289
Li, S. W., Westwick, J., and Poll, C. T. (2002). Receptor‐operated Ca2þ influx channels in leukocytes: A therapeutic target? Trends Pharmacol. Sci. 23, 63–70. Liedtke, W., Choe, Y., Marti‐Renom, M. A., Bell, A. M., Denis, C. S., Sali, A., Hudspeth, A. J., Friedman, J. M., and Heller, S. (2000). Vanilloid receptor‐related osmotically activated channel (VR‐OAC), a candidate vertebrate osmoreceptor. Cell 103, 525–535. Lin, D. C., and Lin, S. (1979). Actin polymerization induced by a motility‐related high aYnity cytochalasin binding complex from human erythrocyte membrane. Proc. Natl. Acad. Sci. USA 76, 2345–2349. Lin, D. C., Tobin, K. D., Grumet, M., and Lin, S. (1980). Cytochalasins inhibit nuclei‐induced actin polymerization by blocking filament elongation. J. Cell Biol. 84, 455–460. Lin, S., Lin, D. C., Spudich, J. A., and Kun, E. (1973). Inhibition of mitochondrial contraction by cytochalasin B. FEBS Lett. 37, 241–243. Lintschinger, B., Balzer‐Geldesetzer, M., Baskaran, T., Graier, W. F., Romanin, C., Zhu, M. X., and Groshner, K. (2000). Coassembly of Trp1 and Trp3 proteins generates diacylglycerol‐ and Ca2þ‐sensitive cation channels. J. Biol. Chem. 275, 27799–27805. Luccardini, C., Casagrande, S., Cupello, A., Pellistri, F., Ramoino, P., and Robello, M. (2004). The combined disruption of microfilaments and microtubules aVects the distribution and function of GABA(A) receptors in rat cerebellum granule cells in culture. Neurosci. Lett. 359, 25–28. Luo, Y., Vassilev, P. M., Li, X., Kawanabe, Y., and Zhou, J. (2003). Native polycystin 2 functions as a plasma membrane Ca2þ‐permeable cation channel in renal epithelia. Mol. Cell. Biol. 23, 2600–2607. MacKnight, A. D. C. (1988). Principles of cell volume regulation. Renal Physiol. Biochem. 3–5, 114–141. MacLean‐Fletcher, S., and Pollard, T. D. (1980). Mechanism of action of cytochalasin B on actin. Cell 20, 329–341. Maguire, G., Connaughton, V., Prat, A. G., Jackson, G. R., Jr., and Cantiello, H. F. (1998). Actin cytoskeleton regulates ion channel activity in retinal neurons. Neuroreport 9, 665–670. Maingret, F., Fosset, M., Lesage, F., Lazdunski, M., and Honore, E. (1999). TRAAK is a mammalian neuronal mechano‐gated Kþ channel. J. Biol. Chem. 274, 1381–1387. Maingret, F., Patel, A. J., Lesage, F., Lazdunski, M., and Honore, E. (2000). Lysophospholipids open the two‐pore domain mechano‐gated Kþ channels TREK‐1 and TRAAK. J. Biol. Chem. 275, 10128–10133. Mallouk, N., and Allard, B. (2000). Stretch‐induced activation of Ca2þ‐activated Kþ channels in mouse skeletal muscle fibers. Am. J. Physiol. Cell Physiol. 278, 473–479. Markey, F., Lindberg, U., and Eriksson, L. (1978). Human platelets contain profilin, a potential regulator of actin polymerisability. FEBS Lett. 88, 75–79. Markin, V. S., and Sachs, F. (2004). Thermodynamics of mechanosensitivity. Phys. Biol. 1, 110–124. Maroto, R., Wood, T. G., Kurosky, A., Martinac, B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nat. Cell Biol. 7, 179–185. Martinac, B. (2004). Mechanosensitive ion channels: Molecules of mechanotransduction. J. Cell Sci. 117, 2449–2460. Maruoka, N. D., Steele, D. F., Au, B. P., Dan, P., Zhang, X., Moore, E. D., and Fedida, D. (2000). Alpha‐actinin‐2 couples to cardiac Kv1.5 channels, regulating current density and channel localization in HEK cells. FEBS Lett. 473, 188–194. Matsudaira, P. (1991). Modular organization of actin crosslinking proteins. Trends Biol. Sci. 16, 87–92.
290
Cantiello et al.
Matsuo, T., and Matsuo, N. (1996). Intracellular calcium response to hydraulic pressure in human trabecular cells. Br. J. Ophthalmol. 80, 561–566. Matsuo, T., Uchida, H., and Matsuo, N. (1996). Bovine and porcine trabecular cells produce prostaglandin F2 alpha in response to cyclic mechanical stretching. Jpn. J. Ophthalmol. 40, 289–296. Mazzanti, M., Assandri, R., Ferroni, A., and DiFrancesco, D. (1996). Cytoskeletal control of rectification and expression of four substates in cardiac inward rectifier Kþ channels. FASEB J. 10, 357–361. Mazzochi, C., Bubien, J. K., Smith, P. R., and Benos, D. J. (2006). The carboxyl terminus of the alpha‐subunit of the amiloride‐sensitive epithelial sodium channel binds to F‐actin. J. Biol. Chem. 281, 6528–6538. Mendelson, M., and Loewenstein, W. R. (1964). Mechanisms of receptor adaptation. Science 144, 554–555. Meyer, R. K., and Aebi, U. (1989). Biochemical and structural analysis of the interaction of a‐actinin with actin filaments and lipids. In ‘‘Cytoskeletal and Extracellular Proteins’’ (U. Aebi and J. Engel, eds.), pp. 57–59. Springer‐Verlag, Berlin. Mikuni‐Takagaki, Y. (1999). Mechanical responses and signal transduction pathways in stretched osteocytes. J. Bone Miner. Metab. 17, 57–60. Minke, B. (2001). The TRP channel and phospholipase C‐mediated signaling. Cell. Mol. Neurobiol. 21, 629–643. Minke, B., and Cook, B. (2002). TRP channel proteins and signal transduction. Physiol. Rev. 82, 429–472. Mitsui, T., Kawajiri, M., Kunishige, M., Endo, T., Akaike, M., Aki, K., and Matsumoto, T. (2000). Functional association between nicotinic acetylcholine receptor and sarcomeric proteins via actin and desmin filaments. J. Cell. Biochem. 77, 584–595. Mobasheri, A., Carter, S. D., Martin‐Vasallo, P., and Shakibaei, M. (2002). Integrins and stretch activated ion channels; putative components of functional cell surface mechanoreceptors in articular chondrocytes. Cell Biol. Int. 26, 1–18. Mochizuki, T., Wu, G., Hayashi, T., Xenophontos, S. L., Veldhuisen, B., Saris, J. J., Reynolds, D. M., Cai, Y., Gabow, P. A., Pierides, A., Kimberling, W. J., Breuning, W. J., et al. (1996). PKD2, a gene for polycystic kidney disease that encodes an integral membrane protein. Science 272, 1339–1342. Montalbetti, N., Li, Q., Gonza´lez‐Perrett, S., Semprine, G., Chen, X.‐Z., and Cantiello, H. F. (2005a). EVect of hydro‐osmotic pressure on polycystin‐2 channel function in the human syncytiotrophoblast. Pflu¨gers Arch. 451, 294–303. Montalbetti, N., Li, Q., Timpanaro, G. A., Gonza´lez‐Perrett, S., Dai, X.‐Q., Chen, X.‐Z., and Cantiello, H. F. (2005b). Cytoskeletal regulation of calcium‐permeable cation channels in the human syncytiotrophoblast: Role of gelsolin. J. Physiol. (Lond.) 566, 309–325. Montalbetti, N., Li, Q., Chen, X.‐Z., and Cantiello, H. F. (2006). Cation channel function in the human syncytiotrophoblast is under control of microtubular structures. Placenta 27, A.58 (Abstract). Montell, C. (2001). Physiology, phylogeny, and functions of the TRP superfamily of cation channels. Sci. STKE 90, RE1. Montell, C. (2005). Drosophila TRP channels. Pflu¨gers Arch. 451, 19–28. Montell, C., Birnbaumer, L., and Flockerzi, V. (2002a). The TRP channels, a remarkably functional family. Cell 108, 595–598. Montell, C., Birnbaumer, L., Flockerzi, V., Bindels, R. J., Bruford, E. A., Caterina, M. J., Clapham, D. E., Harteneck, C., Heller, S., Julius, D., Kojima, I., Mori, I., et al. (2002b). A unified nomenclature for the superfamily of TRP cation channels. Mol. Cells 9, 229–231. Morris, C. E. (1990). Mechanosensitive ion channels. J. Membr. Biol. 113, 93–107.
10. The Cytoskeletal Connection to Ion Channels
291
Morris, C. E., and Sigurdson, W. J. (1989). Stretch‐inactivated ion channel co‐exists with stretch activated ion channels. Science 243, 807–809. Morrow, J. S. (1989). The spectrin membrane skeleton: Emerging concepts. Curr. Opin. Cell Biol. 1, 23–29. Muraki, K., Iwata, Y., Katanosaka, Y., Ito, T., Ohya, S., Shigekawa, M., and Imaizumi, Y. (2003). TRPV2 is a component of osmotically sensitive cation channels in murine aortic myocytes. Circ. Res. 93, 829–838. Nakahira, K., Matos, M. F., and Trimmer, J. S. (1999). DiVerential interaction of voltage‐gated Kþ channel beta‐subunits with cytoskeleton is mediated by unique amino terminal domains. J. Mol. Neurosci. 11, 199–208. Nauli, S. M., and Zhou, J. (2004). Polycystins and mechanosensation in renal and nodal cilia. Bioessays 26, 844–856. Nauli, S. M., Alenghat, F. J., Luo, Y., Williams, E., Vassilev, P., Li, X., Elia, A. E. H., Lu, W., Brown, E. M., Quinn, S. J., Ingber, D. E., and Zhou, J. (2003). Polycystins 1 and 2 mediate mechanosensation in the primary cilium of kidney cells. Nat. Genet. 33, 129–137. Niemeyer, B. A., Suzuki, E., Scott, K., Jalink, K., and Zuker, C. S. (1996). The Drosophila light‐ activated conductance is composed of the two channels TRP and TRPL. Cell 85, 651–659. Nieset, J. E., Redfield, A. R., Jin, F., Knudsen, K. A., Johnson, K. R., and Wheelock, M. J. (1997). Characterization of the interactions of alpha‐catenin with alpha‐actinin and beta‐ catenin/plakoglobin. J. Cell Sci. 110, 1013–1022. O’Neil, R. G., and Heller, S. (2005). The mechanosensitive nature of TRPV channels. Pflu¨gers Arch. 451, 193–203. Okada, M., Matsumori, A., Ono, K., Furukawa, Y., Shioi, T., Iwasaki, A., Matsushima, K., and Sasayama, S. (1998). Cyclic stretch upregulates production of interleukin‐8 and monocyte chemotactic and activating factor/monocyte chemoattractant protein‐1 in human endothelial cells. Arterioscler. Thromb. Vasc. Biol. 18, 894–901. Ong, A. C. M., and Wheatley, D. N. (2003). Polycystic kidney disease: The ciliary connection. The Lancet 361, 774–776. Ornelles, D. A., Fey, E. G., and Penman, S. (1986). Cytochalasin releases mRNA from the cytoskeletal framework and inhibits protein synthesis. Mol. Cell. Biol. 6, 1650–1662. Otey, C. A., Vasquez, G. B., Burridge, K., and Erickson, B. W. (1993). Mapping of the alpha‐ actinin binding site within the beta 1 integrin cytoplasmic domain. J. Biol. Chem. 268, 21193–21197. Otto, J. J. (1994). Actin‐bundling proteins. Curr. Opin. Cell Biol. 6, 105–109. Painter, R. G., and McIntosh, A. T. (1979). The regional association of actin and myosin with sites of particle phagocytosis. J. Supramol. Struct. 12, 369–384. Patel, A. J., Honore, E., Maingret, F., Lesage, F., Fink, M., Duprat, F., and Lazdunski, M. (1998). A mammalian two pore domain mechano‐gated S‐like Kþ channel. EMBO J. 17, 4283–4290. Patel, A. J., and Honore, E. (2001). Properties and modulation of mammalian 2P domain Kþ channels. Trends Neurosci. 24, 339–346. Patel, A. J., Lazdunski, M., and Honore, E. (2001). Lipid and mechano‐gated 2P domain Kþ channels. Curr. Opin. Cell Biol. 13, 422–428. Pazour, G. J. (2004). Intraflagellar transport and cilia‐dependent renal disease: The ciliary hypothesis of polycystic kidney disease. J. Am. Soc. Nephrol. 15, 2528–2536. Pazour, G. J., and Witman, G. B. (2003). The vertebrate primary cilium is a sensory organelle. Curr. Opin. Cell Biol. 15, 105–110. Pazour, G. J., San Agustin, J. T., Follit, J. A., Rosenbaum, J. L., and Witman, G. B. (2002). Polycystin‐2 localizes to kidney cilia and the ciliary level is elevated in orpk mice with polycystic kidney disease. Curr. Biol. 12, R378–R380.
292
Cantiello et al.
Peng, J. B., Chen, X. Z., Berger, U. V., Vassilev, M., Tsukaguchi, H., Brown, E. M., and Hediger, M. (1999). Molecular cloning and characterization of a channel‐like transporter mediating intestinal calcium absorption. J. Biol. Chem. 274, 22739–22746. Perozo, E., Cortes, D. M., Sompornpisut, P., Kloda, A., and Martinac, B. (2002). Open channel structure of MscL and the gating mechanism of mechanosensitive channels. Nature 418, 942–948. Petrov, A. G., and Usherwood, P. N. (1994). Mechanosensitivity of cell membranes: Ion channels, lipid matrix and cytoskeleton. Eur. Biophys. J. 23, 1–19. Phillips, A. M., Bull, A., and Kelly, L. E. (1992). Identification of a Drosophila gene encoding a calmodulin‐binding protein with homology to the trp phototransduction gene. Neuron 8, 631–642. Praetorius, H. A., Frokiaer, J., Nielsen, S., and Spring, K. R. (2003). Bending the primary cilium opens Ca2þ‐sensitive intermediate‐conductance Kþ channels in MDCK cells. J. Membr. Biol. 191, 193–200. Praetorius, H. A., and Spring, K. R. (2001). Bending the MDCK cell primary cilium increases intracellular calcium. J. Membr. Biol. 184, 71–79. Prat, A. G., Ausiello, D. A., and Cantiello, H. F. (1993a). Vasopressin and protein kinase A activate G protein‐sensitive Na channels. Am. J. Physiol. 265, C218–C223. Prat, A. G., Bertorello, A. M., Ausiello, D. A., and Cantiello, H. F. (1993b). Activation of epithelial Naþ channels by protein kinase A requires actin filaments. Am. J. Physiol. 265, C224–C233. Prat, A. G., Jackson, Jr., G. R., Ausiello, D. A., and Cantiello, H. F. (1994). Actin filament organization modulates CFTR‐mediated Cl‐ channel activity. J. Gen. Physiol. 104, 38a. Prat, A. G., Xiao, Y.‐F., Ausiello, D. A., and Cantiello, H. F. (1995). cAMP‐independent regulation of CFTR by the actin cytoskeleton. Am. J. Physiol. 268, C1552–C1561. Prat, A., Holtzman, E., Brown, D., Cunningham, C., Reisin, I., Kleyman, T., McLaughlin, M., Jackson, Jr., G. R., Lydon, J., and Cantiello, H. (1996). Renal epithelial protein (Apx) is an actin cytoskeleton‐regulated Naþ channel. J. Biol. Chem. 271, 18045–18053. Prat, A. G., Cunningham, C. C., Jackson, G. R. J., Borkan, S., Wang, Y., Ausiello, D. A., and Cantiello, H. F. (1999). Actin filament organization is required for proper cAMP‐ dependent activation of CFTR. Am. J. Physiol. 277, C1160–C1169. Pratley, J. N., and McQuillen, N. K. (1973). The role of microfilaments in frog skin ion transport. J. Cell Biol. 56, 850–857. Putney, J. W., Jr., Broad, L. M., Braun, F. J., Lievremont, J. P., and Bird, G. S. (2001). Mechanisms of capacitative calcium entry. J. Cell Sci. 114, 2223–2229. Raychowdhury, M. K., Gonza´lez‐Perrett, S., Montalbetti, N., Timpanaro, G. A., Chasan, B., Goldmann, W. H., Stahl, S., Cooney, A., Goldin, E., and Cantiello, H. F. (2004). Molecular pathophysiology of mucolipidosis type IV. pH dysregulation of the human mucolipin‐1 cation channel. Hum. Mol. Genet. 13, 617–627. Raychowdhury, M. K., McLaughlin, M., Ramos, A., Montalbetti, N., Bouley, R., Ausiello, D. A., and Cantiello, H. F. (2005). Characterization of single channel currents from primary cilia of renal epithelial cells. J. Biol. Chem. 280, 34718–34722. Reichstein, E., and Korn, E. D. (1979). Acanthamoeba profilin. J. Biol. Chem. 254, 6174–6179. Rossier, B. C. (1998). Mechanosensitivity of the epithelial sodium channel (ENaC): Controversy or pseudocontroversy? J. Gen. Physiol. 112, 95–96. Rycroft, B. K., and Gibb, A. J. (2004). Regulation of single NMDA receptor channel activity by alpha‐actinin and calmodulin in rat hippocampal granule cells. J. Physiol. 557, 795–808.
10. The Cytoskeletal Connection to Ion Channels
293
Sackin, H. (1994). Stretch‐activated ion channels. In ‘‘Cellular and Molecular Physiology of Cell Volume Regulation’’ (K. Strange, ed.), pp. 215–240. CRC Press, Boca Raton, FL. Sadeghi, A., Doyle, A. D., and Johnson, B. D. (2002). Regulation of the cardiac L‐type Ca2þ channel by the actin‐binding proteins alpha‐actinin and dystrophin. Am. J. Physiol. 282, C1502–C1511. Sanger, J. W., and Sanger, J. M. (1975). Polymerization of sperm actin in the presence of cytochalasin B. J. Exp. Zool. 193, 441–447. ScheVers, M. S., Le, H., van der Bent, P., Leonhard, W., Prins, F., Spruit, L., Breuning, M. H., de Heer, E., and Peters, D. J. (2002). Distinct subcellular expression of endogenous polycystin‐2 in the plasma membrane and Golgi apparatus of MDCK cells. Hum. Mol. Genet. 11, 59–67. ScheVers, M. S., van der Bent, P., van de Wal, W. A., van Eendenburg, J., Breuning, M. H., de Heer, E., and Peters, D. J. (2004). Altered distribution and co‐localization of polycystin‐ 2 with polycystin‐1 in MDCK cells after wounding stress. Exp. Cell Res. 292, 219–230. Shieh, B.‐H., and Zhu, M.‐Y. (1996). Regulation of the TRP Ca2þ channel by INAD in Drosophila photoreceptors. Neuron 16, 991–998. Shizuta, Y., Shizuta, H., Gallo, M., Davies, P., and Pastan, I. (1976). Purification and properties of filamin, an actin binding protein from chicken gizzard. J. Biol. Chem. 251, 6562–6567. Shoop, R. D., Yamada, N., and Berg, D. K. (2000). Cytoskeletal links of neuronal acetylcholine receptors containing alpha 7 subunits. J. Neurosci. 20, 4021–4029. Skinner, S. J., Somervell, C. E., and Olson, D. M. (1992). The eVects of mechanical stretching on fetal rat lung cell prostacyclin production. Prostaglandins 43, 413–433. Small, D. L., and Morris, C. E. (1994). Delayed activation of single mechanosensitive channels in Lymnea neurons. Am. J. Physiol. 267, C598–C606. Smith, S. J. (1988). Neuronal cytomechanics of growth Cones. Science 242, 708–715. Smith, P. R., Saccomani, G., Joe, E., Angelides, K. J., and Benos, D. J. (1991). Amiloride‐ sensitive sodium channel is linked to the cytoskeleton in renal epithelial cells. Proc. Natl. Acad. Sci. USA 88, 6971–6975. Sotomayor, M., Corey, D. P., and Schulten, K. (2005). In search of the hair‐cell gating spring: Elastic properties of ankyrin and cadherin repeats. Structure 13, 669–682. Spudich, J. A., and Lin, S. (1972). Cytochalasin B, its interaction with actin and actomyosin. Proc. Natl. Acad. Sci. USA 69, 442–446. Srinivasan, Y., Elmer, L., Davis, J., Bennett, V., and Angelides, K. (1988). Ankyrin and spectrin associate with voltage‐dependent Naþ channels in brain. Nature 333, 177–180. Srinivasan, Y., Lewallen, M., and Angelides, K. J. (1992). Mapping the binding site on ankyrin for the voltage‐dependent sodium channel from brain. J. Biol. Chem. 267, 7483–7489. Stamenovic, D., and Ingber, D. E. (2002). Models of cytoskeletal mechanics of adherent cells. Biomech. Model Mechanobiol. 1, 95–108. Staub, O., Verrey, F., Kleyman, T. R., Benos, D. J., Rossier, B. C., and Kraehenbuhl, J.‐P. (1992). Primary structure of an apical protein from Xenopus laevis that participates in amiloride‐sensitive sodium channel activity. J. Cell Biol. 119, 1497–1506. Stendahl, O. I., Hartwig, J. H., Brotschi, E. A., and Stossel, T. P. (1980). Distribution of actin‐ binding protein and myosin in macrophages during spreading and phagocytosis. J. Cell Biol. 84, 215–224. Story, G. M., Peier, A. M., Reeve, A. J., Eid, S., Mosbacher, J., Hricik, T. R., Earley, T. J., Hergarden, A. C., Andersson, D. A., Hwang, S. W., McIntyre, P., Jegla, P., et al. (2003). ANKTM1, a TRP‐like channel expressed in nociceptive neurons, is activated by cold temperatures. Cell 112, 819–829.
294
Cantiello et al.
Stossel, T. P. (1982). The structure of cortical cytoplasm. Philos. Trans. R. Soc. Lond. Ser. B 299, 275–289. Stossel, T. P. (1984). Contribution of actin to the structure of cytomatrix. J. Cell Biol. 99, (Suppl.), 15s–21s. Stossel, T. P. (1993). On the crawling of animal cells. Science 260, 1086–1094. Strange, K. (1994). ‘‘Cellular and Molecular Physiology of Cell Volume Regulation.’’ CRC Press, Boca Raton. Strotmann, R., Harteneck, C., Nunnenmacher, K., Schultz, G., and Plant, T. D. (2000). OTRPC4, a nonselective cation channel that confers sensitivity to extracellular osmolarity. Nat. Cell Biol. 2, 695–702. Sukharev, S., and Corey, D. P. (2004). Mechanosensitive channels: Multiplicity of families and gating paradigms. Sci. STKE 219, re4‐4. Terakawa, K., Itoh, H., Sagawa, N., Yura, S., Yoshida, M., Korita, D., Kakui, K., and Fujii, S. (2002). Site‐specific augmentation of amnion cyclooxygenase‐2 and decidua vera phospholipase‐A2 expression in labor: Possible contribution of mechanical stretch and interleukin‐1 to amnion prostaglandin synthesis. J. Soc. Gynecol. Inves. 9, 68–74. Tracey, W. D., Wilson, R. I., Laurent, G., and Benzer, S. (2003). Painless, a Drosophila gene essential for nociception. Cell 113, 261–273. Tseng, Y., Fedorov, E., McCaVery, J. M., Almo, S. C., and Wirtz, D. (2001). Micromechanics and ultrastructure of actin filament networks crosslinked by human fascin: A comparison with a‐actinin. J. Mol. Biol. 310, 351–366. Tsiokas, L., Kim, E., Arnould, T., Sukhatme, Y. P., and Walz, G. (1997). Homo and hetero‐ dimeric interactions between the gene products of Pkd1 and Pkd2. Proc. Natl. Acad. Sci. USA 94, 6965–6970. Tsiokas, L., Arnould, T., Zhu, C., Kim, E., Walz, G., and Sukhatme, V. P. (1999). Specific association of the gene product of PKD2 with the TRPC1 channel. Proc. Natl. Acad. Sci. USA 96, 3934–3939. Undrovinas, A., Shander, G., and Makielski, J. (1996). Cytoskeleton modulates gating of voltage‐dependent Naþ channels in heart. Am. J. Physiol. 269, H203–H214. van Zundert, B., Alvarez, F. J., Yevenes, G. E., Carcamo, J. G., Vera, J. C., and Aguayo, L. G. (2002). Glycine receptors involved in synaptic transmission are selectively regulated by the cytoskeleton in mouse spinal neurons. J. Neurophysiol. 87, 640–644. van Zundert, B., Alvarez, F. J., Tapia, J. C., Yeh, H. H., Diaz, E., and Aguayo, L. G. (2004). Developmental‐dependent action of microtubule depolymerization on the function and structure of synaptic glycine receptor clusters in spinal neurons. J. Neurophysiol. 91, 1036–1049. Vassilev, P. M., Guo, L., Chen, X. Z., Segal, Y., Peng, J. B., Basora, N., Babakhanlou, H., Cruger, G., Kanazirska, M., Ye, C., Brown, E. M., Hediger, E. M., et al. (2001). Polycystin‐2 is a novel cation channel implicated in defective intracellular Ca2þ homeostasis in polycystic kidney disease. Biochem. Biophys. Res. Commun. 282, 341–350. Vennekens, R., Voets, T., Bindels, R. J. M., Droogmans, G., and Nilius, B. (2002). Current understanding of mammalian TRP homologues. Cell Calcium 31, 253–264. Vittitow, J., and Borras, T. (2002). Expression of optineurin, a glaucoma‐linked gene, is influenced by elevated intraocular pressure. Biochem. Biophys. Res. Commun. 298, 67–74. Voets, T., and Nilius, B. (2003). TRPs Make Sense. J. Membr. Biol. 192, 1–8. Voisin, D. L., and Bourque, C. W. (2002). Integration of sodium and osmosensory signals in vasopressin neurons. Trends Neurosci. 25, 199–205. Volkov, A. G., Deamer, D. W., Tenelian, D. L., and Markin, V. S. (1998). ‘‘Liquid Interfaces in Chemistry and Biology.’’ Wiley, New York.
10. The Cytoskeletal Connection to Ion Channels
295
Vriens, J., Watanabe, H., Janssens, A., Droogmans, G., Voets, T., and Nilius, B. (2004). Cell swelling, heat, and chemical agonists use distinct pathways for the activation of the cation channel TRPV4. Proc. Natl. Acad. Sci. USA 101, 396–401. Walker, R. G., Willingham, A. T., and Zuker, C. S. (2000). A Drosophila mechanosensory transduction channel. Science 287, 2229–2234. Wan, X., Juranka, P., and Morris, C. E. (1999). Activation of mechanosensitive currents in traumatized membrane. Am. J. Physiol. 276, C318–C327. Wang, H., B. F. K., Brandon, N. J., Moss, S. J., and Olsen, R. W. (1999). GABA(A)‐ receptor‐associated protein links GABA(A) receptors and the cytoskeleton. Nature 397, 69–72. Watanabe, H., Vriens, J., Suh, S. H., Benham, C. D., Droogmans, G., and Nilius, B. (2002). Heat‐evoked activation of TRPV4 channels in a HEK293 cell expression system and in native mouse aorta endothelial cells. J. Biol. Chem. 277, 47044–47051. Wegner, A. (1976). Head to tail polymerization of actin. J. Mol. Biol. 108, 139–l50. Weeds, A. (1982). Actin‐binding proteins‐regulators of cell architecture and motility. Nature 296, 811–816. Weeds, A., and Maciver, S. (1993). F‐actin capping proteins. Curr. Opin. Cell Biol. 5, 63–69. Welsh, M. J., Price, M. P., and Xie, J. (2002). Biochemical basis of touch perception: Mechaniosensory function of degenerin/epithelial Naþ channels. J. Biol. Chem. 277, 2369–2372. Wes, P. D., Xu, X.‐Z. S., Li, H.‐S., Chien, F., Doberstein, S. K., and Montell, C. (1999). Termination of phototransduction requires binding of the NINAC myosin III and the PDZ protein INAD. Nat. Neurosci. 2, 447–453. Wessells, N. K., Spooner, B. S., Ash, J. F., Bradley, M. O., Luduena, M. A., Taylor, E. L., Wrenn, J. T., and Yamada, K. M. (1971). Microfilaments in cellular and developmental processes. Science 171, 135–143. Wissenbach, U., Bodding, M., Freichel, M., and Flockerzi, V. (2000). Trp12, a novel Trp related protein from kidney. FEBS Lett. 485, 127–134. Wyszynski, M., Lin, J., Rao, A., Nigh, E., Beggs, A. H., Craig, A. M., and Sheng, M. (1997). Competitive binding of a‐actinin and calmodulin to the NMDA receptor. Nature 385, 439–442. Xu, G. M., Gonza´lez‐Perrett, S., Essafi, M., Timpanaro, G. A., Montalbetti, N., Arnaout, M. A., and Cantiello, H. F. (2003). Polycystin‐1 activates and stabilizes the polycystin‐2 channel. J. Biol. Chem. 278, 1457–1462. Xu, J., Tseng, Y., and Wirtz, D. (2000). Strain hardening of actin filament networks. Regulation by the dynamic cross‐linking protein alpha‐actinin. J. Biol. Chem. 275, 35886–35892. Xu, X. Z., Li, H. S., Guggino, W. B., and Montell, C. (1997). Coassembly of TRP and TRPL produces a distinct store‐operated conductance. Cell 89, 1155–1164. Xu, X. Z., Choudhury, A., Li, X., and Montell, C. (1998). Coordination of an array of signaling proteins through homo‐and heteromeric interactions between PDZ domains and target proteins. J. Cell Biol. 142, 545–555. Xu, X. Z., Moebius, F., Gill, D. L., and Montell, C. (2001). Regulation of melastatin, a TRP‐ related protein, through interaction with a cytoplasmic isoform. Proc. Natl. Acad. Sci. USA 98, 10692–10697. Yoder, B. K., Hou, X., and Guay‐Woodford, L. M. (2002). The polycystic kidney disease proteins, polycystin‐1, polycystin‐2, polaris, and cystin, are co‐localized in renal cilia. J. Am. Soc. Nephrol. 13, 2508–2516. Yokoyama, T., Sekiguchi, K., Tanaka, T., Tomaru, K., Arai, M., Suzuki, T., and Nagai, R. (1999). Angiotensin II and mechanical stretch induce production of tumor necrosis factor in cardiac fibroblasts. Am. J. Physiol. 276, H1968–H1976.
296
Cantiello et al.
Yoshida, M., Westlin, W. F., Wang, N., Ingber, D. E., Rosenzweig, A., Resnick, N., and Gimbrone, M. A., Jr. (1996). Leukocyte adhesion to vascular endothelium induces E‐selectin linkage to the actin cytoskeleton. J. Cell Biol. 133, 445–455. Yue, L., Peng, J. B., Hediger, M. A., and Clapham, D. E. (2001). CaT1 manifests the pore properties of the calcium‐release‐activated calcium channel. Nature 410, 705–709. Yuen, E. Y., Jiang, Q., Feng, J., and Yan, Z. (2005). Microtubule regulation of N‐methyl‐D‐ aspartate receptor channels in neurons. J. Biol. Chem. 280, 29420–29427. Zhang, Y., and Hamill, O. P. (2000). On the discrepancy between whole‐cell and membrane patch mechanosensitivity in Xenopus oocytes. J. Physiol. 523, 101–115. Zhelev, D. V., and Hochmuth, R. M. (1995). Mechanically stimulated cytoskeleton rearrangement and cortical contraction in human neutrophils. Biophys. J. 68, 2004–2014. Zhou, C., Ziegler, C., Birder, L. A., Stewart, A. F., and Levitan, E. S. (2006). Angiotensin II and stretch activate NADPH oxidase to destabilize cardiac Kv4.3 channel mRNA. Circ. Res. 98, 1040–1047. Ziyadeh, F. N., Mills, J. W., and Kleinzeller, A. (1992). Hypotonicity and cell volume regulation in shark rectal gland: Role of organic osmolytes and F‐actin. Am. J. Physiol. 262, F468–F479. Zuckerman, J., Chen, X., Jacobs, J., Hu, B., Kleyman, T. R., and Smith, P. R. (1999). Association of the epithelial Naþ channel with Apx and a‐spectrin in A6 renal epithelial cells. J. Biol. Chem. 274, 23286–23295.
CHAPTER 11 Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels Catherine E. Morris and Peter F. Juranka Neuroscience, Ottawa Health Research Institute, Ottawa Hospital, Ottawa, Ontario K1Y 4E9, Canada
I. Overview II. The System Components A. The Channel Proteins B. Bilayer C. Accessory Proteins III. Big Picture Issues A. Bilayer Mechanics and VGCs B. Prokaryotic VGCs as Ancestral Lipid Stress Detectors? C. MS VGCs and MS TRP Channels: Sharing Insights D. No MS ‘‘Motif ’’ Required: Just Say HMMM E. An Imperturbable K‐Selective Pore Surrounded by MS Voltage Sensors? F. Alcohol and VGCs: Binding Sites or Bilayer Mechanics? IV. Reversible Stretch‐Induced Changes in Particular VGCs A. Kv Channels B. Cav and Kv3 Channels Have Similar Stretch Responses C. Cav: L‐Type Channels in Native Preparations D. Nav Channels E. HCN Channels V. Irreversible Stretch‐Induced Gating Changes in VGCs VI. Technical Issues A. Applying a Stretching Force to Study MS Modulation of VGC Activity B. Gadolinium Strangeness VII. Summary Comments References
Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59011-8
298
Morris and Juranka
I. OVERVIEW Membrane stretch modulates the activity of voltage‐gated channels (VGCs). These channels are nearly ubiquitous among eukaryotes and they are present, too, in prokaryotes, so the potential ramifications of VGC mechanosensitivity are diverse. In situ, traumatic stretch can irreversibly alter VGC activity with lethal results (Iwata et al., 2004), but that is pathology. Of wider biological interest is the inherent reversible stretch sensitivity of all VGC subclasses. Evidently, voltage sensor motions feel the impact of bilayer stretch, with the consequence that reversible stretch‐induced changes in ionic current can be seen in recombinant systems for at least some members of the major types of VGC (Kv, Cav, Nav, HCN). Gating current has yet to be tested with stretch, but kinetic dissection of ionic currents in Kv channels shows that the rates of both independent and concerted voltage sensor motions change with stretch, while the quantity of charge that moves stays fixed (Laitko and Morris, 2004; Laitko et al., 2006). In native preparations, the mechanosensitive (MS) transitions of VGCs could contribute physiologically in mechanoelectric feedback (e.g., in cardiac and smooth muscle), neuronal mechanosensing (e.g., osmosensing, tactility), and so on, but reports on stretch eVects in native cells, while tantalizing, are spotty. Experimentally, a fundamental problem is the impossibility of selectively inhibiting the MS portion of a VGC’s response. This chapter deals principally with the reversible responses of VGCs to stretch, with the general relation of stretch stimuli to other forms of lipid stress, and briefly, with some irreversible stretch eVects (¼stretch trauma). A working assumption throughout is that MS VGC motions (i.e., motions that respond reversibly to bilayer stretch) will be susceptible to other forms of lipid stress such as the stresses produced when amphiphilic molecules (anesthetics, lipids, alcohols, lipophilic drugs) are inserted into the bilayer. Insofar as these molecules change the bilayer’s lateral pressure profile (Cantor, 1997, 1999), they can be termed bilayer mechanical reagents (BMRs). Another aim of this chapter is to delineate MS VGC behavior against the backdrop of eukaryotic channels more widely accepted as ‘‘MS channels,’’ namely the TRP‐based MS cation channels (Kung, 2005; Maroto et al., 2005). We start with some ‘‘big picture issues’’ then focus briefly on particular MS VGCs. A few technical items about recording conditions are inserted (some readers may wish to check this first) before a summary comment.
II. THE SYSTEM COMPONENTS Stretch produces both elastic and plastic changes to membrane structures. Figure 1 briefly inventories the membrane constituents relevant to VGC activity and stretch; as applied to membranes, whether they be native or
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels Auxiliary proteins
Extracellular matrix
Lipid microdomain
Membrane skeleton, and so on
4 (×)
S1 S2
Other membrane proteins
S3
S4
S5
S6
S2
S4 S1
a
S4
b a
S2 S1
c
PD
S3
b
S4
VSD
299
S4
S3 S4
S4
S4 S4
S4 c
Expected nature of the lateral pressure profile at a particular "slice" of the channel−bilayer interface
Selectivity filter
FIGURE 1 Voltage‐gated channels in the bilayer. The illustration at the top of this collage serves as a reminder of the complexity of the environment experienced by VGCs in cellular bilayers. There is a large diversity of lipid species and they are not homogeneously disposed in the plane of the bilayer. Moreover, the inner and outer leaflets diVer in their lipid composition. Channels interact with auxiliary proteins and regulatory proteins (e.g., kinases); various other
300
Morris and Juranka
artificial, ‘‘stretch’’ is a fuzzy term for a fuzzy process. On the other hand, stretch happens.
A. The Channel Proteins The molecular architecture and dynamics of the VGCs themselves are now known in bold outline: four somewhat loosely tethered charge‐bearing sensor domains are arrayed around a central domain that houses a pore (selectivity filter and conduction path), with the sensor domains’ response to voltage determining whether gates in the conduction path occlude the pore. The channels’ cruciform perimeter provides a large lateral surface area where many residues of the (four) voltage sensor domains contact bilayer lipids. Even the central pore makes some contact with bilayer lipids (Long et al., 2005a,b; Tombola et al., 2006).
channel and nonchannel membrane proteins are present. These as well as the channels being studied can make long range mechanical connections via membrane skeleton elements and scaVolding proteins (which have their own dynamics). Not depicted are the highly structured invaginations and evaginations of bilayer (e.g., caveolae, microvilli). VGCs are composed of either four subunits (e.g., Kv and HCN channels) or four domains (Nav and Cav channels) of the Shaker‐like six transmembrane structure labeled ‘‘4 S1, S2, S3, S4, S5, S6 (note that S4 is the location of most of the gating charge; note also the location of the S4–S5 linker). The homotetrameric cruciform arrangement of the Kv1.2 channel resolved by Long et al. (2005a,b) is illustrated at left (modified from Tombola et al., 2005). This is thought to be an open‐like structure. Each monomer contributes a voltage sensor domain (VSD) and a pore domain (PD) (essentially S1–S4 and S5–S6, respectively) and the assembled channel exhibits a domain‐ swapped arrangement as shown (three of the VSDs and PDs are labeled a, b, and c). S4 is only partially sequestered by the rest of the protein; the S4–S5 linker ‘‘reaches out’’ along the inner surface of the bilayer (dotted lines) as part of the domain‐swapping arrangement. At far right is a labeled surface presentation of Kv1.2 (from the extracellular side) and to the left of that is a fragment of the same Kv1.2 structure depicted as a ribbon diagram. At bottom is side‐on (ribbon diagram) view of Kv1.2 tetramer; the cruciform channel is sectioned near one of its two widest aspects. The selectivity filter is indicated (dehydrated K ions are present in the filter) and the arrow pointing to it passes through the gating hinge region. At left is an illustration of the lateral forces that the channel in a bilayer would feel. The lateral pressure would be strongly negative where bilayer surface tension is highest (this tends to pull the protein apart) and more weakly positive (but over a wider region) near the bilayer mid region (this tends to compress the protein). Because the channel is cruciform, the lateral lipid–protein interface is extensive. For any given conformation (assuming a given lipid composition) the details of the lateral pressure profile would change continually as an imaginary pressure profile gauge swept out a 360 arc to take in the whole channel. For fully resting and fully activated states, a 90 arc would suYce for a full description, but once a subunit moved independently, the full 360 would be needed.
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
301
B. Bilayer The three‐dimensional structure of the bilayers in which VGCs reside is also understood, although only in broad outline. In the transverse (z) direction, a surface tension (from headgroups squeezing together to prevent exposure of hydrophobic lipid tails to water) at each planar interface exerts large pulling forces on any integral membrane protein (Fig. 1, bottom). These pulling forces extend only several angstroms deep (Cantor, 1999; Gullingsrud and Schulten, 2004) before giving way to counterbalancing compressive forces from the lipids’ acyl chains (which, to maximize their entropy, occupy as much volume as possible). The mid‐bilayer force thus compresses the mid‐ bilayer region of any integral membrane protein. Lipids are asymmetrically disposed in the z axis (e.g., cholesterol is more abundant in the extracellular than the intracellular leaflet) and inhomogeneous in the x–y plane where it subdivides into lipid microdomains of diverse size (usually measured in nanometers in native membranes) and composition, according to the prevailing mix of lipid and protein species (Baumgart et al., 2003; Devaux and Morris, 2004; Gaus et al., 2006; Kahya and Schwille, 2006).
C. Accessory Proteins VGCs also bind directly to auxiliary subunits and directly or secondarily to a large collection of other proteins (e.g., intracellular membrane skeleton and scaVold proteins plus extracellular matrix proteins) (Folco et al., 2004; Wong and Schlichter, 2004).
III. BIG PICTURE ISSUES A. Bilayer Mechanics and VGCs That membrane stretch modulates VGC activity is evident from a glance at Fig. 2 which shows the robust MS responses of Shaker WTIR. Molecularly speaking, this channel is the prototypical VGC. At left, are Shaker currents monitored near the foot of its g(V ) relation (i.e., at a voltage producing a Popen slightly >0). Note that stretch yields ‘‘stretch‐activated’’ (SA) Shaker channel activity just as it yields SA cation channel activity from the oocyte’s endogenous TRPC1‐based (Maroto et al., 2005) cation channels (to record the Shaker currents, these endogenous channels are usually inhibited). The figure also shows that the increase of IShaker in response to stretch is reversible, repeatable, and dose‐dependent. What does this
302
Morris and Juranka −10 mmHg
1.3 pA 750 ms 2 pA 50 ms
Before Xenopus oocyte MS cation (TRPC1-based) MS channel)
During −8 mmHg (stretch) After Increasing membrane tension Step 1
Shaker WTIR (Kv1 channel)
1 pA 0.5 ms
Step 2
0
−5 150 ms
−10 −15 −20
−25
−30 −35
−40
300 pA
−12 mmHg
FIGURE 2 The prototypical Kv1, Shaker, is an MS VGC. Top left, typical MS cation channel activity from the TRPC1‐based channels endogenous to oocytes (shown at low and high time resolution). Below that, unitary current recordings of Shaker WTIR. It too is an SA channel. At right are data sets (two diVerent patches) showing that Shaker WTIR responses to stretch are not only reversible but repeatable, and that they show dose dependence. Modified from Gu et al. (2001).
general result signify in the wider context of bilayer mechanics and VGC conformation changes? In principle, if bilayer stretch regulates VGC activity, then so should changes in bilayer lipid composition, since lipid molecules have particular bilayer mechanical properties depending on their size, shape, internal flexibility, and charge. A thorough review of this topic appeared (Tillman and Cascio, 2003). Table I provides a sampling of reports of amphiphile (including membrane lipids) modulation of VGC kinetics. No theoretical framework has emerged for such observations, which is more compelling than that of lipid stress perturbations of the conformational equilibria of the channels. As Andersen and colleagues point out (Lundbaek et al., 2004), specific lipid– protein interactions seldom appear to be involved in lipid actions of VGC gating, but rather, ‘‘hydrophobic coupling between a membrane‐spanning protein and the surrounding bilayer means that protein conformational changes may be associated with a reversible, local bilayer deformation . . . the energetic cost of the bilayer deformation contributes to the total energetic cost of the protein conformational change.’’ This group deploys gramicidin A as a molecular force‐transducer for bilayer mechanics. They then alter the composition of human embryonic kidney (HEK) cell bilayer via micelle‐forming amphiphiles (e.g., TX100), cholesterol depletion (Lundbaeck et al., 2004), or addition of the amphiphilic drugs capsaicin and capsazepine (Lundbaeck et al., 2005).
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
303
TABLE I A Sampler: Modulation of VGCs by Bilayer Mechanical Reagents (BMRs) Cholesterol
Kv1.3 (Hajdu et al., 2003) HCN4 (Barbuti et al., 2004) Nav1.4 (Lundbaek et al., 2004) Cav (L‐type) (Toselli et al., 2005)
Propofol
HCN2 (Ying et al., 2006) Nav (neuronal) (Ouyang et al., 2003) Kv (T‐lymphocytes) (Mozrzymas et al., 1996) Cav (neuronal L‐type) (Olcese et al., 1994)
Phenothiazines Chlorpromazine
Cav (a1E) (McNaughton et al., 2001) Nav (cardiac) (Ogata and Narahashi, 1989) HERG (Thomas et al., 2003)
Trifluoperazine Mesoridazine
Kv1.3 (Teisseyre and Michalak, 2003) HERG (Su et al., 2004)
Volatile anesthetics Isoflurane
Kv1 (Shaker) (Correa, 1998) Nav1.2, Nav1.4, Nav1.6 (Shiraishi and Harris, 2004) Cav (cardiac L‐ and T‐type) (Camara et al., 2001) hIK1 (Namba et al., 2000)
Halothane
HERG (Li and Correa, 2002) Cav (L‐type) (Kamatchi et al., 2001) HCN1, HCN2 (Chen et al., 2005)
1‐Alkanols
Kv3 (Shahidullah et al., 2003) BKCa (Chu and Treistman, 1997) HCN (neuronal Ih) (Okamoto et al., 2006)
Fatty acids
BKCa channels (Clarke et al., 2003) Nav1.5 (Xiao et al., 2006)
Cannabinoids
Nav (neuronal) (Nicholson et al., 2003)
HERG (Guizy et al., 2005) Kv1.2 (neuronal) (Poling et al., 1996) P‐type Ca channel (neuronal) (Fisyunov et al., 2006) T‐Type Ca channel (Chemin et al., 2001) Capsaicin and capsazepine
HCN1 (Gill et al., 2004)
Lysophospholipid
Nav1.4 (Lundbaek et al., 2005). HERG (Wang et al., 2001)
Propanolol
HERG (Yao et al., 2005)
Ceramide
Kv1.3 channels (Bock et al., 2003)
Nicotine
Nav1.5 (Liu et al., 2004)
Resveratrol
Nav1.5 (Wallace et al., 2006)
HCN (cardiac If) (Hu et al., 1997)
304
Morris and Juranka
Using this approach in conjuction with a recombinant skeletal muscle Nav channel, they have made a solid case that altered bilayer elasticity, or a parameter tightly correlated with bilayer elasticity, is what underlies the resulting kinetic modulation of the VGC. If bilayer elasticity regulates the conformational equilibria in VGCs, then so should bilayer stretch. And as we already saw (Fig. 2), it does. The promiscuous eVects of a vast number of amphiphilic drugs on VGC kinetics and the promiscuous kinetic eVects of stretch on many of these same channels almost certainly share bilayer mechanical origins.
B. Prokaryotic VGCs as Ancestral Lipid Stress Detectors? In eukaryotic Kvs structurally similar to the prokaryotic Kv channel, KvAP, bilayer stretch increases the probability of voltage sensor motion (Laitko and Morris, 2004). Electrical propagation over a long distance is clearly not what prokaryotic VGCs do for a living, but responding to bilayer stress might be part of their evolutionary raison d’etre. Osmotic safety valve channels detect near lytic bilayer stress in prokaryotes (Kung, 2005); less drastic lipid stress perturbances may be the province of MS VGCs. Benthic vent organisms (like the archaeon from which KvAP was isolated) undoubtedly lead particularly unquiet lives with respect to ambient pressure, temperature, and osmotic stress (Sako et al., 1996) all of which will aVect the physico‐mechanical characteristics of bilayers and hence the behavior of MS VGCs. As gauged from crystal structures (Lee et al., 2005) and EPR spectroscopy‐ based structures (Cuello et al., 2004), the thickness and elasticity of the bilayer immediately adjacent to KvAP voltage sensors might regulate the ease of voltage sensor movements. Thickness and elasticity are physical attributes of the bilayer that would vary in prokaryotes, experiencing pressure and temperature variations, osmotic stress, lipid metabolism (i.e., lipid substitutions in the bilayer), cell division, and even (for a rod‐shaped organism) in the tubular vs hemispherical sectors of the cell. Pressure, temperature, osmotic stress, and lipid composition: all these factors aVect VGC kinetics as an example or two for each illustrates. Pressure: Hyperbaric pressure causes bilayer acyl chains to pack more densely, and this straightens them, thickening and rigidifying the bilayer (Scarlata et al., 1995). Hyperbaric pressure alters the gating of Nav and Kv channels in squid axon and in vertebrate nodes of Ranvier (Conti et al., 1984; Kendig, 1984), Kv1 channels in oocytes (Meyer and Heinemann, 1997) and N‐type Cav channels (Etzion and Grossman, 2000). Since bilayers are more compressible than proteins, hyperbaric pressure presumably acts at least in part, via lipid stress (see discussion in Gu et al., 2001), but even in a review (Macdonald, 2002)
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
305
the focus is on putative eVects of hyperbaric pressure on protein volume without consideration of possible modulatory alterations at the protein–lipid interface. This view is probably due for some revision. Temperature: In contrast to elevated pressure, elevated temperature thins and fluidizes membranes (Pencer et al., 2005). Arrhenius plots for BKCa kinetics with or without cholesterol (artificial bilayers) are consistent with a model in which cholesterol exerts its modulating eVect via lipid stress (Chang et al., 1995). In a similar line of reasoning, Kv VGC sensor motions with the highest Q10 values (i.e., closed–closed transitions far from the open state; Rodriguez et al., 1998) might, therefore, reflect temperature‐induced changes in the microstructure of the lipid bilayer that alter lipid stresses felt by the channels, and in particular, by voltage sensors in the resting state. Osmotic stretch: VGC activity varies with both membrane stretch and osmotic swelling, and for Cav channels the stretch and swelling responses are similar in detail (see discussion and references in Calabrese et al., 2002). Lipid‐dependent bilayer mechanics: Elevated cholesterol (e.g.) increases bilayer rigidity and thickness (Pencer et al., 2005; Czub and Baginski, 2006); enriching or depleting membrane cholesterol modulates various transitions (activation, slow inactivation, and/ or deactivation) in Kv1.3 and HCN4 channels and shifts inactivation in Nav1.4 (Lundbaek et al., 2004). The responses of archeabacterial VGCs in native lipids (these are more ordered, less flexible than eukaryotic bilayer lipids; Bartucci et al., 2005) to voltage and lipid stress perturbations have not been tested. Assuming that prokaryotic VGCs are responsive to bilayer mechanical perturbations (stretch, hyperbaric pressure, BMRs), then bilayer mechanics need to be included along with transbilayer voltage, in the constellation of factors responsible for the evolution of VGCs. Interestingly, displacement currents of prestin, a voltage‐sensitive protein unrelated to the VGCs, are tension sensitive (Dong and Iwasa, 2004); perhaps it is hard to build a voltage‐dependent membrane protein that is not susceptible to lipid stress. Membrane proteins with recognizable ‘‘S4’’ voltage sensors but no pore domains [a phosphatase (Murata et al., 2005) and a proton channel (Ramsey et al., 2006)] are now known; perhaps they too will prove sensitive to lipid stress. Certainly, volatile anesthetics, which are BMRs, modulate Kv channel gating currents (Correa, 1998).
C. MS VGCs and MS TRP Channels: Sharing Insights MS channels, including VGCs, generate both SA and stretch‐inactivated (SI) currents (Fig. 3). The first reported SI current (it was a non‐VGC) occurred in neuronal patches that also had SA channel activity (another
306
Morris and Juranka Stretch activation (V m = −30 mV)
A
4 pA 0 −5 −8
−13 −17 mmHg −20 0 B
Stretch inactivation (V m = +20 mV) 0
−10 −20 −30 −40 −50 mmHg 0
Same patch, stretch activation (V m = −10 mV) 4 pA
−50
FIGURE 3 Shaker WTIR is an SA and SI channels. (A) A patch held at a voltage near the foot of the Shaker g(V ) relation (in the presence of gadolinium) is subjected to successively increasing pipette suction (i.e., greater membrane tension) and then back to resting tension. Dose‐dependent stretch activation results. (B) Another patch held at a more depolarized voltage (which favors slow inactivation) undergoes stretch inactivation. Then, going to 10 mV (i.e., near the foot of the g(V ) for this patch), the same population of channels exhibit stretch activation (compare the eVects of 50 mmHg at the two diVerent voltages). Modified from Gu et al. (2001).
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
307
non‐VGC) (Morris and Sigurdson, 1989) and with their diVerent unitary conductances, these two were probably distinct channel types. But that need not be the case when SA and SI channels activity coexists. In skeletal muscle cells of the mdx (dystrophic) mouse (Franco‐Obregon and Lansman, 2002), traumatic stretch irreversibly transforms MS channels from SA to SI channels. How this might come about is suggested by reference to our somewhat deeper understanding of reversible and irreversible responses to stretch in VGCs. MS cation channels like those of mdx muscle are thought to be TRP‐based (Maroto et al., 2005) and TRP channel structure is globally similar to Shaker structure (Clapham, 2003). The kinetics of the MS VGC, Shaker, are more tractable than those of TRP channels, so insights relevant to the mdx channel ‘‘transformation’’ might emerge from Shaker’s SA/SI behavior. TRP channels share with this Kv channel a ‘‘6TM‐tetramer’’ body plan and they can be voltage‐dependent (Brauchi et al., 2004; Nilius et al., 2005). The skeletal muscle MS ‘‘TRP’’ cation channel has a stretch‐sensitive closed–closed transition that is weakly voltage‐dependent (Guharay and Sachs, 1984, 1985); Kv channels have voltage‐dependent closed–closed transitions that are stretch‐sensitive (Laitko and Morris, 2004; Laitko et al., 2006). And—critical point—a Kv channel can generate both SA and SI currents (Gu et al., 2001), echoing the case of the mdx skeletal muscle channels. What is the nature of this dual SA and SI behavior in Kv1 channels? Moving between closed, open, and slow‐inactivated states, Kv1 channels undergo several voltage‐dependent transitions (the voltage sensor undergoes outward movement with respect to the electric field) plus some voltage‐independent transitions (mode‐shift). In Shaker mutants with well‐characterized rate‐ limiting transitions, we isolated 1. ‘‘SA transition’’ (i.e., a transition whose response to stretch enhances IKv) and 2. ‘‘SI transitions’’ (i.e., a transition whose response to stretch diminishes IKv). The SA transition is activation—an outward motion of sensor charge, independent in each of the tetrameric channel’s subunits (Laitko and Morris, 2004). The SI transitions are a preopening step and a slow inactivation step (preopening is concerted, involving all four subunits, slow inactivation is less well understood; Laitko and Morris, 2004; Laitko et al., 2006). We can take Shaker 5aa as an example of an SA/SI Kv channel. In 5aa, activation and slow inactivation accelerate with stretch, so 5aa channels subjected to depolarization plus stretch exhibit SA IKv followed by SI IKv. This dual SI and SA behavior in a Kv illustrates that having (at least) one open and two closed conformations allows a channel to exhibit both SA and SI behavior.
308
Morris and Juranka
Taken alone, however, this dual behavior does not illuminate the irreversible (posttraumatic stretch) switching of mdx MS channels. For that, we draw on the irreversible eVect of traumatic stretch characterized in recombinant Nav1.4 a‐subunit channels (briefly, strong membrane stretch irreversibly leads to accelerated, left‐shifted activation and inactivation kinetics, the underlying molecular explanation is unknown (see the section on irreversible stretch changes, below). MS TRP channels in mdx muscles presumably experience comparable irreversible stretch‐induced change in the basic rates governing transitions among (at least) three states, while retaining their inherent (reversible) MS responses. This simple scenario could explain an irreversible ‘‘SA‐to‐SI’’ switch. The Kv analogy substitutes an ‘‘irreversible voltage change’’ (¼clamping at two diVerent voltages) for an ‘‘irreversible mechanical change’’ (¼stretch trauma, as in the mdx TRP or the Nav1.4 a‐subunit channels) as the factor controlling the characteristic rates of the system before stretch is applied to examine the reversible MS responses. Clamped at well‐separated voltages, Shaker WTIR, for example, shows SA at one voltage and SI at the other (Fig. 3). The analogy assumes that the TRP channels have (at least) three states and two MS transitions, with diVerent transitions being rate‐limiting in the conditions prevailing before and after the stretch trauma, to yield net SA behavior before and net SI behavior after. As a generic phenomenon (single population of channels with at least two MS transitions, with basic rates governed by diVerent factors at diVerent times), this might explain reports of the MS cation channels from a given preparation being variously SA channels or SI channels (Kirber et al., 1988; Hisada et al., 1993).
D. No MS ‘‘Motif ’’ Required: Just Say HMMM During activation, Shaker WTIR behaves as an SA channel and Shaker ILT behaves as an SI channel (Fig. 4A and C), yet except for three conservatively mutated neutral residues in the S4 voltage sensor, these channels are identical. Should we label these residues a ‘‘stretch motif’’? Of course not. Should we assume the three residues contact the bilayer? Again, no. These two Shaker variants behave as SI or SA channel by virtue of a. the identity of their rate‐limiting voltage‐dependent (and stretch‐ sensitive) steps and b. the particular eVect of stretch on that step. Stretch, it happens, decelerates the rate‐limiting voltage‐dependent step of Shaker ILT. By contrast, it accelerates the (diVerent) rate‐limiting voltage‐dependent step of Shaker WTIR.
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels Ai
B
Shaker WTIR/Gd3+ 10 mV
−20 mV
−40 mV
1 nA
25
10 ms
−10 mV
Shaker 5aa
309
1 nA
1000
10 ms
Aii
Current (pA)
0 50 mV
30 mV
0 mv
40 mV
Shaker WTIR −25 mV −30 mV 5 ms 20 mV
1 nA
5 ms
−20 mV
C
1500
200 pA
−15 mV
−10 mV
1500
*
500 pA
* * 80 mV
0
200 ms
D 50 ms
0
Shaker ILT
0
50 pA
Time (ms) 0
STRETCH Before During After
Aiii −25 mV
0
0.5 nA
120 mV
100 mV 0
200 ms
0
"Hidden" in ILT V (C CA) V (C CA) V V CA) C4A (C V CA) (C
Rate-limiting in 5aa
100
200 ms
(C4AP
Token O)
ILT
FIGURE 4 SA of Shaker WTIR and Shaker 5aa, SI of Shaker ILT. (Ai) Shaker activation before, during (dotted lines), and after stretch in the presence of gadolinium. (Aii) Likewise in another patch but with no gadolinium added. (Aiii) There is a narrow range of voltage below the foot of the g(V ) relation where no current flows unless the membrane is stretched (in this patch, that range included 20 mV) (from Tabarean and Morris 2002). (B) For Shaker 5aa, in this example, the comparable range extended at least from 40 to 20 mV. With larger depolarizations, it is clear that stretch (gray traces) accelerates both activation and slow inactivation (from Laitko and Morris 2004). (C) For Shaker ILT, stretch (asterisks) decelerates activation. The consequence is a right shift of the g(V ) curve with no change in the amount of charge moved; for Shaker WTIR, by contrast, stretch left shifts the g(V ) (from Laitko et al., 2006). (D) A kinetic scheme for Shaker channels (Laitko et al., 2006). The major motions of the voltage sensor are independent and then resolve in a concerted final voltage‐dependent motion, at which point an additional concerted motion opens the channel. For current turn on (the Markov model ‘‘token’’ in our experiments) during depolarizing step, the independent motion is slow enough in Shaker 5aa to be the rate‐limiting step, whereas in Shaker ILT the concerted voltage‐ dependent motion is rate‐limiting. In Shaker WTIR, the two transitions have comparable rates over a wide voltage range. The independent sensor motions are ‘‘hidden’’ in Shaker ILT when ionic current is the token, although not when gating current is the token. Modified from Laikto et al. (2006).
310
Morris and Juranka
This motif‐free way of describing SA and SI gating could be called ‘‘HMMM’’ for Hidden Markov Model Mechanosensitivity (hidden Markov model theory is well described in Wikipedia). Ion channels kinetic schemes are Markov models; formally, in a Markov model, all states are directly visible to the observer, so state transition rates (probabilities) are the only parameters that need to be considered. In a Hidden Markov Model, by contrast, some states are not directly visible, but—and this is the critical point—variables influenced by those states are visible. In reality, this applies for most kinetic analyzes of ion channel behavior. If the MS activity of a VGC (or a ligand‐gated channel, and so on) was monitored by by by by
optically determined protein motions (e.g., IsacoV) and chemically determined protein motions (e.g., Yellen) and gating currents (Ledwell and Aldrich, 1999) and ionic current,
our understanding of which motions are ‘‘stretch‐sensitive’’ (in say, Shaker WTIR or Cav or Kv3 channels) would not be restricted (as currently) to motions limiting for ionic current flow. In a multistate channel, when ionic current alone is used as the token (‘‘token’’ is Hidden Markov Model terminology) for MS channel activity (or for ‘‘molecule‐X’’‐modulated activity), a picture of reduced dimensions emerges. Like the two‐dimensional shadow of, say, a rotating three‐dimensional helix, this picture can be misleading. Stretch globally alters bilayer structure (Gullingsrud and Schulten, 2004), so it will globally aVect the membrane‐embedded regions of membrane proteins. Even for an ideal two‐state MS channel, multiple aspects of the bilayer–channel interaction (Wiggins and Phillips, 2005) would, therefore, contribute to the free energy of MS gating. Mutations enhancing MS gating could be said to belong to ‘‘an MS motif,’’ but in all likelihood such residues would be found scattered about the protein as in the case of MscL gain‐of‐ function mutations (Ou et al., 1998). Given the multiple sources of free energy itemized by Wiggins and Phillips (2005), a ‘‘global MS motif’’ makes sense but ‘‘global motif’’ seems almost oxymoronic. For any given VGC, speculations about what structural features cause stretch to accelerate transition X and decelerate transition Y would be idle. However, it may eventuate that the independent depolarization‐induced transitions of all VGCs (these transitions move voltage sensor charge outward—for example activation in Shaker, deactivation in the sea urchin HCN channel (Mannikko et al., 2002) accelerate with stretch (and slow with hyperbaric pressure and cholesterol). Our preliminary evidence on HCN channels (Lin et al., 2007) would support this view. If so, then some robust property of the independent voltage sensor motion, a property retained in all
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
311
VGCs, may underlie a universal lipid stress eVect on that motion. Two (not mutually exclusive) possibilities are: 1. The outward movement of voltage sensor charge requires a locally thinned membrane. 2. Expansion of the sensor array in the plane of the bilayer (Tabarean and Morris, 2002) occurs during outward movement of voltage sensor charge. In Shaker WTIR and Shaker ILT, overlapping Q(V) curves (Ledwell and Aldrich, 1999) show that the activation motions are not diVerent. Yet WTIR is SA and ILT is SI. Why? Because the next step, after activation, toward the open state—a concerted voltage‐dependent motion—is fast in WTIR but ultraslow (and rate‐limiting) in ILT. And in ILT, it turns out that stretch retards that concerted motion (Laitko et al., 2006). It presumably does so too in WTIR, but evidently not enough to slow the onset of the token we monitor, namely ionic current. Because ionic current is our only monitored parameter, a reduced dimension picture emerges—in other words, we are dealing with HMMM. It is self‐evident that all conformation changes are inherently thermal and thus, to some degree, temperature‐sensitive. In the same way, any conformation changes occurring in the plane of the bilayer must be inherently lipid stress‐sensitive. Special ‘‘MS motifs’’ would be needed only for specialized MS tuning (e.g., for directional or frequency sensitivity or to create a threshold at a specific tension). These same conformation changes (in the plane of the bilayer) should also be susceptible to BMRs (i.e., molecules that change the shape of the lateral pressure profile). No specific ‘‘BMR’’‐binding site (‘‘motif’’) in the protein would be required (except, again, for very special cases where the lipid can be shown to be a ligand for a specific part of the channel such as the domain responsible for fast inactivation; Oliver et al., 2004). And what if Kv channel‐X but not Kv channel‐Y responds to, say, butanol or cholesterol? The parsimonious assumption is that lipid stress‐sensitive rate‐limiting transitions in a HMM dominate the measured response, not that the response is dominated by the presence or absence of butanol or cholesterol or so on binding sites. In other words, the parsimonious assumption invokes a lipid stress version of HMMM. In summary, in Kv channels, SI and SA eVects are not about ‘‘MS motifs’’ but rather, they reflect the stretch‐perturbed operation of particular multistate kinetic schemes. It is likely that this assertion applies, too, for MS TRP channels. The same line of reasoning says that for amphiphilic molecules (BMRs) exerting actions on VGCs via lipid stress, it is probably not ‘‘binding motifs’’ that underlie the action but multistate kinetic schemes with ‘‘hidden’’ transitions. We return to this in the alkanol section below.
312
Morris and Juranka
E. An Imperturbable K‐Selective Pore Surrounded by MS Voltage Sensors? The voltage‐dependent conformation changes of Kv proteins are susceptible to membrane stretch but the selectivity filter at the heart of the assembled Kv tetramer seems inured to stretch. This can be said for a Kv1 (ShakerWTIR and Shaker ILT), Kv3 (Shaw) (Fig. 5), and BKCa channels (Dopico et al., 1994) as well as for Cav channel (Fig. 6). The evidence: (1) when NPopen increases with stretch in Shaker WTIR, in the Shaw channel, or in BKCa, single‐channel amplitude with or without stretch is identical (Dopico et al., 1994; Gu et al., 2001; Laitko et al., 2006) and (2) in both Shaker WTIR and ILT, g(V)max is unaVected by stretch.
SHAW2 F335A
Ai During
2
During
Aii
Difference current at: −25 mmHg
120 mV 1 nA 20 ms
After Before
−20 mmHg
1
−15 mmHg 0
−100 −40
0.5 nA
50 mV −100 0 0
1000
2000
3000
B/D/A −30 mmHg, 0 mV, 10 runs Mean current (pA)
Bi
Bii
−10 mmHg
1s Before
15 During
10 After Before
5 100 ms
20 pA
Biii Before Counts
During During
50 ms
After 10
20
30
40
(pA)
After
FIGURE 5 A Kv3 channel current is enhanced by stretch. This same Kv3 channel is inhibited by alkanols (Shahidullah et al., 2003). Alkanols lower surface tension whereas stretch increases it, but it is not yet clear if these surface tension eVects converge on the same transition. Nevertheless, as the figure shows, stretch reversibly (Ai) and in a dose‐dependent manner (Aii) increase the channels activity. Recordings at 0 mV (Bi) rule out any possibility of interference from the endogenous MS channel. Unitary currents at 0 mV before, during, and after stretch (Bii) show no evidence for increased unitary current (Biii) during stretch. Modified from Laitko et al. (2006).
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
313
N-type Ca channel Cell-attached, a-subunit only Control
Stretch
Recover
Stretch
1 pA 20 ms
0
1
0
1
0
1
0
1
Whole-cell recording
−30 mV
10 mV
Control
Stretch FIGURE 6 Cav channel activity increases reversibly with stretch. Single‐channel (a‐subunit only) and whole‐cell recordings (full complement of subunits) of recombinant N‐type (high voltage activated) Cav channels illustrate that stretch increases the peak and steady state current levels and accelerates inactivation from the open state but has no evident eVect on unitary current and no eVect on the rate of current onset. Modified from Calabrese et al. (2002).
314
Morris and Juranka
Perhaps the circular domain‐swapping arrangement noted by Long et al. (2005b) in the Kv1.2 tetramer contributes to the mechanical stability (Riechmann et al., 2005) of Kv selectivity filters. Other eukaryotic MS channels (e.g., two‐pore‐domain‐K and TRP channels) have stretch‐insensitive unitary conductances; since the TRPs are 6TM channels, it will be interesting to learn if they too have Kv‐like domain‐swapping arrangements. In planar bilayers, BKCa channel activity is modulated by BMRs without changing unitary conductance since cholesterol reduces BKCa activity (Chang et al., 1995) and this is antagonized by ethanol (Crowley et al., 2003). Since cholesterol and ethanol have opposing eVects on surface tension (and line tension) in bilayers, these certainly look like lipid stress‐mediated phenomena. In oocyte membrane, the BMRs ethanol, butanol, hexanol, and heptanol reversibly enhance BKCa current (Chu and Treistman, 1997). They do not, however, detectably perturb the selectivity filter (Chu et al., 1998). But under some conditions, the pathway in series with the selectivity filter—or perhaps even the selectivity filter itself—may succumb to lipid stress. Chang et al. (1995) detected a small (5%) decrease in BKCa channel conductance when cholesterol was increased by 10%. Moczydlowski and colleagues studied BKCa in a barebones artificial bilayer regime (Park et al., 2003) by way of retesting their own longstanding hypothesis that anionic lipid enhances BKCa unitary conductance via surface charge eVects. The original hypothesis did not hold, and they now postulate that ‘‘lipid modulation of Kþ conductance is preferentially coupled through conformational changes of the selectivity filter region . . . [and they] . . . discuss possible mechanisms for the eVect of anionic lipids in the context of . . . general membrane physical properties proposed to regulate membrane protein conformation via energetics of bilayer stress.’’ Ultrasimple artificial bilayers, having fewer degrees of freedom for energy minimization, may allow some amphiphiles more latitude to exert pressure eVects at the channel–protein interface than when present at the same concentration in complex bilayers.
F. Alcohol and VGCs: Binding Sites or Bilayer Mechanics? Alkanols partition into the bilayer headgroup region, being stabilized by hydrogen bonding to carbonyl oxygens of the phospholipid’s glycerol backbone; partitioning increases with increased chain length (Feller et al., 2002). The impact of short chain alkanols on the bilayer lateral pressure profile can be powerful enough to destabilize the tetrameric arrangements of ion channels (Van den Brink‐van der Laan et al., 2004). A short chain alkanol series has become an even better experimental tool for bilayer mechanics now that Ly and Longo (2004), using pipette aspiration of giant unilamellar
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
315
vesicles, confirmed experimentally for bilayers a fact long established for monolayers . . . Area compressibility, which is proportional to interfacial (¼surface) tension, follows Traube’s rule of interfacial tension reduction: for each additional alcohol CH2 group the concentration required to reach the same area compressibility falls 3‐fold.
Whereas stretch increases bilayer surface tension (Gullingsrud and Schulten, 2004), alkanols lower it. Stretch and alkanols are similar, however, in that both thin the bilayer and decompress its acyl chains. To monitor reversible modulation of VGCs by stretch or by ‘‘BMR X,’’ currents obtained ‘‘before, during, and after’’ exposure are needed and while this is readily achieved for membrane patches subjected to stretch, this is not so for slowly exchanging BMR molecules like cholesterol. Short chain alcohols, by contrast, are experimentally tractable, as attested by the Covarrubias group’s concentration jump experiments on Kv3 channels (Shahidullah et al., 2003). Strikingly, this group interprets their extensive data as evidence that bilayer mechanics do not mediate alkanol eVects on channels. Before looking at that work, therefore, it seems germane to look at alkanol actions on two non‐VGC proteins. The first is rhodopsin (a prototypical G‐protein– coupled receptor). In rod outer disk segments, Mitchell et al. (1996) monitored eVects of an alkanol series on a photoactivatable rhodopsin conformation change and found the pattern of altered rates conforms to a model of lipid‐mediated action (in eVect, as if Traube’s rule is the operative factor). Their rhodopsin studies with other BMRs also support an interpretation of lipid mediation, as do their findings (a) that at physiologically relevant levels of osmolality and ethanol, 90% of ethanol’s eVect arises from disordered acyl chain packing (Mitchell and Litman, 2000) and (b) that cholesterol and alkanols have opposite‐going eVects, indicative of reduced acyl chain packing free volume, not of specific cholesterol–rhodopsin interactions (Niu et al., 2002). For rhodopsin reconstituted into vesicles (rhodopsin/lipid ratio from 1:422 to 1:40), they reported that elevated rhodopsin‐packing density minimally impacts rhodopsin’s structural stability yet markedly reduces its activation (Niu and Mitchell, 2005). Extrapolated to VGCs, such a scenario would predict ‘‘lipid tuning’’ of channel dynamics in densely packed lipid rafts and ethanol interference with that tuning (Crowley et al., 2003). This scenario is appealing, obviating as it does, the need to postulate hosts of VGC‐specific binding sites for various lipophilic reagents. Nonetheless, the view that alkanols act on channels (ligand‐gated channels as well as VGCs) via protein sites has considerable currency. The second system to consider, then, is the recombinant glycine
316
Morris and Juranka
receptor channel for which Davies et al. (2004) find that ethanol and butanol actions are antagonized by hyperbaric pressure. The word ‘‘bilayer’’ does not appear in the chapter; results are interpreted as evidence that alkanols target a pressure‐antagonism‐sensitive mechanism in glycine receptors, with ‘‘the mutant a1(A52S) GlyR findings suggest[ing] . . . the N‐terminus as a potential target for ethanol action.’’ The ionic current (their ‘‘token’’: see HMMM section above) alteration in a point‐mutated channel is seen as evidence that the mutated residue is part of a ‘‘motif’’ (¼binding target). By analogy to the SA–SI story for Shaker WTIR and Shaker ILT, a more parsimonious possibility would be that these ligand‐gated channels have hidden kinetic processes, and that lipid stress acts on a hidden transition(s) in a way that aVects the token (ionic current). In other words, this has all the hallmarks of HMMM creating the illusion of a motif (¼binding target). With these alkanol stories in mind, we return to Kv3 (Shaw2) channels. Shortly after the molecular cloning of Shaker, Shaw, Shal, and Shab, in a seminal paper, Covarrubias and Rubin (1993) wrote: There is presently a debate regarding the relative merits of lipid‐based and protein‐based theories of anesthesia and the action of ethanol . . . of four structurally homologous cloned Kþ channels . . . only the Shaw2 channel . . . is rapidly and reversibly blocked by ethanol in a concentration‐dependent manner . . . [this] . . . can be explained by assuming a bimolecular interaction between ethanol and the channel. . . . also . . . [these] channels were selectively blocked by halothane (1 mM). Our results support the ‘‘protein hypothesis’’ of ethanol and anesthetic action.
Others (Chu and Treistman, 1997) found that octanol and decanol potently inhibit Shaw2 but not BKCa and took this channel specificity as further support for the protein theory. When Correa (1998) tested volatile anesthetics on Shaker‐gating current (a diVerent ‘‘token,’’ note, than ionic current), she found that Shaker does, after all, respond to halothane and other BMRs (alkanols were not tested). The steps closest to opening (the steps dominating Shaw2 ionic current onset) are most aVected, and this she took as support for direct protein–BMR interactions. It is now known, however, from the Shaker ILT response to stretch (Laitko et al., 2006), that lipid stress acts on these preopening voltage‐dependent steps. Since Shaw2 kinetics were thought to approximate a ‘‘two‐state’’ situation (Shahidullah et al., 2003), our view was that if BMR actions on Shaw2 are lipid‐mediated, stretch and alkanol eVects could be directly compared. Clearly, Shaw2 is an MS channel; during depolarizing steps, it is an SA channel (Fig. 5). However, Shaw2 kinetics are not ‘‘two‐state’’ and compounding that, the dominant MS transition is not rate‐limiting for activation (Laitko et al., 2006). Consequently, direct comparisons of the stretch (which
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
317
thins the bilayer, lowers mid‐bilayer lateral pressure, and increases the surface tension) with the bilayer mechanical eVects of short chain alcohols (which share the first two eVects, but decrease the surface tension) were not feasible. What is clear, however, is that alkanols decrease steady state IKv3 and stretch increases it. If both outcomes do arise from lipid stress, one possibility is that high surface tension (from stretch) slows Kv3 pore closing and low surface tension (from alkanols) accelerates the same transition. Unfortunately, eVects of stretch on Kv3 pore closing still need to be tested; we were discouraged by finding that tail current rates tested before, during and after stretch in other Kvs are oocyte batch dependent (Laitko et al., 2006). Nevertheless, a direct slowing action of stretch on Kv3 pore closing is one possible outcome that could reconcile a bilayer mechanical model with the Covarrubias ‘‘protein’’ model (Harris et al., 2000). In that model, alkanols stabilize the Kv3 channel‐closed conformation by a direct interaction at a crevice formed by a 13‐amino acid cytoplasmic S4–S5 loop. Covarrubias’ group has provided several strong lines of evidence supporting this view. They monitored steady state current during fast (1 ms rise time) concentration jumps on excised patches (Shahidullah et al., 2003), finding that ‘‘on binding’’ rates (i.e., onset of inhibition of steady state ionic current) changed 3‐fold per alkanol carbon (they took this to reflect ‘‘productive collisions at an alkanol binding site’’). Butanol’s inhibitory action is >1000 faster from the intracellular than the extracellular face (Harris et al., 2000), a result seen as reflecting a cytoplasmic binding site (an alternate view is that the asymmetrically lowered surface tension can perturb the lateral pressure profile to favor a closed state). Circular dichroism spectroscopy on a peptide that forms the S4–S5 loop (a diVerent token from the ionic current data) shows alkanol chain length‐dependent binding (Shahidullah et al., 2003). Chimeras of alkanol‐sensitive and an alcohol‐insensitive Kv3s and site‐directed mutagenesis data further implicate the S4–S5 loop in the inhibition of channels by alkanols. Strikingly, a proline to alanine point mutation in the S6‐gating hinge region changes the Kv3 channel from one whose steady state current decreases with butanol to one whose steady state current increases with butanol (Harris et al., 2000). Recalling the Shaker WTIR/ILT story (a three‐point mutation changes the channel from SA to SI), we wonder if this Shaw2 mutant is hinting at a lipid stress‐mediated example of HMMM. Alkanols reduce the line tension of a bilayer pore (Dan and Safran, 1998; Ly and Longo, 2004) and thus, for any irregular‐shaped channel, they inevitably perturb the lateral pressure profile (Wiggins and Phillips, 2005). As we pointed out (Laitko et al., 2006, Fig. 9C), the diVusional mobility of certain BMRs (including an alkanol) may drop at a channel’s perimeter, and
318
Morris and Juranka
if so their eVects on the lateral pressure profile could exceed expectation (i.e., expectation from the BMR’s line tension eVect). Whenever a BMR’s concentration at the perimeter of channels exceeds the ‘‘bulk’’ bilayer concentration, moreover, distinctions between ‘‘low aYnity BMR‐binding sites’’ vs lipid stress‐mediated BMR eVects could be largely semantic. In the Kv3 channel, Shaw2, the location of the S4–S5 loop is conjectural, but in Kv1.2 channels the equivalent loop ‘‘runs parallel to the intracellular membrane surface just at the level of the inner helix bundle crossing’’ (Long et al., 2005a,b). Since this is precisely adjacent to where alkanols preferentially locate and since they lower bilayer surface tension by partitioning in the headgroup region (with acyl chains aligned normal to the bilayer plane), they could hardly avoid lateral contact with an S4–S5 loop located just there. Two decades ago, Treistman and Wilson (1987) testing an alkanol series and temperature on the Aplysia neurons potassium current, IA, found it ‘‘unlikely [an alkanol] exerts its actions on IA via perturbation of a bulk lipid phase.’’ However, they also followed alcohol eVects on fluorescent lipid probes and from this pointed out that their IA results could ‘‘be consistent with [lipid] domain‐specific actions within a heterogeneously organized lipid environment.’’ In turn, we could add, this could be consistent with alkanol‐ modulated changes in the channel‐specific lateral pressure profile at the protein–lipid interface. The ‘‘protein hypothesis’’ gained ground at a time when (1) bulk bilayer lipid eVects like ‘‘fluidity’’ were being ruled out and (2) the kinetically important motions of VGCs were seen as sequestered from the bilayer. Combining updated structural information with information of the eVects of stretch and BMRs on recombinant VGCs suggests that these two extreme views (‘‘protein’’ vs ‘‘bulk bilayer’’) can be reconciled by an intermediate view that focuses on channel‐specific motions at the complex‐shaped and conformation‐dependent lateral interface between bilayer lipids and the channel protein. The wider issue of the interaction between VGC conformations and bilayer mechanics would be well served by a full description of the kinetic eVects of stretch, hyperbaric pressure, temperature, and BMRs (e.g., an alkanol series and cholesterol) on each of several identified transitions in one structurally characterized VGC. A Kv1 channel would be the best candidate because of emerging Kv1.2 structural information and because MS transitions in a Kv1 (Shaker) are already known to include activation, prepore opening, and slow inactivation. Use of defined bilayers in this endeavor would facilitate computational probing of the findings as is being done for bacterial MS channels (Meyer et al., 2006).
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
319
IV. REVERSIBLE STRETCH‐INDUCED CHANGES IN PARTICULAR VGCs A. Kv Channels The Kv1, Shaker, is a robustly MS channel whether in cell‐attached or excised patches, and whether membrane tension is generated by negative or positive pipette pressure (Gu et al., 2001). Monitored at a fixed voltage, Shaker can behave as either an SA or an SI channel (Fig. 3). Stretch increases the rate of Shaker activation at the foot of the g(V ) curve, and slightly (10 mV) left shifts the entire curve at lytic tension (Tabarean and Morris, 2002). When stretch is applied during a voltage step, both rise and decay times accelerate and peak current increases, a pattern particularly evident in the S3–S4 deletion mutants, including Shaker 5aa. In Shaker 5aa, activation and slow inactivation speeds are similar, allowing eVects of stretch on both to be monitored simultaneously and modeled. Shaker 5aa behavior was described within the framework of a linear ‘‘Aldrich’’ model (appropriately scaled for Shaker 5aa and with slow inactivation added) where, over a wide voltage range, 1.5‐ to 2‐fold rate changes (forward and backward rates for closed–closed transitions changed reciprocally, yielding <10‐fold increase in overall forward rate along the activation path) could account for the observed eVects of moderate stretch over a wide range of voltages. Subsequently (Laitko and Morris, 2004), we focused on the simple kinetics of Shaker 5aa (using the voltage range where 5aa kinetics obey a scheme in which four independent forward steps resolve in a cooperative C---O transition (Fig. 4D). Stretch did not aVect the voltage dependence of the key activation parameter in Shaker 5aa, only the basic rates. Time‐ and amplitude‐scaling transformations showed that while voltage changes the rates of activation and inactivation of Shaker 5aa diVerently, stretch had exactly the same eVect (1.5‐ to 2‐fold acceleration) on the characteristic times of both processes (Fig. 7). This suggests that the protein motions limiting 5aa activation and slow inactivation are similar enough physically to ‘‘feel’’ the stretch‐altered bilayer similarly (e.g., both might involve motions over a substantial fraction of the channel–lipid interface). Interestingly, Pathak et al. (2006) have subsequently reported preliminary fluorescence data showing that slow inactivation in Shaker involves (in addition to the pore domain) a motion that projects throughout the voltage‐sensing domain. As already explained, we used Shaker ILT to examine the concerted voltage‐dependent closed–closed motion just prior to the concerted opening step (Laitko et al., 2006). We found that this motion, which may entail some
320
Morris and Juranka
∆p = 0/−35/0 mmHg
Current (pA)
1000
1.0
Current, normalized
A
80 mV
Amplitudes normalized, times scaled
0.5
0.0
0 0
1
0
1
0
1
1.0 1000
∆p = 0/−10/−20/−30/−40/0 mmHg 0.5
500
40 mV
0
0.0 0
1
∆p = 0/−10/−20/−30/−40/0 mmHg 1.0
800
0.5
400
20 mV
0.0
0 0
1
0
2
Time (s) 20,40, and 80 mV ∆p = 0
Current (pA)
1400
700
0
1
2
Time, scaled Current, normalized
B
∆p = 0
1.0
0.5
0.0
0
1
0
1
FIGURE 7 Activation and slow inactivation transition rates are equally stretch sensitive in Shaker 5aa. This finding is illustrated (all traces from one patch) by comparing the eVect of rescaling currents for diVerent stretch intensities at any given voltage (A) vs rescaling for diVerent voltages at a given stretch intensity (B). (A) First, at one voltage (80 mV), currents before/during/after stretch; at right these are amplitude and time scaled. Next, at 40 mV then at 20 mV, a stretch dose response; at right, amplitude and time scaled (the procedure is to rescale for activation only. The outcome is that inactivation trajectories all overlap if they are all
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
321
expansion in the plane of the bilayer, nevertheless decelerates with stretch, possibly for entropic reasons (stretch may interfere with the tidy orchestration of eight moving parts of the channel). Pore closing in both Shaker ILT and Shaker WTIR (as monitored by tail current rates) showed a consistent eVect of stretch within a batch of oocytes but not between batches (acceleration, no eVect, and deceleration were all obtained). It may be that oocyte bilayer lipid composition diVers enough among frogs to confound direct lipid stress (i.e., stretch) eVects. Kv3 and BKCa channels were discussed in earlier sections and Kv3 is discussed in comparison to Cav channel stretch responses in the next section. Assorted accounts have appeared on swelling‐activated Kv conductances in native preparations, but none test the possibility that membrane stretch is involved. For example, in gastric myocytes, hypoosmotic solutions markedly increase IKCa and a 4‐aminopyridine sensitive IKv (Piao et al., 2001). A trigeminal nerve Kv conductance is described as MS (Piao et al., 2006) based on responses to osmotic stimuli. The actin cytoskeleton is necessary for the response and membrane stretch was not tested. In cardiac cells, swelling enhances a slowly activating delayed rectifier current (IKs) (Kubota et al., 2002). Two proteins constitute IKs (KCNQ1 and KCNE1) {KCNQ1 ¼ KvLQT1; KCNE1 ¼ minK} but recombinant homomeric KCNQ1 channels on their own respond to swelling (increased conductance, no change in kinetics) and protein tyrosine kinase activity is not involved. Grunnet et al. (2003) found that recombinant KCNQ1 and KCNQ4 (but not KCNQ2 and KCNQ3) channels are tightly regulated by cell volume changes. Current amplitude increases with no change in kinetics and here too the actin cytoskeleton is implicated (a similar result was obtained by the same group for HNC2 channels; Calloe et al., 2005). The swelling‐induced eVect does not require the auxiliary subunits (KCNE1–3). Stretch was not tested.
B. Cav and Kv3 Channels Have Similar Stretch Responses Historically, the first indication that VGCs are included among the MS channels was a careful report on a smooth muscle L‐type Ca current in which whole‐cell inflation was used to stress the membrane (Langton, 1993).
accelerated the same‐fold as activation). (B) Same patch, when the same voltages are tested with no stretch and currents normalized to activation, the inactivation trajectories do not overlap. Thus, activation and slow inactivation have diVerent voltage dependences but the same stretch dependence. For more details, see Fig. 8 and related text of Laitko and Morris (2004).
322
Morris and Juranka
Many native myocytes exhibit the pattern of responses described in that paper and comparable responses have been confirmed for recombinant L‐type (Lyford et al., 2002) and N‐type (Calabrese et al., 2002) Cav channels (Fig. 6). In both recombinants, mechanosensitivity resides with the a‐ (pore) subunit. The demonstrable tension‐sensitivity (as opposed to ‘‘pressure’’‐ sensitivity) of N‐type channels allowed us to argue that Cav channels respond directly to bilayer stretch. Whether stretch is applied under whole‐cell conditions (inflation or osmotic swelling), under perforated patch (osmotic swelling), or via stretch of cell‐attached patches, the following stretch eVects are seen for both N‐ and L‐type Cav channels: 1. Stretch reversibly increases peak and steady state ICav 1.5‐fold. 2. Stretch does not change the rate of ICav onset. 3. Stretch reversibly increases gmax with no shift in the midpoint of g(V) (gmax and g(V) midpoint are determined from I/V plots; eVectively stretch scales up the I/V relation). 4. Stretch increases single‐channel activity without changing the unitary conductance. New membrane insertion does not account for the increased gmax for either recombinant N‐type (Calabrese et al., 2002) or native L‐type current (Ben‐Tabou De‐Leon et al., 2006) since membrane capacitance is unaVected. In the case of N‐type channels (but not L‐type), inactivation from the open state accelerates with stretch. A T‐type channel stably transfected in HEK cells, into which were co‐ transfected N‐type channels, exhibited no response to the same stretch stimuli that enhanced N‐type current (Calabrese et al., 2002). On the other hand, native pituitary cell T‐type current responds like L‐type current in the same cells albeit more weakly (Ben‐Tabou De‐Leon et al., 2003, 2006). When reporting the action of stretch on N‐type current (Calabrese et al., 2002), we were struck by a qualitative diVerence between Cav and Kv (Shaker) responses: the rate of current onset does not change with stretch and activation does not left shift. We felt that stretch might not aVect voltage sensor motions in Cav channels as it does in Shaker channels. We have now, however, observed Cav‐channel‐like responses in a Kv3 channel (Laitko et al., 2006): in both, stretch reversibly and repeatably increases the peak and steady state current, with an apparent increase in gmax and with no evidence (from the current onset rate) that activation accelerates with stretch (Fig. 5). Taking an Occam’s razor approach, it seems simplest therefore, to use for the Cav channels a generic working hypothesis consistent with what is known from Shaker (and mutants): That hypothesis is that the net eVect of stretch on ICav results from
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
323
1. stretch acceleration of an MS voltage‐dependent forward transition (not the rate‐limiting transition for onset of ICav) and 2. stretch deceleration of an MS‐concerted voltage‐sensitive forward transition (simply to be consistent with Shaker ILT) and 3. in the N‐type Cav channel, stretch‐accelerated inactivation. These putative MS transitions in Cav channels would be modulated by stretch in such a way that, very robustly, the net result is increased peak and steady state ICav, with no evidence of shift in the g(V ) curve. We hoped that activation kinetics in T‐type Cav channels would be more ‘‘accessible’’ than in N‐ or L‐type, but inexplicably the recombinant T‐type isoform we tested (Calabrese et al., 2002) showed no stretch sensitivity in HEK cells. However, Ben‐Tabou De‐Leon et al. (2003, 2006) find that in pituitary cells, T‐type currents are mildly responsive to swell/shrink stimuli, and although they made no kinetic analysis, their traces suggest that activation and inactivation rates respond. This would be worth a closer look, since access to MS activation could help clarify whether the ‘‘Occam’s razor’’ hypothesis is warranted. C. Cav: L‐Type Channels in Native Preparations L‐type Cav channel mechanosensitivity has been tested in more native myocytes using more types of mechanostimuli than for any other VGC. When the channels respond, the response is stereotyped (as described above: increased peak and steady state ICav, with no evidence of shift in the g(V ) curve). Nevertheless, the substantial collection of reports is full of idiosyncrasies, contradictions, and quirks, some of which we reviewed (Morris and Laitko, 2005; see also references in Calabrese et al., 2002). A few additional items are worthy of mention. For hypo‐ and hyperosmotic stimuli, Ben‐Tabou De‐Leon et al. (2003, 2006) showed that F‐actin reagents (cytochalsin D and phalloidin, eYcacy monitored via microscopy) have no eVect on the swell/shrink modulation of L‐ and T‐type ICav in pituitary cells. In ventricular myocytes, by contrast, Pascarel et al. (2001) found that F‐actin is needed for T‐type current to respond to swelling. Amano et al. (2005) observed that both hypoosmotic solution and fast bath flow modulate L‐type current in smooth muscle myocytes. It is unclear if the flow‐induced enhancement (which is transient except when cells are thapsigargin‐treated) is a membrane mechanical eVect. Xu et al. (2000) tested several stretch stimuli on smooth muscle ICa using whole‐cell clamp. Hypoosmotic bath solution increases ICa as does cell
324
Morris and Juranka
inflation, but direct longitudinal stretch (up to 130%) using two electrodes does not. Belus and White (2003) using axial stretch on ventricular myocytes observed no eVect of stretch on L‐type current, although the same stimulus activated a streptomycin sensitive linear current. Perhaps, in the ventricular myocytes, the plasma membrane regions where L‐type Ca channels reside can avoid lipid stress during a directional mechanical stimulus; cardiac muscle L‐type Ca channels are expressed in longitudinally arrayed puncta (Grabner et al., 1998). Since flow is a directional mechanostimulus, the transience of the flow eVect noted by Amano et al. (2005) might echo the ineVectiveness of directional stretch.
D. Nav Channels Gastrointestinal interstitial cells of Cajal express a Nav1.5 current that appears to be modulated by shear stress and stretch, but this evidently depends on membrane skeleton (Strege et al., 2003). Squid axon excitability is reversibly susceptible to axonal inflation (Terakawa and Nakayama, 1985) and squid axon gating and ionic current are modulated (slowed) by hyperbaric pressure (Conti et al., 1984). Immunolabeling reveals that Nav channels are present in the sensory neurite of Pacinian corpuscles (Pawson and Bolanowski, 2002). TTX‐sensitive Na entry into vascular smooth muscle cells and into Nav1.2 transfected CHO cells inhibits shear stress‐mediated activation of a signaling kinase (Traub et al., 1999). In Nav1.4 channels expressed in oocytes, where there is such an overwhelming irreversible eVect of stretch (a switch to a fast‐gating mode; Tabarean et al., 1999) we did not rigorously test the fast Nav1.4 currents for reversible MS responses. However, this same Nav channel, expressed in HEK cells shows left shifts in its inactivation curves when bilayer elasticity is elevated by adding various amphiphiles or by depleting cholesterol (Lundbaek et al., 2004). We have therefore begun testing Nav1.5 a‐subunit channels in cell‐attached oocyte patches subjected to stretch. Unlike the Nav1.4 a‐subunit, these do not show irreversible changes in response to stretch and our preliminary finding is that both the onset and decline phases of current transients elicited by voltage steps accelerate reversibly with stretch (CEM, unpublished observation).
E. HCN Channels Recombinant sea urchin and mammalian HCN channels (spHCN and HCN2) expressed in oocytes respond directly and reversibly to membrane stretch (Lin et al., 2007); our preliminary findings are that both
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
325
hyperpolarization‐induced turn on of IHCN2 and depolarization‐induced turn oV of IHCN2 accelerate with stretch. It remains to be seen whether these MS responses bear any relation to the responses of HCN2 channels coexpressed with aquaporin in Xenopus laevis oocytes and monitored via two‐electrode voltage‐clamp during swelling (Calloe et al., 2005). The hyperpolarization‐activated current increases 30% with no change in activation kinetics. This response was abolished by cytochalasin D. The only indication that membrane stretch triggers the response is the fact that inflating oocytes with oil mimics the eVect of osmotic swelling.
V. IRREVERSIBLE STRETCH‐INDUCED GATING CHANGES IN VGCs Stretched membrane patches experience not only elastic changes (idealized as reversible far‐field tension in the plane of the membrane) but also plastic (irreversible) changes in patch microstructure (these changes are not characterized but could involve damage to the integrity of the membrane skeleton, protein clusters, caveolae, microvilli, and any organized lipid microdomains). The impact of plastic change becomes self‐evident whenever the responses of channels to stretch depend on a patch’s mechanical ‘‘history.’’ The physical nature of the microstructural changes will vary among cell types. Recombinant Nav1.4 in oocyte patches (Shcherbatko et al., 1999; Tabarean et al., 1999) and N‐type Cav studied under whole‐cell clamp (Calabrese et al., 2002) both exhibit history‐dependent activity changes due to unidentified cumulative mechanical trauma of the membrane. Figure 8A shows, from a single oocyte, Nav1.4 current families from four patches, two of which formed without applied stretch; stretch irreversibly alters the channel’s gating characteristics. Surprisingly, the eVect of this ‘‘trauma’’ on Nav1.4 expressed in Xenopus oocytes is like the eVect of the channel’s auxiliary subunit. In oocytes, Nav1.4 a‐ (pore) subunits show anomalously slow, right‐shifted kinetics because oocytes lack auxiliary b1‐subunits. Coexpression of a‐ and b‐subunits leads to normal gating behavior. Surprisingly, however, applying stretch to patches with only the a‐subunit makes the Nav1.4 a channels switch irreversibly into a gating mode that is not only fast and left‐shifted but indistinguishable from what is seen with the b1‐subunit present (Shcherbatko et al., 1999; Tabarean et al., 1999). How stretch cumulatively and irreversibly alters the a‐subunit environment to achieve the same eVect as having b1‐subunits is unknown. Interestingly, the auxiliary b1‐subunit (which spans the bilayer only once) has an extracellular domain that acts as an adhesion molecule (McEwen and Isom, 2004). On the basis of eVects of
326
Morris and Juranka
A Spontaneous seal
Spontaneous seal
0.1 nA 0.1 nA 5 ms −15 mmHg seal
−25 mmHg seal
0.1 nA 0.1 nA Four patches, same oocyte B 1.0 Inactivation
Inactivation
1.0
Time control
1.5
0.5
t=0 4−16 min later
0.0 Vprepulse (mV)
1.0
Vprepulse (mV)
1.0
Example illustrating progressive irreversible effect
Inactivation
t = 0 (no stretch) 4−19 min later With stretch
Activation
0.0
Effect of stretch
Activation (Same cells)
0.5
0.5 Control After 2 min of stretch After 3 min of stretch 4 min after stretch was released
t=0 4−10 min later with stretch 0.0
0.0 −120
−80 −40 Vprepulse (mV)
0
−80
−40
0
40
80
Test-pulse voltage (mV)
FIGURE 8 Irreversible eVects of stretch on Nav1.4 and N‐type Cav channels. (A) Even ‘‘unintended’’ stretch experienced during seal formation causes an irreversible switch from anomalous slow mode to fast mode gating in Nav1.4 (a‐subunit only) channels expressed in oocytes.
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
327
nocadazole, it was suggested (Shcherbatko et al., 1999) that stretch might disrupt kinetically important interactions of Nav1.4 channels with microtubules. This seems unpersuasive (see Fig. 1B and related discussion of Morris et al., 2006); moreover, in mammalian cells Nav1.4 gating is unaVected by cytoskeletal reagents (Moran et al., 2000). N‐type Cav channels in the whole‐cell configuration show run down and time‐dependent left shift of their inactivation curve (Fig. 8B). In this configuration, stretch that increases ICav (while having no eVect on the voltage dependence of activation) causes a further irreversible left shift of the inactivation curve (see Fig. 6 in Calabrese et al., 2002 for details). Meanwhile, the same channels do exhibit reversible stretch acceleration of open state inactivation. Kinetically, discrete irreversible stretch changes in VGC behavior such as these may correspond to (1) changes that, in an intact cell, would be reversed via enzyme‐mediated processes or (2) changes that in an intact cell would be avoided by virtue of mechanoprotective arrangements. Membrane skeleton is undoubtedly important for mechanoprotection (Morris, 2001) and sequestration of channels in specialized membrane subdomains may also contribute. In skeletal muscle, caveolae are exceedingly diYcult to flatten with stretch (Dulhunty and Franzini‐Armstrong, 1975); perhaps residence in caveolae aVords VGCs a degree of mechanoprotection from both reversible stretch eVects and irreversible stretch damage.
VI. TECHNICAL ISSUES A. Applying a Stretching Force to Study MS Modulation of VGC Activity Elevating the tension in a membrane by applying stretch to a patch of membrane or an intact cell is not a precision process. Various techniques have been used in the context of VGCs and are briefly itemized here. This topic has been reviewed (Morris et al., 2006). 1. Gigaseal Patch Recording Pipette suction (negative pressure) is used to apply stretch to either cell‐ attached or excised patches. Usually, we apply suction 1 s before step pulses begin. When P/N linear subtraction is used, suction is present during
Modified from Tabarean et al. (1999). (B) For N‐type Cav channels recorded under whole‐cell clamp in HEK cells, stretch exacerbates a time‐dependent left shift of inactivation (as in the Nav1.4 irreversible eVect, the eVect is progressive, as illustrated), even though, in the same cells, the voltage dependence of activation is entirely unaVected. Modified from Calabrese et al. (2002).
328
Morris and Juranka
the subtraction pulses. Positive pipette pressure also elevates the membrane tension with the membrane at opposite curvature. Since patches are more durable with suction than with positive pressure, suction is the stimulus most used. Irreversible changes that occur with stretch must inevitably include changes in the lipid organization of the bilayer. We suspect this accounts for some of the nonstationary behavior seen in many patches. Stretch‐induced disruption of the membrane skeleton probably exacerbates the previously mentioned eVect. Sometimes, a patch will show entirely reversible stretch eVects with the first stimulus, and other times, there will be a settling in period. Dose response data are hard to obtain. We deal with this by routinely checking for reversibility (testing the response to a voltage step or ramp before, during, and after a stretch stimulus). 2. Hydrostatic Inflation Positive pipette pressure applied under whole‐cell clamp inflates the cell. This stimulus is not a strictly reversible stimulus. Cell debris can move into the tip during a pressure‐oV step. Applied pressure is an unreliable gauge of membrane tension because mechanical resistances develop in the pipette (Langton, 1993; Calabrese et al., 2002). In HEK cells, elevated membrane tension develops only when a frank increase in cell volume occurs (Calabrese et al., 2002). Long range structures (membrane skeleton?) may disorganize or rupture at this point, causing load to transfer to the bilayer. Inflation may also be coincident with flattening of membrane irregularities, at which point the microdomain organization as well as the local curvatures in the VGC‐ bearing bilayer may change (similar events presumably occur in patches on gigaseal formation). For various MS VGCs, whole‐cell inflation and osmotic swelling may have the same eVect on channel activity, but it is never certain that membrane stress is the source of gating energy in the case of osmotic swelling. 3. Osmotic Swelling For mammalian cell lines and several native cells, both perforated patch recordings (downshock of bath solution) and whole‐cell recording (pipette solution at elevated osmotic pressure or, more commonly, external solution at reduced osmotic pressure) have been used. As with hydrostatic inflation, monitoring cell volume during recordings is critical. Xenopus oocytes, being adapted for prolonged sojourns in pond water, swell very little with downshock, but will swell if the cells are made to express aquaporin (Calloe et al., 2005); under those conditions, VGC currents can be recorded under two microelectrode clamps in conjunction with transient (and visually measureable) osmotic‐swelling stimuli.
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
329
4. Inflation by Oil Injection During two microelectrode voltage clamps, Xenopus oocytes can be inflated by injecting about 50 nl of oil using the same device used to inject cRNA. This procedure detectably increases oocyte volume without dilution of any cytoplasmic molecules (Calloe et al., 2005). 5. Stretching Native Myocytes Using Two Pipettes See Xu et al., 2000; Belus and White, 2003. 6. Shear Flow This complex mechanical stimulus needs careful verification (Levitan et al., 2000); flow eVects can be confounded with mechanochemical signals (Maroto and Hamill, 2001).
B. Gadolinium Strangeness To cleanly measure MS VGC currents, other MS currents need to be eliminated. Gadolinium inhibits some SA cation channels (Yang and Sachs, 1989; although see Hurwitz et al., 2002) making it useful in some biophysical studies. We have found it highly problematic used in connection with recombinant VGCs expressed in Xenopus oocytes (Gu et al., 2001) and so prefer 1‐mM lanthanum chloride whenever possible). Gadolinium right‐shifts voltage‐dependent gating parameters (thought be a surface charge shielding eVect) and reduces maximal conductances (Elinder and Arhem, 1994; Gu et al., 2001). Given that lanthanides also rigidify the exposed bilayer leaflet (Tanaka et al., 2002), they probably exert lipid stress eVect too, but no‐gadolinium controls (Tabarean and Morris, 2002) showed that the eVects of stretch on Shaker are evident with or without a lanthanide. Gadolinium’s side eVects are generally acceptable in biophysical studies but its use is not warranted to ‘‘selectively’’ block SA cation channels in native cell where the final outputs are action potentials or a background cation conductance. At vanishingly small concentrations, carbonate, bicarbonate, phosphate, and hypophosphate ions precipitate gadolinium; when gadolinium is used to block SA cation channels in native cells exposed to CO2‐bicarbonate buVered solutions, it is diYcult to know what to make of the findings. Do cells become coated with a microprecipitate of gadolinium‐(bi)carbonate‐(bi)phosphate that somehow produces the observed eVects?
330
Morris and Juranka
VII. SUMMARY COMMENTS The conformational motions of VGCs respond to changes in lipid stress. Fields as distinct as sensory physiology and anesthesiology and evolutionary biology would be well served by a more precise understanding of this fact and its implications. What is ideally needed at this juncture is to pick one well‐characterized VGC and subject it to a range of the biophysically important physical and chemical lipid stress agents. These include stretch, hyperbaric pressure, and temperature, plus an array of BMRs (e.g., an alkanol series, cholesterol). Kv1 channels are the obvious candidate, given the amassed structure–function information available for them. Already, it has been shown that activation, prepore opening, and slow inactivation transitions in the prototypical Kv1, Shaker, are MS. To facilitate computational probing, bilayer‐based studies would be ideal (KvAP, though not a Kv1— more a ‘‘Kv0’’—could be used here), but preparations that allow for gating current measurements and other means (optical, chemical) of monitoring channel motions will also be needed since many transitions aVected by physical and chemical lipid stressors will not be directly accessible through measurements of ionic current. Our experience with Kv channels, using stretch as the lipid stressor, illustrates that the behavior of ‘‘hidden transitions’’ can be critical; finessing the hidden transitions as well as focusing on the rate‐limiting ones should greatly clarify the actions of amphiphilic molecules on VGCs. Many VGCs (like some TRP channels) are ‘‘MS channels,’’ and what seems likely to bear most fruit now is to dissect the behavior of 6TM channels in terms of their diverse inherent susceptibilities to the physical and chemical agents of lipid stress. Acknowledgments Work in CE Morris’ laboratory is supported by grants from the Canadian Institutes of Health Research.
References Amano, S., Ishikawa, T., and Nakayama, K. (2005). Facilitation of L‐type Ca2þ currents by fluid flow in rabbit cerebral artery myocytes. J. Pharmacol. Sci. 98, 425–429. Barbuti, A., Gravante, B., Riolfo, M., Milanesi, R., Terragni, B., and DiFrancesco, D. (2004). Localization of pacemaker channels in lipid rafts regulates channel kinetics. Circ. Res. 94, 1325–1331. Bartucci, R., Gambacorta, A., Gliozzi, A., Marsh, D., and Sportelli, L. (2005). Bipolar tetraether lipids: Chain flexibility and membrane polarity gradients from spin‐label electron spin resonance. Biochemistry 44, 15017–15023. Baumgart, T., Hess, S. T., and Webb, W. W. (2003). Imaging coexisting fluid domains in biomembrane models coupling curvature and line tension. Nature 425, 821–824.
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
331
Belus, A., and White, E. (2003). Streptomycin and intracel, lular calcium modulate the response of single guinea‐pig ventricular myocytes to axial stretch. J. Physiol. 546, 501–509. Ben‐Tabou De‐Leon, E., Blotnick, E., and Nussinovitch, I. (2003). EVects of osmotic swelling on voltage‐gated calcium channel currents in rat anterior pituitary cells. Am. J. Physiol. Cell Physiol. 285, C840–C852. Ben‐Tabou De‐Leon, S., Ben‐Zeev, G., and Nussinovitch, I. (2006). EVects of osmotic shrinkage on voltage‐gated Ca2þ channel currents in rat anterior pituitary cells. Am. J. Physiol. Cell Physiol. 290, C222–C232. Bock, J., Szabo, I., Gamper, N., Adams, C., and Gulbins, E. (2003). Ceramide inhibits the potassium channel Kv1. 3 by the formation of membrane platforms. Biochem. Biophys. Res. Commun. 305, 890–897. Brauchi, S., Orio, P., and Latorre, R. (2004). Clues to understanding cold sensation: Thermodynamics and electrophysiological analysis of the cold receptor TRPM8. Proc. Natl. Acad. Sci. USA 101, 15494–15499. Calabrese, B., Tabarean, I. V., Juranka, P., and Morris, C. E. (2002). Mechanosensitivity of N‐type calcium channel currents. Biophys. J. 83, 2560–2574. Calloe, K., Elmedyb, P., Olesen, S. P., Jorgensen, N. K., and Grunnet, M. (2005). Hypoosmotic cell swelling as a novel mechanism for modulation of cloned HCN2 channels. Biophys. J. 89, 2159–2169. Camara, A. K., Begic, Z., Kwok, W. M., and Bosnjak, Z. J. (2001). DiVerential modulation of the cardiac L‐ and T‐type calcium channel currents by isoflurane. Anesthesiology 95, 515–524. Cantor, R. S. (1997). The lateral pressure profile in membranes: A physical mechanism of general anesthesia. Biochemistry 36, 2339–2344. Cantor, R. S. (1999). Lipid composition and the lateral pressure profile in bilayers. Biophys. J. 76, 2625–2639. Chang, H. M., Reitstetter, R., Mason, R. P., and Gruener, R. (1995). Attenuation of channel kinetics and conductance by cholesterol: An interpretation using structural stress as a unifying concept. J. Membr. Biol. 143, 51–63. Chemin, J., Monteil, A., Perez‐Reyes, E., Nargeot, J., and Lory, P. (2001). Direct inhibition of T‐type calcium channels by the endogenous cannabinoid anandamide. EMBO J. 20, 7033–7040. Chen, X., Sirois, J. E., Lei, Q., Talley, E. M., Lynch, C., III, and Bayliss, D. A. (2005). HCN subunit‐specific and cAMP‐modulated eVects of anesthetics on neuronal pacemaker currents. J. Neurosci. 25, 5803–5814. Chu, B., and Treistman, S. N. (1997). Modulation of two cloned potassium channels by 1‐alkanols demonstrates diVerent cutoVs. Alcohol Clin. Exp. Res. 21, 1103–1107. Chu, B., Dopico, A. M., Lemos, J. R., and Treistman, S. N. (1998). Ethanol potentiation of calcium‐activated potassium channels reconstituted into planar lipid bilayers. Mol. Pharmacol. 54, 397–406. Clapham, D. E. (2003). TRP channels as cellular sensors. Nature 426, 517–524. Clarke, A. L., Petrou, S., Walsh, J. V., Jr., and Singer, J. J. (2003). Site of action of fatty acids and other charged lipids on BKCa channels from arterial smooth muscle cells. Am. J. Physiol. Cell Physiol. 284, C607–C619. Conti, F., Inoue, I., Kukita, F., and Stuhmer, W. (1984). Pressure dependence of sodium gating currents in the squid giant axon. Eur. Biophys. J. 11, 137–147. Correa, A. M. (1998). Gating kinetics of Shaker Kþ channels are diVerentially modified by general anesthetics. Am. J. Physiol. 275, C1009–C1021. Covarrubias, M., and Rubin, E. (1993). Ethanol selectively blocks a noninactivating Kþ current expressed in Xenopus oocytes. Proc. Natl. Acad. Sci. USA 90, 6957–6960.
332
Morris and Juranka
Crowley, J. J., Treistman, S. N., and Dopico, A. M. (2003). Cholesterol antagonizes ethanol potentiation of human brain BKCa channels reconstituted into phospholipid bilayers. Mol. Pharmacol. 64, 365–372. Cuello, L. G., Cortes, D. M., and Perozo, E. (2004). Molecular architecture of the KvAP voltage‐dependent Kþ channel in a lipid bilayer. Science 306, 491–495. Czub, J., and Baginski, M. (2006). Comparative molecular dynamics study of lipid membranes containing cholesterol and ergosterol. Biophys. J. 90, 2368–2382. Dan, N., and Safran, S. A. (1998). EVect of lipid characteristics on the structure of transmembrane proteins. Biophys. J. 75, 1410–1414. Davies, D. L., Crawford, D. K., Trudell, J. R., Mihic, S. J., and Alkana, R. L. (2004). Multiple sites of ethanol action in alpha1 and alpha2 glycine receptors suggested by sensitivity to pressure antagonism. J. Neurochem. 89, 1175–1185. Devaux, P. F., and Morris, R. (2004). Transmembrane asymmetry and lateral domains in biological membranes. TraYc 5, 241–246. Dong, X. X., and Iwasa, K. H. (2004). Tension sensitivity of prestin: Comparison with the membrane motor in outer hair cells. Biophys. J. 86, 1201–1208. Dopico, A. M., Kirber, M. T., Singer, J. J., and Walsh, J. V., Jr. (1994). Membrane stretch directly activates large conductance Ca(2þ)‐activated Kþ channels in mesenteric artery smooth muscle cells. Am. J. Hypertens. 7, 82–89. Dulhunty, A. F., and Franzini‐Armstrong, C. (1975). The relative contributions of the folds and caveolae to the surface membrane of frog skeletal muscle fibres at diVerent sarcomere lengths. J. Physiol. 250, 513–539. Elinder, F., and Arhem, P. (1994). EVects of gadolinium on ion channels in the myelinated axon of Xenopus laevis: Four sites of action. Biophys. J. 67, 71–83. Etzion, Y., and Grossman, Y. (2000). Pressure‐induced depression of synaptic transmission in the cerebellar parallel fibre synapse involves suppression of presynaptic N‐type Ca2þ channels. Eur. J. Neurosci. 12, 4007–4016. Feller, S. E., Brown, C. A., Nizza, D. T., and Gawrisch, K. (2002). Nuclear overhauser enhancement spectroscopy cross‐relaxation rates and ethanol distribution across membranes. Biophys. J. 82, 1396–1404. Fisyunov, A., Tsintsadze, V., Min, R., Burnashev, N., and Lozovaya, N. A. (2006). Cannabinoids modulate the P‐type high voltage‐activated calcium currents in Purkinje neurons. J. Neurophysiol. 96, 1267–1277. Folco, E. J., Liu, G. X., and Koren, G. (2004). Caveolin‐3 and SAP97 form a scaVolding protein complex that regulates the voltage‐gated potassium channel Kv1. 5. Am. J. Physiol. Heart Circ. Physiol. 287, H681–H690. Franco‐Obregon, A., and Lansman, J. B. (2002). Changes in mechanosensitive channel gating following mechanical stimulation in skeletal muscle myotubes from the mdx mouse. J. Physiol. 539, 391–407. Gaus, K., Zech, T., and Harder, T. (2006). Visualizing membrane microdomains by Laurdan 2‐photon microscopy. Mol. Membr. Biol. 23, 41–48. Gill, C. H., Randall, A., Bates, S. A., Hill, K., Owen, D., Larkman, P. M., Cairns, W., Yusaf, S. P., Murdock, P. R., Strijbos, P. J., Powell, A. J., Benham, C. D., et al. (2004). Characterization of the human HCN1 channel and its inhibition by capsazepine. Br. J. Pharmacol. 143, 411–421. Grabner, M., Dirksen, R. T., and Beam, K. G. (1998). Tagging with green fluorescent protein reveals a distinct subcellular distribution of L‐type and non‐L‐type Ca2þ channels expressed in dysgenic myotubes. Proc. Natl. Acad. Sci. USA 95, 1903–1908.
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
333
Grunnet, M., Jespersen, T., MacAulay, N., Jorgensen, N. K., Schmitt, N., Pongs, O., Olesen, S. P., and Klaerke, D. A. (2003). KCNQ1 channels sense small changes in cell volume. J. Physiol. 549, 419–427. Gu, C. X., Juranka, P. F., and Morris, C. E. (2001). Stretch‐activation and stretch‐inactivation of Shaker‐IR, a voltage‐gated Kþ channel. Biophys. J. 80, 2678–2693. Guharay, F., and Sachs, F. (1984). Stretch‐activated single ion channel currents in tissue‐ cultured embryonic chick skeletal muscle. J. Physiol. 352, 685–701. Guharay, F., and Sachs, F. (1985). Mechanotransducer ion channels in chick skeletal muscle: The eVects of extracellular pH. J. Physiol. 363, 119–134. Guizy, M., Arias, C., David, M., Gonzalez, T., and Valenzuela, C. (2005). {Omega}‐3 and {omega}‐6 polyunsaturated fatty acids block HERG channels. Am. J. Physiol. Cell Physiol. 289, C1251–C1260. Gullingsrud, J., and Schulten, K. (2004). Lipid bilayer pressure profiles and mechanosensitive channel gating. Biophys. J. 86, 3496–3509. Hajdu, P., Varga, Z., Pieri, C., Panyi, G., and Gaspar, R., Jr. (2003). Cholesterol modifies the gating of Kv1. 3 in human T lymphocytes. Pflugers Arch. 445, 674–682. Harris, T., Shahidullah, M., Ellingson, J. S., and Covarrubias, M. (2000). General anesthetic action at an internal protein site involving the S4‐S5 cytoplasmic loop of a neuronal K(þ) channel. J. Biol. Chem. 275, 4928–4936. Hisada, T., Walsh, J. V. Jr, and Singer, J. J. (1993). Stretch‐inactivated cationic channels in single smooth muscle cells. Pflugers Arch. 422, 393–396. Hu, Y. M., Zhang, Z., Gao, R. B., and Xu, Y. Q. (1997). EVect of lysophosphatidylcholine on the pacemaker current I(f) of sheep cardiac Purkinje fibers in ischemia‐like condition. Sheng Li Xue Bao 49, 513–520. Hurwitz, C. G., Hu, V. Y., and Segal, A. S. (2002). A mechanogated nonselective cation channel in proximal tubule that is ATP sensitive. Am. J. Physiol. Renal Physiol. 283, F93–F104. Iwata, A., Stys, P. K., Wolf, J. A., Chen, X.H, Taylor, A. G., Meaney, D. F., and Smith, D. H. (2004). Traumatic axonal injury induces proteolytic cleavage of the voltage‐gated sodium channels modulated by tetrodotoxin and protease inhibitors. J. Neurosci. 24, 4605–4613. Kahya, N., and Schwille, P. (2006). Fluorescence correlation studies of lipid domains in model membranes. Mol. Membr. Biol. 23, 29–39. Kamatchi, G. L., Durieux, M. E., and Lynch, C., III (2001). DiVerential sensitivity of expressed L‐type calcium channels and muscarinic M(1) receptors to volatile anesthetics in Xenopus oocytes. J. Pharmacol. Exp. Ther. 297, 981–990. Kendig, J. J. (1984). Nitrogen narcosis and pressure reversal of anesthetic eVects in node of Ranvier. Am. J. Physiol. 246, C91–C95. Kirber, M. T., Walsh, J. V., Jr., and Singer, J. J. (1988). Stretch‐activated ion channels in smooth muscle: A mechanism for the initiation of stretch‐induced contraction. Pflugers Arch. 412, 339–345. Kubota, T., Horie, M., Takano, M., Yoshida, H., Otani, H., and Sasayama, S. (2002). Role of KCNQ1 in the cell swelling‐induced enhancement of the slowly activating delayed rectifier K(þ) current. Jpn. J. Physiol. 52, 31–39. Kung, C. (2005). A possible unifying principle for mechanosensation. Nature 436, 647–654. Laitko, U., and Morris, C. E. (2004). Membrane tension accelerates rate‐limiting voltage‐ dependent activation and slow inactivation steps in a Shaker channel. J. Gen. Physiol. 123, 135–154. Laitko, U., Juranka, P. F., and Morris, C. E. (2006). Membrane stretch slows the concerted step prior to opening in a Kv channel. J. Gen. Physiol. 127, 687–701.
334
Morris and Juranka
Langton, P. D. (1993). Calcium channel currents recorded from isolated myocytes of rat basilar artery are stretch sensitive. J. Physiol. 471, 1–11. Ledwell, J. L., and Aldrich, R. W. (1999). Mutations in the S4 region isolate the final voltage‐ dependent cooperative step in potassium channel activation. J. Gen. Physiol. 113, 389–414. Lee, S. Y., Lee, A., Chen, J., and MacKinnon, R. (2005). Structure of the KvAP voltage‐ dependent Kþ channel and its dependence on the lipid membrane. Proc. Natl. Acad. Sci. USA 102, 15441–15446. Levitan, I., Helmke, B. P., and Davies, P. F. (2000). A chamber to permit invasive manipulation of adherent cells in laminar flow with minimal disturbance of the flow field. Ann. Biomed. Eng. 28, 1184–1193. Li, J., and Correa, A. M. (2002). Kinetic modulation of HERG potassium channels by the volatile anesthetic halothane. Anesthesiology 97, 921–930. Lin, W., Latiko, U., Juranka, P. F., and Morris, C. E. (2007). Dual stretch responses of mHCN2 pacemaker channels: Accelerated activation, accelerated deactivation. Biophys. J. [Epub ahead of print 2006 Dec 1]. Liu, L., Zhu, W., Zhang, Z. S., Yang, T., Grant, A., Oxford, G., and Simon, S. A. (2004). Nicotine inhibits voltage‐dependent sodium channels and sensitizes vanilloid receptors. J. Neurophysiol. 91, 1482–1491. Long, S. B., Campbell, E. B., and Mackinnon, R. (2005a). Crystal structure of a mammalian voltage‐dependent Shaker family Kþ channel. Science 309, 897–903. Long, S. B., Campbell, E. B., and Mackinnon, R. (2005b). Voltage sensor of Kv1. 2: Structural basis of electromechanical coupling. Science 309, 903–908. Lundbaek, J. A., Birn, P., Hansen, A. J., Sogaard, R., Nielsen, C., Girshman, J., Bruno, M. J., Tape, S. E., Egebjerg, J., Greathouse, D. V., Mattice, G. L., Koeppe, R. E., II, et al. (2004). Regulation of sodium channel function by bilayer elasticity: The importance of hydrophobic coupling. EVects of micelle‐forming amphiphiles and cholesterol. J. Gen. Physiol. 123, 599–621. Lundbaek, J. A., Birn, P., Tape, S. E., Toombes, G. E., Sogaard, R., Koeppe, R. E., II, Gruner, S. M., Hansen, A. J., and Andersen, O. S. (2005). Capsaicin regulates voltage‐dependent sodium channels by altering lipid bilayer elasticity. Mol. Pharmacol. 68, 680–689. Ly, H. V., and Longo, M. L. (2004). The influence of short‐chain alcohols on interfacial tension, mechanical properties, area/molecule, and permeability of fluid lipid bilayers. Biophys. J. 87, 1013–1033. Lyford, G. L., Strege, P. R., Shepard, A., Ou, Y., Ermilov, L., Miller, S. M., Gibbons, S. J., Rae, J. L., Szurszewski, J. H., and Farrugia, G. (2002). a(1C) (Ca(V)1. 2) L‐type calcium channel mediates mechanosensitive calcium regulation. Am. J. Physiol. Cell Physiol. 283, C1001–C1008. Macdonald, A. G. (2002). Ion channels under high pressure. Comp. Biochem. Physiol. A Mol. Integr. Physiol. 131, 587–593. Mannikko, R., Elinder, F., and Larsson, H. P. (2002). Voltage‐sensing mechanism is conserved among ion channels gated by opposite voltages. Nature 419, 837–841. Maroto, R., and Hamill, O. P. (2001). Brefeldin A block of integrin‐dependent mechanosensitive ATP release from Xenopus oocytes reveals a novel mechanism of mechanotransduction. J. Biol. Chem. 276, 23867–23872. Maroto, R., Raso, A., Wood, T. G., Kurosky, A., Martinac., B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nat. Cell Biol. 7, 179–185. McNaughton, N. C., Green, P. J., and Randall, A. D. (2001). Inhibition of human alpha1E subunit‐mediated Ca2þ channels by the antipsychotic agent chlorpromazine. Acta Physiol. Scand. 173, 401–408.
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
335
Meyer, G. R., Gullingsrud, J., Schulten, K., and Martinac, B. (2006). Molecular dynamics study of MscL interactions with a curved lipid bilayer. Biophys. J. 91, 1630–1637. Meyer, R., and Heinemann, S. H. (1997). Temperature and pressure dependence of Shaker Kþ channel N‐ and C‐type inactivation. Eur. Biophys. J. 26, 433–445. McEwen, D. P., and Isom, L. L. (2004). Heterophilic interactions of sodium channel beta1 subunits with axonal and glial cell adhesion molecules. J. Biol. Chem. 279, 52744–52752. Mitchell, D. C., and Litman, B. J. (2000). EVect of ethanol and osmotic stress on receptor conformation. Reduced water activity amplifies the eVect of ethanol on metarhodopsin II formation. J. Biol. Chem. 275, 5355–5360. Mitchell, D. C., Lawrence, J. T., and Litman, B. J. (1996). Primary alcohols modulate the activation of the G protein‐coupled receptor rhodopsin by a lipid‐mediated mechanism. J. Biol. Chem. 271, 19033–19036. Moran, O., Tammaro, P., Nizzari, M., and Conti, F. (2000). Functional properties of sodium channels do not depend on the cytoskeleton integrity. Biochem. Biophys. Res. Commun. 276, 204–209. Morris, C. E. (2001). Mechanoprotection of the plasma membrane in neurons and other non‐ erythroid cells by the spectrin‐based membrane skeleton. Cell Mol. Biol. Lett. 6, 703–720. Morris, C. E., and Sigurdson, W. J. (1989). Stretch‐inactivated ion channels coexist with stretch‐activated ion channels. Science 243, 807–809. Morris, C. E., and Laitko, U. (2005). The mechanosensitivity of voltage‐gated channels may contribute to cardiac M.E.F. In ‘‘Cardiac Mechano‐Electric Feedback and Arrhythmias: From Pipette to Patient’’ (P. Kohl, M. Franz, and F. Sachs, eds.), pp. 33–41. Saunders Philadelphia, Pennsylvania. Morris, C. E., Juranka, P. F., Lin, W., Morris, T. J., and Laitko, U. (2006). Studying the mechanosensitivity of voltage‐gated channels using oocyte patches. In ‘‘Xenopus Protocols: Cell Biology and Signal Transduction’’ (X. J. Liu, ed.). Methods Mol. Biol. 322, 315–329. Mozrzymas, J. W., Teisseyre, A., and Vittur, F. (1996). Propofol blocks voltage‐gated potassium channels in human T lymphocytes. Biochem. Pharmacol. 52, 843–849. Murata, Y., Iwasaki, H., Sasaki, M., Inaba, K., and Okamura, Y. (2005). Phosphoinositide phosphatase activity coupled to an intrinsic voltage sensor. Nature 435, 1239–1243. Namba, T., Ishii, T. M., Ikeda, M., Hisano, T., Itoh, T., Hirota, K., Adelman, J. P., and Fukuda, K. (2000). Inhibition of the human intermediate conductance Ca(2þ)‐activated K(þ) channel, hIK1, by volatile anesthetics. Eur. J. Pharmacol. 395, 95–101. Nicholson, R. A., Liao, C., Zheng, J., David, L. S., Coyne, L., Errington, A. C., Singh, G., and Lees, G. (2003). Sodium channel inhibition by anandamide and synthetic cannabimimetics in brain. Brain Res. 978, 194–204. Nilius, B., Talavera, K., Owsianik, G., Prenen, J., Droogmans, G., and Voets, T. (2005). Gating of TRP channels: A voltage connection? J. Physiol. 567, 35–44. Niu, S. L., Mitchell, D. C., and Litman, B. (2002). Manipulation of cholesterol levels in rod disk membranes by methyl‐beta‐cyclodextrin: Effects on receptor activation. J. Biol. Chem. 277, 20139–20145. Niu, S. L., and Mitchell, D. C. (2005). EVect of packing density on rhodopsin stability and function in polyunsaturated membranes. Biophys. J. 89, 1833–1840. Ogata, N., and Narahashi, T. (1989). Block of sodium channels by psychotropic drugs in single guinea‐pig cardiac myocytes. Br. J. Pharmacol. 97, 905–913. Olcese, R., Usai, C., Maestrone, E., and Nobile, M. (1994). The general anesthetic propofol inhibits transmembrane calcium current in chick sensory neurons. Anesth. Analg. 78, 955–960.
336
Morris and Juranka
Okamoto, T., Harnett, M. T., and Morikawa, H. (2006). Hyperpolarization‐activated cation current (Ih) is an ethanol target in midbrain dopamine neurons of mice. J. Neurophysiol. 95, 619–626. Oliver, D., Lien, C. C., Soom, M., Baukrowitz, T., Jonas, P., and Fakler, B. (2004). Functional conversion between A‐type and delayed rectifier Kþ channels by membrane lipids. Science 304, 265–270. Ou, X., Blount, P., HoVman, R. J., and Kung, C. (1998). One face of a transmembrane helix is crucial in mechanosensitive channel gating. Proc. Natl. Acad. Sci. USA 95, 11471–11475. Ouyang, W., Wang, G., and Hemmings, H. C., Jr. (2003). Isoflurane and propofol inhibit voltage‐gated sodium channels in isolated rat neurohypophysial nerve terminals. Mol. Pharmacol. 64, 373–381. Park, J. B., Kim, H. J., Ryu, P. D., and Moczydlowski, E. (2003). EVect of phosphatidylserine on unitary conductance and Ba2þ block of the BK Ca2þ‐activated Kþ channel: Re‐examination of the surface charge hypothesis. J. Gen. Physiol. 121, 375–397. Pascarel, C., Brette, F., and Le Guennec, J. Y. (2001). Enhancement of the T‐type calcium current by hyposmotic shock in isolated guinea‐pig ventricular myocytes. J. Mol. Cell. Cardiol. 33, 1363–1369. Pathak, M., Agarwal, G., Kurtz, L., Gandhi, C. S., and IsacoV, E. Y. (2006). A fluorescence scan to investigate the structural dynamics of the Shaker Kþ channel. Abstract #2268 Biophysical Society. Pawson, L., and Bolanowski, S. J. (2002). Voltage‐gated sodium channels are present on both the neural and capsular structures of Pacinian corpuscles. Somatosens. Mot. Res. 19, 231–237. Pencer, J., Nieh, M. P., Harroun, T. A., Krueger, S., Adams, C., and Katsaras, J. (2005). Bilayer thickness and thermal response of dimyristoylphosphatidylcholine unilamellar vesicles containing cholesterol, ergosterol and lanosterol: A small‐angle neutron scattering study. Biochim. Biophys. Acta 1720, 84–91. Piao, L., Li, Y., Li, L., and Xu, W. X. (2001). Increment of calcium‐activated and delayed rectifier potassium current by hyposmotic swelling in gastric antral circular myocytes of guinea pig. Acta Pharmacol. Sin. 22, 566–572. Piao, L., Ying Li, H., Park, C. K., Cho, I. H., Piao, Z. G., Jung, S. J., Choi, S. Y., Lee, S. J., Park, K., Kim, J. S., and Oh, S. B. (2006). Mechanosensitivity of voltage‐gated K(þ) currents in rat trigeminal ganglion neurons. J. Neurosci. Res. 83, 1373–1380. Poling, J. S., Rogawski, M. A., Salem, N., Jr., and Vicini, S. (1996). Anandamide, an endogenous cannabinoid, inhibits Shaker‐related voltage‐gated Kþ channels. Neuropharmacology 35, 983–991. Ramsey, I. S., Moran, M. M., Chong, J. A., and Clapham, D. E. (2006). A voltage‐gated proton‐selective channel lacking the pore domain. Nature 440, 1213–1216. Riechmann, L., Lavenir, I., de Bono, S., and Winter, G. (2005). Folding and stability of a primitive protein. J. Mol. Biol. 348, 1261–1272. Rodriguez, B. M., Sigg, D., and Bezanilla, F. (1998). Voltage gating of Shaker Kþ channels. The eVect of temperature on ionic and gating currents. J. Gen. Physiol. 112, 223–242. Sako, Y., Nomura, N., Uchida, A., Ishida, Y., Morii, H., Koga, Y., Hoaki, T., and Maruyama, T. (1996). Aeropyrum pernix gen. nov., sp. nov., a novel aerobic hyperthermophilic archaeon growing at temperatures up to 100 degrees C. Int. J. Syst. Bacteriol. 46, 1070–1077. Scarlata, S., McBath, H., and Haspel, H. C. (1995). EVect of lipid packing on the conformational states of purified GLUT‐1 hexose transporter. Biochemistry 34, 7703–7711. Shahidullah, M., Harris, T., Germann, M. W., and Covarrubias, M. (2003). Molecular features of an alcohol binding site in a neuronal potassium channel. Biochemistry 42, 11243–11252.
11. Lipid Stress at Play: Mechanosensitivity of Voltage‐Gated Channels
337
Shcherbatko, A., Ono, F., Mandel, G., and Brehm, P. (1999). Voltage‐dependent sodium channel function is regulated through membrane mechanics. Biophys. J. 77, 1945–1959. Shiraishi, M., and Harris, R. A. (2004). EVects of alcohols and anesthetics on recombinant voltage‐gated Naþ channels. J. Pharmacol. Exp. Ther. 309, 987–994. Strege, P. R., Holm, A. N., Rich, A., Miller, S. M., Ou, Y., Sarr, M. G., and Farrugia, G. (2003). Cytoskeletal modulation of sodium current in human jejunal circular smooth muscle cells. Am. J. Physiol. Cell Physiol. 284, C60–C66. Su, Z., Martin, R., Cox, B. F., and Gintant, G. (2004). Mesoridazine: An open‐channel blocker of human ether‐a‐go‐go‐related gene Kþ channel. J. Mol. Cell. Cardiol. 36, 151–160. Tabarean, I. V., and Morris, C. E. (2002). Membrane stretch accelerates activation and slow inactivation in Shaker channels with S3‐S4 linker deletions. Biophys. J. 82, 2982–2994. Tabarean, I. V., Juranka, P., and Morris, C. E. (1999). Membrane stretch affects gating modes of a skeletal muscle sodium channel. Biophys. J. 77, 758–774. Tanaka, T., Tamba, Y., Masum, S. M., Yamashita, Y., and Yamazaki, M. (2002). La(3þ) and Gd(3þ) induce shape change of giant unilamellar vesicles of phosphatidylcholine. Biochim. Biophys. Acta 1564, 173–182. Teisseyre, A., and Michalak, K. (2003). The voltage‐ and time‐dependent blocking effect of trifluoperazine on T lymphocyte Kv1.3 channels. Biochem. Pharmacol. 65, 551–561. Terakawa, S., and Nakayama, T. (1985). Are axoplasmic microtubules necessary for membrane excitation? J. Membr. Biol. 85, 65–77. Tillman, T. S., and Cascio, M. (2003). EVects of membrane lipids on ion channel structure and function. Cell Biochem. Biophys. 38, 161–190. Thomas, D., Wu, K., Kathofer, S., Katus, H. A., Schoels, W., Kiehn, J., and Karle, C. A. (2003). The antipsychotic drug chlorpromazine inhibits HERG potassium channels. Br. J. Pharmacol. 139, 567–574. Tombola, F., Pathak, M. M., and IsacoV, E. Y. (2005). How far will you go to sense voltage? Neuron 48, 719–725. Tombola, F., Pathak, M. M., and IsacoV, E. Y. (2006). How does voltage open an ion channel? Annu. Rev. Cell Dev. Biol. May 5 [Epub ahead of print]. Toselli, M., Biella, G., Taglietti, V., Cazzaniga, E., and Parenti, M. (2005). Caveolin‐1 expression and membrane cholesterol content modulate N‐type calcium channel activity in NG108–15 cells. Biophys. J. 89, 2443–2457. Traub, O., Ishida, T., Ishida, M., Tupper, J. C., and Berk, B. C. (1999). Shear stress‐mediated extracellular signal‐regulated kinase activation is regulated by sodium in endothelial cells. Potential role for a voltage‐dependent sodium channel. J. Biol. Chem. 274, 20144–20150. Treistman, S. N., and Wilson, A. (1987). Alkanol eVects on early potassium currents in Aplysia neurons depend on chain length. Proc. Natl. Acad. Sci. USA 84, 9299–9303. Van den Brink‐van der Laan, E., Chupin, V., Killian, J. A., and de KruijV, B. (2004). Small alcohols destabilize the KcsA tetramer via their eVect on the membrane lateral pressure. Biochemistry 43, 5937–5942. Wallace, C. H., Baczko, I., Jones, L., Fercho, M., and Light, P. E. (2006). Inhibition of cardiac voltage‐gated sodium channels by grape polyphenols. Br. J. Pharmacol. 149, 657–665. Wang, J., Wang, H., Han, H., Zhang, Y., Yang, B., Nattel, S., and Wang, Z. (2001). Phospholipid metabolite 1‐palmitoyl‐lysophosphatidylcholine enhances human ether‐a‐go‐ go‐related gene (HERG) K(þ) channel function. Circulation 104, 2645–2648. Wiggins, P., and Phillips, R. (2005). Membrane‐protein interactions in mechanosensitive channels. Biophys. J. 88, 880–902. Wong, W., and Schlichter, L. C. (2004). DiVerential recruitment of Kv1. 4 and Kv4. 2 to lipid rafts by PSD‐95. J. Biol. Chem. 279, 444–452.
338
Morris and Juranka
Xiao, Y. F., Ma, L., Wang, S. Y., Josephson, M. E., Wang, G. K., Morgan, J. P., and Leaf, A. (2006). Potent block of inactivation‐deficient Naþ channels by n‐3 polyunsaturated fatty acids. Am. J. Physiol. Cell Physiol. 290, C362–C370. Xu, W. X., Li, Y., Wu, L. R., and Li, Z. L. (2000). EVects of diVerent kinds of stretch on voltage‐dependent calcium current in antrial circular smooth muscle cells of the guinea‐pig. Sheng Li Xue Bao 52, 69–74. Yang, X. C., and Sachs, F. (1989). Block of stretch‐activated ion channels in Xenopus oocytes by gadolinium and calcium ions. Science 243, 1068–1071. Yao, X., McIntyre, M. S., Lang, D. G., Song, I. H., Becherer, J. D., and Hashim, M. A. (2005). Propranolol inhibits the human ether‐a‐go‐go‐related gene potassium channels. Eur. J. Pharmacol. 519, 208–211. Ying, S. W., Abbas, S. Y., Harrison, N. L., and Goldstein, P. A. (2006). Propofol block of I(h) contributes to the suppression of neuronal excitability and rhythmic burst firing in thalamocortical neurons. Eur. J. Neurosci. 23, 465–480.
CHAPTER 12 Hair Cell Mechanotransduction: The Dynamic Interplay Between Structure and Function Anthony J. Ricci* and Bechara Kachar{ *Department of Otolaryngology, Stanford University, Stanford, California 94305 { Section of Structural Biology, National Institutes of Deafness and Communicative Disorders, Bethesda, Maryland 20892
I. II. III. IV. V. VI. VII. VIII. IX. X. XI.
Overview Auditory System Hair Bundle Structure MET Involves Mechanically Gated Channels Where Are These Channels? The Gating Spring Theory How Are the Channels Activated? To be or Not to be Tethered Characterizing Channel Properties? MET Channel Pore Adaptation A. Motor Adaptation B. Multiple Components of Adaptation C. Fast Adaptation D. Functional Role of Adaptation XII. The Dynamic Hair Bundle XIII. Summary and Future Directions References
I. OVERVIEW Hair cells are capable of detecting mechanical vibrations of molecular dimensions and to do this at frequencies in the 10s to 100s of kHz. This remarkable feat is accomplished by the interplay of mechanically gated ion Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59012-X
340
Ricci and Kachar
channels located near the top of a complex and dynamic sensory hair bundle. The hair bundle is composed of a series of actin filled stereocilia that has both active and passive mechanical components as well as a highly active turnover process whereby the components of the hair bundle are rapidly and continually recycled. Hair bundle mechanical properties will have significant impact on the gating of the mechanically activated channels and delineating between attributes intrinsic to the ion channel and those imposed by the channel’s microenvironment is often diYcult. The goal of this chapter is to delineate between what is known and accepted regarding hair cell mechanotransduction and what remains to be explored, particularly in relation to the interplay between hair bundle properties and mechanotransducer channel response. In addition, the interplay between hair bundle dynamics and mechanotransduction will be discussed.
II. AUDITORY SYSTEM The peripheral auditory system is a remarkable feat of evolutionary biological engineering with a threshold of sensitivity at molecular dimensions; where stimulus energy levels are below energy levels associated with the Brownian motion of the sensory organelle, the hair bundle (Denk and Webb, 1992; Jaramillo and Wiesenfeld, 1998). The dynamic range encompasses more than six orders of magnitude without damage or saturation. The frequency range is from 10s to 100,000s of Hertz and frequency discrimination is less than 1 Hz. For a system that is thermodynamically challenged these characteristics are remarkable (Bialek, 1987). Although hair cells come in a variety of ‘‘flavors,’’ with diVerent innervation patterns and diVerent complements of ion channels, the one commonality to all hair cells is the presence of an apical hair bundle that when deflected activates a mechanically gated ion channel. That activation of mechanically gated channels in the sensory hair bundle underlies sensory processing in the auditory and vestibular system has been known for almost 30 years (Hudspeth and Corey, 1977). Many fundamental principles regarding this mechanoelectric transduction (MET) process, such as the gating spring theory, and the presence of adaptation have been elucidated and are consistent across a variety of species and end organs. Multiple functional consequences, such as providing mechanical amplification and filtering, extending the dynamic range of the hair cell, and setting the hair cell resting potential, have been ascribed to the adaptation and activation process. Much like the Hodgkin–Huxley model of the action potential, the gating spring model of mechanotransduction established a framework from which to investigate the properties of hair cell transduction.
12. Hair Cell Mechanotransduction
341
Similarly, this model provides a microscopic interpretation of macroscopic data that although consistently supported by results across species and end organs remain to be tested at the mechanistic level. The purpose of this chapter is to discuss MET and the dynamic properties of the sensory hair bundle including channel gating and adaptation; addressed in relation to the complex structure of the sensory hair bundle.
III. HAIR BUNDLE STRUCTURE Hair bundles come in a variety of shapes and sizes; however, there are some fundamental commonalities to them. Each hair bundle consists of rows of stereocilia, ordered in height, that form a staircase pattern (Fig. 1B and C ). The hair bundle resides on the apical surface of the hair cell and is positioned between the hair cell body and the overlying tectorial membrane (Fig. 1A), optimally located to sense any shearing motion between structures. Stereocilia are actin‐filled membrane protrusions of up to 100 mm in length (in vestibular hair bundles), where parallel and uniformly polarized actin filaments are tightly cross‐linked to form a paracrystalline structure. This paracrystalline actin core gives the stereocilia a rigidity that allows them to rotate about their base, rather than bend when stimulated (Crawford and Fettiplace, 1985). A variety of extracellular filaments connect stereocilia of adjacent rows. Ankle links, horizontal top connectors, shaft connectors, and the tip‐link (Bashtanov et al., 2004) together anchor the stereocilia so that they move as a unit when stimulated (Crawford and Fettiplace, 1985).
IV. MET INVOLVES MECHANICALLY GATED CHANNELS The hair bundle is the site of MET (Hudspeth and Corey, 1977; Tilney and Saunders, 1983). Initial work in frog saccule demonstrated that hair bundle deflection toward its tall edge reduced input resistance while movement toward the short cilia increased resistance, suggesting that an ion channel was being opened and closed in response to hair bundle deflection (Hudspeth and Corey, 1977). Measurements from a mammalian cochlea preparation similarly demonstrated an electrical response from outer hair cells (OHC) in response to sound pressure changes (Russell and Sellick, 1978). Directional sensitivity of the hair bundle was more quantitatively assessed (Shotwell et al., 1981) and results demonstrated a specific axis of sensitivity that aligned with the graded change in hair bundle height; sensitivity falling oV as a cosine function of the angle of rotation in either direction away from the most sensitive position. Directional sensitivity of the uniquely organized OHC
342
Ricci and Kachar A
B
C
D
FIGURE 1 Hair bundle structure. (A) Freeze fracture image of OHC with sensory hair bundle embedded in overlying tectorial membrane; arrows indicate tallest row of stereocilia in contact with tectorial membrane. (B) Scanning electron micrograph of OHC stereocilia showing the three rows of increasing height. Arrows indicate tip‐link connecting adjacent stereocilia and thought to provide directional sensitivity to the hair bundle. (C) Schematic drawing of OHC bundle showing the typical orientation of stereociliary rows. (D) Schematic of three stereocilia at rest and during stimulation. The stereocilia rotate about their base creating a shearing force near the tips where the tip‐link creates tension between adjacent rows. Arrow indicates movement that opens channels.
bundle has not been directly investigated, though evidence suggests that it would be consistent with what has been found in other hair cell types. The orientation of the OHC bundle (Fig. 1B) is significantly curved so that stimulation toward the tall edge of the hair bundle can encompass up to 180 of rotation. How this might eVect response properties remains to be determined. A direct activation of an ion channel was further supported by field recordings of microphonic potentials from an isolated sensory epithelial preparation; here the kinetics of the response excluded a multistepped signal transduction process (Corey and Hudspeth, 1979b). A modification of this
343
12. Hair Cell Mechanotransduction
preparation allowed voltage clamping of the epithelium, further demonstrating the rapid nature of mechanical sensitivity and illustrating a Ca2þ dependence to the activation kinetics (Corey and Hudspeth, 1983). A comparable set of experiments in mouse cochlea culture demonstrated an electrical response from hair bundle deflection consistent with ion channel gating (Russell et al., 1986). Direct measurements of activation kinetics in hair cells from turtle auditory papilla (Crawford et al., 1989) further validated the presence of mechanically gated ion channels in the sensory hair bundle. Single‐channel recordings of the MET channel (Ohmori, 1984; Crawford et al., 1991; Kros et al., 1992; Ricci, 2002) unequivocally demonstrated that hair bundle deflection activated MET channels in hair cells.
V. WHERE ARE THESE CHANNELS? Where in the hair bundle are these elusive channels located (Fig. 2)? Extracellular recordings first suggested that ion channels were located near the tops of the stereocilia (Hudspeth, 1982). This finding was supported by
A
B
C
Top insertion
Side insertion
FIGURE 2 Tip‐link and possible locations for the MET channel. Tip‐link connects rows of adjacent stereocilia (B). Mechanically gated channels are hypothesized to be located at either (A, C) or both ends of the tip‐link. A third possibility is at a dense region where the stereocilia come close together. Possible channel locations are indicated by white circles in panel (B). Also shown in the transmission electron micrograph of (B) is the tenting of the membrane at the top of the stereocilia, thought to be created by the tip‐link pulling on this membrane. (A, C) show possible configurations of the MET channel tethered to the tip‐link and cytoskeleton when located at either end of this link.
344
Ricci and Kachar
iontophoretic mapping of the location based on sensitivity to the channel blocker gentamycin (Jaramillo and Hudspeth, 1991), by the appearance of a Ca2þ ‘‘blush’’ near to the tops of the stereocilia, presumably created by Ca2þ entry through the channel (Lumpkin and Hudspeth, 1995), and antibody labeling to a putative MET channel‐binding site also suggested a location near to the tops of the stereocilia (Furness et al., 1996). Ca2þ gradients in the stereocilia, presumably created by entry through MET channels, also support a location near the tops of the stereocilia (Lumpkin and Hudspeth, 1998). Although existing data consistently puts the channels near to the top of the stereocilia, the precise location and relationship to accessory structures (like the stereociliary links) remains to be determined. Reports also exist for a channel located near the base of the stereocilia, a conclusion based on Ca2þ imaging experiments (Ohmori, 1988).
VI. THE GATING SPRING THEORY The recognition that mechanotransduction involved the gating of an ion channel driven by shearing forces created by hair bundle deflection led to the initial gating spring hypothesis (Corey and Hudspeth, 1983). In its most simple form, this hypothesis suggests that a tensed elastic element exists between stereocilia and the channel such that force is exerted onto the channel with stereocilia shearing (Fig. 3). Concomitant with the identification of a channel‐driven mechanism and the localization of the ion channel to the sensory hair bundle apical surface was the evaluation of hair bundle mechanics. Work in turtle auditory papilla demonstrated that the hair bundle bends as a unit, pivoting about its base, creating a shearing between stereocilia of adjacent rows (Crawford and Fettiplace, 1985; Fig. 1D). This work also provided the first evidence of active bundle movements—oscillations that might provide a mechanical filter to incoming sound waves. A hypothesis generated from the gating spring theory was that channel opening might alter hair bundle compliance (Hudspeth, 1982). The total hair bundle stiVness at a minimum is the sum of the passive stiVness and the gating compliance (Fig. 3). Estimates of gating compliance suggest that gating provides more than half of the hair bundle stiVness (Marquis and Hudspeth, 1997). The channel gate might be in series with the gating spring so that when channels open there is an increase in overall length, equivalent to the length of the gate, that momentarily slackens the gating spring reducing stiVness (increasing compliance; Fig. 3). A change in hair bundle compliance associated with MET channel gating was first reported in frog saccule (Howard and Hudspeth, 1988). A comparable change in compliance has also been
345
12. Hair Cell Mechanotransduction B
MET channel Gating spring Force
(1)
1.0
F Force (nN)
A
F
Open (3)
0.5 (2) 0.0
Closed (1)
(2) −0.5
(3) F
−40
0 40 80 Displacement (nm)
120
FIGURE 3 Gating spring theory. (A) Gating spring theory posits the presence of a spring through which force is applied to the channel gate. The channel gate opens in series with this spring, thus transiently reducing the force onto this spring and increasing compliance. The force displacement plot of (B) supports this simple model, illustrating a linear (Hookean) component of the plot when channels are either closed (1) or open (3), the slopes of which are the same given that the spring constant remains the same. The nonlinear component (2) represents the transition between closed and open where the channel gate (probability of opening of the channel) is increasing. This region can be described by the Boltzmann function that also describes the activation curve for the MET channel.
identified in mouse cochlea (Russell et al., 1992), turtle auditory papilla (Ricci et al., 2002), and rat cochlea (Kennedy et al., 2005), giving strong support to a common gating mechanism across hair cells of diVerent species and end organs. The gating spring theory has been formalized in a variety of ways (Markin and Hudspeth, 1995a; van Netten and Kros, 2000; Ricci et al., 2002) incorporating two and three state models. In its simplest form, the model suggests that the diVerence in energy for a given stimulus is equivalent to the diVerence in energy between the open and closed states of the channel, which in turn is equivalent to the displacement diVerence times a constant (z) that is composed of the gating spring constant (kgs) times the gating swing, (d) and so takes the form: A ¼ Ac
Ao ¼ zðx
xo Þ
ð1Þ
where A is energy associated with the closed (c) or open (o) state of the channel, x is the displacement, usually referring to movement at the top of the stereocilia. More general formulations can be found that do not require the assumption of two states (Howard and Hudspeth, 1988; van Netten and Kros, 2000; van Netten et al., 2003). From this simple equation it is clear
346
Ricci and Kachar
that no distinct mechanism is implied and that caution must be taken when ascribing physical components to the hypothesis. A practical example of this problem comes from the consistent estimate of d, the gating swing, at values in excess of 7 nm, a value much larger than would be predicted for the mechanical gate of a channel or of a conformational change in the channel as it goes from closed to open state (Howard and Hudspeth, 1988; Markin and Hudspeth, 1995a; van Netten et al., 2003; see schematic in Fig. 3). Similarly, estimates of the number of channels per hair bundle based on the single‐channel gating force estimates are more than an order of magnitude greater than estimates made using single‐channel conductance values (Ricci et al., 2002). These discrepancies do not negate the theory but do question the underlying mechanistic interpretation. It is possible to account for these discrepancies in several ways. First, the assumption in the gating spring model for the whole bundle is that the springs are in parallel, an assumption unlikely to be absolutely true on the microscopic level given the complex interconnections between stereocilia (Howard and Hudspeth, 1988; Markin and Hudspeth, 1995a; van Netten and Kros, 2000; Ricci et al., 2002). Channels in series, a likely outcome of a bundle organization with multiple rows of stereocilia, would mean summing of the gating springs (Fig. 4). At an extreme, the movement of the gating spring would be proportional to the number of rows of stereocilia and therefore might significantly reduce the length of the gating spring (Fig. 4). More quantitative information regarding linkages between stereocilia and the relative movements between stereocilia are needed to resolve this issue. Interestingly, estimates of gating swing in mammalian cochlear hair cells are closer to 2 nm per channel—values more attuned with the molecular dimensions of a channel. Interestingly, such lower values can be obtained by taking into account bundle organization that includes both series and parallel components (van Netten and Kros, 2000). A second possibility is that the gating swing does not solely represent a channel gate but includes elements in series with the channel conformational change. These elements could be protein or lipid. A third possibility is that the scaling value (often termed ) that converts the bundle motion at the top of the stereocilia to that at the channel could be wrong. As we do not know where the channels are located, there are assumptions associated with the estimates. In saccular hair cells,
is about 0.1–0.15 (Markin and Hudspeth, 1995a) while in mammalian hair cells more widely ranging values have been estimated (OHC’s: 0.05–1.0; vestibular hair cells: 0.02–0.04; Geisler, 1993; Pickles, 1993; Markin and Hudspeth, 1995a; Furness et al., 1997; van Netten and Kros, 2000). And finally, it is possible, though somewhat unlikely that the compliance change measured is not actually associated with the MET channel but represents an additional compliant component of the hair bundle.
12. Hair Cell Mechanotransduction
347
FIGURE 4 Stereociliary organization suggests both series and parallel contributions to the gating spring. Freeze fracture (A) of the apical surface of an OHC showing the insertion points of the stereocilia. Blue arrows indicate directional sensitivity and the portion of the bundle diagrammed in (B). (B) is a schematic representation of the circuit presumed to be generated by the orientation of the stereocilia, demonstrating that although most models have considered the gating springs to be in parallel, there is a significant series component to the structure.
VII. HOW ARE THE CHANNELS ACTIVATED? The identification of a link between stereocilia, located near to the tops and connecting stereocilia of adjacent rows and aligned along the axis of sensitivity of the hair bundle, added a morphological correlate to the gating spring theory (Pickles et al., 1984; Fig. 2B). This tip‐link appeared to be localized appropriately and to have properties associated with a putative gating spring. Deflection of the hair bundle would stretch the link, increasing tension onto the channel. Loss of the tip‐link results in loss of transduction (Assad et al., 1991); recovery of this link restored transduction in chick (Zhao et al., 1996). For the tip‐link to be the gating spring it needs to be inherently elastic (like a spring). Ultrastructural investigations revealed a coiled filamentous structure very unlikely to have the appropriate elastic properties (Kachar et al., 2000). Findings have implicated cadherin 23 as a component of the tip‐link complex (Siemens et al., 2004; Sollner et al., 2004). Other investigations have demonstrated cadherin 23 as a component of side‐ links and questioned their role as the tip‐link, in particular because detection of cadherin 23 in mature hair bundles is limited (Gillespie et al., 2005; Michel et al., 2005). Modeling of the elastic properties of cadherin 23 suggests it
348
Ricci and Kachar
cannot be the gating spring because it is too stiV (Sotomayor et al., 2005). Here too the model must be interpreted cautiously because it models only a portion of the molecule and makes assumptions regarding the structural organization of the hair bundle, that is, that macroscopic gating force estimates can be accurately interpreted at the single‐channel level. Together these findings suggest that perhaps the tip‐link serves as a tether, translating the force associated with hair bundle deflection to the ion channel, as some element is required to translate the shearing motion of the stereocilia into a force exerted onto the channel. Determining the role of the tip‐link in transduction requires revisiting the site of the MET channels. Figure 2 depicts two stereocilia with the tip‐link in between. The channel being located near to the top of the stereocilia suggests it might be located at either end or both ends of the tip‐link (Hudspeth, 1989; Hudspeth and Gillespie, 1994; Corey, 2003). Another argument considers the tip‐link simply as a structural element serving to keep the stereocilia in close approximation and places the channels at a location, based on immunocytochemistry, below the tip‐link, where the stereocilia are juxtaposed (Furness et al., 1996, 2002). A third possibility is that the tip‐link serves as a tether but does not directly couple to the MET channels, rather it serves to stretch the membrane and the membrane exerts force onto the channel (Kachar et al., 2000; Fig. 5). Schematic representations of these possibilities are given in Figs. 2 and 4. The only experimental argument suggesting that MET channels exist at both ends of the tip‐link is from Ca2þ imaging data (Denk et al., 1995). The argument states that the geometry of the bundle requires MET channels be located on the side of the tall stereocilia (top end of the tip‐link) if there is a Ca2þ signal present in the tallest row of stereocilia. If a Ca2þ signal were detected in the shortest row of stereocilia, then channels must be located at the tops of the stereocilia as well. Both results were observed. Unfortunately, the quantification of the links near the top of tallest stereocilia is lacking. Horizontal links may provide additional stimulation between tall stereocilia thus making interpretations regarding the channel location at the side questionable. In addition, the argument that channels need be near the sterocilia top (base of the tip‐link) because of a signal in the lowest rank of stereocilia does not require that the channel be directly coupled to the tip‐link. Other than these Ca2þ imaging data, there is no direct evidence to delineate between channel locations at the microscopic level. The gating spring model can accommodate all of these possibilities. The tenting observed in the top of the stereocilia (Fig. 2B) is suggestive of an increased membrane tension (Kachar et al., 2000). This tenting is lost when the tip‐link is disrupted (Rzadzinska et al., 2005). As will be discussed further below, identifying the precise location of the MET channel is critical to unraveling some of the mechanical constraints on how the system might operate at the molecular level.
12. Hair Cell Mechanotransduction
349
FIGURE 5 How is the MET channel gated? (A–C) The direction that force is sensed by these mechanically gated channels is unknown. Force can be represented either perpendicular (B) or parallel to the membrane (C). Mechanically gated channels of other systems have been shown to be sensitive to either so both are real possibilities for hair cells. (D) illustrates the classical hypothesis that the MET channel is directly tethered to the tip‐link; however, direct evidence for this tethering does not exist and the possibility that the channel is nontethered remains to be rigorously explored (E). The tip‐link can exert force onto the channel either directly or indirectly by tenting the membrane.
VIII. TO BE OR NOT TO BE TETHERED As the molecular identity of the MET channel remains unknown, the mechanism by which the channel is activated is also a point of much speculation. Illustrations like those shown in Fig. 2 have reinforced the concept that the MET channel must be tethered extracellularly as well as to the cytoskeleton. In addition, the ubiquitous presence of the tip‐link in diVerent hair cell types furthered the argument that the tip‐link is involved in channel gating and that the channel must be tethered. In fact, adjacent rows of stereocilia need to be tethered in order for force to be translated to the
350
Ricci and Kachar
channel; no data exists suggesting that the channel requires a direct tethering. A mechanically gated channel can be activated by forces exerted either parallel or perpendicular to the lipid bilayer. Many mechanoreceptors work without tethers and sense force via the plasma membrane (Sukharev et al., 1997; Kung, 2005). The bacterial mechanogated channels, the MscL family, are activated by forces generated by the hydrophobic interactions between the lipid and the channel protein—forces that are applied parallel to the membrane (Sukharev et al., 1997; Kung, 2005). On the other hand, in C. elegans touch receptors, evidence exists for tethering from both extracellular and intracellular sites at least in part via microtubules (Huang et al., 1995; Du et al., 1996; O’Hagan and Chalfie, 2006). However, evidence has begun to question the requirement of intracellular tethers, particularly the larger microtubules (microtubules composed of 15 protofilaments; O’Hagan and Chalfie, 2006). Both TREK and TRAAK are two mechano‐gated potassium channels that are sensitive to membrane stretch and do not require cytoskeletal or extracellular tethering (Maingret et al., 1999; Patel et al., 2001). In hair cells, direct evidence regarding gating or tethering is sparse. Schematic representations of tethered versus nontethered activation of the hair cell MET channel are given in Fig. 5. In the tethered version, the tip‐link is shown directly coupled to the channel so that force is exerted perpendicular to the membrane while in the nontethered version the tip‐link is shown serving as a lever pulling on the membrane, exerting force parallel to the bilayer. The two examples represent the extreme cases (despite the tethered version being the current model), the two additional possibilities, tethered only externally or tethered only internally are not shown but could equally explain existing data. Interestingly, the gating spring theory can account for any of these configurations, perhaps slightly better for the nontethered version where the gating swing would now include a lipid bilayer component being pulled and thus may be predicted to be larger (Fig. 5). Hypotheses have suggested that a series of ankyrin repeats at the terminal region of a channel could have spring‐like properties that might represent the molecular correlate of the gating spring (Howard and Bechstedt, 2004; Sotomayor et al., 2005). Without more accurate data regarding single‐ channel gating forces and without molecular identification of the channel, these hypotheses, though novel and exciting, remain to be tested. Presently, data does not exist to determine whether the MET channel senses force exerted perpendicular or parallel to the membrane. The two possibilities are depicted in Fig. 5 and MET channels sensitive to either force direction have been identified (see other chapters). Limiting the ability to determine the mechanism of channel activation are several factors including identifying channel location, the channel molecular nature, and the mechanism of stimulation. Hair bundle stimulation has taken a variety of forms, fluid jet,
12. Hair Cell Mechanotransduction
351
stiV or flexible fibers attached to piezo‐driven actuators, or optical tweezers (Corey and Hudspeth, 1980; Crawford and Fettiplace, 1985; Ohmori, 1985, 1987; Howard and Hudspeth, 1987; Crawford et al., 1989; Kros et al., 1992; Benser et al., 1993; Jaramillo and Hudspeth, 1993; Holt et al., 1997; Ricci et al., 2000; Kros et al., 2002; Kennedy et al., 2003; Vollrath and Eatock, 2003; Ricci et al., 2005; Cheung and Corey, 2006). The major limitation to these methods is that the stimulus to the channel is filtered via hair bundle mechanics, mechanics that are not being controlled or monitored at the molecular level. The most obvious example of this is adaptation, a process that resets the molecular orientation of the channel with reference to the hair bundle position occurs when stiV probes, meant to be a displacement clamp are used (Fig. 8). If the bundle were clamped at the molecular level, adaptation would not occur. That is, adaptation requires tension on the MET channel to be relieved, this can only occur if there is a physical movement within the hair bundle, a movement that should be eliminated by a true displacement clamp. The problem is akin to trying to interpret voltage clamp data that is not properly space clamped. Development of new methodologies is needed for investigations of MET channel and mechanics at the single (or paired) stereocilium level in order to more directly investigate gating mechanisms.
IX. CHARACTERIZING CHANNEL PROPERTIES? Separating properties intrinsic to the MET channel from those imposed onto the channel from accessory proteins and hair bundle mechanics is diYcult largely due to the problems described above. Not having absolute control over the micromechanics of the hair bundle limits the ability to directly probe molecular mechanisms. A clear example of this problem is the investigations of channel kinetics. Activation kinetics has been inferred in frog from macroscopic measurements (Corey and Hudspeth, 1983) and has been measured directly in turtle (Crawford et al., 1989; Ricci et al., 2005) and rat (Crawford et al., 1989; Ricci et al., 2005). In both frog and turtle, the kinetics were Ca2þ‐dependent (Corey and Hudspeth, 1983; Ricci et al., 2005). In turtle, kinetics varied with characteristic frequency of the hair cell, suggesting variations in channel structure. In rat, the kinetics are too fast to accurately measure, implying that they are at least an order of magnitude faster than in turtle or frog (Crawford et al., 1989; Ricci et al., 2005), further suggesting variations in channel structure. However, not having a direct measure of the force exerted onto the channel makes a determination regarding the nature and the underlying mechanism of the kinetic diVerence diYcult. It is as likely that hair bundle mechanics vary allowing a faster force
352
Ricci and Kachar
application to the channel in mammals and in higher frequency hair cells as it is that the intrinsic channel properties vary tonotopically. Without knowing the rate‐limiting step, conclusions regarding mechanism are speculation. So even channel properties normally considered intrinsic to the structure of the channel protein, like activation kinetics, must be evaluated carefully for the hair cell MET channel.
X. MET CHANNEL PORE What channel properties might be considered intrinsic to the channel? Given that the channel identity remains elusive, creating a profile of properties serves as an important tool for correctly identifying the channel protein. Most likely properties associated with the channel pore, like permeation, rectification, and pharmacology, will be intrinsic to the channel protein. The MET channel is a nonspecific cation channel (Corey and Hudspeth, 1979a; Ohmori, 1985, 1989; Crawford et al., 1989; Kros et al., 1992; Farris et al., 2004). MET channels show little rectification in high extracellular Ca2þ solutions but a slight inward rectification appears when Ca2þ is lowered (Crawford et al., 1989; Kros et al., 1992; Farris et al., 2004). The channels have a high Ca2þ permeability (Ohmori, 1985; Crawford et al., 1991; Lumpkin et al., 1997; Ricci and Fettiplace, 1998). Ca2þ both permeates and blocks the channel with a half blocking [Ca2þ] of 1 mM (Crawford et al., 1991; Kros et al., 1992; Lumpkin et al., 1997; Ricci and Fettiplace, 1998; Gale et al., 2001), likely as a function of ion interactions within the channel pore (Lumpkin et al., 1997). Ca2þ binds within the pore at a distance equivalent to about 0.5 of the distance into the electric field (Kros et al., 1992; Gale et al., 2001; Farris et al., 2004). The pharmacology of the MET channel is unusual in that many compounds serve as open channel blockers (Farris et al., 2004). The major properties of molecules thought to be MET channel blockers are their being positively charged so as to be driven into the channel electrochemically and the molecule being of suYcient size to plug the pore (Farris et al., 2004). Aminoglycosides have long been known to block hair cell MET channels (Kroese et al., 1989; Kimitsuki and Ohmori, 1993; Glowatzki et al., 1997; Ricci, 2002; Marcotti et al., 2005; Waguespack and Ricci, 2005). The block appears to hold the channel in its open state (Kroese et al., 1989; Denk et al., 1992; Jaramillo and Hudspeth, 1993;) and the eYcacy of block is directly related to the probability of opening of the channel (Ricci, 2002). Evidence suggests that aminoglycosides are permeable blockers of the channel but also report the unusual finding that the permeability works with external application but not with internal application (Marcotti et al., 2005). This unusual finding was also observed with the permeable blocker FM1–43 (Gale et al., 2001).
353
12. Hair Cell Mechanotransduction
Together these data suggest that the channel may have an open state at positive potentials that is diVerent than that at negative potentials. This hypothesis was first suggested by the complex blocking eVects of amiloride on hair cell MET channels (Jorgensen and Ohmori, 1988; Rusch et al., 1994) and also in oocyte mechanosensitive channels (Lane et al., 1993). The potential for an additional open state is indirectly supported by evidence suggesting channel rectification when the hair bundle is placed in lowered Ca2þ solutions (Crawford et al., 1989) and by single‐channel measurements that support this rectification (Ricci et al., 2003). The pharmacological profile established for the MET channel overlaps with that of several classes of channels including cyclic nucleotide gated channels, transient receptor potential channels, Ca2þ channels, and nicotinic receptor channels (Farris et al., 2004). Using techniques established for investigating nicotinic pore properties (Adams et al., 1980) and applied to other mechanosensitive channels (Cruickshank et al., 1997), sodium channels (Hille, 1971), and NMDA receptor channels (Zarei and Dani, 1994), the hair cell pore dimensions were estimated from the external face. A summary schematic of the MET channel is presented in Fig. 6 that depicts relative pore dimensions (Farris et al., 2004). These estimates suggest a large pore size, befitting the large single‐channel conductance measurements (Ricci et al., 2003). Unusually, there was no diVerence observed in pore dimensions between frequency location despite there being a diVerence in single‐channel
1.8 nm Extracellular
3.1 nm 1.2 nm Intracellular
Ca++ binding site Negative charges FIGURE 6 Illustration representation of the pore of the MET channel with dimensions estimated from ionic permeabilities (Farris et al., 2004). Width of the external face was estimated based on pharmacological antagonists. Ca2þ‐binding sites both intracellular and extracellular are posited based on changes in channel kinetics (Ricci et al., 2005). The negative charges near the central region of the channel have been estimated based on the current–voltage responses.
354
Ricci and Kachar
conductance (Ricci et al., 2003; Farris et al., 2004). Similarly, no diVerence in Ca2þ permeation has been identified between frequency locations (Ricci, 2002). Single‐channel properties have been measured in turtle (Crawford et al., 1991; Ricci et al., 2003), chick (Ohmori, 1984), and mouse (Kros et al., 1992). The single‐channel conductance increases with characteristic frequency (Ricci et al., 2003)—a property thought to in part underlie tonotopic diVerences in adaptation kinetics (see below; Ricci and Fettiplace, 1997). The single‐channel conductance was also sensitive to external Ca2þ, increasing as Ca2þ was lowered (Ricci et al., 2003). All of the single‐channel data must be evaluated carefully in part because of the unusual manner in which the measurements are made. Whole‐cell recordings of single‐channel properties are limited in resolution due to membrane noise and filtering imposed by having the entire cell membrane in the electrical circuit. Obtaining single channels by disrupting the hair bundle with BAPTA may have unrecognized consequences. An example of single‐channel recordings is given in Fig. 7. These results demonstrate the presence of a mechanically sensitive channel in the stereocilia that has an apparent large single‐channel conductance. Because of the filtering and signal:noise diYculties arise resolving subconductance or flickering behavior so that it is possible that other states exist that have not yet been characterized. Measurements using noise analysis obtained much smaller values for this conductance, values likely diYcult to observe with direct measurements (Holton and Hudspeth, 1986). Although noise analysis typically underestimates conductance values, the diVerence (about an order of magnitude) is greater than predicted by the error associated with the technique and may suggest that the sensitivity of the single‐ channel recordings is limited. The single‐channel conductance estimates when compared with macroscopic maximal current responses suggest one or at most two channels per stereocilia (Crawford et al., 1991; Kros et al., 1992; Ricci et al., 2003).
XI. ADAPTATION To this point, a general overview of transduction and the gating spring theory has been presented that depicts the hair bundle as a passive element and channel gating as the active element. These assumptions will be further elucidated and challenged below, but first the third major component to mechanotransduction, adaptation, need be formally introduced. Figure 8 presents an example of MET currents elicited from mechanical deflection of a hair bundle in the rat cochlea and turtle auditory papilla. The currents activate rapidly and then, despite a constant stimulus, decay with a time course that is dependent on the stimulus intensity and typically has two
355
12. Hair Cell Mechanotransduction A
∆x (nm) −76 10 pA
76 114 c o
190 10 ms
300 nm 0
0 P OPEN
Current (pA)
B
−5 1
−10 0 C
10
20
30
40
50 ms
50 nm
10 pA
10 ms FIGURE 7 (A) Single‐channel recordings of MET currents from turtle auditory hair cells (Ricci et al., 2003). (B) Averages of single‐channel recordings from increasing stimuli demonstrating that the single‐channel behavior is similar to the macroscopic behavior (that is open time increases with increased stimulus). (C) is an enlarged example of single‐channel recording meant to demonstrate the limitation of recording single channels in the whole‐cell mode. Flickering behavior was observed (red line) that could indicate additional conductance state or could simply reflect a limited voltage clamp speed. Blue line indicates level used to estimate single‐channel conductance.
356
Ricci and Kachar
A
B
C
Rat outer hair cell
Turtle auditory hair cell 1.2
Current (nA)
1.0 0.8 0.6 0.4 0.2 200 pA 10 ms
0.0 −0.2
0.0 0.2 0.4 Displacement (nm)
FIGURE 8 Transduction and adaptation are similar between species. Voltage clamp recordings of MET currents from a rat OHC preparation (A) and a turtle auditory papilla hair cell (C). Qualitatively these responses are quite similar. The peak current vs voltage plot for the OHC response is shown in (B) and fit with a double Boltzmann function.
components (Wu et al., 1999). The decay in current is termed adaptation. Adaptation was first described in frog saccule (Eatock et al., 1987). The process is Ca2þ‐dependent, underlies several important physiological processes, and involves multiple mechanisms (Eatock, 2000; Hudspeth, 2005; Lemasurier and Gillespie, 2005; Fettiplace and Hackney, 2006). Generally, adaptation is a resetting of the relationship between hair bundle position and force sensed by the MET channel. Deflection of the hair bundle toward the tallest rows increases tension in the hair bundle opening channels, Ca2þ enters driving a reduction in tension sensed by the MET channel resulting in channel closure. A reduction in tension elicited by hair bundle deflection away from the tallest rows results in an opposite phenomenon where channels initially close, reducing Ca2þ in the stereocilia leading to an increase in hair bundle tension that reopens channels. The adaptation processes create a Ca2þ‐ dependent feedback that sets the resting open probability of the channel (Ricci et al., 1998). This feedback system, integrated with other stereociliary Ca2þ homeostatic mechanisms (like Ca2þ ATPases and buVers) serve to
12. Hair Cell Mechanotransduction
357
maintain Ca2þ at a constant steady‐state level. A consequence of adaptation then is the shifting of the current displacement plot depending on whether intraciliary Ca2þ is increased (rightward shift) or decreased (leftward shift). A variety of pieces of data support this basic description of adaptation. The classical experiment compares the current–displacement (I–X) plot about the hair bundle resting position against the plot elicited when the bundle is biased by a known amount (Crawford et al., 1989). Measurements of hair bundle compliance show an increased compliance that correlates with slow adaptation rates (Howard and Hudspeth, 1987; Assad et al., 1989; Ricci et al., 2000; Cheung and Corey, 2006). Disruption of tip‐links results in a bundle ‘‘relaxation’’ movement indicating a standing tension in the bundle, presumably established by adaptive forces (Assad et al., 1991). Voltage‐dependent hair bundle movements also support this basic description (Assad et al., 1989; Cheung and Corey, 2006; Ricci et al., 2002).
A. Motor Adaptation Adaptation was first suggested to be a myosin‐driven process based on its Ca2þ sensitivity and change in hair bundle compliance (Eatock et al., 1987; Howard and Hudspeth, 1987; Assad and Corey, 1992). The premise of the model suggests that the MET channels are tethered to the actin cytoskeleton by myosin, Ca2þ entry triggers the release of myosin from the actin resulting in a slippage of the channel down the stereocilia, reducing tension in the gating spring closing channels. When channels are closed, the myosin climbs the actin restoring tension to the gating spring. Implicit with the classical view of adaptation is that the channels are located along the side insertion of the tip‐link so that myosin can move up and down the actin and also that the channel is tethered to the cytoskeleton. EVects on channels located near the top of the stereocilia would be indirect via translation through the tip‐ link. A variety of evidence exists supporting the basic hypothesis that myosin is involved in adaptation. The process is Ca2þ‐dependent (Corey and Hudspeth, 1983; Crawford et al., 1989, 1991; Hudspeth and Gillespie, 1994; Benser et al., 1996; Walker and Hudspeth, 1996; Ricci and Fettiplace, 1998; Ricci et al., 1998). Interfering with the myosin cycle alters adaptation (Gillespie and Hudspeth, 1993; Wu et al., 1999). Identification of myosin 1C isozymes in the hair bundle and its immunolocalization at the tip‐link insertion sites also implicated this isozyme in adaptation (Gillespie et al., 1993; Metcalf et al., 1994). The calmodulin‐dependence of adaptation (Walker and Hudspeth, 1996) indirectly implicated myosin 1C as direct interactions between calmodulin and myosin 1C have been observed (Cyr et al., 2002). A novel chemical‐ genetic strategy provides the most direct evidence implicating myosin 1C
358
Ricci and Kachar
(Holt et al., 2002). The ATP‐binding site of myosin 1C was altered making it selectively vulnerable to a modified ATP analogue. Incorporation of this modified myosin into hair cells allowed evaluation of its role in adaptation, where adaptation was reduced when the myosin cycle was interrupted (Holt et al., 2002). Recent construction of a mouse permanently modified with this construct has confirmed these initial findings in vestibular hair cells (StauVer et al., 2005). Confirmation of these results in the auditory system remains as does the direct link to hair bundle mechanics that this mouse should provide. Several elegant experiments have investigated myosin 1C force‐generating properties at the single molecule level; this work argues that the myosin properties are ideal for an adaptation motor in that they have a strain‐sensing ADP release mechanism and two movements associated with the head group (Batters et al., 2004a,b). Although a variety of evidence supports a role for myosin 1C in hair cell adaptation, this mechanism is by no means solved. Whether motor adaptation exists in mammalian OHCs may be questioned by the current recordings which show little slow decay in current, being largely the fast component of adaptation (Kennedy et al., 2003; Section XI.C). In addition, immunocytochemistry (Schneider et al., 2006) shows a much more diVuse pattern of labeling along the stereocilia as compared to originally reported (Garcia et al., 1998; Steyger et al., 1998) not necessarily consistent with the conventional interpretation of its role in adaptation. Given that myosins have many roles in cellular function and maintenance care must be taken when ascribing a particular function to these ubiquitous proteins (Hasson and Mooseker, 1997; Friedman et al., 1999; Krendel and Mooseker, 2005).
B. Multiple Components of Adaptation From the early work investigating adaptation, a discrepancy existed, where data from frog implicated a motor mechanism with adaptation rates in the tens of milliseconds (Eatock et al., 1987; Howard and Hudspeth, 1987; Assad et al., 1989; Assad and Corey, 1992), whereas data from turtle auditory papilla suggested millisecond time courses and initially did not find a mechanical correlate of adaptation thereby implicating a channel mechanism (Crawford and Fettiplace, 1985; Crawford et al., 1989). The discrepancy was furthered when the mechanical response of turtle hair bundles was shown to be in the opposite direction to that reported in frog (Assad et al., 1989; Ricci et al., 2000). However, this work also demonstrated that a second mechanical response could be obtained depending on hair bundle resting position (Ricci et al., 2002). In addition, the kinetics of adaptation as well as pharmacological sensitivities suggested perhaps two components of adaptation might exist
12. Hair Cell Mechanotransduction
359
(Wu et al., 1999). More light was shed onto the discrepancy when methods of hair bundle stimulation were compared (Holt et al., 1997). It was shown that stimulus rise‐time altered the rates of adaptation (Wu et al., 1999; Vollrath and Eatock, 2003). When comparable stimuli were used comparable results between turtle and frog were obtained (Eatock, 2000; Vollrath and Eatock, 2003). And finally, evidence suggests that hair bundle mechanics from frog are comparable to those of turtle when the time frame of imaging and recording are comparable (Ricci et al., 2000; Cheung and Corey, 2006). Ultimately, a resolution to the discrepancy may be that multiple forms of adaptation exist and that each form can be found in both auditory and vestibular hair cells but that the apparent contribution of each depends strongly on experimental design (Wu et al., 1999; Eatock, 2000; Holt and Corey, 2000; Vollrath and Eatock, 2003). The conventional form of adaptation (described above) called slow or motor adaptation and a fast adaptation (or Ca2þ‐dependent channel closure) coexist; whether the underlying mechanisms are independent remains to be determined.
C. Fast Adaptation Fast adaptation was observed in frog as a minor component of the hair bundle mechanical response (Howard and Hudspeth, 1987) and later characterized more carefully as a ‘‘notch’’ (Benser et al., 1996). It has been modeled as a Ca2þ‐dependent closed channel state (Crawford et al., 1989, 1991; Choe et al., 1998; Wu et al., 1999). EVects of Ca2þ buVers suggest a site of Ca2þ binding very close to the channel and support the hypothesis that fast adaptation can be understood as a Ca2þ‐dependent feedback that serves to maintain Ca2þ at a constant level near to its binding site within the stereocilia (Ricci et al., 1998). There is a mechanical correlate to fast adaptation (Fig. 9); however, whether this change represents a change in compliance is unclear (Ricci et al., 2002). As the bundle has moved less after adaptation, a decrease in compliance could be predicted. However, if the compliance curve has shifted due to adaptation, without a compliance change, a similar bundle movement would be obtained (Fig. 9). To date, the mechanism underlying fast adaptation remains controversial. A Ca2þ‐ dependent relaxation of the gating spring has been suggested (Martin et al., 2003). Evidence suggests a more direct eVect of Ca2þ onto the channel (Cheung and Corey, 2006). However, a role for myosin 1C has also been reported (StauVer et al., 2005), implicating a role for rocking of the myosin head group without unbinding from the actin and supported by the reported properties for myosin 1C (Batters et al., 2004a). It may be diYcult to delineate between direct mechanistic eVects or indirect eVects by altering a protein in series with the channel. That is, can fast and slow adaptation be
360
Ricci and Kachar A
B
30 pA 5 ms
C
50 nm
FIGURE 9 Hair bundle movements mimic MET currents. Recordings with a flexible fiber allow hair bundle mechanical responses to be imaged with a photodiode motion detector (Crawford and Fettiplace, 1985). (A) shows the stimulus, (B) the MET current response to the stimulus, and (C) the response of the hair bundle to the force stimulus. The large black arrows indicate fast adaptation responses while the three gray arrows point to the slow adaptation response. A mechanical correlate to both adaptation responses are observed in both the current and the hair bundle response.
independently modulated without indirectly altering each other? An additional question remains regarding myosin 1C as the mechanism underlying fast adaptation. The kinetics of fast adaptation vary tonotopically; however, slow adaptation kinetics have yet to be shown to have any frequency‐related variations, thus if one molecule is responsible for both processes it would seem that an important point is missing. Further investigations are needed to clarify this mechanism.
D. Functional Role of Adaptation What are the functional consequences of these complex adaptation mechanisms? Adaptation in any form prevents saturation, extending the dynamic range of the sensory cell. Here adaptation can reset the operating
12. Hair Cell Mechanotransduction
361
range such that the MET dynamic range is extended severalfold (Eatock et al., 1987; Crawford et al., 1989; Fig. 10A). Adaptation also serves to maintain hair bundle sensitivity at its optimal and most linear point, that is the resting hair bundle position is kept at the steepest portion of the activation curve. Adaptation provides a mechanical filter to incoming sound (Ricci and Fettiplace, 1997). Variations in both activation and adaptation rates create a mechanical bandpass filter (Ricci and Fettiplace, 1997; Ricci et al., 2005; Fig. 10B). The time course of adaptation varies by orders of magnitude across species and end organs (Ricci and Fettiplace, 1997; Kennedy et al., 2003, 2005). It is unclear at this point what mechanisms underlie the tonotopic variation in adaptation rate. DiVerences in channel properties, numbers, and also hair bundle mechanics may contribute. Adaptation has also been posited to be part of a mechanical amplification process (Jaramillo et al., 1993; Markin and Hudspeth, 1995b; Hudspeth, 1997; Choe et al., 1998; Jaramillo and Wiesenfeld, 1998; Hudspeth et al., 2000; Indresano et al., 2003). The mechanism for amplification is thought to involve the gating spring compliance, the adaptation motors, and possibly fast adaptation (Martin et al., 2000; Chan and Hudspeth, 2005; Le GoV et al., 2005). Cooperative interactions between MET channels have also been implicated as a mechanism for amplification (Iwasa and Ehrenstein, 2002). As adaptation sets the resting open probability of the MET channel, it also plays an important role in setting the hair cell resting potential (Farris et al., 2006; Fig. 10C). These multiple important roles for adaptation warrant a better understanding of the underlying mechanisms.
XII. THE DYNAMIC HAIR BUNDLE A theme throughout this chapter has been attempting to delineate properties associated with the sensory hair bundle from those associated with the MET channel. Initial investigations have treated the hair bundle as an invariant structure in the transduction process; however, growing evidence suggests the hair bundle is very dynamic. A simple example is considering the number of myosin isoforms found in the hair bundle and cell. Myosin XVa is located near the tops of the stereocilia, forming a cap‐like structure (Rzadzinska et al., 2004). This myosin is critical for proper development of the hair bundle (Liang et al., 1999; Anderson et al., 2000; Liburd et al., 2001; Rzadzinska et al., 2004). Myosin VIIa is also localized along the length of the stereocilia (Hasson et al., 1995; Rzadzinska et al., 2004). Defects in myosin VIIa are associated with Usher’s syndrome (el‐Amraoui et al., 1996; Mburu et al., 1997; Todi et al., 2005) and are typically associated with hair bundle defects (Rhodes et al., 2004). Mice lacking myosin VIIa have MET activation curves
362
Ricci and Kachar
A
100 nm 0.8
Current (nA)
0.0
Current (nA)
−0.2 −0.4 −0.6
0.6 0.4 0.2
−0.8 0.0
−1.0 0
B
10
20 30 Time (ms)
FIp
1.0
40
50
−100
0 100 200 300 Displacement (nm)
Fhp
Gain
0.8 0.6 0.4
Fres
0.2 10
C
100 Frequency (Hz)
1000
100 pA
10
100 Frequency (Hz)
1000
100 pA
MET Po = 0.06
MET Po = 0.24 5 mV 10 mV
FIGURE 10 Adaptation serves multiple functional roles. (A) demonstrates a double pulse protocol used to first characterize adaptation. By generating activation curves about the hair bundles resting position and comparing it to the activation curve generated from a displaced position, the ability of adaptation to extend the dynamic range of the hair cell response is observed. (B) illustrates that the combination of activation kinetics, which generate a low pass filter (Flp) and adaptation kinetics, which generate a high pass filter (Fhp), together produces a bandpass filter with a center frequency similar to the electrical resonant frequency of the hair cells (adapted from Ricci et al., 2005). (C) demonstrates that changing the resting open
12. Hair Cell Mechanotransduction
363
that are shifted to the right so that there is no MET current on at rest which prevents aminoglycoside accumulation (Richardson et al., 1997, 1999; Kros et al., 2002). These findings demonstrate the complexity of the hair bundle– MET channel interaction in that the myosin VIIa, which is unlikely to be directly associated with the channel, has profound eVects on channel function. It is possible that myosin VIIa is involved in establishing the resting tension of the hair bundle; loss of this tension reduces the coupling between hair bundle deflection and force sensed by the MET channel. Mutations of myosin VI also lead to sensorineural hearing loss (Ahmed et al., 2002, 2003; Mohiddin et al., 2004). The subcellular localization of myosin VI is not well established but also thought to be in the stereocilia (Rzadzinska et al., 2004). Myosin VI is known to regulate endocytosis and may play a role in apical endocytosis in hair cells (Swiatecka‐Urban et al., 2004). Apical endocytosis appears to be involved in the turnover and renewal of stereocilia membrane components (Kachar et al., 1997; Grati et al., 2006). Not only are the elements present in the hair bundle for it to play a dynamic role in signal transduction, but the stereocilia and the hair bundle structure also appear to be constantly remodeling (Lin et al., 2005). Length regulation and turnover of the streocilia actin core follows an actin treadmill mechanism (Fig. 11) that involves a variety of molecules, including some myosins and espins (Rzadzinska et al., 2004, 2005). The actin treadmilling appears to work from a top‐down mechanism, with actin polymerization occurring near the top of the stereocilia (Rzadzinska et al., 2004). Onto this continuous turnover is placed the machinery of mechanotransduction (Fig. 11); separating the components of these diVerent processes is an important remaining task. Determining how hair bundle shape is driven or modulated by activity of the MET channel is a question for the future. Given that Ca2þ entry into the stereocilia is largely through MET channels and that many of the structural proteins involved in stereocilia turnover and maintenance are Ca2þ‐dependent, it seems likely that an interaction between these components will be identified. The number of identified proteins required for hair bundle development and function is rapidly expanding. Proteins like harmonin (Verpy et al., 2000; Siemens et al., 2002), whirlin (Mburu et al., 2003; Belyantseva et al., 2005), espins (Zheng et al., 2000; Li et al., 2004; Sekerkova et al., 2004, 2006; Rzadzinska et al., 2005), cadherins (Di Palma et al., 2001; Siemens et al., 2002, 2004; Sollner et al., 2004; Michel et al., 2005), and fimbrin (Tilney et al., 1989; Zine et al., 1995) have
probability of the MET channel by exposing the hair bundle to diVerent external Ca2þ concentrations alters the resting potential of the hair cell and may modulate the frequency selectivity of the filter in this way.
364
Ricci and Kachar
Polymerization and cross-linking of actin Myosin 15a Myosin 1c, 7
Treadmilling
Actin depolymerization
FIGURE 11 Superimposing hair bundle dynamic recycling with MET proteins illustrates the complexities of the system. Schematic representation of stereociliary pair illustrates the dynamics of actin turnover. Included are the various myosins located at specific sites along the stereocilia, the tip‐links, side‐links, and putative location for MET channel. Inset shows myosin 15 immunolabeling (green), actin (red, rhodamine phalloidin). Scale bar is 0.5 mm.
12. Hair Cell Mechanotransduction
365
important roles as scaVolding and cross‐linking elements, but their function in terms of hair cell mechanotransduction remain to be elucidated. To this point, three mechanical bundle movements have been identified and related to MET currents. Gating spring compliance, thought to be associated with the opening and closing of the MET channel, slow motor adaptation, thought to be driven by myosins climbing and slipping along the actin cytoskeleton, and fast adaptation, where the underlying mechanism is less clear but may be directly associated with the channel or with myosins. Several additional hair bundle movements exist that have yet to be explored in terms of function or mechanism. One of these movements which has been called a ‘‘flick’’ (though it remains throughout the duration of a stimulus) is a voltage‐dependent, Ca2þ‐ independent movement that does not require current through the MET channel but does require an intact hair bundle (Ricci et al., 2000; Cheung and Corey, 2006). A second movement, termed a ‘‘sag’’ is often seen with long depolarizations and is a return (negative movement) to the hair bundle’s resting position or even negative to that position during a constant depolarization. This movement has a very slow time course, yet the movement can be large (Ricci et al., 2002). How these additional movements factor into our understanding of hair bundle dynamics remain to be elucidated.
XIII. SUMMARY AND FUTURE DIRECTIONS Over the past 25 years, a great deal of information has been collected regarding the MET process in hair cells. The hair bundle structure and the component proteins that contribute to this structure are rapidly being elucidated. Exploring the functional role of these new components in the transduction process has already revealed previously unrecognized complexities. Long‐standing hypotheses regarding mechanisms of activation and adaptation are being both supported and challenged. New technologies are allowing more detailed experimentation at the physiological, molecular, and protein levels. Identification of all the players will greatly aid in deciphering the mechanisms of mechanotransduction. Physiological measurements at the single molecule or at least single stereocilia level are needed to distinguish between existing models of mechanotransduction. Through all these new developments, the gating spring theory at its simplest is still capable of explaining much existing data. Care, however, must be taken when applying molecular mechanisms to this generalized gating hypothesis. Important questions remain as to how hair bundle dynamics are influenced by the MET process and which proteins are critical for transduction and which are critical for hair bundle maintenance and turnover.
366
Ricci and Kachar
Acknowledgments This work was supported by RO1 DC03896 to AJR and from NIDCD‐IRP from NIH to BK. AJR is grateful to those at NIDCD for providing a safe haven when laboratories in New Orleans were unavailable. Part of this work was accomplished while aYliated with LSU Neuroscience Center in New Orleans.
References Adams, D. J., Dwyer, T. M., and Hille, B. (1980). The permeability of endplate channels to monovalent and divalent metal cations. J. Gen. Physiol. 75, 493–510. Ahmed, Z. M., Smith, T. N., Riazuddin, S., Makishima, T., Ghosh, M., Bokhari, S., Menon, P. S., Deshmukh, D., GriYth, A. J., Friedman, T. B., and Wilcox, E. R. (2002). Nonsyndromic recessive deafness DFNB18 and Usher syndrome type IC are allelic mutations of USHIC. Hum. Genet. 110, 527–531. Ahmed, Z. M., Morell, R. J., Riazuddin, S., Gropman, A., Shaukat, S., Ahmad, M. M., Mohiddin, S. A., Fananapazir, L., Caruso, R. C., Husnain, T., Khan, S. N., GriYth, A. J., et al. (2003). Mutations of MYO6 are associated with recessive deafness, DFNB37. Am. J. Hum. Genet. 72, 1315–1322. Anderson, D. W., Probst, F. J., Belyantseva, I. A., Fridell, R. A., Beyer, L., Martin, D. M., Wu, D., Kachar, B., Friedman, T. B., Raphael, Y., and Camper, S. A. (2000). The motor and tail regions of myosin XV are critical for normal structure and function of auditory and vestibular hair cells [In Process Citation]. Hum. Mol. Genet. 9, 1729–1738. Assad, J. A., and Corey, D. P. (1992). An active motor model for adaptation by vertebrate hair cells. J. Neurosci. 12, 3291–3309. Assad, J. A., Hacohen, N., and Corey, D. P. (1989). Voltage dependence of adaptation and active bundle movement in bullfrog saccular hair cells. Proc. Natl. Acad. Sci. USA 86, 2918–2922. Assad, J. A., Shepherd, G. M., and Corey, D. P. (1991). Tip‐link integrity and mechanical transduction in vertebrate hair cells. Neuron 7, 985–994. Bashtanov, M. E., Goodyear, R. J., Richardson, G. P., and Russell, I. J. (2004). The mechanical properties of chick (Gallus domesticus) sensory hair bundles: Relative contributions of structures sensitive to calcium chelation and subtilisin treatment. J. Physiol. 559, 287–299. Batters, C., Arthur, C. P., Lin, A., Porter, J., Geeves, M. A., Milligan, R. A., Molloy, J. E., and Coluccio, L. M. (2004a). Myo1c is designed for the adaptation response in the inner ear. EMBO J. 23, 1433–1440. Batters, C., Wallace, M. I., Coluccio, L. M., and Molloy, J. E. (2004b). A model of stereocilia adaptation based on single molecule mechanical studies of myosin I. Philos. Trans. R. Soc. Lond. B Biol. Sci. 359, 1895–1905. Belyantseva, I. A., Boger, E. T., Naz, S., Frolenkov, G. I., Sellers, J. R., Ahmed, Z. M., GriYth, A. J., and Friedman, T. B. (2005). Myosin‐XVa is required for tip localization of whirlin and diVerential elongation of hair‐cell stereocilia. Nat. Cell Biol. 7, 148–156. Benser, M. E., Issa, N. P., and Hudspeth, A. J. (1993). Hair‐bundle stiVness dominates the elastic reactance to otolithic‐membrane shear. Hear. Res. 68, 243–252. Benser, M. E., Marquis, R. E., and Hudspeth, A. J. (1996). Rapid, active hair bundle movements in hair cells from the bullfrog’s sacculus. J. Neurosci. 16, 5629–5643. Bialek, W. (1987). Physical limits to sensation and perception. Annu. Rev. Biophys. Biophys. Chem. 16, 455–478. Chan, D. K., and Hudspeth, A. J. (2005). Ca2þ current‐driven nonlinear amplification by the mammalian cochlea in vitro. Nat. Neurosci. 8, 149–155.
12. Hair Cell Mechanotransduction
367
Cheung, E. L., and Corey, D. P. (2006). Ca2þ changes the force sensitivity of the hair‐cell transduction channel. Biophys. J. 90, 124–139. Choe, Y., Magnasco, M. O., and Hudspeth, A. J. (1998). A model for amplification of hair‐ bundle motion by cyclical binding of Ca2þ to mechanoelectrical‐transduction channels. Proc. Natl. Acad. Sci. USA 95, 15321–15326. Corey, D. (2003). Sensory transduction in the ear. J. Cell Sci. 116, 1–3. Corey, D. P., and Hudspeth, A. J. (1979a). Ionic basis of the receptor potential in a vertebrate hair cell. Nature 281, 675–677. Corey, D. P., and Hudspeth, A. J. (1979b). Response latency of vertebrate hair cells. Biophys. J. 26, 499–506. Corey, D. P., and Hudspeth, A. J. (1980). Mechanical stimulation and micromanipulation with piezoelectric bimorph elements. J. Neurosci. Methods 3, 183–202. Corey, D. P., and Hudspeth, A. J. (1983). Kinetics of the receptor current in bullfrog saccular hair cells. J. Neurosci. 3, 962–976. Crawford, A. C., and Fettiplace, R. (1985). The mechanical properties of ciliary bundles of turtle cochlear hair cells. J. Physiol. (Lond.) 364, 359–379. Crawford, A. C., Evans, M. G., and Fettiplace, R. (1989). Activation and adaptation of transducer currents in turtle hair cells. J. Physiol. (Lond.) 419, 405–434. Crawford, A. C., Evans, M. G., and Fettiplace, R. (1991). The actions of calcium on the mechano‐electrical transducer current of turtle hair cells. J. Physiol. (Lond.) 434, 369–398. Cruickshank, C. C., Minchin, R. F., Le Dain, A. C., and Martinac, B. (1997). Estimation of the pore size of the large‐conductance mechanosensitive ion channel of Escherichia coli. Biophys. J. 73, 1925–1931. Cyr, J. L., Dumont, R. A., and Gillespie, P. G. (2002). Myosin‐1c interacts with hair‐cell receptors through its calmodulin‐binding IQ domains. J. Neurosci. 22, 2487–2495. Denk, W., and Webb, W. W. (1992). Forward and reverse transduction at the limit of sensitivity studied by correlating electrical and mechanical fluctuations in frog saccular hair cells. Hear. Res. 60, 89–102. Denk, W., Keolian, R. M., and Webb, W. W. (1992). Mechanical response of frog saccular hair bundles to the aminoglycoside block of mechanoelectrical transduction. J. Neurophysiol. 68, 927–932. Denk, W., Holt, J. R., Shepherd, G. M., and Corey, D. P. (1995). Calcium imaging of single stereocilia in hair cells: Localization of transduction channels at both ends of tip links. Neuron 15, 1311–1321. Di Palma, F., Holme, R. H., Bryda, E. C., Belyantseva, I. A., Pellegrino, R., Kachar, B., Steel, K. P., and Noben‐Trauth, K. (2001). Mutations in Cdh23, encoding a new type of cadherin, cause stereocilia disorganization in waltzer, the mouse model for Usher syndrome type 1D. Nat. Genet. 27, 103–107. Du, H., Gu, G., William, C. M., and Chalfie, M. (1996). Extracellular proteins needed for C. elegans mechanosensation. Neuron 16, 183–194. Eatock, R. A. (2000). Adaptation in hair cells. Annu. Rev. Neurosci. 23, 285–314. Eatock, R. A., Corey, D. P., and Hudspeth, A. J. (1987). Adaptation of mechanoelectrical transduction in hair cells of the bullfrog’s sacculus. J. Neurosci. 7, 2821–2836. el‐Amraoui, A., Sahly, I., Picaud, S., Sahel, J., Abitbol, M., and Petit, C. (1996). Human Usher 1B/mouse shaker‐1: The retinal phenotype discrepancy explained by the presence/absence of myosin VIIA in the photoreceptor cells. Hum. Mol. Genet. 5, 1171–1178. Farris, H. E., LeBlanc, C. L., Goswami, J., and Ricci, A. J. (2004). Probing the pore of the auditory hair cell mechanotransducer channel in turtle. J. Physiol. 558, 769–792.
368
Ricci and Kachar
Farris, H. E., Wells, G. B., and Ricci, A. J. (2006). Steady‐state adaptation of mechanotransduction modulates the resting potential of auditory hair cells, providing an assay for endolymph [Ca2þ]. J. Neurosci. 26, 12526–12536. Fettiplace, R., and Hackney, C. M. (2006). The sensory and motor roles of auditory hair cells. Nat. Rev. Neurosci. 7, 19–29. Friedman, T. B., Sellers, J. R., and Avraham, K. B. (1999). Unconventional myosins and the genetics of hearing loss. Am. J. Med. Genet. 89, 147–157. Furness, D. N., Hackney, C. M., and Benos, D. J. (1996). The binding site on cochlear stereocilia for antisera raised against renal Naþ channels is blocked by amiloride and dihydrostreptomycin. Hear. Res. 93, 136–146. Furness, D. N., Zetes, D. E., Hackney, C. M., and Steele, C. R. (1997). Kinematic analysis of shear displacement as a means for operating mechanotransduction channels in the contact region between adjacent stereocilia of mammalian cochlear hair cells. Proc. Biol. Sci. 264, 45–51. Furness, D. N., Karkanevatos, A., West, B., and Hackney, C. M. (2002). An immunogold investigation of the distribution of calmodulin in the apex of cochlear hair cells. Hear. Res. 173, 10–20. Gale, J. E., Marcotti, W., Kennedy, H. J., Kros, C. J., and Richardson, G. P. (2001). FM1–43 dye behaves as a permeant blocker of the hair‐cell mechanotransducer channel. J. Neurosci. 21, 7013–7025. Garcia, J. A., Yee, A. G., Gillespie, P. G., and Corey, D. P. (1998). Localization of myosin‐ Ibeta near both ends of tip links in frog saccular hair cells. J. Neurosci. 18, 8637–8647. Geisler, C. D. (1993). A model of stereociliary tip‐link stretches. Hear. Res. 65, 79–82. Gillespie, P. G., and Hudspeth, A. J. (1993). Adenine nucleoside diphosphates block adaptation of mechanoelectrical transduction in hair cells. Proc. Natl. Acad. Sci. USA 90, 2710–2714. Gillespie, P. G., Wagner, M. C., and Hudspeth, A. J. (1993). Identification of a 120 kd hair‐ bundle myosin located near stereociliary tips. Neuron 11, 581–594. Gillespie, P. G., Dumont, R. A., and Kachar, B. (2005). Have we found the tip link, transduction channel, and gating spring of the hair cell? Curr. Opin. Neurobiol. 15, 389–396. Glowatzki, E., Ruppersberg, J. P., Zenner, H. P., and Rusch, A. (1997). Mechanically and ATP‐induced currents of mouse outer hair cells are independent and diVerentially blocked by d‐tubocurarine. Neuropharmacology 36, 1269–1275. Grati, M., Schneider, M. E., Lipkow, K., Strehler, E. E., Wenthold, R. J., and Kachar, B. (2006). Rapid turnover of stereocilia membrane proteins: Evidence from the traYcking and mobility of plasma membrane Ca(2þ)‐ATPase 2. J. Neurosci. 26, 6386–6395. Hasson, T., and Mooseker, M. S. (1997). The growing family of myosin motors and their role in neurons and sensory cells. Curr. Opin. Neurobiol. 7, 615–623. Hasson, T., Heintzelman, M. B., Santos‐Sacchi, J., Corey, D. P., and Mooseker, M. S. (1995). Expression in cochlea and retina of myosin VIIa, the gene product defective in Usher syndrome type 1B. Proc. Natl. Acad. Sci. USA 92, 9815–9819. Hille, B. (1971). The permeability of the sodium channel to organic cations in myelinated nerve. J. Gen. Physiol. 58, 599–619. Holt, J. R., and Corey, D. P. (2000). Two mechanisms for transducer adaptation in vertebrate hair cells. Proc. Natl. Acad. Sci. USA 97, 11730–11735. Holt, J. R., Corey, D. P., and Eatock, R. A. (1997). Mechanoelectrical transduction and adaptation in hair cells of the mouse utricle, a low‐frequency vestibular organ. J. Neurosci. 17, 8739–8748. Holt, J. R., Gillespie, S. K., Provance, D. W., Shah, K., Shokat, K. M., Corey, D. P., Mercer, J. A., and Gillespie, P. G. (2002). A chemical‐genetic strategy implicates myosin‐1c in adaptation by hair cells. Cell 108, 371–381.
12. Hair Cell Mechanotransduction
369
Holton, T., and Hudspeth, A. J. (1986). The transduction channel of hair cells from the bull‐frog characterized by noise analysis. J. Physiol. (Lond.) 375, 195–227. Howard, J., and Bechstedt, S. (2004). Hypothesis: A helix of ankyrin repeats of the NOMPC‐ TRP ion channel is the gating spring of mechanoreceptors. Curr. Biol. 14, R224–R226. Howard, J., and Hudspeth, A. J. (1987). Mechanical relaxation of the hair bundle mediates adaptation in mechanoelectrical transduction by the bullfrog’s saccular hair cell. Proc. Natl. Acad. Sci. USA 84, 3064–3068. Howard, J., and Hudspeth, A. J. (1988). Compliance of the hair bundle associated with gating of mechanoelectrical transduction channels in the bullfrog’s saccular hair cell. Neuron 1, 189–199. Huang, M., Gu, G., Ferguson, E. L., and Chalfie, M. (1995). A stomatin‐like protein necessary for mechanosensation in C. elegans. Nature 378, 292–295. Hudspeth, A. J. (1982). Extracellular current flow and the site of transduction by vertebrate hair cells. J. Neurosci. 2, 1–10. Hudspeth, A. J. (1989). Mechanoelectrical transduction by hair cells of the bullfrog’s sacculus. Prog. Brain Res. 80, 129–135. Hudspeth, A. J. (1997). Mechanical amplification of stimuli by hair cells. Curr. Opin. Neurobiol. 7, 480–486. Hudspeth, A. J. (2005). How the ear’s works work: Mechanoelectrical transduction and amplification by hair cells. C. R. Biol. 328, 155–162. Hudspeth, A. J., and Corey, D. P. (1977). Sensitivity, polarity, and conductance change in the response of vertebrate hair cells to controlled mechanical stimuli. Proc. Natl. Acad. Sci. USA 74, 2407–2411. Hudspeth, A. J., and Gillespie, P. G. (1994). Pulling springs to tune transduction: Adaptation by hair cells. Neuron 12, 1–9. Hudspeth, A. J., Choe, Y., Mehta, A. D., and Martin, P. (2000). Putting ion channels to work: Mechanoelectrical transduction, adaptation, and amplification by hair cells. Proc. Natl. Acad. Sci. USA 97, 11765–11772. Indresano, A. A., Frank, J. E., Middleton, P., and Jaramillo, F. (2003). Mechanical noise enhances signal transmission in the bullfrog sacculus. J. Assoc. Res. Otolaryngol. 4, 363–370. Iwasa, K. H., and Ehrenstein, G. (2002). Cooperative interaction as the physical basis of the negative stiVness in hair cell stereocilia. J. Acoust. Soc. Am. 111, 2208–2212. Jaramillo, F., and Hudspeth, A. J. (1991). Localization of the hair cell’s transduction channels at the hair bundle’s top by iontophoretic application of a channel blocker. Neuron 7, 409–420. Jaramillo, F., and Hudspeth, A. J. (1993). Displacement‐clamp measurement of the forces exerted by gating springs in the hair bundle. Proc. Natl. Acad. Sci. USA 90, 1330–1334. Jaramillo, F., and Wiesenfeld, K. (1998). Mechanoelectrical transduction assisted by Brownian motion: A role for noise in the auditory system. Nat. Neurosci. 1, 384–388. Jaramillo, F., Markin, V. S., and Hudspeth, A. J. (1993). Auditory illusions and the single hair cell. Nature 364, 527–529. Jorgensen, F., and Ohmori, H. (1988). Amiloride blocks the mechano‐electrical transduction channel of hair cells of the chick. J. Physiol. (Lond.) 403, 577–588. Kachar, B., Battaglia, A., and Fex, J. (1997). Compartmentalized vesicular traYc around the hair cell cuticular plate. Hear. Res. 107, 102–112. Kachar, B., Parakkal, M., Kurc, M., Zhao, Y., and Gillespie, P. G. (2000). High‐resolution structure of hair‐cell tip links. Proc. Natl. Acad. Sci. USA 97, 13336–13341. Kennedy, H. J., Evans, M. G., Crawford, A. C., and Fettiplace, R. (2003). Fast adaptation of mechanoelectrical transducer channels in mammalian cochlear hair cells. Nat. Neurosci. 6, 832–836.
370
Ricci and Kachar
Kennedy, H. J., Crawford, A. C., and Fettiplace, R. (2005). Force generation by mammalian hair bundles supports a role in cochlear amplification. Nature 433, 880–883. Kimitsuki, T., and Ohmori, H. (1993). Dihydrostreptomycin modifies adaptation and blocks the mechano‐electric transducer in chick cochlear hair cells. Brain Res. 624, 143–150. Krendel, M., and Mooseker, M. S. (2005). Myosins: Tails (and heads) of functional diversity. Physiology (Bethesda) 20, 239–251. Kroese, A. B., Das, A., and Hudspeth, A. J. (1989). Blockage of the transduction channels of hair cells in the bullfrog’s sacculus by aminoglycoside antibiotics. Hear. Res. 37, 203–217. Kros, C. J., Rusch, A., and Richardson, G. P. (1992). Mechano‐electrical transducer currents in hair cells of the cultured neonatal mouse cochlea. Proc. R. Soc. Lond. B Biol. Sci. 249, 185–1993. Kros, C. J., Marcotti, W., van Netten, S. M., Self, T. J., Libby, R. T., Brown, S. D., Richardson, G. P., and Steel, K. P. (2002). Reduced climbing and increased slipping adaptation in cochlear hair cells of mice with Myo7a mutations. Nat. Neurosci. 5, 41–47. Kung, C. (2005). A possible unifying principle for mechanosensation. Nature 436, 647–654. Lane, J. W., McBride, D. W., Jr., and Hamill, O. P. (1993). Ionic eVects on amiloride block of the mechanosensitive channel in Xenopus oocytes. Br. J. Pharmacol. 108, 116–119. Le GoV, L., Bozovic, D., and Hudspeth, A. J. (2005). Adaptive shift in the domain of negative stiVness during spontaneous oscillation by hair bundles from the internal ear. Proc. Natl. Acad. Sci. USA 102, 16996–17001. Lemasurier, M., and Gillespie, P. G. (2005). Hair‐cell mechanotransduction and cochlear amplification. Neuron 48, 403–415. Li, H., Liu, H., Balt, S., Mann, S., Corrales, C. E., and Heller, S. (2004). Correlation of expression of the actin filament‐bundling protein espin with stereociliary bundle formation in the developing inner ear. J. Comp. Neurol. 468, 125–134. Liang, Y., Wang, A., Belyantseva, I. A., Anderson, D. W., Probst, F. J., Barber, T. D., Miller, W., Touchman, J. W., Jin, L., Sullivan, S. L., Sellers, J. R., Camper, S. A., et al. (1999). Characterization of the human and mouse unconventional myosin XV genes responsible for hereditary deafness DFNB3 and shaker 2. Genomics 61, 243–258. Liburd, N., Ghosh, M., Riazuddin, S., Naz, S., Khan, S., Ahmed, Z., Liang, Y., Menon, P. S., Smith, T., Smith, A. C., Chen, K. S., Lupski, J. R., et al. (2001). Novel mutations of MYO15A associated with profound deafness in consanguineous families and moderately severe hearing loss in a patient with Smith‐Magenis syndrome. Hum. Genet. 109, 535–541. Lin, H. W., Schneider, M. E., and Kachar, B. (2005). When size matters: The dynamic regulation of stereocilia length. Curr. Opin. Neurobiol. 17, 55–61. Lumpkin, E. A., and Hudspeth, A. J. (1995). Detection of Ca2þ entry through mechanosensitive channels localizes the site of mechanoelectrical transduction in hair cells. Proc. Natl. Acad. Sci. USA 92, 10297–10301. Lumpkin, E. A., and Hudspeth, A. J. (1998). Regulation of free Ca2þ concentration in hair‐cell stereocilia. J. Neurosci. 18, 6300–6318. Lumpkin, E. A., Marquis, R. E., and Hudspeth, A. J. (1997). The selectivity of the hair cell’s mechanoelectrical‐transduction channel promotes Ca2þ flux at low Ca2þ concentrations. Proc. Natl. Acad. Sci. USA 94, 10997–11002. Maingret, F., Fosset, M., Lesage, F., Lazdunski, M., and Honore, E. (1999). TRAAK is a mammalian neuronal mechano‐gated Kþ channel. J. Biol. Chem. 274, 1381–1387. Marcotti, W., van Netten, S. M., and Kros, C. J. (2005). The aminoglycoside antibiotic dihydrostreptomycin rapidly enters mouse outer hair cells through the mechano‐electrical transducer channels. J. Physiol. 567, 505–521.
12. Hair Cell Mechanotransduction
371
Markin, V. S., and Hudspeth, A. J. (1995a). Gating‐spring models of mechanoelectrical transduction by hair cells of the internal ear. Annu. Rev. Biophys. Biomol. Struct. 24, 59–83. Markin, V. S., and Hudspeth, A. J. (1995b). Modeling the active process of the cochlea: Phase relations, amplification, and spontaneous oscillation. Biophys. J. 69, 138–147. Marquis, R. E., and Hudspeth, A. J. (1997). EVects of extracellular Ca2þ concentration on hair‐bundle stiVness and gating‐spring integrity in hair cells. Proc. Natl. Acad. Sci. USA 94, 11923–11928. Martin, P., Mehta, A. D., and Hudspeth, A. J. (2000). Negative hair‐bundle stiVness betrays a mechanism for mechanical amplification by the hair cell [In Process Citation]. Proc. Natl. Acad. Sci. USA 97, 12026–12031. Martin, P., Bozovic, D., Choe, Y., and Hudspeth, A. J. (2003). Spontaneous oscillation by hair bundles of the bullfrog’s sacculus. J. Neurosci. 23, 4533–4548. Mburu, P., Liu, X. Z., Walsh, J., Saw, D., Jr., Cope, M. J., Gibson, F., Kendrick‐Jones, J., Steel, K. P., and Brown, S. D. (1997). Mutation analysis of the mouse myosin VIIA deafness gene. Genes Funct. 1, 191–203. Mburu, P., Mustapha, M., Varela, A., Weil, D., El‐Amraoui, A., Holme, R. H., Rump, A., Hardisty, R. E., Blanchard, S., Coimbra, R. S., Perfettini, I., Parkinson, N., et al. (2003). Defects in whirlin, a PDZ domain molecule involved in stereocilia elongation, cause deafness in the whirler mouse and families with DFNB31. Nat. Genet. 34, 421–428. Metcalf, A. B., Chelliah, Y., and Hudspeth, A. J. (1994). Molecular cloning of a myosin I beta isozyme that may mediate adaptation by hair cells of the bullfrog’s internal ear. Proc. Natl. Acad. Sci. USA 91, 11821–11825. Michel, V., Goodyear, R. J., Weil, D., Marcotti, W., Perfettini, I., Wolfrum, U., Kros, C. J., Richardson, G. P., and Petit, C. (2005). Cadherin 23 is a component of the transient lateral links in the developing hair bundles of cochlear sensory cells. Dev. Biol. 280, 281–294. Mohiddin, S. A., Ahmed, Z. M., GriYth, A. J., Tripodi, D., Friedman, T. B., Fananapazir, L., and Morell, R. J. (2004). Novel association of hypertrophic cardiomyopathy, sensorineural deafness, and a mutation in unconventional myosin VI (MYO6). J. Med. Genet. 41, 309–314. O’Hagan, R., and Chalfie, M. (2006). Mechanosensation in Caenorhabditis elegans. Int. Rev. Neurobiol. 69, 169–203. Ohmori, H. (1984). Mechanoelectrical transducer has discrete conductances in the chick vestibular hair cell. Proc. Natl. Acad. Sci. USA 81, 1888–1891. Ohmori, H. (1985). Mechano‐electrical transduction currents in isolated vestibular hair cells of the chick. J. Physiol. (Lond.) 359, 189–217. Ohmori, H. (1987). Gating properties of the mechano‐electrical transducer channel in the dissociated vestibular hair cell of the chick. J. Physiol. (Lond.) 387, 589–609. Ohmori, H. (1988). Mechanical stimulation and Fura‐2 fluorescence in the hair bundle of dissociated hair cells of the chick. J. Physiol. (Lond.) 399, 115–137. Ohmori, H. (1989). Mechano‐electrical transduction of the hair cell. Jpn. J. Physiol. 39, 643–657. Patel, A. J., Lazdunski, M., and Honore, E. (2001). Lipid and mechano‐gated 2P domain K(þ) channels. Curr. Opin. Cell Biol. 13, 422–428. Pickles, J. O. (1993). A model for the mechanics of the stereociliar bundle on acousticolateral hair cells. Hear. Res. 68, 159–172. Pickles, J. O., Comis, S. D., and Osborne, M. P. (1984). Cross‐links between stereocilia in the guinea pig organ of Corti, and their possible relation to sensory transduction. Hear. Res. 15, 103–112.
372
Ricci and Kachar
Rhodes, C. R., Hertzano, R., Fuchs, H., Bell, R. E., de Angelis, M. H., Steel, K. P., and Avraham, K. B. (2004). A Myo7a mutation cosegregates with stereocilia defects and low‐frequency hearing impairment. Mamm. Genome 15, 686–697. Ricci, A. J. (2002). DiVerences in mechano‐transducer channel kinetics underlie tonotopic distribution of fast adaptation in auditory hair cells. J. Neurophysiol. 87, 1738–1748. Ricci, A. J., and Fettiplace, R. (1997). The eVects of calcium buVering and cyclic AMP on mechano‐electrical transduction in turtle auditory hair cells. J. Physiol. (Lond.) 501, 111–124. Ricci, A. J., and Fettiplace, R. (1998). Calcium permeation of the turtle hair cell mechanotransducer channel and its relation to the composition of endolymph. J. Physiol. (Lond.) 506, 159–173. Ricci, A. J., Wu, Y. C., and Fettiplace, R. (1998). The endogenous calcium buVer and the time course of transducer adaptation in auditory hair cells. J. Neurosci. 18, 8261–8277. Ricci, A. J., Crawford, A. C., and Fettiplace, R. (2000). Active hair bundle motion linked to fast transducer adaptation in auditory hair cells. J. Neurosci. 20, 7131–7142. Ricci, A. J., Crawford, A. C., and Fettiplace, R. (2002). Mechanisms of active hair bundle motion in auditory hair cells. J. Neurosci. 22, 44–52. Ricci, A. J., Crawford, A. C., and Fettiplace, R. (2003). Tonotopic variation in the conductance of the hair cell mechanotransducer channel. Neuron 40, 983–990. Ricci, A. J., Kennedy, H. J., Crawford, A. C., and Fettiplace, R. (2005). The transduction channel filter in auditory hair cells. J. Neurosci. 25, 7831–7839. Richardson, G. P., Forge, A., Kros, C. J., Fleming, J., Brown, S. D., and Steel, K. P. (1997). Myosin VIIA is required for aminoglycoside accumulation in cochlear hair cells. J. Neurosci. 17, 9506–9519. Richardson, G. P., Forge, A., Kros, C. J., Marcotti, W., Becker, D., Williams, D. S., Thorpe, J., Fleming, J., Brown, S. D., and Steel, K. P. (1999). A missense mutation in myosin VIIA prevents aminoglycoside accumulation in early postnatal cochlear hair cells. Ann. NY Acad. Sci. 884, 110–124. Rusch, A., Kros, C. J., and Richardson, G. P. (1994). Block by amiloride and its derivatives of mechano‐electrical transduction in outer hair cells of mouse cochlear cultures. J. Physiol. (Lond.) 474, 75–86. Russell, I. J., and Sellick, P. M. (1978). Intracellular studies of hair cells in the mammalian cochlea. J. Physiol. (Lond.) 284, 261–290. Russell, I. J., Richardson, G. P., and Cody, A. R. (1986). Mechanosensitivity of mammalian auditory hair cells in vitro. Nature 321, 517–519. Russell, I. J., Kossl, M., and Richardson, G. P. (1992). Nonlinear mechanical responses of mouse cochlear hair bundles. Proc. Biol. Sci. 250, 217–227. Rzadzinska, A. K., Schneider, M. E., Davies, C., Riordan, G. P., and Kachar, B. (2004). An actin molecular treadmill and myosins maintain stereocilia functional architecture and self‐ renewal. J. Cell Biol. 164, 887–897. Rzadzinska, A. K., Schneider, M. E., Noben‐Trauth, K., Bartles, J. R., and Kachar, B. (2005). Balanced levels of Espin are critical for stereociliary growth and length maintenance. Cell. Motil. Cytoskeleton 62, 157–165. Schneider, M. E., Dose, A. C., Salles, F. T., Chang, W., Erickson, F. L., Burnside, B., and Kachar, B. (2006). A new compartment at stereocilia tips defined by spatial and temporal patterns of myosin IIIa expression. J. Neurosci. 26, 10243–10252. Sekerkova, G., Zheng, L., Loomis, P. A., Changyaleket, B., Whitlon, D. S., Mugnaini, E., and Bartles, J. R. (2004). Espins are multifunctional actin cytoskeletal regulatory proteins in the microvilli of chemosensory and mechanosensory cells. J. Neurosci. 24, 5445–5456.
12. Hair Cell Mechanotransduction
373
Sekerkova, G., Zheng, L., Mugnaini, E., and Bartles, J. R. (2006). DiVerential expression of espin isoforms during epithelial morphogenesis, stereociliogenesis and postnatal maturation in the developing inner ear. Dev. Biol. 291, 83–95. Shotwell, S. L., Jacobs, R., and Hudspeth, A. J. (1981). Directional sensitivity of individual vertebrate hair cells to controlled deflection of their hair bundles. Ann. NY Acad. Sci. 374, 1–10. Siemens, J., Kazmierczak, P., Reynolds, A., Sticker, M., Littlewood‐Evans, A., and Muller, U. (2002). The Usher syndrome proteins cadherin 23 and harmonin form a complex by means of PDZ‐domain interactions. Proc. Natl. Acad. Sci. USA 99, 14946–14951. Siemens, J., Lillo, C., Dumont, R. A., Reynolds, A., Williams, D. S., Gillespie, P. G., and Muller, U. (2004). Cadherin 23 is a component of the tip link in hair‐cell stereocilia. Nature 428, 950–955. Sollner, C., Rauch, G. J., Siemens, J., Geisler, R., Schuster, S. C., Muller, U., and Nicolson, T. (2004). Mutations in cadherin 23 aVect tip links in zebrafish sensory hair cells. Nature 428, 955–959. Sotomayor, M., Corey, D. P., and Schulten, K. (2005). In search of the hair‐cell gating spring elastic properties of ankyrin and cadherin repeats. Structure (Camb.) 13, 669–682. StauVer, E. A., Scarborough, J. D., Hirono, M., Miller, E. D., Shah, K., Mercer, J. A., Holt, J. R., and Gillespie, P. G. (2005). Fast adaptation in vestibular hair cells requires Myosin‐1c activity. Neuron 47, 541–553. Steyger, P. S., Gillespie, P. G., and Baird, R. A. (1998). Myosin Ibeta is located at tip link anchors in vestibular hair bundles. J. Neurosci. 18, 4603–4615. Sukharev, S. I., Blount, P., Martinac, B., and Kung, C. (1997). Mechanosensitive channels of Escherichia coli: The MscL gene, protein, and activities. Annu. Rev. Physiol. 59, 633–657. Swiatecka‐Urban, A., Boyd, C., Coutermarsh, B., Karlson, K. H., Barnaby, R., Aschenbrenner, L., Langford, G. M., Hasson, T., and Stanton, B. A. (2004). Myosin VI regulates endocytosis of the cystic fibrosis transmembrane conductance regulator. J. Biol. Chem. 279, 38025–38031. Tilney, L. G., and Saunders, J. C. (1983). Actin filaments, stereocilia, and hair cells of the bird cochlea. I. Length, number, width, and distribution of stereocilia of each hair cell are related to the position of the hair cell on the cochlea. J. Cell Biol. 96, 807–821. Tilney, M. S., Tilney, L. G., Stephens, R. E., Merte, C., Drenckhahn, D., Cotanche, D. A., and Bretscher, A. (1989). Preliminary biochemical characterization of the stereocilia and cuticular plate of hair cells of the chick cochlea. J. Cell Biol. 109, 1711–1723. Todi, S. V., Franke, J. D., Kiehart, D. P., and Eberl, D. F. (2005). Myosin VIIA defects, which underlie the Usher 1B syndrome in humans, lead to deafness in Drosophila. Curr. Biol. 15, 862–868. van Netten, S. M., and Kros, C. J. (2000). Gating energies and forces of the mammalian hair cell transducer channel and related hair bundle mechanics [In Process Citation]. Proc. R. Soc. Lond. B Biol. Sci. 267, 1915–1923. van Netten, S. M., Dinklo, T., Marcotti, W., and Kros, C. J. (2003). Channel gating forces govern accuracy of mechano‐electrical transduction in hair cells. Proc. Natl. Acad. Sci. USA 100, 15510–15515. Verpy, E., Leibovici, M., Zwaenepoel, I., Liu, X. Z., Gal, A., Salem, N., Mansour, A., Blanchard, S., Kobayashi, I., Keats, B. J., Slim, R., and Petit, C. (2000). A defect in harmonin, a PDZ domain‐containing protein expressed in the inner ear sensory hair cells, underlies usher syndrome type 1C [In Process Citation]. Nat. Genet. 26, 51–55.
374
Ricci and Kachar
Vollrath, M. A., and Eatock, R. A. (2003). Time course and extent of mechanotransducer adaptation in mouse utricular hair cells: Comparison with frog saccular hair cells. J. Neurophysiol. 90, 2676–2689. Waguespack, J. R., and Ricci, A. J. (2005). Aminoglycoside ototoxicity: Permeant drugs cause permanent hair cell loss. J. Physiol. 567, 359–360. Walker, R. G., and Hudspeth, A. J. (1996). Calmodulin controls adaptation of mechanoelectrical transduction by hair cells of the bullfrog’s sacculus. Proc. Natl. Acad. Sci. USA 93, 2203–2207. Wu, Y.‐C., Ricci, A. J., and Fettiplace, R. (1999). Two components of transducer adaptation in auditory hair cells. J. Neurophysiol. 82, 2171–2181. Zarei, M. M., and Dani, J. A. (1994). Ionic permeability characteristics of the N‐methyl‐ D‐aspartate receptor channel. J. Gen. Physiol. 103, 231–248. Zhao, Y., Yamoah, E. N., and Gillespie, P. G. (1996). Regeneration of broken tip links and restoration of mechanical transduction in hair cells. Proc. Natl. Acad. Sci. USA 93, 15469–15474. Zheng, L., Sekerkova, G., Vranich, K., Tilney, L. G., Mugnaini, E., and Bartles, J. R. (2000). The deaf jerker mouse has a mutation in the gene encoding the espin actin‐bundling proteins of hair cell stereocilia and lacks espins. Cell 102, 377–385. Zine, A., Hafidi, A., and Romand, R. (1995). Fimbrin expression in the developing rat cochlea. Hear. Res. 87, 165–169.
CHAPTER 13 Insights into the Pore of the Hair Cell Transducer Channel from Experiments with Permeant Blockers Sietse M. van Netten* and Corne´ J. Kros{ *Department of Neurobiophysics, University of Groningen, 9747AG, Groningen, The Netherlands { School of Life Sciences, University of Sussex, Falmer, Brighton BN1 9QG, United Kingdom
I. II. III. IV.
Overview Introduction Ionic Selectivity of the Transducer Channel Permeation and Block of Mechanoreceptor Channels by FM1‐43 A. Evidence for Permeation of FM1‐43 Through the Hair Cell Transducer Channel B. Permeation of FM1‐43 Through Other Mechanoreceptors C. FM1‐43 as a Screen for Functional Transducer Channels and Mechanoreceptors V. Permeation and Block of the Hair Cell Transducer Channel by Aminoglycoside Antibiotics A. Evidence for Permeation of Aminoglycoside Antibiotics Through the Transducer Channel B. Inferences About the Functional Geometry of the Transducer Channel Pore VI. Transducer Channel Block by Amiloride and Its Derivatives A. Amiloride and Amiloride Derivatives as Permeant Transducer Channel Blockers: A Reinterpretation B. Structure–Activity Sequences for Amiloride and Its Derivatives VII. Conclusions References
Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59013-1
376
van Netten and Kros
I. OVERVIEW This chapter considers recent experiments designed to infer properties of the ion‐conducting pore of the mechanoelectrical transducer channel of sensory hair cells using permeant blockers. By combining results from experiments with three classes of large cationic permeant blockers, the fluorescent dye FM1‐43, the aminoglycoside antibiotics, and the potassium‐sparing diuretic amiloride, information has been obtained on the free energy profile along the transducer channel’s pore as sensed by these blocker molecules. These energy profiles provide information about the position of the negatively charged binding site for the blockers (likely to be the channel’s selectivity filter) as well as about positively charged barriers near the extracellular and intracellular faces of the channel that impede the blockers’ permeation. The extracellular barrier is relatively modest and allows almost diVusion‐limited entry of blockers from the extracellular side. A larger intracellular energy barrier eVectively prevents exit of the blocking molecules from the intracellular side, trapping the blockers inside the hair cells. A putative geometrical model of the transducer channel pore is presented that draws on results from all three classes of permeant blockers. The pore contains a large vestibule that is easily accessible from the extracellular side. The negatively charged selectivity filter is located at the end of the vestibule, about 2 nm inside the pore from the extracellular side, with the total pore length estimated as 3 nm. Further experiments with permeant blockers may help toward understanding the molecular nature of the hair cell transducer channel and its relation with mechanoreceptor channels in other sensory systems.
II. INTRODUCTION The molecular identity of the hair cell transducer channel is, despite intensive research eVorts, still in question at the time of writing (Corey, 2006). Nevertheless, recent pharmacological characterization is beginning to yield information about key properties of the channel that may help toward its eventual molecular identification by comparison with other known channel types. Historically, the pharmacology of the transducer channel has been studied to try and find the mode of action of ototoxic drugs. Particularly for the widely used aminoglycoside antibiotics, large polycationic molecules which lead to permanent hearing loss (reviewed by Forge and Schacht, 2000) associated with hair cell degeneration (Wersa¨ll et al., 1973), the transducer channel has been implicated early on as a possible target. This is because aminoglycosides were found to block extracellularly recorded electrical responses when administered into the endolymph (into which the hair bundles
13. Permeant Blockers of the Transducer Channel
377
that contain the transducer channels protrude) but not the perilymph (which surrounds the basolateral membrane of the hair cells), both in the goldfish sacculus (Matsuura et al., 1971) and the guinea pig cochlea (Konishi, 1979). The first studies in which the eVects of aminoglycoside antibiotics were tested directly on the responses of individual hair cells showed that the drugs blocked the transducer channel from the extracellular but not the intracellular side (Ohmori, 1985; Kroese et al., 1989). What remained puzzling was how the drug molecules might enter the hair cells where they are believed to exert their ototoxic eVects (Schacht, 1986; Hiel et al., 1993). The possibility that the transducer channels might be slightly permeable to aminoglycosides was considered but thought unlikely. This was because at the time the diameter of the transducer channel pore was estimated as about 0.7 nm, based on its finite permeability to the ion channel blocker tetraethylammonium (TEA; Ohmori, 1985; Howard et al., 1988). The minimal diameter of a number of diVerent aminoglycoside antibiotics was estimated as about 1 nm. It was therefore considered uncertain at the time whether the channel block was a necessary step in the chain of events leading to ototoxicity (Kroese et al., 1989). Experiments with a large number of blockers, however, have led to a revision of the minimum pore size of the transducer channel to about 1.25 nm (Farris et al., 2004). Following on from the discovery that an unrelated large polycationic molecule, the styryl dye FM1‐43, permeates through the transducer channel (Gale et al., 2001), aminoglycoside block was reexamined, leading to the conclusion that these drugs also can enter the hair cells through the transducer channel after all (Marcotti et al., 2005). In this chapter, we will discuss experiments with a variety of permeant blockers and the insights these experiments provide into the pore structure of the transducer channel and its relation to other ion channel types.
III. IONIC SELECTIVITY OF THE TRANSDUCER CHANNEL The hair cell transducer channel is a nonselective cation channel with very similar permeabilities for the alkali cations and a much higher permeability for divalent metal ions, the highest for Ca2þ (Ohmori, 1985; Jorgensen and Kroese, 1995). The permeability sequence of the alkali metal ions corresponds to Eisenman sequence XI, pointing to a high negative charge density of the selectivity filter (Hille, 2001). The high aYnity of the Ca2þ ions for the selectivity filter has the eVect that Ca2þ, although more permeable, has a lower conductance than monovalent metal cations. The Ca2þ ions thus tend to linger around in the pore and act as permeant blockers of the transducer current with a half‐blocking concentration (KD) of about 1 mM and a Hill coeYcient of 1 (Crawford et al., 1991; Ricci and Fettiplace, 1998;
378
van Netten and Kros
Fettiplace and Ricci, 2006). Due to the composition of the endolymph (high Kþ, micromolar Ca2þ) most of the transducer current is in vivo carried by Kþ ions, with the Ca2þ concentration optimized to ensure a large transducer current with suYcient Ca2þ entering to drive adaptation (Ricci and Fettiplace, 1998). The current–voltage curves of transducer currents of mammalian outer hair cells (OHCs) in the presence of 1.3‐mM extracellular Ca2þ show inward rectification at negative potentials and outward rectification at positive potentials. This double rectification is compatible with Ca2þ acting as a permeant blocker that is dislodged from its binding site and extruded from the channel at extreme potentials (Kros et al., 1992; Kros, 1996). The simplest quantitative model that fits these observations (a single energy barrier formed by bound Ca2þ ions that impedes flow of monovalent cations) suggests that the binding site for Ca2þ is situated at an electrical distance about halfway through the pore (Kros et al., 1992).
IV. PERMEATION AND BLOCK OF MECHANORECEPTOR CHANNELS BY FM1‐43 A. Evidence for Permeation of FM1‐43 Through the Hair Cell Transducer Channel FM1‐43 (Fig. 1) is a fluorescent styryl dye with a divalent cationic head group related to TEA and a long lipophilic tail which enables it to partition reversibly into the outer leaflet of the cell membrane when present in the extracellular solution (Betz et al., 1992). On incorporation into the membrane, its fluorescence increases by two orders of magnitude. It cannot cross the lipid bilayer so that when it becomes internalized into cells by endocytosis, it remains trapped in the inner leaflet of endocytosed vesicles, the basis for its use as a tracker of endo‐ and exocytosis in living cells (Betz and Bewick, 1992; Cochilla et al., 1999). The dye readily labels sensory hair cells from their apical surface, leading to the conclusion that a rapid endocytotic process operates in their apical region (Seiler and Nicolson, 1999). However, a number of observations did not easily fit with this interpretation. Aminoglycosides and amiloride (also a transducer channel blocker; see Section VI.A) as well as high concentrations of extracellular Ca2þ inhibited dye labeling of hair cells in the Xenopus lateral line, leading to the suggestion that the dye might enter through the transducer channels instead (Nishikawa and Sasaki, 1996). These findings were confirmed and extended in the mammalian cochlea in which confocal microscopy showed rapid FM1‐43 entry in inner hair cells (IHCs) and OHCs (but not nearby supporting cells) from the apical surface (Gale et al., 2001). FM1‐43 loading was prevented by
13. Permeant Blockers of the Transducer Channel
379
FIGURE 1 Structure of five permeant blocker molecules of the mechanoelectrical transducer channel, as discussed and compared in this chapter. The styryl dyes FM1‐43 and FM3‐25 are related in that the two butyl end groups of the lipophilic tail in FM1‐43 are replaced by two octadecyl groups in FM3‐25 giving it the appearance of a fishhook. The aminoglycoside antibiotic dihydrostreptomycin has the largest end‐on diameter of the five molecules shown, while the potassium‐sparing diuretics amiloride and its analogue benzamil are the smallest molecules.
380
van Netten and Kros
pretreatment with a high concentration of the Ca2þ chelator EGTA, which would abolish transducer currents by cutting the tip links that gate the transducer channels (Crawford et al., 1991). Dye loading recovered with time hours after EGTA was removed, suggesting tip links regenerated over time (Zhao et al., 1996). The degree of FM1‐43 loading was larger in OHCs than in adjacent IHCs and larger in hair cells located more basally along the cochlear coils, thus showing correlation with the expected size of the (resting) transducer current (Kros et al., 1992; Kros, 1996; He et al., 2004). The most direct evidence for entry of FM1‐43 through the transducer channel came from recordings of transducer currents (Gale et al., 2001, replotted in Fig. 2). FM1‐43 blocked transducer currents in a voltage‐ dependent manner, such that the block was maximal at intermediate and reduced at extreme negative and positive membrane potentials. The current‐ voltage curves in the presence of FM1‐43 were in eVect exaggerated versions of the double rectification attributed to Ca2þ ions acting as permeant blockers of the transducer channel (Kros et al., 1992). The rectification was much steeper in the presence of FM1‐43, but the barrier for permeation of monovalent ions was about halfway through the transducer channel, just as for Ca2þ. The KD of FM1‐43 block was three orders of magnitude lower than that of Ca2þ, at 1‐3 mM (obviously varying with membrane potential) in the presence of 1.3‐mM extracellular Ca2þ, making it the most potent transducer channel blocker known to date (Gale et al., 2001; Farris et al., 2004).
FIGURE 2 Block of mechanoelectrical transducer currents in mouse OHCs by extracellularly applied FM1‐43. Data are from Gale et al. (2001), with permission, but blocked current is plotted here as a fraction of the control current. The voltage dependence of the block, in the presence of 1.3‐mM extracellular Ca2þ, is shown for diVerent extracellular concentrations ranging from 0.3 to 20 mM, as indicated on the right‐hand side of each trace. The blockage is most pronounced at membrane potentials close to zero and reduces at both hyperpolarizing and depolarizing membrane potentials, the former is evidence of permeation through the channel (Section IV.A). Copyright (2001) by the Society for Neuroscience.
13. Permeant Blockers of the Transducer Channel
381
FM1‐43 block was even more eVective when extracellular Ca2þ was lowered to 100 mM and much less so in 10‐mM Ca2þ. The Hill coeYcient was around 2 at intermediate potentials, reducing to near 1 at extreme negative and positive potentials, pointing to possible ionic interactions in the pore. Overall, these findings suggest that FM1‐43 competes with Ca2þ for a minimum of two binding sites about halfway through the pore of the transducer channel. When applied intracellularly, even 50‐mM FM1‐43 did not block the transducer currents at any potential, despite the low intracellular free Ca2þ concentration. This shows that FM1‐43 passes the channel preferentially in one direction, enabling the drug to accumulate in the hair cells. The kinetics of FM1‐43 block suggest that the drug molecules may be trapped inside the closed channel (Gale et al., 2001). FM1‐43 is an elongated molecule with a maximum end‐on diameter of about 0.78 nm and a length of about 2.2 nm, putting a minimum of about 0.8 nm on the size of the selectivity filter and suggesting a vestibule facing the extracellular side of the channel that is suYciently large to accommodate at least two FM1‐43 molecules. FM3‐25 is a derivative of FM1‐43 in which the two butyl end groups of the lipophilic tail (at the opposite end of the molecule to the divalent cationic head group) are replaced by two much longer octadecyl groups (Fig. 1). This increase from 4 to 18 carbons gives FM3‐25 a shape rather like a fishhook (Meyers et al., 2003), which would limit the depth to which the head group can penetrate the transducer channel pore to about 1.8 nm. FM3‐25 dye did not enter the hair cells or block the transducer current at any concentration tested (up to 30 mM). This suggests that the FM3‐25 cationic head group was unable to reach its binding site inside the channel pore, putting a lower limit of a distance of 1.8 nm into the pore on the position of the binding site, which may well be the selectivity filter.
B. Permeation of FM1‐43 Through Other Mechanoreceptors Using cell lines transfected with various ion channels, FM1‐43 dye has been demonstrated to permeate through TRPV1 channels when activated by capsaicin or heat, and through P2X2 purinoreceptors activated by ATP (Meyers et al., 2003). This is a useful observation as finding molecularly characterized ion channel types with structural similarities to the hair cell transducer channel may help with the search for its molecular identity. When injected into living mice, FM1‐43 was moreover able to label many diVerent types of sensory cells throughout the body, while not labeling nonsensory cells (Meyers et al., 2003). It is possible that FM1‐43 may have entered these cell types through their sensory transducer channels, again opening
382
van Netten and Kros
up new ways to put the hair cell mechanoelectrical transducer channel in context.
C. FM1‐43 as a Screen for Functional Transducer Channels and Mechanoreceptors As FM1‐43 requires transducer channels to be open for it to label hair cells and probably other sensory cells, it provides a rapid and convenient optical means to screen for the proper function of these channels. For example, mutations in the gene encoding the unconventional myosin, myosin VIIA, lead to deafness in humans and shaker1 mice. The hair bundles of Myo7a mutant mice need to be displaced beyond their physiological operating range for the transducer channels to open and they hyperadapt to excitatory stimuli (Kros et al., 2002). This points to a role for myosin VIIA in anchoring membrane proteins to the actin core of the stereocilia. The lack of transducer channels open at rest (normally 5–10%) in homozygous Myo7a mutant mice prevents FM1‐43 loading of the hair cells and labeling was only restored when the hair bundles were stimulated by large excitatory force steps (Gale et al., 2001). FM1‐43 could thus be used as a screen for mutations aVecting hair cell transduction. Other applications that have been used include screening for the onset during development of mechanoelectrical transduction by hair cells (Ge´le´oc and Holt, 2003; Si et al., 2003) and monitoring the recovery of hair cell function following ototoxic damage by aminoglycoside antibiotics (Taura et al., 2006).
V. PERMEATION AND BLOCK OF THE HAIR CELL TRANSDUCER CHANNEL BY AMINOGLYCOSIDE ANTIBIOTICS A. Evidence for Permeation of Aminoglycoside Antibiotics Through the Transducer Channel The studies of Ohmori (1985) and Kroese et al. (1989) showed that extracellularly applied aminoglycoside antibiotics reversibly blocked transducer currents at negative but not at positive potentials. Kroese et al. (1989) also found that intracellularly applied aminoglycosides did not block the transducer currents even at concentrations of 500 mM, one to two orders of magnitude higher than the concentrations with which they achieved block from the extracellular side. The KD for extracellularly applied aminoglycosides increased when the extracellular Ca2þ concentration was increased, pointing to competition between the two ions. Aminoglycoside block was
13. Permeant Blockers of the Transducer Channel
383
quantitatively interpreted as an interaction with a binding site inside the transducer channel, reachable only from the extracellular side, that blocks the flow of cations through the channel at negative potentials (Kroese et al., 1989). The Hill coeYcient was 1, implying that an interaction of one aminoglycoside molecule with the binding site is suYcient to block monovalent cation current through the channel. Assuming two positive charges for the aminoglycoside dihydrostreptomycin (DHS; see Fig. 1 for its structure), the binding site was estimated to be at an electrical distance of 20% into the channel from its extracellular side. Experiments by Marcotti et al. (2005) provided several advances over these early studies. By extending the range of membrane potentials used to voltage clamp individual hair cells, it became evident that block of the transducer current by extracellular DHS became progressively reduced at large negative potentials beyond 80 mV (Fig. 3). This could be explained if the blocking site was cleared due to the drug molecules being dragged through the channel pore into the cells by the large electrical driving force, in other words, if DHS acted as a permeant channel blocker like FM1‐43. Intracellular DHS, when tested at extremely high concentrations (up to 10 mM) and over a larger range of membrane potentials than before, was found to block the transducer current, but now at positive potentials, so with opposite polarity to
FIGURE 3 Mechanoelectrical transducer currents in mouse OHCs blocked by extracellularly applied DHS and depicted as a fraction of control currents (from Marcotti et al., 2005, with permission). The voltage dependence of the block, in the presence of 1.3‐mM extracellular Ca2þ, is shown for diVerent extracellular concentrations ranging from 1 to 100 mM DHS, as indicated with the various symbols. Continuous lines were calculated with the 2B1BS model [Section V.B.1; Eq. (6)]. Parameters used in the model are the same for all concentrations shown: E ¼ E2 E1 ¼ 4.63kT; ¼ 2 1¼ 0.91; Eb ¼ 8.27kT; b ¼ 0.79; nH ¼ 1, and z ¼ 2. The reduction of blockage at potentials more negative than about 92 mV [Eq. (8)] signifies an increased permeation of DHS through the channel at these voltages.
384
van Netten and Kros
extracellular DHS. This indicated that the DHS binding site in the channel could also be reached from the intracellular solution after all, further strengthening the case for DHS being a permeant blocker. The fact that the KD for intracellular block was two to three orders of magnitude higher than for extracellular block, together with the normally large negative driving force across the transducer channel (about 150 mV) brought about by the endocochlear potential explains why the transducer channel eVectively acts as a one‐way valve for aminoglycoside antibiotics so that they can accumulate in the hair cells. The kinetics of the block in response to mechanical steps and voltage steps showed that the transducer channel has to open first before DHS block can occur (Marcotti et al., 2005). In contrast to FM1‐43, DHS is thus an open channel blocker of the transducer channel. This may be the reason why the eYcacy of DHS block is reduced in Myo7a mutant mice, which have very low resting open probability of their transducer channels (Kros et al., 2002), and why the KD for rapidly adapting high‐frequency hair cells in the turtle cochlea is higher than for the more slowly adapting low‐frequency hair cells (Ricci, 2002). Moreover, the transducer current shuts oV more slowly than normal in the presence of DHS, implying that the channel cannot close with the blocker present in the pore. This fits with the reported disappearance of the gating compliance when the transducer channels are blocked by aminoglycosides (Howard and Hudspeth, 1988). This behavior of DHS is again in contrast to FM1‐43, which is likely to be able to reside inside the closed channel (Section IV.A). Perhaps this is due to DHS (molecular weight of the free base 583.6) being a somewhat bulkier molecule than FM1‐43 (molecular weight 451), with an end‐on diameter of about 1 nm and a length of about 1.5 nm. On the other hand, benzamil is smaller than both and also interferes with closure of the channel (Section VI.A), so other factors may be involved. The ability to enter hair cells through their transducer channels is likely to be a general property of aminoglycosides as qualitatively similar observations to those using DHS have been made with neomycin and gentamicin (Kros et al., 2006).
B. Inferences About the Functional Geometry of the Transducer Channel Pore To quantitatively describe the voltage and concentration dependences of blockage and permeation of the hair cell transducer channel by permeant blocker molecules, like DHS, a two‐barrier one‐binding site (2B1BS) model of the transducer channel will be considered. A similar approach has been utilized previously to describe voltage‐dependent proton block (Woodhull, 1973) and has been applied more generally to the binding of ions to, and
385
13. Permeant Blockers of the Transducer Channel
permeation through, an ion channel pore (Hille, 2001). An equivalent ‘‘punch‐through’’ model was considered for blockage of amiloride and several analogues of mechanosensitive channels in Xenopus oocytes, but discarded because the measured behavior at large hyperpolarizing potentials did not suggest permeant block (Lane et al., 1991). The model described here provides a more complete treatment of the 2B1BS model used by Marcotti et al. (2005) and introduces two new elements: the possibility to apply the model to blocking interactions with Hill coeYcients other than 1 and a quantitative analysis of drug binding from the intracellular side. The rate constants that are applied in the 2B1BS model description can be determined via Eyring’s rate theory applied to the two energy barriers (Fig. 4). This yields results that are equivalent to a stochastic description of diVusion, which can be interpreted as a random hopping of particles, surmounting energy barriers (Hille, 2001). The blocker molecules are
E2
E1
Eb
0
d1
db
d2
1
FIGURE 4 Free energy profiles of the 2B1BS model used to describe the voltage and concentration dependent block and permeation of extracellularly applied DHS (Section V) and amiloride (Section VI). The free energy diVerences are shown for three membrane voltages þ150 mV (top, dark gray curve), 0 mV (middle, black curve), and 150 mV (bottom, light gray curve). Changing the membrane voltage eVectively tilts the energy profile. The barriers are indicated with E1 and E2 and are located at relative electrical distance 1 and 2, respectively, as measured across the membrane from the extracellular side. The binding site with relative free energy Eb is located at b. The actual values of E1, E2, Eb, 1, 2, and b as shown are the results obtained for DHS at 1.3‐mM extracellular Ca2þ.
386
van Netten and Kros
attracted to a preferred position with low free energy which can therefore be considered as the binding site. Rate constants in Eyring’s rate theory depend exponentially on the ratio of Gibbs free energy of the barriers and the thermal energy kT, similar to Arrhenius’ equation of chemical reactions in which the rate constants kr are related to the ratio of activation energy EA and kT [kr / expð EA =ðkTÞÞ]. The free energy related to the membrane potential V contributes to EA and thus causes the rate constants to be voltage dependent. The equilibrium reached will therefore also depend on the membrane potential. In the following sections, we discuss some specific consequences of the 2B1BS model in relation to permeation of blocker molecules.
1. General Model Equations Assuming that the unblocked transducer channels are indicated by C, the blocked channels by CB, the extra‐ and intracellular blocker by Bo and Bi, and using the forward (k1, k2) and reverse (k 1, k 2) rate constants, the reaction equation is given by: k1
k2
k
k
C þ Bo , CB , C þ Bi 1
(1)
2
Experiments showed that DHS is an open channel blocker (Section V.A). In the further discussion below of the reaction equation [Eq. (1)], we assume that the blocker can reach the binding site from the moment that the channels are opened via a mechanical stimulus. Alternatively, the same dynamics may be obtained via a sudden change of the blocker concentration on either the extra‐ or intracellular side of the membrane or of the membrane voltage. Assigning t to the description of the dynamics as the time variable, C(t) and CB(t) ¼ 1 C(t) denote the time‐dependent fractions of unblocked and blocked channels. The rate of change of C(t) is dependent on the four rate constants k1, k 1, k2, and k 2 of the transitions across the barriers and the intra‐ and extracellular blocker concentrations ½Bi and ½Bo according to: dCðtÞ ¼ ð k1 ½Bo nH dt
k 2 ½Bi nH Þ CðtÞ þ ðk
1
þ k2 ÞCBðtÞ
(2)
Here, the Hill coeYcient nH is the degree of cooperativity of the binding process of the blocker to the binding site. A general solution of Eq. (2), describing the approach to equilibrium can be written as: CðtÞ ¼ Cð1Þ þ ½Cð0Þ
Cð1Þexpð t=tÞ;
(3a)
387
13. Permeant Blockers of the Transducer Channel
where the time constant t is given by: t¼
1 k1 ½Bo nH þ k 2 ½Bi nH þ k
1
(3b)
þ k2
The steady state or equilibrium solution of Eq. (3a) (reached if, t!1, so that dCðtÞ=dt ¼ 0) is then given by: Cð1Þ ¼ tðk
1
þ k2 Þ;
(4a)
which is assumed to equal the ratio of blocked current I to the control (unblocked) current Ic, yielding: I k 1 þ k2 ¼ Cð1Þ ¼ Ic k1 ½Bo nH þ k 2 ½Bi nH þ k
1
þ k2
(4b)
To incorporate voltage dependence in the model description, the rate constants (k1, k2, k 1, and k 2) can be quantified using Eyring’s rate theory, and thus expressed as exponential functions of the diVerence between the free energy of the barrier maxima and the binding site minimum (Hille, 2001). The free energy diVerences have both a chemical and a voltage‐dependent component. The voltage is assumed to vary with a fixed gradient so that it linearly changes the free energy across the membrane, eVectively tilting the overall free energy profile depending on the membrane potential V, as illustrated in Fig. 4 for three diVerent membrane potentials. Defining the maxima of the free energy of the barriers as E1 and E2 and the free energy of the binding site as Eb, with respect to V ¼ 0 (Fig. 4), and denoting their fractional positions across the membrane by1,2, and db the rate constants are: E1 d1 V 0 (5a) k1 ðV Þ ¼ k0 exp ðs molnH Þ 1 Vs kT E1 Eb ðd1 db ÞV k 1 ðV Þ ¼ k0 exp (5b) s 1 Vs kT E2 Eb ðd2 db ÞV k2 ðV Þ ¼ k0 exp (5c) s 1 Vs kT E2 ðd2 1ÞV 0 (5d) ðs molnH Þ 1 k 2 ðV Þ ¼ k0 exp Vs kT kT 21 J) is the thermal noise energy, Here Vs ¼ RT zF ¼ ze0 , where kT (4.1 10 z the blocker molecule’s valence, and e0 the elementary charge (1.6 10 21 C). The proportionality constant k0 of the first order rate constants k 1 and k2, and
388
van Netten and Kros
k0 ¼ k0 =molnH of the second order rate constants k1 and k 2 are related to the thermal noise energy divided by Planck’s constant (h ¼ 6.63 10 34 J s). This assumes that the transmission coeYcient is one (k0 ¼ kT=h; Hille, 2001). Inserting Eq. (5) into Eq. (4b) then yields: 0
I 1 ; ðV Þ ¼ nH Ic 1 þ ½Bo þ K2 ðV Þ ½Bi nH =K1 ðV Þ
(6a)
with K1 ðV Þ ¼ exp
Eb db V þ Vs kT
E 1 þ exp kT
dV Vs
;
(6b)
and K2 ðV Þ ¼ exp with E ¼ E2
E1 and d ¼ d2
E kT
ðd
1ÞV Vs
;
(6c)
d1
2. ‘‘Punch‐Through’’ of Extracellular Blocker Molecules A frequently applied experimental approach is that blocker molecules are applied extracellularly, while the intracellular concentrations remain low so that [Bi] ¼ 0 can be assumed. If, in addition, a blocker molecule binds to the binding site in the channel pore with a Hill coeYcient equal to 1 (nH ¼ 1, i.e., no cooperativity), Eq. (6) reduces to the Hill equation: Cð1Þ ¼
I 1 ; ðV Þ ¼ Ic 1 þ ½Bo =K1 ðV Þ
(7)
with K1 given by Eq. (6b). Then, K1(V) can be interpreted as a voltage‐ dependent half‐blocking concentration (KD). It can be shown that K1(V ) may assume a minimum at a specific membrane potential V0 with the consequence that at this voltage I/Ic is minimal so that at V0 the blocking eVect is maximal, independent from the applied extracellular concentration. This maximal block is reached at: Vs E db þ ln V0 ¼ (8) d kT d db As is obvious from Eq. (8), a minimum of K1 at V0 is only defined if d db > 0. Put into words, this condition means that a maximum in
389
13. Permeant Blockers of the Transducer Channel
blocking eYcacy is only reached if the distance between the energy barriers exceeds the distance of the binding site measured from the extracellular side. A maximum block at V0 means that the block of positively charged molecules decreases if the membrane potential becomes more hyperpolarized than V0. This eVect has been termed ‘‘punch‐through,’’ as it can be interpreted as the electric field clearing a blocker molecule from the binding site and forcing it into the cytoplasm. From Eq. (8), it is clear that for positively charged blocker molecules and increased energy barrier height E2 (i.e., increased E ), V0 will shift to more negative potentials. This means that the ‘‘punch‐through’’ is eVectively resisted by increasing E2 so that more blocker molecules will reside at the binding site. Inserting V0 into the expression for K1 [Eq. (6b)] gives: Eb Edb =d d =d K1 ðV0 Þ ¼ exp (9) d ðd db Þðdb þdÞ=d db b kT Apart from the dependence on the locations of energy barriers and binding site, given by the three most right‐hand factors of Eq. (9), the first right‐ hand (exponential) factor indicates that the free energy of the binding site Eb is the most important parameter governing the half‐blocking concentration K1. If Eb is low (or negative), the molecules are strongly attracted to the binding site, in‐line with Eq. (9), which shows that K1 exponentially falls with Eb. A low K1 means that even relatively low blocker concentrations will lead to a significant blocking eVect [Eq. (7)].
3. Permeation Rate The eVective rate of entry or ‘‘punch‐through’’ of blocker molecules at a specific membrane potential can be calculated from Eqs. (7) and (6b). In the absence of an intracellular concentration ([Bi] ¼ 0), the entry rate R, that is the number of molecules entering per open channel per second, equals the blocked fraction CB times k2: RðV Þ ¼ CB k2 ¼ ½1
Cð1Þk2 ¼
k2 ðV Þ
1 þ K1 ðV Þ=½Bo
(10)
If the block is of the ‘‘open channel’’ type, the total number of blocker molecules entering a cell per second through Nch channels with an average open probability of po is given by: Nentry ¼ Nch po RðV Þ ¼
Nch po k2 ðV Þ 1 þ K1 ðV Þ=½Bo
(11)
390
van Netten and Kros
Applying Eq. (11) to the OHC, an entry rate of 9000 DHS molecules per second per hair cell was calculated for a clinically relevant DHS concentration of 1 mM in the endolymph (further assuming 80 transducer channels per cell with a resting open probability of 0.3 due to the low endolymphatic Ca2þ concentration and an electrical driving force of 150 mV; Marcotti et al., 2005). In contrast to the entry limiting eVect of the strong binding (Eb ¼ 8.24kT ) and the intracellularly facing barrier having a considerable free energy (E2 ¼ 15.84kT ), the extracellular barrier, because of its relatively low free energy (E1 ¼ 10kT ), is easy to surmount by DHS. From the dynamics of DHS binding to the channel, a Ca2þ‐dependent rate constant (k1 ¼ 3.52 108 s 1 M 1) could be determined under low Ca2þ (100 mM) conditions (Marcotti et al., 2005). This relatively high value is indicative of an almost diVusion‐limited influx of extracellular cations (Hille, 2001; Li and Aldrich, 2004). 4. Asymmetry in Blocking Potency of Extracellularly and Intracellularly Applied DHS The salient features of the block by intracellular DHS reported by Marcotti et al. (2005; see Section V.A), namely the much lower potency of the block and block occurring at positive rather than negative membrane potentials can also be explained by the 2B1BS model. A reduction in eYcacy in reaching the binding site from the intracellular side is evident from Eq. (6), which shows that the intracellular concentration is eVectively multiplied by K2(V ) in comparison to the extracellular concentration of the blocker molecules. The main contribution of K2(V ) consists of the exponential factor exp( E/kT ), which amounts to about 8.8 10 3, in‐line with a reduced blocking eVect of two to three orders of magnitude for intracellularly applied blockers. Another interesting feature of the 2B1BS model is that if both the rate constants related to the second barrier (k2 and k 2) are taken zero, it reduces to the Langmuir isothermal relation, describing the classical Woodhull blockage model, used by Kroese et al. (1989) to describe their measured blocking eVects of aminoglycoside antibiotics. In the 2B1BS model, this corresponds to an infinite energy barrier height E2, preventing the molecules to surmount it. This leads to a saturating constant block at strongly hyperpolarizing potentials, unlike the new observations of extracellularly applied DHS by Marcotti et al. (2005), which can only be described with a more complete permeant (k2 6¼ 0) version of the 2B1BS model (Section V.B.2). The observation that the block of intracellularly applied DHS occurs at positive potentials can be fitted by the 2B1BS model if a small value of is applied: 0. This can be interpreted as the two barriers coinciding at the intracellular side so that eVectively no pathway to the extracellular space exists with the consequence that the drug cannot escape from its binding site
13. Permeant Blockers of the Transducer Channel
391
into the extracellular space. It is evident indeed from the measured block of intracellularly applied DHS (Marcotti et al., 2005) that no signs of release at positive potentials are present like they are observed at negative potentials for extracellularly applied DHS. A physical asymmetry between extra‐ and intracellularly applied DHS may arise from the orientation that the molecules possess while their positive head groups are bound to the binding site. As will be described in more detail in Section VII, the pore most likely has a negatively charged binding site at about two‐third of its length measured from the extracellular side, attracting the positive head group of a DHS molecule. When entering the pore from the intracellular side, the head group may bind to the same site, with the consequence that the DHS molecule is lined up with the pore so that its orientation is approximately at 180 with the geometry it has when entering from the extracellular side. In addition, part of the about 1.5‐nm‐long molecule could then still be in the cytoplasm, while the positive electrical field applied would not exert suYcient force to ‘‘punch’’ it out. This behavior would be consistent with DHS plugging the pore when applied from the inside and is therefore best described with the reduced version of the 2B1BS model that is equivalent to the classical Woodhull blockage model, as discussed above.
VI. TRANSDUCER CHANNEL BLOCK BY AMILORIDE AND ITS DERIVATIVES A. Amiloride and Amiloride Derivatives as Permeant Transducer Channel Blockers: A Reinterpretation The synthetic drug amiloride and related compounds find clinical application as potassium‐sparing diuretics, thanks to their high‐aYnity blocking action (at submicromolar concentrations) on epithelial Naþ channels in the distal and collecting tubules of the kidney (Kleyman and Cragoe, 1988). Amiloride has been found to reversibly inhibit the hair cell transducer current, but at higher concentrations (KD around 50 mM; Jorgensen and Ohmori, 1988; Ru¨sch et al., 1994), which is probably why no adverse eVects of clinical treatment with amiloride on the auditory system have been reported. The drug has been used in early eVorts to search for similarities of the hair cell transducer channel with other known ion channels and to gain information about the channel’s gating mechanism. Amiloride and most of its derivatives are weak bases with a large fraction being protonated at physiological pH and hence carrying positive charges. With a molecular weight of 229.6 for the free base, amiloride is rather smaller than FM1‐43 and the aminoglycosides and the maximum end‐on diameter is about 0.6 nm
392
van Netten and Kros
(Fig. 1). Studies using amiloride and benzamil (Fig. 1; a larger derivative of amiloride, with a molecular weight of 319.8, that blocks the transducer channel an order of magnitude more strongly) showed that, as for aminoglycosides, block is relieved at positive potentials. Furthermore, the drug‐ binding site can only be reached when the channel is open and the channel cannot close with the drug bound (Ru¨sch et al., 1994). The Hill coeYcients for amiloride and related compounds were close to 2 (Ru¨sch et al., 1994), suggesting that two drug molecules are needed to block the channel, as for FM1‐43 (Gale et al., 2001), but diVerent from aminoglycosides for which one molecule suYces to block the channel (Kroese et al., 1989; Marcotti et al., 2005). Intracellularly applied amiloride at a high concentration of 1 mM did not appear to aVect transducer currents (Ru¨sch et al., 1994). At the time, the voltage dependence of the transducer channel block by amiloride was interpreted as being due to a conformational change on depolarization of the channel obscuring two amiloride‐binding sites on the extracellular face of the channel, outside the electric field. At negative membrane potentials, the binding sites would be accessible and drug binding would result in an allosteric block of the channel (Ru¨sch et al., 1994; Kros, 1996). This ‘‘conformation model’’ of amiloride block was adopted to explain the incompleteness of the block at extreme negative potentials, a feature incompatible with the blocking mechanism that had been proposed for aminoglycosides by Kroese et al. (1989), namely that it would obstruct cation flow through the transducer channel by binding to a site within the pore reachable from the extracellular side only. A similar conformation model had already been applied to quantitatively explain amiloride block of mechanosensitive channels in frog oocytes (Lane et al., 1991), so this seemed to point to a satisfying similarity in the design of two diVerent types of mechanoreceptor channel. Ru¨sch et al. (1994) considered amiloride binding inside the channel pore unlikely but not impossible: The incompleteness of the block at extreme negative potentials was also compatible with the drug molecules being punched through the channel to the intracellular side. This explanation was not pursued further, mainly because permeation of the transducer channel by large molecules was considered inconceivable. The evidence that even larger molecules such as FM1‐43 and DHS could permeate the transducer channel, and the various similarities between amiloride and aminoglycoside block that became apparent in the study of Marcotti et al. (2005) prompts a reexamination of amiloride block. The block by amiloride and benzamil can in fact be seen to be partially released at the largest negative potentials tested (see Figs. 3 and 4 in Ru¨sch et al., 1994). Consequently, their data can be better fitted with the 2B1BS model (Section V.B) than with the original conformation model, which does not accommodate the reduction of the block at hyperpolarized potentials seen in
13. Permeant Blockers of the Transducer Channel
393
the data (Fig. 5). The fitting parameters are qualitatively similar to those for DHS, but with the binding site positioned about halfway through the pore, less diVerence in the height and distance of the two energy barriers, and a stronger attraction to the binding site in the pore. Most likely this reflects diVerences in size and position of the electrical charges between the amiloride and benzamil molecules on the one hand and DHS on the other. The finding that amiloride, like the aminoglycosides, can alter the mechanics of the transducer channel corroborates the idea that both act on the channel in essentially the same way (Denk et al., 1992; Wiersinga‐Post and Van Netten, 1998). Novel testable predictions from the new interpretation of the interaction of amiloride with the hair cell transducer channel include that the positively charged rather than the neutral form blocks the channel, that lowering extracellular Ca2þ will increase the potency of the block, and that at higher concentrations than those tested before, it will block transduction from the intracellular side. It remains to be seen whether amiloride turns out to be a permeant blocker of other mechanoreceptors, including those of frog oocytes.
FIGURE 5 Mechanoelectrical transducer currents in mouse OHCs blocked by extracellularly applied amiloride and benzamil, depicted as a fraction of control currents (data are from Fig. 4 in Ru¨sch et al., 1994, with permission). The voltage dependence of the block, in the presence of 1.3‐mM extracellular Ca2þ, is shown for three concentrations of amiloride (30, 60, and 100 mM) and one concentration of benzamil (10 mM). Continuous lines were calculated with the 2B1BS model [Section V.B.1; Eq. (6)], in contrast with an earlier interpretation of the same data in terms of a conformation model of blockage (Ru¨sch et al., 1994). Parameters used in the model are the same for all concentrations shown for amiloride: E ¼ E2 E1 ¼ 3.2kT; ¼ 2 1 ¼ 0.66; Eb ¼ 18.1kT; b ¼ 0.44, nH ¼ 2, and z ¼ 2. For benzamil, the parameters used were E ¼ E2 E1 ¼ 1.37kT; ¼ 2 1 ¼ 0.71; Eb ¼ 24.4kT; b ¼ 0.55, nH ¼ 2, and z ¼ 2. The better fits to the data with the 2B1BS model than those obtained with the conformational model indicates that also amiloride and benzamil are permeant blockers of the mechanoelectrical transducer channel.
394
van Netten and Kros
B. Structure–Activity Sequences for Amiloride and Its Derivatives Because of their clinical application and hence commercial interest, about a thousand bulkier compounds related to amiloride have been synthesized (Kleyman and Cragoe, 1988). This allows for a comparison as to how the addition of various side chains at diVerent positions of the molecule aVect the KD for its diVerent targets. For example, the so‐called structure‐activity sequence is very diVerent between the kidney epithelial Naþ channel and the hair cell transducer channel (Ru¨sch et al., 1994). Of the various ion channels and transporters that can be blocked by amiloride and its analogues, only the mechanosensitive channel of frog oocytes had a similar sequence to that of the hair cell transducer channel (Lane et al., 1992; Ru¨sch et al., 1994). It would be interesting to extend this analysis to mechanosensitive channels in other sensory systems to seek for common properties.
VII. CONCLUSIONS The results discussed in the preceding sections may be summarized by constructing a putative geometrical model of the transducer channel with a specific charge distribution lining the pore (Fig. 6). The interactions of the channel with the alkali metal cations, divalent cations (in particular Ca2þ), the permeant cationic blocker molecules discussed in this chapter as well as other polycationic blocker molecules that have been investigated (Farris et al., 2004), all support the conclusion that at about halfway the pore a negative charge distribution forms a binding site. A binding site’s location that is obtained from fitting barrier models (including the 2B1BS model) is usually estimated in terms of electrical distances. These cannot directly be translated into physical distances along the pore. The diVerences in the binding site’s apparent position as obtained from fitting the 2B1BS model to DHS (b ¼ 0.79), amiloride (b ¼ 0.44), and benzamil (b ¼ 0.55) could be due to diVerences in the blockers and may very well indicate the presence of a single dominating negatively charged region in the pore. These permeant molecules are structurally diVerent so that, when bound, the positive groups responsible for binding are located at diVerent distances from the negative charges. This would, under the assumption of equal valence (þ2), lead to significantly diVerent binding energies Eb as supported by the model fits (DHS ¼ 8.27kT; amiloride ¼ 18.1.1kT; benzamil ¼ 24.4kT ). The Eisenman sequence established for the transducer channel suggests this negatively charged region to be the selectivity filter. On the basis of the length of the nonblocking FM3‐25 molecule, the negative binding site was estimated to be located at about 2 nm inside the pore, measured from the extracellular side,
395
13. Permeant Blockers of the Transducer Channel
In
Out
1 nm FIGURE 6 Putative model of the hair cell transducer channel pore. The charge distribution and dimensions depicted are based on a combination of results from fitting the 2B1BS model to DHS blocking data and results on other permeant blockers and other data. Positive charges reflect energy barriers, while negative charges indicate a binding site, most likely related to the selectivity filter (Section III). The large vestibule on the extracellular side (Out) must reach at least 1.8 nm into the pore as to prevent FM3‐25 molecules to reach the negatively charged binding site depicted at about 2 nm from the extracellular side (Section IV.A). The narrowest part of the pore, facing the intracellular side (In), must have at least an eVective diameter of 1 nm in order to accommodate permeation of the large aminoglycoside molecules into the cytoplasm (Farris et al., 2004, who estimate its narrowest diameter as 1.25 nm). The total length of the channel pore is estimated to be 3 nm. The transducer channel’s gate, engaged by the gating spring, is schematically depicted at the extracellular side. Gating of the transducer channel, however, is most likely related to conformational changes of the channel that eVectively closes the pore to ions and permeant molecules, some of which (e.g., Ca2þ and FM1‐43) may become trapped inside the vestibule.
while the total membrane spanning pore length was approximated by about 3 nm (Farris et al., 2004). The two barriers assumed in the 2B1BS model can for all three types of molecules described with the 2B1BS model be associated with positive free energies and are thus indicated with positive charges (Fig. 6). Also, for these three types of molecules, the barrier energy at the intracellular side (E2) exceeds that of the extracellularly facing one (E1) with one to several times kT. The lower energy of the extracellular barrier is associated in the case of DHS with a second‐order (i.e., concentration dependent) rate constant that suggests a high, extracellular Ca2þ‐dependent influx of extracellular cations that is almost diVusion limited. This points to a relatively large vestibule, as depicted in Fig. 6. The vestibule should at least be large enough to accommodate two FM1‐43 molecules (maximum end‐on diameter 0.78 nm and length about 2.2 nm) that can be trapped if the channel’s gate closes. The gate is schematically indicated as an extracellularly located element. Gating of the transducer channel, however, is most likely related to conformational changes of the vestibular region of the pore of the
396
van Netten and Kros
channel that eVectively open and close the pore to ions and permeant molecules. Taken together, a functional channel pore model of the transducer channel is emerging that facilitates the entry into the hair cell of relatively large extracellular molecules that all seem to interact or compete with Ca2þ. The physiological significance of this design is that it is likely to be important for the extremely large single‐channel conductance of the transducer channel: a range of 150–300 pS has been reported for endolymphatic (i.e., low) extracellular Ca2þ concentrations (Ricci et al., 2003). These large conductances are similar to those reported for large‐conductance Ca2þ‐activated Kþ channels (BK channels) in which a large vestibule (but placed on the intracellular side of the channel) lined with a ring of negative charges is proposed to promote a large single‐channel conductance by allowing near‐diVusion limited access of intracellular Kþ ions to the selectivity filter (Brelidze et al., 2003; Li and Aldrich, 2004; Brelidze and Magleby, 2005). The single‐channel conductance of the transducer channel is tonotopically organized in that it gradually and systematically changes along the cochlea, with high‐frequency hair cells having the largest conductance (Ricci et al., 2003). Future work is likely to establish the molecular identity of the channel, which can then be investigated with several permeant molecules that have given us already some insights into the channel pore. For example, it would be interesting to test whether the tonotopic gradient in transducer channel conductance is due to a systematic change in the dimensions or number of negative charges in the vestibule. Acknowledgments Supported by the Netherlands Organisation for Scientific Research (S.M.v.N) and the MRC (C.J.K.). The authors thank Dr. Ce´cil J. W. Meulenberg for his comments on an early version of this chapter and his help with the preparation of Fig. 1.
References Betz, W. J., and Bewick, G. S. (1992). Optical analysis of synaptic vesicle recycling at the frog neuromuscular junction. Science 255, 200–203. Betz, W. J., Mao, F., and Bewick, G. S. (1992). Activity dependent fluorescent staining and destaining of living vertebrate motor nerve terminals. J. Neurosci. 12, 363–375. Brelidze, T. I., and Magleby, K. L. (2005). Probing the geometry of the inner vestibule of BK channels with sugars. J. Gen. Physiol. 126, 105–121. Brelidze, T. I., Niu, X., and Magleby, K. L. (2003). A ring of eight conserved negatively charged amino acids doubles the conductance of BK channels and prevents inward rectification. Proc. Natl. Acad. Sci. USA 100, 9017–9022. Cochilla, A. J., Angelson, J. K., and Betz, W. J. (1999). Monitoring secretory membrane with FM1‐43 fluorescence. Ann. Rev. Neurosci. 22, 1–10. Corey, D. P. (2006). What is the hair cell transduction channel? J. Physiol. 576, 23–28. Crawford, A. C., Evans, M. G., and Fettiplace, R. (1991). The actions of calcium on the mechano‐electrical transducer current of turtle hair cells. J. Physiol. 434, 369–398.
13. Permeant Blockers of the Transducer Channel
397
Denk, W., Keolian, R. M., and Webb, W. W. (1992). Mechanical response of frog saccular hair bundles to the aminoglycoside block of mechanoelectrical transduction. J. Neurophysiol. 68, 927–932. Farris, H. E., LeBlanc, C. L., Goswami, J., and Ricci, A. J. (2004). Probing the pore of the auditory hair cell mechanotransducer channel in turtle. J. Physiol. 558, 769–792. Fettiplace, R., and Ricci, A. J. (2006). Mechanoelectrical transduction in auditory hair cells. In ‘‘Vertebrate Hair Cells’’ (R. A. Eatock, R. R. Fay, and A. N. Popper, eds.), pp. 154–203. Springer Science and Business Media, New York. Forge, A., and Schacht, J. (2000). Aminoglycoside antibiotics. Audiol. Neurootol. 5, 3–22. Gale, J. E., Marcotti, W., Kennedy, H. J., Kros, C. J., and Richardson, G. P. (2001). FM1‐43 dye behaves as a permeant blocker of the hair‐cell’s mechanotransducer channel. J. Neurosci. 21, 7013–7025. Ge´le´oc, G. S. G., and Holt, J. R. (2003). Developmental acquisition of sensory transduction in hair cells of the mouse inner ear. Nat. Neurosci. 6, 1019–1020. He, D. Z. Z., Jia, S., and Dallos, P. (2004). Mechanoelectrical transduction of adult outer hair cells studied in a gerbil hemicochlea. Nature 429, 766–770. Hiel, H., Erre, J. P., Aurousseau, C., Bouali, R., Dulon, D., and Aran, J. M. (1993). Gentamicin uptake by cochlear hair cells precedes hearing impairment during chronic treatment. Audiology 32, 78–87. Hille, B. (2001). ‘‘Ion Channels of Excitable Membranes,’’ 3rd ed. Sinauer Associates, Sunderland, MA. Howard, G., and Hudspeth, A. J. (1988). Compliance of the hair bundle associated with gating of mechanoelectrical transduction channels in the bullfrog’s saccular hair cell. Neuron 1, 189–199. Howard, J., Roberts, W. M., and Hudspeth, A. J. (1988). Mechanoelectrical transduction by hair cells. Ann. Rev. Biophys. Biophys. Chem. 17, 99–124. Jorgensen, F., and Kroese, A. B. A. (1995). Calcium selectivity of the transducer channel in hair cells of the frog sacculus. Acta Physiol. Scand. 155, 363–376. Jorgensen, F., and Ohmori, H. (1988). Amiloride blocks the mechano‐electrical transduction channel of hair cells in the chick. J. Physiol. 403, 577–588. Kleyman, T. R., and Cragoe, E. J., Jr. (1988). Amiloride and its analogs as tools in the study of ion transport. J. Membrane Biol. 105, 1–21. Konishi, T. (1979). EVects of local application of ototoxic antibiotics on cochlear potentials in guinea pigs. Acta Otolaryngol. 88, 41–46. Kroese, A. B. A., Das, A., and Hudspeth, A. J. (1989). Blockage of the transduction channels of hair cells in the bullfrog’s sacculus by aminoglycoside antibiotics. Hear. Res. 37, 203–218. Kros, C. J. (1996). Physiology of mammalian cochlear hair cells. In ‘‘The Cochlea’’ (P. Dallos, A. N. Popper, and R. R. Fay, eds.), pp. 318–385. Springer‐Verlag, New York. Kros, C. J., Ru¨sch, A., and Richardson, G. P. (1992). Mechano‐electrical transducer currents in hair cells of the cultured neonatal mouse cochlea. Proc. R. Soc. Lond. B 249, 185–193. Kros, C. J., Marcotti, W., van Netten, S. M., Self, T. J., Libby, R. T., Brown, S. D. M., Richardson, G. P., and Steel, K. P. (2002). Reduced climbing and increased slipping adaptation in cochlear hair cells of mice with Myo7a mutations. Nat. Neurosci. 5, 41–47. Kros, C., Marcotti, W., and van Netten, S. (2006). Aminoglycoside ototoxicity depends on drug entry through the hair‐cell transducer channels. Assoc. Res. Otolaryngol. Abs.: p. 85. Lane, J. W., McBride, D. W., Jr., and Hamill, O. P. (1991). Amiloride block of the mechanosensitive cation channel in Xenopus oocytes. J. Physiol. 441, 347–366. Lane, J. W., McBride, D. W., Jr., and Hamill, O. P. (1992). Structure‐activity relations of amiloride and its analogues in blocking the mechanosensitive channel in Xenopus oocytes. Br. J. Pharmacol. 106, 283–286.
398
van Netten and Kros
Li, W., and Aldrich, R. W. (2004). Unique inner pore properties of BK channels revealed by quaternary ammonium block. J. Gen. Physiol. 124, 43–57. Marcotti, W., van Netten, S. M., and Kros, C. J. (2005). The aminoglycoside antibiotic dihydrostreptomycin rapidly enters hair cells through the mechano‐electrical transducer channels. J. Physiol. 567, 505–521. Matsuura, S., Ikeda, K., and Furukawa, T. (1971). EVects of streptomycin, kanamycin, quinine, and other drugs on the microphonic potentials of goldfish sacculus. Jpn. J. Physiol. 21, 579–590. Meyers, J. R., MacDonald, R. B., Duggan, A., Lenzi, D., Standaert, D. G., Corwin, J. T., and Corey, D. P. (2003). Lighting up the senses: FM1‐43 loading of sensory cells through nonselective ion channels. J. Neurosci. 23, 4054–4065. Nishikawa, S., and Sasaki, F. (1996). Internalization of styryl dye FM1‐43 in the hair cells of lateral line organs in Xenopus larvae. J. Histochem. Cytochem. 44, 733–741. Ohmori, H. (1985). Mechano‐electrical transduction currents in isolated vestibular hair cells of the chick. J. Physiol. 359, 189–217. Ricci, A. J. (2002). DiVerences in mechano‐transducer channel kinetics underlie tonotopic distribution of fast adaptation in auditory hair cells. J. Neurophysiol. 87, 1738–1748. Ricci, A. J., and Fettiplace, R. (1998). Calcium permeation of the turtle hair cell mechanotransducer channel and its relation to the composition of endolymph. J. Physiol. 506, 159–173. Ricci, A. J., Crawford, A. C., and Fettiplace, R. (2003). Tonotopic variation in the conductance of the hair cell mechanotransducer channel. Neuron 40, 983–990. Ru¨sch, A., Kros, C. J., and Richardson, G. P. (1994). Block by amiloride and its derivatives of mechano‐electrical transduction in outer hair cells of mouse cochlear cultures. J. Physiol. 474, 75–86. Schacht, J. (1986). Molecular mechanisms of drug‐induced hearing loss. Hear. Res. 22, 297–304. Seiler, C., and Nicolson, T. (1999). Defective calmodulin‐dependent rapid apical endocytosis in zebrafish sensory hair cell mutants. J. Neurobiol. 41, 424–433. Si, F., Brodie, H., Gillespie, P. G., Vazquez, A. E., and Yamoah, E. N. (2003). Developmental assembly of transduction apparatus in chick basilar papilla. J. Neurosci. 23, 10815–10826. Taura, A., Kojima, K., Ito, J., and Ohmori, H. (2006). Recovery of hair cell function after damage induced by gentamicin in organ culture of rat vestibular maculae. Brain Res. 1098, 33–48. Wersa¨ll, J., Bjorkroth, B., Flock, A., and Lundquist, P. G. (1973). Experiments on ototoxic eVects of antibiotics. Adv. Otorhinolaryngol. 20, 14–41. Wiersinga‐Post, J. E., and van Netten, S. M. (1998). Amiloride causes changes in the mechanical properties of hair cell bundles in the fish lateral line similar to those induced by dihydrostreptomycin. Proc. R. Soc. Lond. B 265, 615–623. Woodhull, A. M. (1973). Ionic blockage of sodium channels in nerve. J. Gen. Physiol. 61, 687–708. Zhao, Y.‐D., Yamoah, E. N., and Gillespie, P. G. (1996). Regeneration of broken tip links and restoration of mechanical transduction in hair cells. Proc. Natl. Acad. Sci. USA 93, 15469–15474.
CHAPTER 14 Models of Hair Cell Mechanotransduction Susanne Bechstedt and Jonathon Howard Max‐Planck‐Institute of Molecular Cell Biology and Genetics (MPI‐CBG), 01307 Dresden, Germany
I. Overview II. Introduction III. Transduction Channel Properties A. Localization and Number of Transduction Channels in Stereocilia B. Pore Properties C. Molecular Identity of the Transduction Channel IV. Gating A. Transduction Channel Kinetics and Thermodynamics B. Biophysical Concept of the Gating Spring C. Molecular Representation of the Gating Spring V. Active Hair Bundle Motility A. Adaptation B. Spontaneous Oscillations VI. Conclusions References
I. OVERVIEW Hair cell mechanotransduction is based on a finely tuned machinery residing in the hair bundle, the hair cell’s receptive organelle. The machinery consists of a transduction channel, an adaptation motor, the tip link, and many other components that reside in the stereocilia. The transduction channel is connected to and opened by a gating spring for which there are several molecular candidates. The interplay between the motor, the spring, the channel, and the tip link assures that the channel is always working at Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59015-5
400
Bechstedt and Howard
its most sensitive point of this machine, allowing very fast responses to a force stimulus. This chapter addresses the mechanisms and molecular components underlying mechanotransduction, adaptation, and motility in the hair bundle.
II. INTRODUCTION Hair cells are specialized receptor cells that transduce mechanical force (e.g., from sound waves, gravity, or vibrations) into an electrical signal. The receiving subcellular organelle, the hair bundle, exhibits a delicate and unique architecture (Fig. 1A). Rows of stereocilia, actin‐filled protrusions emerging from the apical surface of the hair cell, show a staircase‐like arrangement in height. In nonmammalian and vestibular hair cells, the kinocilium, a true cilium with an arrangement of nine concentric doublet microtubules surrounding two singlet microtubules, is found adjacent to the tallest stereocilia. Deflection of the hair bundle toward the largest stereocilium (Fig. 1B) gives rise to an excitatory receptor potential. In addition to many lateral links, the stereocilia are also interconnected by an apical tip link (Fig. 1C). Bundle deflection in the excitatory direction is thought to increase tension in the tip link, which leads to opening of the mechanoelectrical transduction channel (Fig. 1C and D), located close to the insertion site of the tip link. The channel must be directly gated by force because the gating time of about 10 ms is too short for second messenger signaling. The tension is thought to be conveyed to the channel via an elastic element termed the gating spring. The compliance of the gating spring allows the channel to rapidly fluctuate between open and closed positions even when the bundle is fixed, thereby allowing small displacements of the hair bundle to be detected as small changes of the probability of the channel being open. Following a large excitatory (or inhibitory) stimulus, the transduction machinery is able to adapt so that it can again respond sensitively to small hair bundle deflections. Adaptation is thought to be an active process, driven by myosin motors interacting with actin filaments that form the core of the stereocilia. The active process is also thought to lead to spontaneous oscillation, which may play a role in increasing the sensitivity of hair cells to sounds of particular frequencies. In this chapter, we review the electrophysiological, mechanical, and biochemical mechanisms underlying mechanoelectrical transduction. Of great current interest is the molecular identification of the transducer components—the channel, tip link, gating spring, and adaptation motor—and we discuss evidence for and against recently proposed candidates.
14. Models of Hair Cell Mechanotransduction
401
FIGURE 1 (A) Scanning electron micrograph of the bullfrog’s sacculus hair bundle (A. J. Hudspeth and R. A. Jacobs). (B) Deflection of the hair bundle causes the stereocilia to slide with respective to each other. (C) The tip link is tensed by the stereocilia‐sliding motion, which conveys the tension to the transduction channel. (B and C drawn after Fig. S1 from Sotomayor et al., 2005). (D) According to the gating spring hypothesis, the gate of the channel is coupled to an elastic element, which allows the channel to open and close rapidly without moving the whole bundle. The open question is where the gating spring resides.
III. TRANSDUCTION CHANNEL PROPERTIES A. Localization and Number of Transduction Channels in Stereocilia Many diVerent studies locate the transduction channel near the tip of the stereocilium (Hudspeth, 1982; Jaramillo and Hudspeth, 1991; Denk et al., 1995; Lumpkin and Hudspeth, 1995). Potential locations are near the end of the tip link in the shorter or the taller stereocilium, or at both ends (Fig. 1C).
402
Bechstedt and Howard
The attachment site at the lower stereocilium is characterized by electron dense plaques at the membrane and the cytoskeleton, a potentially stretched membrane and filaments that run from the membrane toward the actin cytoskeleton. At the taller stereocilium, the site of tip link contact is also characterized by electron‐dense material, called the insertional plaque (Kachar et al., 2000). Denk et al. (1995) observed calcium entry into the shortest as well as into the tallest stereocilium on bundle deflection in frog saccular hair cells, supporting the model of hair cell transducer channels at both ends of the tip link. Localization of components that are associated with the transduction channel support that theory as well: Myosin 1c, a candidate for the adaptation motor, as well as calmodulin, which may confer calcium sensitivity to the motor, show immunolocalization at both sides of the tip link (Garcia et al., 1998). However, the significant ultrastructural diVerences at the two sides of insertion of the tip links as well as the diVerent geometry with respect to the actin cytoskeleton argue against the same molecular channel complex acting on both sides. Channels on both sides connected by the tip link also imply a negative cooperativity between the two channels. Such negative cooperativity would prevent the channel apparatus from displaying the observed negative stiVness (Martin et al., 2000). Localization at the sites of calcium entry in other hair cells, especially from mammalian cochlea is necessary to resolve the issue of whether there are channels at both ends of the tip link. The number of transduction channels opened by hair bundle deflection is rather low. It has been estimated that there are around 50–100 functional channels per bundle, translating into 1–2 channels per stereocilium only, from the fluctuation analysis of the transduction current (Holton and Hudspeth, 1986), from the mechanical compliance of the hair bundle (Howard and Hudspeth, 1988), and from the relative size of the single channel current (Crawford et al., 1991; Ricci et al., 2003). This number is consistent with localization at either or both ends of the tip links.
B. Pore Properties The hair cell transduction channel is a nonselective cation channel that allows calcium and potassium, as well as other small mono‐ and divalent cations, to pass (Corey and Hudspeth, 1979a; Ohmori, 1985). The conductance of 100 pS is quite large (Ohmori, 1985; Crawford et al., 1991; Denk et al., 1995; Geleoc et al., 1997; van Netten and Kros, 2000; Ricci et al., 2003). The unusually large single‐channel conductance suggests a wide pore and this is supported by the fact that large organic compounds such as choline, tetraethylammonium (TEA), and dihydrostreptomycin (DHS) are
14. Models of Hair Cell Mechanotransduction
403
able to permeate the transduction channel (Marcotti et al., 2005). The conductivity is reduced to about 50% by high extracellular calcium and nearly doubled when lowering extracellular Ca2þ from 2.8 mM to 50 mM (Ricci et al., 2003); these results indicate that calcium interacts strongly with the pore. Ricci et al. (2003) also found that the conductance is positively correlated with the hair cell’s characteristic frequency: transduction channel properties such as conductance, activation, and adaptation vary with respect to the tonotopic organization of the sensory epithelia. No high aYnity blockers for the hair cell transduction channel have been found. However, there are several low aYnity blockers. These include inorganic cations such as Ca2þ(KD ¼ 1 mM), Mg2þ, La3þ, and Gd3þ (KD ¼ 10 mM) (Ohmori, 1985; Crawford et al., 1991; Kimitsuki et al., 1996; Ricci and Fettiplace, 1998). Transduction channels are also blocked by several aminoglycoside antibiotics including gentamicin and DHS at 1 mM concentrations (Kroese et al., 1989; Kimitsuki et al., 1996), as well as amiloride with KD ¼ 50 mM (Jorgensen and Ohmori, 1988; Rusch et al., 1994). The interaction site with the channel probably lies at the negatively charged selectivity filter: the polycationic aminoglycosides only block the receptor current at negative potentials and have only little eVect at positive potentials (Ohmori, 1985; Kroese et al., 1989). The electrical and pharmacological properties of the transduction channel— nonspecific selectivity to cations, large conductance, and weak block by polycations—are not suYciently unique to place it into any specific channel family (Hille, 2001).
C. Molecular Identity of the Transduction Channel 1. Candidate Families A number of channels from diVerent channel families have been suspected to comprise the transduction channel in hair cells. Members of the ENaC/DEG/ASIC family are known to be involved in many diVerent types of mechanotransduction. ENaC channels are involved in baroreception (Drummond et al., 1998, 2001) and the DEG/ENaC family members MEC‐4 and MEC‐10 have been implicated in touch reception in Caenorhabditis elegans. MEC‐4 has been proven to be part of the mechanotransducer channel in touch receptor cells in C. elegans: in vivo whole‐cell patch clamp recordings of C. elegans touch neurons showed that the MEC‐4 channel complex is directly activated by mechanical stimuli (O’Hagan et al., 2005). UNC‐8 in C. elegans (Tavernarakis et al., 1997) and pickpocket in Drosophila melanogaster (Ainsley et al., 2003) detect locomotion and body stretch.
404
Bechstedt and Howard
An anti‐ENaC antibody was found to label stereocilia tips (Hackney et al., 1992), but no further proof in that direction was found since then. Interestingly, DEG/ENaC channels are blocked by amiloride (Benos et al., 1995; Garty and Palmer, 1997). Taken together, this evidence makes the DEG/ENaC family a possible candidate for being the transduction channel. On the other hand, arguments against DEG/ENaC channels are that they have neither a large conductance nor the calcium permeability that are characteristic to the hair cell transduction channel. Furthermore, the amiloride block shows a much higher aYnity and has a diVerent mechanism (Rusch et al., 1994; Benos et al., 1995). Finally, in situ localization studies showed ENaC expression in the cochlea, but never in hair cells (Couloigner et al., 2001; Grunder et al., 2001). Thus, DEG/ENaC channels are unlikely to be the hair cells mechanotransduction channel. A member of the P2X family of ATP‐gated channels, P2X(2), is expressed in hair cells and localizes to the apical region of stereocilia (Housley et al., 1999). P2X(2) shows similar behavior with respect to Ca2þ ions as the transduction channel (Evans et al., 1996; Virginio et al., 1998; Ding and Sachs, 1999). There are a few thousands active P2X(2) channels in hair cells (Raybould and Housley, 1997), although they are not gated by mechanical stimulation. Because P2X(2) is upregulated after sustained loud noise, leading to a measurable increase of ATP‐gated inward current, it has been suggested that P2X(2) receptors have a regulatory role in hair cells (Wang et al., 2003). The transmembrane cochlear‐expressed gene TMC1 encodes a novel transmembrane protein that does not belong into any known channel family. Recessive and dominant mutations in TMC1 lead to congenital (DFNB7/ B11) and progressive hearing loss (DFNA36) in mice (Kurima et al., 2002; Vreugde et al., 2002). TMC1 is expressed in both inner and outer hair cells from an early stage in development. These lines of evidence make TMC1 a possible candidate for being the hair cell transduction channel, although there is no evidence that TMC1 actually forms a pore. Rather than being the transduction channel, it has been proposed that TMC1 is involved in traYcking of molecules to the plasma membrane or that it serves as an intracellular regulatory signal for diVerentiation of immature hair cells into fully functional auditory receptors (Marcotti et al., 2006). 2. TRP Channel Family The TRP channel family is the biggest and most divergent family of ion channels involved in sensory transduction (Clapham et al., 2001). Members of this family sense light, pain, stretch, fluid flow, heat, cold, pheromones, capsaicin as well as sweet, bitter, and umami taste. Besides acting as sensory channels, they are involved in a wide variety of cellular functions from Ca2þ
14. Models of Hair Cell Mechanotransduction
405
and Mg2þ homeostasis to cell‐cycle control (Reuter et al., 1992; Clapham et al., 2001; Nilius and Voets, 2005; Ramsey et al., 2006). The huge variability in function is accompanied by a similar diversity in selectivity, gating mechanisms, and other channel properties. The involvement of TRPs in mechanotransduction processes ranging from stretch and touch in invertebrates to kidney fluid flow in vertebrates as well as the variability within the family, make TRP channels an attractive group in which to search for the hair cell transducer channel. a. NompC. The first evidence for TRP channels involved in mechanotransduction was a mechanoreception defective mutant in Drosophila called NompC, which also showed defects in hearing (Walker et al., 2000). NompC is also expressed in hair cells in zebrafish (Danio rerio) and Xenopus (Sidi et al., 2003, 2005). It localizes to stereocilia and most prominently to the kinocilial bulb in Xenopus (Shin et al., 2005). Morpholino‐mediated knockdown and exon‐deletion in zebrafish give phenotypes such as missing acoustic startle reflex and tilted or circular swim behavior (Sidi et al., 2003). Furthermore, both the uptake of FM1–43 into neuromasts and microphonic potentials were abolished in morphants, indicating a role for NompC in auditory and vestibular function. The results point toward NompC being the transduction channel. There are several questions that arise when thinking about NompC as the transduction channel. First, it has not been found in higher vertebrates, leading to speculation that the gene has been lost from the genome (Corey, 2003). Second, in the fly NompC works in microtubule‐based mechanoreceptors (Fig. 2A–C), in contrast to the actin‐based stereocilia found in hair bundles. Third, NompC in fly mechanoreceptors is opened by compression of the dendrite at the site of mechanotransduction (Thurm et al., 1983), in contrast to the hair cell transduction channel which is opened by tension conveyed by the tip link. Although the latter diVerence could potentially be overcome by the presence of diVerent cytoskeleton adaptor molecules in diVerent types of mechanoreceptors, these discrepancies argue against NompC being the hair cell transducer channel. What then is the role of NompC in the hair cells of lower vertebrates? If NompC never made the transition from microtubule‐ to actin‐based mechanotransducers, then it is likely that NompC plays a role in kinocilium function, which is in good agreement with its localization to kinocilia in Xenopus (Shin et al., 2005). Perhaps the kinocilium of lower vertebrate hair cells is mechanically sensitive, providing a second mechanosensory system in these cells. But a role in higher vertebrates is unlikely, especially in the cochlea where the hair cells lose their kinocilium during development.
406
Bechstedt and Howard A
B
(i)
MIC
Tubular body
MT
Cuticle 25 nm (ii) K+
9 + 0 cilium
Support cells Neuron
(iii)
MIC
CTC
C Dendritic sheath
Transduction channel Microtubule
Cone−tubule connection FIGURE 2 (A) Schematic drawing of the campaniform receptor, a bristle type receptor in the fly. A bipolar neuron extends its single dendrite into the cuticle. The distal part of the dendrite is separated by a cilium (9 þ 0) from the cell body. The dendrite is filled with a very regular arrangement of microtubules called the tubular body. (B) A detailed view of the region between the membrane and microtubules in the tubular body. In the membrane, electron dense structures (MIC, membrane‐integrated cones) are visible. They connect via cone–tubule connections (CTCs) to microtubules (MTs). Panel (ii) was taken after bending the cuticle in the excitatory direction (excitatory stimulus). Clearly the distance between the membrane and the microtubule is shortening. The lower panel was taken after unphysiological bending of the cuticle. This situation nicely illustrates the cone–tubule connection (CTC). An extendable filamentous‐like protein spans the distance between the membrane and the microtubule cytoskeleton.
407
14. Models of Hair Cell Mechanotransduction
b. TRPA1. TRPA1 has been a hot candidate for the hair cell transducer channel. TRPA1, also known as ANKTM1 in Drosophila, was first found in nociceptive neurons and described as a cold‐sensing channel (Story et al., 2003). Later, the channel was reported to have a more general role in nociception and to be gated by pungent chemicals such as mustard oils and isothiocyanates (Bandell et al., 2004; Jordt et al., 2004). Like NompC, the channel possesses a large number (17) of intracellular ankyrin domains (Fig. 3). Many lines of evidence—localization, RNAi, and morpholino‐mediated knockdowns in mice and zebrafish, as well as the channel’s electrophysiological signature—support the role of TRPA1 as the transduction channel (Corey et al., 2004; Nagata et al., 2005). However, TRPA1 knockout mice (Bautista et al., 2006; Kwan et al., 2006) do not have any auditory or vestibular defects. Instead mice lacking TRPA1 show reduced sensitivity to mustard oil and bradykinin, as well as to painful cold and mechanical cutaneous stimuli.
NompC
TRP box
N 29 ankyrin repeats
1 2 3 4 5P6
C
5 nm FIGURE 3 Upper panel: schematic representation of the NompC channel structure. A large N‐terminus consists of 29 ankyrin repeats, followed by the six transmembrane domains and a short C‐terminus, containing the TRP box. Lower panels: ankyrin spring structure deduced from 12 ankyrin repeat structure (1N11; Michaely et al., 2002). Left panel: side view showing the spring‐like conformation of a 29 ankyrin repeat structure. Right panel: top view showing an almost perfect full turn of the 29 ankyrin structure.
408
Bechstedt and Howard
These results prove the suggested role of TRPA1 as a pain sensor in DRG neurons, but give no evidence of an auditory function for TRPA1. There is no hint of a transduction channel gene among the inherited deafness genes, even though a large gene necessary to encode a channel should be hit by mutagenesis with a relatively high frequency. Perhaps any interference with hair cell transduction channel function leads to early lethality, implying a second important function for the channel. Where do we go from here? The hunt for the transduction channel is wide open.
IV. GATING A. Transduction Channel Kinetics and Thermodynamics Transduction channel gating is fast and direct (Corey and Hudspeth, 1979b, 1983b). The extremely short delay, estimated to be only 13 ms at 37 C, excludes any diVusible second messenger involvement in channel gating. Instead, the channel has to be directly gated by mechanical forces acting on the hair bundle. Figure 4A–C shows an early experiment from Howard and Hudspeth (1987) in which the hair bundle is stimulated with a glass fiber. The fast onset of receptor current (Fig. 4C) is followed by a decline in current called adaptation. A fast and a slow adaptation phase can be distinguished (arrow and arrowhead, respectively). The position of the hair bundle shows a fast rebound (twitch, Fig. 4B arrow) in the opposite direction of the initial stimulus and then a slower relaxation in the stimulatory direction (Fig. 4B, arrowhead). These movements are mechanical correlates of fast and slow adaptation, respectively. The transducer is sensitive over a range of about 100 nm (Fig. 4E), corresponding to 1 of angular rotation of the stereocilia (Corey and Hudspeth, 1983a; Holton and Hudspeth, 1986; Ohmori, 1987; Howard and Hudspeth, 1988). At the perceptual threshold of hearing, stereocilia bundles are deflected by about 0.1–1 nm (Rhode and Geisler, 1967; Rhode, 1984), a stimulus corresponding to a current response of about 1 pA in the hair cell. The open probability at the bundle’s resting position is not zero. In experimental setups, it has been estimated that about 10–20% of the transduction channels are open (Corey and Hudspeth, 1983a; Ohmori, 1987). The system is already under tension and set near its point of maximal responsiveness such that a small deflection at the resting position give a large change in receptor response.
409
14. Models of Hair Cell Mechanotransduction A Mechanical stimulus
D 500 nm
Bundle stiffness (pN/nm) 2000 1600
100 ms B Bundle displacement 400
20 nm −200 C Receptor current
10 pA
E
−100 100 0 Displacement (nm)
200
Receptor potential (mV) 3 2 1
−2 −200
−100
0
100
200
Displacement (nm) FIGURE 4 (A) Mechanical stimulus exerted on a bullfrog’s sacculus hair bundle with an elastic glass fiber. (B) Bundle displacement due to the stimulus given in A. (C) Subsequent receptor current response (A–C from Howard and Hudspeth, 1987). (D) Bundle stiVness determined from experiments similar to A–C. The bundle’s stiVness is lowest in the region of a natural stimulus (few nanometers in the excitatory direction). (E) Receptor potential as a function of displacement determined by experiments as in A–C. For D and E, the black curve resembles experiments done without prior manipulation of the bundle, while the red curves show responses after a sustained stimulus in the excitatory direction and the blue curves show bundle responses after sustained deflection in the opposite direction. It is clear that bundle responses are the same each time just shifted by the amount of prior sustained deflection. This demonstrates the action of slow adaptation, mediated by myosin motors that are climbing or slipping along the actin cytoskeleton during the adaptation process, setting the bundle to its most sensitive point according to the new stimulus. D and E from Howard and Hudspeth (1988).
Transduction channel activation involves two steps, a fast one and a slow one (Corey and Hudspeth, 1983b; Crawford et al., 1989; van Netten and Kros, 2000). Gating kinetics are dependent on the magnitude of hair bundle deflection (Corey and Hudspeth, 1983b; Crawford et al., 1989). In response to large stimuli, the activation time constant is a few microseconds in mammals (Ricci et al., 2005), whereas time constants are a couple of hundred microseconds for small stimuli. Activation time constants also depend on the calcium concentration. Calcium ions also aVect adaptation: under low
410
Bechstedt and Howard
external Ca2þ (0.05 mM instead of 2.8 mM), the time constant of fast adaptation doubles (Fettiplace et al., 2003). This result might indicate another direct interaction of Ca2þ with the transducer channel in addition to its interaction with the pore or another component of the transduction machinery that is connected mechanically to the channel. The tuning of hair cells to diVerent frequencies is thought to involve diVerent kinetics of channel activation and fast adaptation (Ricci, 2002; Ricci et al., 2005). Activation and adaptation time courses depend on the frequency range of hearing of the species. For example, they are slower in turtle than in rat. In mammals, kinetics of channel activation (and adaptation) are more than one order of magnitude faster than in nonmammals, consistent with the need for higher frequency detection. The mammalian cochlea is tonotopically organized with hair cells tuned to higher frequency at the base and cells tuned to lower frequency at the apex. The variation in characteristic frequency of hair cells along the cochlea cannot be exclusively explained by the variation of mechanical properties of the cochlea tissue (e.g., basilar membrane). Frequency tuning appears to be augmented by intrinsic properties of the hair cells: for example, basal hair cells display faster kinetics than apical ones (Ricci et al., 2005). What causes these diVerences in kinetics between morphological similar hair bundles? The answer probably lies in the transduction channel complex itself. Splice variants, diVerent accessory subunits, or alternative channel composition could provide a toolbox for building the kinetic gradient in the cochlea. In order to make channel kinetics faster the gating spring has to become stiVer. In Section IV.C.3, we discuss the myosin light‐chain‐binding domain as a potential molecular representation of the gating spring. Myosin has the attractive feature that the light chains can be readily exchanged on multiple binding sites within the light‐chain‐binding domain. If diVerent hair cells have diVerent mixtures of light chains, and if diVerent light chains confer diVerent stiVness to the domain (Howard and Spudich, 1996), then gradual tonotopic organization could be achieved.
B. Biophysical Concept of the Gating Spring Fluid movement in the inner ear leads to bundle deflection, which has to be translated into a change of open probability of the transduction channel. As we noted earlier, the very short latency of opening of these channels implies that external mechanical forces must directly couple to the channel without involving a second messenger. To match the mechanical impedance of the channel molecule with that of the hair bundle, Corey and Hudspeth (1983b) postulated the existence of a gating spring, an elastic structure that transmits forces generated by the shearing of the stereocilia (Fig. 1B) to the molecular
14. Models of Hair Cell Mechanotransduction
411
gate of the channel (Fig. 1C and D). By compressing and extending the gating spring, the channel can flicker rapidly between its open and closed states without necessitating the much slower movements of the whole hair bundle (Fig. 1D). The advantage of such a mechanism is that because the current is proportional to the open probability, even one channel can convey graded information about the stimulus strength (Holton and Hudspeth, 1986). Sensitive mechanical measurements confirmed the existence of the gating spring. The gating spring postulate predicts that the open probability, p, depends approximately on the displacement of the hair bundle X (in the positive excitatory direction) according to p¼
1 1þe
zðX X0 Þ=kT
ð1Þ
where z is the single‐channel gating force, X0 is the displacement at which half the channels are open, and kT the Boltzmann constant times absolute temperature (Corey and Hudspeth, 1983a). The equation provides an approximal fit to the experimental data, yielding z 0.6 pN (Howard and Hudspeth, 1988; Hudspeth et al., 2000; Martin et al., 2000). The molecular interpretation of the gating force is z ¼ gkd
ð2Þ
where ¼ 0.14 is the geometrical gain between hair bundle displacement and gating spring extension (Howard et al., 1988), is the stiVness of a single gating spring, and d is the distance by which a gating spring shortens as a channel opens. Without addition of mechanical or structural data it is not possible to determine or d. According to the gating spring postulate, the opening and closing of the channel make the bundle less stiV within the range of displacements that the channels are most sensitive. This additional compliance is observed (Fig. 4D, black curve) and allows one to estimate the stiVness of the gating spring as well as providing another independent estimate of channel number. The number of channels agrees with that estimated from electrophysiology with about 1–2 channels per stereocilium, and the stiVness of each gating spring is estimated at 0.5 pN/nm. This also allows an estimate of the swing of the gate d to be 4–8 nm (Howard and Hudspeth, 1988; Martin et al., 2000). This distance implies that a force of 1 pN acting on the channel does 1–2kT (kT 4 10 21 J at room temperature) work during the closed to open transition. The swing of the gate is large compared to the size of the structural change associated with the opening of potassium channels, 1 nm from structural studies (Jiang et al., 2002). This indicates that the hair cell transduction channel may possess or be connected to a rigid lever that, by increasing the eVective swing of the gate, makes the opening more sensitive to force.
412
Bechstedt and Howard
Several experimental observations deviate from the two‐state model described by Eq. (1). For example, some studies (Corey and Hudspeth, 1983b; Kros et al., 2002) indicate that the open probability of the channels is not well described by the Boltzmann curve [Eq. (1)] and the gating compliance deviates from that predicted by the two‐state model (Jaramillo et al., 1993; Kros et al., 2002). These results fit better to a three‐state model with two distinct closed channel states (C1 ! C2 ! O) (Corey and Hudspeth, 1983b; Markin and Hudspeth, 1995).
C. Molecular Representation of the Gating Spring The tip link, the transduction channel itself, the lipid bilayer, or any accessory protein that attaches the channel to the actin cytoskeleton are potential molecular candidates for the gating spring. In the following sections, three possible candidates are discussed. 1. Cadherins Because transmission electron micrographs show the tip link to be a thin, delicate strand (Pickles et al., 1984), it was hypothesized that the tip link might have suYcient compliance to form the gating spring (Howard and Hudspeth, 1988). This hypothesis has been called into question by higher resolution microscopy showing that the tip link consists of two extended polypeptide filaments that form a helically coiled rope‐like structure that does not appear to possess suYcient compliance to form the gating spring (Kachar et al., 2000). Two diVerent cadherins have been suggested to be tip link components. Cadherin 23, also called otocadherin, is a calcium‐dependent cell adhesion molecule, containing a single transmembrane domain and 27 cadherin domains. It has been linked genetically to hearing and was suggested to be the tip link (Siemens et al., 2004; Sollner et al., 2004). However, a study by Michel et al. (2005) reports cadherin 23 as a component of transient lateral links during development, with no detection of cadherin 23 in the mature cochlea. A recent paper provides evidence that protocadherin‐15 might be a component of the tip link (Ahmed et al., 2006). For both cadherins the following argument applies: if the tip link is composed of cadherin, then the tip link is almost certainly not the gating spring, because the cadherin domains are expected to be almost inextensible in response to piconewton‐scale forces typical of physiological stimulation (Sotomayor et al., 2005). Thus, the molecular identity of the tip link is still uncertain. But irrespective of whether the tip link is cadherin 23 or protocadherin‐15, it is unlikely that cadherins constitute the gating spring.
14. Models of Hair Cell Mechanotransduction
413
2. Ankyrin Repeat Domain in TRP Channels Recently, the channel itself has come into focus in the search for the gating spring. In invertebrate mechanoreceptors, as well as in zebrafish hair cells, the candidate transduction channel is NompC, a member of the TRP channel family. In higher vertebrates another TRP channel family member, TRPA1, has been identified as a candidate transduction channel as discussed above (Corey et al., 2004). Both channel proteins share a large cytosolic domain containing 29 (NompC) or 17 (TRPA1) ankyrin domains (Fig. 3). What is a huge N‐terminal ankyrin repeat doing on a potential mechanotransducer channel? One possibility is that it targets the channel to its correct location, as shown for other membrane proteins (Bennett and Chen, 2001). Alternatively, it could be a protein‐interaction domain. Another possibility, however, is that it transmits mechanical forces to the channel’s gate. Extrapolating from a 12 ankyrin repeat protein (Michaely et al., 2002), 29 ankyrin repeats are expected to form approximately one turn of a helix with a pitch of about 20 nm (Howard and Bechstedt, 2004). Such a helical geometry is expected to confer compliance to the structure, even if the protein itself is quite rigid: a Young’s modulus of 1 GPa for the protein, typical for structural proteins such as actin and tubulin (Howard, 2001), would yield a stiVness of the helix of 1 pN/nm, similar to the gating spring. This order of stiVness has also been inferred from molecular dynamics simulations (Sotomayor et al., 2005). Indeed, direct mechanical measurements on ankyrin repeat proteins of various lengths by atomic force microscopy (AFM) confirm the stiVness to be on the order of 1 pN/nm (Lee et al., 2006). Thus, the ankyrin repeat domain is a good candidate for the gating spring (Corey and Sotomayor, 2004; Howard and Bechstedt, 2004). The actual stiVness of the channel complex may be two or four times larger than that of a single ankyrin repeat domain depending on whether a heterotetrameric or a homotetrameric channel architecture is assumed. This stiVness is in good agreement with the stiVness of 3 pN/nm measured for fly bristle receptors (Thurm et al., 1983). An appealing feature of the ankyrin spring hypothesis for invertebrate mechanoreceptors is that it can account for the filament‐like connection between the membrane, where the fly transduction channels are located, and the microtubule cytoskeleton (Fig. 2). The connection, called the membrane‐integrated cone, has a dimension of about 20 nm in a resting campaniform receptor dendrite (Thurm et al., 1983). The structure can be compressed such that the gap between the microtubules and the membrane becomes narrower (Fig. 2Bii). This is thought to occur during excitatory stimulation (Thurm, 1983). Bending of the cuticle in the inhibitory direction leads to stretching of the microtubule– membrane connection up to 65 nm (Vo¨lker, 1982). The observed length for the membrane–microtubule connection in the unstimulated, stimulated, and
414
Bechstedt and Howard
stretched situation (Fig. 2Bi–iii) matches very well the size of the ankyrin spring. The large elongation could be accommodated by unfolding of the ankyrin repeats, which is found to occur at forces >10 pN (Lee et al., 2006; Li et al., 2006). Thus, the case that the ankyrin helix forms the gating spring is strong in fly mechanoreceptors. The case that the gating spring in hair cells is an ankyrin helix is less strong than it is in flies. The main problem is that it is unclear whether the hair cell transduction channel is a member of the TRP family of channel proteins that contain ankyrin repeats. On the other hand, there are many similarities between the microtubule‐based mechanoreceptors found in insect hearing and touch organs and the actin‐based mechanoreceptors found in vertebrate hair cells. For example, fast transduction kinetics and adaptation have been found for the fly bristle receptor (Walker et al., 2000). Active amplification and spontaneous oscillations are also found in the fly’s ear, the Johnston’s organ (Gopfert and Robert, 2002, 2003). The potassium‐rich receptorlymph in fly bristle‐type receptors (Grunert and Gnatzy, 1987) is similar to the endolymph in the vertebrate inner ear (Wangemann and Schacht, 1996). The same molecules atonal (math1), delta, and notch are required for mechanoreceptor development in flies and mice. Thus, it seems that general biophysical and developmental principles are conserved from insects to mammals. Considering the morphological presence of a spring‐like molecule in fly mechanoreceptors, an ankyrin spring is still an attractive, though unproven hypothesis, for the vertebrate hair cell. 3. Myosin Lever Arm As mentioned before, any compliant protein in series with the gate of the transduction channel and the actin cytoskeleton could in principle act as the gating spring. One potential candidate is the motor protein myosin 1c. It consists of a head, a long light‐chain‐binding domain called the neck or lever domain, and a small C‐terminus. The neck region is formed by an ‐helix that contains three calmodulin‐binding IQ motifs. The neck can act as a lever arm to amplify small movements in the motor domains associated with changes in the nucleotide state (Uyeda et al., 1996; RuV et al., 2001). Of particular interest is that myosin II, which binds two calmodulin‐like light chains, has a stiVness of 0.7–2 pN/nm (Veigel et al., 1998) and myosin V, which has six light chains, has a stiVness of 0.2 pN/nm. Thus, myosins have suYcient compliance to act as the gating spring. Calculations suggest that the bending of the neck might contribute some or perhaps most of the compliance (Howard and Spudich, 1996). Myosin 1c in addition to being the adaptation motor (Holt et al., 2002; StauVer et al., 2005), might therefore also act as the gating spring. Thus, oVers an alternative to the tip link and the ankyrin helix as a compliant element in the transduction complex.
14. Models of Hair Cell Mechanotransduction
415
V. ACTIVE HAIR BUNDLE MOTILITY A. Adaptation As discussed earlier, there are two types of adaptation in hair cells. Fast adaptation, occurring on a timescale of 1 ms or less is thought to be due to the direct binding of Ca2þto the transducer channel or to an element directly in series with the channel, causing the channel to close. Slow adaptation occurring at a timescale of tens of milliseconds is thought to be due to myosin motors that climb up or slip back along the actin cytoskeleton (Howard and Hudspeth, 1987; Eatock, 2000; LeMasurier and Gillespie, 2005). Thus, the prevailing hypothesis is that fast and slow adaptations are caused by diVerent underlying mechanisms. These mechanisms can be distinguished by diVerent mechanical consequences for the hair bundle. The twitch in a fast movement of the hair bundle in opposite direction to the stimulus (Fig. 4B arrow) is thought to be the mechanical correlate of fast adaptation and represents the fast reclosure of transduction channels. The slow mechanical relaxation in the same direction as the stimulus (Fig. 4B, arrowhead) is thought to be the mechanical correlate of slow adaptation and represents the slipping (or climbing for stimuli of opposite direction) of adaptation motors down the actin cytoskeleton to decrease tension in the gating spring. The best candidate for the adaptation motor is myosin 1c (Holt et al., 2002; Batters et al., 2004; StauVer et al., 2005). Interestingly, recent studies suggest that both types of adaptation may be due to the same underlying mechanism: Ca2þ acting on the adaptation motor may trigger a rapid bundle movement and channel reclosure as a result of the negative stiVness of the gating apparatus (Tinevez, 2006; Tinevez et al., 2007). This is the same mechanism thought to drive spontaneous oscillations as described in the next section.
B. Spontaneous Oscillations Hair bundles have been observed to oscillate spontaneously (Crawford and Fettiplace, 1985; Howard and Hudspeth, 1987; Rusch and Thurm, 1990; Denk and Webb, 1992; Martin and Hudspeth, 1999; Martin et al., 2003). The hair bundle itself contains a motor that can give rise to hair bundle movements even in the absence of stimuli. It has been hypothesized that these active movements might be involved in cochlea amplification and be the cause for otoacoustic emission (Hudspeth, 1989). It has been shown experimentally that bullfrog hair bundles display spontaneous movements (example in Fig. 5A), which are noisy oscillations between 5 and 50 Hz (Martin et al., 2003). These spontaneous oscillations are
416
Bechstedt and Howard
driven by an active process (Martin and Hudspeth, 1999; Martin et al., 2001). The prevailing idea is that the system consisting of the transduction channel, gating spring, and adaptation motor together with Ca2þ, which acts as a negative feedback signal between channel opening and motor forces, is able to undergo spontaneous oscillations. Such processes require a region of negative stiVness in the force–displacement curve (Fig. 5B) and an element, such as the adaptation motor, that forces the system into this unstable region (Martin et al., 2000). Theoretical studies have shown that negative stiVness can produce the observed hair bundle movements, and that it can enhance sensitivity and frequency selectivity (Choe et al., 1998;
A 10 nm
200 ms
Force B
1
2 3
4
1
Displacement
2
C 3
4
Spontaneous oscillations 3
D
2
20 nm 4 1
100 ms FIGURE 5 Spontaneous oscillations. (A) Example for spontaneous oscillation in the bullfrog’s sacculus (from Martin et al., 2003). (B) Force–displacement curve showing the region of negative stiVness. Due to the negative slope, the bundle is bistable around zero displacement. The two stable points are indicated by green stars. (C) Scheme from the force–displacement curve showing the points between the bundle moves during spontaneous oscillations. (D) Model of a spontaneous oscillation. The numbers are indicating movement according the points in the force–displacement curve in C. B–D from Martin et al. (2000).
14. Models of Hair Cell Mechanotransduction
417
Camalet et al., 2000; Martin et al., 2003; Vilfan and Duke, 2003; Nadrowski et al., 2004). In Fig. 5B–D, a model for hair bundle oscillation is shown. Without stimulation, the bundle should reside at the point of zero force. The force– displacement curve (Fig. 5B) shows a negative slope at this point (black curve), indicating that the bundle is bistable. The two stable points are shown by green stars in Fig. 5B. If the bundle settles at the left stable point (negative bundle displacement) the transduction channel’s open probability will be zero, leading to a decrease in Ca2þ concentration. This causes adaptation and a shift of the force–displacement curve in the negative direction (blue curve) along the green dotted line. The stable point vanishes, and the bundle must jump (1 ! 2) to its positive stable point to maintain zero force conditions. At the new position, channel open probability increases and Ca2þ levels inside the stereocilium rise, shifting the force–displacement curve to the right (red) (slow process 2 ! 3). Here the opposite eVect happens. When the local minimum (3) reaches the abscissa, any further movement would violate the zero force condition and the bundle jumps (3 ! 4) to the left stable point. From here the sequence starts again giving rise to the characteristic oscillation pattern. In summary, the hair bundle is acting as a relaxation oscillator at a point of instability close to its maximum mechanical sensitivity, which is set by the adaptation motor. The interplay between a negative stiVness region in the force–displacement curve, the transduction channel and Ca2þ as a feedback signal can account for the observed hair bundle oscillations and possibly frequency‐selective amplification of hair cells (Tinevez, 2006; Tinevez et al., 2007). The adaptive shifts correspond to slow adaptation caused by myosins for lower frequency oscillations (Ju¨licher and Prost, 1997; Camalet et al., 2000) and fast adaptation for higher frequencies (Hudspeth, 1997; Ricci et al., 2000). The hair bundle’s active processes have been suggested to function in mammalian cochlea amplification (Chan and Hudspeth, 2005; Kennedy et al., 2005; Cheung and Corey, 2006). Clearly, the bundle is able to generate force due to the activity of the channel, the gating spring, calcium‐dependent channel reclosure, and the action of a myosin adaptation motor, which places the transduction channel complex at the point of maximal sensitivity. This mechanism is almost certainly involved in amplification in nonmammals that do not hear at such high frequencies as mammals. In mammals, somatic hair cell electromotility (Brownell et al., 1985) mediated by prestin (Zheng et al., 2000) has been suggested to be the cause of amplification (for review see Dallos and Fakler, 2002; Geleoc and Holt, 2003). As shown by Fettiplace and colleges (Kennedy et al., 2005), both somatic and hair bundle motilities can be found in outer hair cells of rats, indicating that both processes could take part in cochlea amplification.
418
Bechstedt and Howard
VI. CONCLUSIONS Work on hair cells has provided us with extensive knowledge about mechanoelectrical transduction. We know that the transduction channel has a large pore that is permeable to cations with preference for calcium, and that calcium interacts with the pore. Activation kinetics are very fast, implying that the channel is directly gated by mechanical force, which is transmitted to the channel via the gating spring. We also begin to understand how the interplay between the channel, calcium, and the adaptation motor is able to cause spontaneous hair bundle motility, a possible mechanism underlying cochlea amplification. Calcium is a key player, modulating channel activation, adaptation, and spontaneous oscillations. Calcium interacts with the transduction channel pore and may regulate myosin function by binding to its calmodulin light chains (Batters et al., 2004). Despite many years of research, we still do not know the molecular identity of many key players in the transduction complex. The transduction channel properties and evidence from invertebrates points toward the channel being a member of the TRP family. The tip link is probably at least partially formed by cadherins. How it couples to the channel and whether there is a channel at either side of the tip link is still unknown. The gating spring, a well‐characterized biophysical element in hair cell channel gating, might be formed from the light‐ chain‐binding domain of myosin or another compliant protein domain such as an ankyrin repeat domain. Genetic approaches have proved highly eVective in identifying transduction molecules, but may fail to identify essential proteins of the transduction complex. Gene expression analysis and proteomic approaches using mass spectrometry might oVer alternative ways forward. References Ahmed, Z. M., Goodyear, R., Riazuddin, S., Lagziel, A., Legan, P. K., Behra, M., Burgess, S. M., Lilley, K. S., Wilcox, E. R., GriYth, A. J., Frolenkov, G. I., Belyantseva, I. A., et al. (2006). The tip‐link antigen, a protein associated with the transduction complex of sensory hair cells, is protocadherin‐15. J. Neurosci. 26, 7022–7034. Ainsley, J. A., Pettus, J. M., Bosenko, D., Gerstein, C. E., Zinkevich, N., Anderson, M. G., Adams, C. M., Welsh, M. J., and Johnson, W. A. (2003). Enhanced locomotion caused by loss of the Drosophila DEG/ENaC protein Pickpocket1. Curr. Biol. 13, 1557–1563. Bandell, M., Story, G. M., Hwang, S. W., Viswanath, V., Eid, S. R., Petrus, M. J., Earley, T. J., and Patapoutian, A. (2004). Noxious cold ion channel TRPA1 is activated by pungent compounds and bradykinin. Neuron 41, 849–857. Batters, C., Arthur, C. P., Lin, A., Porter, J., Geeves, M. A., Milligan, R. A., Molloy, J. E., and Coluccio, L. M. (2004). Myo1c is designed for the adaptation response in the inner ear. EMBO J. 23, 1433–1440.
14. Models of Hair Cell Mechanotransduction
419
Bautista, D. M., Jordt, S. E., Nikai, T., Tsuruda, P. R., Read, A. J., Poblete, J., Yamoah, E. N., Basbaum, A. I., and Julius, D. (2006). TRPA1 mediates the inflammatory actions of environmental irritants and proalgesic agents. Cell 124, 1269–1282. Bennett, V., and Chen, L. (2001). Ankyrins and cellular targeting of diverse membrane proteins to physiological sites. Curr. Opin. Cell Biol. 13, 61–67. Benos, D. J., Awayda, M. S., Ismailov, I. I., and Johnson, J. P. (1995). Structure and function of amiloride‐sensitive Naþchannels. J. Membr. Biol. 143, 1–18. Brownell, W. E., Bader, C. R., Bertrand, D., and de Ribaupierre, Y. (1985). Evoked mechanical responses of isolated cochlear outer hair cells. Science 227, 194–196. Camalet, S., Duke, T., Julicher, F., and Prost, J. (2000). Auditory sensitivity provided by self‐tuned critical oscillations of hair cells. Proc. Natl. Acad. Sci. USA 97, 3183–3188. Chan, D. K., and Hudspeth, A. J. (2005). Ca2þ current‐driven nonlinear amplification by the mammalian cochlea in vitro. Nat. Neurosci. 8, 149–155. Cheung, E. L., and Corey, D. P. (2006). Ca2þ changes the force sensitivity of the hair‐cell transduction channel. Biophys. J. 90, 124–139. Choe, Y., Magnasco, M. O., and Hudspeth, A. J. (1998). A model for amplification of hair‐bundle motion by cyclical binding of Ca2þ to mechanoelectrical‐transduction channels. Proc. Natl. Acad. Sci. USA 95, 15321–15326. Clapham, D. E., Runnels, L. W., and Strubing, C. (2001). The TRP ion channel family. Nat. Rev. Neurosci. 2, 387–396. Corey, D. P. (2003). New TRP channels in hearing and mechanosensation. Neuron 39, 585–588. Corey, D. P., and Hudspeth, A. J. (1979a). Ionic basis of the receptor potential in a vertebrate hair cell. Nature 281, 675–677. Corey, D. P., and Hudspeth, A. J. (1979b). Response latency of vertebrate hair cells. Biophys. J. 26, 499–506. Corey, D. P., and Hudspeth, A. J. (1983a). Analysis of the microphonic potential of the bullfrog’s sacculus. J. Neurosci. 3, 942–961. Corey, D. P., and Hudspeth, A. J. (1983b). Kinetics of the receptor current in bullfrog saccular hair cells. J. Neurosci. 3, 962–976. Corey, D. P., and Sotomayor, M. (2004). Hearing: Tightrope act. Nature 428, 901–903. Corey, D. P., Garcia‐Anoveros, J., Holt, J. R., Kwan, K. Y., Lin, S. Y., Vollrath, M. A., Amalfitano, A., Cheung, E. L., Derfler, B. H., Duggan, A., Geleoc, G. S., Gray, P. A., et al. (2004). TRPA1 is a candidate for the mechanosensitive transduction channel of vertebrate hair cells. Nature 432, 723–730. Couloigner, V., Fay, M., Djelidi, S., Farman, N., Escoubet, B., Runembert, I., Sterkers, O., Friedlander, G., and Ferrary, E. (2001). Location and function of the epithelial Na channel in the cochlea. Am. J. Physiol. Renal Physiol. 280, F214–F222. Crawford, A. C., and Fettiplace, R. (1985). The mechanical properties of ciliary bundles of turtle cochlear hair cells. J. Physiol. 364, 359–379. Crawford, A. C., Evans, M. G., and Fettiplace, R. (1989). Activation and adaptation of transducer currents in turtle hair cells. J. Physiol. 419, 405–434. Crawford, A. C., Evans, M. G., and Fettiplace, R. (1991). The actions of calcium on the mechano‐electrical transducer current of turtle hair cells. J. Physiol. 434, 369–398. Dallos, P., and Fakler, B. (2002). Prestin, a new type of motor protein. Nat. Rev. Mol. Cell. Biol. 3, 104–111. Denk, W., and Webb, W. W. (1992). Forward and reverse transduction at the limit of sensitivity studied by correlating electrical and mechanical fluctuations in frog saccular hair cells. Hear. Res. 60, 89–102.
420
Bechstedt and Howard
Denk, W., Holt, J. R., Shepherd, G. M., and Corey, D. P. (1995). Calcium imaging of single stereocilia in hair cells: Localization of transduction channels at both ends of tip links. Neuron 15, 1311–1321. Ding, S., and Sachs, F. (1999). Ion permeation and block of P2X(2) purinoceptors: Single channel recordings. J. Membr. Biol. 172, 215–223. Drummond, H. A., Price, M. P., Welsh, M. J., and Abboud, F. M. (1998). A molecular component of the arterial baroreceptor mechanotransducer. Neuron 21, 1435–1441. Drummond, H. A., Welsh, M. J., and Abboud, F. M. (2001). ENaC subunits are molecular components of the arterial baroreceptor complex. Ann. NY Acad. Sci. 940, 42–47. Eatock, R. A. (2000). Adaptation in hair cells. Annu. Rev. Neurosci. 23, 285–314. Evans, R. J., Lewis, C., Virginio, C., Lundstrom, K., Buell, G., Surprenant, A., and North, R. A. (1996). Ionic permeability of, and divalent cation eVects on, two ATP‐gated cation channels (P2X receptors) expressed in mammalian cells. J. Physiol. 497(Pt. 2), 413–422. Fettiplace, R., Crawford, A. C., and Ricci, A. (2003). The eVects of calcium on mechanotransducer channel kinetics in auditory hair cells. In ‘‘Biophysics of the Cochlea’’ (A. W. Gummer, ed.). World Scientific Publishing, Singapore. Garcia, J. A., Yee, A. G., Gillespie, P. G., and Corey, D. P. (1998). Localization of myosin‐ Ibeta near both ends of tip links in frog saccular hair cells. J. Neurosci. 18, 8637–8647. Garty, H., and Palmer, L. G. (1997). Epithelial sodium channels: Function, structure, and regulation. Physiol. Rev. 77, 359–396. Geleoc, G. S., and Holt, J. R. (2003). Auditory amplification: Outer hair cells pres the issue. Trends Neurosci. 26, 115–117. Geleoc, G. S., Lennan, G. W., Richardson, G. P., and Kros, C. J. (1997). A quantitative comparison of mechanoelectrical transduction in vestibular and auditory hair cells of neonatal mice. Proc. Biol. Sci. 264, 611–621. Gopfert, M. C., and Robert, D. (2002). The mechanical basis of Drosophila audition. J. Exp. Biol. 205, 1199–1208. Gopfert, M. C., and Robert, D. (2003). Motion generation by Drosophila mechanosensory neurons. Proc. Natl. Acad. Sci. USA 100, 5514–5519. Grunder, S., Muller, A., and Ruppersberg, J. P. (2001). Developmental and cellular expression pattern of epithelial sodium channel alpha, beta and gamma subunits in the inner ear of the rat. Eur. J. Neurosci. 13, 641–648. Grunert, U., and Gnatzy, W. (1987). Kþ and Caþþ in the receptor lymph of arthropod cuticular mechanoreceptors. J. Comp. Physiol. [A] 161, 329–333. Hackney, C. M., Furness, D. N., Benos, D. J., Woodley, J. F., and Barratt, J. (1992). Putative immunolocalization of the mechanoelectrical transduction channels in mammalian cochlear hair cells. Proc. Biol. Sci. 248, 215–221. Hille, B. (2001). ‘‘Ion Channels of Excitable Membranes.’’ Sinauer Associates, Inc., Sunderland. Holt, J. R., Gillespie, S. K., Provance, D. W., Shah, K., Shokat, K. M., Corey, D. P., Mercer, J. A., and Gillespie, P. G. (2002). A chemical‐genetic strategy implicates myosin‐1c in adaptation by hair cells. Cell 108, 371–381. Holton, T., and Hudspeth, A. J. (1986). The transduction channel of hair cells from the bull‐frog characterized by noise analysis. J. Physiol. 375, 195–227. Housley, G. D., Kanjhan, R., Raybould, N. P., Greenwood, D., Salih, S. G., Jarlebark, L., Burton, L. D., Setz, V. C., Cannell, M. B., Soeller, C., Christie, D. L., Usami, S., et al. (1999). Expression of the P2X(2) receptor subunit of the ATP‐gated ion channel in the cochlea: Implications for sound transduction and auditory neurotransmission. J. Neurosci. 19, 8377–8388. Howard, J. (2001). ‘‘Mechanics of Motor Proteins and the Cytoskeleton.’’ Sinauer Associates, Inc., Sunderland.
14. Models of Hair Cell Mechanotransduction
421
Howard, J., and Bechstedt, S. (2004). Hypothesis: A helix of ankyrin repeats of the NOMPC‐TRP ion channel is the gating spring of mechanoreceptors. Curr. Biol. 14, R224–R226. Howard, J., and Hudspeth, A. J. (1987). Mechanical relaxation of the hair bundle mediates adaptation in mechanoelectrical transduction by the bullfrog’s saccular hair cell. Proc. Natl. Acad. Sci. USA 84, 3064–3068. Howard, J., and Hudspeth, A. J. (1988). Compliance of the hair bundle associated with gating of mechanoelectrical transduction channels in the bullfrog’s saccular hair cell. Neuron 1, 189–199. Howard, J., and Spudich, J. A. (1996). Is the lever arm of myosin a molecular elastic element? Proc. Natl. Acad. Sci. USA 93, 4462–4464. Howard, J., Roberts, W. M., and Hudspeth, A. J. (1988). Mechanoelectrical transduction by hair cells. Annu. Rev. Biophys. Biophys. Chem. 17, 99–124. Hudspeth, A. (1989). How the ear’s works work. Nature 341, 397–404. Hudspeth, A. (1997). Mechanical amplification of stimuli by hair cells. Curr. Opin. Neurobiol. 7, 480–486. Hudspeth, A. J. (1982). Extracellular current flow and the site of transduction by vertebrate hair cells. J. Neurosci. 2, 1–10. Hudspeth, A. J., Choe, Y., Mehta, A. D., and Martin, P. (2000). Putting ion channels to work: Mechanoelectrical transduction, adaptation, and amplification by hair cells. Proc. Natl. Acad. Sci. USA 97, 11765–11772. Jaramillo, F., and Hudspeth, A. J. (1991). Localization of the hair cell’s transduction channels at the hair bundle’s top by iontophoretic application of a channel blocker. Neuron 7, 409–420. Jaramillo, F., Markin, V. S., and Hudspeth, A. J. (1993). Auditory illusions and the single hair cell. Nature 364, 527–529. Jiang, Y., Lee, A., Chen, J., Cadene, M., Chait, B. T., and MacKinnon, R. (2002). The open pore conformation of potassium channels. Nature 417, 523–526. Jordt, S. E., Bautista, D. M., Chuang, H. H., McKemy, D. D., Zygmunt, P. M., Hogestatt, E. D., Meng, I. D., and Julius, D. (2004). Mustard oils and cannabinoids excite sensory nerve fibres through the TRP channel ANKTM1. Nature 427, 260–265. Jorgensen, F., and Ohmori, H. (1988). Amiloride blocks the mechano‐electrical transduction channel of hair cells of the chick. J. Physiol. 403, 577–588. Ju¨licher, F., and Prost, J. (1997). Spontaneous oscillations of collective molecular motors. Phys. Rev. Lett. 78, 4510–4513. Kachar, B., Parakkal, M., Kurc, M., Zhao, Y., and Gillespie, P. G. (2000). High‐resolution structure of hair‐cell tip links. Proc. Natl. Acad. Sci. USA 97, 13336–13341. Kennedy, H. J., Crawford, A. C., and Fettiplace, R. (2005). Force generation by mammalian hair bundles supports a role in cochlear amplification. Nature 433, 880–883. Kimitsuki, T., Nakagawa, T., Hisashi, K., Komune, S., and Komiyama, S. (1996). Gadolinium blocks mechano‐electric transducer current in chick cochlear hair cells. Hear. Res. 101, 75–80. Kroese, A. B., Das, A., and Hudspeth, A. J. (1989). Blockage of the transduction channels of hair cells in the bullfrog’s sacculus by aminoglycoside antibiotics. Hear. Res. 37, 203–217. Kros, C. J., Marcotti, W., van Netten, S. M., Self, T. J., Libby, R. T., Brown, S. D., Richardson, G. P., and Steel, K. P. (2002). Reduced climbing and increased slipping adaptation in cochlear hair cells of mice with Myo7a mutations. Nat. Neurosci. 5, 41–47. Kurima, K., Peters, L. M., Yang, Y., Riazuddin, S., Ahmed, Z. M., Naz, S., Arnaud, D., Drury, S., Mo, J., Makishima, T., Ghosh, M., Menon, P. S., et al. (2002). Dominant and recessive deafness caused by mutations of a novel gene, TMC1, required for cochlear hair‐cell function. Nat. Genet. 30, 277–284.
422
Bechstedt and Howard
Kwan, K. Y., Allchorne, A. J., Vollrath, M. A., Christensen, A. P., Zhang, D. S., Woolf, C. J., and Corey, D. P. (2006). TRPA1 Contributes to cold, mechanical, and chemical nociception but is not essential for hair‐cell transduction. Neuron 50, 277–289. Lee, G., Abdi, K., Jiang, Y., Michaely, P., Bennett, V., and Marszalek, P. E. (2006). Nanospring behaviour of ankyrin repeats. Nature 440, 246–249. LeMasurier, M., and Gillespie, P. G. (2005). Hair‐cell mechanotransduction and cochlear amplification. Neuron 48, 403–415. Li, L., Wetzel, S., Pluckthun, A., and Fernandez, J. M. (2006). Stepwise unfolding of ankyrin repeats in a single protein revealed by atomic force microscopy. Biophys. J. 90, L30–L32. Lumpkin, E. A., and Hudspeth, A. J. (1995). Detection of Ca2þ entry through mechanosensitive channels localizes the site of mechanoelectrical transduction in hair cells. Proc. Natl. Acad. Sci. USA 92, 10297–10301. Marcotti, W., van Netten, S. M., and Kros, C. J. (2005). The aminoglycoside antibiotic dihydrostreptomycin rapidly enters mouse outer hair cells through the mechano‐electrical transducer channels. J. Physiol. 567, 505–521. Marcotti, W., Erven, A., Johnson, S. L., Steel, K. P., and Kros, C. J. (2006). Tmc1 is necessary for normal functional maturation and survival of inner and outer hair cells in the mouse cochlea. J. Physiol. 574(Pt. 3), 677–698. Markin, V. S., and Hudspeth, A. J. (1995). Gating‐spring models of mechanoelectrical transduction by hair cells of the internal ear. Annu. Rev. Biophys. Biomol. Struct. 24, 59–83. Martin, P., and Hudspeth, A. J. (1999). Active hair‐bundle movements can amplify a hair cell’s response to oscillatory mechanical stimuli. Proc. Natl. Acad. Sci. USA 96, 14306–14311. Martin, P., Mehta, A. D., and Hudspeth, A. J. (2000). Negative hair‐bundle stiVness betrays a mechanism for mechanical amplification by the hair cell. Proc. Natl. Acad. Sci. USA 97, 12026–12031. Martin, P., Hudspeth, A. J., and Julicher, F. (2001). Comparison of a hair bundle’s spontaneous oscillations with its response to mechanical stimulation reveals the underlying active process. Proc. Natl. Acad. Sci. USA 98, 14380–14385. Martin, P., Bozovic, D., Choe, Y., and Hudspeth, A. J. (2003). Spontaneous oscillation by hair bundles of the bullfrog’s sacculus. J. Neurosci. 23, 4533–4548. Michaely, P., Tomchick, D. R., Machius, M., and Anderson, R. G. (2002). Crystal structure of a 12 ANK repeat stack from human ankyrinR. EMBO J. 21, 6387–6396. Michel, V., Goodyear, R. J., Weil, D., Marcotti, W., Perfettini, I., Wolfrum, U., Kros, C. J., Richardson, G. P., and Petit, C. (2005). Cadherin 23 is a component of the transient lateral links in the developing hair bundles of cochlear sensory cells. Dev. Biol. 280, 281–294. Nadrowski, B., Martin, P., and Julicher, F. (2004). Active hair‐bundle motility harnesses noise to operate near an optimum of mechanosensitivity. Proc. Natl. Acad. Sci. USA 101, 12195–12200. Nagata, K., Duggan, A., Kumar, G., and Garcia‐Anoveros, J. (2005). Nociceptor and hair cell transducer properties of TRPA1, a channel for pain and hearing. J. Neurosci. 25, 4052–4061. Nilius, B., and Voets, T. (2005). TRP channels: A TR(I)P through a world of multifunctional cation channels. Pflugers Arch. 451, 1–10. O’Hagan, R., Chalfie, M., and Goodman, M. B. (2005). The MEC‐4 DEG/ENaC channel of Caenorhabditis elegans touch receptor neurons transduces mechanical signals. Nat. Neurosci. 8, 43–50. Ohmori, H. (1985). Mechano‐electrical transduction currents in isolated vestibular hair cells of the chick. J. Physiol. 359, 189–217. Ohmori, H. (1987). Gating properties of the mechano‐electrical transducer channel in the dissociated vestibular hair cell of the chick. J. Physiol. 387, 589–609.
14. Models of Hair Cell Mechanotransduction
423
Pickles, J. O., Comis, S. D., and Osborne, M. P. (1984). Cross‐links between stereocilia in the guinea pig organ of Corti, and their possible relation to sensory transduction. Hear. Res. 15, 103–112. Ramsey, I. S., Delling, M., and Clapham, D. E. (2006). An introduction to trp channels. Annu. Rev. Physiol. 68, 619–647. Raybould, N. P., and Housley, G. D. (1997). Variation in expression of the outer hair cell P2X receptor conductance along the guinea‐pig cochlea. J. Physiol. 498(Pt. 3), 717–727. Reuter, G., Gitter, A. H., Thurm, U., and Zenner, H. P. (1992). High frequency radial movements of the reticular lamina induced by outer hair cell motility. Hear. Res. 60, 236–246. Rhode, W. S. (1984). Cochlear mechanics. Annu. Rev. Physiol. 46, 231–246. Rhode, W. S., and Geisler, C. D. (1967). Model of the displacement between opposing points on the tectorial membrane and reticular lamina. J. Acoust. Soc. Am. 42, 185–190. Ricci, A. (2002). DiVerences in mechano‐transducer channel kinetics underlie tonotopic distribution of fast adaptation in auditory hair cells. J. Neurophysiol. 87, 1738–1748. Ricci, A. J., and Fettiplace, R. (1998). Calcium permeation of the turtle hair cell mechanotransducer channel and its relation to the composition of endolymph. J. Physiol. 506(Pt. 1), 159–173. Ricci, A. J., Crawford, A. C., and Fettiplace, R. (2000). Active hair bundle motion linked to fast transducer adaptation in auditory hair cells. J. Neurosci. 20, 7131–7142. Ricci, A. J., Crawford, A. C., and Fettiplace, R. (2003). Tonotopic variation in the conductance of the hair cell mechanotransducer channel. Neuron 40, 983–990. Ricci, A. J., Kennedy, H. J., Crawford, A. C., and Fettiplace, R. (2005). The transduction channel filter in auditory hair cells. J. Neurosci. 25, 7831–7839. RuV, C., Furch, M., Brenner, B., Manstein, D. J., and Meyhofer, E. (2001). Single‐molecule tracking of myosins with genetically engineered amplifier domains. Nat. Struct. Biol. 8, 226–229. Rusch, A., and Thurm, U. (1990). Spontaneous and electrically induced movements of ampullary kinocilia and stereovilli. Hear. Res. 48, 247–263. Rusch, A., Kros, C. J., and Richardson, G. P. (1994). Block by amiloride and its derivatives of mechano‐electrical transduction in outer hair cells of mouse cochlear cultures. J. Physiol. 474, 75–86. Shin, J. B., Adams, D., Paukert, M., Siba, M., Sidi, S., Levin, M., Gillespie, P. G., and Grunder, S. (2005). Xenopus TRPN1 (NOMPC) localizes to microtubule‐based cilia in epithelial cells, including inner‐ear hair cells. Proc. Natl. Acad. Sci. USA 102, 12572–12577. Sidi, S., Friedrich, R. W., and Nicolson, T. (2003). NompC TRP channel required for vertebrate sensory hair cell mechanotransduction. Science 301, 96–99. Siemens, J., Lillo, C., Dumont, R. A., Reynolds, A., Williams, D. S., Gillespie, P. G., and Muller, U. (2004). Cadherin 23 is a component of the tip link in hair‐cell stereocilia. Nature 428, 950–955. Sollner, C., Rauch, G. J., Siemens, J., Geisler, R., Schuster, S. C., Muller, U., and Nicolson, T. (2004). Mutations in cadherin 23 affect tip links in zebrafish sensory hair cells. Nature 428, 955–959. Sotomayor, M., Corey, D. P., and Schulten, K. (2005). In search of the hair‐cell gating spring elastic properties of ankyrin and cadherin repeats. Structure (Camb.) 13, 669–682. StauVer, E. A., Scarborough, J. D., Hirono, M., Miller, E. D., Shah, K., Mercer, J. A., Holt, J. R., and Gillespie, P. G. (2005). Fast adaptation in vestibular hair cells requires myosin‐1c activity. Neuron 47, 541–553. Story, G. M., Peier, A. M., Reeve, A. J., Eid, S. R., Mosbacher, J., Hricik, T. R., Earley, T. J., Hergarden, A. C., Andersson, D. A., Hwang, S. W., McIntyre, P., Jegla, T., et al. (2003). ANKTM1, a TRP‐like channel expressed in nociceptive neurons, is activated by cold temperatures. Cell 112, 819–829.
424
Bechstedt and Howard
Tavernarakis, N., ShreZer, W., Wang, S., and Driscoll, M. (1997). unc‐8, a DEG/ENaC family member, encodes a subunit of a candidate mechanically gated channel that modulates C. elegans locomotion. Neuron 18, 107–119. Thurm, U. (1983). Mechano‐electric transduction. In ‘‘Biophysics’’ (W. Hoppe, W. Lohmann, H. Markl, and H. Ziegler, eds.), pp. 666–671. Springer‐Verlag, Berlin. Thurm, U., Erler, G., Goedde, J., Kastrup, T., Keil, T., Voelker, W., and Vohwinkel, B. (1983). Cilia specialized for mechanotransduction. J. Submicrosc. Cytol. 15, 151–155. Tinevez, J.‐Y. (2006). Mouvents actifs, regules par le calcium, de la touVe ciliaire des cellules ciliees mecano‐sensorielles de l’orieille interne. In ‘‘UFR de physique,’’ p. 244. Univesite Denis Diderot (Paris 7), Paris. Tinevez, J.‐Y., Julicher, F., and Martin, P. (2007). Myosin‐based adaptation in hair cells can mediate the various incarnations of active hair‐bundle motility. (in preparation). Uyeda, T. Q., Abramson, P. D., and Spudich, J. A. (1996). The neck region of the myosin motor domain acts as a lever arm to generate movement. Proc. Natl. Acad. Sci. USA 93, 4459–4464. van Netten, S. M., and Kros, C. J. (2000). Gating energies and forces of the mammalian hair cell transducer channel and related hair bundle mechanics. Proc. Biol. Sci. 267, 1915–1923. Veigel, C., Bartoo, M. L., White, D. C., Sparrow, J. C., and Molloy, J. E. (1998). The stiVness of rabbit skeletal actomyosin cross‐bridges determined with an optical tweezers transducer. Biophys. J. 75, 1424–1438. Vilfan, A., and Duke, T. (2003). Two adaptation processes in auditory hair cells together can provide an active amplifier. Biophys. J. 85, 191–203. Virginio, C., North, R. A., and Surprenant, A. (1998). Calcium permeability and block at homomeric and heteromeric P2X2 and P2X3 receptors, and P2X receptors in rat nodose neurones. J. Physiol. 510(Pt. 1), 27–35. Vo¨lker, W. (1982). Lebendbeobachtungen an kutikula¨ren Reizu¨bertragungsstrukturen campaniformer Sensillen und Hochauflo¨sungs‐Elektronenmikroskopie der reizaufnehmenden Sinneszellregion Dissertation. Mathematisch‐Naturwissenschaftliche Fakulta¨t, p. 126. Westfa¨lische Wilhelms‐Universita¨t, Mu¨nster. Vreugde, S., Erven, A., Kros, C. J., Marcotti, W., Fuchs, H., Kurima, K., Wilcox, E. R., Friedman, T. B., GriYth, A. J., Balling, R., Hrabe De Angelis, M., Avraham, K. B., et al. (2002). Beethoven, a mouse model for dominant, progressive hearing loss DFNA36. Nat. Genet. 30, 257–258. Walker, R. G., Willingham, A. T., and Zuker, C. S. (2000). A Drosophila mechanosensory transduction channel. Science 287, 2229–2234. Wang, J. C., Raybould, N. P., Luo, L., Ryan, A. F., Cannell, M. B., Thorne, P. R., and Housley, G. D. (2003). Noise induces up‐regulation of P2X2 receptor subunit of ATP‐gated ion channels in the rat cochlea. Neuroreport 14, 817–823. Wangemann, P., and Schacht, J. (1996). Homeostatic mechanisms in the cochlea. In ‘‘The Cochlea’’ (P. Dallos, A. N. Popper, and R. R. Fay, eds.), pp. 130–185. Springer, New York. Zheng, J., Shen, W., He, D. Z., Long, K. B., Madison, L. D., and Dallos, P. (2000). Prestin is the motor protein of cochlear outer hair cells. Nature 405, 149–155.
CHAPTER 15 Touch Liam J. Drew, Francois Rugiero, and John N. Wood Molecular Nociception Group, Biology Department, University College London, London WC1E 6BT, United Kingdom
I. Overview II. Introduction III. Structure of Skin and Touch Receptors A. Epidermis B. Dermis C. Mechanosensory Receptors IV. Physiology of Mechanoreceptive Nerve Fibers A. Low‐Threshold Mechanoreceptors B. High‐Threshold Mechanoreceptors V. Quantitating Mechanical Responses in Animal Models VI. Electrophysiological Approaches to Mechanosensation in Rodents VII. Mechanosensitive Ion Channels in Cultured Sensory Neurons VIII. Gating MS Ion Channels in DRG Neurons IX. Candidate Ion Channels A. DEG/ENaC Ion Channels B. TRP Ion Channels C. Mechanosensitive Potassium Channels X. Voltage‐Gated Channels and Mechanosensation A. Sodium Channels B. Calcium Channels XI. Indirect Signaling Between Sensory Neurons and Nonneuronal Cells XII. Conclusions References
Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59016-7
426
Drew et al.
I. OVERVIEW Light touch, a sense of muscle position, and the responses to tissue‐ damaging levels of pressure all involve mechanosensitive sensory neurons that originate in the dorsal root or trigeminal ganglia. A variety of mechanisms of mechanotransduction have been proposed. These range from direct activation of mechanically activated channels at the tips of sensory neurons to indirect eVects of intracellular mediators, or chemical signals released from distended tissues, or specialized mechanosensory end organs. In this chapter, we describe the properties of mechanosensitive channels present in sensory neurons and the potential molecular candidates that may underlie this type of activity.
II. INTRODUCTION In mammals, a number of diVerent types of mechanoreceptors respond to distinct mechanical stimuli. There are four principle mechanosensory systems each with specialized receptor cells that have evolved to detect diverse forms of mechanical events. These are: (1) touch (detection of mechanical events impacting on the skin, including noxious mechanosensation), (2) kinesthesia, or the awareness of position, location, and orientation of the body and its parts (a branch of proprioception originating from receptors in the muscles, joints, and bones), (3) body motion and balance (a branch of proprioception originating in the inner ear), and (4) hearing (the detection of sound waves by hair cells of the inner ear). In this chapter, we will focus on touch, with some reference to kinesthesia. Excellent reviews on proprioception, balance, and hearing can be found in the literature (Day and Fitzpatrick, 2005; LeMasurier and Gillespie, 2005; Macefield, 2005; and Howard in Chapter 15). Touch has been studied anatomically, electrophysiologically, pharmacologically, and most recently using brain imaging techniques (Hlushchuk and Hari, 2006). Remarkably, the primary transduction events that underlie this modality in mammals remain unknown despite extensive eVorts to identify the channels and receptors that are likely to be involved. In this chapter, we focus on aspects of touch and noxious mechanosensation mediated by primary mammalian sensory neurons that innervate the skin. We review the specialized end organs present in the skin that have been implicated in mechanosensation, discuss the properties of channels present on sensory neurons that are mechanosensitive, and describe the candidate molecules that may underlie light touch and noxious mechanosensation.
15. Touch
427
III. STRUCTURE OF SKIN AND TOUCH RECEPTORS Touch receptors are localized in the skin, the body’s largest organ. Anatomically, the skin is the main part of the integumentary system that comprises the nails, hair, glands (sweat and sebaceous), and specialized nerve structures detecting tactile stimuli, temperature changes, and tissue damage. The skin accounts for 12–15% of the body weight, an average area of 1.8 m2 in the adult human and a thickness of 2–3 mm on most of the body surface. The skin on the palms of the hands and the soles of the feet is particularly thick (owing to a high keratin content) and its surface is hairless and rich in papillae—ridges that are important for grip. The skin on the rest of the body is hairy, has fewer ridges, and is thinner and softer. Apart from being the site of tactile‐, temperature‐, and tissue‐damage‐evoked pain sensations, the skin also has a role as an anatomical barrier between external and internal environments preventing disease‐causing microorganisms to enter the body, and plays a role in temperature regulation. The skin comprises two tissue layers, the epidermis (outer layer) and the dermis (inner layer). A third deeper layer, the hypodermis, connects the skin to the underlying bones and muscles.
A. Epidermis The outermost layer of the skin is composed of several strata or stratified layers rimmed by an underlying basal membrane. From the outside to the inside, the strata making up the epidermis are the corneum, lucidum, granulosum, spinosum, and basale (Fig. 1A). The stem cells are produced in the innermost layers and diVerentiate while they move up to the distal layers. Apart from the stratum basale, the epidermis has no direct blood supply. Hence, the cells that migrate away from this layer are bound to die and when they eventually reach the corneum they are sloughed oV, a process that takes 35–45 days and that is known as desquamation. In these diVerent layers four types of cells are encountered: (1) Keratinocytes, whose role is to synthesize keratin, a protein that is the source of the skin’s strength and flexibility and that waterproofs the skin surface. (2) Langerhans cells, derived from a macrophage‐ monocyte precursor in the bone marrow, constitute an epithelial component of the immune system and play a role in the recognition and processing of antigens in order to present them to either lymphocytes and/or macrophages. (3) Melanocytes produce melanin, the dark pigment that gives the skin its color and that acts as a sunscreen to protect the skin from ultraviolet light. (4) Merkel cells, involved with pressure sensation (see below).
428
Drew et al. A Stratum corneum Stratum lucidum Stratum granulosum Stratum spinosum Stratum basale Basement membrane Papillary layer Dermis Reticular layer
B
Hairy skin Merkel’s disk
Glabrous skin
Epidermis
Epidermal− dermal border Free nerve ending Meissner’s corpuscle
Dermis
Hair follicle receptor Pacinian corpuscle Ruffini’s ending FIGURE 1 The skin and its receptors. (A) The layered structure of the skin. (B) The diVerent types of skin receptors. Adapted from Bear, M. F., Connors, B. W., Paradiso, M. A. ‘‘Neuroscience: Exploring the Brain,’’ 2nd ed. Lippincott Williams & Wilkins, 2nd Bk&Cdr ed. (March 15, 2002).
15. Touch
429
B. Dermis The dermis which lies beneath the epidermis is the main part of the skin containing sensory nerve terminals, blood supply, smooth muscle, hair follicles, glands, and lymphatic tissue. The dermis is enriched in fibroblasts, adipocytes, and macrophages and is the site of production of excreted substances by glands. The dermis consists of two layers, one made of loose connective tissue (papillary layer) and the other made of dense connective tissue (reticular layer) (Fig. 1A). These two layers are very tightly connected and rich in collagen for strength, reticular fibers for support, and elastin fibers for flexibility. The papillary layer lies beneath the epidermis and contacts this layer through fingerlike projections called papillae. These papillae have diVerent functions; some supply the epidermis with blood while others have a sensory function as they contain Meissner’s corpuscles (see below). The skin on the palms of the hands and on the soles of the feet contain two rows of papillae resulting in the finger‐ and footprints which protect the skin from tearing and help gripping. The reticular layer is denser, contains less organized fibers and is rich in collagen, hence resisting stretch. Pacinian corpuscles, the sensory receptors for deep pressure (see below) are found in this layer. The dermis contains also sweat glands, hair follicles, lymph vessels, and smooth muscle.
C. Mechanosensory Receptors Receptors in the skin can detect three diVerent sorts of sensations: tactile sensations, temperature changes, and tissue damage resulting in pain. Tactile sensations can be divided into three modalities (touch, pressure, and vibration), temperature into two (hot and cold), whereas pain may be evoked by mechanical, thermal, or chemical stimuli. All these so‐called ‘‘superficial’’ sensory modalities constitute, together with proprioceptive sensations (arising from muscles, joints, and ligaments) and visceral pain, the somatovisceral senses. The terminals of the sensory nerves express peripheral receptors and they exist either in a free form or embedded in more complex and integrated anatomical structures. Free nerve endings generally express receptors for damage sensing and thermosensation, whereas the nerve terminals associated with specific end organs are specialized mechanoreceptors. 1. Specialized Mechanoreceptors The skin can detect three diVerent modalities of tactile sensations. Touch is detected by Meissner’s corpuscles and hair follicle receptors. These two types of receptors respond to sudden light touch and therefore the critical
430
Drew et al.
factor in activating them is the velocity rather than the intensity of the stimulus. Vibration is detected by nerve fibers terminating in Pacinian corpuscles. On the basis of their firing properties, these three classes of aVerents are said to be low‐threshold, rapidly adapting (or phasic) mechanoreceptors. Finally, pressure is sensed superficially by Merkel cell–neurite complexes and deeper in the skin by RuYni’s endings. The responses of sensory fibers connected to these structures are proportional to the pressure applied on the skin. Both Merkel’s discs and RuYni’s endings are defined as low‐threshold, slowly adapting mechanoreceptors, but they diVer slightly in their mode of activation. Meissner’s corpuscles are low‐threshold, rapidly adapting (or phasic) mechanoreceptors and are found in the dermal papillae of the glabrous skin, mainly hand palms and foot soles but also lips, tongue, face, nipples, and genitals (Fig. 1B). Anatomically, they consist of an encapsulated nerve ending, the capsule being made of flattened supportive cells arranged as horizontal lamellae embedded in connective tissue. There is one single nerve fiber per corpuscle. Any physical deformation of the corpuscle triggers a volley of action potentials that quickly ceases. When the stimulus is removed, the corpuscle regains its shape and while doing so produces another volley of action potentials. Due to their superficial location in the dermis, these corpuscles are particularly sensitive to touch and vibrations, being able to respond to low‐frequency vibrations in the range of 20–40 Hz. Hair follicle receptors (G hairs) are unmyelinated sensory nerve terminals which coil around the shaft of a hair within the external root sheath (Fig. 1B). They respond to hair motion and its direction by firing trains of action potentials at the onset and removal of the stimulus. Merkel’s discs consist of a Merkel cell in close apposition to an enlarged nerve terminal. A single sensory fiber can branch to contact up to 90 Merkel cells. Merkel’s discs are found in the basal layer of the epidermis in fingers, lips, and genitals (Fig. 1B). Functionally, they are sensitive to very low‐ frequency vibrations (5–15 Hz) and respond to a low‐intensity constant pressure by a nonadapting volley of action potentials (static response) for up to 30 min. RuYni’s endings are thin cigar‐shaped encapsulated sensory endings that detect pressure when the skin is stretched. They are broadly expressed in the dermis (Fig. 1B). Pacinian corpuscles are low‐threshold mechanosensors that display a very rapid adaptation in response to the indentation of the skin. They are also called acceleration detectors because they can detect changes in the strength of the stimulus and if, as happens in vibrations, the rate of change in the stimulus is altered, that is the acceleration of the skin movement changes, their response becomes proportional to this change. Pacinian corpuscles are
15. Touch
431
expressed in the deep dermis (Fig. 1B) as well as in vessel walls. Anatomically similar to Meissner’s corpuscles, they are large ovoid corpuscles (1 mm in length) made of concentric lamellae of fibrous connective tissue and fibroblasts lined by flat modified Schwann cells. In the center of the corpuscle, in a fluid‐filled cavity called inner bulb, terminates one single aVerent unmyelinated nerve ending. Pacinian corpuscles sense gross pressure changes and most of all vibrations (in the range 150–300 Hz), which they can detect even centimeters away. They have a large receptive field on the skin’s surface with a particularly sensitive center. Notably, these corpuscles function also as proprioceptive detectors and are therefore highly expressed in ligaments, muscles, and joint capsules. 2. Free Nerve Endings In contrast to the majority of low‐threshold mechanoreceptors (LTMs) that terminate in specialized end organs, the sensory terminals of nociceptive neurons (and a number of temperature receptors) exist as bare nerve endings in the skin. Temperature receptors react to local changes in skin temperature. Six ion channels expressed on sensory nerve terminals and belonging to the transient receptor potential (TRP) channel family have been ascribed a function in thermosensation (for a thorough review see Patapoutian et al., 2003). The heat receptor function is supported by four channels from the vanilloid receptor subfamily: TRPV1, TRPV2, TRPV3, and TRPV4. TRPV3 and TRPV4 are activated by temperatures in the physiological range (above 33 C for TRPV3 and between 27 and 42 C for TRPV4) while TRPV1 and TRPV2, activated by temperatures above 42 and 52 C respectively, are the receptors for noxious heat. Two more TRP channels, the melastatin‐related TRPM8 and the ankyrin repeat‐rich TRPA1 are cold receptors. TRPM8 senses temperatures below 25 C whereas TRPA1 serves as a noxious cold receptor (activated by temperatures below 17 C). Between 20 and 40 C, the temperature receptors adapt quite rapidly but more extreme temperatures are continuously sensed as hot and cold to protect the skin from damage. Nociceptors are often polymodal, responding to chemical, mechanical, and temperature stimuli separately or together, when they are of a level of intensity that has the potential to cause tissue damage. Nociceptors are broadly of two classes, those that are peptidergic (i.e., they express the neuropeptides substance P and CGRP) and express TrkA and those that are nonpeptidergic, express c‐Ret receptor complexes in adulthood and bind isolectinB4 (IB4). Both classes of nociceptors innervate the skin whereas nociceptive innervation of joints and the viscera is almost exclusively by peptidergic neurons (Keller and Marfurt, 1991; Ivanavicius et al., 2004; Robinson et al., 2004). Interestingly, in the skin, IB4‐positive neurons appear to terminate more
432
Drew et al.
superficially, in the stratum granulosum, than do peptidergic neurons that end in the stratum spinosum (Zylka et al., 2005).
IV. PHYSIOLOGY OF MECHANORECEPTIVE NERVE FIBERS Sensory nerves have their cell bodies in the dorsal root ganglia (DRG) or trigeminal ganglia and transmit sensory information into the dorsal horn of the spinal cord or the brain stem at the level of the pons, respectively. Sensory nerve fibers can be classified according to their conduction velocities (determined by their diameters and their degree of myelination) into three groups. The nociceptor and temperature‐related fibers are mainly associated with two of these three classes of fibers. They are the C fibers (unmyelinated, diameter <1 mm, conduction velocity <1 m/s) and the Ad fibers (thinly myelinated, 2.5 mm diameter, <15 m/s conduction velocity) which terminate in laminae I, II and I, II, V of the spinal dorsal horn, respectively. However, a review also highlights the existence of nociceptors functioning in the Ab range (Djouhri and Lawson, 2004). The mechanosensation‐associated fibers are predominantly Ab fibers which are myelinated fibers with a large diameter (10 mm) and a conduction velocity in man of up to 60 m/s. These fibers terminate in laminae III, IV, and V of the spinal cord.
A. Low‐Threshold Mechanoreceptors LTMs are generally associated with myelinated Ab fibers, the largest and fastest type of sensory fibers. Nevertheless, two particular types of LTMs are also found on Ad and C fibers (see later ‘‘D‐hair’’ and ‘‘C‐LT’’). Microelectrode recordings from awake human subjects (microneurography) have greatly advanced our knowledge of the physiology of LTMs (Vallbo and Hagbarth, 1968). They have led to the identification of four distinct classes of Ab‐LTM aVerents in the glabrous skin of the human hand (Macefield, 2005) that strictly correlate to the four types of mechanoreceptors encountered in this type of skin (Meissner’s and Pacinian corpuscles, RuYni’s endings, and Merkel cells). Among the four types of LTM aVerents, a further subclassification can be established according to their rate of adaptation. Two types of LTM fibers fire only at the onset of a sustained stimulus and are therefore called fast or rapid adapting (FAI or RA1 and FAII or RA11) while two other types fire throughout the stimulus and are referred to as slowly adapting (SAI and SAII; Fig. 2). FAI and FAII aVerents are connected to Meissner’s and Pacinian corpuscles, respectively, while SAI and SAII fibers innervate Merkel cells and RuYni’s endings, respectively. Type I aVerents have small
433
15. Touch
Receptive field size Small
Large
Adaptation
RA-Meissner corpuscles/ G hair PC-Pacinian corpuscle
Fast
Stimulus response SAI-Merkel cell
Slow
SAII-Ruffini corpuscle
Stimulus response
FIGURE 2 Response properties of low‐threshold cutaneous mechanoreceptors. The stimulus trace represents a mechanical displacement. The response trace shows the pattern of action potentials (vertical lines) evoked by such a stimulus.
receptive fields that contain several ‘‘hot spots’’ of highest sensitivity corresponding to the individual Meissner’s corpuscles and Merkel cells. In contrast, type II aVerents display only one zone of maximal sensitivity in receptor fields which are large and unevenly shaped. Therefore, FAI aVerents (Meissner) respond to light mechanical stimulation in a very delimited area whereas FAII (Pacinian) can sense brisk and transient stimuli applied remotely from the area of optimal sensitivity. The same applies for SA aVerents as SAI fibers (Merkel) are sensitive to stimuli applied to a discrete zone of the skin while SAII (RuYni) can be activated by lateral stretch of the skin. SAI fibers can be further diVerentiated from SAII aVerents by their interspike intervals during their sustained firing not being constant, as opposed to the very regular pattern of interspike intervals observed in SAII receptors (Olausson et al., 2000). In human arm hairy skin, five types of large myelinated aVerents are encountered. SAI and SAII aVerents are the same as those in glabrous skin but three types of aVerents make up the RA versant of the hairy skin (Macefield, 2005). These are the hair (also called G1 hairs) and field (or G2 hairs) units (Leem et al., 1993; Lewin and Moshourab, 2004) and Pacinian corpuscles. Hair units (G1) respond to the movements of hairs and are linked to receptor fields that can be either well defined or irregular but that always contain several ‘‘hot spots’’ corresponding each to an individual hair. On the other hand, field units (G2) are sensitive only to actual skin contact and are connected to receptor fields similar to those of G1 receptors except that their ‘‘hot spots’’ are larger and less isolated.
434
Drew et al.
It should be noted that the distribution of types of LTMs varies between diVerent species and also between diVerent parts of the body. Hence, it has been shown that the hairy skin of the human face was supplied by the same kind of SA and RA fibers as those found in the glabrous skin of the hand (Macefield, 2005). The D‐hair (Brown and Iggo, 1967), connected to large receptor fields, is the only known example of an LTM aVerent belonging to the thinly myelinated class of Ad fibers and most surprisingly constitutes the most sensitive form of LTMs, being able to respond to hair movements as small as 1 mm in the mouse (Lewin and Moshourab, 2004). It is considered a subclass of hair follicle receptor (Stucky et al., 1998). Finally, LTMs can also be found on unmyelinated C fiber aVerents. The expression of these tactile C aVerents or low‐threshold C fibers (C‐LT) is highly variable from one species to another but is high enough in humans to be considered as an important part of the touch receptor machinery responsive to caress‐like and skin‐to‐skin contact between individuals that lead to pleasant sensations (Olausson et al., 2002). B. High‐Threshold Mechanoreceptors High‐threshold mechanoreceptors (HTM) are the receptors for noxious mechanosensation. These receptors are predominantly found on two types of sensory nerve fibers that respond also to temperature variations, namely the thinly myelinated Ad and unmyelinated C fibers (see above and Fig. 2). Ad fibers functioning as HTM encode the degree of skin indentation, firing throughout the duration of the stimulus, with little response to the moving part of the stimulus (Garell et al., 1996) and are believed to be the receptors for fast mechanical pain. Structurally, they have free unmyelinated nerve terminals projecting to the deep epidermis where they associate with keratinocytes and Schwann cells and possess large receptor fields displaying ‘‘hot spots’’ of higher sensitivity (Lewin and Moshourab, 2004). C fibers represent the predominant group of sensory nerve fibers innervating the skin. Most C fibers are polymodal receptors capable of responding to diVerent stimuli, notably noxious mechanosensation and temperature change. Nevertheless, some of the C fibers sensitive to noxious mechanosensation are unresponsive to thermal stimuli. Hence, can we distinguish among C fiber aVerents that respond both to mechanical and thermal stimuli— which are the C‐mechanoheat (C‐MH), the C‐mechanocold (C‐MC), and the C‐mechanoheatcold (C‐MHC) aVerents—and aVerents that respond
15. Touch
435
uniquely to painful mechanical stimulation, the C‐mechanonociceptors (C‐M) fibers (Lewin and Moshourab, 2004). It is not clear, however, whether there are any diVerences in mechanosensitivity between these diVerent groups of C fibers. In addition, ‘‘silent’’ or ‘‘sleeping’’ nociceptors do not respond to a mechanical or thermal stimulation in physiological conditions but can acquire thermo‐ and/or mechanosensitivity following inflammation (Schmidt et al., 1995). They account for a significant part of C fibers and are also expressed in viscera and joints. These fibers are termed C‐MiHi (Lewin and Moshourab, 2004).
V. QUANTITATING MECHANICAL RESPONSES IN ANIMAL MODELS The two most commonly used behavioral tests of mechanosensation are the von Frey test (von Frey, 1894) and the Randall–Selitto assay (Randall and Selitto, 1957). The von Frey test involves applying a punctate stimulus to a given region of the rodent’s body, usually the plantar surface of the hind paw, and recording the stimulus intensity that evokes a withdrawal reflex. Stimuli are typically applied using calibrated fibers with a specific bending force. The Randall‐Selitto test uses a device that applies an ascending pressure ramp to either the animal’s paw or tail and the point when a specific pain‐related behavior is evoked (e.g., vocalization or writhing) is used as the pain threshold. Typically, thresholds in the Randall–Selitto test are around an order of magnitude greater than those observed in the von Frey test suggesting that diVerent aspects of mechanosensory processing are being measured. This is further supported by observations of rodents that have had the majority of their nociceptors abolished by neonatal capsaicin treatment; these animals exhibit increased tolerance in the Randall–Selitto test but not in the von Frey test (Saumet and Duclaux, 1982; Nagy and van der Kooy, 1983; Shir and Seltzer, 1990). While experimentally induced changes in withdrawal thresholds in behavioral tests are consistent with aVects on the transduction machinery of sensory neurons, care must be taken in extrapolating from such observations as changes in the electrical excitability of the neurons, the central synaptic properties of the neurons, or the general motivational/motor state of the animal may aVect the outcome. It is reasonable to argue that the von Frey hair threshold is a measure of when the animal becomes aware of the presence of a mechanical stimulus that causes irritation, in contrast to the obvious pain evoked by a Randell–Sellito apparatus. However, there are no established tests for responses to light touch.
436
Drew et al.
VI. ELECTROPHYSIOLOGICAL APPROACHES TO MECHANOSENSATION IN RODENTS Mechanoreceptive neurons of the DRG synapse on to second order sensory neurons in the dorsal horn of the spinal cord. The activity of single spinal neurons (either superficial nociception‐specific or deeper wide dynamic range neurons) can be recorded in the intact animal while stimuli are applied to the receptive field of that cell (Matthews et al., 2006), thus giving a read out of the convergent input of multiple primary sensory neurons. For example, this approach has been used to show that administration of cannabinoid receptor agonists to the peripheral receptive field can inhibit responses to mechanical stimulation (Kelly et al., 2003; Elmes et al., 2004). In addition, it has been used to assess the thermal and mechanical sensitivity of null mutants with genes for sensory neuron ion channels ablated (Souslova et al., 2000). Such studies have revealed a separation of the encoding of thermal and mechanical input by the somatosensory system (Matthews et al., 2006). A number of groups have developed techniques for recording from either DRG somata or nerve fibers in more or less intact preparations. The skin‐ nerve preparation (Reeh, 1986) is a technique for recording from teased single‐nerve fibers of the saphenous nerve. The nerve is dissected out attached to an area of hind limb skin to which ‘‘natural’’ or electrical stimuli are applied. Using this technique, the firing properties of nerve fibers, classified by their conduction velocities and firing patterns to established subtypes of cutaneous aVerents, can be recorded in response to calibrated mechanical stimuli. This approach has been used to assess the mechanosensitivity of nerve fibers in null mutant mice (Price et al., 2000) and sensitization of nociceptive fibers to mechanical stimulation (Steen et al., 1995). Koerber and Woodbury (Woodbury et al., 2001) developed a related approach in which the thoracic and/or sciatic nerves are dissected out with their spinal cord and peripheral cutaneous connections intact, and recordings are made from the DRG somata. In addition, Sally Lawson’s group has extensively used in situ recordings from DRG neurons in anesthetized rodents and guinea pigs (Djouhri et al., 1998). These approaches have the advantage that the recorded neuron can be phenotyped by immunocytochemical labeling and its projections examined. Such approaches can be used to determine if genetic ablation of a candidate mechanotransducer eVects mechanically evoked firing in cutaneous receptors. However, observed firing rates will be determined by the eYciency of transduction and the electrical properties of the nerve fiber, while it should be determined that the morphology of the peripheral terminals is unchanged. The electrical properties of the fibers can be tested directly (although
15. Touch
437
electrical stimulation of the skin may bypass the sensory terminal in exciting the nerve fiber) and terminal morphology can be observed with standard labeling procedures. Gary Lewin’s group, using the skin‐nerve preparation, has used a probe fixed to a linear stepping motor under computer control to apply defined mechanical stimuli to the receptive fields of neurons (Shin et al., 2003; Dubreuil et al., 2004), whereas other groups have used von Frey hairs to assess fibers mechanical thresholds (Albers et al., 2006) or broad classes of stimuli (e.g., brush, pinch, pinprick; Djouhri et al., 1998) to classify receptive properties.
VII. MECHANOSENSITIVE ION CHANNELS IN CULTURED SENSORY NEURONS Ideally, recordings would be made of mechanically evoked receptor potentials from the nerve terminal. Loewenstein and coworkers in the 1960s used Pacinian corpuscles with the aVerent nerve attached extracted from the cat mesentery to record such events (Loewenstein and Mendelson, 1965; Loewenstein and Skalak, 1966). Enzymatic removal of the corpuscle suggested that these structures act as a mechanical filter determining the dynamics of the mechanical stimulus that reach the mechanoreceptive nerve ending. Nonselective cation channels that conduct mainly sodium in physiological conditions mediate these receptor potentials. In addition, Katz (1950) and Ottoson and coworkers (Ottoson, 1964; Husmark and Ottoson, 1971), using extracellular electrodes, have characterized the receptor potentials of frog muscle spindles exposed to stretch. These responses are defined by an initial RA component followed by a static phase. Observations following the removal of sodium again suggested that this ion carries the majority of charge in normal Ringer’s solution (Husmark and Ottoson, 1971). Receptor potentials were also decreased by removal of calcium and potassium (Ottoson, 1964; Husmark and Ottoson, 1971); however, these data are diYcult to interpret due to the likely nonspecific aVects on membrane potential that such manipulations would cause. Hunt and Ottoson (1975, 1976) also achieved similar results using muscle spindles from the cat tail, although the ionic basis of transduction was not investigated. No one has obtained similar results in rodents and certainly not from the bare nerve endings of nociceptors. One approach to this problem has been developed by Brock et al. (1998) who used a suction electrode attached to the guinea pig cornea to record activity in the terminals of nociceptors, they showed that sensory terminals could generate TTX‐resistant action potentials in response to mechanical stimulation. Mechanical stimulation was achieved by ‘‘pushing the recording electrode gently against the corneal surface with a displacement
438
Drew et al.
of the micromanipulator’’; it remains to be seen if this stimulation technique can be used to quantify mechanosensitivity and applied to transgenic mice. More recently, a number of groups have used cultured sensory neurons to study the responses of these cells to mechanical stimulation. These studies have used Ca2þ imaging or electrophysiology as the read out of the neuron’s response. Such studies rely on the redistribution of ion channels expressed by neurons so that molecules normally localized at their peripheral terminals are found in the somatic cell membrane (Baccaglini and Hogan, 1983; Cesare and McNaughton, 1996; Reid et al., 2002). In studies using ratiometric Ca2þ imaging neuronal stimulation using a rounded micropipette (Sharma et al., 1995; Gotoh and Takahashi, 1999; Raybould et al., 1999; Gschossmann et al., 2000), fluid jet (Sullivan et al., 1997), and hypoosmolarity‐induced cell swelling (Viana et al., 2001) has been shown to induce an increase in cytosolic Ca2þ. In each case, the rise in Ca2þ was dependent on extracellular Ca2þ suggesting that it was mediated via a calcium permeable membrane channel rather than release from intracellular stores. One problem with such studies has been the time course of the observed responses. In each case, the rise in Ca2þ levels has been slow (occurring over many seconds or even minutes) and the return to baseline has consistently occurred considerably after the end of the stimulus. How these slow responses relate to the rapid encoding of mechanical stimulation by nerve endings in vivo is unclear. Viana et al. (2001) showed that under voltage clamp, the [Ca2þ]i elevation produced by hypotonic stimulation was accompanied by the development of an inward current and a conductance increase. The time course and amplitude of the [Ca2þ]i response to hypoosmotic stimulation showed a close correlation with electrophysiological properties of trigeminal neurons. Fast [Ca2þ]i responses were associated with short duration action potentials characteristic of Ab fibers. Subsequently, a role for GAP‐43 in sensing osmolarity changes was demonstrated. By recruiting PKC to the cell membrane, GAP‐43 can induce a rise in IP3 levels and intracellular calcium in response to hypoosmotic stimuli (Caprini et al., 2003). However, GAP‐43 is sparsely expressed in adult DRG neurons and this phenomenon is distinct from the influx of extracellular calcium in DRG cultures stimulated osmotically. Although some studies reported that calcium channel antagonists did not inhibit mechanically evoked responses (Sullivan et al., 1997; Gotoh and Takahashi, 1999; Gschossmann et al., 2000) as sensory neurons express a heterogeneous population of voltage‐gated calcium channels (Yusaf et al., 2001), it cannot be guaranteed that all of them will be blocked. Viana et al. (2001) showed that the calcium channel blocker Ni2þ does indeed inhibit swelling‐evoked responses. Calcium imaging studies have consistently reported that Gd3þ, a blocker of a range of mechanosensitive ion channels (Hamill and McBride, 1996), inhibited responses to mechanical stimulation. However, the eVective blocking
15. Touch
439
concentration ranged from 5 mM (Gotoh and Takahashi, 1999) to 250 mM (Gschossmann et al., 2000). Moreover, Gd3þ might also antagonize voltage‐ gated Ca2þ channels (Boland et al., 1991). Gschossmann et al. (2000) reported that amiloride (100 mM) and k‐opioid agonists reduced mechanically evoked Ca2þ increases while not aVecting increases induced by capsaicin. Drummond et al. (1998; using the stimulation method of Sullivan et al., 1997) reported that increases in intracellular Ca2þ were antagonized by 100 nM amiloride. Another variable that fluctuated widely between studies was the proportion of cells that responded, ranging from 25% (Raybould et al., 1999) to 93% (Gotoh and Takahashi, 1999), although this may simply reflect diVerences in the degree of pressure applied using diVerent stimulation protocols. The specificity of these mechanically evoked Ca2þ increases remains to be established through studies on populations of non‐sensory neurons. Electrophysiological studies of mechanotransduction by sensory neurons have used a number of stimulation protocols sometimes making comparison of the data diYcult. Cho et al. (2002) investigated the expression of stretch‐ activated ion channels expressed by sensory neurons by applying negative and positive pressure to membrane patches of neonatal DRG neurons. This extensive study characterized two classes of MS ion channels; low‐threshold (LT) channels (present in 26% patches) which had an activation threshold of around 10 to 20 mmHg and a P1/2 of 60.6 mmHg and high‐threshold (HT) channels (24% of patches) with a P1/2 of 83.1 mmHg and a threshold of >60 mmHg. The activity of both channel types increased with ascending pressure and both nonselectively conducted cations with significant Ca2þ permeability. HT channels displayed a relatively linear current–voltage (I–V) relationship whereas LT channels were outwardly rectifying. Channel activity in both cases was inhibited by disruption of the cytoskeleton using either cytochalasin D (a disruptor of actin polymerization) or colchicine (a microtubule disruptor) and also by patch excision, which presumably disturbs the normal membrane–cytoskeleton interaction. Gd3þ inhibited HT and LT channels but neither were sensitive to amiloride nor arachidonic acid (AA). Finally, PGE2, acting via a PKA‐dependent pathway, selectively potentiated the activity of HT ion channels, in part by decreasing their activation threshold. At the whole‐cell level, Cho et al. (2002) mechanically stimulated DRG neurons, in a method similar to that reported by Takahashi and Gotoh (2000), by applying positive pressure through the patch pipette. Both studies reported the activation of cationic currents in a subset of sensory neurons with diameters greater than 20 mm; Takahashi and Gotoh (2000), however, observed currents in a substantially higher proportion of cells (75%) than did Cho et al. (2002; around 25%). A substantial lag was apparent between the application of the stimulus and the activation of the ensuing current
440
Drew et al.
(ranging from 0.5 to 3 s). The reasons for this delay are unclear as the increase in pressure should rapidly result in membrane stretch and so the possibility of an intermediate biochemical gating mechanism cannot be excluded. McCarter et al. (1999) and ourselves (Drew et al., 2002, 2004; Di Castro et al., 2006) have taken the approach of stimulating voltage‐clamped neurons using a heat‐polished glass probe. In their initial characterization, McCarter et al. (1999) reported that such a stimulus evoked nonselective cationic currents that were inhibited by Gd3þ and high concentrations of, the amiloride analogue, benzamil. We used a piezoelectric crystal device to control a glass probe that is used to stimulate the center of the neuron’s cell body (Fig. 3). The mechanically activated (MA) currents we observe using this approach have characteristics consistent with those reported by McCarter et al. (1999) reported. MA currents using this stimulation protocol are typically observed in >90% of neurons, consistent with the prevalence of mechanoreceptive neurons in the DRG. Currents activate at a specific membrane displacement and using a series of incremental ( ¼ 2 mm) mechanical steps showed that currents are proportional in amplitude to the stimulus size. We hypothesized that if this were an appropriate stimulation technique then diVerent sensory neuron subpopulations would have diVerential responses to stimulation consistent with their presumptive in vivo functions. In two studies we found evidence to support this. First, we categorized neonatal rat neurons as presumptive nociceptors or LTMs based on their response to capsaicin. Nociceptors (i.e., capsaicin‐sensitive neurons) displayed smaller MA currents that activated at higher thresholds than those seen in capsaicin‐insensitive neurons (Drew et al., 2002). The majority of currents were initially RA with a persistent late phase, although in a subset of capsaicin‐insensitive neurons MA currents were observed that displayed SA kinetics. Among smaller neurons, it was observed that IB4‐negative neurons generated MA currents but that IB4‐positive neurons were largely refractory to mechanical stimulation. In a second study (Fig. 4), we classified adult mouse neurons according to size, action potential width, and capsaicin sensitivity (Drew et al., 2004). Large neurons with narrow action potentials (i.e., LTM neurons) almost exclusively displayed rapidly MA currents that were larger than and kinetically distinct from MA currents in nociceptive neurons (wide action potentials, large or small cell body). Nociceptors (Fig. 5) displayed either SA responses or, what we termed, intermediately adapting currents that were clearly distinguishable from RA and SA currents. Action potential duration, a reliable indicator of a neuron’s receptive properties in vivo (Djouhri et al., 1998), was the strongest predictor of MA current properties; capsaicin sensitivity was associated only with wide action potential neurons but itself was not indicative of mechanosensitivity. Although a number of neurons did not respond within the stimulus range used, a greater proportion
A
Stimulating probe
Recording electrode
Neuron
Stimulus (mm)
C
12 10 8 6 4 2 0
B
0
100
200
300
Current (nA)
0
−1
−2
−3 0
100
200 Time (ms)
300
FIGURE 3 Technical setup and recordings. Mechanical stimulation of neuronal cell bodies was achieved using a heat‐polished glass probe controlled by a piezocrystal drive. (A) Schematic diagram of the experimental setup and (B) Photo of a neuron during the application of a 6‐mm stimulus. The recording electrode contacts the neuron on the top left and the stimulating probe comes from the top right of the picture to the center of the neuron. (C) Mechanical stimulation of cultured DRG neurons evokes graded cationic currents.
442
Drew et al.
A
B 1.00 Neurons (%)
100 mV/ms
2 ms
20 mV
2 ms 50 mV/ms
0.75 0.5 0.25 0.00
APs > 1 ms
APs < 1 ms
No res RA IA SA C
D
MA current (nA)
2.0
APs < 1 ms APs > 1 ms
1.5 1.0 0.5 0.0 6
8
10 12 Stimulus (mm)
14
16
FIGURE 4 Action potentials and mechanically activated currents of large, wild‐type mouse DRG neurons. (A) Examples of narrow (top) and wide (bottom) action potentials of large DRG neurons. Action potential traces are shown on the left and the diVerentials of these waveforms, which allow inflections to be more easily observed, are on the right. (B) Frequency histograms indicating the proportion of neurons with narrow and wide action potentials that respond to mechanical stimulation with rapidly adapting (RA), slowly adapting (SA), intermediately adapting (IA), or no (No res) currents. (C) Example traces of MA currents. Left: RA current
15. Touch
443
responded if the intensity was increased (unpublished data, LJD); no clear diVerences between IB4‐positive and IB4‐negative neurons were seen in adult mouse neurons. Interestingly, compression of the neurites of cultured sensory using a glass probe also evokes MA currents that are kinetically distinct and apparently associated with diVerent cell types (Hu and Lewin, 2006). Overall, these data confirmed a separation of MA current characteristics with neuronal phenotype consistent with the in vivo function of these cells. The MA currents we observe are mediated by nonselective cation channels permeant to calcium and magnesium that reverse at around 0 mV in quasi‐ physiological solutions and have a relatively linear I–V relationship. Replacement of external sodium with the impermeant organic monovalent cation N‐methyl‐D‐glucamine reduced current amplitude to a significantly greater degree in nociceptors (by 80%) than in capsaicin insensitive neurons (by 68%). Currents in both cell types are inhibited by external Ca2þ, in a manner compatible with this cation acting as a permeant blocker, but to a markedly higher degree in non‐nociceptors. This would suggest diVerent channels are operational in the diVerent cell types. To test this further, we looked at current inhibition by low micromolar levels of Gd3þ and ruthenium red in diVerent neuronal populations but the level of block by these compounds is indistinguishable between current types. We have found that FM1‐43 (Gale et al., 2001) acts as a permeant blocker of MS ion channels with greater potency at SA currents. These experiments therefore provide some evidence that molecularly distinct ion channels may transduce mechanical stimuli in diVerent DRG neurons. When we looked for signaling molecules that regulate MA currents, we found that nerve growth factor (NGF) and activation of PKC (but not PKA) acting through distinct mechanisms both increased mechanical responsiveness (Di Castro et al., 2006). Working with neonatal and adolescent rat neurons, it was found that MA currents were potentiated by PKC activation selectively in IB4‐negative nociceptors with a relatively rapid time course and in a tetanus toxin‐sensitive manner, suggesting that PKC activation induced the insertion of extra MS channels into the cell membrane. NGF, again only acting in IB4‐ negative neurons consistent with an action on TrkA receptors, increased MA current amplitudes but with a slower time course. Inhibitors of mRNA transcription and protein translation blocked this action suggesting that NGF induces the synthesis of new MS ion channels. Again in this study IB4‐positive neurons were essentially insensitive to mechanical stimulation and sensitivity was not increased by PKC activation or application of either NGF or GDNF. from a narrow action potential neuron. Right: intermediately adapting current from a neuron with a wide action potential. (D) Relationship between stimulus size and MA current amplitude in neurons with narrow and wide action potentials.
A
MA current amplitude (nA)
B
Neurons (%)
1.00 0.75 0.50 0.25 0.00
IB4-negative IB4-positive
0.5 0.4 0.3 IB4-negative
0.2 0.1
IB4-positive
0.0 4
No Res RA IA SA C
Rapidly adapting
50 pA
Intermediately adapting
12
Slowly adapting
100 pA
200 pA
IB4-negative
6 8 10 Stimulus (mm)
IB4-negative + IB4-positive IB4-negative 100 ms
15. Touch
445
Overall, these studies indicate that in cultured sensory neurons a number of diVerent types of mechanical stimulation are capable of causing either an increase in intracellular Ca2þ or a cationic current, that is excitatory responses consistent with what occurs at the sensory terminal. An overview of these investigations, however, suggests that in many cases the transduction process activated by each stimulus type is distinct, and therefore it must be asked how many of these mechanisms are normally active in the sensory terminal in response to physiologically relevant mechanical events. It remains to be determined if the ion channels activated by cell swelling through positive pipette pressure correspond to those gated by compression of the cell by a glass probe. Both evoke a nonselective cationic current but the long activation latency associated with the former stimulus is in stark contrast to the submillisecond delay in gating by an external probe. Furthermore, in the Cho et al. (2002) study, this form of pipette pressure activated currents in only around 30% of neurons and neurons with diameters less than 30 mm, that is those most likely to be LTMs did not respond. Similarly, both classes of stretch‐activated ion channels recorded at the single‐channel level were absent from neurons over 25 mm in diameter arguing against a role in light touch sensation. In contrast, currents evoked by compression of the somatic cell membrane are largest and have the lowest threshold in large neurons, likely derived from LTMs. The initial rapid adaptation of these responses followed by a sustained component is similar to the kinetics of the receptor potential at cat muscle spindles (Hunt and Ottoson, 1975, 1976). RA currents (where MS ion channels close soon after opening) will encode both the magnitude and velocity of a mechanical stimulus; this latter aspect of a stimulus being important in both touch and in monitoring muscle position (Hunt and Ottoson, 1975, 1976). In addition, MA currents are mediated by nonselective cation channels (like all other mechanically evoked currents in DRG neurons) consistent with the limited data available on the ionic basis of transduction at the sensory terminal. Finally, MA currents evoked by membrane compression activate with very short latencies and also inactivate very rapidly when the stimulus is withdrawn; such rapidity is expected of a transduction mechanism that gives accurate information on mechanical stimuli and can encode high frequency vibrations. In contrast, LT ion channels (Cho et al., 2002) remained active for a considerable time (>5 min) after the cessation of the stimulus.
FIGURE 5 MA currents exhibited by wild‐type small to medium mouse DRG neurons. (A) Frequency histograms for responses of IB4‐negative and IB4‐positive neurons; responses were of four types‐slowly (SA), rapidly (RA), or intermediately (IA) adapting currents, or no response (No res). (B) Stimulus‐response relationships for pooled data from IA and SA currents (IB4‐negative neurons) and IA currents (IB4‐positive neurons). (C) Example traces of RA, IA, and SA currents.
446
Drew et al.
On balance, the ion channels underlying MA currents evoked by external compression of the membrane represent the strongest candidates for being mechanotransducers in mammalian sensory neurons.
VIII. GATING MS ION CHANNELS IN DRG NEURONS Estimating the force required to gate mechanosensitive ion channels is diYcult especially for whole‐cell stimuli. For example, increases in membrane tension in response to hypotonicity is distinct in diVerent types of neurons depending on their membrane reserves and the degree of membrane insertion and retrieval during volume changes (Zhang and Bourque, 2003). Localized stimulation of neuronal somata with a glass probe will result in a focused area of membrane stretch surrounded by a concentric pressure gradient; as the stimulus size is increased, the area of neuronal membrane stretched above resting tension will expand and the membrane stretch at any point within that area will increase. Therefore, the population of MS ion channels in the membrane is exposed to a range of tensions and increasing the stimulus intensity will increase current amplitude due to activation of more channels (i.e., the area of suprathreshold membrane tension increases) and potentially by aVecting the behavior (e.g., slowing the rate of closing/increasing the rate of reopening) of channels exposed to higher tensions (Drew et al., 2004). [Cho et al. (2002) found that increasing membrane tension increased channel activity primarily by a reduction in the duration of long closings.] These concerns make kinetic analysis of channel opening and closing using such protocols limited; in addition, the relatively slow movement of the probe means that the area of suYcient stretch develops over several milliseconds. Finally, if sensory neurons diVer in their amount of membrane reserve and degree of crenellation, then the degree of membrane stretch for a given displacement will vary. The use of membrane patches allows more accurate prediction of the tension reaching the channel under observation (Cho et al., 2002). Indeed, high‐speed, pressure‐clamp devices have been developed for the study of ion channels in membrane patches (Besch et al., 2002), which allow for rapid actuation of membrane stretch that is important given the rapid activation and adaptation/inactivation of many MS ion channels. Current models of the gating of mechanosensitive ion channels suggests that they are activated either by the direct sensing of membrane tension, as is the case for bacterial MS channels, or due to tethering to intracellular and/or extracellular structures that the channel moves relative to (Kung, 2005). Direct gating by membrane tension can be demonstrated when a known
15. Touch
447
protein is reconstituted in a lipid bilayer and gated by stretch. Unequivocal evidence of the second mechanism is more diYcult because if gating is dependent on cytoskeletal elements it remains possible that these structures exert their eVects indirectly by interactions with the cell membrane. MA currents evoked by external membrane compression (Drew et al., 2002) are inhibited by cytochalasin whereas stretch‐activated ion channels (Cho et al., 2002) in DRG neurons are reduced by both cytochalasin and colchicine. These data suggest that the activity of these channels may be reliant on cytoskeletal anchoring. However, actin disruption could exert its eVects not by disturbing an interaction between the ion channel and an intracellular anchor but by reducing membrane tension in the vicinity of the channel by abolishing the normal interaction between the cortical cytoskeleton and the plasmalemma. The behavior of MS channels in their nonnative environment (e.g., in the cell soma or on the neurite) may be diVerent from that at the sensory terminal if the resting membrane tension adjacent to the channel varies. It is not known if these specialized mechanosensory regions have local areas of prestressed membrane that could aVect channel gating. If membrane tension is the key gating factor, then channel activation would be due to both the intrinsic sensitivity of the channel and the resting membrane tension determined by the cytoarchitecture of the terminal. These two factors will have coevolved for optimal activation of the ion channel in response to relevant stimuli and could mean that diVerent membrane arrangements at subtypes of mechanoreceptor endings aVect channel behavior. In addition to cytoskeleton–membrane interactions, the lipid composition of the membrane may also be important. For example, the cholesterol content of the membrane, particularly in microdomains, will aVect local membrane stiVness (Lundbaek et al., 2004). In addition, PIP2 in the cell membrane could be a key molecule in coupling MS channels to membrane stretch; Chemin et al. (2005) showed that high levels of PIP2 can gate the mechanosensitive ion channel TREK‐1 and that this molecule appears to be necessary for mechanogating. Molecular cloning of the DRG mechanotransduction channel will facilitate the answering of these questions.
IX. CANDIDATE ION CHANNELS Studies of invertebrate mechanosensation using screens of mechanosensitive mutants in Caenorhabditis elegans and Drosophila melanogaster (Ernstrom and Chalfie, 2002; Tracey et al., 2003) have focused attention on two classes of ion channels as potential mechanotransducers in mammals: the DEG/ENaC (degenerin/epithelial sodium channels) and TRP superfamilies.
448
Drew et al.
A. DEG/ENaC Ion Channels 1. DEG/ENaC Ion Channels in C. Elegans Mechanosensation The current model of mechanotransduction in C. elegans is that an ion channel lies at the center of a multiprotein complex tethered to intracellular and extracellular structures. The ion channel is made up of two subunits, MEC‐4 and MEC‐10, of the DEG/ENaC family. These channels physically interact with MEC‐2, a stomatin‐like protein (Goodman et al., 2002), and MEC‐6, a single membrane pass protein, with a low homology to human paraoxonases (Chelur et al., 2002). The functional ion channel is an amiloride‐sensitive sodium channel. Evidence that this ion channel complex directly transduces mechanical stimuli came from Suzuki et al. (2003a) using a transgenic nematode line that expresses a cameleon Ca2þ‐indicator protein. Low levels of mechanical stimulation evoked Ca2þ transients in wild‐type touch receptor cells but not in those from mec‐4, mec‐6, and mec‐2 loss‐of‐function mutants. Subsequently, O’Hagan et al. (2005) definitively made direct, in situ recordings of mechanically evoked receptor currents in patch‐clamped touch receptor cells. These neurons generated rapidly activating and inactivating inward currents at the application and the withdrawal of a mechanical stimulus. Mechanically activated currents were absent in nematodes with loss‐of‐function mutations in mec‐2, mec‐4, and mec‐6. Interestingly, mec‐7 mutants displayed currents significantly smaller than those in wild‐type animals, suggesting this microtubule is required for normal mechanosensitivity, but not for channel gating. Mechanosensory functions have been proposed for two other nematode DEG channels, both of which are in the same subgroup as the MEC channels and have a conserved extracellular regulatory domain (Goodman and Schwarz, 2003). They are UNC‐105 (Liu et al., 1996), which is expressed in muscle cells, and UNC‐8 (Tavernarakis et al., 1997), which is expressed in sensory neurons, motor neurons, and interneurons. In Drosophila, Adams et al. (1998) identified PPK1, a DEG/ENaC homologue, expressed in the sensory dendrites of type II sensory receptors. When Ainsley et al. (2003) ablated the gene for PPK1, they found that mutants had normal larval touch sensitivity but showed locomotor abnormalities possibly due to a reduction in mechanosensory feedback during movement. 2. ASICs and Mammalian Mechanosensation Following the extensive work defining the molecular components of the mechanotransduction complex in the touch receptors of C. elegans, the structurally related ENaCs and ASICs appeared to be plausible candidates for mammalian mechanotransducing channels. In particular, ASICs (acid sensing ion channels, so named as most members of this subfamily are activated by rapid drops in pH) were seen as strong candidate mechanosensors because
15. Touch
449
they are highly expressed in sensory neurons and two isoforms, ASIC3 and ASIC1b, are almost exclusively found in these cells. However, the evidence that these channels mediate mechanotransduction in mammalian sensory neuron is unconvincing. The mechanically evoked firing patterns of cutaneous aVerent fibers in ASIC1 (Page et al., 2004), ASIC2 (Price et al., 2000), and ASIC3 null mutants (Price et al., 2001) have all been characterized using the skin‐nerve preparation. Overall, these analyses have shown very subtle diVerences between knockout animals and wild‐type controls. The behavior of cutaneous mechanoreceptors in mice lacking the ASIC1 gene was entirely normal. In ASIC2 nulls, Ab‐fibers had reduced suprathreshold responses to mechanical stimuli; the decrease was small in SA LTMs, but in RA LTMs firing was reduced by around 50%. Conversely, in ASIC3 nulls, RA LTMs displayed increased rates of firing in response to mechanical stimulation whereas there was a reduction in the mechanosensitivity of Ad‐nociceptors. While the phenotypes of ASIC2 and ASIC3 nulls could be consistent with a role for these channels in transduction, it is diYcult to reconcile the broad expression of ASICs in the DRG with small changes in subpopulations of receptors. The voltage insensitivity of ASICs and the apparently normal membrane properties of ASIC2 and ASIC3 knockout (KO) neurons (Price et al., 2000; Drew et al., 2004) suggest such phenotypes are not due to decreased nerve excitability. However, the firing rates of fibers in response to electrical stimulation were not reported and ASIC3 mutants also showed reduced firing in response to heat stimuli. Additionally, a number of other papers have also cast doubt on the likelihood of ASICs transducing mechanical stimuli in sensory neurons. Roza et al. (2004), using a diVerent strain of ASIC2 null mice to Price et al. (2000), failed to find a significant diVerence in mechanically evoked firing patterns in RA LTMs and this group also found indicators of auditory and intestinal mechanosensation to be unchanged in these animals. Additionally, transgenic mice expressing a dominant‐negative form of ASIC3, which knocked down expression of any functional ASICs (as assessed by application of low‐pH stimuli), had normal behavioral responses to mechanical stimuli (Mogil et al., 2005). In a study of cultured DRG neurons from ASIC2 and ASIC3 KOs (Drew et al., 2004), MA currents evoked in neurons derived from LTMs (based on the generation of narrow action potentials) were found to be normal in both single KOs and cells lacking both ASIC2 and ASIC3. In addition, MA currents in nociceptive neurons were normal in the double KOs. The stretch‐ activated ion channels observed by Cho et al. (2002) in DRG neurons were insensitive to amiloride which blocks ASICs. These data together demonstrate that neither of these forms of mechanical stimulation activate ASICs in sensory neurons and no data has been published showing direct gating
450
Drew et al.
of ASICs by pressure. While the caveat exists that ASICs would only be mechanosensitive when expressed in the correct environment at the nerve terminal, a number of evolutionary observations argue against them functioning in mechanosensation. First, the phylogenetic tree of structurally related ENaC/ASIC/DEG ion channels reveals a substantial diversification of these molecules in C. elegans and D. melanogaster that is not apparent in vertebrates and that those channels most strongly implicated in nematode mechanotransduction form a small subgroup that is distantly related to mammalian ASICs (Goodman and Schwarz, 2003). Second, ASIC homologues are present in zebrafish and are expressed across the nervous system but are absent from sensory neurons negating a role in sensory transduction (Paukert et al., 2004). Furthermore, this group showed that proton sensitivity of ASICs arose recently in evolution (Coric et al., 2005). This observation and uncertainty that pH falls substantially and rapidly enough to activate these channels in mammals means that the function of ASICs in both the peripheral and central nervous systems remains enigmatic. However, on balance, the available evidence suggests they are not mechanotransducing channels.
B. TRP Ion Channels 1. TRP Channels in Invertebrates Genetic screens of invertebrates have identified a number of members of the TRP ion channel family as either mechanotransducers or essential for the function of mechanosensory cells (Lin and Corey, 2005). OSM‐9 (Colbert et al., 1997) and OCR‐2 (Tobin et al., 2002) appear to form a heteromeric ion channel expressed in the sensory processes of the ciliated ASH neurons in C. elegans that is required for the detection of nose touch by these cells. This channel appears to be polymodal in function, responding also to osmotic and chemical stimuli, and it remains to be determined if the channel is directly gated by each stimulus class or acts downstream of the true transducer. Also in the nematode, TRP‐4 has been identified as the likely mechanosensory channel of a proprioceptive neuron, DVA (Li et al., 2006). TRP‐4 mutants exhibit abnormal movement and Ca2þ transients evoked by body bending in wild‐type DVA neurons are absent in these animals. TRP‐4 is the C. elegans homologue of TRPN1, which was originally identified (as NompC) by Walker et al. (2000) as a candidate mechanotransduction channel in bristle receptors of Drosophila. Mutations causing premature stop codons in nompC led to a loss of all but a small nonadapting part of the mechanoreceptor potential recorded extracellularly from the bristle. Additionally, one mutant (cysteine to tyrosine at residue 1400, close
15. Touch
451
to the fourth predicted transmembrane domain) was isolated that increased significantly the rate of adaptation of mechanically evoked potentials, providing strong evidence that this channel is itself mechanically gated. There is no close mammalian homologue of TRPN1 but, surprisingly, Sidi et al. (2003) found that zebrafish selectively express an orthologue of TRPN1 in hair cells and that downregulation of its expression using morpholino‐ antisense oligonucleotides abolishes extracellularly recorded microphonic potentials. In addition, an orthologue of TRPN1 is found in Xenopus, where it is expressed in structures with mechanosensory functions, although its precise distribution was not consistent with it being the transduction channel itself (Shin et al., 2005). Drosophila nompC mutants have only moderate deficits in auditory responses (Eberl et al., 2000). In Drosophila, responses to auditory stimuli are detected by structures analogous to bristle receptors in Johnston’s organ, within the antennae of the fly. Transduction here is likely mediated by a heteromeric complex of two TRP channel subunits: Nanchung (NAN; Kim et al., 2003) and Inactive (IAV; Gong et al., 2004). NAN is related to OSM‐9 whereas IAV is more closely related to OCR‐4. In flies lacking functional genes for either channel, sound‐evoked antennal aVerent nerve activity is absent. These observations along with the distribution of these channels make them very strong candidates for being mechanogated transduction channels. Finally, in Drosophila, the TRP ion channel, painless, has been implicated in noxious mechanosensation by Tracey et al. (2003). This group developed a genetic screen for studying nocifensive behavior in Drosophila larvae and showed that animals lacking functional painless expression showed defective behavioral responses to noxious temperatures and noxious pressure. Painless is expressed in a discrete punctate fashion on the dendrites of putative nociceptors, and the sensory deficits of mutants suggested that painless functions as a transducer of noxious stimuli in multiple modalities. However, channel activation by physical stimuli was not demonstrated and so it is possible that the protein acts up‐ or downstream of transduction. 2. TRP Candidates in Mammals Several mammalian TRP channels have been described as being mechanically activated. Among them TRPV4, the closest mammalian homologue of OSM‐9, seems to be a very attractive candidate as an HTM as it appears to be activated by both osmotic and mechanical stimuli (Liedtke and Kim, 2005) and is expressed in some DRG neurons from both LTM and HTM groups (Alessandri-Haber et al., 2003; Suzuki et al., 2003b). Studies have shown that
452
Drew et al.
TRPV4 knockout animals display altered osmotic regulation and pressure sensation (Liedtke and Friedman, 2003; Mizuno et al., 2003; Suzuki et al., 2003c) while other works have described the involvement of TRPV4 in nociception in response to hyper‐ and hypoosmotic stimuli and in inflammation‐induced mechanical hyperalgesia (Alessandri-Haber et al., 2003, 2005, 2006). Nevertheless, it is important to keep in mind that if TRPV4 is eVectively implicated in mechanosensation, its activation by mechanical and/or osmotic stimuli seems to be indirect and requires the synthesis of 50 ,60 ‐ epoxyeicosatrienoic acid [50 ,60 ‐EET], a metabolite of AA (Vriens et al., 2004). This suggests an activating mechanism for TRPV4 whereby a membrane mechanical sensor is coupled to the phospholipase A2 (PLA2) which in turn metabolizes membrane phospholipids to synthesize metabolites that in the end lead to the opening of TRPV4. Other members of the TRPV subfamily have been implicated in osmoregulation and mechanosensation processes. Birder et al. (2002) demonstrated that TRPV1 was necessary for normal bladder function in that it is essential to the purinergic release triggered by the mechanical distension of the bladder. An N‐terminal splice variant of TRPV1 has been found to be expressed in the osmosensitive arginine‐vasopressin‐ releasing neurons of the supraoptic nucleus, where it is proposed to be part of the central osmoreceptor (Sharif Naeini et al., 2006). Finally, there is evidence that TRPV2 can be activated by an osmotic challenge in mouse aortic myocytes and can also be activated by membrane stretch when expressed in heterologous systems (Muraki et al., 2003). Bearing in mind that TRPV1, TRPV2, and TRPV4 are all expressed in DRG neurons, all three channels may participate in mechanosensation. Another candidate TRP channel is TRPA1, the mammalian homologue of the fly gene painless (Chapter 8, this volume). Corey et al. showed in 2004 that TRPA1 was expressed in the hair cells of the inner ear and that the mRNA for TRPA1 appeared at embryonic day E17, exactly matching the onset of mechanosensitivity in these cells. Furthermore, they also showed that the downregulation of the TRPA1 protein induced an alteration of the receptor cell function. This was seen as a major breakthrough as TRPA1 is known to be expressed in DRG neurons (although there is a debate over the proportion of neurons expressing it), where it is believed to act as a noxious cold sensor (Story et al., 2003). But two recent studies using TRPA1 knockout mice oppose the assumption that TRPA1 might be the mechanosensor in both hair cells and DRG neurons. Bautista et al., (2006) could not find any deficit in cold sensation and sound detection in these animals. On the other hand, Kwan et al. (2006) observed alterations in cold sensitivity but not in hearing and balance functions. Nevertheless, altered responses to punctate mechanical stimuli were reported, suggesting that TRPA1 might be part of a broader mechanosensitive complex in DRG neurons. Hence, the channel or
15. Touch
453
channel complex acting as the mechanotransducer of physiological and/or noxious tactile stimuli remains to be discovered. Other mechanically activated mammalian TRP subunits might be involved but their expression in DRG neurons has yet to be determined. Among these channels are the polycystins TRPP2 and TRPC1. TRPP2 is a cation‐permeable channel involved in the autosomal dominantly inherited polycystic kidney disease in which mutations in either the TRPP1 or TRPP2 gene cause the occurrence of cysts in the kidney and the liver (Delmas, 2005). TRPP2, in association with TRPP1 which is not a channel protein, form functional membrane channels tightly associated to actin filaments (Montalbetti et al., 2005) and located to the primary cilia of kidney epithelial cells where they function as calcium permeable channels opening in response to fluid flow (Nauli et al., 2003). TRP2 expression is widespread including high levels in the heart (Volk et al., 2003) as well as in the embryonic nodal cilia that determine the left–right body axis (McGrath et al., 2003). It is not known yet whether TRPP2 is directly activated by a mechanical stimulus but it has been already demonstrated that TRPP1 is able to activate TRPP2 when these two channels are associated (Delmas et al., 2004), suggesting that the mechanosensitivity of TRPP2 may depend on TRPP1. Nevertheless, the functionality of TRPP2 alone in responding to mechanical stimuli has also been shown in epithelial cells (Montalbetti et al., 2005). TRPC1 has also been shown to be mechanically activated (Maroto et al., 2005). In this study, the authors sought to describe the molecular identity of the mechanosensitive cation channel (MscCa) located in cytoskeleton‐deficient membrane vesicles of the Xenopus oocytes. When the human TRPC1 was expressed in CHO‐K1 cells, it also showed a mechanoresponsive behavior. These data demonstrated first that TRPC1 was mechanically activated but also that it did not need any connection to the cytoskeleton to be functional as opposed to TRPP2. Two more mammalian TRP subunits may be considered as mechanosensor candidates. TRPML3 is expressed in hair cells and was shown to be the gene responsible for the semidominant mouse mutant varitint‐waddler which displays early‐onset hearing loss and vestibular defects (Di Palma et al., 2002). Finally, another kidney‐located TRP channel, TRPM3, was shown to be activated by a decrease in osmolarity causing cell swelling (Grimm et al., 2003).
C. Mechanosensitive Potassium Channels The two‐pore potassium channels TASK and TREK‐1 are known to be mechanically gated (Patel et al., 2001). TREK‐1 is present in small diameter sensory neurons that are usually assumed to be nociceptors. TREK‐1 is a
454
Drew et al.
polymodal ion channel that is activated by lipids, membrane stretch, G‐protein–subunits and heat. In addition, the volatile anesthetics nitrous oxide, xenon, and cyclopropane (that may act in part through inhibition of NMDA receptors) have been shown to be potent activators of TREK‐1 at clinically relevant concentrations (Gruss et al., 2004). Interestingly, these activators require the presence of a particular amino acid (Glu‐306) that also has been implicated in channel activation by AA and membrane stretch. A study of knockout mice showed a pronounced pain phenotype (Alloui et al., 2006). Mice with a disrupted TREK‐1 gene were more sensitive to painful heat and mechanical stimuli (albeit low‐intensity stimuli in the von Frey test and not in the Randall–Selitto test) and showed enhanced hyperalgesia in conditions of inflammation, demonstrating a role for TREK‐1 in inhibiting noxious input into the CNS. However, less obviously, osmotic stress in inflamed knockout animals resulted in a lowered pain phenotype. These observations suggest that TREK‐1 may be physiologically activated by polymodal noxious stimuli and shape the form of the receptor potential. If this is the case then TREK‐1 would act as an excitability brake, although it remains to be determined if general baseline neuronal excitability is decreased in these animals.
X. VOLTAGE‐GATED CHANNELS AND MECHANOSENSATION Measuring mechanically gated channels in sensory neurons in voltage clamp has been very informative, but in normal circumstances, sufficiently large mechanically evoked depolarizations will alter the activity of voltage‐ dependent channels possibly resulting in action potential generation. Interestingly, a number of voltage‐gated channels (and some ligand‐gated channels) are directly influenced by mechanical stimuli. NMDA receptors, for example, have been shown to be modulated by changes in osmolarity being potentiated in hypoosmotic conditions and inhibited by external hyperosmotic solutions (Paoletti and Ascher, 1994).
A. Sodium Channels Of the nine voltage‐gated sodium channel a‐subunits (Nav1–9), three are strongly associated with expression in sensory neurons (Nav 1.7, Nav 1.8, and Nav 1.9), while most other subunits (apart from Nav1.4) are expressed to some extent in these cells. Interestingly, there is some evidence that mechanical stimuli can eVect the peak current density of Nav1.5 in intestinal cells, and this eVect depends on interactions of the C‐terminus of the sodium
15. Touch
455
channel with the PDZ domains of syntrophin, an actin‐binding protein (Ou et al., 2003). As voltage‐gated sodium channels propagate action potentials, alterations in their threshold of activation have profound eVects on peripheral pain thresholds, and may also be involved in the phenomenon of allodynia, where nonnoxious touch may be perceived as a painful stimulus (Wood et al., 2004). Posttranslational modifications of sodium channels by inflammatory mediators involving phosphorylation have been shown to underlie some of these eVects, for example in Nav1.8. This sodium channel is specifically associated with nociceptors, and its I–V relationship is shifted to more negative potentials and its peak current increased in response to PKA‐ mediated phosphorylation of serine residues in the second intracellular loop (Fitzgerald et al., 1999). Of all sodium channels, Nav1.7 seems to play the most significant role in altering inflammatory pain thresholds and by analogy noxious mechanical thresholds, although the underlying mechanism is unknown (Nassar et al., 2004). Nav1.7 is present at high densities at the terminals of nociceptive neurons (as well as sympathetic neurons) and its function has been addressed by the study of tissue‐specific knockout mice. When Nav1.7 is deleted in nociceptors, thermal pain thresholds and nonnoxious mechanosensation are apparently normal. Strikingly, noxious mechanosensation is almost completely abolished (Nassar et al., 2004). In the Nav1.8 null mutant, a similar phenotype is apparent (Akopian et al., 1999). Investigations of sensory processing in the spinal cord of Nav1.8 null mutants by Matthews et al. (2006) also revealed a selective deficit in mechanical over thermal input. This suggests that nerve fibers expressing Nav1.7 and Nav1.8 are necessary for noxious mechanosensation. In Nav1.8 null mutants, the mechanosensitivity of neuromas is also attenuated, consistent with this idea (Roza et al., 2004). Capsaicin killing of TRPV1‐positive neurons which include many nociceptors expressing Nav1.7 and Nav1.8 also results in a similar phenotype, without much eVect on thermal pain thresholds (Hayes and Tyers, 1980). This suggests that there is either a quantitative diVerence in sensory coding for thermal and mechanical stimuli, where far fewer functional nociceptors are required to convey information about noxious heat, or suggests that a diVerent set of fibers are involved in signaling the extent of noxious mechanical events from damaging levels of heat. An alternative explanation could be that the channel associated with noxious mechanosensation are expressed in a membrane domain at sensory nerve terminals close to Nav1.7 and Nav1.8. Uncoupling of the link between these sodium channels and mechanosensory channels would then lead to the specific loss of noxious mechanosensation. Identifying the proteins that interact with these sodium channels may thus provide us with clues to the identity of the noxious mechanotransducing channels.
456
Drew et al.
B. Calcium Channels An association between the expression of the voltage‐gated calcium channel Cav3.2 in NT‐4‐dependent D‐hair cells has been claimed recently, and pharmacological data accumulated to suggest this calcium channel may have a role to play in these highly mechanosensitive cells (Shin et al., 2003). The authors suggest that this low‐threshold calcium channel may be able to amplify small receptor potentials and account in part for the high sensitivity of D‐hair receptors. A broader role for these channels in sensory neuron function is suggested by antisense studies, which demonstrated a pronociceptive role for these channels in hyperalgesia and allodynia (Bourinet et al., 2005). Bouskila and Bostock (1998) showed direct eVects of mechanical stimulation on calcium currents in sensory neurons. N‐type calcium channel activity increased by over 70% when sensory neurons were subjected to a stream of buVer, while T‐type currents show a small decrease.
XI. INDIRECT SIGNALING BETWEEN SENSORY NEURONS AND NONNEURONAL CELLS Do the specialized end organs in which LTM aVerent fibers terminate transduce mechanical stimuli and signal to sensory neurons (Ogawa, 1996)? While the function of Pacinian corpuscles as mechanical filters (Bell et al., 1994) is well established, the possibility that Merkel cells transduce stimuli has been more contentious. Merkel cells have a number of attributes of a neurosecretory cell; they contain dense core vesicles close to the region apposing the nerve terminal (Haeberle et al., 2004). This led to speculation that Merkel cells respond to mechanical stimulation and chemically communicate sensory information to the nerve terminal. Toxic ablation of these Merkel cells has given conflicting data (probably due to both incomplete eradication of Merkel cells and/or secondary damage to the sensory nerves); however, when Mills and Diamond (1995) ablated Merkel cells with near UV light and carefully mapped the aVected touch domes, they showed that normally functioning SA1 fibers were present in the absence of Merkel cells. Moreover, Kinkelin et al. (1999) have shown that Merkel cells are almost entirely eliminated postnatally in p75 null mutant mice with no change in the physiology of SA mechanoreceptors. A role for ATP in mechanosensation has been suggested by a number of studies that show that this mediator may be released from distorted tissue and act on the terminals of sensory neurons (for example, Cockayne et al., 2000). ATP is released from Xenopus oocytes vesicular stores in response to mechanical stimulation in an exquisitely sensitive fashion (Maroto and
15. Touch
457
Hamill, 2001). Expression cloning in Xenopus oocytes lead to the identification of P2Y1 receptors as potential mechanosensors which were activated by low levels of ATP released from oocytes (Nakamura and Strittmatter, 1996). Interestingly, P2Y1 is expressed at high levels in large diameter sensory neurons. Cocultures of human keratinocytes with mouse DRG neurons have revealed that mechanical activation of the keratinocytes results in intracellular calcium waves that depend on the presence of extracellular ATP (Koizumi et al., 2004). A similar increase in calcium could be measured in the DRG neurons, and this depended on the presence of P2Y2 receptors, suggesting a role for keratinocyte‐released ATP in signaling to sensory neurons. However, direct application of ATP to toad skin preparations support a modulatory rather than excitatory role for ATP. At high concentrations, ATP suppressed impulse discharge from SA mechanoreceptors without eVect on RA mechanoreceptors (Fallon et al., 2002). Finally, a role for the AA metabolite 50 ,60 ‐epoxyeicosatrienoic acid (50 ,60 ‐ EET) in gating TRPV4 has been demonstrated (Vriens et al., 2004). As eicosanoids are membrane permeant, it is possible that external sources of this metabolite generated after the activation of PLA2 by tissue damage could activate TRPV4. XII. CONCLUSIONS Mechanically regulated electrical activity by touch and tissue damaging levels of pressure in sensory neurons seems to involve a variety of direct and indirect mechanisms and ion channels, and the involvement of specialized end organs in mechanotransduction complicates matters even more. Imaging studies are providing useful information about the events in the central nervous system associated with touch pain and allodynia (a pathological state where touch becomes painful; Naito and Ehrsson, 2006; Ruehle et al., 2006). In contrast, although a variety of TRP channels are potential candidate mechanosensors, and there is evidence of a role for TRPV4 in some aspects of mechanosensation, the channels underlying touch and noxious mechanosensation remain to be identified. Acknowledgments We thank the MRC, BBSRC, and Wellcome Trust for support, and Paolo Cesare for his many contributions to this work.
References Adams, C. M., Anderson, M. G., Motto, D. G., Price, M. P., Johnson, W. A., and Welsh, M. J. (1998). Ripped pocket and pickpocket, novel Drosophila DEG/ENaC subunits expressed in early development and in mechanosensory neurons. J. Cell Biol. 140, 143–152.
458
Drew et al.
Ainsley, J. A., Pettus, J. M., Bosenko, D., Gerstein, C. E., Zinkevich, N., Anderson, M. G., Adams, C. M., Welsh, M. J., and Johnson, W. A. (2003). Enhanced locomotion caused by loss of the Drosophila DEG/ENaC protein Pickpocket1. Curr. Biol. 13, 1557–1563. Akopian, A. N., Souslova, V., England, S., Okuse, K., Ogata, N., Ure, J., Smith, A., Kerr, B. J., McMahon, S. B., Boyce, S., Hill, R., Stanfa, L. C., et al. (1999). The tetrodotoxin‐ resistant sodium channel SNS has a specialized function in pain pathways. Nat. Neurosci. 2(6), 541–548. Albers, K. M., Woodbury, C. J., Ritter, A. M., Davis, B. M., and Koerber, H. R. (2006). Glial cell‐line‐derived neurotrophic factor expression in skin alters the mechanical sensitivity of cutaneous nociceptors. J. Neurosci. 26, 2981–2990. Alessandri‐Haber, N., Yeh, J. J., Boyd, A. E., Parada, C. A., Chen, X., Reichling, D. B., and Levine, J. D. (2003). Hypotonicity induces TRPV4‐mediated nociception in rat. Neuron 39, 497–511. Alessandri‐Haber, N., Joseph, E., Dina, O. A., Liedtke, W., and Levine, J. D. (2005). TRPV4 mediates pain‐related behavior induced by mild hypertonic stimuli in the presence of inflammatory mediator. Pain 118, 70–79. Alessandri‐Haber, N., Dina, O. A., Joseph, E. K., Reichling, D., and Levine, J. D. (2006). A transient receptor potential vanilloid 4‐dependent mechanism of hyperalgesia is engaged by concerted action of inflammatory mediators. J. Neurosci. 26, 3864–3874. Alloui, A., Zimmermann, K., Mamet, J., Duprat, F., Noe¨l, J., Chemin, J., Guy, J., Blondeau, N., Voilley, N., Rubat‐Coudert, C., Borsotto, M., Romey, G., et al. (2006). TREK‐1, a Kþ channel involved in polymodal pain perception. The EMBO Journal advance online publication. 4 May 7601116. Baccaglini, P. I., and Hogan, P. G. (1983). Some rat sensory neurons in culture express characteristics of differentiated pain sensory cells. Proc. Natl. Acad. Sci. USA 80, 594–598. Bautista, D. M., Jordt, S. E., Nikai, T., Tsuruda, P. R., Read, A. J., Poblete, J., Yamoah, E. N., Basbaum, A. I., and Julius, D. (2006). TRPA1 mediates the inflammatory actions of environmental irritants and proalgesic agents. Cell 124, 1269–1282. Bell, J., Balanowski, S., and Holmes, M. H. (1994). The structure and function of pacinian corpuscles: A review. Prog. Neurobiol. 42, 79–128. Besch, S. R., Suchyna, T., and Sachs, F. (2002). High‐speed pressure clamp. Pflugers Arch. 445, 161–166. Birder, L. A., Nakamura, Y., Kiss, S., Nealen, M. L., Barrick, S., Kanai, A. J., Wang, E., Ruiz, G., De Groat, W. C., Apodaca, G., Watkins, S., and Caterina, M. J. (2002). Altered urinary bladder function in mice lacking the vanilloid receptor TRPV1. Nat. Neurosci. 5, 856–860. Boland, L. M., Brown, T. A., and Dingledine, R. (1991). Gadolinium block of calcium channels: Influence of bicarbonate. Brain Res. 563, 142–150. Bourinet, E., Alloui, A., Monteil, A., Barrere, C., Couette, B., Poirot, O., Pages, A., McRory, J., Snutch, T. P., Eschalier, A., and Nargeot, J. (2005). Silencing of the Cav3.2 T‐type calcium channel gene in sensory neurons demonstrates its major role in nociception. EMBO J. 24(2), 315–324. Bouskila, Y., and Bostock, H. (1998). Modulation of voltage‐activated calcium currents by mechanical stimulation in rat sensory neurons. J. Neurophysiol. 804, 1647–1652. Brock, J. A., McLachlan, E. M., and Belmonte, C. (1998). Tetrodotoxin‐resistant impulses in single nociceptor nerve terminals in guinea‐pig cornea. J. Physiol. 512, 211–217. Brown, A. G., and Iggo, A. (1967). A quantitative study of cutaneous receptors and aVerent fibres in the cat and rabbit. J. Physiol. 193, 707–733. Caprini, M., Gomis, A., Cabedo, H., Planells‐Cases, R., Belmonte, C., Viana, F., and Ferrer‐ Montiel, A. (2003). GAP43 stimulates inositol trisphosphate‐mediated calcium release in response to hypotonicity. EMBO J. 22(12), 3004–3014.
15. Touch
459
Cesare, P., and McNaughton, P. (1996). A novel heat‐activated current in nociceptive neurons and its sensitization by bradykinin. Proc. Natl. Acad. Sci. USA 93, 15435–15439. Chelur, D. S., Ernstrom, G. G., Goodman, M. B., Yao, C. A., Chen, L., O’Hagan, R., and Chalfie, M. (2002). The mechano‐sensory protein MEC‐6 is a subunit of the C. elegans touch‐cell degenerin channel. Nature 420, 669–673. Chemin, J., Patel, A. J., Duprat, F., Lauritzen, I., Lazdunski, M., and Honore, E. (2005). A phospholipid sensor controls mechanogating of the Kþchannel TREK‐1. EMBO J. 24, 44–53. Cho, H., Shin, J., Shin, C. Y., Lee, S.‐Y., and Oh, U. (2002). Mechanosensitive ion channels in cultured sensory neurons of neonatal rats. J. Neurosci. 22, 1238–1247. Cockayne, D. A., Hamilton, S. G., Zhu, Q. M., Dunn, P. M., Zhong, Y., Novakovic, S., Malmberg, A. B., Cain, G., Berson, A., Kassotakis, L., Hedley, L., Lachnit, W. G., et al. (2000). Urinary bladder hyporeflexia and reduced pain‐related behaviour in P2X3‐deficient mice. Nature 407, 1011–1015. Colbert, H. A., Smith, T. L., and Bargmann, C. I. (1997). OSM‐9, a novel protein with structural similarity to channels is required for olfaction, mechanosensation, and olfactory adaptation in Caenorhabditis elegans. J. Neurosci. 17, 8259–8269. Coric, T., Zheng, D., Gerstein, M., and Canessa, C. M. (2005). Proton sensitivity of ASIC1 appeared with the rise of fishes by changes of residues in the region that follows TM1 in the ectodomain of the channel. J. Physiol. 568, 725–735. Corey, D. P., Garcia‐Anoveros, J., Holt, J. R., Kwan, K. Y., Lin, S. Y., Vollrath, M. A., Amalfitano, A., Cheung, E. L., Derfler, B. H., Duggan, A., Geleoc, G. S., Gray, P. A., et al. (2004). TRPA1 is a candidate for the mechanosensitive transduction channel of vertebrate hair cells. Nature 432, 723–730. Day, B. L., and Fitzpatrick, R. C. (2005). The vestibular system. Curr. Biol. 15, R583–R586. Delmas, P. (2005). Polycystins: Polymodal receptor/ion‐channel cellular sensors. Pflugers Arch. 451, 264–276. Delmas, P., Nauli, S. M., Li, X., Coste, B., Osorio, N., Crest, M., Brown, D. A., and Zhou, J. (2004). Gating of the polycystin ion channel signaling complex in neurons and kidney cells. FASEB J. 18, 740–742. Di Castro, A., Drew, L. J., Wood, J. N., and Cesare, P. (2006). Modulation of sensory neuron mechanotransduction by PKC and NGF dependent pathways. Proc. Natl. Acad. Sci. USA 103, 4699–4704. Di Palma, F., Belyantseva, I. A., Kim, H. J., Vogt, T. F., Kachar, B., and Noben‐Trauth, K. (2002). Mutations in Mcoln3 associated with deafness and pigmentation defects in varitint‐ waddler (Va) mice. Proc. Natl. Acad. Sci. USA 99, 14994–14999. Djouhri, L., and Lawson, S. N. (2004). A beta‐fiber nociceptive primary aVerent neurons: A review of incidence and properties in relation to other aVerent A‐fiber neurons in mammals. Brain Res. Brain Res. Rev. 46(2), 131–145. Djouhri, L., Bleazard, L., and Lawson, S. N. (1998). Association of somatic action potential shape with sensory receptive properties in guinea‐pig dorsal root ganglion neurones. J. Physiol. 513, 857–872. Drew, L. J., Wood, J. N., and Cesare, P. (2002). Distinct mechanosensitive properties of capsaicin‐sensitive and ‐insensitive sensory neurons. J. Neurosci. 22, RC228. Drew, L. J., Rohrer, D. K., Price, M. P., Blaver, K., Cockayne, D. A., Cesare, P., and Wood, J. N. (2004). ASIC2 and ASIC3 do not contribute to mechanically activated current in mammalian sensory neurons. J. Physiol. 556, 691–710. Drummond, H. A., Price, M. P., Welsh, M. J., and Abboud, F. M. (1998). A molecular component of the arterial baro‐receptor mechanotransducer. Neuron 21, 1435–1441. Dubreuil, A. S., Boukhaddaoui, H., Desmadryl, G., Martinez‐Salgado, C., Moshourab, R., Lewin, G. R., Carroll, P., Valmier, J., and Scamps, F. (2004). Role of T‐type calcium
460
Drew et al.
current in identified D‐hair mechanoreceptor neurons studied in vitro. J. Neurosci. 24, 8480–8484. Eberl, D. F., Hardy, R. W., and Kernan, M. J. (2000). Genetically similar transduction mechanisms for touch and hearing in Drosophila. J. Neurosci. 20, 5981–5988. Elmes, S. J., Jhaveri, M. D., Smart, D., Kendall, D. A., and Chapman, V. (2004). Cannabinoid CB2 receptor activation inhibits mechanically evoked responses of wide dynamic range dorsal horn neurons in naive rats and in rat models of inflammatory and neuropathic pain. Eur. J. Neurosci. 20, 2311–2320. Ernstrom, G. G., and Chalfie, M. (2002). Genetics of sensory mechanotransduction. Annu. Rev. Genet. 36, 411–453. Fallon, J. B., Maroto, R., Hamill, O. P., and Proske, U. (2002). ATP suppresses activity in slowly adapting but not rapidly adapting mechanoreceptors in toad skin. Neuroreport 1311, 1443–1446. Fitzgerald, E. M., Okuse, K., Wood, J. N., Dolphin, A. C., and Moss, S. J. (1999). cAMP‐ dependent phosphorylation of the tetrodotoxin‐resistant voltage‐dependent sodium channel SNS. J. Physiol. 516, 433–446. Gale, J. E., Marcotti, W., Kennedy, H. J., Kros, C. J., and Richardson, G. P. (2001). FM1‐43 dye behaves as a permeant blocker of the hair‐cell mechanotransducer channel. J. Neurosci. 21, 7013–7025. Garell, P. C., McGillis, S. L., and Greenspan, J. D. (1996). Mechanical response properties of nociceptors innervating feline hairy skin. J. Neurophysiol. 75, 1177–1189. Goodman, M. B., and Schwarz, E. M. (2003). Transducing touch in Caenorhabditis elegans. Annu. Rev. Physiol. 65, 429–452. Goodman, M. B., Ernstrom, G. G., Chelur, D. S., O’Hagan, R., Yao, C. A., and Chalfie, M. (2002). MEC‐2 regulates C. elegans DEG/ENaC channels needed for mechanosensation. Nature 415, 1039–1042. Gong, Z., Son, W., Chung, Y. D., Kim, J., Shin, D. W., McClung, C. A., Lee, Y., Lee, H. W., Chang, D. J., Kaang, B. K., Cho, H., Oh, U., et al. (2004). Two interdependent TRPV channel subunits inactive and Nanchung, mediate hearing in Drosophila. J. Neurosci. 24, 9059–9066. Gotoh, H., and Takahashi, A. (1999). Mechanical stimuli induce intracellular calcium response in a subpopulation of cultured rat sensory neurons. Neuroscience 92, 1323–1329. Gruss, M., Bushell, T. J., Bright, D. P., Lieb, W. R., Mathie, A., and Franks, N. P. (2004). Two‐ pore‐domain Kþ channels are a novel target for the anesthetic gases xenon, nitrous oxide, and cyclopropane. Mol. Pharmacol. 65, 443–452. Gschossmann, J. M., Chaban, V. V., McRoberts, J. A., Raybould, H. E., Young, S. H., Ennes, H. S., Lembo, T., and Mayer, E. A. (2000). Mechanical activation of dorsal root ganglion cells in vitro: Comparison with capsaicin and modulation by kappa‐opioids. Brain Res. 856, 101–110. Grimm, C., Kraft, R., Sauerbruch, S., Schultz, G., and Harteneck, C. (2003). Molecular and functional characterization of the melastatin‐related cation channel TRPM3. J. Biol. Chem. 278, 21493–21501. Haeberle, H., Fujiwara, M., Chuang, J., Medina, M. M., Panditrao, M. V., Bechstedt, S., Howard, J., and Lumpkin, E. A. (2004). Molecular profiling reveals synaptic release machinery in Merkel cells. Proc. Natl. Acad. Sci. USA 101(40), 14503–14508. Hamill, O. P., and McBride, D. W., Jr. (1996). The pharmacology of mechanogated membrane ion channels. Pharmacol. Rev. 48, 231–252. Hayes, A. G., and Tyers, M. B. (1980). EVects of capsaicin on nociceptive heat pressure and chemical thresholds. Brain Res. 189(2), 561–564. Hu, J., and Lewin, G. R. (2006). Mechanosensitive currents in the neurites of cultured mouse sensory neurons. J. Physiol. 577, 815–828.
15. Touch
461
Hu, J., Milenkovic, N., and Lewin, G. R. (2006). The high threshold mechanotransducer: A status report. Pain 120(1–2), 3–7[Epub (2005) Dec 15]. Hunt, C. C., and Ottoson, D. (1975). Impulse activity and receptor potential of primary and secondary endings of isolated mammalian muscle spindles. J. Physiol. 252, 259–281. Hunt, C. C., and Ottoson, D. (1976). Initial burst of primary endings of isolated mammalian muscle spindles. J. Neurophysiol. 39, 324–330. Hlushchuk, Y., and Hari, R. (2006). Transient suppression of ipsilateral primary somatosensory cortex during tactile finger stimulation. J. Neurosci. 26(21), 5819–5824. Husmark, I., and Ottoson, D. (1971). Ionic eVects on spindle adaptation. J. Physiol. 218, 257–269. Ivanavicius, S. P., Blake, D. R., Chessell, I. P., and Mapp, P. I. (2004). Isolectin B4 binding neurons are not present in the rat knee joint. Neuroscience 128, 555–560. Katz, B. (1950). Depolarization of sensory terminals and the initiation of impulses in the muscle spindle. J. Physiol. 11, 261–282. Keller, J. T., and Marfurt, C. F. (1991). Peptidergic and serotoninergic innervation of the rat dura mater. J. Comp. Neurol. 309, 515–534. Kelly, S., Jhaveri, M. D., Sagar, D. R., Kendall, D. A., and Chapman, V. (2003). Activation of peripheral cannabinoid CB1 receptors inhibits mechanically evoked responses of spinal neurons in noninflamed rats and rats with hindpaw inflammation. Eur. J. Neurosci. 18, 2239–2243. Kim, J., Chung, Y. D., Park, D. Y., Choi, S., Shin, D. W., Soh, H., Lee, H. W., Son, W., Yim, J., Park, C. S., Kernan, M. J., and Kim, C. (2003). A TRPV family ion channel required for hearing in Drosophila. Nature 424, 81–84. Kinkelin, I., Stucky, C. L., and Koltzenburg, M. (1999). Postnatal loss of Merkel cells, but not of slowly adapting mechanoreceptors in mice lacking the neurotrophin receptor p75. Eur. J. Neurosci. 11, 3963–3969. Koizumi, S., Fujishita, K., Inoue, K., Shigemoto‐Mogami, Y., Tsuda, M., and Inoue, K. (2004). Ca2þ waves in keratinocytes are transmitted to sensory neurons: The involvement of extracellular ATP and P2Y2 receptor activation. Biochem. J. 380(Pt. 2), 329–338. Kung, C. (2005). A possible unifying principle for mechanosensation. Nature 436, 647–654. Kwan, K. Y., Allchorne, A. J., Vollrath, M. A., Christensen, A. P., Zhang, D. S., Woolf, C. J., and Corey, D. P. (2006). TRPA1 contributes to cold, mechanical, and chemical nociception but is not essential for hair‐cell transduction. Neuron 50, 277–289. Leem, J. W., Willis, W. D., and Chung, J. M. (1993). Cutaneous sensory receptors in the rat foot. J. Neurophysiol. 69, 1684–1699. LeMasurier, M., and Gillespie, P. G. (2005). Hair‐cell mechanotransduction and cochlear amplification. Neuron 48, 403–415. Lin, S. Y., and Corey, D. P. (2005). TRP channels in mechanosensation. Curr. Opin. Neurobiol. 15, 350–357. Loewenstein, W. R., and Mendelson, M. (1965). Components of receptor adaptation in a Pacinian corpuscle. J. Physiol. 177, 377–397. Loewenstein, W. R., and Skalak, R. (1966). Mechanical transmission in a Pacinian corpuscle. An analysis and a theory. J. Physiol. 182, 346–378. Lundbaek, J. A., Birn, P., Hansen, A. J., Sogaard, R., Nielsen, C., Girshman, J., Bruno, M. J., Tape, S. E., Egebjerg, J., Greathouse, D. V., Mattice, G. L., Koeppe, R. E., II, et al. (2004). Regulation of sodium channel function by bilayer elasticity: The importance of hydrophobic coupling. EVects of Micelle‐forming amphiphiles and cholesterol. J. Gen. Physiol. 123, 599–621. Lewin, G. R., and Moshourab, R. (2004). Mechanosensation and pain. J. Neurobiol. 61, 30–44. Liedtke, W., and Friedman, J. M. (2003). Abnormal osmotic regulation in trpv4 / mice. Proc. Natl. Acad. Sci. USA 100, 13698–13703.
462
Drew et al.
Liedtke, W., and Kim, C. (2005). Functionality of the TRPV subfamily of TRP ion channels: Add mechano‐TRP and osmo‐TRP to the lexicon! Cell. Mol. Life Sci. 622985–623001. Li, W., Feng, Z., Sternberg, P. W., and Xu, X. Z. (2006). A C. elegans stretch receptor neuron revealed by a mechanosenstive TRP channel homologue. Nature 440, 684–687. Liu, J., Schrank, B., and Waterston, R. H. (1996). Interaction between a putative mechanosensory membrane channel and a collagen. Science 273, 361–364. McCarter, G. C., Reichling, D. B., and Levine, J. D. (1999). Mechanical transduction by rat dorsal root ganglion neurons in vitro. Neurosci. Lett. 273, 179–182. Macefield, V. G. (2005). Physiological characteristics of low‐threshold mechanoreceptors in joints, muscle and skin in human subjects. Clin. Exp. Pharmacol. Physiol. 32, 135–144. Maroto, R., and Hamill, O. P. (2001). Brefeldin A block of integrin‐dependent mechanosensitive ATP release from Xenopus oocytes reveals a novel mechanism of mechanotransduction. J. Biol. Chem. 276(26), 23867–23872. Maroto, R., Raso, A., Wood, T. G., Kurosky, A., Martinac, B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nat. Cell Biol. 7, 179–185. Matthews, E. A., Wood, J. N., and Dickenson, A. H. (2006). Nav 1.8‐null mice show stimulus‐ dependent deficits in spinal neuronal activity. Mol. Pain 2, 5. McGrath, J., Somlo, S., Makova, S., Tian, X., and Brueckner, M. (2003). Two populations of node monocilia initiate left‐right asymmetry in the mouse. Cell 114, 61–73. Mills, L. R., and Diamond, J. (1995). Merkel cells are not the mechanosensory transducers in the touch dome of the rat. J. Neurocytol. 24, 117–134. Mizuno, A., Matsumoto, N., Imai, M., and Suzuki, M. (2003). Impaired osmotic sensation in mice lacking TRPV4. Am. J. Physiol. Cell Physiol. 285, C96–C101. Mogil, J. S., Breese, N. M., Witty, M. F., Ritchie, J., Rainville, M. L., Ase, A., Abbadi, N., Stucky, C. L., and Seguela, P. (2005). Transgenic expression of a dominant‐negative ASIC3 subunit leads to increased sensitivity to mechanical and inflammatory stimuli. J. Neurosci. 25(43), 9893–9901. Montalbetti, N., Li, Q., Gonzalez‐Perrett, S., Semprine, J., Chen, X. Z., and Cantiello, H. F. (2005). EVect of hydro‐osmotic pressure on polycystin‐2 channel function in the human syncytiotrophoblast. Pflugers Arch. 451, 294–303. Muraki, K., Iwata, Y., Katanosaka, Y., Ito, T., Ohya, S., Shigekawa, M., and Imaizumi, Y. (2003). TRPV2 is a component of osmotically sensitive cation channels in murine aortic myocytes. Circ. Res. 93, 829–838. Naito, E., and Ehrsson, H. H. (2006). Somatic sensation of hand‐object interactive movement is associated with activity in the left inferior parietal cortex. J. Neurosci. 26(14), 3783–3790. Nakamura, F., and Strittmatter, S. M. (1996). P2Y1 purinergic receptors in sensory neurons: Contribution to touch‐induced impulse generation. Proc. Natl. Acad. Sci. USA 93(19), 10465–10470. Nassar, M. A., Stirling, L. C., Forlani, G., Baker, M. D., Matthews, E. A., Dickenson, A. H., and Wood, J. N. (2004). Nociceptor‐specific gene deletion reveals a major role for Nav1.7 (PN1) in acute and inflammatory pain. Proc. Natl. Acad. Sci. USA 101(34), 12706–12711. Nagy, J. I., and van der Kooy, D. (1983). EVects of neonatal capsaicin treatment on nociceptive thresholds in the rat. J. Neurosci. 3, 1145–1150. Nauli, S. M., Alenghat, F. J., Luo, Y., Williams, E., Vassilev, P., Li, X., Elia, A. E., Lu, W., Brown, E. M., Quinn, S. J., Ingber, D. E., and Zhou, J. (2003). Polycystins 1 and 2 mediate mechanosensation in the primary cilium of kidney cells. Nat. Genet. 33, 129–137.
15. Touch
463
Ogawa, H. (1996). The Merkel cell as a possible mechanoreceptor cell. Prog. Neurobiol. 49(4), 317–334. O’Hagan, R., Chalfie, M., and Goodman, M. B. (2005). The MEC‐4 DEG/ENaC channel of Caenorhabditis elegans touch receptor neurons transduces mechanical signals. Nat. Neurosci. 8, 43–50. Olausson, H., Wessberg, J., and Kakuda, N. (2000). Tactile directional sensibility: Peripheral neural mechanisms in man. Brain Res. 866(1–2), 178–187. Olausson, H., Lamarre, Y., Backlund, H., Morin, C., Wallin, B. G., Starck, G., Ekholm, S., Strigo, I., Worsley, K., Vallbo, A. B., and Bushnell, M. C. (2002). Unmyelinated tactile aVerents signal touch and project to insular cortex. Nat. Neurosci. 5, 900–904. Ottoson, D. (1964). The eVect of sodium deficiency on the response of the isolated muscle spindle. J. Physiol. 171, 109–118. Ou, Y., Strege, P., Miller, S. M., Makielski, J., Ackerman, M., Gibbons, S. J., and Farrugia, G. (2003). Syntrophin gamma 2 regulates SCN5A gating by a PDZ domain‐mediated interaction. J. Biol. Chem. 278, 1915–1923. Page, A. J., Brierley, S. M., Martin, C. M., Martinez‐Salgado, C., Wemmie, J. A., Brennan, T. J., Symonds, E., Omari, T., Lewin, G. R., Welsh, M. J., and Blackshaw, L. A. (2004). The ion channel ASIC1 contributes to visceral but not cutaneous mechanoreceptor function. Gastroenterology 127, 1739–1747. Paoletti, P., and Ascher, P. (1994). Mechanosensitivity of NMDA receptors in cultured mouse central neurons. Neuron 13(3), 645–655. Patapoutian, A., Peier, A. M., Story, G. M., and Viswanath, V. (2003). ThermoTRP channels and beyond: Mechanisms of temperature sensation. Nat. Rev. Neurosci. 4, 529–539. Patel, A. J., Lazdunski, M., and Honore, E. (2001). Lipid and mechano‐gated 2P domain K(þ) channels. Curr. Opin. Cell Biol. 13(4), 422–428. Paukert, M., Sidi, S., Russell, C., Siba, M., Wilson, S. W., Nicolson, T., and Grunder, S. (2004). A family of acid‐sensing ion channels from the zebrafish: Widespread expression in the central nervous system suggests a conserved role in neuronal communication. J. Biol. Chem. 279, 18783–18791. Price, M. P., Lewin, G. R., McIlwrath, S. L., Cheng, C., Xie, J., Heppenstall, P. A., Stucky, C. L., Mannsfeldt, A. G., Brennan, T. J., Drummond, H. A., Qiao, J., Benson, C. J., et al. (2000). The mammalian sodium channel BNC1 is required for normal touch sensation. Nature 407(6807), 1007–1011. Price, M. P., McIlwrath, S. L., Xie, J., Cheng, C., Qiao, J., Tarr, D. E., Sluka, K. A., Brennan, T. J., Lewin, G. R., and Welsh, M. J. (2001). The DRASIC cation channel contributes to the detection of cutaneous touch and acid stimuli in mice. Neuron 32, 1071–1083. Randall, L. O., and Selitto, J. (1957). A method for measurement of analgesic activity on inflamed tissue. Arch. Int. Pharmacodyn. Ther. 111, 409–419. Raybould, H. E., Gschossman, J. M., Ennes, H., Lembo, T., and Mayer, E. A. (1999). Involvement of stretch‐sensitive calcium flux in mechanical transduction in visceral aVerents. J. Auton. Nerv. Syst. 75, 1–6. Reeh, P. W. (1986). Sensory receptors in mammalian skin in an in vitro preparation. Neurosci. Lett. 66, 141–146. Reid, G., Babes, A., and Pluteanu, F. (2002). A cold‐ and menthol‐activated current in rat dorsal root ganglion neurons: Properties and role in cold transduction. J. Physiol. 545, 595–614. Robinson, D. R., McNaughton, P. A., Evans, M. L., and Hicks, G. A. (2004). Characterization of the primary spinal aVerent innervation of the mouse colon using retrograde labelling. Neurogastroenterol. Motil. 16, 113–124.
464
Drew et al.
Roza, C., Puel, J. L., Kress, M., Baron, A., Diochot, S., Lazdunski, M., and Waldmann, R. (2004). Knockout of the ASIC2 channel in mice does not impair cutaneous mechanosensation, visceral mechanonociception and hearing. J. Physiol. 558, 659–669. Ruehle, B. S., Handwerker, H. O., Lennerz, J. K., Ringler, R., and Forster, C. (2006). Brain activation during input from mechanoinsensitive versus polymodal C‐nociceptors. J. Neurosci. 26(20), 5492–5499. Saumet, J. L., and Duclaux, R. (1982). Analgesia induced by neonatal capsaicin treatment in rats. Pharmacol. Biochem. Behav. 16, 241–243. Schmidt, R., Schmelz, M., Forster, C., Ringkamp, M., Torebjork, E., and Handwerker, H. (1995). Novel classes of responsive and unresponsive C nociceptors in human skin. J. Neurosci. 15, 333–341. Sharma, R. V., Chapleau, M. W., Hajduczok, G., Wachtel, R. E., Waite, L. J., Bhalla, R. C., and Abboud, F. M. (1995). Mechanical stimulation increases intracellular calcium concentration in nodose sensory neurons. Neuroscience 66, 433–441. Shin, J. B., Martinez‐Salgado, C., Heppenstall, P. A., and Lewin, G. R. (2003). A T‐type calcium channel required for normal function of a mammalian mechanoreceptor. Nat. Neurosci. 6, 724–730. Shin, J. B., Adams, D., Paukert, M., Siba, M., Sidi, S., Levin, M., Gillespie, P. G., and Grunder, S. (2005). Xenopus TRPN1 (NOMPC) localizes to microtubule‐based cilia in epithelial cells, including inner‐ear hair cells. Proc. Natl. Acad. Sci. USA 102, 12572–12577. Shir, Y., and Seltzer, Z. (1990). A‐fibers mediate mechanical hyperesthesia and allodynia and C‐fibers mediate thermal hyperalgesia in a new model of causalgiform pain disorders in rats. Neurosci. Lett. 115, 62–67. Sidi, S., Friedrich, R. W., and Nicolson, T. (2003). NompC TRP channel required for vertebrate sensory hair cell mechanotransduction. Science 301, 96–99. Souslova, V., Cesare, P., Ding, Y., Akopian, A. N., Stanfa, L., Suzuki, R., Carpenter, K., Dickenson, A., Boyce, S., Hill, R., Nebenuis‐Oosthuizen, D., Smith, A. J., et al. (2000). Warm‐coding deficits and aberrant inflammatory pain in mice lacking P2X3 receptors. Nature 407, 1015–1017. Steen, K. H., Steen, A. E., and Reeh, P. W. (1995). A dominant role of acid pH in inflammatory excitation and sensitization of nociceptors in rat skin, in vitro. J. Neurosci. 15, 3982–3989. Sullivan, M. J., Sharma, R. V., Wachtel, R. E., Chapleau, M. W., Waite, L. J., Bhalla, R. C., and Abboud, F. M. (1997). Non‐voltage‐gated Ca2þinflux through mechanosensitive ion channels in aortic baroreceptor neurons. Circ. Res. 80, 861–867. Sharif Naeini, R., Witty, M. F., Seguela, P., and Bourque, C. W. (2006). An N‐terminal variant of Trpv1 channel is required for osmosensory transduction. Nat. Neurosci. 9, 93–98. Story, G. M., Peier, A. M., Reeve, A. J., Eid, S. R., Mosbacher, J., Hricik, T. R., Earley, T. J., Hergarden, A. C., Andersson, D. A., Hwang, S. W., McIntyre, P., Jegla, T., et al. (2003). ANKTM1, a TRP‐like channel expressed in nociceptive neurons, is activated by cold temperatures. Cell 112, 819–829. Stucky, C. L., DeChiara, T., Lindsay, R. M., Yancopoulos, G. D., and Koltzenburg, M. (1998). Neurotrophin 4 is required for the survival of a subclass of hair follicle receptors. J. Neurosci. 18, 7040–7046. Suzuki, H., Kerr, R., Bianchi, L., Frokjaer‐Jensen, C., Slone, D., Xue, J., Gerstbrein, B., Driscoll, M., and Schafer, W. R. (2003a). In vivo imaging of C. elegans mechanosensory neurons demonstrates a specific role for the MEC‐4 channel in the process of gentle touch sensation. Neuron 39, 1005–1017.
15. Touch
465
Suzuki, M., Watanabe, Y., Oyama, Y., Mizuno, A., Kusano, E., Hirao, A., and Ookawara, S. (2003b). Localization of mechanosensitive channel TRPV4 in mouse skin. Neurosci. Lett. 353, 189–192. Suzuki, M., Mizuno, A., Kodaira, K., and Imai, M. (2003c). Impaired pressure sensation in mice lacking TRPV4. J. Biol. Chem. 278, 22664–22668. Takahashi, A., and Gotoh, H. (2000). Mechanosensitive whole‐cell currents in cultured rat somatosensory neurons. Brain Res. 869, 225–230. Tavernarakis, N., ShreZer, W., Wang, S., and Driscoll, M. (1997). unc‐8, a DEG/ENaC family member, encodes a subunit of a candidate mechanically gated channel that modulates C. elegans locomotion. Neuron 18, 107–119. Tobin, D., Madsen, D., Kahn‐Kirby, A., Peckol, E., Moulder, G., Barstead, R., Maricq, A., and Bargmann, C. (2002). Combinatorial expression of TRPV channel proteins defines their sensory functions and subcellular localization in C. elegans neurons. Neuron 35, 307–318. Tracey, W. D., Jr, Wilson, W. D., Laurent, G., and Benzer, S. (2003). Painless, a Drosophila gene essential for nociception. Cell 113(2), 261–273. Vallbo, A. B., and Hagbarth, K. E. (1968). Mechnoreceptor activity recorded from human peripheral nerves. Electroencephalogr. Clin. Neurophysiol. 25, 407. Viana, F., de la Pena, E., Pecson, B., Schmidt, R. F., and Belmonte, C. (2001). Swelling‐ activated calcium signalling in cultured mouse primary sensory neurons. Eur. J. Neurosci. 13, 722–734. Volk, T., Schwoerer, A. P., Thiessen, S., Schultz, J. H., and Ehmke, H. (2003). A polycystin‐2‐ like large conductance cation channel in rat left ventricular myocytes. Cardiovasc. Res. 58, 76–88. von Frey, M. (1894). Beitra¨ge zur Physiologie des Schmerzsinns (2. Mitteilung). Berichte u¨ber die Verhandlungen der Ko¨niglich Sa¨chsischen Gesellshaft der Wissenschaften 46, 283–297. Vriens, J., Watanabe, H., Janssens, A., Droogmans, G., Voets, T., and Nilius, B. (2004). Cell swelling, heat, and chemical agonists use distinct pathways for the activation of the cation channel TRPV4. Proc. Natl. Acad. Sci. USA 101, 396–401. Walker, R. G., Willingham, A. T., and Zuker, C. S. (2000). A Drosophila mechanosensory transduction channel. Science 287, 2229–2234. Wood, J. N., Boorman, J. P., Okuse, K., and Baker, M. D. (2004). Voltage‐gated sodium channels and pain pathways. J. Neurobiol. 61(1), 55–71. Woodbury, C. J., Ritter, A. M., and Koerber, H. R. (2001). Central anatomy of individual rapidly adapting low‐threshold mechanoreceptors innervating the ‘‘hairy’’ skin of newborn mice: Early maturation of hair follicle aVerents. J. Comp. Neurol. 436, 304–323. Yusaf, S. P., Goodman, J., Pinnock, R. D., Dixon, A. K., and Lee, K. (2001). Expression of voltage‐gated calcium channel subunits in rat dorsal root ganglion neurons. Neurosci. Lett. 311, 137–141. Zhang, Z., and Bourque, C. W. (2003). Osmometry in osmosensory neurons. Nat. Neurosci. 6, 1021–1022. Zylka, M. J., Rice, F. L., and Anderson, D. J. (2005). Topographically distinct epidermal nociceptive circuits revealed by axonal tracers targeted to Mrgprd. Neuron 45, 17–25.
This page intentionally left blank
CHAPTER 16 Mechanosensitive Ion Channels in Dystrophic Muscle JeVry B. Lansman Department of Cellular and Molecular Pharmacology, School of Medicine, University of California, San Francisco, California 94143
I. II. III. IV.
Overview Introduction MS Channel Expression During Myogenesis Permeabilty Properties of MS Channels in Skeletal Muscle A. Permeability to Monovalent Cations B. Permeability to Divalent Cations V. Gating A. SA Gating B. Voltage‐Sensitive Gating C. Modal Gating in mdx Muscle VI. Pharmacology A. Block by Gadolinium Ion B. Aminoglycoside Antibiotics VII. Conclusions References
I. OVERVIEW Mechanosensitive (MS) ion channels are expressed abundantly in skeletal muscle at all stages of development. In wild‐type muscle, MS channels show primarily stretch‐activated (SA) gating. In dystrophic myotubes from the mdx mouse, a loss‐of‐function mutant that lacks dystrophin, there are two types of MS channels. In addition to conventional SA channels, some channels shift into a novel gating mode in which channels stay open for extended periods of time and are stretch‐inactivated (SI). The shift in gating Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59017-9
468
JeVry B. Lansman
mode can occur slowly at the start of an experiment or more abruptly in response to strong pressure or voltage steps. SA and SI gating have similar energetic requirements which likely reflect the energy required to deform the bilayer. The pharmacological properties of MS channels show similarities to other cation‐selective channels. Highly charged pharmacological probes such as lanthanide cations and aminoglycoside antibiotics act by plugging the channel pore rather than interacting with the lipid bilayer.
II. INTRODUCTION The ability to detect mechanical forces is shared by all living organisms. Mechanical sensitivity underlies the detection of auditory and tactile stimuli as well as distension pressures, shear stresses, and osmotic gradients in various cell types. Despite the importance of mechanical sensitivity, it remains the least understood sensory process. MS ion channels are thought to constitute primary sensors of mechanical stimuli. In most cases, however, their molecular identity and mechanism of force transduction remain largely unknown. SA MS channels were first detected in skeletal muscle cells using single‐ channel recording methods (Brehm et al., 1984; Guharay and Sachs, 1984). Guharay and Sachs (1984) found that treating cells with drugs that disrupt actin filaments or microtubules increased the sensitivity of SA channels to membrane tension. This finding suggested a role of the cytoskeleton in transduction of membrane tension. Skeletal muscle possesses a complex cytoskeleton that plays a key role in maintaining its shape and surface morphology, as well as supporting stresses generated during contraction. How cytoskeletal structures regulate mechanotransduction has been a diYcult question to address experimentally. One approach to this problem is to use loss‐of‐function mutants in which a specific cytoskeletal protein is absent. Dystrophin is a large submembrane cytoskeletal protein that is a member of the b‐spectrin/a‐actinin protein family (Koenig et al., 1988). Dystrophin is linked to the membrane by a glycoprotein complex composed of the sarcoglycans, dystroglycan, syntrophin, and dystrobrevin (Ervasti et al., 1990; Yoshida and Ozawa, 1990; Ibraghimov-Beskrovnaya et al., 1992, 1993; Adams et al., 1993; Ahn et al., 1996). The glycoprotein complex spans the membrane and connects the actin cytoskeleton to laminin in the extracellular basement membrane (Ervasti and Campbell, 1993). In the mdx mouse, a point mutation in exon 23 of the dystrophin gene leads to the loss of full‐length dystrophin in skeletal muscle (HoVman et al., 1987; Sicinski et al., 1989). The mdx mouse model makes it possible to study the biophysical consequences of the loss of dystrophin on the mechanotransduction process.
469
16. Mechanosensitive Ion Channels in Dystrophic Muscle
III. MS CHANNEL EXPRESSION DURING MYOGENESIS During myogenesis, myoblastic stem cells diVerentiate and form multinucleated skeletal muscle cells (Fig. 1A). In vitro, myoblasts proliferate until stimulated to withdraw from the cell cycle and begin diVerentiation. Subsequently, myoblasts align oriented with their long axis in parallel and fuse to A
B
Single myoblasts
Aligned myoblasts
AChγ 60
AChγ 40
Voltage sensitive SA
AChγ 40
Voltage sensitive SA
0
AChγ 40
50
Voltage sensitive SA
Percentage of patches
100
Myotubes
FIGURE 1 Expression of MS channels during myogenesis. (A) Micrograph showing the stages of myogenesis in vitro during which patch clamp recordings were made: single myoblasts 24 h after plating C2 mouse muscle cells at low density (left); myoblasts that had proliferated in culture and had aligned prior to fusion (middle); and multinucleated myotubes during the first week after myoblast fusion (right). (B) Fraction of patches containing the acetylcholine receptor channel, SA channel, and a voltage‐insensitive cation channel. AChg40 and AChg60 represent the small and large conductance acetylcholine receptor channels, respectively. AChg60 is the adult form of the AChR channel that appears in fully diVerentiated muscle fibers. Adapted from Franco and Lansman (1990a).
470
JeVry B. Lansman
form electrically excitable multinucleated myotubes. Figure 1B shows that MS channels can be detected at all stages of myogenesis in vitro. SA MS channels are highly expressed in myoblasts (80% patches) but expression declines during skeletal muscle diVerentiation (30% patches in myotubes). In all of the recordings, 15–20 patches contained activity of a cation‐ selective channel that showed an increase in open probability with membrane depolarization, but was otherwise insensitive to membrane stretch. This activity was originally thought to represent MS channels in patches where membrane geometry prevented normal force transduction. More recent data suggests this activity arises from a cation channel regulated by IGF‐1 (unpublished data).
IV. PERMEABILTY PROPERTIES OF MS CHANNELS IN SKELETAL MUSCLE A. Permeability to Monovalent Cations MS channels in skeletal muscle are permeable to monovalent cations, but show rather low selectivity among the alkali metal cations. MS channel currents are well resolved when extracellular Naþ is replaced with 155‐mM Liþ, Naþ, Kþ, Rbþ, or Csþ. The single‐channel conductance is largest with Rbþ as the charge carrier (38 pS) and smallest with Liþ (27 pS); (Franco and Lansman, 1990a). The conductance selectivity sequence is Rb > K > Na > Cs > Li, which corresponds to Eisenman sequence III for a weak field strength site (Eisenman, 1962). There is little change in the reversal potential in the presence of the diVerent alkali cations, indicating little discrimination among the alkali metal cations.
B. Permeability to Divalent Cations MS channels in skeletal muscle have a relatively high permeability to Ca2þ and other divalent cations. Figure 2 shows records of the activity of single MS channels in the presence of Ca2þ‐ or Ba2þ‐containing solution. The single‐channel current–voltage relation in the presence of either 110‐mM Ca2þ or 110‐mM Ba2þ as the extracellular solution gives single‐channel conductances of 13 and 24 pS, respectively. The reversal potential in the presence of Ca2þ‐containing solutions (þ22 mV) was used to calculate the relative permeability of Ca2þ to Kþ (PCa/PK), which was 7. MS channels in skeletal muscle, thus, have a relatively high permeability to Ca2þ. The high Ca2þ permeability allows relatively large Ca2þ fluxes at negative membrane potentials.
471
16. Mechanosensitive Ion Channels in Dystrophic Muscle A 110-mM CaCl 2
110-mM BaCl2
1 pA 200 ms B
Ba2+
Ca2+
−100
100
−100
100
mV
−2
−4
mV
−2
pA
−4
pA
FIGURE 2 Ca2þ permeability of MS channels in skeletal muscle. (A) Single‐channel currents carried by Ca2þ (top) or Ba2þ (bottom). The holding potential was 60 mV. The patch electrode contained either 110‐mM CaCl2 or BaCl2. (B) The single‐channel current–voltage relation. Filled symbols, recordings from cultured myotubes; open symbols, recordings from myoblasts. DiVerent symbols represent recordings from diVerent patches. The conductance was 13.1 1 pS and current reversed at þ22 6 mV (S.D., n ¼ 8) with Ca2þ; the conductance was 24 4 pS and current reversed at þ17 8 mV (S.D., n ¼ 7) with Ba2þ. Adapted from Franco and Lansman (1990a).
V. GATING A. SA Gating SA MS channels are found in myotubes grown in tissue culture (Franco and Lansman, 1990a; Franco-Obrego´n and Lansman, 1994). SA with identical conductance properties can also be detected in skeletal muscle fibers acutely isolated from the flexor digitorum brevis (FDB), a small, fast twitch fiber (Franco-Obrego´n and Lansman, 1994). Patch clamp recordings from FDB fibers from 2‐week‐old mice show that 70% patches contain SA channels, although these have a much lower resting open probability in the absence of a pressure stimulus. Despite these diVerences, SA channels in
472
JeVry B. Lansman
myotubes and fibers in wild‐type muscle activate over a similar pressure range in response to either positive or negative pressure. The absence of dystrophin in mdx FDB fibers alters the pressure sensitivity of SA channels. Figure 3 shows the activity of SA channels recorded from FDB fibers from wild‐type (Fig. 3A) or mdx (Fig. 3B) mice. In both cases, channel activity increases on applying negative pressure to the electrode. Figure 3C shows the pressure dependence of channel opening. Channel opening was somewhat less sensitive to pressure in mdx fibers. In addition, channel activity after a step of negative pressure in recordings from mdx fibers was generally smaller than the activity before the pressure step (data not shown, see Franco-Obrego´n and Lansman, 1994). By contrast, channel activity was higher after a pressure step compared with activity before the step in wild‐type fibers. These diVerences in SA gating may reflect diVerences in the
A
Wild-type fibers
mmHg 0
−5
−10
2 pA B
400 ms mdx fibers
0
Channel open probability
C
1.0 0.8 0.6 0.4 0.2 0.0
−5
−30
−20 −10 Suction (mmHg)
0
−10 FIGURE 3 SA channels in acutely isolated wild‐type and mdx FDB fibers. (A) Single‐ channel activity recorded from wild‐type FDB fiber. Pressure applied to the electrode is indicated to the right of each current record. (B) Single‐channel activity recorded from an mdx FDB fiber. (C) The relationship between pressure and channel open probability. Recordings were made from wild‐type (open symbols; n ¼ 22) and mdx (filled symbols; n ¼ 25) FDB fibers. The fit through the experimental points represent the Boltzmann relationship with half‐ activation pressure, P1/2 of 14.0 and 20.0 mmHg and steepnesses 3.0 and 5.0 mmHg, for wild‐type and mdx fibers, respectively. FDB fibers were isolated from 2‐week‐old mice. At this age, 17% of the recordings from wild‐type and 22% of the recordings from mdx fibers had cation channel activity that failed to respond to suction. Adapted from Franco‐Obrego´n and Lansman (1994).
16. Mechanosensitive Ion Channels in Dystrophic Muscle
473
mechanical properties of the membrane of wild‐type and mdx muscle (Pasternak et al., 1995).
B. Voltage‐Sensitive Gating SA channels in skeletal muscle have a voltage‐sensitive gating mechanism in which channel burst duration increases with depolarization (Franco and Lansman, 1990). This is shown for the SA channels in FDB fibers in Fig. 4A. Figure 4B shows that channel open probability increases with membrane depolarization, but the intrinsic voltage sensitivity is similar for MS channels in wild‐type (open symbols) and mdx (filled symbols) fibers. It is interesting that SA channels in wild‐type and mdx fibers have diVerent sensitivities to stretch, but not to voltage. A simple explanation for this is that diVerences in sensitivity to membrane stretch reflect diVerences in the mechanical properties of the membrane of dystrophin‐containing and dystrophin‐deficient muscle. Voltage sensitivity, on the other hand, is apparently unaVected by the presence or absence of dystrophin, suggesting there is a direct eVect of voltage on channel gating that does not depend on membrane mechanics.
−30 −40 −50 −60 −70 −80 2 pA
B Channel open probability
mV −20
A
10
1
0.1
0.01 −120
−80 −40 Holding potential (mV)
0
400 ms FIGURE 4 Voltage‐dependent gating of SA channels. (A) Records of single‐channel activity obtained from a recording on an mdx FDB fiber. The patch holding potential is indicated at the right of each current record. (B) EVect of the holding potential on channel open probability in recordings from wild‐type (open symbols) and mdx (filled symbols) fibers. Channel activity increased with depolarization e‐fold per 56 and 53 mV in wild‐type and mdx fibers, respectively.
474
JeVry B. Lansman
C. Modal Gating in mdx Muscle 1. SI Gating Mode Patch clamp recordings from tissue cultured myotubes from mdx mice have revealed striking changes in MS channel behavior that cannot be explained in terms of simple changes in membrane mechanical properties associated with dystrophin deficiency (Franco and Lansman, 1990b, 2002). MS channels typically are open for only a few milliseconds at rest and a pressure stimulus causes stretch activation. In some recordings from mdx myotubes (20–50% patches), however, MS channels remain open for many tens of seconds. The persistent opening of MS channels has been interpreted as a shift in gating mode where the open state is energetically favored at rest. Figure 5 shows an example of a change in MS channel gating mode that occurred at the start of an experiment. In mdx muscle, MS channel open probability was low at the start of the recording, but gradually increased over the next several minutes (Fig. 5A). Channels remained open almost continuously for the duration of the experiment. By contrast, MS channels in wild‐type myotubes open only very briefly at the beginning of the recording but then close and remain closed (Fig. 5B).
A
mdx myotubes
B
Wild-type myotubes
2 pA 10 s FIGURE 5 A novel MS channel gating mode characterized by persistent channel opening in mdx myotubes. (A) Recording from a patch on an mdx myotube showing a slow, persistent increase in channel opening following the start of the recording (arrowhead). (B) Recordings from two diVerent wild‐type myotubes showing that seal formation (arrowhead) caused a brief inward current, but channel activity remained negligible for the duration of the recording in the absence of a pressure stimulus. Single‐channel currents were filtered at 0.5 kHz and sampled at 1.25 kHz. Adapted from Franco-Obrego´n and Lansman (2002).
475
16. Mechanosensitive Ion Channels in Dystrophic Muscle
MS channels that become persistently open also acquire a SI gating mechanism. Figure 6A shows the open probability measured during consecutive sweeps plotted as a function of time during the experiment that lasted for about 10 min. Applying negative pressure suppressed channel activity. Figure 6B shows representative currents during an individual sweep with either 0‐mm applied pressure (first and third traces) or 30 mmHg (second and fourth traces). The data show that SI gating is readily reversible and quite stable over many minutes.
NP
A 0
0
B
mdx myotube
1
2 pA 10 s
−30 mmHg
−30 mmHg
FIGURE 6 The SI gating mode. (A) MS channel open probability (Npo) during consecutive 1 s intervals of a recording lasting 13 min. This patch contained only a single channel. Suction was applied to the patch electrode at the indicated times. Mean channel open probability during the first 143‐s interval was 0.54, and applying 30 mmHg of pressure to the electrode for 153 s reduced channel open probability to 0.044. After releasing the pressure, channel open probability returned to 0.85 and a subsequent application of 30 mmHg for 137 s reduced channel open probability to 0.01; open probability returned to 0.63 after suction was released. (B) Representative current records during the indicated periods. Single‐channel currents were filtered at 0.5 kHz and sampled at 1.25 kHz. Adapted from Franco-Obrego´n and Lansman (2002).
476
JeVry B. Lansman
An analysis of the pressure dependence of SA and SI showed an important similarity in the transduction process, despite the fact that pressure causes opposite changes in channel opening. Figure 7 shows the relationship between pressure and channel open probability for SA (left) and SI channels (right). The data were fit with a Boltzmann relation with half‐maximal activation pressures P1/2 of 36.5 and 13 mmHg and steepnesses 6 and 6.5 for SA and SI channels, respectively. Although, stretch‐inactivation is shifted to more negative pressures compared with stretch‐activation, the slopes of the Boltzmann fit are similar. This suggests that both types of gating involve a single energetic process, such as thinning of the lipid membrane adjacent to the channel during membrane deformation (see Section VII).
A
mdx
SA
mdx
SI
No suction Suction No suction 2 pA 400 ms B
mdx SI
mdx SA
0.8 0.6
0.4
0.4 0.2 0.2 0.0 −30
−20 −10 Suction (mmHg)
0
−60
−40 −20 Suction (mmHg)
0
Open probability
Open probability
0.6
0.0
FIGURE 7 Pressure dependence of SA and SI gating. (A) Single‐channel currents showing SA (left) and SI (right) mechanotransduction modes. Recordings made with standard saline in the patch electrode. Top records, channel activity in the absence of applied pressure; middle records, after applying 15 mmHg suction to the electrode; bottom records, after releasing the pressure stimulus to 0 mmHg. (B) Relationship between the pipette pressure and channel open probability for SA (left, n ¼ 6) and SI channels (right, n ¼ 14). Data were fit with a Boltzmann relation with half‐maximal activation pressures P1/2 of 36.5 and 13 mmHg and steepnesses 6 and 6.5 for SA and SI channels, respectively. Adapted from Franco-Obrego´n and Lansman (2002).
477
16. Mechanosensitive Ion Channels in Dystrophic Muscle
2. Stretch‐Induced Gating Mode Transitions The hypothesis that SA and SI gating represent two gating modes of a single type of MS channel was strengthened by the finding that membrane stretch was a suYcient stimulus to cause conversion between the two gating modes. The transition always involved a shift from SA to SI gating and was essentially irreversible over the time course of the recording. Figure 8 (top)
Np0
A 0
B
mdx myotube
1 0 mmHg 2 pA 300 ms
−5 mmHg
−5 mmHg 0 mmHg
−15 mmHg
−15 mmHg 0 mmHg
FIGURE 8 Induction of the SI gating mode by membrane stretch. (A) Channel open probability (Npo) measured in consecutive 300‐ms sweeps. The bars indicate the time during which the indicated pressure stimulus was applied to the patch electrode. (B) Representative current records obtained during the experiment. Npo ¼ 0.04 at the beginning of the experiment, 0.20 after applying 5mmHg of suction, 0.15 after subsequently releasing the pressure stimulus, 0.01 after application of a second suction stimulus of 15mmHg, and 0.10 after releasing the pressure stimulus. Adapted from Franco-Obrego´n and Lansman (2002).
478
JeVry B. Lansman
shows an example of a stretch‐induced transition in MS channel gating mode. Figure 8A shows a plot of channel open probability for consecutive 1‐s sweeps lasting for the entire recording which lasted 10 min. The plot of open probability vs time during the experiment shows that the response of an MS channel to membrane stretch changed. This is seen more clearly in the records in Fig. 8B. For example, applying suction initially increased channel opening (Fig. 8B, second pair, 5 mmHg) and channels remained open. When suction was applied for a second time, channels closed as is characteristic of SI gating.
VI. PHARMACOLOGY A. Block by Gadolinium Ion The lanthanide Gd3þ has an ionic radius of 0.105 nm, close to Ca2þ with an ionic radius of 0.106 nm, and is useful as a transition state analogue for studying Ca2þ‐binding sites in ion channels and other proteins. Gd3þ blocks SA channels with high aYnity (Yang and Sachs, 1989; Franco and Lansman, 1990) but the mechanism of block is not well understood. Gd3þ binds to charged phospholipids with high aYnity and produces strong electrostatic eVects that modify bilayer properties (Ermakov et al., 1998). This has suggested that Gd3þ inhibits MS channels by its eVects on lipids rather than by binding to the MS channel pore. However, an analysis of the block of persistently open SI channels in mdx muscle by Gd3þ indicates that it acts by simply plugging the open channel thereby preventing ion flow (Franco et al., 1991). Figure 9 shows Gd3þ block of high po MS channels in mdx myotubes. Entry and exit of Gd3þ from the open channel is resolved as the discrete interruptions of the single‐channel current. As expected for a simple bimolecular reaction between a single Gd3þ ion and a site in the open channel, increasing the concentration of Gd3þ increased the number of interruptions of the single‐channel current (Fig. 9B). Measurements of the durations of the open and blocked times showed Gd3þ entry into the channel (blocking rate) is insensitive to membrane potential, while Gd3þ exit from the pore (unlocking rate) is faster at negative potentials. An increased rate of unblocking with hyperpolarization indicates Gd3þ binds within the channel and is swept into the cell interior at negative voltages where the applied electric field exceeds the chemical binding energy. If Gd3þ acts within the lipid bilayer to modify channel gating, then it would be expected to change the pressure sensitivity of channel opening, reflecting a mechanism of inhibition at the level of the mechanotransduction process. Measurements were made of the pressure‐open probability relation for SI channels in the presence of Gd3þ. The Boltzmann parameters used to
479
16. Mechanosensitive Ion Channels in Dystrophic Muscle Open channel block A
C Blocking rate (s−1)
10,000
B 0-mM Gd3+
D
1000 −200 −160 −120
−80
−40
0
100 −200 −160 −120 −80 −40 Patch potential (mV)
0
5.0
Unblocking rate (s−1)
10,000 2.5
7.5
1 pA 50 ms
FIGURE 9 Gadolinium block of MS channels. (A) Illustration showing the open channel blocking model. Gd3þ is present in the extracellular solution at a low concentration. The right figure shows an open channel that is freely permeable to Naþ and Ca2þ in the extracellular solution (open circles). Binding of the blocking ion to a site in the channel prevents ion conduction by physically occluding the channel permeation pathway. (B) Single‐channel currents recorded from mdx myotubes with electrodes containing physiological saline and the indicated concentration of Gd3þ. The holding potential was 60 mV. Currents were filtered at 2 kHz. (C) Dependence of the blocking rate on the patch potential. Open symbols represent the inverse of the mean open time (blocking rate) obtained from a single exponential fit to the histograms of open times. Filled symbols are obtained from an analysis of the distribution of current amplitudes. (D) Dependence of the unblocking rate on the patch potential. Filled squares represent the inverse of the mean blocked time obtained from the exponential fit to the histogram of closed times. Open circles are obtained from an analysis of the distribution of current amplitudes. The line through the experimental points is the fit to a single exponential. The unblocking rate changed e‐fold per 85 mV, which corresponds to an eVective electrical distance d ¼ 0.09 for a trivalent blocking particle. Adapted from Franco et al. (1991).
fit the pressure‐open probability curve for the SI channel in the presence of Gd3þ (P1/2 ¼ 17 mmHg and steepness ¼ 6 mmHg) were virtually the same as the mdx SI channels in the absence of blocker (see above). It is likely that Gd3þ exerts its blocking actions within the MS channel pore and, apparently, does not alter the properties of the bilayer suYciently to alter mechanotransduction.
480
JeVry B. Lansman
B. Aminoglycoside Antibiotics Aminoglycoside antibiotics are positively charged molecules that interact with both membrane lipids and ion‐binding sites of proteins. In skeletal muscle, aminoglycosides block both the L‐type channel and MS channel (Haws et al., 1996; Winegar et al., 1996). Block of MS channels occurs in the submillimolar range (KD ¼ 200 mM for neomycin at pH 7.4) and involves a partial occlusion of the channel pore at high concentrations. Figure 10 shows the partial block of Partial occlusion of the open channel
1.0
No drug
0.8 i/imax (−60 mV)
Neomycin 100 mM
100 mM
0.6 0.4 0.2
400 mM
1 pA 50 ms
0 −1 −2 Current (pA)
0.0
1
10
100 1000 Neomycin (mM)
10,000
FIGURE 10 Block of MS channels by aminoglycoside antibiotics. (Top) Illustration showing the molecular interpretation of the partial occlusion blocking model used to describe the substrate blocking behavior. In this model, an aminoglycoside antibiotic can enter the channel conduction pathway; however, it only partially blocks the channel. (Bottom, right) Records showing the block of single MS channels with increasing concentrations of neomycin. Increasing the concentration of neomycin reduces the amplitude of the single‐channel current and causes the transition of the channel to a level that is 40% of the fully open level. The amplitude distribution of the open channel current is shown at the right of the records. There is a progressive reduction in the amplitude of the fully open state and a parallel occupancy of the subconductance level. (Bottom, left) Concentration dependence of the reduction of the single‐channel current by neomycin. The amplitude of the single‐channel current in the presence of drug is normalized to that in the absence of drug (i/imax) for the full conductance state (filled symbols) and the subconductance state (open symbols). The smooth curve drawn through the open symbols is the fit to a model in which drug binding to a single site is modified by the presence of fixed negative charges. Holding potential was 60 mV. Adapted from Winegar et al. (1996).
16. Mechanosensitive Ion Channels in Dystrophic Muscle
481
MS channels in FDB fibers by neomycin. Increasing the concentration of neomycin had two eVects that occurred in parallel: it reduced the amplitude of the single‐channel current and caused the channel to occupy a subconductance state of approximately one‐third the amplitude of the fully open channel. Analysis of the blocking mechanism suggested that only one of the positively charged amino groups on the aminoglycoside molecule entered into the channel pore.
VII. CONCLUSIONS Dystrophin‐deficient muscle shows characteristic cytoskeletal abnormalities relevant to understanding MS channel gating behavior. Ultrastructural studies show that the absence of dystrophin is associated with irregularities of the spectrin cytoskeleton (Williams and Bloch, 1999a,b). For example, there are regions in which spectrin is lost over M lines and in longitudinal strands that leaves discrete areas of the plasma membrane without structural support from the cytoskeleton. We suspect that mechanical disturbances would likely cause changes in membrane composition or structure at such sites, and such changes could underlie the gating mode shifts in dystrophin‐deficient muscle. The localization of spectrin abnormalities to discrete surface domains would be consistent with the observation of gating mode conversion at only a fraction of the recording sites. A switch in MS channel gating from SA to SI has been described for MS channel in other systems, notably Shaker Kþ channels (Gu et al., 2001) and gramacidin A channels incorporated into pure lipid membranes (Martinac and Hamill, 2002). Any number of mechanisms could account for shifts between SA and SI gating but recent work has focused on changes in the extent of hydrophobic mismatch between the bilayer and exterior hydrophobic length of the channel protein, which can influence ion channel conformational state (Cantor, 1994, 1999; Lundbaek and Andersen, 1994). In particular, a change in bilayer thickness that results from accumulation of lipids with longer or shorter acyl chain lengths relative to the embedded channel length can account for the key features of SA and SI gating in mdx muscle (Martinac and Hamill, 2002). The model assumes that channels open when hydrophobic mismatch is minimal. If bilayer thickness is greater than the channel hydrophobic length, membrane stretch would cause the bilayer to thin, thus reducing the mismatch, and produce SA gating. This transduction process would occur primarily in wild‐type muscle, but also in some patches on mdx muscle. A shift to the SI gating mode would occur when there is a change in local lipid composition near the channel that reduces hydrophobic mismatch.
482
JeVry B. Lansman
A reduced mismatch favors opening and channels would remain open in the absence of a pressure stimulus. SI gating occurs since membrane stretch causes bilayer thinning and increased mismatch. Although a change in hydrophobic mismatch provides a simple model for explaining the complex changes in MS channel gating in mdx muscle, the dystrophin‐dependent processes that control bilayer composition and/or thickness are not known. The molecular identity of the MS channel in skeletal muscle is not known. Recent work has indicated that skeletal muscle expresses several members of the TRPC and TRPV channel families, including TRPC1, TRPC4, TRPC6 (Vandebrouck et al., 2002), and TRPV2, TRPV3, TRPV4, and TRPV6 (Kru¨ger et al., 2004). The conductance and selectivity of MS channels in skeletal muscle fall within the range of those for known TRP channels. Several of the TRP channels that are expressed in skeletal muscle are thought to be MS channels including TRPV2 (Iwata et al., 2002), TRPV4 (Liedtke, 2005), and TRPC1 (Maroto et al., 2005). A future goal will be to understand the functional diversity of MS channels in skeletal muscle using transgenic and small interfering RNA methods. References Adams, M. E., Butler, M. H., Dwyer, T. M., Peters, M. F., Murnane, A. A., and Froehner, S. C. (1993). Two forms of mouse syntrophin, a 58 kd dystrophin‐associated protein, diVer in primary structure and tissue distribution. Neuron 11, 531–540. Ahn, A. H., Freener, C. A., Gussoni, E., Yoshida, M., Ozawa, E., and Kunkel, L. M. (1996). The three human syntrophin genes are expressed in diverse tissues, have distinct chromosomal locations, and each bind to dystrophin and its relatives. J. Biol. Chem. 271, 2724–2730. Brehm, P., Kullberg, R., and Moody‐Corbett, F. (1984). Properties of non‐junctional acetylcholine receptor channels on innervated muscle of Xenopus laevis. J. Physiol. 350, 631–648. Cantor, R. S. (1999). Lipid composition and the lateral pressure profile in bilayers. Biophys. J. 76(5), 2625–2639. Eisenman, G. (1962). Cation selective glass electrodes and their mode of operation. Biophys. J. 2, 259–323. Ermakov, Yu., A., Averbakh, A. Z., Arbuzova, A. B., and Sukharev, S. I. (1998). Lipid and cell membranes in the presence of gadolinium and other ions with high aYnity to lipids. 2. A dipole component of the boundary potential on membranes with diVerent surface charge. Membr. Cell Biol. 12(3), 411–426. Ervasti, J. M., and Campbell, K. P. (1993). A role for the dystrophin‐glycoprotein complex as a transmembrane linker between laminin and actin. J. Cell. Biol. 122(4), 809–823. Ervasti, J. M., Ohlendieck, K., Kahl, S. D., Gaver, M. G., and Campbell, K. P. (1990). Deficiency of a glycoprotein component of the dystrophin complex in dystrophic muscle. Nature 345(6273), 315–319. Franco, A., Jr., and Lansman, J. B. (1990a). Stretch‐sensitive channels in developing muscle cells from a mouse cell line. J. Physiol. 427, 361–380. Franco, A., Jr., and Lansman, J. B. (1990b). Calcium entry through stretch‐inactivated ion channels in mdx myotubes. Nature 344, 670–673.
16. Mechanosensitive Ion Channels in Dystrophic Muscle
483
Franco, A., Jr., Winegar, B. D., and Lansman, J. B. (1991). Open channel block by gadolinium ion of the stretch‐inactivated ion channel in mdx myotubes. Biophys. J. 59(6), 1164–1170. Franco‐Obrego´n, A., Jr., and Lansman, J. B. (1994). Mechanosensitive ion channels in skeletal muscle from normal and dystrophic mice. J. Physiol. 481, 299–309. Franco‐Obrego´n, A., and Lansman, J. B. (2002). Changes in mechanosensitive channel gating following mechanical stimulation in skeletal muscle myotubes from the mdx mouse. J. Physiol. 539, 391–407. Gu, C. X., Juranka, P. F., and Morris, C. E. (2001). Stretch‐activation and stretch‐inactivation of Shaker‐IR, a voltage‐gated Kþ channel. Biophys. J. 80(6), 2678–2693. Guharay, F., and Sachs, F. (1984). Stretch‐activated single ion channel currents in tissue‐ cultured embryonic chick skeletal muscle. J. Physiol. 352, 685–701. Haws, C. M., Winegar, B. D., and Lansman, J. B. (1996). Block of single L‐type Ca2þ channels in skeletal muscle fibers by aminoglycoside antibiotics. J. Gen. Physiol. 107(3), 421–432. HoVman, E. P., Brown, R. H., Jr., and Kunkel, L. M. (1987). Dystrophin: The protein product of the Duchenne muscular dystrophy locus. Cell 51, 919–928. Ibraghimov‐Beskrovnaya, O., Ervasti, J. M., Leveille, C. J., Slaughter, C. A., Sernett, S. W., and Campbell, K. P. (1992). Primary structure of dystrophin‐associated glycoproteins linking dystrophin to the extracellular matrix. Nature 355, 696–702. Ibraghimov‐Beskrovnaya, O., Milatovich, A., Ozcelik, T., Yang, B., Koepnick, K., Francke, U., and Campbell, K. P. (1993). Human dystroglycan: Skeletal muscle cDNA, genomic structure, origin of tissue specific isoforms and chromosomal localization. Hum. Mol. Genet. 2, 1651–1657. Iwata, Y., Katanosaka, Y., Arai, Y., Komamura, K., Miyatake, K., and Shigekawa, M. (2002). A novel mechanism of myocyte degeneration involving the Ca2þ permeable growth factor‐ regulated channel. J. Cell Biol. 161(5), 957–967. Koenig, M., Monaco, A. P., and Kunkel, L. M. (1988). The complete sequence of dystrophin predicts a rod‐shaped cytoskeletal protein. Cell 53(2), 219–226. Kru¨ger, J., Kunert‐Keil, C., and Brinkmeir, H. (2004). RNA transcripts coding for members of the TRP cation family in mouse skeletal and heart muscle. Deutsche Physiol. Gessel. Abstracts. Liedtke, W. (2005). TRPV4 plays an evolutionary conserved role in the transduction of osmotic and mechanical stimuli in live animals. J. Physiol. 567(Pt. 1), 53–58. Lundbaek, J. A., and Andersen, O. S. (1994). Lysophospholipids modulate channel function by altering the mechanical properties of lipid bilayers. J. Gen. Physiol. 104(4), 645–673. Maroto, R., Raso, A., Wood, T. G., Kurosky, A., Martinac, B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nat. Cell. Biol. 7(2), 179–185. Martinac, B., and Hamill, O. P. (2002). Gramicidin A channels switch between stretch activation and stretch inactivation depending on bilayer thickness. Proc. Natl. Acad. Sci. USA 99(7), 4308–4312. Pasternak, C., Wong, S., and Elson, E. L. (1995). Mechanical function of dystrophin in muscle cells. J. Cell Biol. 128, 355–361. Sicinski, P., Geng, Y., Ryder‐Cook, A. S., Barnard, E. A., Darlison, M. G., and Barnard, P. J. (1989). The molecular basis of muscular dystrophy in the mdx mouse: A point mutation. Science 244, 1578–1580. Vandebrouck, C., Martin, D., Colson‐Van Schoor, D., Debaix, H., and Gailly, P. (2002). Involvement of TRPC in the abnormal calcium influx observed in dystrophic (mdx) mouse skeletal muscle fibers. J. Cell Biol. 158(6), 1089–1096. Williams, M. W., and Bloch, R. J. (1999a). Extensive but coordinate reorganization of the membrane skeleton in myofibers of dystrophic (mdx) mice. J. Cell. Biol. 144(6), 1259–1270.
484
JeVry B. Lansman
Williams, M. W., and Bloch, R. J. (1999b). DiVerential distribution of dystrophin and beta‐ spectrin at the sarcolemma of fast twitch skeletal muscle fibers. J. Muscle Res. Cell Motil. 20(4), 383–393. Winegar, B. D., Haws, C. M., and Lansman, J. B. (1996). Subconductance block of single mechanosensitive ion channels in skeletal muscle fibers by aminoglycoside antibiotics. J. Gen. Physiol. 107(3), 433–443. Yang, X. C., and Sachs, F. (1989). Block of stretch‐activated ion channels in Xenopus oocytes by gadolinium and calcium ions. Science 243, 1068–1071. Yoshida, M., and Ozawa, E. (1990). Glycoprotein complex anchoring dystrophin to sarcolemma. J. Biochem. (Tokyo) 108(5), 748–752.
CHAPTER 17 MscCa Regulation of Tumor Cell Migration and Metastasis Rosario Maroto and Owen P. Hamill Department of Neuroscience and Cell Biology, University of Texas Medical Branch, Galveston, Texas 77555
I. Overview II. Introduction III. DiVerent Modes of Migration A. Amoeboid Migration B. Mesenchymal Migration C. Collective Cell Migration D. Mechanisms for Switching Migration Modes IV. Ca2þ Dependence Of Cell Migration A. Measuring [Ca2þ]i B. Identifying Ca2þ Influx Pathways C. Ca2þ Dependence of Amoeba Locomotion D. Ca2þ Dependence of Vertebrate Cell Amoeboid Migration E. The Role of [Ca2þ]i Gradients and Transients in Mesenchymal Cell Migration V. The Role of MscCa in Cell Migration VI. Can Extrinsic Mechanical Forces Acting on MscCa Switch on Cell Migration? References
I. OVERVIEW The acquisition of cell motility is a required step in order for a cancer cell to migrate from the primary tumor and spread to secondary sites (metastasize). For this reason, blocking tumor cell migration is considered a promising approach for preventing the spread of cancer. However, cancer cells like normal cells can migrate by several diVerent modes referred to as Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59019-2
486
Maroto and Hamill
‘‘amoeboid,’’ ‘‘mesenchymal,’’ and ‘‘collective cell.’’ Furthermore, under appropriate conditions a single cell can switch between modes. A consequence of this plasticity is that a tumor cell may be able to avoid the eVects of an agent that targets only one mode by switching modes. Therefore, a preferred strategy would be to target mechanisms that are shared by all modes. Here we review the evidence that Ca2þ influx via the mechanosensitive Ca2þ‐permeable channel (MscCa) is a critical regulator of all modes of cell migration and therefore represents a very good therapeutic target to block metastasis.
II. INTRODUCTION Cancer is a multistep process that results in a normal cell, often an epithelial cell lining a gland, duct, or organ surface, undergoing abnormally increased multiplication to produce a localized primary tumor that with time invades and spreads (metastasizes) to surrounding tissues and eventually causes death. However, in order for a tumor to metastasize, the tumor cell must migrate from the primary tumor, pass through blood vessels, penetrate into the secondary tumor site, and migrate through the tissue to establish a metastasis. Therefore, the acquisition of cell motility is a necessary although not a suYcient step for tumor invasion and metastasis, which also require the additional steps of barrier matrix breakdown, tumor cell adherence, growth, and angiogenesis at the secondary sites. Nevertheless, because metastasis will only be achieved if the tumor cell completes every step in the metastatic cascade, identifying the most sensitive and susceptible step that regulates tumor cell migration should provide a promising target to block metastasis (Grimstad, 1987; Stracke et al., 1991; Kassis et al., 2001). There are currently two models used to explain tumor progression to the metastatic disease. One is the traditional ‘‘multi‐hit’’ genetic model that proposes a sequence of mutations that triggers the various stages of cancer (e.g., initiation, promotion) with the final mutation(s) promoting increased tumor cell invasiveness and metastasis (Emmelot and Scherer, 1977; Cahill et al., 2000; Hanahan and Weinberg, 2000; Zhou et al., 2005). Evidence supporting this model includes the existence of several stable human tumor cell lines that demonstrate high invasiveness when implanted in animals (Kaighn et al., 1979; Sung et al., 1998), and the recent discovery that many primary tumor cells already express a genetic signature that predicts their metastatic potential (Ramaswamy et al., 2003; Varambally et al., 2005). The second model is an epigenetic one based on the discovery that growth factors that trigger the epithelial–mesenchymal transition (EMT), in which nonmotile epithelial cells are converted into motile mesenchymal cells (e.g., during
17. MscCa and Metastasis
487
normal embryogenesis and wound healing), are also released by stromal cells surrounding the tumor and promote increased tumor cell invasiveness and metastasis (Thiery, 2002; Thompson and Newgreen, 2005; but see Tarin, 2005). Specific cancers may utilize one or a combination of the two mechanisms since the mechanisms are not exclusive (e.g., one aspect of the metastatic genetic signature may include the potential to undergo EMT). In any case, the regulatory molecules involved in transforming a tumor cell from a nonmotile to a motile phenotype need to be identified. In this chapter we focus on the role of the MscCa, which is identified as a member of the transient receptor potential channel family (Maroto et al., 2005; Saimi et al., 2007) and shown to be essential for prostate tumor cell migration (Maroto et al., 2007). Because MscCa is expressed by both nonmotile and motile cells, we review the evidence for the idea that changes in MscCa properties triggered by events associated with cancer progression may contribute to increased tumor invasiveness and metastasis.
III. DIFFERENT MODES OF MIGRATION Normal cells and tumor cells move according to one of three major modes of migration referred to as ‘‘amoeboid,’’ ‘‘mesenchymal,’’ and ‘‘collective cell.’’ Furthermore, under specific circumstances a single cell can switch between these modes (Friedl and Wolf, 2003; Sahai and Marshall, 2003; Friedl, 2004; Wolf and Friedl, 2006). Because of this plasticity, a tumor cell may be able to avoid the eVects of an agent that blocks only one migratory mode by switching to another mode. Therefore, a preferred strategy would be to identify and target molecular mechanisms that are shared by all modes. With this in mind, we consider the diVerent modes of migration, their similarities and diVerences, and in particular their possible common dependence on Ca2þ influx via MscCa.
A. Amoeboid Migration Amoeboid movement is expressed by a variety of invertebrate and vertebrate cells, but has been the most intensely studied in the amoeba Dictyostelium discoideum. This cell displays an ellipsoidal profile with either a monopodal or polypodal form, and undergoes a rapid (e.g., >20 mm/min) gliding movement that involves repetitive cycles of protrusion and contraction with little adhesiveness to the substrate. This lack of adhesiveness is consistent with the absence of integrin expression by the amoeba (Friedl, 2004). The amoeba uses two mechanically distinct mechanisms to push itself
488
Maroto and Hamill
forward (Yoshida and Soldati, 2006) a filopodia–lamellipodia mechanism that depends on actin polymerization and a bleb mechanism in which a local region of membrane where the cortical‐CSK has been disrupted is pushed outward by cytoplasmic pressure generated by myosin II. Both protrusion mechanisms involve significant mechanical distortions of the membrane at the front of the cell that could activate MscCa to provide feedback (via Ca2þ influx and/or membrane polarization) between the force‐generating mechanisms and resultant membrane distortions. Neutrophils, eosinophils, lymphocytes, stem cells, and specific tumor cells associated with leukemia, lymphoma, and small cell lung carcinoma also display amoeboid movement. Furthermore, specific cell types that display a mesenchymal mode of migration when crawling on a two‐dimensional (2D) substrate can switch to an amoeboid mode when migrating through a 3D substrate (Friedl, 2004). Vertebrate cells undergoing amoeboid migration also display both blebbing and filopodia–lamellipodia mechanisms of forward protrusion (Sahai and Marshall, 2003; Blaser et al., 2006). Fish and amphibian keratocyes may represent a hybrid form of amoeboid/mesenchymal locomotion because they normally show a smooth gliding movement but also express a broad flat lamellipodium. Furthermore, when they become stuck on their substrate they tend to pull out a rear tether and display a more discontinuous ‘‘mesenchymal‐like’’ locomotion (Lee et al., 1999). Interestingly, an amoeba can be induced to develop a broad lamellipodium and undergo keratocyte‐like migration by knocking out a gene that regulates the amoeba’s aggregation process (Asano et al., 2004). However, a double knockout of myosin II and the aggregation gene does not block keratocyte‐like migration, indicating that myosin II may be dispensable for this mode of movement.
B. Mesenchymal Migration Mesenchymal movement is displayed by fibroblasts, neurons, smooth muscle, and endothelial cells, as well as by specific cancer cells from epithelial tumors, gliomas, and sarcomas. In this mode, the cell typically displays a highly polarized morphology with a front lamellipodium, immediately behind which is the lamella, followed by the cell body with the nucleus, and usually ending with a rear tail or tether. Compared with the smooth, gliding amoeboid movement, mesenchymal migration is relatively discontinuous and slower (<1 mm/min) because of its greater adhesiveness and strong dependence on integrin engagement and disengagement from the substrate. Mesenchymal migration can be divided into five steps involving: (1) forward protrusion of the cell’s leading edge, (2) formation of adhesions at the front
17. MscCa and Metastasis
489
of the cell with the extracellular matrix (‘‘gripping’’), (3) pulling against the ECM via the cell adhesions as the myosin–cytoskeleton (CSK) contracts and exerts traction force against the substrate, (4) progressive stretching of the cell as the traction force develops at the cell front and pulls against the cell rear, and (5) finally, detachment of the rear adhesions from the ECM allowing net cell displacement and relaxation of membrane stretch (LauVenburger and Horwitz, 1996; Sheetz et al., 1999; Ridley et al., 2003). The important aspect of this mode of migration in relation to MscCa is that the membrane bilayer of the whole cell will tend to experience a slow ramp of increasing tension for as long as the rate of forward protrusion exceeds the rate or rear retraction (Lee et al., 1999; Maroto et al., 2007).
C. Collective Cell Migration In the collective cell mode of migration, the cells are connected by cell junctions formed by cadherins and integrins, and move in a mass with the motile cells at the leading invasive edge generating the adhesion and traction forces (likely via the mesenchymal mode) that tend to pull the rear nonmotile tumor cells along passively. This pattern of migration represents the predominate migration mode for most epithelial cancers in situ, and provides the advantage of increased heterogeneity by allowing nonmotile, proliferating cells along with motile path‐finding cells to invade the new tissues (Friedl and Wolf, 2003; Wolf and Friedl, 2006).
D. Mechanisms for Switching Migration Modes Cells that normally express mesenchymal and/or collective cell migration can be converted to the amoeboid mode by reducing the eVectiveness of integrin‐ECM adhesion (i.e., with integrin‐blocking antibodies or arginine‐ glycine aspartate (RGD) peptides that compete for integrin‐ECM‐binding sites), by blocking matrix proteases, or by stimulating the Rho‐associated serine/threonine kinase (ROCK) that increases cortical contraction, thereby promoting cell rounding and forward protrusion by membrane blebbing (Friedl, 2004). With this switch, the cell becomes more deformable due to its lack of adhesiveness and can now squeeze between matrix barriers. This lessens the dependence on the actions of matrix‐degrading metalloproteinases and increases resistance to metalloproteinase inhibitors. The weakened dependence on integrin adhesion also results in a loss of dependence on calpain proteolytic cleavage important for integrin‐linked adhesion turnover (Carragher et al., 2005). In neutrophils, rear integrins tend to be endocytosed
490
Maroto and Hamill
rather than dissembled by calpain activity, and in contrast to mesenchymal cells, inhibition of calpain actually promotes, rather than inhibits, migration by enhancing cell protrusion and cell spreading (Lokuta et al., 2003). On the other hand, amoeboid movement retains a strong dependence on myosin II contractility as indicated by increased sensitivity to ROCK inhibition (Sahai and Marshall, 2003). Since that both calpain and myosin II are Ca2þ sensitive, one would expect that both modes of migration would display Ca2þ dependence. Another mechanism that appears to promote mode switching relates to the relocation of cavoelin‐1 (Cav‐1), a lipid raft‐ associated protein that colocalizes with MscCa/TRPC1 (Lockwich et al., 2000; Brazier et al., 2003; Maroto et al., 2005). For example, when endothelial cells switch from migration in a 2D to a 3D matrix there is a redistribution of Cav‐1, and possibly MscCa, from the back to the front of the cell (Parat et al., 2003). As described below, this shift would be consistent with intracellular [Ca2þ] ([Ca2þ]i) transients being initiated in the front of the amoeboid like neutrophils (Kindzelskii et al., 2004) but in the rear of mesenchymal‐like cells (Maroto et al., 2007). IV. Ca2þ DEPENDENCE OF CELL MIGRATION Although a variety of signaling pathways may regulate cell migration, Ca2þ signaling has always been considered a significant player because many of the eVector molecules that mediate migration are Ca2þ sensitive, including myosin light chain kinase (i.e., that regulates myosin II), calpain, gelsolin, a‐actinin, and phosphatase (calcineurin) and integrins (Hendey and Maxfield, 1993; Arora and McCulloch, 1996; Eddy et al., 2000; Mamoune et al., 2003; Franco and Huttenlocher, 2005). The Ca2þ regulatory role has been reinforced by the finding that a variety of Ca2þ transport proteins including pumps, exchangers, and various gated Ca2þ channels can modulate cell migration (Dreval et al., 2005). A. Measuring [Ca2þ]i The most convenient and common method used to measure [Ca2þ]i involves using fluorescent microscopy and Ca2þ‐sensitive fluorescent dyes like fura‐2 and its membrane permeable form fura‐2 AM (Grynkiewicz et al., 1985). The main advantage of the approach is that changes in [Ca2þ]i can be monitored while simultaneously measuring cell migration (i.e., by time‐lapse videomicrosopy). As a consequence, one can relate specific spatio‐temporal changes in [Ca2þ]i to specific events occurring during migration. However,
17. MscCa and Metastasis
491
there are also some practical limitations associated with the method, including the diYculty of detecting local vs global [Ca2þ]i changes and the possibility of compartmentalization of the dyes in organelles. The first limitation has been somewhat overcome by recent technical developments that includes the use of total internal reflectance fluorescence microscopy that oVers added spatial resolution to allow detection of single‐channel [Ca2þ]i fluctuations at the ventral membrane surface adhering with the glass surface (Demuro and Parker, 2005). In addition, the development of Ca2þ‐sensor ‘‘cameleons’’ that operate by fluorescence energy transfer and can be targeted to the plasma membrane or the ER can be used to measure [Ca2þ]i changes in these membrane microdomains (Miyawaki et al., 1997; Isshiki et al., 2002). In the case of fura‐2 compartmentalization, there are discrepant views on its occurrence and significance. For example, one group has proposed that the apparent [Ca2þ]i gradient seen in T lymphocytes is due to fura‐2 accumulation in mitochondria (Quintana and Hoth, 2004), whereas another group found that the [Ca2þ]i gradient seen in fibroblasts was not associated with mitochondria but instead colocalized with the Golgi apparatus in the perinuclear region (Wahl et al., 1992). A further complication is that mitochondria are motile, and their motility varies inversely with [Ca2þ]i so that they move fastest in lower [Ca2þ]i (100‐300 nM) but stop movement in higher [Ca2þ]i (i.e., 1 mM) (Yi et al., 2004). As a consequence, one would expect mitochondria to migrate up a [Ca2þ]i gradient and accumulate in regions of highest [Ca2þ]i where they may function as Ca2þ buVers and/or prevent the spread of local [Ca2þ]i transients (Tinel et al., 1999; Yi et al., 2004; Levina and Lew, 2006). However, in apparent contradiction of this idea, mitochondria accumulate in the lamellipodium of migrating fibroblasts and prostate tumor cells (DeBiasio et al., 1987; Maroto et al., 2007), and yet these cells develop a global [Ca2þ]i gradient that increases from front to back of the cell (Hahn et al., 1992; Matoto et al., 2007). The stimulus that promotes this accumulation remains unclear but could involve the added requirement for ATP and/or an elevated [Ca2þ]i in membrane subdomains of the lamellipodium. In any case, it would appear that compartmentalization of fura‐2 dye cannot alone explain the sustained, and in some cases rapidly reversible, [Ca2þ]i gradients seen in a variety of migrating cells (see Section IV.E.2). B. Identifying Ca2þ Influx Pathways The simplest method to demonstrate a requirement for Ca2þ influx is to show that migration requires the presence of external Ca2þ (Strohmeier and Bereiter‐Hahn, 1984). Patch clamp recording can then be used to characterize the kinetics, conductance, surface distribution, and pharmacological
492
Maroto and Hamill
properties of the Ca2þ channels expressed in the migrating cell (Lee et al., 1999; Maroto et al., 2007). With this knowledge one can then use various treatments to relate particular [Ca2þ]i changes to specific Ca2þ channels activities. One perceived practical limitation of patch clamping is that channel current measurements are restricted to the dorsal surface because it is not possible to patch the ventral ‘‘adherent’’ surface, at least with the traditional patch clamp method (Hamill et al., 1981). In this case, one might argue that because CSK‐generated mechanical (traction) forces are transmitted to the substrate purely at ventral surface adhesions, then only mechanosensitive processes in these sites will experience mechanical force and become activated (Mobasheri et al., 2002). However, the traction forces that pull on the substrate via the ventral surface adhesions will also tend to stretch the whole cell for as long as the rear of the cell remains firmly attached to the substrate. Apart from causing the cell to become extended, there are other manifestations of these stretching forces including the smoothing out of membrane folds and microvilli in spreading cells (Erickson and Trinkhaus, 1976), an elastic recoil seen occasionally in some migrating cells as presumably stretching forces exceed adhesive forces (Mandeville and Maxfield, 1997), and even cell rupture/fragmentation that can occur when cell retraction is blocked and the pulling forces exceed the elastic limits of the bilayer (Verkhovsky et al., 1989). Galbraith and Sheetz (1999) have elegantly and directly addressed the issue of force distribution on the ventral and dorsal surfaces by using optical tweezers to measure the membrane tension on the dorsal membrane, and a micromachined device to measure tension generated on the ventral membrane. Their measurements indicate that the dorsal matrix is as eVectively linked to the force‐generating CSK as the ventral adhesions so tension‐ sensitive channels located in both the dorsal and ventral surfaces should experience the same stretch. In this case, the MscCa properties measured on the dorsal surface (i.e., their gating kinetics and subsurface distribution) should be important in defining the [Ca2þ]i dynamics measured during cell migration (Maroto et al., 2007). C. Ca2þ Dependence of Amoeba Locomotion One of the earliest observations implicating Ca2þ in amoeboid migration was that lanthanum, a known Ca2þ channel inhibitor, blocked movement of Amoeba discoides (Hawkes and Hoberton, 1973). Subsequently, microinjection of aequorin (a photoprotein that emits light on Ca2þ binding) was used to demonstrate a sustained [Ca2þ]i elevation in the tail of the amoeba, as well as transient Ca2þ influxes in the tips of advancing pseudopods—lowering external [Ca2þ]o did not immediately reduce rear [Ca2þ]i but it did block
17. MscCa and Metastasis
493
continued migration (Taylor et al., 1980). This was interpreted as indicating that rear [Ca2þ]i can be maintained by Ca2þ release from internal stores, but migration is more sensitive to Ca2þ influx into the pseuodopod tips (Taylor et al., 1980). In another study, direct injection of fura‐2 was used to show that monopodal amoebae developed a continuous [Ca2þ]i gradient increasing from front to rear, whereas polypodal amoebae showed a decrease in [Ca2þ]i in extending pseudopodia, and an increase in retracting pseudopodia (Gollnick et al., 1991; Yumura et al., 1996). Subsequently, intracellular BAPTA, a fast Ca2þ buVer, was shown to reduce cell spreading, pseudopodia formation, and amoebae locomotion, and these eVects could be reversed by raising [Ca2þ]o (Unterweger and Schlatterer, 1995). On the other hand, the same study found that chelation of [Ca2þ]o by the relatively slow Ca2þ buVer EGTA did not block pseudopod formation, although it did block the development of any [Ca2þ]i gradient and cell migration. Nebl and Fischer (1997) used recombinant aequorin to demonstrate that chemoattractants induced an increase in [Ca2þ]i that was entirely dependent on Ca2þ influx, and speculated that Ca2þ‐induced actin depolymerization in the rear acted to prevent the formation of stable pseudopod formation in this region of the cell. [Ca2þ]o was shown to be required for shear‐flow‐induced amoebae motility (but not directionality) and that addition of either EGTA or Gd3þ stopped cell movement (Fache et al., 2005). In this case, the eVects of external Ca2þ were shown to stimulate cell speed by increasing the amplitude, but not the frequency, of both protrusion and retraction events at the cell’s leading edge (Fache et al., 2005). Another study based on mutants lacking two major Ca2þ‐binding proteins in the ER (calreticulum and calnexin) concluded that chemotaxis depended on both Ca2þ influx and Ca2þ‐ induced Ca2þ release from internal stores (Fisher and Wilczynska, 2006). Despite the above results, there are also several studies that seem to discount a critical role for Ca2þ in amoeboid migration. For example, based on normal chemotaxis seen in a mutant amoeba lacking an IP3‐like receptor, it was concluded that Ca2þ signaling was not required for chemotaxis (Traynor et al., 2000). However, diVerent groups studying the same mutant found that [Ca2þ]i transients dependent on Ca2þ influx were not only retained but were required for both chemotaxis and electrotaxis (Schaloske et al., 2005; Shanley et al., 2006). In a diVerent study, it was reported that amoebae can continue their random locomotion with the same speed in the absence of [Ca2þ]o and the presence of 50‐mM EGTA or EDTA, apparently ruling out any role for Ca2þ influx (Korohoda et al., 2002). However, a more trivial explanation may relate to inadvertent Ca2þ leaching from the low profile glass chamber in which both the ventral and dorsal surfaces of the migrating cell make close contact with the glass. Under these conditions, Ca2þ may build up in the narrow gaps between the adherent cell and glass
494
Maroto and Hamill
surfaces and reach levels (1 mM) suYcient to support migration (Fisher and Wilczynska, 2006). A similar phenomenon may also account for the apparent lack of external Ca2þ dependence of human leukocyte locomotion when they are ‘‘chimneying’’ between closely apposed glass slide and cover slip (Malawista and Boisfleury‐Chevance, 1997). In summary, while most studies indicate that both Ca2þ influx and [Ca2þ]i elevations are required for an amoeba to migrate, the exact role of Ca2þ influx in forward protrusion and rear retraction needs to be better defined. There also remains the unresolved issue on whether the reports of amoeba’s migration in the absence of [Ca2þ]o are real or artifactual. In particular, it will be interesting to test whether migration by chimneying is retained in the presence of the faster Ca2þ‐buVering capacity of BAPTA. D. Ca2þ Dependence of Vertebrate Cell Amoeboid Migration Newt neutrophils, which are relatively large (100 mm in diameter) and comparable in size to an amoeba, develop a sustained [Ca2þ]i gradient that increases from front to rear of the cell as they migrate. Furthermore, spontaneous changes in [Ca2þ]i gradient direction result in changes in migration direction (Brundage et al., 1991; Gilbert et al., 1994). In contrast, the smaller human neutrophils do not develop a detectable [Ca2þ]i gradient but instead display [Ca2þ]i transients when migrating on adhesive substrates (e.g., polylysine, fibronectin, or vitronectin), but not on nonadhesive substrates (Marks and Maxfield, 1990; Hendey and Maxfield, 1993). These [Ca2þ]i transients can be blocked, along with neutrophil migration, by either removing [Ca2þ]o or buVering [Ca2þ]i. The [Ca2þ]i‐buVered neutrophils apparently become immobilized because they are unable to retract their rear, which remains anchored to the adhesive substrate. However, they are still capable of spreading, assuming a polarized morphology, and extending their plasma membrane. Furthermore, their motility can be restored by using RGD peptides to block specific integrin attachments to the substrate. Since a similar block of motility could be induced by inhibitors of the Ca2þ‐dependent phosphatase, calcineurin, it was proposed that this enzyme mediated Ca2þ ‐dependent detachment of the integrin–substrate adhesions (Hendey and Maxfield, 1993). However, the same group latter suggested that a more general mechanism for rear detachment may involve Ca2þ‐increased myosin II contractility (Eddy et al., 2000). A similar Ca2þ and RGD sensitivity was seen for neutrophils migrating through a 3D matrigel substrate (Mandeville and Maxfield, 1997), whereas neutrophils migrating on nonadhesive substrates (e.g., glass in the presence of albumin/serum or through cellulose filters) did not display Ca2þ transients nor did they require the presence of
17. MscCa and Metastasis
495
external Ca2þ or elevations in [Ca2þ]i in order to migrate (Zigmond et al., 1988; Marks and Maxfield, 1990; Hendey and Maxfield, 1993; LaVafian and Hallet, 1995; Alterafi and Zhelev, 1997). A similar phenomena may occur in the normally gliding fish keratocytes that show an increased frequency of [Ca2þ]i transients when their rear becomes transiently stuck on the substrate (Lee et al., 1999). An apparently diVerent role for Ca2þ signaling involves Ca2þ influx‐mediated ‘‘priming’’ of nonmotile eosinophils that enables them to undergo transepithelial migration. However, once the cells are primed, they can migrate in the absence of [Ca2þ]o, although they still depend on [Ca2þ]i elevations (Liu et al., 1999, 2003). In summary, some of the discrepancies in the Ca2þ dependence of neutrophil migration may arise through diVerences in substrate adhesiveness with the strongest Ca2þ dependence seen on sticky substrates, but little or no Ca2þ dependence on nonadhesive substrates. At least in this respect, vertebrate cells that display the amoeboid mode may diVer from the amoeba, which retains Ca2þ dependence even though the amoeba does not depend on integrin adhesion. At least for human neutrophils, [Ca2þ]i transients rather than gradients appear to be more important in regulating cell migration by promoting rear retraction possibly by increased adhesion disassembly via increases in calcineurin, MLCK, and/or calpain activity. E. The Role of [Ca2þ]i Gradients and Transients in Mesenchymal Cell Migration Cells migrating in the mesenchymal mode can also display sustained [Ca2þ]i gradients and/or fast transients. Since these diVerent spatio‐temporal [Ca2þ]i dynamics may regulate diVerent steps associated with the mesenchymal migratory cycle, they will be discussed separately below. 1. A Model for Sustained [Ca2þ]i Gradients A basic question from the onset is how any cell can maintain a sustained [Ca2þ]i gradient for as long as several hours in a cytoplasm that allows free diVusion of Ca2þ. In particular, the existence of any stable regions of diVerent [Ca2þ]i within a continuous aqueous medium would seem to disobey the second law of thermodynamics according to which solutes should passively diVuse down their concentration gradient until they reach equilibrium—in the case of Ca2þ, this equilibration should occur in seconds or at most minutes. To explain this apparent paradox, Braiman and Priel (2001) proposed that the cell uses energy to actively take up Ca2þ uptake into internal stores that can then be passively allowed to leak out into localized regions of the cytoplasm. By this process, combined with a polarized distribution of Ca2þ release
496
Maroto and Hamill
channels on a contiguous ER Ca2þ store, the cell could create a sustained [Ca2þ]i elevation in specified subdomains of the cell (Petersen et al., 2001). The interesting aspect of this model is that, one could have uniform Ca2þ influx across the cell surface and uniform active uptake by the internal Ca2þ stores as long as there was a gradient of Ca2þ release from the stores. A further prediction of this model is that if both active uptake and passive leak occur in very close proximity of the membrane, then a subcortical membrane domain of elevated [Ca2þ]i could be maintained that might go undetected by techniques that only measure global [Ca2þ]i. 2. [Ca2þ]i Gradients Determine Migrational Directionality In several cells undergoing mesenchymal migration, [Ca2þ]i gradients have been shown to be important in determining migration directionality. In particular, Xu et al. (2004) observed that migrating cerebellar granule cells develop a [Ca2þ]i gradient (low front–high back) according to their migration direction. Furthermore, experimental reversal of the [Ca2þ]i gradient by the application to the front of the cell, an external gradient of various agents that cause [Ca2þ]i elevation (e.g., chemo‐repellant slit2, acetylcholine, and ryanodine) was found to be accompanied by a reversal in migration direction. Similarly, if an external gradient of BAPTA‐AM was applied to the back of the cell, again the [Ca2þ]i gradient and migration direction was reversed. Although some of the same neurons also displayed occasional [Ca2þ]i transients, no causal relationship was noted between the transients and migration direction (Xu et al., 2004). Similar [Ca2þ]i gradients related to migration direction have been seen in migrating fibroblasts, kidney epithelial tumor cells, vascular endothelial cells, and prostate tumor cells (Hahn et al., 1992; Schwab et al., 1997; Kimura et al., 2001; Maroto et al., 2007). Moreover, Schwab and colleagues have proposed that the relatively high Ca2þ‐ activated Kþ activity evident in the rear of migrating kidney epithelial tumor cells was a direct consequence of a [Ca2þ]i gradient rather than polarized surface expression of the Kþ channels (Schwab et al., 1995, 2006). They also proposed that the underlying basis for the [Ca2þ]i gradient was due to a combination of higher density of Ca2þ influx pathways and ER [Ca2þ]i stores in the cell body compared with the lamellipodia (Schwab et al., 1997). Studies of the highly motile prostate tumor cell line, PC‐3, have confirmed some of these ideas (Maroto et al., 2007). [Ca2þ]i gradients are seen not only in migrating cells but also in polarized exocrine acinar gland cells where they may regulate unidirectional fluid secretion. In particular, a time‐dependent reversal of the [Ca2þ]i gradient from the luminal to blood side of the acinar cell after acetylcholine (ACh) stimulation has been proposed to be the main basis for a push‐pull model for unidirectional fluid secretion (Kasai and Augustine, 1990). In this model,
17. MscCa and Metastasis
497
[Ca2þ]i elevation, first on the luminal cytoplasmic side of the cell causes Cl and water eZux into the lumen, then [Ca2þ]i elevation on the blood side of the cell causes Cl and water influx from the blood side. Although both cell surfaces express the same Ca2þ‐activated Cl channel, the depolarization that follows ACh stimulation shifts the Cl driving force from eZux to influx. A somewhat similar mechanism could presumably underlie the role of ion and water movements in coordinating cell locomotion (Schwab et al., 2006). This possibility seems to be reinforced by the demonstration that aquaporins are selectively expressed in the leading edge of migrating cells (Saadoun et al., 2005). A quite diVerent cell function related to a sustained [Ca2þ]i gradient involves tip growth of fungi in which elevated [Ca2þ]i in the growing tip has been proposed to promote increased insertion of new membrane via exocytosis (Silverman‐Gavrila and Lew, 2003). This mechanism would seem unlikely to account for migration directionally since exocytosis predominates at the cell front while endocytosis occurs mainly at the cell rear (Bretscher and Aguado‐Velasco, 1998). A more plausible eVect of the [Ca2þ]i gradient in promoting cell migration would be to induce polarization of the activities of enzymes regulating actin polymerization/ depolymerization, integrin activation/assembly/disassembly, and myosin II contractility (LauVenburger and Horwitz, 1996; Sheetz et al., 1999; Ridley et al., 2003). 3. [Ca2þ]i Transients [Ca2þ]i transients have been associated with an even wider variety of other processes including fertilization, cell diVerentiation, exocytosis, muscle contraction, phagocytosis, and neuronal outgrowth and migration (Berridge et al., 2003). This may be because a [Ca2þ]i transient provides a more eYcient and safe way to achieve high levels of [Ca2þ]i compared with steady‐state elevations. Furthermore, the temporal component of the signal provides an added dimension in terms of encoding information. [Ca2þ]i transients can take a number of forms in motile cells—they can be highly localized and associated with pseudopod (or bleb) protrusion or retraction, they can spread throughout the cell as a regenerative [Ca2þ]i wave, or they can circumnavigate the perimeter of a cell in a clockwise or anticlockwise direction (Kindzelskii et al., 2004). [Ca2þ]i transients can be generated spontaneously or can be induced experimentally by electrical, chemical, and mechanical stimuli. In particular, it has been shown that direct mechanical stretch of fibroblasts and keratocytes, and osmotic swelling of endothelial cells can induce [Ca2þ]i transients (Arora et al., 1994; Oike et al., 1994; Lee et al., 1999; Wu et al., 1999). [Ca2þ]i transients may also have diVerent initiation sites on diVerent cells and these site may vary within a single cell during the course of the migratory cycle. In particular, the initiation sites of
498
Maroto and Hamill
[Ca2þ] transients have been related to the distribution of membrane rafts and caveolae (i.e., invaginated membrane structures), which contain the molecular signaling machinery required for Ca2þ signaling, and can undergo redistribution during migration and specific forms of stimulation. Membrane raft‐ and caveolae‐dependent Ca2þ signaling has been observed in cells undergoing both mesenchymal migration (Manes et al., 1999; Isshiki et al., 2002; Parat et al., 2003; Rizzo et al., 2003) and amoeboid migration (Gomez‐Mouton et al., 2001; Pierini et al., 2003; Kindzelskii et al., 2004). For example, Isshiki et al. (2002) found that the caveolae in quiescent endothelial cells are clustered around the edge of the cell but when stimulated to migrate, either by wounding a cell monolayer or by exposing the cells to laminar shear stress, the caveolae move to the trailing edge of the cell, concomitant with this relocation the sites of Ca2þ waves initiation move to the same location (see also Rizzo et al., 2003; Beardsley et al., 2005). In contrast, in human neutrophils lipid rafts and [Ca2þ]i transient initiation sites have been localized to the leading edge of the migrating cells, and cholesterol depletion, which disrupts raft structure, was found to block both [Ca2þ]i transient initiation and cell migration (Manes et al., 1999; Kindzelskii et al., 2004). Some insight into the diVerent results may be related to the demonstration that both the leading edge and rear of lymphocytes are enriched in lipid components that partition into diVerent raft‐like domains (Gomez‐Mouton et al., 2001) and that Cav‐1, a raft maker, shows a diVerent polarized distribution in endothelial cells depending on whether the cells were migrating on 2D substrate or through a 3D matrix (Parat et al., 2003). In particular, Cave‐1 moves from the cell’s rear to the cell’s front during the switch from the 2D/mesenchymal to the 3D/amoeboid migration modes. These findings are highly intriguing giving that TRPC1, a structural subunit of MscCa (Maroto et al., 2005), colocalizes with Cave‐1‐associated membrane lipid rafts (Lockwich et al., 2000; Brazier et al., 2003) and has been localized at the leading edge of migrating neutrophils (Kindzelskii et al., 2004) and the rear of migrating prostate tumor cells (Maroto et al., 2007). Together these results indicate that MscCa may redistribute to diVerent regions of the cell surface and perform diVerent, yet critical functions depending on the mode of migration. In this case, MscCa seems to meet the critical criterion of modulating all modes of migration, and unlike integrins, myosin II, calpain, and metalloproteases should not become dispensable following a switch in migration mode. 4. [Ca2þ]i Transients Promote Cell Migration but Inhibit Neurite Outgrowth [Ca2þ]i transients have been positively correlated with cell migration in cerebellar granular cells, neutrophils, vascular smooth muscle, keratocytes and astrocytoma cells (Komuro and Rakic, 1996; Lee et al., 1999; Ronde
17. MscCa and Metastasis
499
et al., 2000; Scherberich et al., 2000; Giannone et al., 2002). Furthermore, the cessation of [Ca2þ]i transients has been correlated with the termination of granule cell migration (Kumuda and Komuro, 2004). In contrast, high‐ frequency [Ca2þ]i transients cause nerve growth cone stalling and axon retraction, while the inhibition of [Ca2þ]i transients stimulates the extension of axonal growth cones and the outgrowth of axonal and dendritic filipodia (Gomez and Spitzer, 1999; Gomez et al., 2001; Robles et al., 2003; Lohmann et al., 2005). The [Ca2þ]i transients in all cases appear to depend on MscCa‐ mediated Ca2þ influx because they are blocked by anti‐MscCa agents (Lee et al., 1999; Jacques‐Fricke et al., 2006). Furthermore, the opposite effects both appear to depend on calpain activity (Huttenlocher et al., 1997; Robles et al., 2003). However, whereas calpain activity in the cell rear acts to cleave integrin–CSK linkages and in this way promotes rear retraction and cell migration (Huttenlocher et al., 1997), calpain activity in the nerve growth cone and filopodia acts by promoting actin–integrin disengagement at the front of the process, thereby reducing the traction forces required for lamellar protrusion and growth cone translocation (Robles et al., 2003). Interestingly, calpain inhibition in resting neutrophils promotes polarization and random migration whereas it reduces the neutrophil’s capacity for directional migration toward chemotactic stimuli (Lokuta et al., 2003). This may occur because constitutive calpain activity in resting neutrophils acts as a negative regulator of polarization and migration, whereas the polarized calpain activity in chemotaxing neutrophils promotes directional persistence in a chemo‐attractant gradient.
V. THE ROLE OF MscCa IN CELL MIGRATION A key issue for all modes of cell migration is the nature of the mechanosensitive molecules that act to coordinate forward cell protrusion with rear cell retraction. An attractive candidate is MscCa that because of its unique ability to transduce membrane stretch/cell extension and transduce this into a Ca2þ influx (Guharay and Sachs, 1984; Sachs and Morris, 1998; Hamill and Martinac, 2001; Hamill, 2006) can provide feedback between mechanical forces that tend to extend the cell and the Ca2þ‐sensitive regulators of force generation and cell–substrate adhesion. The first indirect evidence for a role of MscCa in cell migration was provided by the demonstration that the nonspecific MscCa blocker Gd3þ (Yang and Sachs, 1989; Hamill and McBride, 1996) blocked fish keratocyte migration (Lee et al., 1999; Doyle and Lee, 2004; Doyle et al., 2004). Subsequent studies, also using Gd3þ, further implicated MscCa in migration of a mouse fibroblast cell line, NIH3T3 (Munevar et al., 2004), and the human fibrosarcoma cell line, HT1080 (Huang et al., 2004).
500
Maroto and Hamill
However, these studies indicated diVerent sites (i.e., front or back) and diVerent actions (i.e., rear retraction, development of tractions forces, and disassembly of focal adhesions) for MscCa mediated Ca2þ influx, which may partly depend upon diVerent modes of cell migration. Significant limitations in these early studies were the lack of protein identity of MscCa and the absence of MscCa‐ specific reagents, which have been overcome by the recent identification of the canonical transient receptor potential (TRPC1) (Wes et al., 1995) as an MscCa subunit (Maroto et al., 2005), and the discovery of a highly selective MS channel blocker, GsMTx4 a peptide isolated from the tarantula (Grammostola spatulata) venom (Suchyna et al., 2004). Several studies have already implicated TRPC1 in regulating cell migration. For example, Huang et al. (2003) showed immunohistologically that TRPC1 was expressed in a punctuate pattern around the cell periphery, and based on Gd3þ block proposed that TRPC1 supported [Ca2þ]i transients and cell migration. Rao et al. (2006) while studying an intestinal epithelial cell line demonstrated that suppression of TRPC1 inhibited cell migration, whereas TRPC1 overexpression of TRPC1 enhanced cell migration as measured by an in vitro wound closure assay. Maroto et al. (2007) characterized MscCa in both motile (PC‐3) and nonmotile (LNCaP) human prostate tumor cell lines and found that MscCa displayed the same single‐channel conductance, Mg2þ and Gd3þ sensitivity as the MscCa endogenously expressed in Xenopus oocytes identified as formed by TRPC1 (Maroto et al., 2005). Furthermore, MscCa activity was shown to be required for cell migration based on the block by anti‐MscCa/TRPC1 agents including GsMTx4, an anti‐TRPC1 antibody raised against the external pore region of the channel, siRNA suppression, and overexpression of TRPC1. Apart from MscCa, there are other Ca2þ channels that have been implicated in regulating cell migration including both the T‐type (Huang et al., 2004) and L‐type voltage‐gated Ca2þ channels (Yang and Huang, 2005) that may also display mechanosensitivity (Morris and Juranka, Chapter 11, this volume). Also in addition to the TRPCs, which have been implicated in forming MscCa, other TRP subfamily members are expressed in tumor cells and have been implicated in diVerent steps associated with cancer (Peng et al., 2001; Wissenbach et al., 2001; Nilius et al., 2005; Sa´nchez et al., 2005). Of particular interest is TRPM7 that has been shown to regulate cell adhesion by regulating calpain via Ca2þ influx through the channel (Su et al., 2006) and actomyosin contractility via intrinsic kinase activity of TRPM7 (Clark et al., 2006). Although TRPM7 stretch sensitivity has not been directly demonstrated, it has been shown that fluid shear stress‐applied human kidney epithelial cells promote membrane traYcking of TRPM7 to the cell surface (Oancea et al., 2006). Given that fluid shear stress can also trigger cell migration (Isshiki et al., 2002), this may provide an additional MS mechanism to regulate cell motility. In this case, it will be interesting to
17. MscCa and Metastasis
501
determine whether the shear‐induced increase in TRPM7 surface expression is also dependent on specific integrin engagement (Maroto and Hamill, 2001) and/or related to the flow‐induced recruitment of caveolae to specific regions of the migrating cell (Rizzo et al., 2003; Navarro et al., 2004). There are other classes of gated channels that have been implicated in regulating cell migration including voltage‐gated Naþ (Grimes et al., 1995; Bennett et al., 2004; Onganer and Djamgoz, 2005) and Kþ channels (Laniado et al., 2001) and Ca2þ‐activated Kþ channels (Schwab et al., 1994). These diVerent channels may participate in a variety of processes to modulate the pattern of cell migration in the same way as diVerent channels act to produce specific patterns of firing and synaptic release in excitable cells. One would expect that MscCa plays a central role in orchestrating the other channels because of its unique ability to transduce internally and externally generated forces into both depolarization and Ca2þ influx.
VI. CAN EXTRINSIC MECHANICAL FORCES ACTING ON MscCa SWITCH ON CELL MIGRATION? A key question is what causes a cell to switch from a nonmotile to a motile phenotype and vice versa? Although there are numerous studies indicating that growth factors including tumor necrosis factor‐a and transforming growth factor‐b can increase cell motility by promoting the EMT (Bates and Mercurio, 2003; Masszi et al., 2004; Montesano et al., 2005; Nawshad et al., 2005), less well studied is the potential role of extrinsic mechanical forces in turning on cell motility. However, there are at least two key observations that support such a role. In the first place, it has been demonstrated that stationary cell fragments formed from fish keratocytes and lacking a cell nucleus or a microtubular CSK can be stimulated to polarize and undergo persistent locomotion by the application of fluid shear stress or direct mechanical poking (Verkhovsky et al., 1989). Similarly, the application of shear stress to quiescent Dictyostelium can cause CSK reorganization and stimulate cell migration (De´cave´ et al., 2003; Fache et al., 2005). Furthermore, these latter mechanical eVects were shown to be critically dependent on the presence of external Ca2þ (Fache et al., 2005). One possible explanation is that mechanical forces alter the membrane traYcking (Maroto and Hamill, 2001; Isshiki et al., 2002; Rizzo et al., 2003) and/or the MscCa‐gating properties (Hamill and McBride, 1992, 1997; McBride and Hamill, 1992), which in turn alters the [Ca2þ]i dynamics generated by intrinsic mechanical forces and contributes to further polarization of the cell and directional migration. Several previous studies have already discussed the possible role of the changing mechanical environment in terms of
502
Maroto and Hamill
promoting tumor malignancy, including the possible role of increasing interstitial stress and fluid pressure within a growing tumor (Sarntinoranont et al., 2003) and the increased tumor stiVness due to perturbed vasculature and fibrosis (Paszek et al., 2005) of stimulating increased cell motility and escape from the encapsulated tumor. In this case, MscCa may serve as both a trigger and mediator of tumor progression to malignancy. Note Added in Proof Numata, T., Shimizu, T., and Okada, Y. (Am. J. Physiol. 292, C460–C467, 2007) have recently reported that TRPM7 is a stretch‐ and swelling‐activated cation channel expressed in human epithelial cells and is blocked by Gd3þ. These results are consistent with the notion that several classes of mechanosensitive channels may regulate different aspects of tumor cell migration (i.e., forward protrusion and rear retraction) depending upon their differential surface distribution and interaction with downstream Ca2þ‐sensitive effectors.
Acknowledgments We thank the Department of Defense, Prostate Cancer Research Program and the National Cancer Institute for their funding support.
References Alterafi, A., and Zhelev, D. (1997). Transient increase of cytosolic calcium during neutrophil motility responses. J. Cell. Sci. 110, 1967–1977. Arora, P. D., Bibby, K. J., and McCulloch, C. A. G. (1994). Slow oscillations of free Intracellular calcium ion concentration in human fibroblasts responding to mechanical stretch. J. Cell. Physiol. 161, 187–200. Arora, P. D., and McCulloch, C. A. (1996). Dependence of fibroblast migration on actin severing activity of gelsolin. J. Biol. Chem. 271, 20516–20523. Asano, Y., Mizuno, T., Kon, T., Nagasaki, A., Sutoh, K., and Uyeda, T. Q. P. (2004). Keratocyte‐like locomotion of amiB‐null Dictyostelium cells. Cell Motil. Cytoskel. 59, 17–27. Bates, R. C., and Mercurio, A. M. (2003). Tumor necrosis factor‐a stimulates the epithelial‐to‐ mesenchymal transition of human colonic organoids. Mol. Biol. Cell 14, 1790–1800. Beardsley, A., Fang, K., Mertz, H., Castranova, V., Friedn, S., and Liu, J. (2005). Loss of caveolin‐1 polarity impedes endothelial cells polarization and directional movement. J. Biol. Chem. 280, 3541–3547. Bennett, E. S., Smith, B. A., and Harper, J. M. (2004). Voltage‐gated Naþ channels confer invasive properties on human prostate cancer cells. Pflu¨gers Arch. 447, 908–914. Berridge, M. J., Bootman, M. D., and Roderick, H. L. (2003). Calcium signaling: Dynamics homeostasis and remodeling. Nature Revs. Mol. Cell. Biol. 4, 517–529. Blaser, H., Reichman‐Fried, M., Castanon, I., Dumstrel, K., Marlow, F. L., Kawakami, K., Soinica‐Krezel, L., Heisenberg, C. P., and Raz, E. (2006). Migration of Zebrafish primordial germ cells: A role for myosin contraction and cytoplasmic flow. Develop. Cell 11, 613–627. Braiman, A., and Priel, Z. (2001). Intracellular stores maintain stable cytosolic Ca2þ gradients in epithelial cells by active Ca2þ redistribution. Cell Cal. 30(6), 361–371. Brazier, S. C., Singh, B. B., Liu, X., Swaim, W., and Ambudkar, I. S. (2003). Caveolin‐1 contributes to assembly of store‐operated Ca2þ influx channels by regulating plasma membrane localization of TRPC1. J. Biol. Chem. 278, 27208–27215.
17. MscCa and Metastasis
503
Bretscher, M. S., and Aguado‐Velasco, C. (1998). Membrane traYc during cell locomotion. Curr. Opin. Cell Biol. 10, 537–541. Brundage, R. A., Fogarty, K. E., Tuft, R. A., and Fay, F. S. (1991). Calcium gradients underlying polarization and chemotaxis of eosinophils. Science 254, 703–706. Cahill, D. P., Kinzler, K. W., Vogelstein, B., and Lengauer, C. (2000). Genetic instability and Darwinian selection in tumors. Trends Genet. 15, M57–M60. Carragher, N. O., Walker, S. M., Scott, L. A., Harris, F., Sawyer, T. K., Brunton, V. G., Ozanne, B. W., and Frame, M. C. (2005). Calpain 2 and Src dependence distinguishes mesenchymal and amoeboid modes of tumor cell invasion linking to integrin function. Oncogene 25, 5726–5740. Clark, K., Langeslag, M., Van Leeuwen, B., Ran, L., Ryazanov, A. G., Figdor, C. G., Moolenaar, W. H., Jalink, K., and Van Leeuwen, F. N. (2006). TRPM7, a novel regulator of actomyosin contractility and cell adhesion. EMBO J. 25, 290–301. DeBiasio, R., Bright, G. R., Ernst, L. A., Waggoner, A. S., and Taylor, D. L. (1987). Five parameter fluorescent imaging: Wound healing of living Swiss 3T3 cells. J. Cell Biol. 105, 1613–1622. De´cave´, E., Rieu, D., Dalous, J., Fache, S., Brechet, Y., Fourcade, B., Satre, M., and Bruckert, F. (2003). Shear flow‐induced motility of Dictyostelium discoideum cells on solid substrate. J. Cell Sci. 116, 4331–4343. Demuro, A., and Parker, I. (2005). ‘‘Optical Patch‐clamping’’: Single‐channel recording by imaging Ca2þ flux through individual muscle acetylcholine receptor channels. J. Gen. Physiol. 126, 179–192. Doyle, A., Marganski, W., and Lee, J. (2004). Calcium transients induce spatially coordinated increases in traction force during the movement of fish keratocytes. J. Cell Sci. 117, 2203–2214. Doyle, A. D., and Lee, J. (2004). Cyclic changes in keratocyte speed and traction stress arise from Ca2þ‐dependent regulation of cell adhesiveness. J. Cell Sci. 118, 369–379. Dreval, V., Dieterrich, P., Stock, C., and Schwab, A. (2005). The role of Ca2þ transport across the plasma membrane for cell migration. Cell Physiol. Biochem. 16, 119–128. Eddy, R. J., Pierini, L. M., Matsumura, F., and Maxfield, F. R. (2000). Ca2þ‐dependent myosin II activation is required for uropod retraction during neutrophil migration. J. Cell Sci. 113, 1287–1298. Emmelot, P., and Scherer, E. (1977). Multi‐hit kinetics of tumor formation with special reference to experimental liver and human lung carcinogenesis and some general conclusions. Cancer Res. 37, 1702–1708. Erickson, C. A., and Trinkhaus, J. P. (1976). Microvilli and blebs as sources of reserve surface membrane during cell spreading. Exp. Cell Res. 99, 375–384. Fache, S., Dalous, J., Engelund, M., Hansen, C., Chamaraux, F., Fourcade, B., Satre, M., Devreotes, P., and Bruckert, F. (2005). Calcium mobilization stimulates Dictyostelium discoideum shear‐flow‐induced cell motility. J. Cell Sci. 118, 3445–3457. Fisher, P. R., and Wilczynska, Z. (2006). Contribution of endoplasmic reticulum to Ca2þ signals in Dictyostelium depends on extracellular Ca2þ. FEMS Microbial. Lett. 257, 268–277. Franco, S. J., and Huttenlocher, A. (2005). Regulating cell migration: Calpains make the cut. J. Cell Sci. 118, 3829–3838. Friedl, P. (2004). Prespecification and plasticity: Shifting mechanisms of cell migration. Curr. Opin. Cell Biol. 16, 14–23. Friedl, P., and Wolf, K. (2003). Tumor cell invasion: Diversity and escape mechanisms. Nature Rev. Cancer 3, 362–374.
504
Maroto and Hamill
Galbraith, C. G., and Sheetz, M. P. (1999). Keratocytes pull with similar forces on their dorsal and ventral surfaces. J. Cell. Biol. 147, 1313–1323. Giannone, G., Ronde, P., Gaire, M., Haiech, J., and Takeda, K. (2002). Calcium oscillations trigger focal adhesion disassembly in human U87 astrocytoma cells. J. Biol. Chem. 277, 26364–26371. Gilbert, S. H., Perry, K., and Fay, F. S. (1994). Mediation of chemoattractant‐induced changes in [Ca2þ]i and cell shape, polarity and locomotion by InsP3, DAG, and protein kinase C in Newt Eosinophils. J. Cell Biol. 127, 489–503. Gollnick, F., Meyer, R., and Stockem, W. (1991). Visualization and measurement of calcium transients in Amoeboid proetus by fura‐2 fluorescence. Eur. J. Cell Biol. 55, 262–271. Gomez, T. M., and Spitzer, N. C. (1999). In vivo regulation of axon extension and pathfinding by growth cone calcium transients. Nature 397, 350–355. Gomez, T. M., Robles, E., Poo, M., and Spitzer, N. C. (2001). Filopoidal calcium transients promote substrate‐dependent growth cone turning. Science 291, 1983–1987. Gomez‐Mouton, C., Abad, J. L., Mira, E., Lacalle, R. A., Gallardo, E., Jimenez‐Baranda, S., Illa, I., Bernad, A., Manes, S., and Martinez‐A, C. (2001). Segregation of leading‐edge and uropod components into specific lipid rafts during T cell polarization. Proc. Natl. Acad. Sci. USA 98, 9642–9647. Grimes, J. A., Fraser, S. P., Stephens, G. J., Downing, J. E. G., Laniado, M. E., Foster, C. S., Abel, P. D., and Djamgoz, M. B. A. (1995). DiVerential expression of voltage‐activated Naþ currents in two prostatic tumor cell lines: Contribution to invasiveness in vitro. FEBS Letts. 369, 290–294. Grimstad, I. A. (1987). Direct evidence that cancer cell locomotion contributes importantly to invasion. Exp. Cell Res. 173, 515–523. Grynkiewicz, G., Poenie, M., and Tsien, R. (1985). A new generation of Ca2þ indicators with greatly improved fluorescence properties. J. Biol. Chem. 260, 3440–3450. Guharay, F., and Sachs, F. (1984). Stretch activated single ion channel currents in tissue cultured embryonic chick skeletal muscle. J. Physiol. 352, 685–701. Hahn, K., Debiasio, R., and Taylor, D. L. (1992). Patterns of elevated free calcium and calmodulin in living cells. Nature 359, 736–738. Hamill, O. P. (2006). Twenty odd years of stretch activated channels. Pflu¨gers Arch. 453, 333–351. Hamill, O. P., and Martinac, B. (2001). Molecular basis of mechanotransduction in living cells. Physiol. Rev. 81, 685–740. Hamill, O. P., and McBride, D. W., Jr. (1992). Rapid adaptation of single mechanosensitive channels in Xenopus oocytes. Proc. Natl. Acad. Sci. USA 89, 7462–7466. Hamill, O. P., and McBride, D. W., Jr. (1996). The pharmacology of mechanogated membrane ion channels. Pharmacol. Rev. 48, 231–252. Hamill, O. P., and McBride, D. W., Jr. (1997). Induced membrane hypo/hyper‐ mechanosensitivity: A limitation of patch‐clamp recording. Ann. Rev. Physiol. 59, 621–631. Hamill, O. P., Marty, A., Neher, E., Sakmann, B., and Sigworth, F. J. (1981). Improved patch‐ clamp techniques for high‐resolution current recording from cells and cell‐free membrane patches. Pflu¨gers Arch. 391, 85–100. Hanahan, D., and Weinberg, R. A. (2000). The hallmarks of cancer. Cell 100, 57–70. Hawkes, R. B., and Hoberton, D. V. (1973). A calcium‐sensitive lanthanum inhibition of amoeboid movement. J. Cell. Physiol. 81(3), 365–370. Hendey, B., and Maxfield, F. R. (1993). Regulation of neutrophil motility and adhesion by intracellular calcium transients. Blood Cells 19, 143–164.
17. MscCa and Metastasis
505
Huang, J. B., Kindzelskii, A. L., Clark, A. J., and Petty, H. R. (2004). Identification of channels promoting calcium spikes and waves in HT1080 Tumor cells: Their apparent roles in cell motility and invasion. Cancer Res. 64, 2482–2489. Huttenlocher, A., Palecek, S. P., Lu, Q., Zhang, W., Mellgren, R. L., LauVenburger, D. A., Ginsberg, M., and Horowitz, A. (1997). Regulation of cell migration by calcium‐dependent protease calpain. J. Biol. Chem. 272, 32719–32722. Isshiki, M., Ando, J., Yamamoto, K., Fujta, T., and Ying, Y. (2002). Sites of Ca2þ wave initiation move with caveola to the trailing edge of migrating cells. J. Cell Sci. 115, 475–484. Jacques‐Fricke, B. T., Seow, Y., Gottlieb, P. A., Sachs, F., and Gomez, T. M. (2006). Ca2þ influx through mechanosensitive channels inhibits neurite outgrowth in opposition to other influx pathways and release of intracellular stores. J. Neurosci. 26, 5656–5664. Kaighn, M. E., Narayan, K. S., Ohnuki, Y., Lechner, J. F., and Jones, L. W. (1979). Establishment and characterization of a human prostatic carcinoma cell line (PC‐3). Invest. Urol. 17, 16–23. Kasai, H., and Augustine, G. J. (1990). Cytosolic Ca2þ gradients triggering unidirectional fluid secretion from exocrine pancreas. Nature 348, 735–738. Kassis, J., LauVenburger, D. A., Turner, T., and Wells, A. (2001). Tumor invasion as dysregulated cell motility. Semin. Cancer Biol. 11, 105–117. Kimura, C., Oike, M., Koyama, T., and Ito, Y. (2001). Alterations of Ca2þ mobilizing properties in migrating endothelial cells. Am. J. Physiol. 281, H745–H754. Kindzelskii, A. L., Sitrin, R. G., and Petty, H. R. (2004). Cutting edge: Optical microspectrophotometry supports the existence of gel phase lipid rafts at the lamellipodium of neutrophils: Apparent role in calcium signaling. J. Immunol. 172, 4681–4685. Komuro, H., and Rakic, P. (1996). Intracellular Ca2þ fluctuations modulate the rate of neuronal migration. Neuron 17, 275–285. Korohoda, W., Madeja, Z., and Sroka, J. (2002). Diverse chemotactic responses of Dictystelium discoideum Amoeba in the developing (temporal) and stationary (spatial) concentration gradients of folic acid, cAMP, Ca2þ and Mg2þ. Cell Motil. Cytoskel. 53, 1–25. Kumuda, T., and Komuro, H. (2004). Completion of neuronal migration regulated by loss of Ca2þ transients. Proc. Natl. Acad. Sci. USA 101, 8479–8484. LaVafian, I., and Hallet, M. B. (1995). Does cytosolic free Ca2þ signal neutrophil chemotaxis in response to formylated chemotactic peptide? J. Cell Sci. 108, 3199–3205. Laniado, M. E., Fraser, S. P., and Djamgoz, M. B. (2001). Voltage‐gated Kþ channel activity in human prostate cancer cell lines of markedly diVerent metastatic potential: Distinguishing characteristics of PC‐3 and LNCaP cells. Prostate 46, 262–274. LauVenburger, D. A., and Horwitz, A. F. (1996). Cell migration: A physically integrated molecular process. Cell 84, 359–369. Lee, J., Ishihara, A., Oxford, G., Johnson, B., and Jacobson, K. (1999). Regulation of cell movement is mediated by stretch‐activated calcium channels. Nature 400, 382–386. Levina, N. N., and Lew, R. R. (2006). The role of tip localized mitochondria in hyphal growth. Fungal Genet. Biol. 43, 65–74. Liu, L., Ridefelt, P., Ha˚kansson, L., and Venge, P. (1999). Regulation of human eosinophil migration across lung epithelial monolayers by distinct calcium signaling mechanisms in the two cells types. J. Immunol. 163, 5649–5655. Liu, L., Ha˚kansson, L., Ridefelt, P., Garcia, R. C., and Venge, P. (2003). Priming of eosinophil migration across lung epithelial cell monolayers and upregulation of CD11b/CD18 are elicited by extracellular Ca2þ. Am. J. Respir. Cell Mol. Biol. 28, 713–721. Lockwich, T. P., Liu, X., Singh, B. B., Jadlowiec, J., Weiland, S., and Ambudkar, I. S. (2000). Assembly of Trp1 in a signaling complex associated with caveolin‐scaVolding lipid raft domains. J. Biol. Chem. 275, 11934–11942.
506
Maroto and Hamill
Lohmann, C., Finski, A., and BonhoeVer, T. (2005). Local calcium transients regulate the spontaneous motility of dendritic filopodia. Nature Neurosci. 8, 305–312. Lokuta, M. A., Nuzzi, P. A., and Huttnelocher, A. (2003). Calpain regulates neutrophils chemotaxis. Proc. Natl. Acad. Sci. USA 100, 4006–4011. Malawista, S. E., and Boisfleury‐Chevance, A. (1997). Random locomotion and chemotaxis of human blood polymorphonuclear leukocytes (PMN) in the presence of EDTA: PMN in close quarters require neither leukocyte integrins nor external divalent cations. Proc. Natl. Acad. Sci. USA 94, 11577–11582. Mamoune, A., Luo, J. H., LauVenburger, D. A., and Wells, A. (2003). Calpain‐2 as a target for limiting prostate cancer invasion. Cancer Res. 63, 4632–4640. Mandeville, J. T. H., and Maxfield, F. R. (1997). EVects of buVering intracellular free calcium on neutrophil migration through three‐dimensional matrices. J. Cell. Physiol. 171, 168–178. Manes, S., Mira, E., Gomez‐Mouton, C., Lacalle, R. A., Keller, P., Labrador, J. P., and Martinez‐A, C. (1999). Membrane raft domains mediate front‐rear polarity in migrating cells. EMBO J. 18, 6211–6220. Marks, P. W., and Maxfield, F. R. (1990). Transient increases in cytosolic free calcium appear to be required for the migration of adherent human neutrophils. J. Cell Biol. 110, 43–52. Maroto, R., and Hamill, O. P. (2001). Brefeldin A block of integrin‐dependent mechanosensitive ATP release from Xenopus oocytes reveals a novel mechanism of mechanotransduction. J. Biol. Chem. 276, 23867–23872. Maroto, R., Raso, A., Wood, T. G., Kurosky, A., Martinac, B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nature Cell Biol. 7, 179–185. Maroto, R., Kurosky, A., and Hamill, O. P. (2007). The role of MscCa in prostate tumor cell migration. (submitted for publication.) Masszi, A., Fan, L., Rosivall, L., McCulloch, C. A., Rotstein, O. D., Mucsi, I., and Kapus, A. (2004). Integrity of cell‐cell contacts is a critical regulator of TGF‐beta 1‐induced epithelial‐ to‐myofibroblast transition: Role for beta‐catenin. Am. J. Pathol. 165, 1955–1967. McBride, D. W., Jr., and Hamill, O. P. (1992). Pressure‐clamp: A method for rapid step perturbation of mechanosensitive channels. Pflu¨gers Arch. 421, 606–612. Miyawaki, A., Llopis, J., Heim, R., McCaVery, J. M., Adams, J. A., Ikura, M., and Tsien, R. Y. (1997). Fluorescent indicators for Ca2þ based on green fluorescent proteins and calmodulin. Nature 388, 882–887. Mobasheri, A., Carter, S. D., Martin‐Vasallo, P., and Shakibaei, M. (2002). Integrins and stretch‐activated ion channels; putative components of functional cell surface mechanoreceptors in articular chondrocytes. Cell Biol. Int. 26, 1–18. Montesano, R., Soulie, P., Eble, J. A., and Carrozzino, F. (2005). Tumor necrosis factor‐a confers an invasive transformed phenotype on mammary epithelial cells. J. Cell Sci. 118, 3487–3500. Munevar, S., Wang, Y. L., and Dembo, M. (2004). Regulation of mechanical interactions between fibroblasts and the substratum by stretch‐activated Ca2þ entry. J. Cell Sci. 117, 85–92. Navarro, A., Anand‐Apte, A. B., and Parat, M. O. (2004). A role for caveola in cell migration. FASEB J. 18, 1801–1811. Nawshad, A., Lagamba, D., Polad, A., and Hay, E. D. (2005). Transforming growth factor‐b‐ signaling during epithelial‐mesenchymal transformation: Implications for embryogenesis and tumor metastasis. Cells Tissues Organs 179, 11–23. Nebl, T., and Fisher, P. R. (1997). Intracellular Ca2þ signals in Dictyostelium chemotaxis are mediated exclusively by Ca2þ influx. J. Cell Sci. 110, 2845–2853.
17. MscCa and Metastasis
507
Nilius, B., Voets, T., and Peters, J. (2005). TRP channels in disease. Sci. STRKE, re8. Oancea, E., Wolfe, J. T., and Clapham, D. E. (2006). Functional TRPM7 channels accumulate at the plasma membrane in response to fluid flow. Circ. Res. 98, 245–253. Oike, M., Droogmans, G., and Nilius, B. (1994). Mechanosensitive Ca2þ transients in endothelial cells from human umbilical vein. Proc. Natl. Acad. Sci. USA 91, 2940–2944. Onganer, P. U., and Djamgoz, M. B. A. (2005). Small‐cell lung cancer (human); potentiation of endocytoc membrane activity by voltage‐gated Naþ channel expression in vitro. J. Membr. Biol. 204, 67–75. Parat, M. O., Anand‐Apte, B., and Fox, P. L. (2003). DiVerential caveolin‐1 polarization in endothelial cells during migration in two and three dimensions. Mol. Biol. Cell 14, 3156–3168. Paszek, M. J., Zahir, N., Johnson, K. R., Lakins, J. N., Rozenberg, G. I., Gefen, A., Reinhardt‐ King, C. A., Margulies, S. S., Dembo, M., Boettiger, D., Weaver, D. A., and hammer, V. A. (2005). Tensional homeostasis and the malignant phenotype. Cancer Cell 8, 241–254. Peng, J. B., Zhuang, L., Berger, U. V., Adam, R. M., Williams, B. J., Brown, E. M., Hediger, M. A., and Freeman, M. R. (2001). CaT1 expression correlates with tumor grade in prostate cancer. Biochem. Biophys. Res. Commun. 282, 729–734. Petersen, O. H., Tepikin, A., and Park, M. K. (2001). The endoplasmic reticulum: One continuous or several separate Ca2þ stores? Trends Neurosci. 24, 271–276. Pierini, L.M, Eddy, R. J., Fuortes, M., Seveau, S., Casulo, C., and Maxfield, F. R. (2003). Membrane lipid organization is critical for human neutrophil polarization. J. Biol. Chem. 278, 10831–10841. Quintana, A., and Hoth, M. (2004). Apparent cystosolic calcium gradients in T‐lymphocytes due to fura‐2 accumulation in mitochondria. Cell Calc. 36, 99–109. Ramaswamy, S., Ross, K. N., Lander, E. S., and Golub, T. R. (2003). A molecular signature in primary solid tumors. Nature Genet. 33, 49–54. Rao, J. N., Platoshyn, O., Golovina, V. A., Liu, L., Zou, T., Marasa, B. S., Turner, D. J., Yuan, J. X. L., and Wang, J. Y. (2006). TRPC1 functions as a store‐operated Ca2þ channel in intestinal epithelial cells and regulates early mucosal restitution after wounding. Am. J. Physiol. 290, G782–G792. Ridley, A. J., Schwartz, M. A., Burridge, K., Firtel, R. A., Ginsberg, M. H., Borisy, G., Parsons, J. T., and Horwitz, A. R. (2003). Cell migration: Integrating signals from front to back. Science 302, 1704–1709. Rizzo, V., Morton, C., Depaola, N., Schnitzer, J. E., and Davies, P. F. (2003). Recruitment of endothelial caveola into mechanotransduction pathways in flow conditioning in vitro. Am. J. Physiol. 285, H1720–H1729. Robles, E., Huttenlocher, A., and Gomez, T. M. (2003). Filopodial calcium transients regulate growth cone motility and guidance through local activation of calpain. Neuron 38, 597–609. Ronde, P., Giannone, G., Gerasymova, I., Stoeckel, H., Takeda, K., and Haiech, J. (2000). Mechanism of Calcium oscillations in human astrocytoma cells. Biochim. Biophys. Acta 1498, 273–280. Saadoun, S., Papadopoulos, M. C., Hara‐Chikuma, M., and Verkman, A. S. (2005). Impairmant of angiogeneisis and cell migration by targeted aquaporin‐1 gene disruption. Nature 434, 786–792. Sachs, F., and Morris, C. E. (1998). Mechanosensitive ion channels in nonspecialized cells. Revs. Physiol. Biochem. Pharmacol. 132, 1–77. Sahai, E., and Marshall, C. J. (2003). DiVerent modes of tumor cell invasion have distinct requirements for Rho/ROCK signaling and extracellular proteolysis. Nat. Cell Biol. 5, 711–719.
508
Maroto and Hamill
Saimi, Y., Zhou, X., Loukin, S. H., Haynes, W. J., and Kung, C. (2007). Microbial TRP channels and their mechanosensitivity. Curr. Top. Memb. 58, 311–327. Sa´nchez, M. G., Sanchez, A. M., Collado, B., Malagarie‐Cazenave, S., Olea, N., Carmena, M. J., Prieto, J. C., and Diaz‐Laviada, I. I. (2005). Expression of the transient receptor potential vanilliod 1 (TRPV1) in LNCaP and PC‐3 prostate cancer cells and in prostate tissue. Eur. J. Pharmacol. 515, 20–27. Sarntinoranont, M., Rooney, F., and Ferrari, M. (2003). Interstitial stress and fluid pressure within a growing tumor. Ann. Biomed. Eng. 31, 327–335. Schaloske, R. H., Lusche, D. F., Bezares‐Roder, K., Happle, K., Malchow, D., and Schlatterer, C. (2005). Ca2þ regulation in the absence of the iplA gene product in Dictyostelium discoideum. BMC Cell Biol. 6, 13–30. Shanley, L. J., Walczysko, P., Bain, M., MacEwan, D. J., and Zhao, M. (2006). Influx of extracellular Ca2þ is necessary for electrotaxis in Dictyostelium. J. Cell Sci. 119, 4741–4748. Scherberich, A., Campos‐Toimil, M., Ronde, P., Takeda, K., and Baretz, A. (2000). Migration of human vascular smooth muscle cells involves serum‐dependent repeated cytosolic calcium transients. J. Cell Sci. 113, 653–662. Schwab, A., Wojnowski, L., Gabriel, K, and Oberleither, H. (1994). Oscillating activity of a Ca2þ‐ sensitive Kþ channel: A prerequisite for migration of transformed Mardin‐Darby canine kidney focus cells. J. Clin. Invest. 93, 1631–1636. Schwab, A., Finsterwalder, F., Kersting, U., Danker, T., and Oberleithner, H. (1997). Intracellular Ca2þ distribution in migrating transformed renal epithelial cells. Pflu¨gers Arch. 434, 70–76. Schwab, A., Nechyporuk‐Zloy, V., Fabian, A., and Stock, C. (2006). Cells move when ions and water flow. Pflu¨gers Arch. 453, 421. Sheetz, M. P., Felsenfeld, D., Galbraith, C. G., and Choquet, D. (1999). Cell migration as a five‐step cycle. Biochem. Soc. Sympos. 65, 233–243. Silverman‐Gavrila, L. B., and Lew, R. R. (2003). Calcium gradient dependence of Neurospora crassa hyphal growth. Microbiology 149, 2475–2485. Stracke, M. L., Aznavoorian, S. A., Beckner, M. E., Liotta, L. A., and Schiffmann, E. (1991). Cell motility, a principal requirement for metastasis. EXS 59, 147–162. Strohmeier, R., and Bereiter‐Hahn, J. (1984). Control of cell shape and locomotion by external calcium. Exp. Cell Res. 154, 412–420. Su, L. T., Agapito, M. A., Li, M., Simonson, W. T. N., Huttenlocher, A., Habas, R., Yue, L., and Runnels, L. W. (2006). TRPM7 regulates cell adhesion by controlling the calcium‐ dependent protease calpain. J. Biol. Chem. 281, 11260–11270. Suchyna, T. M., Tape, S. E., Koeppe, R. E., II, Andersen, O. S., Sachs, F., and Gottlieb, P. A. (2004). Bilayer‐dependent inhibition of mechanosensitive channels by neuroactive peptide enantiomers. Nature 430, 235–240. Sung, V., Gilles, C., Clarke, M. R., Aaron, A. D., Azumi, N., and Thompson, E. W. (1998). The LCC15‐MB human breast cancer cell line expresses osteopontin and exhibits an invasive and metastatic phenotype. Exp. Cell Res. 241, 273–284. Tarin, D. (2005). The fallacy of epithelial mesenchymal transition in neoplasia. Cancer Res. 65, 5996–6001. Taylor, D. L., Blinks, J. R., and Reynolds, G. (1980). Contractile basis of amoeboid movement VIII. Aequorin Luminescence during amoeboid movement, endocytosis and capping. J. Cell Biol. 86, 599–607. Thiery, J. P. (2002). Epithelial‐mesenchymal transitions in tumor progression. Nature Rev. Cancer 2, 442–454. Thompson, E. W., and Newgreen, D. F. (2005). Carcinoma invasion and metastasis: A role for epithelial‐mesenchymal transition? Cancer Res. 65, 5991–5995.
17. MscCa and Metastasis
509
Tinel, H., Cancela, J. M., Mogami, H., Geraimenko, J. V., Gerasimenko, O. V., Tepikin, A. V., and Petersen, O. H. (1999). Active mitochondria surrounding the pancreatic acinar granule region prevent spreading of inositol trisphosphate‐evoked local cytosolic Ca2þ signals. EMBO J. 18, 4999–5008. Traynor, D., Milne, J. L. S., Insall, R. H., and Kay, R. R. (2000). Ca2þ signaling is not required for chemotaxis in. Dictyostelium. EMBO J. 19, 4846–4854. Unterweger, N., and Schlatterer, C. (1995). Introduction of calcium buVers into the cytosol of Dictyostelium discoideum amoeba alters cell morphology and inhibits chemotaxis. Cell Cal. 17, 97–110. Varambally, S., Yu, J., Laxman, B., Rhodes, D. R., Mehra, R., Tomlins, S. A., Shah, R. B., Chandran, U., Monzon, F. A., Becich, M. J., Wei, J. T., Pienta, K. J., et al. (2005). Integrative genomic and proteomic analysis of prostate cancer reveals signatures of metastatic progression. Cancer Cell 8, 393–406. Verkhovsky, A. B., Svitkina, T. M., and Borisy, G. G. (1989). Self‐polarization and directional motility of cytoplasm. Curr. Biol. 9, 11–20. Wahl, M., Sleight, R. G., and Gurenstein, E. (1992). Association of cytoplasmic free Ca2þ gradients with subcellular organelles. J. Cell Physiol. 150, 593–609. Wes, P. D., Chevesich, J., Jeromin, A., Rosenberg, C., Stetten, G., and Montell, C. (1995). TRPC1, a human homolog of a Drosophila store‐operated channel. Proc. Natl. Acad. Sci. USA 92, 9652–9656. Wissenbach, U., Niemeyer, B. A., Fixemer, T., Schneidewind, A., Trost, C., Cavalie´, A., Reus, K., Mee se, E., Bonkhoff, H., and Flockerzi, V. (2001). Expression of CaT‐like, a novel calcium‐selective channel, correlates with the malignancy of prostate cancer. J. Biol. Chem. 276, 19461–19468. Wolf, K., and Friedl, P. (2006). Molecular mechanisms of cancer cell invasion and plasticity. Brit. J. Dermatol. 154, 11–15. Wu, Z., Wong, K., Glogauer, M., Ellen, R. P., and McCulloch, C. A. G. (1999). Regulation of stretch‐activated intracellular calcium transients by actin filaments. Biochem. Biophys. Res. Commun. 261, 419–425. Xu, H., Yuan, X., Guan, C., Duan, S., Wu, C., and Feng, L. (2004). Calcium signaling in chemorepellant Slit2‐dependnet regulation of neuronal migration. Proc. Natl. Acad. Sci. USA 101, 4296–4301. Yang, S., and Huang, X. Y. (2005). Ca2þ influx through L‐type Ca2þ channels controls the trailing tail contraction in growth factor‐induced fibroblast cell migration. J. Biol. Chem. 280, 27130–27137. Yang, X. C., and Sachs, F. (1989). Block of stretch‐activated ion channels in Xenopus oocytes by gadolinium and calcium ions. Science 243, 1068–1071. Yi, M., Weaver, D., and Hajnoczky, G. (2004). Control of mitochondrial motility and distribution by the calcium signal: A homeostatic circuit. J. Cell Biol. 167, 661–672. Yoshida, K., and Soldati, T. (2006). Dissection of amoeboid movement into two mechanically distinct modes. J. Cell Sci. 119, 3833–3844. Yumura, S., Furuya, K., and Takeuchi, I. (1996). Intracellular free calcium responses during chemotaxis of Dictyostelium cells. J. Cell Sci. 109, 2673–2678. Zhou, X., Rao, N. P., Cole, S. W., Mok, S. C., Chen, Z., and Wong, D. T. (2005). Progress in concurrent analysis of loss of heterozygosity and comparative genomic hybridization utilizing high density nucleotide polymorphism arrays. Cancer Genet. Cytogenet. 159, 53–57. Zigmond, S. H., Slonczewski, J. L. M., Wilde, M. W., and Carson, M. (1988). Polynuclear leukocyte locomotion is insensitive to lowered cytoplasmic calcium levels. Cell Motil. Cytoskel. 9, 184–189.
This page intentionally left blank
CHAPTER 18 Stretch‐Activated Conductances in Smooth Muscles Kenton M. Sanders and Sang Don Koh Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, Nevada 89557
I. Overview II. Introduction III. Mechanosensitive Conductances that Generate Inward Currents A. Vascular Smooth Muscle B. Bladder Myocytes C. GI Myocytes IV. Mechanosensitive Conductances That Generate Outward Currents A. Vascular Muscles B. Bladder Smooth Muscle C. Uterine Smooth Muscle D. GI Smooth Muscle References
I. OVERVIEW The excitability of smooth muscle cells is regulated, in part, by stretch‐ activated ion channels in the plasma membrane. The response to stretch of a particular muscle or organ (i.e., enhancement or stabilization of excitability) is tuned to specific functional needs by the types of ion channels expressed. Mechanosensitive ionic conductances that yield either inward or outward currents have been observed in and characterized in studies of smooth muscles. In vascular muscles the dominant response to stretch is muscle contraction (the myogenic response). Several mechanisms for the myogenic response have been proposed, and one of these hypotheses involves stretch‐dependent
Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59020-9
512
Sanders and Koh
activation of nonselective cation channels. The inward current resulting from activation of these channels causes plasma membrane depolarization, activation of voltage–gated Ca2þ channels, Ca2þ entry, and excitation– contraction coupling. Thus, increasing vascular pressure and distension of blood vessels cause responsive vasoconstriction. Other conductances have also been proposed as participants in the myogenic response, and progress characterizing the inward current channels responsive to stretch is summarized. Outward currents responding to muscle stretch are also present in smooth muscles. For example, expression of stretch‐sensitive two‐pore domain Kþ (K2P) channels has been reported in visceral smooth muscles. These organs resist contraction on filling and provide a reservoir function. Stretch‐ dependent outward current channels are hypothesized to help stabilize membrane potential until it becomes desirable to empty the stored contents. Mechanosensitive conductances participate in the integrated responses of smooth muscle tissues and this chapter summarizes the state of knowledge about this interesting class of channels found in smooth muscles.
II. INTRODUCTION Smooth muscle organs are often referred to as ‘‘volume organs’’ because they experience rather dramatic changes in volume during normal physiological processes. The importance of the length or stretch of smooth muscles has long been known to contribute to the regulation of the electrical and contractile states of these muscles. The first experiments linking changes in muscle length to electrophysiological responses were reported by Bu¨lbring (1955), who demonstrated that stretch of the taenia coli of rabbits caused cell membrane depolarization. This response was not due to activation of nerves but was an intrinsic response of the smooth muscle cells (myogenic). Many smooth muscle organs, including blood vessels, urinary bladder, uterus, and gastrointestinal (GI) organs, display intrinsic (myogenic), nonneural responses to stretch, and these responses are tuned to the physiology and functional needs of specific organs. Experiments on isolated smooth muscle cells have shown that a number of ion channels can be activated on cell elongation or cell deformation. In some cases, stretch is transduced to achieve a contractile response through activation of mechanosensitive channels, induction of inward currents, depolarization, and entry of Ca2þ through either the mechanosensitive conductance or through voltage‐gated Ca2þ channels that are ubiquitous in smooth muscles and typically responsive to small levels of depolarization. Mechanosensitive conductances responsible for inward currents include: stretch‐activated nonselective cation channels (Kirber et al., 1988; Davis et al., 1992a; Wellner and Isenberg, 1993b), swelling‐activated
18. Stretch‐Activated Conductances in Smooth Muscles
513
Cl channels (Dick et al., 1998; Yamazaki et al., 1998), and Ca2þ channels (Langton, 1993; Farrugia et al., 1999). In some smooth muscles, contraction in response to stretch would be deleterious to the physiological function of the organ or tissue region. Thus, in some cases stretch activates a net outward current or a balance to stretch‐activated inward currents to accomplish membrane potential stabilization and/or relaxation. Kþ channels are the dominant ionic conductances that respond to stretch and stabilize membrane potentials of smooth muscles. Numerous types of Kþ channels have been characterized in studies of smooth muscle cells, including large conductance Ca2þ‐activated Kþ (BK) channels, small and intermediate conductance‐activated Kþ (SK and IK) channels, voltage‐dependent Kþ (Kv) channels, ATP‐sensitive Kþ (KATP) channels, inward rectifier Kþ (Kir) channels, and K2P channels (Nelson and Quayle, 1995; Standen and Quayle, 1998; Brayden, 2002; Calderone, 2002; Cole et al., 2005; Jackson, 2005; Sanders and Koh, 2006). Some of these conductances respond to stretch, either directly or in response to secondary cellular signaling. One basic truth about smooth muscles is that the cells express an abundance of synergistic and contradictory pathways, so nearly every stimulus leads to a highly integrated response. In the case of stretch, channels promoting either inward or outward currents can be activated. There is always a balancing act, and the net conductance determines the eVect of stretch on membrane potential. As above, membrane potential is a critical factor, since Ca2þ entry and intracellular Ca2þ ([Ca2þ]i) are dramatically aVected by rather small depolarization or hyperpolarization responses in many smooth muscles. This chapter reviews the major conductances that have been linked to cellular stretch or membrane deformation in smooth muscles from vascular and visceral tissues. First, there is a description of the inward conductances that have been imbued with mechanosensitivity. Activation of these channels tends to increase membrane excitability and increased contractile activity. We will also discuss outward conductances that may stabilize membrane potentials in organs or regions of organs that serve as reservoirs. There have been several modes of mechanical stimulation employed to activate mechanosensitive ionic conductances in smooth muscle cells. These include: longitudinal elongation of cells, stretching of cell membranes by applying negative pressure to pipettes during on‐cell patch recording, inflation of cells during whole‐cell recording by application of positive pressure to patch pipettes, osmotic swelling by bathing cells in hypoosmotic solutions, and rapid perfusion of extracellular solutions. At present, it is not clear that the stress applied to the cytoskeleton of smooth muscle cells by the various forms of mechanical perturbation is equivalent, and relatively few studies have compared responses to multiple stimulus modalities. In general, cell elongation, while diYcult to perform on smooth muscle cells, must be considered the gold
514
Sanders and Koh
standard for activation of ‘‘stretch‐activated’’ ion channels because this form of perturbation is likely to best simulate the type of cellular deformation that occur during filling or increasing the pressure within smooth muscle organs.
III. MECHANOSENSITIVE CONDUCTANCES THAT GENERATE INWARD CURRENTS An increase in luminal pressure in resistance arteries and arterioles causes an initial passive distension and then constriction that is due to active stress development by vascular smooth muscle cells. This phenomenon, termed the myogenic response by Bayliss (1902), has been investigated for many years by vascular biologists and the mechanism is still not fully understood. Myogenic contractions are associated with a sustained depolarization of smooth muscle cells (Harder, 1984), and the sustained depolarization is insensitive to L‐type Ca2þ channel‐blocking drugs (Knot and Nelson, 1995). Many investigators believe that the membrane potential responses to vascular distension is due to activation of mechanosensitive ion channels in smooth muscle cells (Kirber et al., 1988; Wu and Davis, 2001). It should be noted that the stretch‐ activated channels in smooth muscles (including those of bladder and stomach) are similar to channels originally described in studies of skeletal muscle and Xenopus ooctyes in terms of their single channel conductance, ionic permeability, and blocking eVects of multivalent cations [e.g., compare properties of channels described in this chapter with channels described in references Yang and Sachs (1993) and Morris (1990)].
A. Vascular Smooth Muscle 1. Nonselective Cation Conductances Negative pressure applied to the interior of patch pipettes during on‐cell recording of membrane currents activated nonselective cation channels in pig coronary artery smooth muscle cells (Davis et al., 1992a). Membrane stretch increased the open probability of the stretch‐activated channels, but did not aVect the unitary conductance. These authors compared the eVects of stretching the membrane patch with cell elongation. Stretching cells was accomplished by attaching additional patch pipettes and using the strong mechanical properties of giga‐seals to increase the length of the cells. These elegant experiments showed that stretching cells caused depolarization and, in some cases, generation of action potentials. Under conditions of voltage clamp stretch caused activation of sustained inward currents. The reversal potential for this current was about 15 mV. Voltage‐gated Ca2þ channels
18. Stretch‐Activated Conductances in Smooth Muscles
515
were not aVected when cells were stretched by about 15–20% of the resting cell length in these studies. Associating stretch‐activated unitary currents with changes in membrane potential when smooth muscle cells are stretched is somewhat diYcult because: (1) there are no specific pharmacological blockers for stretch‐ activated cation channels and (2) the molecular identity of the channels responsible for inward currents is in question so gene inactivation or knockout experiments are not yet possible. At present comparisons of properties of single channel and whole‐cell currents may be the best way to show that specific unitary conductances are responsible for whole‐cell and tissue responses. The whole‐cell current activated by longitudinal stretch of coronary artery myocytes was a function of the degree of stretch (Fig. 1) and did not experience significant rundown during the course of experiments (Wu and Davis, 2001). Repeated stretches yielded currents of approximately the same magnitude, and phasic stretches at approximately the rate of the heartbeat resulted in sustained inward currents. The current–voltage relationship for the conductance activated by stretch showed that the conductance was weakly, outwardly rectifying positive to 10 mV and the reversal potential was 18 mV. The conductance responsible for the stretch‐activated current did not appear to carry much Ca2þ under physiological ionic gradients and the current was not reduced when Ca2þ was removed from the bath solution. At negative potentials, Naþ appeared to be the predominant charge carrier, but at positive potentials, the current was dominated by Kþ since it did not decrease as potentials approached the Naþ equilibrium potential. The properties of the stretch‐activated nonselective cation current in coronary myocytes are summarized in Table I. Wu and Davis (2001) also found an outward current that was enhanced by stretch in vascular smooth muscles. Large‐conductance BK channels were responsible for the outward current, and this conductance activated in response to activation of the nonselective cation conductance (Section IV) because activation of the outward current conductance was blocked by reducing extracellular Ca2þ during cell elongation and by iberiotoxin (IbTX). Block of BK channels resulted in enhanced depolarization in response to stretch. Thus, stretch‐dependent depolarization is an integrated response that is determined by the inward current caused by activation of stretch‐sensitive nonselective cation channels, Ca2þ entry due mainly to depolarization and activation of voltage‐gated Ca2þ channels, and activation of BK channels. Activation of Kþ channels during myogenic responses opposes the dominant depolarization and contractile responses. Thus, blocking Kþ channels tends to increase the gain of machanosensitivity in vascular muscles. Stretching cells by greater than 15% of their slack length caused [Ca2þ]i to increase in vascular muscle myocytes (Davis et al., 1992b). The threshold
516
Sanders and Koh C 120%
Length
25 Control Stretch
10 s
Current
I(pA/pF)
A
20 15 10
120%
5
10 s −100
−60
−20
20
60
V (mV) −5
B −6
D
6 I(pA/pF)
I(pA/pF)
−5 −4 −3
−100
−2
−60
4 2
−20
20
60
V (mV) −2
−1 130 110 120 Length change (%)
−4
FIGURE 1 Stretch‐activated nonselective cation currents in coronary smooth muscle cells. The amplitude of the stretch‐activated current in coronary myocytes increased as a function of the magnitude of longitudinal stretch. Panel A shows responses under voltage clamp as cell length was increased to 120%, 125%, and 130% of the resting length. Panel B shows a plot of the amplitude of the stretch‐activated current as a function of cell length. The relationship between current and length was fit by a Boltzmann equation and the half‐maximal response occurred at 123.3% of the resting length. This experiment was performed with physiological salt concentrations in the bathing solution and the holding potential was 60 mV. Panel C shows the current– voltage relationship for the stretch‐activated currents in five cells. Average membrane current was evoked by ramping cells from 100 to þ60 mV. Current density was recorded before (open circles) and during a cell elongation to 115% of the resting cell length (filled circles). Panel D shows the diVerence currents obtained by subtracting the currents elicited before and during stretch. The stretch‐activated current reversed at 18 mV. Figure is redrawn and used with permission from Wu and Davis (2001).
for significant changes in [Ca2þ]i was elongation of cells by about 10%. Extracellular Ca2þ was the source for stretch‐induced rise in [Ca2þ]i, but nifedipine blocked only a portion of Ca2þ entry stimulated by stretch. Gadolinium, which can block stretch‐activated channels in oocytes (Yang and Sachs, 1989), blocked the stretch‐induced increase in Ca2þ in pig coronary artery myocytes. These data can be interpreted in the following way: stretching cells activate nonselective cation channels, and some Ca2þ may enter cells by this pathway. The inward current caused by stretch leads to depolarization and this activates voltage‐gated Ca2þ channels. These channels facilitate greater entry of Ca2þ. This mechanism suggests that stretch‐activated
517
18. Stretch‐Activated Conductances in Smooth Muscles TABLE I
Properties of Nonselective Cation Conductance Activated by Stretch in Vascular Myocytes (Summary of Findings from Wu and Davis, 2001) 1. EVects of stretch were reversible; current could be activated repeatedly 2. Stretch‐activated current was graded and increased by increasing stretch up to 135% of resting cell length 3. Stretch‐activated current was activated with same time course as depolarization in response to stretch of myocytes 4. Reversal potential in physiological ionic gradients was between
15 and
20 mV
5. Stretch‐activated current persisted in Ca2þ‐free bathing solution 6. Current was unaVected by Kþ channel‐blocking drugs 7. Stretch‐activated current was blocked by gadoliniuma and Grammostola spatulata venoma a It should be noted that these blockers are not specific for non‐specific cation channels (NSCC) and can also block L‐type Ca2þ channels.
cation channels do not have to be the dominant source of Ca2þ entry in smooth muscle cells to achieve the required rise in [Ca2þ]i for excitation‐contraction coupling. Most vascular muscles have resting potentials within the window current range for L‐type Ca2þ channels (Vogalis et al., 1991; Fleischmann et al., 1994), so small changes in membrane potential can have a significant impact on Ca2þ entry. Nonselective cation channels can also be activated by cell inflation in mesenteric resistance arteries (Setoguchi et al., 1997). These authors found that positive pressure applied to the patch pipette under whole‐cell recording conditions caused noticeable cell inflation and increased a nonselective cation conductance in arterial myocytes. The degree of conductance increase was related to the cross‐sectional area of the cells (i.e., amount of cellular inflation). The conductance activated by cell inflation depended on the Naþ gradient, and extacellular Ca2þ reduced the amount of current carried by the stretch‐activated channels. Gd3þ blocked the stretch‐activated conductance with an IC50 of 14 mM. Gdþ also blocked stretch‐dependent depolarization of myocytes. Stretch‐activated currents were also recorded from myocytes from resistance arteries of spontaneously hypertensive rats (Ohya et al., 1998). These studies showed, as with myocytes from normotensive animals, that cell inflation evoked Gd3þ‐sensitive nonselective cation currents. The conductance responsible for these currents showed greater sensitivity to stretch and was increased in amplitude in myocytes from hypertensive rats. Activation of 32‐pS cation channels in cell‐attached patches was also more sensitive to negative pipette pressure in cells from hypertensive animals. The increase in
518
Sanders and Koh
sensitivity of stretch‐activated nonselective cation channels in hypertensive muscles could lead to enhanced myogenic responses and increased arterial resistance. Physiologically, activation of nonselective cation conductances in vascular myocytes seems to be a major factor in mediating myogenic responses; however, other mechanisms may contribute. Endothelia cells, once thought to be important, appear to contribute relatively little to the responses because myogenic responses can still be observed in arteries with the endothelium removed (Nelson et al., 1997). In studies of renal arterioles, which exhibit strong myogenic responses, increased perfusion pressure caused graded aVerent arteriolar constriction. This response was largely blocked by 10‐mM Gd3þ (Takenaka et al., 1998). The myogenic response was also attenuated by reducing extracellular Naþ, which appears to be the major charge carrier for stretch‐activated channels in vascular myocytes (see above). Entry of Naþ is likely linked to depolarization and activation of voltage‐gated Ca2þ channels, and myogenic constrictions in renal resistance vessels were also reduced in Ca2þ‐free bathing solutions or by diltiazem to block L‐type Ca2þ channels. The molecular identity of stretch‐activated nonselective cation channels is unclear and this has limited progress to determine how stretch influences the open probability of channels. Recent studies have tried to link myogenic responses in smooth muscles to specific molecular entities. For example, myogenic tone and depolarization responses to elevated perfusion pressure in cerebral arteries were reduced by antisense oligonucleotides to TRPC6 (Welsh et al., 2002). Park et al. (2003) followed up on these observations by comparing the properties of mechanosensitive channels in arterial smooth muscle cells with the properties of canonical transient receptor potential channels (TRPCs). In these studies, negative pressure in the patch pipette during cell‐attached recording activated nonselective cation channels of about 30 pS. The open probability of the channels increased as a function of negative pressure. The channels were blocked by Gd3þ and an inhibitor of phospholipase C (U73122), but facilitated by diacylglycerol (DAG) and cyclopiazonic acid (CPA). In the presence of DAG or CPA, channels could not be activated without membrane stretch, showing that these drugs facilitate channel activity but stretch is an obligatory stimulus for activation. Park et al. (2003) concluded that the 30‐pS stretch‐activated channels in vascular smooth muscles have properties similar to TRPCs. Unfortunately, the criteria to link whole‐cell and single‐channel currents to TRPCs is complicated by the extremely nonspecific nature of the pharmacology of TRPCs. Further attempts to test the role of TRPCs in the myogenic response, using specific gene inactivation in smooth muscle cells and other molecular manipulations, are needed to identify the molecular correlates of stretch‐activated nonselective cation channels.
18. Stretch‐Activated Conductances in Smooth Muscles
519
Another report suggested that a vanilloid receptor homologue (TRPV2) might contribute to swelling‐activated nonselective cation currents in vascular myocytes (Muraki et al., 2003). These authors found that osmotic swelling activated a nonselective cation current and increased [Ca2þ]i in aortic myocytes. The responses were blocked by ruthenium red, a blocker of TRPV2 channels. Aortic myocytes expressed immunoreactivity for TRPV2 antibodies, and TRPV2 immunoreactivity was also observed in mesenteric and basilar arterial myocytes. Treating mouse aorta with TRPV2 antisense oligonucleotides suppressed osmotic swelling‐induced nonselective cation currents and the increase in [Ca2þ]i and reduced expression of TRPV2 protein. These authors concluded that TRPV2 is an important mechanosensor in vascular smooth muscle cells that couples membrane stretch to activation of a nonselective cation conductance. Unfortunately, experiments comparing the eYcacy of activation of TRPV2 by osmotic stretching of cells were not compared with negative pipette pressure or cell elongation, stimuli that have been linked by previous experiments to the activation of channels responsible for the myogenic responses of vascular muscles. Thus, it is unclear, at present, whether TRPV2 is a central molecular component of the myogenic response. 2. Other Mechanosensitive Inward Currents in Vascular Myocytes There have been a few reports describing mechanosensitive properties of voltage‐gated Ca2þ channels in smooth muscles. Basilar artery cells, under whole‐cell recording conditions, were exposed to positive and negative pressure applied through the pipette. In most cells, positive pressure caused visible cell inflation and increased L‐type Ca2þ currents (Langton, 1993; McCarron et al., 1997). There were similar increases in L‐type Ca2þ currents in perforated‐patch recordings when cells were swelled under hypoosmotic conditions (Langton, 1993). It is unlikely that the increase in current was due to incorporation of new channels in response to membrane stretch because there was not an increase in total capacitance while current amplitude increased by up to 50% in some experiments. While it is likely that L‐type Ca2þ currents are very important in mediating the increase in tone in the myogenic eVect in vascular muscles, stretch‐dependent depolarization appears to be more dependent on other conductances (e.g., nonselective cation channels) since stretch‐induced depolarization persists in the presence of Ca2þ channel blockade (Knot and Nelson, 1995). Stretch‐induced depolarization could also be accomplished by eZux of Cl ions if longitudinal stretch of cells is coupled to activation of anion selective conductances. The reversal potentials of the stretch‐activated conductances characterized in several studies, however, have not been consistent with activation of a Cl conductance and are independent of changes in extracellular Cl concentration (Davis et al., 1992a; Setoguchi et al., 1997; Wu and Davis, 2001). However, along with nonselective cation conductances, swelling‐activated
520
Sanders and Koh
Cl channels are found in vascular smooth muscles that could, under certain circumstances, contribute to the mechanosensitivity of these cells. Hypoosmotic solutions caused cell swelling and activation of an outwardly rectifying Cl conductance with a permeability sequence of SCN > I > Br > Cl in pulmonary and renal arterial myocytes (Yamazaki et al., 1998). Tamoxifen and 4,40 ‐diisothiocyanostilbene‐2,20 ‐disulfonic acid (DIDS) blocked the swelling‐ activated Cl current. Expression of the volume‐regulated Cl conductance was also evaluated and PCR showed that both pulmonary and renal myocytes express ClC‐3. Levels of ClC‐3 gene expression were slightly higher in pulmonary artery cells than in renal cells; however, there were no quantitative diVerences in current density in the two cell types. Further studies (e.g., gene inactivation or knockout studies) are needed to be certain that ClC‐3 is responsible for the selling‐activated Cl conductance in vascular myocytes. Nelson and colleagues (1997) showed that indanyloxyacetic acid (IAA‐94) and DIDS hyperpolarized and dilated pressurized cerebral arteries; however, niflumic acid had no eVect. The drugs tested had no eVect on vascular diameter or membrane potential when the perfusion pressure was low or when myogenic tone was absent. These observations suggested that a Cl conductance, but probably not a Ca2þ‐activated conductance, might participate in myogenic responses in some arteries. It is, however, quite diYcult to link Cl conductances to physiological responses because the drugs used to block Cl channels are notoriously nonselective. If these drugs inhibit myogenic responses, it is possible that part of the inhibitory eVect could be due to eVects on L‐type Ca2þ channels (Doughty et al., 1998; Dick et al., 1999) or nonselective cation channels (Park et al., 2003).
B. Bladder Myocytes There are also stretch‐activated channels in myocytes of the urinary bladder. Application of negative pressure to the inside of patch electrodes activated channels with a slope conductance of 39 pS and a reversal potential of 2 mV (Wellner and Isenberg, 1993a). With physiological ionic gradients, the slope conductance of the stretch‐activated channels was similar to the conductance of stretch‐activated channels in toad gastric myocytes (Kirber et al., 1988). The conductance was nonselective and carried a variety of cations with a selectivity sequence of Kþ > Naþ > Csþ > Ba2þ > Ca2þ. The presence of divalent ions reduced the unitary conductance of the channels with monovalent cations as charge carriers by about half. The single channel conductance and the reversal potential were not aVected by substitution of Cl with aspartate. The channels, like in vascular muscles, were blocked by Gd3þ at mM concentrations. Gd3þ reduced long open times of the channels. The authors considered the possibility that the stretch‐activated nonselective cation conductance in bladder
18. Stretch‐Activated Conductances in Smooth Muscles
521
myocytes might contribute to the relatively depolarized membrane potentials of intact bladder muscles (i.e., 35 mV in the guinea pig) and the occurrence of spontaneous action potentials. The open probability of stretch‐activated nonselective channels in bladder myocytes increased with hyperpolarization, which is usually considered a property of a pacemaker‐like conductance; however, the stretch‐activated channels in bladder myocytes required mechanical stress for activation even during hyperpolarization. Longitudinal stretch of guinea pig bladder myocytes by up to 20% caused significant depolarization that was dependent on the change in length (Wellner and Isenberg, 1994). The depolarization was suYcient to increase the frequency of action potentials in many cells. Under voltage clamp, stretch induced an inward current at a holding potential of 50 mV that was blocked by Gd3þ. With Kþ currents blocked, the current activated by stretch reversed at 0 mV. The stretch‐activated inward current in guinea pig myocytes adapted slowly, decaying in amplitude with time (Wellner and Isenberg, 1995). Internal dialysis of cells with solutions containing dibutyryl cAMP increased the rate of decay, and this was found to be due to activation of BK channels. The authors suggested that entry of Ca2þ through the stretch‐activated channels and secondary activation of BK channels was responsible for the slow adaptation. Thus, bladder myocytes possess a feedback mechanism to limit the excitatory eVects of stretch. Openings of stretch‐activated channels were also increased by a protein kinase A (PKA)‐ dependent mechanism, possibly phosphorylation of the channels. In some cases, treatment of the intracellular surface with dibutyryl cAMP caused openings of nonselective cation channels without application of suction. Filling of the bladder causes elongation of smooth muscle cells. The activation of stretch‐activated inward current channels would have the eVect, as in vascular muscles, of stimulating excitation–contraction coupling. As above, Wellner and Isenberg (1994) showed that the frequency of action potentials increased in response to stretch. This response, if unimpeded, would then to defeat the storage function of the bladder since filling would lead rapidly to a contractile response. New evidence suggests that the inward current activated by stretch may be complimented by activation of stretch‐sensitive Kþ channels (Section IV).
C. GI Myocytes 1. Nonselective Cation Conductances Stretch‐activated channels in smooth muscles were first identified in toad stomach myocytes (Kirber et al., 1988). Using either cell‐attached or excised patch configurations of the patch‐clamp technique, channels were activated
522
Sanders and Koh
when negative pressure was applied to the inside of patch pipettes. The current–voltage relationship for these channels showed inward rectification that was not aVected by Kþ substitution for Naþ. The apparent conductance of the stretch‐activated channels was about 60 pS with either Naþ or Kþ as charge carriers. Replacement of Cl with aspartate at the intracellular surface of patches had no eVect on reversal potentials. Addition of physiological levels of Ca2þ to the pipette solutions (i.e., to the extracellular surface of patches) greatly decreased slope conductance and unitary current amplitude. The stretch‐activated channels also carried current when Ca2þ ions were the only cations in the patch pipette, but the conductance of the channels under this condition was only about 20 pS. While the cation channels observed in smooth muscle resembled stretch‐ activated channels previously described in other cell types (cf. Guharay and Sachs, 1984; Morris, 1990, Kirber et al. (1988) must be credited with providing the first data supporting the hypothesis that stretch‐activated cation channels are responsible for depolarization and contractile responses to stretch in smooth muscles. Thus, their observations contributed significantly to the current state of understanding important physiological responses such as the myogenic response in vascular smooth muscles (Bayliss, 1902) and stretch‐sensitive depolarization in visceral smooth muscle (Bu¨lbring, 1955). Kirber et al. (1988) calculated that under physiological conditions unitary currents through stretch‐activated cation channels at the resting potentials of cells would be about 2.5 pA. Most patches recorded from contained at least two stretch‐activated channels. Thus, even a small increase in the activation of these channels would contribute significantly to the total conductance of smooth muscle cells, which normally have input resistances in excess of 1 GO. Kirber et al. (1988) suggested that inward current through stretch‐activated channels might drive membrane potential to more depolarized levels and activate voltage‐gated Ca2þ currents. Ca2þ entry through Ca2þ channels, and possibly Ca2þ entry through the stretch‐activated cation channels, might provide suYcient Ca2þ entry to initiate contraction. Further study of toad gastric myocytes also revealed expression of a hyperpolarization‐activated (HA) channel (Hisada et al., 1991). Stretching membrane patches with negative pipette pressure caused a shift in the hyperpolarization sensitivity of the HA channels. These channels were permeable to both Naþ and Kþ, but were not permeable to Cl . The activity of HA channels increased transiently after patch excision, but the ability of hyperpolarization to activate these channels disappeared within 3–5 min in excised patches. The HA channels had similar sensitivity to stretch and similar conductance (e.g., 60 pS) to the stretch‐activated channels these authors had described previously (Kirber et al., 1988). Thus, it was not easy to distinguish between channels that were and were not activated by hyperpolarization.
18. Stretch‐Activated Conductances in Smooth Muscles
523
The density of stretch‐activated channels that were not activated by hyperpolarization far exceeded the HA and stretch‐activated type of channels, and stretch activation of the former subclass of channels persisted after patch excision. These observations suggest that there are at least two populations of stretch‐activated nonselective cation channels in toad gastric myocytes: one type of these channels may also be activated or sensitized to stretch by hyperpolarization. HA and stretch‐sensitive cation channels were also activated by aluminofluoride (Hisada et al., 1993), which has actions such as activation of guanosine triphosphate (GTP)‐binding proteins and inhibition of phosphatases (Chabre, 1990). The question of whether smooth muscle stretch‐activated cation channels conduct appreciable amounts of Ca2þ and how global Ca2þ increases suYciently to accomplish excitation–contraction coupling when smooth muscle cells are stretched were addressed in elegant studies using digital imaging to monitor [Ca2þ]i levels with fura‐2 (Kirber et al., 2000). Unitary currents elicited by stretch of a patch of membrane were monitored simultaneously with [Ca2þ]i. Application of negative pressure to patch pipettes activated Ca2þ‐permeable nonselective cation channels and increased global [Ca2þ]i (Fig. 2). The Ca2þ transients recorded from cells also displayed a large focal increase in [Ca2þ]i near the tips of pipettes. When Ca2þ was buVered to low levels in the pipette solution, only the global increase in [Ca2þ]i was observed on application of negative pressure to the pipettes. In these experiments, stretch‐activated inward currents and depolarization‐induced Ca2þ entry were preserved, but the focal rise due to local Ca2þ entry was inhibited. Removal of Ca2þ from the bathing solution with Ca2þ present in the pipette caused just the opposite phenomenon: the focal rise in [Ca2þ]i was preserved, but the global transient was absent. Unloading of Ca2þ stores prior to stretch of the membrane patch greatly reduced the intensity of the focal rise in [Ca2þ]i. These studies show that at least two mechanisms contribute to the rise in [Ca2þ]i and contraction in response to stretch. Stretch‐activated channels increased [Ca2þ]i by depolarization and activation of voltage‐gated Ca2þ channels and by amplification of the focal increase in stretch‐induced [Ca2þ]i by release of Ca2þ from internal Ca2þ stores. The latter appeared due to Ca2þ‐induced Ca2þ release from ryanodine receptors that was induced by Ca2þ entering cells via the stretch‐activated nonselective cation channels. The local Ca2þ transients caused by stretch‐activated channels were later characterized by imaging with high temporal and spatial resolution (Zou et al., 2002). While it was known from previous studies that stretch‐activated nonselective cation channels conduct Ca2þ when Ca2þ is the only charge carrier, these authors provided direct evidence that appreciable Ca2þ enters cells via stretch‐ activated channels in physiological ionic gradients. With fluo‐3‐loaded cells, localized Ca2þ transients were observed at the tips of patch pipettes used to
524
Sanders and Koh
Ca2+
Pressure
Ca2+
Current
1s
50 mmHg 25 pA
Shutter
Pipette
300
0
FIGURE 2 Stretch‐activated channels in a patch of membrane lead to focal and global increases in [Ca2þ]i. [Ca2þ]i was imaged in fura‐2 AM‐loaded cells. Drawing shows recording configuration where Ca2þ was present in both the bath and pipette. First trace shows negative pressure applied to the inside of the pipette and the stretch‐activated inward currents activated by negative pressure. Third trace is a blowup of the region of the record denoted. Fourth trace shows the times at which the images below were obtained (vertical lines). Images were taken just before and after negative pressure was applied, and three sequential images were collected during the application of negative pressure (time denoted by wavy marks in fourth trace). During membrane stretch, there was an intense focal increase in [Ca2þ]i near the tip of the pipette and a smaller global increase in [Ca2þ]i. From Kirber et al. (2000).
18. Stretch‐Activated Conductances in Smooth Muscles
525
measure currents and stretch membrane patches before the global increase in [Ca2þ]i were initiated. The localized Ca2þ transients were observed even when Ca2þ stores were inactivated. The authors calculated that about 18% of the current through stretch‐activated nonselective cation channels in toad myocytes was carried by Ca2þ at membrane potentials more negative than resting potential. In the gut, muscarinic stimulation activates a Ca2þ‐facilitated nonselective cation conductance in smooth muscle cells (Benham et al., 1985). One study evaluated the sensitivity of this conductance to cell swelling and found that inward current evoked by carbachol increased in peak amplitude by about 50% in hypotonic solutions (Waniishi et al., 1997). Hypertonicity had opposite eVects. These eVects were not due to changes in muscarinic receptor binding since current activated by GTP S were also potentiated by hypotonicity. The current activated by hypotonicity was blocked by procaine and Zn (400 mM). The authors also noted that hypotonicity increased the amplitude and duration of depolarization responses to carbachol in cells under current clamp. Studies using hypotonicity to swell cells are always complicated by the possibility that cellular components might be diluted when water enters cells. In the study by Waniishi et al. (1997), there was a small shift in equilibrium potential for currents activated by muscarinic stimulation, suggesting changes in ionic concentrations. It is also possible that cell swelling applies fundamentally diVerent forces on the cytoskeleton than cell elongation and it is not clear whether cell swelling and stretch are equivalent stimuli. The mechanism for swelling activation of agonist‐sensitive nonselective cation channels was not determined and the possibility existed that stretch‐dependent Ca2þ release could be responsible for potentiation of currents during hypotonic conditions; this conductance is strongly facilitated by [Ca2þ]i (Pacaud and Bolton, 1991). The authors addressed this problem by taking steps to limit influx of Ca2þ and changes in [Ca2þ]i with strong Ca2þ buVering, but it is diYcult to entirely eliminate eVects due to changes in Ca2þ in the restricted spaces between sarcoplasmic reticulum (SR) and the plasma membranes of smooth muscle cells. If the eVects of hypotonicity in these experiments can be compared to stretch, the data suggest that responses to muscarinic stimulation could be enhanced in muscles that are elongated by filling of GI organs. This would be useful augmentation of muscarinic responses because more force might be needed to do the work of emptying contents. At present, however, neither the mechanism for activation of muscarinic‐activated nonselective cation channels nor the significance of these channels to the behavior of GI muscles is understood.
526
Sanders and Koh
2. Other Mechanosensitive Conductances Expressed in Visceral Smooth Muscle Myocytes Mechanosensitive Ca2þ currents have also been reported in human jejunal circular smooth muscle cells. L‐type Ca2þ channel currents can be enhanced in these cells by rapid bath perfusion, which was used as a stimulus of mechanosensitive responses (Farrugia et al., 1999; Holm et al., 2000). In jejunal myocytes, entry of Ca2þ was linked to activation of Ca2þ‐activated Kþ channels since increased perfusion enhanced outward current (in voltage clamp) and membrane hyperpolarization (in current clamp). Nifedipine (to block L‐type channels) or iberiotoxin (blocker of large‐conductance BK channels) blocked the development of outward current and depolarization. The authors suggested that the link between mechanosensitive Ca2þ channels and BK channels might provide feedback to limit contractile responses to stretch in jejunal smooth muscles. The ‐subunit of CaV1.2 channels (molecular basis for L‐type Ca2þ currents in smooth muscles) may have intrinsic mechanosensitivity. CaV1.2 channels were cloned from human intestine and expressed in HEK‐293 or Chinese hamster ovary cells either alone or with b2‐subunits (Lyford et al., 2002). Currents from the expressed channels were enhanced by rapid bath perfusion. When a proline‐rich domain of the C‐terminus of CaV1.2 that may facilitate interactions with integrins was removed, mechanosensitive responses were not disrupted. The authors suggested that the mechanosensitivity may reside in the pore‐forming region of the 1C‐subunit, but experiments to test this hypothesis were not provided. Swelling‐activated Cl channels are also found in visceral smooth muscle cells. For example, the ClC‐3 gene is expressed in canine colonic smooth muscle cells and a current similar to that generated by expressed ClC‐3 channels (outwardly rectifying and activated by reduced hypoosmotic extracellular solutions and inhibited by PKC) is present in colonic myocytes (Dick et al., 1998). This conductance was blocked by tamoxifen and DIDS, but niflumic acid, nicardipine, and removal of Ca2þ did not aVect currents. The swelling‐activated Cl conductance was inhibited by extracellular ATP and negatively regulated by PKC. The phorbol ester, phorbol 12,13 dibutyrate, decreased the swelling‐activated Cl and chelerythrine activated current, even in the presence of isotonic extracellular solutions. As discussed in the Section III.A.2, it is diYcult to evaluate the physiological significance of the Cl conductances because blockers have significant nonspecific eVects. For example, tamoxifen, a promising inhibitor of the swelling‐activated Cl current in colonic myocytes, blocked L‐type Ca2þ currents and DIDS had similar eVects (Dick et al., 1999). These Cl channel‐blocking drugs also inhibited delayed rectifier Kþ currents. Thus, the eVects of compounds on intact tissues would tend to be obscure.
18. Stretch‐Activated Conductances in Smooth Muscles
527
IV. MECHANOSENSITIVE CONDUCTANCES THAT GENERATE OUTWARD CURRENTS A. Vascular Muscles As discussed in Section III, a general increase in vascular luminal pressure induces depolarization (Nelson et al., 1990), and mechanosensitive channels play key functions in depolarization and contractile responses (Meininger and Davis, 1992; Setoguchi et al., 1997). In contrast to this dominant response, several studies have discussed the functional role of BK channels (Benham and Bolton, 1986; Brayden and Nelson, 1992; Nelson and Quayle, 1995; Nelson et al., 1995; Brenner et al., 2000; Pluger et al., 2000) and delayed rectifier Kþ channels (Kv, Cheong et al., 2001a,b; Albarwani et al., 2003; Ahmed et al., 2004; Cole et al., 2005) in myogenic tone and vasoconstriction. BK channels are voltage‐and Ca2þ‐dependent, but some investigators have suggested that these channels are also regulated by stretch. For example, application of negative pressure to membrane patches increased openings of BK channels in mesenteric arterial smooth muscle cells without aVecting the unitary conductance or the voltage sensitivity (Kirber et al., 1992; Dopico et al., 1994). BK channels were also activated by intracellular application of fatty acids, but this occurred only in the presence of basal channel openings. When [Ca2þ ]i is low or when membrane potential is at negative levels (i.e., when the open probability of BK channels is low), fatty acids would not enhance the open probability of BK channels (Kirber et al., 1992). Thus, stretch activation of BK channels from low open probability was independent of fatty acid generation. Pretreatment with fatty acids did not prevent activation of BK channels by stretch in coronary arterial smooth muscle cells (Lee et al., 2000), and pretreatment with albumin, which binds to free fatty acids, did not aVect activation of BK channels by stretch. These data suggest that the activation of BK by stretch is independent of fatty acid production in rabbit coronary artery. In other experiments on coronary myocytes, repetitive negative pressure pulses were applied to patches. In these experiments, BK channel activity progressively increased, but there was no increase in stretch‐ activated cation channels with repetitive negative pressure pulses. The authors suggested that this might be due to diVerent sensitivities of these channels to dynamic changes in membrane tension (Wu et al., 2003), but it is not clear how [Ca2þ]i was held constant in these experiments. To date, there have been no reports demonstrating that Kv channels express mechanosensitivity in vascular smooth muscles. Even though the activation of BK channels has been demonstrated in many vascular smooth muscle cells (SMCs), the mechanism of activation of these channels has not been clarified. Two possible mechanisms have been
528
Sanders and Koh
discussed: (1) direct activation by stretch and/or (2) inhibitory control of BK channel openings by interactions with the cytoskeleton. Evidence for cytoskeleton regulation of BK channels is based on the use of cytochalasin‐D (actin depolymerizer) and phalloidin (actin stabilizer). Treatment with cytochalasin‐ D increased the open probability of BK channels, and this eVect was reversed by treatment with phalloidin. These findings suggest that interactions with actin filaments may inhibit the open probability of BK channels and stretch might reduce this form of negative regulation (Piao et al., 2003). These are preliminary observations, however, and the actual mechanism of actin regulation of channel open probability needs further study. There has been some discussion regarding the physiological significance of stretch‐activation of BK channels. Activation of BK channels in response to stretch may oppose or provide feedback for the contractile response to cell elongation mediated by nonselective cation channel activation (Kirber et al., 1992). Thus, activation of BK channels may limit stretch‐induced vasoconstriction. There is controversy, however, about the role of BK channels in the myogenic responses in vessels of diVerent sizes. Blockers of BK channels (iberiotoxin or tetraethylammonium) did not aVect the resting diameter of arterioles (Jackson and Blair, 1998) as others have shown in small arteries (Paterno et al., 1996). Therefore, the contributory role of BK channels in stretch‐dependent responses may vary as a function of vascular region and also with species.
B. Bladder Smooth Muscle The bladder can store considerable volumes of urine before voiding becomes necessary. During filling, the bladder wall stretches to accommodate the increase in volume. The compliance of the detrusor muscle to stretch is of critical importance for the storage function of the bladder. For adequate filling to occur, the detrusor smooth muscle must be able to stretch and accommodate the increase in bladder volume without a significant rise in pressure. It is known that smooth muscle from the bladder detrusor exhibits phasic contractions in response to spontaneous action potentials or transmural nerve stimuli (Levin et al., 1986; Brading, 1992). Since spontaneous contractions occur locally and do not readily spread throughout the tissue (Hashitani et al., 2000), it has been suggested that the poor electrical coupling of detrusor smooth muscle facilitates muscle bundles to adjust their length to achieve the minimum surface area/volume ratio during bladder filling without contraction or rise in intravesicular pressure (Levin et al., 1986; Kinder and Mundy, 1987; Brading, 1994).
18. Stretch‐Activated Conductances in Smooth Muscles
529
As discussed above, stretch of bladder myocytes activates nonselective cation channels in guinea pig myocytes (Wellner and Isenberg, 1993a,b). This mechanism, which is excitatory in nature, does not correlate with the observation that stretch does not activate contraction in bladder but leads to stabilization of membrane potential or even relaxation. Thus, influx and/or release of Ca2þ must couple to other mechanisms that blunt the excitatory responses to stretch. Ji et al. (2002) failed to find stretch‐activated nonselective channels in mouse and rabbit bladder myocytes. Instead, stretching these muscles caused gating of ryanodine receptors, release of Ca2þ from the SR, and generation of Ca2þ sparks or propagated Ca2þ waves. This mechanism coupled to activation of a Ca2þ‐activated Cl conductance and occurred in the absence of Ca2þ influx. Activation of Cl channels would have much the same eVect as activation of nonselective cation channels (i.e., depolarization and enhanced open probability of voltage‐gated Ca2þ channels; Ji et al., 2002). This mechanism also fails to explain the stabilization of membrane potential in response to filling in the bladder. Phosphorylation of BK channels might increase coupling between Ca2þ sparks and BK channels in bladder myocytes (Wellner and Isenberg, 1993a). This might switch the coupling of sparks from activation of net inward to net outward current, but this phenomenon has not been demonstrated. BK channels clearly participate in the regulation of membrane potential and repolarization of action potentials in bladder smooth muscles (Heppner et al., 1997; Hashitani and Brading, 2003), but this does not mean BK channels are involved in stabilization of membrane potential during stretch‐ dependent responses unless it can be shown that a specific stretch‐dependent mechanism regulates BK channel open probability in bladder myocytes. Other mechanisms, such as other classes of stretch‐activated Kþ channels (see below) are likely to mediate the membrane stabilization response to stretch in bladder muscles.
C. Uterine Smooth Muscle The uterus is a unique smooth muscle organ that undergoes extreme changes in volume during pregnancy. Muscle quiescence during pregnancy is another example of stabilization of excitability during conditions of stretch. How muscle relaxation is maintained during pregnancy has been investigated widely, and some studies have suggested that the expression of BK channels is regulated by female steroid hormones (Mironneau and Savineau, 1980; Kihira et al., 1990; Toro et al., 1990; Anwer et al., 1992; Khan et al., 1993). The functional role of BK channels and the response to
530
Sanders and Koh
stretch in controlling basal contractility is controversial (Khan et al., 1993, 1997; Aaronson et al., 2006). Recently, there is evidence showing that K2P channels are expressed in myometrium. K2P channels are a family of Kþ channels that may contribute to regulation of membrane potential in both electrically excitable and nonexcitable cells.Within the K2P channel family, TREK‐1, TREK‐2, and TRAAK have unique functional properties and represent mechanosensitive Kþ channels (Fink et al., 1996; Patel et al., 1998; Maingret et al., 1999a,b; Lesage and Lazdunski, 2000; Lesage et al., 2000). The expression of TREK‐1 in human myometrium has been shown in molecular, immunoblot, and immunohistochemistry studies (Bai et al., 2005). Transcripts for TREK‐1 occur at higher levels than for TREK‐2, and a significant increase in TREK‐1 expression was seen in pregnant tissues (Tichenor et al., 2005). These data suggest the hypothesis TREK channels could participate in the regulation of muscle excitability during pregnancy, but this hypothesis has not yet been tested.
D. GI Smooth Muscle There are mechanosensitive inward current channels GI myocytes (as described in Section III.C), but contraction is not the dominant response to stretch in many regions of the GI tract. For example, the colon and fundus display a reservoir function that allows volume expansion (muscle elongation) without increasing intraluminal pressure. Filling of these organs does not cause myogenic contraction. Thus, in addition to neural reflexes that contribute to volume accommodation, a myogenic mechanism is likely to exist that stabilizes membrane potential and limits excitability during cell elongation. Stretch‐activated Kþ channels have been observed in studies of toad gastric myocytes (Petrou et al., 1994; Ordway et al., 1995; Kirber et al., 2000), guinea pig gastric myocytes (Li et al., 2002), and murine and canine colonic myocytes (Koh and Sanders, 2001). Single‐channel studies in toad gastric myocytes revealed 20‐pS Kþ channel at 0 mV in an asymmetrical Kþ gradient (3/130 mM) that were activated by negative pipette pressure. Open probability was augmented by fatty acids. These authors suggested that fatty acids may be liberated from membrane phospholipids in response to stretch to enhance Kþ channel activity. The properties of the channels in toad gastric myocyte channels are similar to TREK‐1 channel, in terms of the activation by fatty acids, stretch, and intracellular acidic pH (Zou et al., 2001). In murine colonic myocytes, negative patch pipette pressure activated Kþ channels that were voltage‐independent with a slope conductance of
531
18. Stretch‐Activated Conductances in Smooth Muscles −40 cmH2O
A −20 cmH2O −20 cmH2O
−40 cmH2O
−60 cmH2O
−60 cmH2O
I−O
−80 cmH2O
10 s 10 pA Probability density
B
1.0 0.8 0.6 0.4 0.2 0.0 −80
−60
−40 cmH2O
−20
0
FIGURE 3 Relationship between pressure and open probability of channels activated by negative pressure in murine colonic myocytes. Patches from five cells were exposed to pressure ranging from 20 to 80 cmH2O. In order to be sure that the eVects of negative pressure were reversible and lacked desensitization, diVerent levels of negative pressure were applied to the same patch and each pressure was tested twice. Panel A, a negative pressure of 20 cmH2O had little eVect on channel activity. However, greater negative pressures ( 40 cmH2O) applied to the same patch increased NPo to 6.2. Further negative pressure ( 60 and 80 cmH2O) increased NPo to the maximal level. After restoration of atmospheric pressure in each step, the open probability returned to near zero. After application of pressure pulses, the patch was excised. This caused maximal activation of channels in the patch. Panel B, the graph summarizes the relationship between pressure and NPo in patches from five cells. I‐O denotes inside‐out patches. Figure redrawn from Koh and Sanders (2001).
95 pS in symmetrical Kþ gradients (Fig. 3). The eVects of negative pressure on open probability were graded as a function of pressure and reversed when atmospheric pressure was restored. Cell elongation also caused activation of
532
Sanders and Koh A
B
C −60 cmH2O
10 mm Elongation
2s 10 pA FIGURE 4 Activation of stretch‐dependent Kþ (SDK) channels via cell elongation in murine colonic myocytes. On‐cell patches of murine colonic myocytes were held at 0 mV in asymmetrical Kþ gradients (5/140 mM). The pipette solution contained charybdotoxin (200 nM) to inhibit large conductance Ca2þ‐activated Kþ channels. Panel A, two patch pipettes were sealed to the same cell. Single‐channel currents were measured via one pipette, and the other pipette was used to stretch the cell. Panels B and C, after confirming that negative patch pressure ( 60 cmH2O) activated SDK channels in this patch, the cells were elongated (in this example by 8 mm). Cell elongation caused activation of channels with the same properties as negative pressure. Figure redrawn from Koh and Sanders (2001).
Kþ channels with the same properties as those activated by negative pressure (Koh and Sanders, 2001) (Fig. 4). The stretch‐activated channels were maximally activated by patch excision, suggesting that either an intracellular messenger or interactions between the channels and the cytoskeleton regulate open probability. Sodium nitroprusside (SNP), an NO donor, activates stretch‐activated Kþ channels in smooth muscle myocytes, and 8‐Br‐cGMP mimics these eVects (Koh and Sanders, 2001). Thus, the eVects of NO may be mediated through activation of cGMP‐dependent protein kinase (PKG). These findings revealed a novel property of stretch‐activated Kþ channels in GI myocytes: the activity of these channels may be dually regulated by both stretch and the primary enteric inhibitory neurotransmitter, NO. The hypothesis that K2P channels might be responsible for the stretch‐ activated Kþ conductance in visceral myocytes was investigated. Six functional subfamilies of K2P channels have been described (e.g., TWIK, TREK, TASK, TASK‐2, THIK, and TRESK), and these are classified based on functional domains (Franks and Honore, 2004). TREK family channels are activated by stretch and include TREK‐1, TREK‐2, and TRAAK (KCNK2,
18. Stretch‐Activated Conductances in Smooth Muscles
533
KCNK4, and KCNK10, respectively). RT‐PCR revealed that TREK channels, but not TRAAK channels, are expressed in murine colonic smooth muscle (Koh et al., 2001). TREK channels appear to be good molecular candidates for stretch‐activated Kþ channels (Sanders and Koh, 2006), based on the following similarities between homologously expressed TREK‐1 channels and native stretch‐activated Kþ channels: both channels exhibit similar unitary conductances, pharmacology, stretch sensitivity, and regulation by NO, cGMP, and PKG (Koh et al., 2001). TREK‐1 channels are inhibited by PKA and PKC. Maingret et al. (2002) used site‐directed mutagenesis of a PKA consensus sequence in TREK‐1 and showed that mutation at S333 abolished regulation by PKA. Exposure of cells expressing murine TREK‐1 channels to SNP or 8‐Br‐cGMP, however, resulted in a sustained increase in open probability of mTREK‐1 channels. Mutation of the PKG consensus site (S351A) in mTREK‐1 eliminated the increase in open probability evoked by SNP or 8‐Br‐cGMP (Koh et al., 2001). S351 is potentially phosphorylated by PKG or PKA, but the data demonstrated that the initial decrease in channel activity after exposure to 8‐Br‐cAMP is due to PKA phosphorylation at S333, and channel activation is due to PKG or PKA phosphorylation at S351. The initial decrease in channel activity on exposure to 8‐Br‐cAMP, due to phosphorylation at S333, remained intact in S351A mutant channels. 8‐Br‐cGMP did not cause an initial or sustained decrease in open probability of either wild‐type or S351A channels, suggesting that PKG cannot phosphorylate S333. Thus, phosphorylation near the C‐terminus of TREK‐1, via diVerent second‐messenger signaling pathways, diVerentially regulates and finely tunes channel open probability. Stretch of cell membranes is an activator of TREK family channels (Fig. 5), but others stimuli are also eVective, including shear stress, cell swelling, and negative pressure in patch pipettes (Maingret et al., 1999a,b; Patel et al., 2001). Disruption of the cytoskeleton by either biochemical or mechanical means aVects responses of TREK channels to stretch, and it appears that these channels are normally inhibited by the cytoskeleton (Lesage and Lazdunski, 2000). Regulation of the channels through interactions with the cytoskeleton may be the predominant mechanism that increases channel open probability when membranes are excised from cells (Sanders and Koh, 2006). The specific domain of K2P channels that confers mechanosensitivity has not been determined. Indeed, studies suggest that two separate domains may be involved: one domain may facilitate cytoskeletal interactions and another domain may actively sense membrane stretch (Kim et al., 2001; Lauritzen et al., 2005). Lauritzen et al. (2005) have studied the associations between channels and the cytoskeleton in TREK‐1 channel protein by performing mutagenesis on residues within the C‐terminus. Even in the presence of these mutations, sensitivity to stretch was retained. Kim et al. (2001) made chimeric
534
Sanders and Koh mTREK-1
B
Nontransfected
C
I−V 3 I (nA)
A
2 1
D
Control
E
Stretch
F
I−V 4 I (nA)
500 pA 50 ms
3 2 1
−80
−40
0 40 V (mV)
80
FIGURE 5 Characterization of the mTREK‐1 current expressed in COS cells. Currents were recorded in whole‐cell voltage clamp. The membrane potential was stepped from 80 to þ70 mV in 10‐mV increments for 400 ms. COS cells were stretched by micromanipulator. Panel A shows representative currents mTREK‐1 transfected COS cells. Panel B shows nontransfected COS cells. Panel C shows average current–voltage (I‐V) relationships in transfected () and nontransfected (d) COS cells. Panel D shows representative currents mTREK‐1 transfected COS cells. Panel E is after cell elongation using manipulator. Panel F shows the average current‐voltage relationship in control (s) and after stretch (d). Figure redrawn from Koh et al. (2001).
channels to define functionally important regions. These authors found that replacing the C‐terminal tail of mechanosensitive channels with nonmechanosensitive channels caused loss of arachidonic acid regulation, but the channels retained sensitivity to stretch. Thus, more mutational studies are needed to understand the basis for mechanosensitivity in this class of channels. The role of stretch‐activated Kþ channels (and specifically the role of TREK family channels) in physiological responses has been complicated by the lack of highly specific blockers of these channels. The channels in toad myocytes were inhibited by very high concentrations of TEA (25 mM), but stretch‐dependent Kþ channels in colonic myocytes were not blocked by TEA (up to 10 mM) or 4‐aminopyridine (5 mM). Stretch‐activated Kþ channels in colonic myocytes were also insensitive to changes in [Ca2þ]i (Koh and Sanders, 2001). We found that sulfur‐containing amino acids (L‐cysteine, L‐methionine, or DL‐homocysteine) block stretch‐activated
18. Stretch‐Activated Conductances in Smooth Muscles
535
Kþ channels in GI muscles and myocytes and inhibited hyperpolarization responses to nitrergic nerve stimulation (Park et al., 2005). These compounds also blocked expressed TREK‐1 currents in COS cells. L‐methionine was the most selective blocker of this series of compounds, and it had no eVect on the other types of Kþ channels expressed in colonic myocytes. In conclusion, investigation of mechanosensitive Kþ channels is an exciting new area of research for the regulation of function in visceral smooth muscles. The possibility that these channels are one of the main eVectors for nitrergic enteric inhibitory neural regulation is potentially of fundamental importance to understanding the regulation of whole organ function and how volume accommodation and reservoir functions occur in whole visceral organs. A better understanding of these channels, interactions with cytoskeletal elements, and changes in expression during development and pathophysiological circumstances may provide new methods of therapy for motility disorders. References Aaronson, P. I., Sarwar, U., Gin, S., Rockenbauch, U., Connolly, M., Tillet, A., Watson, S., Liu, B., and Tribe, R. M. (2006). A role for voltage‐gated, but not Ca2þ‐activated, Kþ channels in regulating spontaneous contractile activity in myometrium from virgin and pregnant rats. Br. J. Pharmacol. 147, 815–824. Ahmed, A., Waters, C. M., LeZer, C. W., and Jaggar, J. H. (2004). Ionic mechanisms mediating the myogenic response in newborn porcine cerebral arteries. Am. J. Physiol. Heart Circ. Physiol. 287, H2061–H2069. Albarwani, S., Nemetz, L. T., Madden, J. A., Tobin, A. A., England, S. K., Pratt, P. F., and Rusch, N. J. (2003). Voltage‐gated Kþ channels in rat small cerebral arteries: Molecular identity of the functional channels. J. Physiol. 551, 751–763. Anwer, K., Toro, L., Oberti, C., Stefani, E., and Sanborn, B. M. (1992). Ca2þ‐activated Kþ channels in pregnant rat myometrium: Modulation by a beta‐adrenergic agent. Am. J. Physiol. 263, C1049–C1056. Bai, X., Bugg, G. J., Greenwood, S. L., Glazier, J. D., Sibley, C. P., Baker, P. N., Taggart, M. J., and Fyfe, G. K. (2005). Expression of TASK and TREK, two‐pore domain Kþ channels, in human myometrium. Reproduction 129, 525–530. Bayliss, W. N. (1902). On the local reactions of the arterial wall to changes of internal pressure. J. Physiol. 28, 220–231. Benham, C. D., and Bolton, T. B. (1986). Spontaneous transient outward currents in single visceral and vascular smooth muscle cells of the rabbit. J. Physiol. 381, 385–406. Benham, C. D., Bolton, T. B., and Lang, R. J. (1985). Acetylcholine activates an inward current in single mammalian smooth muscle activates an inward current in single mammalian smooth muscle cells. Nature (Lond.) 316, 345–347. Brading, A. F. (1992). Ion channels and control of contractile activity in urinary bladder smooth muscle. Jpn. J. Pharmacol. 58(Suppl. 2), 120P–127P. Brading, A. F. (1994). The pathophysiological changes in the bladder obstructed by benign prostatic hyperplasia. Br. J. Urol. 74, 133. Brayden, J. E. (2002). Functional roles of KATP channels in vascular smooth muscle. Clin. Exp. Pharmacol. Physiol. 29, 312–316.
536
Sanders and Koh
Brayden, J. E., and Nelson, M. T. (1992). Regulation of arterial tone by activation of calcium‐ dependent potassium channels. Science 256, 532–535. Brenner, R., Perez, G. J., Bonev, A. D., Eckman, D. M., Kosek, J. C., Wiler, S. W., Patterson, A. J., Nelson, M. T., and Aldrich, R. W. (2000). Vasoregulation by the beta1 subunit of the calcium‐activated potassium channel. Nature 407, 870–876. Bu¨lbring, E. (1955). Correlation between membrane potential, spike discharge and tension in smooth muscle. J. Physiol. 128, 200–221. Calderone, V. (2002). Large‐conductance, Ca2þ‐activated Kþ channels: Function, pharmacology and drugs. Curr. Med. Chem. 9, 1385–1395. Chabre, M. (1990). Aluminofluoride and beryllofluoride complexes: A new phosphate analogs in enzymology. Trends Biochem. Sci. 15, 6–10. Cheong, A., Dedman, A. M., Xu, S. Z., and Beech, D. J. (2001a). KV alpha1 channels in murine arterioles: DiVerential cellular expression and regulation of diameter. Am. J. Physiol. Heart Circ. Physiol. 281, H1057–H1065. Cheong, A., Dedman, A. M., and Beech, D. J. (2001b). Expression and function of native potassium channel [K(V)alpha1] subunits in terminal arterioles of rabbit. J. Physiol. 534, 691–700. Cole, W. C., Chen, T. T., and Clement‐Chomienne, O. (2005). Myogenic regulation of arterial diameter: role of potassium channels with a focus on delayed rectifier potassium current. Can. J. Physiol. Pharmacol. 83, 755–765. Davis, M. J., Donovitz, J. A., and Hood, J. D. (1992a). Stretch‐activated single‐channel and whole cell currents in vascular smooth muscle cells. Am. J. Physiol. 262, C1083–C1088. Davis, M. J., Meininger, G. A., and Zawieja, D. C. (1992b). Stretch‐induced increases in intracellular calcium of isolated vascular smooth muscle cells. Am. J. Physiol. 263, H1292–H1299. Dick, G. M., Bradley, K. K., Horowitz, B., Hume, J. R., and Sanders, K. M. (1998). Functional and molecular identification of a novel chloride conductance in canine colonic smooth muscle. Am. J. Physiol. 275, C940–C950. Dick, G. M., Kong, I. D., and Sanders, K. M. (1999). EVects of anion channel antagonists in canine colonic myocytes: Comparative pharmacology of Cl , Ca2þ and Kþ currents. Br. J. Pharmacol. 127, 1819–1831. Dopico, A. M., Kirber, M. T., Singer, J. J., and Walsh, J. V., Jr. (1994). Membrane stretch directly activates large conductance Ca2þ‐activated Kþ channels in mesenteric artery smooth muscle cells. Am. J. Hypertens. 7, 82–89. Doughty, J. M., Miller, A. L., and Langton, P. D. (1998). Non‐specificity of chloride channel blockers in rat cerebral arteries: Block of the L‐type calcium channel. J. Physiol. (Lond.) 507, 433–439. Farrugia, G., Holm, A. N., Rich, A., Sarr, M. G., Szurszewski, J. H., and Rae, J. L. (1999). A mechanosensitive calcium channel in human intestinal smooth muscle cells. Gastroenterology 117, 900–905. Fink, M., Duprat, F., Lesage, F., Reyes, R., Romey, G., Heurteaux, C., and Lazdunski, M. (1996). Cloning, functional expression and brain localization of a novel unconventional outward rectifier Kþ channel. EMBO J. 15, 6854–6862. Fleischmann, B. K., Murray, R. K., and KotlikoV, M. I. (1994). Voltage window for sustained elevation of cytosolic calcium in smooth muscle cells. Proc. Natl. Acad. Sci. USA 91, 11914–11918. Franks, N. P., and Honore, E. (2004). The TREK K2P channels and their role in general anaesthesia and neuroprotection. Trends Pharmacol. Sci. 25, 601–608. Guharay, F., and Sachs, F. (1984). Stretch‐activated single ion channel currents in tissue‐cultured embryonic chick skeletal muscle. J. Physiol. 352, 685–701.
18. Stretch‐Activated Conductances in Smooth Muscles
537
Harder, D. R. (1984). Pressure‐dependent membrane depolarization in cat middle cerebral artery. Circ. Res. 55, 197–202. Hashitani, H., and Brading, A. F. (2003). Electrical properties of detrusor smooth muscles from the pig and human urinary bladder. Br. J. Pharmacol. 140, 146–158. Hashitani, H., Bramich, N. J., and Hirst, G. D. (2000). Mechanisms of excitatory neuromuscular transmission in the guinea‐pig urinary bladder. J. Physiol. 524, 565–579. Heppner, T. J., Bonev, A. D., and Nelson, M. T. (1997). Ca2þ‐activated Kþ channels regulate action potential repolarization in urinary bladder smooth muscle. Am. J. Physiol. 273, C110–C117. Hisada, T., Ordway, R. W., Kirber, M. T., Singer, J. J., and Walsh, J. V., Jr. (1991). Hyperpolarization‐activated cationic channels in smooth muscle cells are stretch sensitive. Pflugers Arch. 417, 493–499. Hisada, T., Singer, J. J., and Walsh, J. V., Jr. (1993). Aluminofluoride activates hyperpolarization‐ and stretch‐activated cationic channels in single smooth muscle cells. Pflugers Arch. 422, 397–400. Holm, A. N., Rich, A., Sarr, M. G., and Farrugia, G. (2000). Whole cell current and membrane potential regulation by a human smooth muscle mechanosensitive calcium channel. Am. J. Physiol. Gastrointest. Liver Physiol. 279, G1155–G1161. Jackson, W. F. (2005). Potassium channels in the peripheral microcirculation. Microcirculation 12, 113–127. Jackson, W. F., and Blair, K. L. (1998). Characterization and function of Ca2þ‐activated Kþ channels in arteriolar muscle cells. Am. J. Physiol. 274, H27–H34. Ji, G., Barsotti, R. J., Feldman, M. E., and KotlikoV, M. I. (2002). Stretch‐induced calcium release in smooth muscle. J. Gen. Physiol. 119, 533–544. Khan, R. N., Smith, S. K., Morrison, J. J., and Ashford, M. L. (1993). Properties of large‐ conductance Kþ channels in human myometrium during pregnancy and labour. Proc. Biol. Sci. 251, 9–15. Khan, R. N., Smith, S. K., Morrison, J. J., and Ashford, M. L. (1997). Ca2þ dependence and pharmacology of large‐conductance Kþ channels in nonlabor and labor human uterine myocytes. Am. J. Physiol. 273, C1721–C1731. Kihira, M., Matsuzawa, K., Tokuno, H., and Tomita, T. (1990). EVects of calmodulin antagonists on calcium‐activated potassium channels in pregnant rat myometrium. Br. J. Pharmacol. 100, 353–359. Kim, Y., Gnatenco, C., Bang, H., and Kim, D. (2001). Localization of TREK‐2 Kþ channel domains that regulate channel kinetics and sensitivity to pressure, fatty acids and pHi. Pflugers Arch. 442, 952–960. Kinder, R. B., and Mundy, A. R. (1987). Pathophysiology of idiopathic detrusor instability and detrusor hyper‐reflexia. An in vitro study of human detrusor muscle. Br. J. Urol. 60, 509–515. Kirber, M. T., Walsh, J. V., Jr., and Singer, J. J. (1988). Stretch‐activated ion channels in smooth muscle: A mechanism for the initiation of stretch‐induced contraction. Pflu¨gers Arch. 412, 339–345. Kirber, M. T., Ordway, R. W., Clapp, L. H., Walsh, J. V., Jr., and Singer, J. J. (1992). Both membrane stretch and fatty acids directly activate large conductance Ca2þ‐activated Kþ channels in vascular smooth muscle cells. FEBS Lett. 297, 24–28. Kirber, M. T., Guerrero‐Hernandez, A., Bowman, D. S., Fogarty, K. E., Tuft, R. A., Singer, J. J., and Fay, F. S. (2000). Multiple pathways responsible for the stretch‐induced increase in Ca2þ concentration in toad stomach smooth muscle cells. J. Physiol. 524, 3–17.
538
Sanders and Koh
Knot, H. J., and Nelson, M. T. (1995). Regulation of membrane potential and diameter by voltage‐dependent Kþ channels in rabbit myogenic cerebral arteries. Am. J. Physiol. Heart Circ. Physiol. 269, H348–H355. Koh, S. D., and Sanders, K. M. (2001). Stretch‐dependent potassium channels in murine colonic smooth muscle cells. J. Physiol. 533, 155–163. Koh, S. D., Monaghan, K., Sergeant, G. P., Ro, S., Walker, R. L., Sanders, K. M., and Horowitz, B. (2001). TREK‐1 regulation by nitric oxide and CGMP dependent protein kinase: An essential role in smooth muscle inhibitory neurotransmission. J. Biol. Chem. 276, 44338–44346. Langton, P. D. (1993). Calcium channel currents recorded from isolated myocytes of rat basilar artery are stretch sensitive. J. Physiol. (Lond.) 471, 1–11. Lauritzen, I., Chemin, J., Honore, E., Jodar, M., Guy, N., Lazdunski, M., and Jane, P. A. (2005). Cross‐talk between the mechano‐gated K2P channel TREK‐1 and the actin cytoskeleton. EMBO Rep. 6, 642–648. Lee, C. J., Kwon, S., Lee, Y. H., Ahn, D. S., and Kang, B. S. (2000). Membrane stretch increases the activity of Ca2þ‐activated Kþ channels in rabbit coronary vascular smooth muscles. Yonsei Med. J. 41, 266–272. Lesage, F., and Lazdunski, M. (2000). Molecular and functional properties of two‐pore‐domain potassium channels. Am. J. Physiol. Renal Physiol. 279, F793–F801. Lesage, F., Maingret, F., and Lazdunski, M. (2000). Cloning and expression of human TRAAK, a polyunsaturated fatty acids‐activated and mechano‐sensitive Kþ channel. FEBS Lett. 471, 137–140. Levin, R. M., Ruggieri, M. R., Velagapudi, S., Gordon, D., Altman, B., and Wein, A. J. (1986). Relevance of spontaneous activity to urinary bladder function: An in vitro and in vivo study. J. Urol. 136, 517–521. Li, L., Jin, N. G., Piao, L., Hong, M. Y., Jin, Z. Y., Li, Y., and Xu, W. X. (2002). Hyposmotic membrane stretch potentiated muscarinic receptor agonist‐induced depolarization of membrane potential in guinea‐pig gastric myocytes. World J. Gastroenterol. 8, 724–727. Lyford, G. L., Strege, P. R., Shepard, A., Ou, Y., Ermilov, L., Miller, S. M., Gibbons, S. J., Rae, J. L., Szurszewski, J. H., and Farrugia, G. (2002). alpha(1C) (Ca(V)1.2) L‐type calcium channel mediates mechanosensitive calcium regulation. Am. J. Physiol. Cell Physiol. 283, C1001–C1008. Maingret, F., Fosset, M., Lesage, F., Lazdunski, M., and Honore, E. (1999a). TRAAK is a mammalian neuronal mechano‐gated Kþ channel. J. Biol. Chem. 274, 1381–1387. Maingret, F., Patel, A. J., Lesage, F., Lazdunski, M., and Honore, E. (1999b). Mechano‐ or acid stimulation, two interactive modes of activation of the TREK‐1 potassium channel. J. Biol. Chem. 274, 26691–26696. Maingret, F., Honore, E., Lazdunski, M., and Patel, A. J. (2002). Molecular basis of the voltage‐ dependent gating of TREK‐1, a mechano‐sensitive K(þ) channel. Biochem. Biophys. Res. Commun. 292, 339–346. McCarron, J. G., Crichton, C. A., Langton, P. D., Mackenzie, A., and Smith, G. L. (1997). Myogenic contraction by modulation of voltage‐dependent calcium currents in isolated rat cerebral arteries. J. Physiol. (Lond.) 498, 371–379. Meininger, G. A., and Davis, M. J. (1992). Cellular mechanisms involved in the vascular myogenic response. Am. J. Physiol. 263, H647–H659. Mironneau, J., and Savineau, J. P. (1980). EVects of calcium ions on outward membrane currents in rat uterine smooth muscle. J. Physiol. 302, 411–425. Morris, C. E. (1990). Mechanosensitive ion channels. J. Membr. Biol. 113, 93–107.
18. Stretch‐Activated Conductances in Smooth Muscles
539
Muraki, K., Iwata, Y., Katanosaka, Y., Ito, T., Ohya, S., Shigekawa, M., and Imaizumi, Y. (2003). TRPV2 is a component of osmotically sensitive cation channels in murine aortic myocytes. Circ. Res. 93, 829–838. Nelson, M. T., and Quayle, J. M. (1995). Physiological roles and properties of potassium channels in arterial smooth muscle. Am. J. Physiol. 268, C799–C822. Nelson, M. T., Patlak, J. B., Worley, J. F., and Standen, N. B. (1990). Calcium channels, potassium channels, and voltage dependence of arterial smooth muscle tone. Am. J. Physiol. 259, C3–C18. Nelson, M. T., Cheng, H., Rubart, M., Santana, L. F., Bonev, A. D., Knot, H. J., and Lederer, W. J. (1995). Relaxation of arterial smooth muscle by calcium sparks. Science 270, 633–637. Nelson, M. T., Conway, M. A., Knot, H. J., and Brayden, J. E. (1997). Chloride channel blockers inhibit myogenic tone in rat cerebral arteries. J. Physiol. (Lond.) 502, 259–264. Ohya, Y., Adachi, N., Nakamura, Y., Setoguchi, M., Abe, I., and Fujishima, M. (1998). Stretch‐activated channels in arterial smooth muscle of genetic hypertensive rats. Hypertension 31, 254–258. Ordway, R. W., Petrou, S., Kirber, M. T., Walsh, J. V., Jr., and Singer, J. J. (1995). Stretch activation of a toad smooth muscle Kþ channel may be mediated by fatty acids. J. Physiol. (Lond.) 484, 331–337. Pacaud, P., and Bolton, T. B. (1991). Relation between muscarinic receptor cationic current and internal calcium in jejunal smooth muscle cells. J. Physiol. 441, 477–499. Park, K. J., Baker, S. A., Cho, S. Y., Sanders, K. M., and Koh, S. D. (2005). Sulfur‐containing amino acids block stretch‐dependent Kþ channels and nitrergic responses in the murine colon. Br. J. Pharmacol. 144, 1126–1137. Park, K. S., Kim, Y., Lee, Y. H., Earm, Y. E., and Ho, W. K. (2003). Mechanosensitive cation channels in arterial smooth muscle cells are activated by diacylglycerol and inhibited by phospholipase C inhibitor. Circ. Res. 93, 557–564. Patel, A. J., Honore, E., Maingret, F., Lesage, F., Fink, M., Duprat, F., and Lazdunski, M. (1998). A mammalian two pore domain mechano‐gated S‐like Kþ channel. EMBO J. 17, 4283–4290. Patel, A. J., Lazdunski, M., and Honore, E. (2001). Lipid and mechano‐gated 2P domain K(þ) channels. Curr. Opin. Cell Biol. 13, 422–428. Paterno, R., Faraci, F. M., and Heistad, D. D. (1996). Role of Ca2þ‐dependent Kþ channels in cerebral vasodilatation induced by increases in cyclic GMP and cyclic AMP in the rat. Stroke 27, 1603–1607. Petrou, S., Ordway, R. W., Hamilton, J. A., Walsh, J. V., Jr., and Singer, J. J. (1994). Structural requirements for charged lipid molecules to directly increase or suppress Kþ channel activity in smooth muscle cells. EVects of fatty acids, lysophosphatidate, acyl coenzyme A and sphingosine. J. Gen. Physiol. 103, 471–486. Piao, L., Ho, W. K., and Earm, Y. E. (2003). Actin filaments regulate the stretch sensitivity of large‐conductance, Ca2þ‐activated Kþ channels in coronary artery smooth muscle cells. Pflugers Arch. 446, 523–528. Pluger, S., Faulhaber, J., Furstenau, M., Lohn, M., Waldschutz, R., Gollasch, M., Haller, H., Luft, F. C., Ehmke, H., and Pongs, O. (2000). Mice with disrupted BK channel beta1 subunit gene feature abnormal Ca2þ spark/STOC coupling and elevated blood pressure. Circ. Res. 87, E53–E60. Sanders, K. M., and Koh, S. D. (2006). Two‐pore‐domain potassium channels in smooth muscles: New components of myogenic regulation. J. Physiol. 570, 37–43. Setoguchi, M., Ohya, Y., Abe, I., and Fujishima, M. (1997). Stretch‐activated whole‐cell currents in smooth muscle cells from mesenteric resistance artery of guinea‐pig. J. Physiol (Lond.) 501, 343–353.
540
Sanders and Koh
Standen, N. B., and Quayle, J. M. (1998). Kþ channel modulation in arterial smooth muscle. Acta. Physiol. Scand. 164, 549–557. Takenaka, T., Suzuki, H., Okada, H., Hayashi, K., Kanno, Y., and Saruta, T. (1998). Mechanosensitive cation channels mediate aVerent arteriolar myogenic constriction in the isolated rat kidney. J. Physiol. 511, 245–253. Tichenor, J. N., Hansen, E. T., and Buxton, I. L. (2005). Expression of stretch‐activated potassium channels in human myometrium. Proc. West. Pharmacol. Soc. 48, 44–48. Toro, L., Ramos‐Franco, J., and Stefani, E. (1990). GTP‐dependent regulation of myometrial KCa channels incorporated into lipid bilayers. J. Gen. Physiol. 96, 373–394. Vogalis, F., Publicover, N. G., Hume, J. R., and Sanders, K. M. (1991). Relationship between calcium current and cytosolic calcium in canine gastric smooth muscle cells. Am. J. Physiol. 260, C1012–C1018. Waniishi, Y., Inoue, R., and Ito, Y. (1997). Preferential potentiation by hypotonic cell swelling of muscarinic cation current in guinea pig ileum. Am. J. Physiol. 272, C240–C253. Wellner, M. C., and Isenberg, G. (1993a). Properties of stretch‐activated channels in myocytes from the guinea‐pig urinary bladder. J. Physiol. 466, 213–227. Wellner, M. C., and Isenberg, G. (1993b). Stretch‐activated nonselective cation channels in urinary bladder myocytes: Importance for pacemaker potentials and myogenic response. EXS 66, 93–99. Wellner, M. C., and Isenberg, G. (1994). Stretch eVects on whole‐cell currents of guinea‐pig urinary bladder myocytes. J. Physiol. (Lond.) 480, 439–448. Wellner, M. C., and Isenberg, G. (1995). cAMP accelerates the decay of stretch‐activated inward currents in guinea‐pig urinary bladder myocytes. J. Physiol. 482, 141–156. Welsh, D. G., Morielli, A. D., Nelson, M. T., and Brayden, J. E. (2002). Transient receptor potential channels regulate myogenic tone of resistance arteries. Circ. Res. 90, 248–250. Wu, S. N., Lin, P. H., Hsieh, K. S., Liu, Y. C., and Chiang, H. T. (2003). Behavior of nonselective cation channels and large‐conductance Ca2þ‐activated Kþ channels induced by dynamic changes in membrane stretch in cultured smooth muscle cells of human coronary artery. J. Cardiovasc. Electrophysiol. 14, 44–51. Wu, X., and Davis, M. J. (2001). Characterization of stretch‐activated cation current in coronary smooth muscle cells. Am. J. Physiol. Heart Circ. Physiol. 280, H1751–H1761. Yamazaki, J., Duan, D., Janiak, R., Kuenzli, K., Horowitz, B., and Hume, J. R. (1998). Functional and molecular expression of volume‐regulated chloride channels in canine vascular smooth muscle cells. J. Physiol. 507, 729–736. Yang, X. C., and Sachs, F. (1989). Block of stretch‐activated channels in Xenopus oocytes by gadolinium and calcium ions. Science 243, 1068–1071. Yang, X. C., and Sachs, F. (1993). Mechanically sensitive, nonselective cation channels. In ‘‘Nonselective Cation Channels’’ (D. Siemen and J. Hescheler, eds.), pp. 79–90. Birkhauser, Basel, Switzerland. Zou, H., Ugur, M., Drummond, R. M., and Singer, J. J. (2001). Coupling of a P2Z‐like purinoceptor to a fatty acid‐activated Kþ channel in toad gastric smooth muscle cells. J. Physiol. 534, 59–70. Zou, H., Lifshitz, L. M., Tuft, R. A., Fogarty, K. E., and Singer, J. J. (2002). Visualization of Ca2þ entry through single stretch‐activated cation channels. Proc. Natl. Acad. Sci. USA 99, 6404–6409.
CHAPTER 19 Mechanosensitive Ion Channels in Blood Pressure‐Sensing Baroreceptor Neurons Mark W. Chapleau,*,{,{,} Yongjun Lu,*,{ and Francois M. Abboud*,{,{ *The Cardiovascular Center, The University of Iowa Carver College of Medicine, Iowa City, Iowa 52242 { Department of Internal Medicine, The University of Iowa Carver College of Medicine, Iowa City, Iowa 52242 { Department of Molecular Physiology & Biophysics, The University of Iowa Carver College of Medicine, Iowa City, Iowa 52242 } Veterans AVairs Medical Center, Iowa City, Iowa 52246
I. Overview II. Introduction III. BR Sensory Transduction A. Vascular Compliance and Viscoelastic Coupling B. Mechanoelectrical Transduction C. Encoding of Depolarization into Frequency of Action Potential Discharge IV. Mechanosensitive Channels in BR Neurons A. Epithelial Naþ Channels B. Acid Sensing Ion Channels C. TRP Channels V. Methodological Limitations and Challenges A. Need for Selective Pharmacological Antagonists B. Complexity of Mechanosensitive Ion Channel Complex(es) C. Heterogeneity of Sensory Neurons VI. Summary and Future Directions References
I. OVERVIEW Baroreceptors (BRs) are mechanosensitive nerve endings in carotid sinuses and aortic arch that function as arterial blood pressure (BP) sensors. Changes in BR activity evoke reflex circulatory adjustments that reduce BP Current Topics in Membranes, Volume 59 Copyright 2007, Elsevier Inc. All right reserved.
1063-5823/07 $35.00 DOI: 10.1016/S1063-5823(06)59021-0
542
Chapleau et al.
variability and its adverse consequences. BR activation during increases in BP involves three processes: (1) vascular distension and deformation of BR nerve endings, (2) depolarization of the nerve terminals consequent to opening of mechanosensitive ion channels (mechanoelectrical transduction), and (3) translation of mechanically induced depolarization into action potential discharge mediated by voltage‐dependent Naþ and Kþ channels. The mechanism of mechanoelectrical transduction (process 2) has been elusive. Recent studies have applied a variety of physiological, pharmacological, and molecular approaches to this problem, including studies in animals and isolated BR neurons in culture. In this chapter, we provide an overview of the molecular basis of BR mechanoelectrical transduction. Emerging evidence points to members of three evolutionarily conserved ion channel families in mediating BR activation: epithelial Na channels (ENaCs), acid sensing ion channels (ASICs), and transient receptor potential (TRP) channels. The precise composition of the BR mechanosensitive ion channel complex and the mechanism of channel gating remain to be determined. Translation of discoveries in lower ‘‘model’’ organisms to studies of mammalian BR function and use of multidisciplinary approaches including state‐of‐the‐art spatial and temporal gene targeting are encouraged in order to move the field forward.
II. INTRODUCTION Arterial BP provides the driving force for delivery of blood flow to tissues and is therefore essential for organ system function and life. Maintenance of a relatively normal BP is particularly important for the brain and heart due to the high metabolic rate of these organs and the need for a continuous supply of blood flow and oxygen to maintain their vital functions. Even transient decreases in BP can compromise cerebral and coronary blood flow with risk of losing consciousness (syncope), stroke, and myocardial infarction. Abnormally high levels of BP or increased BP variability cause ‘‘target organ’’ damage, for example, impairment of vascular endothelial function, vascular and cardiac hypertrophy, kidney disease, and stroke (Mancia and Parati, 2003). Many factors can alter BP, including acute stressors (e.g., hemorrhage, assumption of the upright posture, emotional stress) and chronic disease (e.g., hypertension, autonomic failure). The arterial BR reflex is a key BP regulatory mechanism (Kirchheim, 1976; Chapleau et al., 2001; Chapleau, 2003; Chapleau and Abboud, 2004). BRs are mechanosensitive nerve endings located primarily in adventitia of carotid sinuses and aortic arch (Fig. 1). Changes in arterial BP alter the degree of vascular distension, which is sensed by BR nerve endings via mechanical
543
19. Baroreceptor Mechanoelectrical Transduction Afferents
CNS
Efferents
Carotid sinuses Arterial baroreceptors
Regulatory centers Para-SNA Heart SNA
Aortic arch
SNA SNA
Blood vessels Kidney
FIGURE 1 Baroreceptor reflex pathways. Shown are sites of aVerent BR innervation (filled circles), and eVerent parasympathetic (Para‐SNA) and sympathetic (SNA) projections from the central nervous system (CNS) to the heart, blood vessels, and kidneys. Adapted and reprinted from Chapleau (2003, Fig. 1, p. 104) and Chapleau and Abboud (2004, Fig. 1, p. 2) with permission.
deformation. AVerent BR activity is transmitted along the carotid sinus and aortic depressor nerves (ADNs) to the nucleus tractus solitarii in the medulla oblongata where the signals are integrated and relayed through a network of central neurons controlling eVerent parasympathetic and sympathetic nerve activity to the heart, vasculature, kidney, and other organs (Fig. 1). Changes in the frequency of BR aVerent discharge in response to changes in BP trigger reflex adjustments that buVer or oppose the change in BP (Kirchheim, 1976; Chapleau and Abboud, 2004). For example, a rise in BP increases BR activity leading to reflex inhibition of sympathetic activity, parasympathetic activation, and subsequent decreases in vascular resistance and heart rate. Conversely, a decrease in BP reduces BR activity thereby triggering a reflex increase in sympathetic activity, parasympathetic inhibition, and increases in vascular resistance and heart rate. Changes in BR activity also influence release of vasopressin and renin that contribute to the circulatory adjustment. Thus, the BR reflex provides a powerful moment‐to‐moment negative feedback regulation of BP. In addition to regulating BP, the BR reflex exerts a significant influence on the electrical properties of the heart through modulation of parasympathetic and sympathetic nerve activity (Podrid et al., 1990). Ventricular arrhythmias are a common cause of death during myocardial ischemia and after myocardial infarction. Animal and clinical studies have demonstrated that decreased BR
544
Chapleau et al.
reflex sensitivity predicts susceptibility to arrhythmias and sudden death after acute myocardial infarction and heart failure (Kaye and Esler, 1995; LaRovere et al., 1998). These findings suggest that the reflex may protect the heart from arrhythmias by providing appropriate and rapid modulation of cardiac autonomic tone. The BR reflex is composed of three general components: aVerent sensory transduction, central mediation of the reflex, and eVerent neurocardiac and neurovascular transmission (Fig. 1). The focus of this chapter is on mechanisms mediating mechanoelectrical transduction at the BR sensory nerve terminals.
III. BR SENSORY TRANSDUCTION BR activation during increase in BP involves three processes: (1) vascular distension and deformation of BR nerve endings, (2) subsequent depolarization of the nerve endings (mechanoelectrical transduction), and (3) translation of mechanically induced depolarization into action potential discharge mediated by voltage‐dependent Naþ and Kþ channels (Fig. 2).
Mechanoelectrical transduction
Spike initiating zone
Voltage-dependent channels Na+ Mechanosensitive Na+ channels Ca2+
Na+
K+
K+ Na+,K+ pump
Membrane potential FIGURE 2 Mechanisms of BR activation in response to increase in arterial BP. Opening of mechanosensitive channels depolarizes the sensory nerve terminals (mechanoelectrical transduction). Depolarization of the SIZ suYcient to open voltage‐dependent Naþ and Kþ channels triggers action potential discharge. Adapted from Chapleau et al. (2001, Fig. 1, p. 3) with permission.
19. Baroreceptor Mechanoelectrical Transduction
545
A. Vascular Compliance and Viscoelastic Coupling The compliance of large arteries defines the extent of blood vessel distension for a given increase in BP and therefore is a major determinant of the magnitude of deformation and BR activity (Kirchheim, 1976). Decreased arterial compliance contributes to decreased BR sensitivity in diseases, such as atherosclerosis and chronic hypertension, and with aging (Kirchheim, 1976; Andresen and Yang, 1989). Viscoelastic coupling between elements in the arterial wall and the nerve endings also importantly influences BR activity (Coleridge et al., 1984).
B. Mechanoelectrical Transduction Elucidation of the mechanism of transducing mechanical deformation into membrane depolarization is of great interest in many areas of biology. The depolarization of mechanoreceptors, referred to as the ‘‘receptor’’ or ‘‘generator’’ potential, is graded in relation to the magnitude of mechanical stimulation and decays with time and distance from the point of stimulation (Katz, 1950; Grigg, 1986). The mechanoelectrical transduction is generally considered to involve opening of mechanosensitive ion channels gated by changes in membrane tension (Hamill and Martinac, 2001; Fig. 2). 1. Assessment of BR Sensitivity The traditional approach to assess BR sensory function is to directly record the frequency or integrated voltage of action potential discharge from BR aVerent nerve fibers in vivo, in situ, or in vitro (Kirchheim, 1976; Andresen and Yang, 1989; Chapleau et al., 2001). The first approach used to test whether mechanosensitive channels mediate BR activation was to measure BR activity during ramp increases in pressure in the vascularly isolated carotid sinus of rabbits before and after intraluminal injection of gadolinium (Hajduczok et al., 1994), an established blocker of mechanosensitive channels (Yang and Sachs, 1989). Gadolinium markedly attenuated the pressure‐ induced increases in BR activity without changing the compliance of the carotid sinus (Hajduczok et al., 1994). Gadolinium did not block action potential discharge evoked by chemical stimulation of carotid sinus aVerents with the Naþ channel opener veratrine, indicating that the inhibition of pressure‐induced BR activity by gadolinium was not caused by nonspecific suppression of neuronal excitability. In a separate study, gadolinium did not decrease the activity of rat aortic arch BR fibers (Andresen and Yang, 1992). The reason for the discrepant results is not known.
546
Chapleau et al.
Action potential discharge in aVerent fibers is measured away from the site of mechanoelectrical transduction in the sensory nerve terminals. The small size and complex architecture of the BR terminals embedded in the vascular wall have generally prevented the direct measurement of membrane potential in the terminals and limited investigation into mechanisms of BR activation in vivo. To our knowledge, only one study has reported direct measurements of membrane potential in BRs near the site of mechanoelectrical transduction (Matsuura, 1973). Furthermore, the presence of endothelium and vascular muscle along with other cell types in the vascular wall make it diYcult to attribute changes in BR activity solely to direct actions on the nerve terminals. Factors released from nearby cells and changes in vascular smooth muscle tone can alter BR activity (Chapleau et al., 2001; Chapleau and Abboud, 2004). These limitations motivated us and others to develop an in vitro preparation of isolated BR neurons in culture (Section III.B.2). 2. Study of Isolated BR Neurons in Culture Aortic BR neurons can be labeled in vivo by application of a fluorescent dye (e.g., 1,10 ‐dioctadecyl‐3,3,30 30 ‐tetramethylindocarbocyanineperchlorate, diI) to the aortic arch adventitia or ADN of rats and mice (Mendelowitz and Kunze, 1992; Cunningham et al., 1995, 1997; Li et al., 1997; Sullivan et al., 1997;
Membrane potential
Picopump
Mechanical stimulation Nodose ganglion
1−3 wk dil
Dissociation and culture Isolated nodose neurons in culture
FIGURE 3 Study of isolated BR neurons in culture. BR neurons in nodose ganglia are labeled with a fluorescent dye (e.g., diI) applied to the aortic arch adventitia 1–3 weeks beforehand in vivo. Neurons are dissociated from nodose ganglia and fluorescently labeled BR neurons are studied in vitro. Adapted and reprinted from Chapleau et al. (2001, Fig.7, p.10) and Chapleau and Abboud (2004, Fig. 3, p. 8) with permission.
19. Baroreceptor Mechanoelectrical Transduction
547
Drummond et al., 1998; Kraske et al., 1998; Li et al., 1998; Fig. 3). One to 3 weeks later, nodose neurons are dissociated from nodose ganglia and maintained in culture. Functional studies are performed on individual diI‐labeled aortic BR neurons (Fig. 3). An important consideration is whether molecules expressed at the sensory terminals in vivo are also expressed in the cell membrane of isolated BR neurons in culture. Specific ligand receptors and ion channels present on the sensory nerve endings have been shown to be present on the soma of cultured nodose neurons (Fowler et al., 1985; Stansfeld et al., 1986; Christian et al., 1989; Drummond et al., 1998; Kraske et al., 1998). Furthermore, spike frequency adaptation of dorsal root ganglion (DRG) neurons during sustained mechanical stimulation of cutaneous mechanoreceptor endings correlates with the adaptation during sustained depolarization of the same neuron by current injection into the soma (Harper, 1991). Importantly, we have demonstrated that cultured BR neurons are mechanosensitive. Mechanical stimulation of isolated BR neurons evokes an inward cationic current (voltage clamp), depolarizes the membrane (current clamp), and increases cytosolic calcium concentration (Cunningham et al., 1995, 1997; Sullivan et al., 1997; Drummond et al., 1998; Snitsarev et al., 2002). Mechanosensitive channels have been identified in isolated BR neurons at the single‐ channel level (Kraske et al., 1998). Similar to our findings in vivo, gadolinium blocks responses to mechanical stimulation of BR neurons in vitro (Cunningham et al., 1995, 1997; Sullivan et al., 1997). Gadolinium also inhibits mechanically induced responses in diI‐labeled cardiac sensory neurons isolated from nodose ganglia (Linz and Veelken, 2002; Ditting et al., 2003), and a subpopulation of neurons isolated from DRG (Gotoh and Takahashi, 1999; McCarter et al., 1999; Raybould et al., 1999; Cho et al., 2002; Drew et al., 2002). DiVerences in mechanosensitivity among diVerent types of neurons isolated from nodose and DRG correspond to diVerences in mechanosensitivity of their respective nerve terminals (Cunningham et al., 1995, 1997; Sharma et al., 1995; Sullivan et al., 1997; Drew et al., 2002, 2004). EVerent autonomic neurons isolated from sympathetic ganglia do not generate inward currents or depolarize in response to mechanical stimulation (McCarter et al., 1999; our unpublished observations). Therefore, despite probable diVerences in expression and regulation of sensory molecules in sensory terminals vs isolated neuron somata, the isolated BR neuron appears to be a valid model for investigation of mechanisms of sensory transduction. The in vivo and in vitro findings reviewed above support the hypothesis that mechanosensitive channels mediate BR mechanoelectrical transduction (Fig. 2).
548
Chapleau et al.
C. Encoding of Depolarization into Frequency of Action Potential Discharge Depolarization of mechanosensitive nerve endings is localized to the sensory terminals and rapidly decays with distance from the site of stimulation (Katz, 1950; Grigg, 1986). The mechanically induced depolarization is translated into action potential discharge at the ‘‘spike initiating zone’’ (SIZ) near the nerve terminals (Katz, 1950; Grigg, 1986; Fig. 2). Action potentials are generated when the depolarization reaches a specific threshold for opening of voltage‐ dependent Naþ and Kþ channels. The frequency of action potential discharge increases with further depolarization and is critically dependent on the expression and properties of voltage‐dependent channels and membrane pumps near the SIZ. The role of voltage‐dependent channels in modulating BR activity and their modulation by autocrine/paracrine factors have been reviewed elsewhere (Schild and Kunze, 1997; Chapleau et al., 2001; Chapleau and Abboud, 2004; Schild et al., 2005).
IV. MECHANOSENSITIVE CHANNELS IN BR NEURONS Recent advances in gene discovery in diverse organisms and the development of methods to measure and selectively manipulate gene expression have enabled identification of several candidate genes that may encode BR mechanosensitive channels. Emerging evidence points to members of three evolutionarily conserved ion channel families.
A. Epithelial Naþ Channels The degenerin (DEG) genes MEC‐4 and MEC‐10 strongly influence touch sensitivity in the roundworm Caenorhabditis elegans (C. elegans) and null mutations in MEC‐4 or accessory ion channel subunit genes (MEC‐2 and MEC‐6) eliminate mechanoreceptor currents in C. elegans sensory neurons (Tavernarakis and Driscoll, 2001; Goodman and Schwarz, 2003; O’Hagan et al., 2005). Homology between C. elegans DEG genes and mammalian ENaC suggested that proteins of the DEG/ENaC superfamily may function as mechanosensors in mammals (Canessa et al., 1993). Mammalian ENaCs are heteromultimers composed of a, b, and g subunits that play an essential role in epithelial Naþ transport (Kellenberger and Schild, 2002). Evidence that ENaCs may function as mechanosensors in a variety of cell types is accumulating (Awayda and Subramanyam, 1998; Kellenberger and Schild, 2002; Carattino et al., 2004; Drummond et al., 2004).
549
19. Baroreceptor Mechanoelectrical Transduction A
Nodose ganglion
B
Dil (red) Remove aortic arch, and stain nerve terminals
Aortic arch Baroreceptor nerve terminals
50 mm
C
NF160
50 mm
g ENaC
FIGURE 4 Gamma (g) ENaC expression in aortic BR nerve terminals. (A) DiI was injected into rat nodose ganglion to label aortic BR fibers innervating aortic arch. (B) Two weeks later, colocalization of diI and gENaC is evident in BR nerve endings in aortic arch adventitia (150‐mm optical section; diI—red, gENaC—green, diI þ gENaC—yellow). (C) BR terminal in aortic arch immunolabeled with anti‐NF160, a neuronal marker, and anti‐gENaC. Reprinted from Neuron, Vol. 21, Drummond et al., A molecular component of the arterial baroreceptor mechanotransducer (1998, Fig. 3A–C, p. 1438) with permission from Elsevier.
1. ENaCs Are Expressed in Nodose Neurons and BR Sensory Terminals We hypothesized that the composition of the mechanosensitive ion channel complex in BR neurons may include ENaC subunits and that these subunits may therefore contribute to BR mechanoelectrical transduction. The first step was to determine if ENaC subunits are expressed in BR neurons. We demonstrated by RT‐PCR the presence of mRNA for b and g subunits of ENaC in rat nodose ganglia (Drummond et al., 1998, 2001). Interestingly, the a subunit was not detected. Expression of gENaC protein was evident in diI‐labeled BR neurons in nodose ganglia and, more importantly, was localized in BR sensory terminals in the adventitia of the aortic arch (Drummond et al., 1998, 2001; Fig. 4). The apparent absence of the a subunit may reflect the expectation that mechanosensitive channels would be closed at rest and open only during mechanical stimulation; ENaCs containing a subunits are constitutively open (Kellenberger and Schild, 2002).
550
Chapleau et al.
2. ENaC Blockers Attenuate BR Responses In Vivo and In Vitro A common approach to determine whether ENaC or related channels contribute to functional responses is to obtain measurements before and after application of the DEG/ENaC blocker amiloride (Hamill et al., 1992). Exposure of the isolated carotid sinus of rabbits to the amiloride analogue benzamil decreases BR activity in a dose‐related manner and significantly attenuates the baroreflex‐mediated fall in systemic BP evoked by increases in carotid sinus pressure (Drummond et al., 1998, 2001; Fig. 5).
Systemic BP (mmHg)
A 120 100 mM 100 50 mM 10 mM Control
80 0
50 100 150 200 Carotid sinus pressure (mmHg)
Carotid sinus nerve activity (AU)
B 40 Wash control 50 mM
30
100 mM
20 10
1 mM 0 0
25
50
75
100
Carotid sinus pressure (mmHg) FIGURE 5 ENaC blocker benzamil attenuates BR activity and baroreflex control of BP. Exposure of isolated carotid sinus of rabbits to benzamil causes concentration‐dependent inhibition of baroreflex‐mediated decreases in systemic BP (A), and attenuation of pressure‐ induced increases in aVerent carotid sinus BR activity (B). The BP curves represent average data from five experiments. Carotid sinus nerve activity curves are from one representative experiment. Reprinted from Neuron, Vol. 21, Drummond et al., A molecular component of the arterial baroreceptor mechanotransducer (1998, Fig. 5, p. 1439) with permission from Elsevier.
551
19. Baroreceptor Mechanoelectrical Transduction
The reflex inhibition of sympathetic nerve activity during blood volume expansion achieved by intravenous injection of saline in rats is also inhibited by benzamil injected into the pericardial sac surrounding the heart (Ditting et al., 2003). The reflex response to blood volume expansion is mediated by activation of cardiac mechanoreceptors similar in structure and function to arterial BRs (Bishop et al., 1983). Thus, amiloride‐sensitive channels, presumably containing ENaC subunits, may contribute to activation of both arterial BRs and mechanosensitive aVerents innervating the heart. Consistent with our in vivo findings, amiloride essentially abolishes membrane depolarization and increases in cytosolic calcium evoked by puYng fluid onto isolated BR neurons (Drummond et al., 1998; Snitsarev et al., 2002; Fig. 6). Importantly, the concentration of amiloride used to inhibit mechanically induced responses (0.1–1.0 mM) does not attenuate action potential discharge evoked by depolarizing current injection (Snitsarev et al., 2002; Fig. 6).
Current injection (0.1 nA for 800 ms)
50 mV
Mechanical stimulation (5 psi for 60 s)
Control
Amiloride
Recovery
FIGURE 6 Amiloride (1 mM) selectively blocks mechanically induced depolarization of isolated rat nodose sensory neuron (left) without suppressing neuronal excitability (right). The neuron was mechanically stimulated by a stream of buVer ejected under pressure (5 psi) from a micropipette placed 50 mm from the cell surface. Membrane potential was recorded from a sharp microelectrode. Neuronal excitability was assessed by measuring the action potential response to depolarizing current injections (0.1 nA). Reprinted from Snitsarev et al. (2002, Fig. 1, p. 60) with permission.
552
Chapleau et al.
Although amiloride can potentially inhibit neuronal excitability through eVects on voltage‐gated channels (Carr et al., 2001; Ditting et al., 2003), our results suggest that at lower concentrations (0.1–1.0 mM) amiloride is selective for mechanosensitive channels. Inhibitors of voltage‐dependent channels do not attenuate inward ionic currents evoked by mechanical stimulation of isolated nodose neurons (Cunningham et al., 1997). Interestingly, amiloride and benzamil fail to block inward currents generated by exposure of isolated diI‐labeled cardiac nodose sensory neurons to hypoosmotic stress (Ditting et al., 2003). DiVerent types of mechanosensitive channels may mediate responses to rapid mechanical stimulation with puVs of fluid vs slower cell swelling under hypoosmotic stress.
B. Acid Sensing Ion Channels ASICs represent an additional mammalian subfamily of the DEG/ENaC superfamily (Price et al., 1996; Waldmann et al., 1997; Krishtal, 2003). As their name indicates, ASICs are Hþ‐gated ion channels. Several ASIC subunits have been identified including ASIC1 (1a and 1b), ASIC2 (2a and 2b), ASIC3, and ASIC4. ASIC1a and ASIC1b are splice variants of the same gene, as are ASIC2a and ASIC2b. Unlike ENaCs that are expressed in epithelium and a variety of other cell types, ASICs are expressed primarily in neurons. The ASIC subunits show varying degrees of acid sensitivity with ASIC3 and ASIC1a being the most sensitive and ASIC2 the least sensitive (Lingueglia et al., 1997; Waldmann et al., 1997; Hesselager et al., 2004). The subunits can form homomultimers or heteromultimers; the specific combination of which influences the kinetics of channel inactivation and desensitization, Hþ sensitivity, cation selectivity, and susceptibility to amiloride and gadolinium blockade (Waldmann et al., 1997; Babinski et al., 2000; Alvarez de la Rosa et al., 2002; Benson et al., 2002; Hesselager et al., 2004). ASICs have been implicated in mechanotransduction in sensory nerves innervating a variety of tissues. ASIC2 (also named BNC1 and BNaC1) is expressed in DRG neurons innervating skin and has been shown to be transported from soma to mechanosensory nerve terminals (Price et al., 2000; Garcia‐Anoveros et al., 2001). ASIC2 / mice exhibit a selective decrease in sensitivity of low‐threshold, rapidly adapting cutaneous mechanoreceptors (Price et al., 2000; Welsh et al., 2002), although this finding was not confirmed in a subsequent study (Roza et al., 2004). In contrast, in ASIC3 / mice, high‐threshold mechanonociceptor sensitivity is impaired accompanied by increased sensitivity of the rapidly adapting cutaneous aVerents (Price et al., 2001). ASIC1, ASIC2, and ASIC3 have been shown to contribute diVerentially to mechanotransduction in diVerent types of
19. Baroreceptor Mechanoelectrical Transduction
553
visceral aVerents innervating the gastrointestinal (GI) system (Jones et al., 2005; Page et al., 2005). Interestingly, ASIC3 contributes to mechanosensitivity in most types of GI aVerents, ASIC1a exerts an inhibitory influence on mechanosensitivity in all types of GI aVerents, and ASIC2 may either contribute to or restrain mechanosensitivity depending on the type of aVerent (Jones et al., 2005; Page et al., 2005). Thus, ASIC subunits may contribute diVerentially to mechanotransduction and their contributions vary depending on the type of neuron and nature of the mechanical stimulus. 1. ASICs Are Expressed in Nodose Neurons and BR Terminals ASIC subunits are expressed in nodose ganglia of mice at both mRNA and protein levels (Page et al., 2005; Lu et al., 2006). Preliminary results show similar relative mRNA levels of ASIC1a, ASIC1b, ASIC2a, and ASIC3, with higher expression of the ASIC2b subunit measured by real‐time PCR (Lu et al., 2006). ASIC2b alone does not form a pH‐gated channel, suggesting that its functional significance may relate to modulation of other ASIC subunits (Lingueglia et al., 1997; Hesselager et al., 2004). ASIC1, ASIC2, and ASIC3 proteins are colocalized (immunohistochemistry) in some nodose neurons consistent with the presence of heteromultimeric channels (Lu et al., 2006). ASIC2 staining is also evident in the ADN and nerve terminals in aortic arch. The presence of ASIC2 in BR and somatic aVerent sensory terminals not sensitive to acid are consistent with it serving a mechanoreceptor function in these nerve endings (Price et al., 2000; Garcia‐Anoveros et al., 2001; Welsh et al., 2002; Lu et al., 2006). 2. BR Function Is Impaired in ASIC2 / Mice Both ENaCs and ASICs are inhibited by amiloride, although ASICs are less sensitive with ASIC3 showing incomplete inhibition even at high concentrations of amiloride (Waldmann et al., 1997; Kellenberger and Schild, 2002). Thus, inhibition of mechanically induced responses by amiloride cannot distinguish a specific contribution of ASICs. Therefore, we have chosen to investigate the role of ASICs in BR activation through study of ASIC‐deficient mice. This approach necessitated development of methods to assess BR function in mice. The sensitivity of BR aVerents to changes in BP can be examined by directly recording aVerent BR activity from the ADN and measuring changes in activity during pharmacologically induced changes in BP (Ma et al., 2002; Fig. 7). Preliminary results obtained from ASIC2 / mice support the hypothesis that ASIC2 contributes to BR activation (Ma et al., 2001). The vasoconstrictor phenylephrine was injected intravenously in ASIC2 / and wild‐type mice in order to cause a sustained increase in BP. While the immediate increase in BR activity accompanying the rise in BP appears relatively
554 Arterial pressure (mmHg)
Chapleau et al. 220
SNP
PE
160 100 40
ADNA (mV)
30 15 0 −15 10 s
Arterial pressure (mmHg)
−30
Baseline
SNP
PE
160 120 80 40
ADNA (mV)
20 10 0 −10 −20
0.2 s
FIGURE 7 Assessment of BR aVerent sensitivity to changes in BP in mice in vivo. Shown are recordings of BP and BR activity in ADN (ADNA) in an anesthetized mouse under baseline conditions, during sodium nitroprusside (SNP)‐induced decreases in BP, and during phenylephrine (PE)‐induced increases in BP. Reprinted from Am. J. Physiol. Reg. Integr. Comp. Physiol., Vol. 283, Ma et al., Analysis of aVerent, central, and eVerent components of the baroreceptor reflex in mice (2002, Fig. 4, p. R1037) with permission.
normal, the ability to sustain the increase in BR activity as BP is maintained at a high level is impaired in ASIC2 / mice (Ma et al., 2001). The results suggest that ASIC2 is essential for normal BR sensing of sustained increases in BP. Impaired BR aVerent sensitivity should translate to a defect in BR reflex control of heart rate and BP. Preliminary data suggest that baroreflex sensitivity for control of heart rate is decreased in conscious ASIC2 / mice
19. Baroreceptor Mechanoelectrical Transduction
555
(Sabharwal et al., 2006). To evaluate reflex control of BP, we have measured the BP response to bilateral carotid artery occlusion (BCO) in anesthetized mice. BCO reduces carotid sinus pressure and BR activity thereby triggering a baroreflex‐mediated increase in systemic BP. Activation of carotid body chemoreceptors consequent to ischemia during BCO can contribute to the reflex rise in BP (Alcayaga et al., 1986). Therefore, measurements of the BP response to BCO were repeated while ventilating the mice with 100% oxygen to suppress chemoreceptor activity. This approach enables analysis of the relative contributions of the baroreflex and chemoreflex to the BCO‐induced increase in BP (Sun et al., 2000). Our preliminary results indicate that the baroreflex component of the BCO reflex is significantly impaired in ASIC2 / mice, while the chemoreflex component is enhanced (Sabharwal et al., 2005a). The enhanced chemoreflex component of the BP rise suggests that eVerent sympathetic‐mediated vasoconstriction is preserved in ASIC2 / mice. Thus, the defect in baroreflex control of BP likely resides in BR aVerent nerves, or possibly within the central nervous system. The BR component of the BCO reflex was not altered in ASIC3 / mice (Sabharwal et al., 2005b). The reciprocal relationship between baro‐ and chemoreflex sensitivity in ASIC2 / mice is reminiscent of what has been observed in hypertension, heart failure, hypercholesterolemia, and aging (Trzebski et al., 1982; Somers et al., 1988; Franchini et al., 1996; Ponikowski et al., 1997; Sun et al., 1999; Sun et al., 2001a,b, 2002). We speculate that the functional reciprocity may reflect, in part, compensatory upregulation of expression or function of other ASIC subunits in ASIC2 / mice and that dysregulation of ASICs may contribute to decreased baroreflex sensitivity and increased chemoreflex sensitivity in pathological states.
C. TRP Channels TRP channels represent a superfamily of cation‐selective channels (Clapham et al., 2003; Desai and Clapham, 2005). These evolutionarily conserved channels are very weakly voltage dependent, are expressed in many types of cells, and have been implicated in sensing a variety of stimuli including light, temperature, pheromones, acidity, and osmolarity. TRP subfamilies include TRPC(1–7), TRPV(1–6), TRPM(1–8), TRPA, TRPN, TRPP, and TRPML. The subunits within subfamilies can form heteromultimers that influence the electrophysiological characteristics of the channels. TRPV1 is the well‐known vanilloid receptor sensitive to capsaicin, heat, Hþ, and endogenous cannabinoids. Several TRP channels have been implicated in mechanotransduction (Lin and Corey, 2005; O’Neil and Heller, 2005). TRPV homologues mediate
556
Chapleau et al.
osmosensory and mechanosensory responses in C. elegans, Drosophila, and mammals. Mammalian TRPV1 has been implicated in mechanosensory transduction in specific types of sensory aVerents innervating the intestine and bladder (Birder et al., 2002; Rong et al., 2004; Jones et al., 2005). TRPV4, cloned based on homology with C. elegans OSM‐9, is activated by hypotonic cell swelling, shear stress, acidic pH, warm temperature, arachidonic acid, and 50 ,60 ‐EET (50 ,60 ‐epoxyeicosatrienoic acid), and is expressed in cutaneous mechanosensitive aVerents (Liedtke et al., 2003; Suzuki et al., 2003a,b; Lin and Corey, 2005; O’Neil and Heller, 2005). TRPV2 is also activated by cell swelling and membrane stretch (Muraki et al., 2003). The responsiveness of TRPV channels to both mechanical stimuli and chemical second messengers (e.g., 50 ,60 ‐EET) suggests that their ‘‘mechanosensitivity’’ may involve indirect chemical activation of the channel. TRPA1 and TRPC1 are components of mechanotransducer channels in vertebrate hair cells and Xenopus oocytes, respectively (Corey et al., 2004; Maroto et al., 2005). Activation of TRPA1 channels may involve transmission of force to the channel through accessory proteins (Corey et al., 2004; Lin and Corey, 2005). In contrast, TRPC1 may be directly activated by lipid tension (Maroto et al., 2005). Thus, several TRP proteins should be considered as prime candidates for involvement in BR mechanoelectrical transduction. 1. TRP Proteins Are Expressed in Nodose Neurons and BR Terminals TRPV1 and several TRPC proteins (TRPC1, TRPC3–7) are expressed in sensory neurons in rat nodose ganglia (Helliwell et al., 1998; Glazebrook et al., 2005). TRPV1 appears to be selectively expressed in C‐fiber vagal aVerents (Jin et al., 2004). In contrast, both myelinated and unmyelinated fibers in the ADN contain TRPC1, TRPC3, TRPC4, and TRPC5 proteins with TRPC1, TRPC4, and TRPC5 being distinctly localized in sensory terminals in aortic arch adventitia, although the distribution of specific TRPC subunits may diVer in sensory terminals of myelinated and unmyelinated BR aVerents (Glazebrook et al., 2005; Fig. 8). In preliminary experiments, we have confirmed expression of TRPC1 in nodose ganglia of adult rats and mice (our unpublished observation). Although these results are consistent with TRP channels contributing to BR mechanoelectrical transduction, further studies are warranted. In addition to possibly functioning as a mechanosensor, TRP channels can be activated by ligand binding to G‐protein–coupled receptors, tyrosine kinase activation, and second messengers (Clapham et al., 2003; Desai and Clapham, 2005; Lin and Corey, 2005). Therefore, TRP channels may modulate BR sensitivity indirectly through their sensitivity to chemical factors and second messengers. Future studies of BR sensitivity in TRPV1‐ and TRPC‐deficient mice and
19. Baroreceptor Mechanoelectrical Transduction
557
FIGURE 8 Expression of TRPC subunits in aortic BR nerve terminals. Shown are Z‐series confocal images of BR terminals in aortic arch immunolabeled with antibodies to TRPC1, TRPC3, TRPC4, and TRPC5 (left panels). The same sections were stained with either the neuronal marker PGP9.5 to delineate the terminal or myelin basic protein (MBP) to show the edge of the myelin sheath (corresponding right panels). Scale bar ¼ 20 mm. Reprinted from Glazebrook et al. (2005, Fig. 3, p. 128) with permission.
558
Chapleau et al.
mechanoelectrical transduction in BR neurons lacking TRP subunits are needed to address the functional role of these proteins in BR aVerent nerves.
V. METHODOLOGICAL LIMITATIONS AND CHALLENGES The explosion of knowledge of genome sequences, technological advances, use of model systems in lower organisms amenable to genetic analysis (e.g., Drosophila, C. elegans), and identification of mammalian homologues as candidate mechanosensors have provided new and vast opportunities for discovery. Studies exploring the molecular basis of BR activation have begun to utilize these approaches. For example, ENaC, ASIC, and TRP subunits have been localized in BR sensory nerve terminals by immunohistochemistry. mRNA expression of these molecules has been confirmed in nodose and petrosal ganglia where BR somata reside by RT‐PCR and quantitative real‐ time PCR. The impact of gene deletion on BR function is beginning to be explored using mutant mice. Recent and ongoing studies continue to rely heavily on the use of pharmacological blockers of mechanosensitive channels (e.g., gadolinium and amiloride). While these approaches have been productive, significant limitations of the methods are apparent and new challenges have arisen.
A. Need for Selective Pharmacological Antagonists The absence or limited availability of selective pharmacological antagonists of ENaC, ASIC, and TRP subunits is a major limitation. The selectivity of drugs like gadolinium and amiloride within an experimental paradigm vary depending on a variety of factors, including the array of ion channels expressed in the neuron under study, the endpoint being measured (e.g., receptor potential vs cytosolic Ca2þ vs action potential firing), and the concentration of antagonist used. The utility of amiloride and its analogues to distinguish between ENaC and ASIC subunits is limited. ASIC3 is relatively resistant to blockade by amiloride (Waldmann et al., 1997). Compound A‐317567 has been reported to block ASICs with greater potency than amiloride (Dube et al., 2005). Moreover, indirect evidence suggests that A‐317567 does not block ENaCs (Dube et al., 2005). Psalmotoxin 1 isolated from the tarantula spider and APETx2 from sea anemone selectively block ASIC1a and ASIC3 channels, respectively (Escoubas et al., 2000; Diochot et al., 2004). Future development of selective channel blockers should facilitate studies of BR mechanoelectrical transduction.
19. Baroreceptor Mechanoelectrical Transduction
559
B. Complexity of Mechanosensitive Ion Channel Complex(es) The complexity of the mechanosensitive ion channel complex and the potential involvement of multiple gene families create obstacles to both pharmacological and genetic approaches. ENaC, ASIC, and TRP subunits can form heteromultimeric channels of varying subunit composition. The diverse channels diVer in many respects, including their sensitivity to diVerent types of sensory input (e.g., mechanical, chemical, acidity, temperature) and their susceptibility to blockade by channel blockers (e.g., gadolinium and amiloride). The chemosensitivity of some of these proteins, particularly the TRP channels, creates the possibility that their activation by mechanical stimulation may be indirect, that is, via mechanically induced production of metabolites that subsequently open the channels (Lin and Corey, 2005; O’Neil and Heller, 2005). Heteromultimerization may occur not only between subunits within subfamilies with relatively homologous structures (Schaefer, 2005) but also between subunits from diVerent families. For example, a preliminary report suggests heteromultimerization of ENaCs and ASICs (Meltzer et al., 2006). In addition to the mechanosensitive channel, mechanoelectrical transduction may depend on accessory tethering molecules that link the channel to the cytoskeleton and extracellular matrix (Tavernarakis and Driscoll, 2001; Chelur et al., 2002; Goodman et al., 2002; Goodman and Schwarz, 2003; Lin and Corey, 2005). For example, MEC‐2 and MEC‐6 are required for generation of mechanoreceptor currents in C. elegans sensory neurons (O’Hagan et al., 2005). Homologous proteins such as stomatin and PICK1 may interact with and modulate mammalian ENaCs and ASICs (Mannsfeldt et al., 1999; Fricke et al., 2000; Duggan et al., 2002; Deval et al., 2004; Price et al., 2004). The complexity of the ion channel complexes poses significant challenges to experimental design and interpretation of data. The absence of obligatory accessory molecules in expression systems may prevent reconstitution of functional mechanosensitive channels. Tethering proteins or specific subunits may be selectively targeted to BR sensory terminals, thereby limiting the utility of studying channels in the soma. In fact, assessment of mechanosensitivity in neurons isolated from DRG of ASIC2‐ and ASIC3‐deficient mice failed to detect an impaired response (Drew et al., 2004). In knockout mice, other subunits may replace a deleted subunit to preserve channel function. Alternatively, increased expression of other heteromeric channels may compensate for loss of channel function. The consequences of manipulating one subunit of a heteromultimeric complex can be diYcult to predict. The above factors may be particularly problematic in mice with life‐long gene deletions.
560
Chapleau et al.
C. Heterogeneity of Sensory Neurons An additional factor that must be considered is the tremendous heterogeneity of mechanosensitive nerves. Mechanosensitive aVerents innervating diVerent organs (e.g., blood vessels, GI system, heart, bladder, skin, and muscle) diVer to varying extent in their electrophysiological properties and ability to detect diVerent modes of sensory stimuli. DiVerences in the molecular composition of the mechanosensitive channels likely contribute to the heterogeneity. For example, ASIC1a, ASIC2, and ASIC3 may each promote, oppose, or not influence mechanosensitivity depending on the tissue innervated (e.g., skin vs intestine) and functional type of sensory nerve within a given tissue (Price et al., 2000, 2001; Jones et al., 2005; Page et al., 2005). Subtypes of BR aVerents can be distinguished based on conduction velocity, neuropeptide content, adaptation properties, and action potential discharge characteristics (Kirchheim, 1976; Chapleau et al., 2001). The molecular composition of mechanosensitive channels may diVer in subtypes of BR neurons.
VI. SUMMARY AND FUTURE DIRECTIONS The process of BR activation by increases in BP involves vascular distension and BR deformation, depolarization of the nerve terminals, and encoding of the mechanically induced depolarization into action potential discharge. Emerging evidence suggests that members of the ENaC, ASIC, and TRP ion channel families mediate the BR depolarization (mechanoelectrical transduction), but the precise composition of the mechanosensitive ion channel complex (or complexes) in BR sensory terminals and the mechanism of channel gating remain to be determined. Future studies should move toward more rapid translation of discoveries in lower ‘‘model’’ organisms to investigation of candidate genes mediating BR activation in mammals. Multidisciplinary state‐of‐the‐art approaches including site‐specific and temporal gene targeting are encouraged (Bockamp et al., 2002). For example, use of inducible knockout technology can demonstrate reversibility of functional changes and avoid lethal phenotypes and long‐ term compensatory adaptations that may occur with life‐long gene deletion. Generation of double and triple knockout mice may be necessary to entirely disrupt BR mechanosensitive channel function. Investigators are encouraged to make use of more selective ion channel blockers (e.g., for ENaC, ASIC, and TRP channel subunits) as they become available. The presence of numerous types of neurons and other cells in nodose and petrosal ganglia and the recognition of the existence of BR subtypes underscore
19. Baroreceptor Mechanoelectrical Transduction
561
the importance of studying individual identified BR neurons, preferably combining measurements of function (e.g., ionic currents, membrane depolarization, single‐fiber discharge) with measurements of gene expression (e.g., single‐ cell PCR, immunocytochemistry). In addition to identifying the key molecules involved in BR mechanoelectrical transduction, it will be necessary to distinguish mechanisms of channel activation, albeit direct or indirect via chemical second messengers. The recent discovery of genes regulating mechanosensitivity in lower organisms and the availability of powerful new technologies and experimental approaches make this a truly exciting time to pursue the goal of understanding the molecular basis of BR activation. The challenges are great, but so are the opportunities. Acknowledgments The authors would like to acknowledge their colleagues who over the years have made important contributions to portions of the work summarized in this chapter including Drs. Xiuying Ma, Vladislav Snitsarev, Rasna Sabharwal, Heather Drummond, J. Thomas Cunningham, Margaret Sullivan, George Hajduczok, Ram V. Sharma, Margaret Price, Christopher Benson, and Michael J. Welsh. This publication was made possible by grant HL14388 from the National Institutes of Health, a VA Merit Award to MWC from the Department of Veterans AVairs, and funds from the Heartland AYliate of the American Heart Association. Its contents are solely the responsibility of the authors and do not necessarily represent the oYcial views of the NIH, VA, or AHA.
References Alcayaga, J., Iturriaga, R., and Zapata, P. (1986). Carotid body chemoreceptor excitation produced by carotid occlusion. Acta Physiol. Pharmacol. Latinoam. 36, 199–215. Alvarez de la Rosa, D., Zhang, P., Shao, D., White, F., and Canessa, C. M. (2002). Functional implications of the localization and activity of acid‐sensitive channels in rat peripheral nervous system. Proc. Natl. Acad. Sci. USA 99, 2326–2331. Andresen, M. C., and Yang, M. (1989). Arterial baroreceptor resetting: Contributions of chronic and acute processes. Clin. Exp. Pharmacol. Physiol. Suppl. 15, 19–30. Andresen, M. C., and Yang, M. (1992). Gadolinium and mechanotransduction of rat aortic baroreceptors. Am. J. Physiol. 262(31), H1415–H1421. Awayda, M. S., and Subramanyam, M. (1998). Regulation of the epithelial Naþ channel by membrane tension. J. Gen. Physiol. 112, 97–111. Babinski, K., Catarsi, S., Biagini, G., and Seguela, P. (2000). Mammalian ASIC2a and ASIC3 subunits co‐assemble into heteromeric proton‐gated channels sensitive to Gd3þ. J. Biol. Chem. 275, 28519–28525. Benson, C. J., Xie, J., Wemmie, J. A., Price, M. P., Henss, J. M., Welsh, M. J., and Snyder, P. M. (2002). Heteromultimers of DEG/ENaC subunits form Hþ‐gated channels in mouse sensory neurons. Proc. Natl. Acad. Sci. USA 99, 2338–2343. Birder, L. A., Nakamura, Y., Kiss, S., Nealen, M. L., Barrick, S., Kanai, A. J., Wang, E., Ruiz, G., de Groat, W. C., Apodaca, G., Watkins, S., and Caterina, M. J. (2002). Altered urinary bladder function in mice lacking the vanilloid receptor TRPV1. Nat. Neurosci. 5, 856–860.
562
Chapleau et al.
Bishop, V. S., Malliani, A., and Thoren, P. (1983). Cardiac mechanoreceptors. In ‘‘Handbook of Physiology: The Cardiovascular System’’ (J. T. Shepherd and F. M. Abboud, eds.), Vol. III, pp. 497–555. American Physiological Society, Bethesda, MD. Bockamp, E., Maringer, M., Spangenberg, C., Fees, S., Fraser, S., Eshkind, L., Oesch, F., and Zabel, B. (2002). Of mice and models: Improved animal models for biomedical research. Physiol. Genomics 11, 115–132. Canessa, C. M., Horisberger, J.‐D., and Rossier, B. C. (1993). Epithelial sodium channel related to proteins involved in neurodegeneration. Nature 361, 467–470. Carattino, M. D., Sheng, S., and Kleyman, T. R. (2004). Epithelial Naþ channels are activated by laminar shear stress. J. Biol. Chem. 279, 4120–4126. Carr, M. J., Gover, T. D., Weinreich, D., and Undem, B. J. (2001). Inhibition of mechanical activation of guinea‐pig airway aVerent neurons by amiloride analogues. Br. J. Pharmacol. 133, 1255–1262. Chapleau, M. W. (2003). Arterial baroreflexes. In ‘‘Hypertension Primer,’’ (J. L. Izzo and H. R. Black, eds.), 3rd Ed., pp. 103–106. American Heart Association, Dallas, TX. Chapleau, M. W., and Abboud, F. M. (2004). The baroreceptor reflex: Novel methods and mechanisms. In ‘‘Neural Mechanisms of Cardiovascular Regulation’’ (N. J. Dun, B. H. Machado, and P. M. Pilowsky, eds.), pp. 1–29. Kluwer Academic Publishers, Boston, MA. Chapleau, M. W., Li, Z., Meyrelles, S. S., Ma, X., and Abboud, F. M. (2001). Mechanisms determining sensitivity of baroreceptor aVerents in health and disease. Ann. NY Acad. Sci. 940, 1–19. Chelur, D. S., Ernstrom, G. G., Goodman, M. B., Yao, C. A., Chen, L., O’Hagan, R., and Chalfie, M. (2002). The mechanosensory protein MEC6 is a subunit of the C. elegans touch‐cell degenerin channel. Nature 420, 669–673. Cho, H., Shin, J., Shin, C. Y., Lee, S.‐Y., and Oh, U. (2002). Mechanosensitive ion channels in cultured sensory neurons of neonatal rats. J. Neurosci. 22, 1238–1247. Christian, E. P., Taylor, G. E., and Weinreich, D. (1989). Serotonin increases excitability of rabbit C‐fiber neurons by two distinct mechanisms. J. Appl. Physiol. 67, 584–591. Clapham, D. E., Montell, C., Schultz, G., and Julius, D. (2003). International Union of Pharmacology. XLIII. Compendium of voltage‐gated ion channels: Transient receptor potential channels. Pharmacol. Rev. 55, 591–596. Coleridge, H. M., Coleridge, J. C. G., Poore, E. R., Roberts, A. M., and Schultz, H. D. (1984). Aortic wall properties and baroreceptor behavior at normal arterial pressure and in acute hypertensive resetting in dogs. J. Physiol.(Lond.) 350, 309–326. Corey, D. P., Garcia‐Anoveros, J., Holt, J. R., Kwan, K. Y., Lin, S.‐Y., Vollrath, M. A., Amalfitano, A., Cheung, E. L.‐M., Derfler, B. H., Duggan, A., Geleoc, G. S. G., Gray, P. A., et al. (2004). TRPA1 is a candidate for the mechanosensitive transduction channel of vertebrate hair cells. Nature 432, 723–730. Cunningham, J. T., Wachtel, R. E., and Abboud, F. M. (1995). Mechanosensitive currents in putative aortic baroreceptor neurons in vitro. J. Neurophysiol. 73, 2094–2098. Cunningham, J. T., Wachtel, R. E., and Abboud, F. M. (1997). Mechanical stimulation of neurites generates an inward current in putative aortic baroreceptor neurons in vitro. Brain Res. 757, 149–154. Desai, B. N., and Clapham, D. E. (2005). TRP channels and mice deficient in TRP channels. Pflugers Arch. 451, 11–18. Deval, E., Salinas, M., Baron, A., Lingueglia, E., and Lazdunski, M. (2004). ASIC2b‐ dependent regulation of ASIC3, an essential acid‐sensing ion channel subunit in sensory neurons via the partner protein PICK‐1. J. Biol. Chem. 279, 19531–19539.
19. Baroreceptor Mechanoelectrical Transduction
563
Diochot, S., Baron, A., Rash, L. D., Deval, E., Escoubas, P., Scarzello, S., Salinas, M., and Lazdunski, M. (2004). A new sea anemone peptide, APETx2, inhibits ASIC3, a major acid‐ sensitive channel in sensory neurons. EMBO J. 23, 1516–1525. Ditting, T., Linz, P., Hilgers, K. F., Jung, O., Geiger, H., and Veelken, R. (2003). Putative role of epithelial sodium channels (ENaC) in the aVerent limb of cardio renal reflexes in rats. Basic Res. Cardiol. 98, 388–400. Drew, L. J., Wood, J. N., and Cesare, P. (2002). Distinct mechanosensitive properties of capsaicin‐sensitive and ‐insensitive sensory neurons. J. Neurosci. 22, RC228. Drew, L. J., Rohrer, D. K., Price, M. P., Blaver, K. E., Cockayne, D. A., Cesare, P., and Wood, J. N. (2004). Acid‐sensing ion channels ASIC2 and ASIC3 do not contribute to mechanically activated currents in mammalian sensory neurones. J. Physiol. 556(3), 691–710. Drummond, H. A., Price, M. P., Welsh, M. J., and Abboud, F. M. (1998). A molecular component of the arterial baroreceptor mechanotransducer. Neuron 21, 1435–1441. Drummond, H. A., Welsh, M. J., and Abboud, F. M. (2001). ENaC subunits are molecular components of the arterial baroreceptor complex. Ann. NY Acad. Sci. 940, 42–47. Drummond, H. A., Gebremedhin, D., and Harder, D. R. (2004). Degenerin/epithelial Naþ channel proteins: Components of a vascular mechanosensor. Hypertension 44, 643–648. Dube, G. R., Lehto, S. G., Breese, N. M., Baker, S. J., Wang, X., Matulenko, M. A., Honore, P., Stewart, A. O., Moreland, R. B., and Brioni, J. D. (2005). Electrophysiological and in vivo characterization of A‐317567, a novel blocker of acid sensing ion channels. Pain 117, 88–96. Duggan, A., Garcia‐Anoveros, J., and Corey, D. P. (2002). The PDZ domain protein PICK1 and the sodium channel BNaC1 interact and localize at mechanosensory terminals of dorsal root ganglion neurons and dendrites of central neurons. J. Biol. Chem. 277, 5203–5208. Escoubas, P., De Weille, J. R., Lecoq, A., Diochot, S., Waldmann, R., Champigny, G., Moinier, D., Menez, A., and Lazdunski, M. (2000). Isolation of a tarantula toxin specific for a class of proton‐gated Naþ channels. J. Biol. Chem. 275, 25116–25121. Fowler, J. C., Wonderlin, W. F., and Weinreich, D. (1985). Prostaglandins block a Ca2þ‐ dependent slow spike after hyperpolarization independent of eVects on Ca2þ influx in visceral aVerent neurons. Brain Res. 345, 345–349. Franchini, K. G., Moreira, E. D., Ida, F., and Krieger, E. M. (1996). Alterations in the cardiovascular control by the chemoreflex and the baroreflex in old rats. Am. J. Physiol. 270, R310–R313. Fricke, B., Lints, R., Stewart, G., Drummond, H., Dodt, G., Driscoll, M., and von During, M. (2000). Epithelial Naþ channels and stomatin are expressed in rat trigeminal mechanosensory neurons. Cell Tissue Res. 299, 327–334. Garcia‐Anoveros, J., Samad, T. A., Zuvela‐Jelaska, L., Woolf, C. J., and Corey, D. P. (2001). Transport and localization of the DEG/ENaC ion channel BNaC1a to peripheral mechanosensory terminals of dorsal root ganglia neurons. J. Neurosci. 21, 2678–2686. Glazebrook, P. A., Schilling, W. P., and Kunze, D. L. (2005). TRPC channels as signal transducers. Pflugers Arch. 451, 125–130. Goodman, M. B., and Schwarz, E. M. (2003). Transducing touch in Caenorhabditis elegans. Annu. Rev. Physiol. 65, 429–452. Goodman, M. B., Ernstrom, G. G., Chelur, D. S., O’Hagan, R., Yao, C. A., and Chalfie, M. (2002). MEC‐2 regulates C. elegans DEG/ENaC channels needed for mechanosensation. Nature 415, 1039–1042. Gotoh, H., and Takahashi, A. (1999). Mechanical stimuli induce intracellular calcium response in a subpopulation of cultured rat sensory neurons. Neuroscience 92, 1323–1329. Grigg, P. (1986). Biophysical studies of mechanoreceptors. J. Appl. Physiol. 60, 1107–1115.
564
Chapleau et al.
Hajduczok, G., Chapleau, M. W., Ferlic, R. J., Mao, H. Z., and Abboud, F. M. (1994). Gadolinium inhibits mechano‐electrical transduction in rabbit carotid baroreceptors: Implication of stretch‐activated channels. J. Clin. Invest. 94, 2392–2396. Hamill, O. P., and Martinac, B. (2001). Molecular basis of mechanotransduction in living cells. Physiol. Rev. 81, 685–740. Hamill, O. P., Lane, J. W., and McBride, D. W., Jr. (1992). Amiloride: A molecular probe for mechanosensitive channels. Trends Pharmacol. Sci. 13, 373–376. Harper, A. A. (1991). Similarities between some properties of the soma and sensory receptors of primary aVerent neurones. Exp. Physiol. 76, 369–377. Helliwell, R. J. A., McLatchie, L. M., Clarke, M., Winter, J., Bevan, S., and McIntyre, P. (1998). Capsaicin sensitivity is associated with the expression of the vanilloid (capsaicin) receptor (VR1) mRNA in adult rat sensory ganglia. Neurosci. Lett. 250, 177–180. Hesselager, M., Timmermann, D. B., and Ahring, P. K. (2004). pH dependency and desensitization kinetics of heterologously expressed combinations of acid‐sensing ion channel subunits. J. Biol. Chem. 279, 11006–11015. Jin, Y.‐H., Bailey, T. W., Li, B.‐Y., Schild, J. H., and Andresen, M. C. (2004). Purinergic and vanilloid receptor activation releases glutamate from separate cranial aVerent terminals in nucleus tractus solitarius. J. Neurosci. 24, 4709–4717. Jones, R. C. W., III, Xu, L., and Gebhart, G. F. (2005). The mechanosensitivity of mouse colon aVerent fibers and their sensitization by inflammatory mediators require transient receptor potential vanilloid 1 and acid‐sensing ion channel 3. J. Neurosci. 25, 10981–10989. Katz, B. (1950). Depolarization of sensory terminals and the initiation of impulses in the muscle spindle. J. Physiol.(Lond.) 111, 261–282. Kaye, D. M., and Esler, M. D. (1995). Abnormalities of the autonomic nervous system in heart failure. Cardiol. Rev. 3, 184–195. Kellenberger, S., and Schild, L. (2002). Epithelial sodium channel/degenerin family of ion channels: A variety of functions for a shared structure. Physiol. Rev. 82, 735–767. Kirchheim, H. R. (1976). Systemic arterial baroreceptor reflexes. Physiol. Rev. 56, 100–176. Kraske, S., Cunningham, J. T., Hajduczok, G., Chapleau, M. W., Abboud, F. M., and Wachtel, R. E. (1998). Mechanosensitive ion channels in putative aortic baroreceptor neurons. Am. J. Physiol. 275(44), H1497–H1501. Krishtal, O. (2003). The ASICs: Signaling molecules? Modulators? Trends Neurosci. 26, 477–483. LaRovere, M., Bigger, J. T., Jr., Marcus, F. I., Mortara, A., and Schwartz, P. J. (1998). Baroreflex sensitivity and heart‐rate variability in prediction of total mortality after myocardial infarction. Lancet 351, 478–484. Li, Z., Lee, H. C., Bielefeldt, K., Chapleau, M. W., and Abboud, F. M. (1997). The prostacyclin analogue carbacyclin inhibits Ca2þ‐activated Kþ current in aortic baroreceptor neurones of rats. J. Physiol. (Lond.) 501(Pt. 2), 275–287. Li, Z., Chapleau, M. W., Bates, J. N., Bielefeldt, K., Lee, H.‐C., and Abboud, F. M. (1998). Nitric oxide as an autocrine regulator of sodium currents in baroreceptor neurons. Neuron 20, 1039–1049. Liedtke, W., Tobin, D. M., Bargmann, C. I., and Friedman, J. M. (2003). Mammalian TRPV4 (VR‐OAC) directs behavioral responses to osmotic and mechanical stimuli in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 100, 14531–14536. Lin, S.‐Y., and Corey, D. P. (2005). TRP channels in mechanosensation. Curr. Opin. Neurobiol. 15, 350–357. Lingueglia, E., de Weille, J. R., Bassilana, F., Heurteaux, C., Sakai, H., Waldmann, R., and Lazdunski, M. (1997). A modulatory subunit of acid sensing ion channels in brain and dorsal root ganglion cells. J. Biol. Chem. 272, 29778–29783.
19. Baroreceptor Mechanoelectrical Transduction
565
Linz, P., and Veelken, R. (2002). Serotonin 5‐HT3 receptors on mechanosensitive neurons with cardiac aVerents. Am. J. Physiol. Heart Circ. Physiol. 282, H1828–H1835. Lu, Y., Whiteis, C. A., Benson, C. J., Chapleau, M. W., and Abboud, F. M. (2006). Expression and localization of acid‐sensing ion channels in mouse nodose ganglia (abstract). FASEB J. 20(4) Pt. 1, A775. Ma, X., Price, M. P., Drummond, H. A., Welsh, M. J., Chapleau, M. W., and Abboud, F. M. (2001). The DEG/ENaC ion channel family member BNC1 mediates mechanical transduction of arterial baroreceptor nerve activity in vivo (abstract). FASEB J. 15(5) Pt. 2, A1146. Ma, X., Abboud, F. M., and Chapleau, M. W. (2002). Analysis of aVerent, central, and eVerent components of the baroreceptor reflex in mice. Am. J. Physiol. Regul. Integr. Comp. Physiol. 283, R1033–R1040. Mancia, G., and Parati, G. (2003). The role of blood pressure variability in end‐organ damage. J. Hypertens. 21, S17–S23. Mannsfeldt, A. G., Carroll, P., Stucky, C. L., and Lewin, G. R. (1999). Stomatin, a MEC‐2 like protein, is expressed by mammalian sensory neurons. Mol. Cell. Neurosci. 13, 391–404. Maroto, R., Raso, A., Wood, T. G., Kurosky, A., Martinac, B., and Hamill, O. P. (2005). TRPC1 forms the stretch‐activated cation channel in vertebrate cells. Nat. Cell Biol. 7, 179–185. Matsuura, S. (1973). Depolarization of sensory nerve endings and impulse initiation in common carotid baroreceptors. J. Physiol. 235, 31–56. McCarter, G. C., Reichling, D. B., and Levine, J. D. (1999). Mechanical transduction by rat dorsal root ganglion neurons in vitro. Neurosci. Lett. 273, 179–182. Meltzer, R. H., Kapoor, N., Anderson, S. J., Fuller, C. M., and Benos, D. J. (2006). Heteromultimerization of ASIC and ENaC channels determined by FRET microscopy (abstract). FASEB J. 20(4) Pt. 1, A795. Mendelowitz, D., and Kunze, D. L. (1992). Characterization of calcium currents in aortic baroreceptor neurons. J. Neurophysiol. 68, 509–517. Muraki, K., Iwata, Y., Katanosaka, Y., Ito, T., Ohya, S., Shigekawa, M., and Imaizumi, Y. (2003). TRPV2 is a component of osmotically sensitive cation channels in murine aortic myocytes. Circ. Res. 93, 829–838. O’Hagan, R., Chalfie, M., and Goodman, M. B. (2005). The MEC‐4 DEG/ENaC channel of Caenorhabditis elegans touch receptor neurons transduces mechanical signals. Nat. Neurosci. 8, 43–50. O’Neil, R. G., and Heller, S. (2005). The mechanosensitive nature of TRPV channels. Pflugers Arch. 451, 193–203. Page, A. J., Brierley, S. M., Martin, C. M., Price, M. P., Symonds, E., Butler, R., Wemmie, J. A., and Blackshaw, L. A. (2005). DiVerent contributions of ASIC channels 1a, 2, and 3 in gastrointestinal mechanosensory function. Gut 54, 1408–1415. Podrid, P. J., Fuchs, T., and Candinas, R. (1990). Role of the sympathetic nervous system in the genesis of ventricular arrhythmia. Circulation 82(Suppl. 1), I103–I113. Ponikowski, P., Chua, T. P., Piepoli, M., Ondusova, D., Webb‐Peploe, K., Harrington, D., Anker, S. D., Volterrani, M., Colombo, R., Mazzuero, G., Giordano, A., and Coats, A. J. (1997). Augmented peripheral chemosensitivity as a potential input to baroreflex impairment and autonomic imbalance in chronic heart failure. Circulation 96, 2586–2594. Price, M. P., Snyder, P. M., and Welsh, M. J. (1996). Cloning and expression of a novel human brain Naþ channel. J. Biol. Chem. 271, 7879–7882. Price, M. P., Lewin, G. R., McIlwrath, S. L., Cheng, C., Xie, J., Heppenstall, P. A., Stucky, C. L., Mannsfeldt, A. G., Brennan, T. J., Drummond, H. A., Qiao, J., Benson, C. J., et al.
566
Chapleau et al.
(2000). The mammalian sodium channel BNC1 is required for normal touch sensation. Nature 407, 1007–1011. Price, M. P., McIlwrath, S. L., Xie, J., Cheng, C., Qiao, J., Tarr, D. E., Sluka, K. A., Brennan, T. J., Lewin, G. R., and Welsh, M. J. (2001). The DRASIC cation channel contributes to the detection of cutaneous touch and acid stimuli in mice. Neuron 32, 1071–1083. Price, M. P., Thompson, R. J., Eshcol, J. O., Wemmie, J. A., and Benson, C. J. (2004). Stomatin modulates gating of acid‐sensing ion channels. J. Biol. Chem. 279, 53886–53891. Raybould, H. E., Gschossman, J. M., Ennes, H., Lembo, T., and Mayer, E. A. (1999). Involvement of stretch‐sensitive calcium flux in mechanical transduction in visceral aVerents. J. Auton. Nerv. Syst. 75, 1–6. Rong, W., Hillsley, K., Davis, J. B., Hicks, G., Winchester, W. J., and Grundy, D. (2004). Jejunal aVerent nerve sensitivity in wild‐type and TRPV1 knockout mice. J. Physiol. 560 (Pt. 3), 867–881. Roza, C., Puel, J.‐L., Kress, M., Baron, A., Diochot, S., Lazdunski, M., and Waldmann, R. (2004). Knockout of the ASIC2 channel in mice does not impair cutaneous mechanosensation, visceral mechanonociception and hearing. J. Physiol. 558(Pt. 2), 659–669. Sabharwal, R., Chapleau, M. W., Price, M. P., Welsh, M. J., and Abboud, F. M. (2005a). Subunits of acid‐sensing ion channels (ASICs) diVerentially mediate baroreceptor and chemoreceptor reflexes (abstract). FASEB J. 19(5) Pt. 2, A1617–A1618. Sabharwal, R., Chapleau, M. W., Price, M. P., Welsh, M. J., and Abboud, F. M. (2005b). Molecular mechanisms of baro‐ and chemoreceptor activation: Evidence that ASIC1 and ASIC3 contribute to chemoreceptor activation (abstract). Hypertension 46(4), 832. Sabharwal, R., Stauss, H. M., Lazartigues, E., Whiteis, C. A., Davisson, R. L., Price, M. P., Welsh, M. J., Abboud, F. M., and Chapleau, M. W. (2006). Abnormalities in baroreflex sensitivity and autonomic control in conscious ASIC2 / mice (abstract). FASEB J. 20(5) Pt. 2, A1186. Schaefer, M. (2005). Homo‐ and heteromeric assembly of TRP channel subunits. Pflugers Arch. 451, 35–42. Schild, J. H., and Kunze, D. L. (1997). Experimental and modeling study of Naþ current heterogeneity in rat nodose neurons and its impact on neuronal discharge. J. Neurophysiol. 78, 3198–3209. Schild, J. H., Alfrey, K. D., and Li, B. Y. (2005). Voltage‐gated ion channels in vagal aVerent neurons. In ‘‘Advances in Vagal AVerent Neurobiology’’ (B. J. Undem and D. Weinreich, eds.), pp. 77–99. CRC Press, Boca Raton, FL. Sharma, R. V., Chapleau, M. W., Hajduczok, G., Wachtel, R. E., Waite, L. J., Bhalla, R. C., and Abboud, F. M. (1995). Mechanical stimulation increases intracellular calcium concentration in nodose sensory neurons. Neuroscience 66, 433–441. Snitsarev, V., Whiteis, C. A., Abboud, F. M., and Chapleau, M. W. (2002). Mechanosensory transduction of vagal and baroreceptor aVerents revealed by study of isolated nodose neurons in culture. Autonomic Neurosci. Basic Clin. 98, 59–63. Somers, V. K., Mark, A. L., and Abboud, F. M. (1988). Potentiation of sympathetic nerve responses to hypoxia in borderline hypertensive subjects. Hypertension 11, 608–612. Stansfeld, C. E., Marsh, S. J., Halliwell, J. V., and Brown, D. A. (1986). 4‐Aminopyridine‐ and dendrotoxin‐induced repetitive firing in rat visceral sensory neurons by blocking a slowly inactivating outward current. Neurosci. Lett. 64, 299–304. Sullivan, M. J., Sharma, R. V., Wachtel, R. E., Chapleau, M. W., Waite, L. J., Bhalla, R. C., and Abboud, F. M. (1997). Non‐voltage‐gated calcium influx through mechanosensitive ion channels in aortic baroreceptor neurons. Circ. Res. 80, 861–867.
19. Baroreceptor Mechanoelectrical Transduction
567
Sun, S.‐Y., Wang, W., Zucker, I. H., and Schultz, H. D. (1999). Enhanced peripheral chemoreflex in conscious rabbits with pacing‐induced heart failure. J. Appl. Physiol. 86, 1264–1272. Sun, W., Abboud, F. M., and Chapleau, M. W. (2000). Evaluation of baroreflex and chemoreflex by carotid artery occlusion in mice: A method for phenotypic analysis of deletion of candidate sensory molecules (abstract). Circulation 102, II700. Sun, W., Abboud, F. M., and Chapleau, M. W. (2001a). Altered baro‐ and chemoreflex sensitivity revealed by carotid artery occlusion reflex in apo‐E knockout mice (abstract). FASEB J. 15(5) Pt. 2, A1146. Sun, W., Ma, X., Abboud, F. M., and Chapleau, M. W. (2001b). Impaired baroreflex and enhanced chemoreflex sensitivity in normotensive atherosclerotic mice (abstract). Clin. Auton. Res. 11(3), 187. Sun, W., Ma, X., Abboud, F. M., and Chapleau, M. W. (2002). DiVerential eVects of aging on baro‐ and chemoreflex regulation of arterial pressure: Selective impairment of baroreflex in senescent mice (abstract). Clin. Auton. Res. 12(4), 323. Suzuki, M., Mizuno, A., Kodaira, K., and Imai, M. (2003a). Impaired pressure sensation in mice lacking TRPV4. J. Biol. Chem. 278, 22664–22668. Suzuki, M., Watanabe, Y., Oyama, Y., Mizuno, A., Kusano, E., Hirao, A., and Ookawara, S. (2003b). Localization of mechanosensitive channel TRPV4 in mouse skin. Neurosci. Lett. 353, 189–192. Tavernarakis, N., and Driscoll, M. A. (2001). Degenerins: At the core of the Metazoan mechanotransducer? Ann. NY Acad. Sci. 940, 28–41. Trzebski, A., Tafil, M., Zoltowski, M., and Przybylski, J. (1982). Increased sensitivity of the arterial chemoreceptor drive in young men with mild hypertension. Cardiovasc. Res. 16, 163–172. Waldmann, R., Bassilana, F., de Weille, J. R., Champigny, G., Heurteaux, C., and Lazdunski, M. (1997). Molecular cloning of a non‐inactivating proton‐gated Naþ channel specific for sensory neurons. J. Biol. Chem. 272, 20975–20978. Welsh, M. J., Price, M. P., and Xie, J. (2002). Biochemical basis of touch perception: Mechanosensory function of degenerin/epithelial Naþ channels. J. Biol. Chem. 277, 2369–2372. Yang, X. C., and Sachs, F. (1989). Block of stretch‐activated ion channels in Xenopus oocytes by gadolinium and calcium ions. Science 243, 1068–1071.
This page intentionally left blank
Index A A. astacus, 23 ENaC channel, 245 Acetylcholine (ACh) receptors, 15–17 Acid sensing ion channels (ASICs), 12, 403, 542, 558–560 BR function impaired in ASIC / mice, 553–555 in BR neurons, 552–558 expressed in nodose neurons and BR terminals, 553 mammalian mechanosensation and, 448–450 protein, 128, 130, 136, 141 Actin‐binding proteins (ABPs) features, 273 interactions identification with polycystin‐2 (PC2) channel, 261–265 ‐actinin interaction, 261–263, 281 functional modulation by ‐actinin interaction, 265–266, 268 in vitro and in vivo binding with ‐actinin interaction, 263–264 yeast two‐hybrid system, 261–263 Actin cytoskeleton dynamics role in polycystin‐2 (PC2) channel functioning of, 253–255, 256–260 sensory role in polycystin‐2 (PC2) channel functioning of, 280–281 Actin filaments and their disruption, in mechanosensitive channels, 241–246 ‐Actinin and actin, eVect on isolated polycystin‐2 (PC2) channel, 268 ‐Actinin interaction eVect on, polycystin‐2 channel functioning, 265, 266, 268 ‐Actinin‐2–PC2N interaction, 262–263 Action potential discharge, encoding of depolarization into frequency of, 548
Active hair bundle motility in transduction channels, 415–417 adaptation, 415 spontaneous oscillations, 415–417 Adaptation in active hair bundle motility, 415 in hair cell mechanotransduction, 354, 356, 360–362 fast adaptation, 359–360 functional role of, 360–362 motor adaptation, 357–358 multiple components of, 358–359 transduction and, 356 A fibers, 432, 449 ADF neurons, 51, 70 ADL neurons, 70 ADPKD, 253–254 Alcohol, and voltage‐gated channels, 314–318 ALM neurons, 51–52 Amiloride and its derivatives hair cell transducer channel block by, 391–394 permeant blockers, 391–393 structure‐activity sequences for, 394
‐aminobutyric acid (GABA), 15–16 Aminoglycoside antibiotics eVects on hair cell transducer channel, 377 hair cell transducer channel permeation and block by, 382–391 through transducer channel, 382–384 Amoeba locomotion, Ca2þ dependence of, 492–494 Amoeboid migration, 487–488 Ancestral lipid stress detectors, voltage‐gated channels as, 304–305 ANKTM1 channel, 175, 406. See also TRPA‐1 channels Ankyrin repeat domain, in TRP channels, 413–414 569
Index
570 Aortic BR neurons, 546 gamma ENaC expression in, 547 Apx channels, 240, 245 Arachidonic acid (AA), 193, 215, 249 Arthropods cuticular mechanoreceptor neurons, 5, 8 mechanoreceptors, 1–2 ASH neurons, 51, 70–71 in C. elegans, 450 ASIC1, 449, 558, 560 ASIC2, 449, 553–554 ASIC3, 449, 553, 558–560 ASIC–/–mice, BR function impaired in, 553–555 ASICs. See Acid sensing ion channels (ASICs) ASICs neurons, 239 Astacus astacus. See Crayfish Astacus fluviatilis. See Crayfish Astrocytes, GsMTx4 peptide in, 100–102 ATP‐gated channels, 404 Atrial fibrillation, GsMTx4 peptide use in, 97–99 Auditory and vestibular function of TRPA1, 180 Auditory system, in hair cell mechanotransduction, 340–341 Autosomal dominant polycystic kidney disease (ADPKD), polycystin‐2 (PC2) channel and, 234 AVD neurons, 51 AVM neurons, 51–52, 59 AWA neurons, 70–71 AWC neurons, 70–71 Axon pathfinding calcium‐dependent, 113–116 requirement of TRP channels in, 114–116 encoding of guidance cues in, 112–114 ion channels as molecular integrators, 119–120 mechanosensitive ion channel role, 116–118 physical guidance cues, 116–118
B Band 7 protein, 132 Baroreceptors (BRs). See also BR neurons activation during increases in BP, 542, 544 mechanisms of, 544
complexity of mechanosensitive ion channel complex(es), 559 future directions, 560–561 heterogeneity of sensory neurons, 560 isolated neurons in culture, 546–548 methodological limitation and challenges, 558–560 need for selective pharmacological antagonists, 558 neurons in culture study, 546–548 reflex pathways, 542–544 sensitivity assessment, 545–546 sensory transduction, 544–548 encoding of depolarization into frequency of action potential discharge, 548 mechanoelectrical transduction, 545–548 vascular compliance and viscoelastic coupling, 545 BDNF, 116 Bilayer mechanics and voltage‐gated channels, 301–304, 314–318 Bilayer, voltage‐gated channels in, 299, 301 Bilayer mechanical reagents (BMRs), 298, 311, 316–318 modulation of VGCs by, 303 BKCa channels, 312, 314, 316, 321 BK channels, 513, 515, 521, 526–529 Bladder myocytes, mechanosensitive conductances in, 520–521 Bladder smooth muscle, mechanosensitive conductances in, 528–529 Blood pressure‐sensing BR neurons, mechanosensitive ion channels in, 521–561 BNaCs neurons, 239 Body motion and balance mechanosensory systems, 426 BR neurons, 546. See also Baroreceptors (BRs) gamma ENaC expression in, 547 mechanosensitive channels in, 549–558 acid sensing ion channels, 552–558 epithelial Naþ channels, 549–552 TRP channels, 555–558 mechanosensitive ion channels in, 541–561 Bullfrog’s sacculus hair bundle mechanical stimulus exerted with elastic glass fiber on, 409 scanning electron micrograph of, 401
Index
571 C
C. elegans, 129–130, 248, 250, 350, 403, 447–448, 450, 549, 556, 558 ASH neurons in, 450 DEG/ENaC ion channels, 55, 56, 58, 63–64 cysteine‐rich domains (CRDs), 56–57 DEL‐1, 62–64, 65 extracellular regulatory domain (ERD), 57 MEC‐4, 57–63 MEC‐10, 57–63 membrane‐spanning domains (MSDs), 56–57 neurotoxin‐related domain (NTD), 57 subunit structure and topology, 56 UNC‐8, 62–65 UNC‐105, 65–66 dopaminergic neurons in, 72 ion channels implicated in mechanosensation in, 54 mechanosensitive behaviors, 51–55 mechanosensitive ion channels in, 49–73 mechanosensor, 131, 143, 145 mechanotransducer model, 131–133 associated proteins, 133 mechanosensor core, 131–133 mechanotransducing complex in, 63 mechanotransduction, 131–133 touch receptor neurons, 52, 63 touch receptors in, 27 TRP ion channels, 11–12, 66–72 ORC‐2, 70–71 OSM‐9, 70–71 TRP‐4, 72–73 UNC‐105 muscle mechanotransducer, 133 Ca2þ channels, 26, 30–31, 41–44, 67, 72, 197, 238, 244, 247–249, 265, 275, 279, 353, 396, 456, 500–501, 512–517, 518–519, 526–527, 529 Ca2þ dependence of cell migration, 490–499. See also Cell migration; Migration amoeba locomotion, 492–494 identifying Ca2þ influx pathways, 491–492 measuring [Ca2þ]i, 490–491 role of [Ca2þ]i gradients and transients in mesenchymal cell migration, 495–499 vertebrate cell amoeboid migration, 494–495
Cadherins receptors, 236, 412, 418 Caenorhabditis elegans. See C. elegans [Ca2þ]i gradients in mesenchymal cell migration, 495–499 migrational directionality determination, 496–497 model for sustained [Ca2þ]i gradients, 495–496 [Ca2þ]i measurement, 490–491 Ca2þ influx pathways identification, 491–492 [Ca2þ]i transients in mesenchymal cell migration, 495–499 promote cell migration but inhibit neurite outgrowth, 498–499 Calcitonin gene‐related peptide (CGRP), 177 Calcium‐dependent axon pathfinding. See also Axon pathfinding requirement of TRP channels in, 114–116 Calcium signaling during transduction, by mechanoreceptors in spider, 14–15 Calmodulin, 402 Campaniform receptor in fly, 406 CaM protein, 249 Ca2þ‐permeant cation channels (MscCa). See MscCa channel Cardiac myocytes and atrial fibrillation, GsMTx4 peptide use in, 97–99 Cav channels, 298, 313, 321–324 Cav3.2 channels, 456 CcTRPN1 channel, 173 CD. See Cytochalasin D (CD) Cell migration, 490–499. See also Migration amoeba locomotion, 492–494 Ca2þ dependence of, 490–499 extrinsic mechanical forces acting on MscCa impact on, 501–502 identifying Ca2þ influx pathways, 491–492 measuring [Ca2þ]i, 490–491 MscCa role in, 499–501 role of [Ca2þ]i gradients and transients in mesenchymal cell migration, 495–499 vertebrate cell amoeboid migration, 494–495 Cell physiology, 235 Cellular sites, for GsMTx4 peptide TRPC‐1 channel, 95–96 TRPC‐6 channel, 97 CEP neuron, 51 C fibers, 432, 434–435 CFTR channels, 240, 246
Index
572 Chelicerata, 22 Chicken ‐actinin, 276 Chicken fimbrin, 276 CHO channels, 96 Choline, 402 Cholinergic innervation, 17 CLC channels, 240 Cl– channels, 240, 520–521, 526, 529 C‐mechanocold (C‐MC) aVerents, 434 C‐mechanoheat (C‐MH) aVerents, 434 C‐mechanoheatcold (C‐MHC) aVerents, 434 C‐MiHi fibers, 435 Collective cell migration, 489 Coronary smooth muscle cells, stretch‐activated nonselective cation currents in, 516 COS cells, mTREK‐1 current expressed in, 534–535 Crayfish abdomen and thorax of, 24 pharmacology of MSCs, 33–36 stretch receptor organ (SRO), 21–22 (See also Stretch receptor organ (SRO) of crayfish) adaptation, 41–43 functional properties, 24–43 general behavior, 24–26 impulse response generation, 36–41 ion channels for mechanotransduction in, 21–45 macroscopic receptor currents in stretch receptor neurons, 31–33 mechanosensory neurons of, 22 morphology of, 23–34 MSCs in receptor neurons, 27–31 pharmacology of MSCs, 33–36 transduction processes in, 25 viscoelastic properties of receptor muscles, 26–27 voltage‐gated ion channels, 36–41 Cultured sensory neurons, mechanoreceptive ion channels in, 437–446 Cupiennius salei, 3–4 Cuticular mechanoreceptors in spider, 4–5 in spiders, 4–5 Cyclopiazonic acid (CPA), 519 Cytochalasin B (CB), 241–244 Cytochalasin D (CD), 241–245, 528
eVect on mechanosensitive channels, 241–246 eVect on polycystin‐2 (PC2) channel, 255–260 Cytochrome P450, 193 Cytoskeletal connections, in TRP channels, 251–253 Cytoskeletal interactions with polycystin‐2 (PC2) channel, 273–274 actin cytoskeleton sensory role, 280–281 actin networks elastic properties, 275–280 mechanosensitivity and lipid bilayer, 272, 275 molecular link, 274–275 Cytoskeletal regulation and osmosensory control of polycystin‐2 (PC2) channel, 278 Cytoskeleton mechanosensitive channels connection, 239–241 interface, 272, 279–281 polycystin‐2 (PC2) channel interface, 272, 274–281
D DEG channels, 12, 55, 129–130, 171, 238–239, 403–404, 447–448, 450 DEG/ENaC ion channels in C. elegans, 55–56, 58, 63–64 cysteine‐rich domains (CRDs), 56–57 DEL‐1, 62–65 extracellular regulatory domain (ERD), 57 MEC‐4, 57–63 MEC‐10, 57–63 membrane‐spanning domains (MSDs), 56–57 neurotoxin‐related domain (NTD), 57 subunit structure and topology, 56 UNC‐8, 62–65 UNC‐105, 65–66 DEG/ENaC proteins, phylogenetic relations among, 58 Degenerin channel (DEG). See DEG channels Degenerin (DEG) genes, 549 DEL‐1 ion channel, 54, 62–65
Index
573
Depolarization encoding, into frequency of action potential discharge, 548 Dermis papillary layer, 428–429 reticular layer, 428–429 somatic mechanosensation, 429 Diacylglycerol (DAG), 156, 248–249, 519 Dictyostelium discoideum, 487, 501 Dihydropyridine receptors (DHPR), 215 Dihydrostreptomycin (DHS), 383–384, 402 blocking potency of, 390–391 permeation by, 392–393 DmTRPN1 channel, 173 Dopaminergic neurons, in C. elegans, 72 Dorsal root ganglion (DRG), 547–548 DRG neurons, 408 action potentials and mechanically activated currents of, 442 expression of ASICs in, 449 gating MS ion channels in, 439, 446–447, 457 MA currents exhibited by, 444 stretch activated ion channels in, 449 DRG synapse, mechanoreceptive neurons of, 436 Drosophila melanogaster, 50, 71–73, 96, 129, 131, 198, 239, 250, 403, 405–406, 447–448, 450–451, 556, 558 mechanosensitivity in, 35 TRP channels in, 11–12 TRP protein, 66 DrTRPAI channel, 173 DrTRPN1 channel, 173 Duchenne muscular dystrophy (DMD), 99 abnormal TRPC1/MscCa activity in, 205–206 DVA neurons, 51, 450 Dynamic hair bundle, in hair cell mechanotransduction, 361, 363–365 Dystrophic muscle, MS channels in, 467–482 Dystrophin protein, 468
E Enhanced GFP (eGFP)‐human TRPC1, 208–210 ENaC channel, 129, 134–135, 140–143 ENaC channel, 134–136, 140–141 ENaC channel, 129, 134, 140–143
ENaC channel, 135–136, 140–141 ENaC channel, 129 "ENaC channel, 129
ENaC channel, 129, 134, 140–143 ENaC channels, 129, 171, 238–240, 403–404, 447–448, 450
ENaC expression, in aortic BR neurons, 547 ENaC proteins ENaC and ENaC channel in absence of
ENaC, 140, 142 assessment of, 139 characteristics, 137–138 electrophysiological evidence, 140 expression of, 130 gene silencing of ENaC and ENaC, 142–143 importance of, 138–154 inhibition abolishes pressure‐mediated vasoconstriction, 138–140 mechanosensitive tissue and activity, 135–136 mechanotransduction and, 133–136 physiological importance of, 144–145 in pressure‐mediated myogenic constriction quantification, 138 role of, 144 subunit expression in VSMCs, 140 transduce vessel strain in vivo, 143–144 vascular smooth muscle cell (VSMC) mechanotransduction, 127–145 in vascular smooth muscle mechanotransduction, 137–145 ENAD protein, 247 Endogenous mechanotransducers and TRPA1, 181–184 Endothelin‐1 (ET‐1), 100–101 eVects of cyclic stretch on, 101 Epidermis basement membrane, 427–428 somatic mechanosensation, 427–428 stratum basale, 427–428 stratum corneum, 427–428 stratum granulosum, 427–428 stratum lucidum, 427–428 stratum spinosum, 427–428 Epithelial‐mesenchymal transition (EMT), 486 Epithelial Naþ channels (ENaCs), 12, 55, 192, 542, 558–560. See also ENaC proteins blockers attenuate BR responses in vivo and in vitro, 550–551
Index
574 Epithelial Naþ channels (ENaCs) (cont.) in BR neurons, 549–552 expressed in BR sensory terminals, 549 expressed in nodose neurons, 549 50 ,60 ‐Epoxyeicosatrienoic acid (50 ,60 –EET), 193, 215, 452, 457, 556 Escherichia coli, 30, 238 Extracellular blocker molecule, punch‐through of, 388–389 Eyring’s rate theory, 387
F Filamin/ABP, 280, 273 Filiform hairs mechanoreceptors, in spiders, 3 Flexor digitoum brevis (FDB) fibers, 471–473, 481 FLP neurons, 51, 70 Fly, campaniform receptor in, 406 FM1‐43, 376–377, 405 as permeant blocker of MS ion channels, 443 permeation and mechanoreceptor channels block by, 378–380, 382 as screen for functional transducer channel and mechanoreceptor, 382 through hair cell transducer channel, 378–381 through other mechanoreceptor, 381–382 structure of permeant blocker molecules, 379 Focal segmental glomerulosclerosis (FSGS), 214 Frey test, 434
G G. spatulata, 85, 121 GABAB receptor, 16 GABA receptors, 240 Gadolinium strangeness, in voltage‐gated channels, 329 Gating in MS channels, 471–473, 475–478 in DRG neurons, 446–447 modal gating in mdx muscle, 474–478 SA gating, 467–468, 471–477 voltage‐sensitive gating, 473
Gating spring biophysical concept of, 410–412 molecular representation of, 412–414, 418 theory and hair cell mechanotransduction, 344–347, 365 in transduction channels, 408–414 Gene and protein characteristics of TRPA1, 175–176 Gene silencing of ENaC and ENaC pressure‐mediated vasoconstriction in mouse, 142–143 Gentamicin, eVects of MscCa on, 117–118 G hairs. See Hair follicle receptor Gigaseal patch recording, 327–328 GI myocytes mechanosensitive conductances, 521–523, 525–526 expressed in visceral smooth muscle myocytes, 526 nonselective cation conductances, 521–523, 525 GIRK channels, G protein activation of, 157 GI smooth muscle, mechanosensitive conductances in, 530–535 Gliosis, GsMTx4 peptide use in, 100–102 Glutamate neurotransmitter, 160 Glutamate receptors, 15–16 Glycine receptors, 240 GON‐2 ion channel, 68 Gramicidin (gA), GsMTx4 peptide eVect on, 93, 94 Grammostola rosea, GsMTx4 peptide from, 85 Grammostola spatulata, 35 Green fluorescence protein (eGFP). See EGFP‐human TRPC1 GsMTx4, 500 enGsMTx4 synthesized with only D amino acids, 90 GsMTx4 peptide, 121 from adult astrocyte SAC currents, 89 cDNA of gene encoding, 86 cellular sites for, 95–97 TRPC‐1 channel, 95–96 TRPC‐6 channel, 97 dissociation rates of, 89 eVect on Gramicidin (gA) in lipid bilayers, 93–94 eVect on gramicidin gating, 94 from Grammostola rosea, 85 properties of, 85–94
Index
575
biochemical and structural, 85–88 biophysical and mechanistic, 89–94 solution structure of, 87 specifity of, 93–95 therapeutic uses, 97–98, 101–102 astrocytes and gliosis, 100–102 cardiac myocytes and atrial fibrillation, 97–99 muscular dystrophy, 99–100 neurite growth extension, 102 GTL–1 ion channel, 68
H Hair bundle architecture, 400–401 deflection toward the largest stereocilium, 400–401 structure in hair cell mechanotransduction, 341–342 tip link and gating spring, 400–401, 418 Hair cell mechanotransduction, 339–365. See also Transduction channels adaptation, 354, 356, 360–362 fast adaptation, 359–360 functional role of, 360–362 motor adaptation, 357–358 multiple components of, 358–359 transduction and, 356 auditory system, 340–341 channel activation, 347–349 location, 343–344 properties, 351–352 dynamic hair bundle, 361, 363–365 gating spring theory, 344–347, 365 hair bundle structure, 341–342 MET channel pore, 352–355 MET involving mechanically gated channels, 341–343, 349 models of, 399–418 single‐channel recordings of MET currents, 355–356 tethered channels, 349–351 Hair cell transducer channel aminoglycoside antibiotics eVects, 377 block by amiloride and its derivatives, 391–394 permeant blockers, 391–393
structure‐activity sequences for, 394 evidence for permeation of aminoglycoside antibiotics through, 382–384 FM1–43 through, 378–381 extracellular blocker molecule, 388–389 inferences about functional geometry of pore asymmetry in blocking potency of extracellulary and intracellularly applied DHS, 390–391 general model equations, 386–388 permeation rate, 389–390 punch‐through of extracellular blocker molecule, 388–389 ionic selectivity of, 377–378 permeant blockers of, 376–396 permeation and block by aminoglycoside antibiotics, 382–391 putative model of pore, 395 Hair cell transduction channel. See also Transduction channels pore properties, 402–403 Hair follicle receptor, 428–430 HCN channels, 298, 311, 324–325 HCN2 channels, 324–325 Health and disease, polycystin‐2 (PC2) channel role in, 253–255 Hearing mechanosensory systems, 426 Helix aspersa, 129 Heterologous TRPA1 channel properties, 182 HflK/C proteins, 61 Hidden Markov Model Mechanosensitivity (HMMM), 308, 310–311, 316–317 High‐threshold mechanoreceptors (HTM), 434–435 Hirudo medicinalis, 119 HPETE, 249 HST polycystin‐2 (PC2) channel activity in, 256 actin and associated proteins presence eVect on, 255–257 CD eVect on, 255–257 colocalization and regulation by ‐actinin, 266 filamin eVect on, 277 gelsolin and actin eVect on, 257–260 hydroosmotic pressure eVect on, 269 hydrostatic pressure eVect on, 271
Index
576 HT ion channels, 439 HTM, 451 HTRPC1. See Human TRPC1 (hTRPC1) Human dytrophin, 276 Human fimbrin, 276 Humans, mechanotransduction sense in, 2 Human T‐plastin, 276 Human TREK‐1 mRNA, 156 Human TRPC1 (hTRPC1), 203, 207, 209 MS current activity in, 207–208 Hydrostatic and osmotic pressure eVect on regulation of polycystin‐2 (PC2) channel, 265, 267, 269–270, 271–273 Hydrostatic inflation, 328 Hyperpolarization‐activated (HA) channel, 523–524
I IAV channel, 172 IAV (Inactive) protein, 71 ILI neuron, 51 Imperturbable K‐selective pore, in voltage‐gated channels, 312–314 Inactivation no afterpotential‐D. See INAD protein INAD protein, 249, 251 Indanyloxyacetic acid (IAA‐94), 521 Inhibitory ACh receptor, 16 Inhibitory cysteine knot (ICK), 86–87 Inhibitory GABA receptor, 16 Inhibitory glutamate receptor, 16 Inner ear, TRPA1 expression, 177 Inner hair cells (IHCs), 378 Inositol 1,4,5 trisphosphate (IP3), 248 Ins3P receptor, 249 Integrin receptors, 236 Intracellular acidosis, 157, 162 Intracellular recording from VS‐3 neurons, in spiders, 5 Invertebrates mechanosensation, 447–448 TRP ion channels in, 450–451 Ion channels, mechanosensitivity of, 234 Ionic selectivity, of transducer channel, 377–378 Ionotropic inhibitory receptors, 17 IRK potassium channels, 157
Isolated BR neurons in culture, study of, 546–548
K KCa channel, 31, 42 Kþ channels, 22, 26, 30–31, 38–44, 88, 119–120, 155, 195, 197, 238, 240, 244, 257, 259, 265, 275, 396, 501, 512–513, 515, 527, 529–530, 532–535, 542, 544, 548 analysis of ensemble average currents from, 40 interactions with ‐actinin, 240 PIP2 interactions, 158 in slowly adapting stretch receptor neurons, 39 KCNK2 channels, 532. See also TREK‐1 channels KCNK4 channels, 533 KCNK10 channels, 533 Keratinocytes, 130 Kinesthesia mechanosensory systems, 426 Kinocilium, 400 K2P2.1 channel. See TREK‐1 channels Kv3 channel, 312, 316–318, 321–323 Kv channels, 298, 319–323
L LET‐2 collagen, 66 Lipid bilayer, mechanosensitivity and, 272, 274 LOV‐1 ion channel, 68–69 Low‐threshold mechanoreceptors (LTM), 432–434, 440, 445–456 Lymnaea neurons, 244
M MACh receptor, 16 Macroscopic receptor currents, in stretch receptor neurons, 31–33 Mammalian mechanosensation, and ASICs, 448–450 Mammals, TRP ion channels in, 451–453 Mdx FDB fibers, absence of dystrophin in, 472–473
Index Mdx muscle modal gating in, 474–478 SI gating mode, 474–476, 481–482 stretch‐induced gating mode transitions, 477–478 MEC‐4, 403 MEC‐10, 403 MEC‐2 channel, 448, 549, 559–560 MEC‐4 channel, 448, 549 MEC‐6 channel, 448, 549, 559 MEC‐7 channel, 448 MEC‐10 channel, 448, 549 Mechanical coupling by mechanoreceptors, in spiders, 3–6 Mechanoelectrical transduction, in baroreceptors sensory transduction, 545–548 Mechanoelectric transduction (MET) process. See MET process Mechano‐gated K2P channel TREK‐1. See also TREK‐1 channels regulation by membrane phospholipids, 155–168 Mechano‐operated channels (MOCs), 191 Mechanoreceptive ion channels, in cultured sensory neurons, 437–440, 441–446 Mechanoreceptive nerve fibers high‐threshold mechanoreceptors, 434–435 low‐threshold mechanoreceptors, 432–434 physiology of, 432–435 Mechanoreceptor channels permeation and block by FM1‐43, 378–382 as screen for functional transducer channel and mechanoreceptor, 382 through hair cell transducer channel, 378, 379–381 through other mechanoreceptor, 381–382 structure of five permeant blocker molecules, 379 Mechanosensation ASICs and mammalian mechanosensation, 448–450 electrophysiological approaches to, 436–437, 439 Frey test, 434 gating MS ion channels in DRG neurons, 446–447 invertebrate mechanosensation, 447–448 mechanoreceptive ion channels in cultured sensory neurons, 437–440, 441–446
577 mechanoreceptive nerve fibers physiology of, 432–433, 435 mechanoreceptive potassium channels, 453–454 nonneuronal cells indirect signaling, 456–457 quantitative mechanical responses in animal models, 434 Randall‐Selitto test, 434 in rodents, 436–437 sensory neurons cells indirect signaling, 456–457 somatic mechanosensation, 427, 428–432 dermis, 429 epidermis, 427–428 skin and touch receptors, 429–432 touch receptors somatic, 428–432 TRP ion channels, 249–251, 450–453 voltage‐gated channels and, 454–456 calcium channels, 456 sodium channels, 454–455 Mechanosensing molecular devices, 235–237 Mechanosensitive Ca2þpermeable channel. See MscCa regulation Mechanosensitive channels, 236–239. See also Mechanosensitive ion channels; MS channels actin filaments and their disruption, 241–246 BR neurons, 549–551, 554–558 acid sensing ion channels, 552–558 epithelial Naþ channels, 549–552 TRP channels, 555–558 channel‐cytoskeleton connection, 239–241 interface, 272, 274–281 cytoskeletal connections in TRP channels, 251–253 eVect of cytochalasins, 241–246 in neurite outgrowth, 111–121 calcium‐dependent axon pathfinding, 113–116 encoding of guidance cues, 112–114 ion channels as molecular integrators, 119–120 physical guidance cues, 116, 118 requirement of TRP channels, 114–116 TRP channels mechanosensation and, 249–251 superfamily, 114–116, 247–249, 251–253
578 Mechanosensitive conductances inward currents generation, 514–526 bladder myocytes, 520–521 GI myocytes, 521–523, 525–526 vascular smooth muscle, 514–520 outward currents generation, 527–535 bladder smooth muscle, 528–529 GI smooth muscle, 530–535 uterine smooth muscle, 529–530 vascular muscles, 527–528 Mechanosensitive (gated) ion channels (MSCs). See MSCs Mechanosensitive ion channel inhibitor GsMTx4. See also GsMTx4 peptide cellular sites for, 95, 97 TRPC‐1 channel, 95–96 TRPC‐6 channel, 97 mechanism of, 81–103 properties of, 85–94 biochemical and structural, 85–88 biophysical and mechanistic, 89–94 specifity of, 93, 95 therapeutic uses, 97–99, 101–102 astrocytes and gliosis, 100–102 cardiac myocytes and atrial fibrillation, 97–99 muscular dystrophy, 99–100 neurite growth extension, 102 Mechanosensitive ion channels, 21–22. See also Mechanosensitive channels; MS channels; MSCs baroreceptor neurons, 541–561 in Caenorhabditis elegans, 49–73 conductance, 9–11 density, 9–11 ionic selectivity of, 7–8 molecular characterization of, 11–13 pH sensitivity, 9–11 in slit sensilla, 6–13 of spiders, 1–17 temperature sensitivity, 11 VS‐3 location, 8–9 Mechanosensitivity, 235 lipid bilayer and, 272, 275 Mechanosensory organs, TRPA1 expression in, 176–177 Mechanosensory systems body motion and balance, 426 hearing, 426
Index kinesthesia, 426 touch, 426 Mechanotransduction, 235–236. See also ENaC proteins action potential encoding and, 13–14 in C. elegans, 131–133 channels of hair cells properties, 182–183 dynamic properties of, 13–14 ENaC proteins and, 129–130, 133–136 mechanically gated ENaC activity, 134–135 shear stress activation of ENaC, 134–135 stretch activation of ENaC, 134 vascular smooth muscle, 137–145 genetic link to, 130–131 ion channels in crayfish stretch receptor for, 21–45 model of C. elegans mechanotransducer, 131–133 proteins family involved in, 129–136 sense in humans and spiders, 2 sense in spiders, 2 in slit sensilla, 6–13 conductance, 9–11 density, 9–11 ionic selectivity, 7–8 molecular characterization, 11–13 pH sensitivity, 9–11 receptor current carried by sodium ions in VS‐3 neurons, 8 temperature sensitivity, 11 VS‐3 mechanosensitive channels location, 8–9 MechanoTRPs, 171–186 TRPA1 and, 171–186 MEC‐1 ion channel, 61–62, 133 MEC‐2 ion channel, 57, 60–61, 132–133, 141, 143 MEC‐4 ion channel, 54, 57–62, 63, 66, 70, 132–133, 140 in Xenopus, 61 MEC‐5 ion channel, 61–62, 66, 133 MEC‐6 ion channel, 61, 132, 141 MEC‐7 ion channel, 133 MEC‐9 ion channel, 61–62, 133 MEC‐10 ion channel, 54, 57–63, 66, 70, 132–133 MEC‐12 ion channel, 61, 133 MEC‐1 protein, 239 MEC‐4 protein, 239
Index MEC‐5 protein, 239 MEC‐7 protein, 239 MEC‐9 protein, 239 MEC‐10 protein, 239 MEC‐12 protein, 239 Meissner’s corpuscle, 428–433 Membrane blebbing process, 216 Membrane phospholipids regulation of mechano‐gated K2P channel TREK‐1 by, 155–168 TREK‐1 channel inhibition by, 161–168 stimulation by, 158–160 Merkel’s disk, 428, 430, 432–433 Mesenchymal cell migration, role of [Ca2þ]i gradients and transients in, 495–499 Mesenchymal migration, 488–489 MET channel involving mechanically gating in hair cell mechanotransduction, 341–343, 349 pore in hair cell mechanotransduction, 352–355 tethered to actin cytoskeleton, 357 tip‐link and possible locations for, 343 MET process, 340 Mg2þ channel, 67 Migration amoeboid migration, 487–488 cell migration, 490–499 collective cell migration, 489 mechanisms for switching migration modes, 489–490 mesenchymal migration, 488–489 modes of, 487–490 Mitogen‐activated protein kinase (MAPK), 102 Mutant mouse TRPA1 (mm TRPA1) channel, 173 Myelinated A fibers, 432 Modal gating in mdx muscle, 474–478 Molecular architecture and dynamics, of voltage‐gated channels, 300 Molecular identity of transduction channels, 403–408 candidate families, 404 TRP channel family, 404–408 Molecular representation of gating spring, 412–414 ankyrin repeat domain in TRP channels, 413–414
579 cadherins, 412 myosin lever arm, 414 Mouse DRG neurons, action potentials and mechanically activated currents of, 442 Mouse outer hair cells (OHCs), mechanoelectrical transducer currents in, 380, 383, 393 Mouse renal interlobar arteries inhibition pressure‐mediated vasoconstriction in, 142–143 MS Ca2þ channel. See MscCa channel MscCa channel, 111, 115, 117, 120–121, 192–193, 196, 207, 209, 215 abnormal activity in duchenne muscular dystrophy, 205–206 eVects of gentamicin on, 117–118 maitotoxin activation, 203–205 MscCa regulation. See also Cell migration role in cell migration, 499–501 of tumor cell migration, 485–502 MS channels abnormalities in, 191 aminoglycoside antibiotics block, 480–481 direct and indirect, 195–196 in dystrophic muscle, 467–482 expression during myogenesis, 469–470 extrinsic regulation stretch sensitivity, 196–197 fast turn‐on and turn‐oV, 194 gadolinium ion block, 478–479 gating, 471–478, 481–482 modal gating in mdx muscle, 474–478 SA gating, 471–473 voltage‐sensitive gating, 473 permeability to divalent cations, 470–471 to monovalent cations, 470 properties in skeletal muscle, 470–471 pharmacology, 478–481 aminoglycoside antibiotics block, 480–481 gadolinium ion block, 478–479 protein identification strategies, 197 SI gating mode in, 474–476, 481–482 stretch‐activated gating, 467 stretch‐induced gating mode transitions in, 477–478 TRPC1 reconstitution as, 207, 208–210
Index
580 MscK channel, 192–193, 195 MscL channel, 30, 43, 83, 193, 238, 350 MscNa channel, 192 MSCs, 81–83. See also GsMTx4 peptide in crayfish O. limosus neuron, 28 eVect of enGsMTx4 on, 91 in Escherichia coli, 30 permeability, 26, 30 pharmacology of crayfish, 33, 35–36 physiological function of, 84–85 from rat astrocytes eVect of enGsMTx4 on, 91 in receptor neurons, 27–30, 43–45 MscS channel, 43, 83, 193 MS ion channels in DRG neurons, 446–447 MS Kþ channel. See MscK channel MS modulation activity of voltage‐gated channels stretching force application to study, 327–329 gigaseal patch recording, 327–328 hydrostatic inflation, 328 inflation by oil injection, 329 osmotic swelling, 328 shear flow, 329 stretching native myocytes using two pipettes, 329 MS Naþ channel. See MscNa channel MS TRP channels and MS VGCs, 305–308 MS VGCs and MS TRP channels, 305–308 Muscarinic ACh receptors, 17 Muscular dystrophy, GsMTx4 peptide use in, 99–100 Myogenesis, MS channels expression during, 469–470 Myosin 1c, 402, 410 Myosin lever arm, 414
N Naþ channels, 22, 26, 30–31, 36–37, 41–44, 197, 239–240, 245, 279, 391, 394, 454–455, 501, 515, 519, 523, 542, 544–545, 548 distribution in slowly and rapidly adapting neurons, 37 NAN channel, 172 NAN (Nanchung) protein, 71 Nav channels, 298, 324
Nav1.4 channels, 325, 327, 454 irreversible eVects of stretch on, 326 Nav1.5 channels, 454 Nav1.7 channels, 454–455 Nav1.8 channels, 454–455 Nav1.9 channels, 454 Neither inactivation nor afterpotential‐C. See NINAC myosin III protein Nematode, proprioception in, 64 Neonatal mouse cardiac myocytes (NMCM), 244 Netrin, 116 Neurite growth extension, GsMTx4 peptide use in, 102 Neurite outgrowth mechanosensitive channels in, 111–121 calcium‐dependent axon pathfinding, 113–116 encoding of guidance cues, 112–114 ion channels as molecular integrators, 119–120 physical guidance cues, 116–118 requirement of TRP channels, 114–116 Neuronal cell bodies, mechanical stimulation of, 441 Neurotoxin‐related domain (NTD), 57 NINAC myosin III protein, 249 NMDA receptor channels, 240, 265, 353, 454 Nociception function of TRPA1, 177–180 No mechanoreceptor potential C (NompC) channel, 250 NompC channel, 250, 252, 405–406, 413, 450 structure representation of, 407 NompC mechanosensory ion channel, 72 Nonselective cation conductance, properties of, 517 Noxious mechanosensation, 426, 432–434 N‐type Cav channels irreversible eVects of stretch on, 326
O O. limosus neuron of crayfish mechanosensory channels observed in, 29 MSCs in, 28 OCR‐1 ion channel, 70 OCR‐2 ion channel, 54, 68–71, 450 OCR‐3 ion channel, 70
Index
581
OCR‐4 ion channel, 70 Octopamine application of, 17 receptor, 16 OHC, 341–342 freeze fracture image of, 342 freeze fracture of apical surface of, 347 OLQ neurons, 51 Oocyte, fluorescence images of, 209 Orconectes limosus, 23 OSM‐9 channels, 54, 68–69, 71, 172, 250, 450–451, 556 Osteoblasts, 130 Outer hair cells (OHCs), 378. See also OHC block of mechanoelectrical transducer currents in mouse, 380, 383
P Pacifastacus leniusculus, 23 Pacinian corpuscle, 428, 430–431 PC1. See Polycystin‐1( PC1) PC2C‐‐actinin interaction, 262–263 P120 channel, 175 PDEG‐1 protein, 239 PDE neurons, 51 PDK‐1 ion channel, 68–69 PDK‐2 ion channel, 68–69 PDZ protein, 249 Permeability properties of MS channels in skeletal muscle, 470–471 permeability to divalent cations, 470–471 permeability to monovalent cations, 470 Permeant transducer channel blockers amiloride and its derivatives as, 391–393 structure of, 379 Permeation and block of hair cell transducer channel by aminoglycoside antibiotics, 382–392 mechanoreceptor channels by FM1–43, 378–382 PHB neurons, 70 Phosphatidic acid (PA), 156 Phosphatidylethanolamine (PE), 156 Phosphatidylinositol‐4,5‐bisphosphate (PIP2), 248 Phosphatidylinositol (PI), 156 Phosphatidylserine (PS), 156
Phospholipase A2, 236 Phospholipase A2 (PLA2), 193, 195–196, 216 Phospholipase C, 236 PIP2. See Membrane phospholipids PKC, 443, 526 PKD1 channel, 172 PKD2 channel, 172 PKD1 gene, 253 PKD2 gene, 253–254 Placental trophoblasts, 130 PLM neurons, 51 Polycystic kidney disease (PKD), TRPC1 and, 205–206 Polycystin‐1( PC1), 253 Polycystin‐2 (PC2) channel. See also TRPP2 channel actin‐binding protein interactions identification with, 261–265 ‐actinin interaction, 261–263, 281 functional modulation by ‐actinin interaction, 265–266, 268 in vitro and in vivo binding with ‐actinin interaction, 263–264 yeast two‐hybrid system, 261–263 ‐actinin and actin eVect on isolated, 268 activity in hST, 256 actin and associated proteins presence eVect on, 255–257 CD eVect on, 255–257 colocalization and regulation by ‐actinin, 266 filamin eVect on, 277 gelsolin and actin eVect on, 257–260 hydroosmotic pressure eVect on, 269 hydrostatic pressure eVect on, 271 autosomal dominant polycystic kidney disease (ADPKD) and, 234 channel‐cytoskeleton interface, 272, 274–277, 279–281 cytochalasin D eVect on, 255–256, 258–260 cytoskeletal interactions with, 273–274 actin cytoskeleton sensory role, 280–281 actin networks elastic properties, 275–280 mechanosensitivity and lipid bilayer, 272–273 molecular link, 274–275
Index
582 Polycystin‐2 (PC2) channel (cont.) cytoskeletal regulation and osmosensory control of, 278 functioning of actin cytoskeletal dynamics role in, 280–260 actin cytoskeleton sensory role in, 280–281 ‐actinin interaction eVect, 265–266, 268 hydrostatic and osmotic pressure eVect on regulation, 265, 267, 269, 270–273 mechanosensitivity and lipid bilayer, 272–273 mediated channel function, 253–256, 258–260 osmosensory function hydrostatic and osmotic pressure eVect on regulation, 265, 267, 269, 270–273 role in health and disease, 253–255 Polyunsaturated fatty acids (PUFAs), 157 PPK1 channel, 448 Pressure‐mediated myogenic constriction ENaC proteins in, 137–145 ENaC and ENaC channel in absence of ENaC, 140, 142 assessment of, 139 characteristics, 137–138 electrophysiological evidence, 140 gene silencing of ENaC and ENaC, 142–143 importance of, 138–139, 141–155 inhibition abolishes pressure‐mediated vasoconstriction, 138–140 physiological importance of, 144–145 quantification, 138 role of, 144 subunit expression in VSMCs, 140 transduce vessel strain in vivo, 143–144 Pressure‐mediated vasoconstriction ENaC proteins in ENaC and ENaC channel in absence of ENaC, 140–142 electrophysiological evidence, 140 ENaC inhibition abolition of, 138–140 ENaC subunit expression in VSMCs, 140 gene silencing of ENaC and ENaC, 142–143 importance of, 138–139, 141–143, 145
physiological importance of, 144–145 transduce vessel strain in vivo, 143–144 in mouse renal interlobar arteries inhibition gene silencing of ENaC and ENaC, 142–143 Procambarus clarkii, 23 Prokaryotic VGCs, as ancestral lipid stress detectors, 304–305 Protein kinase C (PKC), 102 PVC neurons, 51 PVD neurons, 51 PVM neurons, 51–52 P2X(2) channels, 404 P2Y1 receptors, 457 P2Y2 receptors, 457
Q QLP neurons, 70
R Rabbit atrial cells, GsMTx4 aVect on action potential of, 94–95 Randall‐Selitto test, 434 Rat astrocytes, eVect of enGsMTx4 on MSCs from, 91 Receptor currents in adapting neurons, 32 eVect of lidocaine and bupivacain on, 34 Receptor muscles, viscoelastic properties of, 26–27 Receptor neurons, MSCs in, 27–30 Receptor‐operated (ROCs) channels, 192, 199–203, 210–212, 214 Rectifying SA (RSA) channel, 28–30 Reversible stretch‐induced changes in voltage‐gated channels, 319–325 Cav channels, 321–323 Cav L‐type channels in native preparations, 323–324 HCN channels, 324–325 Kv channels, 319–321 Kv3 channels, 321–323 Nav channels, 324 Reversible stretch‐induced gating changes, in voltage‐gated channels, 325–327
Index
583
Rho‐associated serine/threonine kinase (ROCK), 489–490 Rodents, mechanosensation in, 436–437 RuYni’s ending, 428, 430, 432–433 Ryanodine receptors (RyR1), 215
S SA channel, 28–31 SA MS channels, 468, 470 Selectins receptors, 236 Sensory neurons and nonneuronal cells signaling, 456–457 Sensory receptors, 235 Shaker 5aa, activation and inactivation transition rates in, 319–320 Shaker Kþ channels, 481 Shaker WTIR, MS responses of, 301–302, 306–312, 316–317, 321 Shaw2 channel. See Kv3 channel Single‐channel recordings of MET currents, in hair cell mechanotransduction, 355–356 Skin dermis, 429 epidermaldermal border, 428 epidermis, 427–428 free nerve ending, 428, 431–432 glabrous skin, 428 hair follicle receptor, 428–430 hairy skin, 428 and its receptors, 428 layered structure of, 428 Meissner’s corpuscle, 428–431 Merkel’s disk, 428, 430 Pacinian corpuscle, 428, 430–431 RuYni’s ending, 428, 430 specialized mechanoreceptors, 429–431 hair follicle receptor, 428–430 Meissner’s corpuscle, 428–431 Merkel’s disk, 428, 430 Pacinian corpuscle, 428, 430–431 RuYni’s ending, 428, 430 and touch receptors somatic mechanosensation, 428–432 Slit sensilla mechanoreceptors innervated by pairs of functionally diVerent neurons, 13 mechanical coupling by, 5–6
mechanotransduction in, 6–13 conductance, 9–11 density, 9–11 ionic selectivity, 7–8 molecular characterization, 11–13 pH sensitivity, 9–11 receptor current carried by sodium ions in VS‐3 neurons, 8 temperature sensitivity, 11 VS‐3 mechanosensitive channels location, 8–9 in spiders, 1–3, 5–6 Sodium channels. See Naþ channels Somatic mechanosensation, 427–432 dermis, 429 epidermis, 427–428 skin and touch receptors, 429–432 Somatosensory neurons of TRPA1, 176–177 B‐spectrin/ actinin protein, 468 Spider mechanoreceptors calcium signaling during transduction by, 14–15 cuticular mechanoreceptors, 4–5 filiform hairs, 3 intracellular recording from VS‐3 neurons, 5 mechanical coupling by, 3–6 slit sensilla, 1–3, 5–6 synaptic modulation of, 15–17 tactile hairs, 1, 3, 5 trichobothria, 1, 3, 5–6 types of, 3–5 mechanosensitive ion channels, 1–17 conductance, 9–11 density, 9–11 ionic selectivity of, 7–8 molecular characterization of, 11–13 pH sensitivity, 9–11 temperature sensitivity, 11 VS‐3 location, 8–9 mechanotransduction sense, 2 Spider venoms, sequence comparison of peptides derived from, 87 Spider VS‐3 neuron, arrangement of eVerent neurons and transmitter receptors on, 16 Spike initiating zone (SIZ), 548 Spontaneous oscillations in active hair bundle motility, 415–417 model, 416
Index
584 Src protein tyrosine kinase, 193, 196, 213 SRO. See Stretch receptor organ (SRO) of crayfish Stereocilia, localization and number of transduction channels in, 401–402 Stomatin protein, 132 Store‐operated (SOCs) channels, 192, 199–203, 205–206, 210–212, 215–216 Stretch‐activated channels (SACs), 83, 524 Stretch‐activated conductances. See also Mechanosensitive conductances in coronary smooth muscle cells, 516 in smooth muscles, 511–535 Stretch‐activated (SA) gating, 468 mode in MS channel, 474–476 in MS channels, 467–468, 471–476, 481–482 Stretch‐dependent Kþ (SDK) channels, activation of, 532 Stretch‐inactivated channel (SIC) activity, 83 Stretch‐induced (SI) gating mode transitions, in MS channel, 477–478, 481–482 Stretching force application, to study MS modulation activity of voltage‐gated channels, 327–329 Stretch receptor neurons macroscopic receptor currents in, 31–33 transduction processes in, 25 Stretch receptor organ (SRO) of crayfish, 21–22 functional properties, 24–43 adaptation, 41–43 general behavior, 24–26 impulse response generation, 36–41 macroscopic receptor currents in stretch receptor neurons, 31–33 MSCs in receptor neurons, 27–31 multifactor property, 41–43 pharmacology of MSCs, 33, 35–36 viscoelastic properties of receptor muscles, 26–27 voltage‐gated ion channels, 36–41 future research directions, 43–45 mechanosensory neurons of, 22 morphology of, 23–34 transduction processes in, 25
T TASK channels, 453 Tactile hairs mechanoreceptors mechanical coupling by, 6 in spiders, 1, 3, 5 Tarantula venom, therapeutic peptide derived from, 81–103 TASK channels, 532 TASK‐2 channels, 532 TASK K2P channels, 158 Tetraethylammonium (TEA), ion channel blocker, 377, 402 Therapeutic uses of GsMTx4 peptide, 97–98, 100–102 astrocytes and gliosis, 100–102 cardiac myocytes and atrial fibrillation, 97–99 muscular dystrophy, 99–100 neurite growth extension, 102 THIK channels, 532 Tip linkand gating spring, in hair bundle, 400, 401, 418 Touch mechanosensory systems, 426. See also Mechanosensation TRAAK channels, 195, 238, 244–245, 350, 530, 532–533 activation by PUFA, 157 Transducer channels. See also Hair cell transducer channel ionic selectivity of, 377–378 two‐barrier one‐binding site (2B1BS) model of, 384–386, 391, 394–395 Transduction channels activation, 409–410 active hair bundle motility, 415–417 adaptation, 415 spontaneous oscillations, 415–417 gating, 408–414 biophysical concept of, 410–412 molecular representation of, 412–414 kinetics and thermodynamics, 408–409 properties, 401–408 localization and number in stereocilia, 401–402 molecular identity, 403–408 pore properties, 402–403 Transduction in hair cell mechanotransduction and adaptation, 356
Index Transduction processes, in stretch receptor neuron, 25 Transient receptor potential (TRPC) channels, 11, 35, 128, 542. See also TRPC channels; TRP ion channels Traube’s rule, 315 TREK channels, 192, 238, 350, 532–535 activation by mechanical stress, 156 activation by PUFA, 157 TREK‐1 channels, 83, 93, 95, 238, 244, 447, 453–454, 530, 532–535 activation by intracellular acidosis, 162 PIP2 inhibition, 162–163 by PUFA, 157 conduction, 160 dual regulation by PIP2, 162, 167 gating model of, 159 inhibition by membrane phospholipids, 161–168 PIP2, 162–164, 166 wortmannin, 160 membrane phospholipids inhibition by, 161–168 regulation by, 155–168 specificity eVect on, 165 stimulation by, 158–160 PIP2 inhibition acidic intracellular pH and polyunsaturated fatty acid activation, 163 eVect of, 166 lowering by polylysine, 164 membrane stretch activation, 162 regulation by membrane phospholipids, 155–168 specificity of eVect of membrane phospholipids on, 165 stimulation by membrane phospholipids, 158–160 TREK‐2 channels, 530, 532 TRESK channels, 532 Trichobothria mechanoreceptors mechanical coupling by, 5–6 in spiders, 1, 3, 5–6 TrkA receptors, 443 TRP‐4, 450 TRPA (ankyrin), 198 TRPA channels, 173, 248, 252, 555
585 TRPA‐1 channels, 35, 95, 251–252, 407, 408, 431, 452, 556 auditory and vestibular function, 180 biological roles, 185–186 biophysical properties of heterologously expressed, 184 endogenous mechanotransducers and, 181–184 expression in mechanosensory organs, 176–177 function of, 177–178, 182–184 gene characteristics, 175–176 heterologous channel and endogenous mechanotransducers, 181–184 inner ear, 177 mechanoTRPs and, 171–186 nociception function, 177–180 protein characteristics, 175–176 somatosensory neurons, 176–177 TRPA ion channel, 68 TRPC channels, 68, 111, 114–118, 120–121, 173, 247–249, 482, 519 activation and function, 199 conformational coupling mechanism, 192, 215 expression, 198–199 interactions with scaVolding protein, 200–202 mechanosensitivity, 203–210, 215 mechanotransduction, 192 MS functions of, 191 pharmacology, 202 properties of, 197–202 single channel conductance, 202 TRPC1 channels, 198–211, 213, 215 TRPC2 channels, 198, 202, 209–211 TRPC3 channels, 198, 200–202, 206, 209, 211–212, 215 TRPC4 channels, 198, 200–203, 205–206, 209, 211–212 TRPC5 channels, 198, 200–202, 206, 209, 212–215 TRPC6 channels, 198–202, 205, 209, 211–215 TRPC7 channels, 198–200, 202, 205, 209 TRPC interactions, 199–200 TRPC1 channels, 83, 95–96, 115–119, 173–174, 249–250, 301, 453, 482, 500, 519, 556–557
586 TRPC1 channels (cont.) abnormal activity in duchenne muscular dystrophy, 205–206 cell swelling, 205 expressed in specialized mechanosensory nerve endings, 206–207 expression in, 203 involvement in wound closure and cell migration, 207 maitotoxin activation, 203–205 polycystic kidney disease, 205–206 reconstitution as MS channel, 207, 208–210 TRPC2 channels, 198, 202, 209–211, 249 TRPC3 channels, 115–116, 198, 200–202, 206, 209, 211–212, 215, 249, 556–557 TRPC4 channels, 198, 200–203, 205–206, 209, 211–212, 249, 482, 556–557 TRPC5 channels, 114, 116, 198, 200–202, 206, 209, 212–215, 249, 252–253, 556–557 TRPC6 channels, 97, 115, 198–202, 205, 209, 211–215, 252–253, 482 regulator of kidney slit diaphragm, 214–215 regulator of myogenic tone, 211–214 TRPC7 channels, 198–200, 202, 205, 209, 249 TRP‐4 channel, 173 TRP channels, 247–249, 404, 406–408, 418, 431, 447–448, 558–560. See also Transient receptor potential (TRPC) channels ankyrin repeat domain in, 413–414 in BR neurons, 555, 557–558 in Caenorhabditis elegans, 11 cytoskeletal connections in, 251–253 in Drosophila, 11–12 expressed in nodose neurons and BR terminals, 556–558 mechanosensation and, 249–251 and mechanosensitive channels, 247–249 nompC channel, 405–406 superfamily, 247–249 TRPA1, 407–408 TRP1 channels, as pain receptors, 11 TRP‐1 ion channel, 68 TRP‐2 ion channel, 68 TRP‐3 ion channel, 68 TRP‐4 ion channel, 54, 68, 69, 72 TRP ion channels in C. elegans, 55, 66–72 ORC‐2, 70–71
Index OSM‐9, 70–71 TRP‐4, 72–73 in candidates in mammals, 451–453 in invertebrates, 450–451 in mechanosensation, 450–453 phylogenetic relations among, 69 structure and topology of mechanosensitive in C. elegans, 67 TRPM channel, 198, 247–249 TRPM3 channel, 453 TRPM7 channel, 500 TRPM8 channel, 431 TRPM ion channel, 68 TRPML channel, 173, 198, 248, 555 TRPML3 channel, 173, 186, 453 TRPML ion channel, 70 TRPML1 ion channel, 70 TRPML2 ion channel, 70 TRPML3 ion channel, 70 TRPN channel, 35, 173, 248, 250–252, 555 TRPN1 channel, 174–175, 250, 450 TRPN1 mRNA, 175–177, 180, 186 TRPN (NompC), 198 TRPP channel, 172, 248, 253, 555 TRPP1 channel, 249, 453. See also Polycystin‐1( PC1) TRPP2 channel, 172, 249, 453. See also Polycystin‐2 (PC2) channel TRPP ion channel, 68, 71 TRPP (polycystin), 198 TRPs channels, implicated in mechanical sensitivity, 172–173 TRPV channel, 172, 247–249, 482 TRPV1 channel, 172, 174, 177, 179, 186, 197, 240, 249–250, 381, 431, 452, 455, 555–556 TRPV2 channel, 172, 249–251, 431, 452, 482, 520, 556 TRPV3 channel, 250, 431, 482 TRPV4 channel, 35, 172, 174, 193, 195, 206, 212, 215–216, 249–252, 431, 451–452, 457, 482, 556 TRPV5 channel, 248–250 TRPV6 channel, 248–250, 482 TRPV ion channel, 68, 198 TRPY channel, 172, 174 TRPY1 channel, 172 TRPY ion channel, 68, 198 ‐tubulin protein, 133 ‐tubulin protein, 133 Tubulins, 133, 239–240
Index
587
Tumor cell migration, MscCa regulation of, 485–502 TWIK channels, 532 Two‐barrier one‐binding site (2B1BS) model free energy profiles of, 385 transducer channel, 384–386, 391, 394–395 Tyrosine kinases, 236 Tyrosine phosphorylation, 193
U UNC‐8 channel, 448 UNC‐105 channel, 448 UNC‐1 ion channel, 65 UNC‐8 ion channel, 54, 62–65 UNC‐24 ion channel, 65 UNC‐105 ion channel, 54, 65–66, 133 UNC‐105 protein, 239 Uniramia, 22 Uroepithelia, 130 Uterine smooth muscle, mechanosensitive conductances in, 529–530
V Vanilloid transient receptor 4 channel. See TRPV4 channel Vascular compliance and viscoelastic coupling in baroreceptors sensory transduction, 545 Vascular muscles, mechanosensitive conductances in, 527–528 Vascular myocytes, mechanosensitive inward currents in, 519–520 Vascular smooth muscle, mechanosensitive conductances in, 514–520 mechanosensitive inward currents in vascular myocytes, 520–521 nonselective cation conductances, 514, 516–519 Vascular smooth muscle cell (VSMC) mechanotransduction, 128, 130, 136–138 ENaC proteins in, 127–145 ENaC and ENaC channel in absence of ENaC, 140–141, 142 assessment of, 139 characteristics, 137–138
electrophysiological evidence, 140 gene silencing of ENaC and ENaC, 142–143 importance of, 138–154 inhibition abolishes pressure‐mediated vasoconstriction, 138–140 physiological importance of, 144–145 quantification, 138 role of, 144 subunit expression in VSMCs, 140 transduce vessel strain in vivo, 143, 144 Vasoconstriction in mouse, gene silencing of ENaC and ENaC, 142–143 Ventral nerve cord neurons, 51, 62 Vertebrate cell amoeboid migration, Ca2þ dependence of, 494–495 Visceral smooth muscle myocytes, mechanosensitive conductances expressed in, 526 Viscoelastic properties, of receptor muscles, 26–27 VNO neurons, 211 Voltage‐dependent Ca2þchannels (VDCCs), 113, 115, 119 Voltage‐gated calcium channels, 456 Voltage‐gated channels (VGCs) alcohol and, 314–318 ancestral lipid stress detectors, 304–305 in bilayer, 299, 301 bilayer mechanics, 301–304, 314–318 binding sites, 314–318 Ca2þ channels, 44 Cav channels, 298, 321–324 eVect on kinetics of lipid composition, 305 osmotic stress, 305 pressure, 304–305 temperature, 305 gadolinium strangeness, 329 HCN channels, 298, 324–325 imperturbable K‐selective pore, 312–314 impulse response generation and, 36–41 Kþ channels, 38–41, 44 Kv channels, 298, 319–323 mechanosensation, 454–456 calcium channels, 456 sodium channels, 454–455 mechanosensitive (MS) transitions of, 298 mechanosensitivity of, 298–330
Index
588 Voltage‐gated channels (VGCs) (cont.) modulation by bilayer mechanical reagents, 303 molecular architecture and dynamics of, 300 Naþ channels, 36, 37, 44 prokaryotic, 304–305 reversible stretch‐induced changes in, 319–325 Cav channels, 321–323 Cav L‐ype channels in native preparations, 323–324 HCN channels, 324–325 Kv channels, 319–321 Kv3 channels, 321–323 Nav channels, 324 reversible stretch‐induced gating changes in, 325–327 stretching force application to study MS modulation activity of, 327–329 system components, 298–301 accessory proteins, 301 bilayer, 301, 399 channel proteins, 300 Voltage‐gated sodium channels, 454–455 Voltage‐sensitive gating, in MS channels, 473
VS‐3 mechanosensitive ion channels location intracellular recording from, 5 noise analysis and pH sensitivity of, 10 in spider, 8–9 VS‐3 neurons in spiders, intracellular recording from, 5
W Woodhull blockage model, 391
X Xenopus neurons, 102, 113–115, 121 Xenopus oocytes, 66, 129, 134, 136, 197, 202–203, 205, 239, 246, 249, 378, 405, 451, 453, 456–457, 500, 514, 556 MEC‐4 ion channel in, 61 XTRPC1, 203–205, 207, 209
Y Yeast two‐hybrid system, 261–263